AN ABSTRACT OF THE THESIS OF

Casey E. Rush for the degree of Master of Science in Food Science and Technology
presented on May 28, 2021

Title: Rework Practices in the Dairy Industry and Their Potential Contribution to Microbial
Spoilage

Abstract approved: ______________________________________________________

Joy G. Waite-Cusic

Rework is a common practice in the dairy industry as a way for processors to

minimize waste while recovering costs for dairy products that are unsaleable. The

regulations available for reworking fluid dairy products only outline product eligibility

requirements with a focus on repasteurization of the products containing rework as a way

to control the safety of the product by eradicating pathogenic bacteria. However, the

pasteurization of milk is not a sterilization process. The quality and shelf life of

pasteurized milk is often limited by the presence of psychrotrophic bacterial spores that

can survive and/or germinate following the HTST-pasteurization (72°C, 15 seconds)

process and subsequently grow in refrigeration temperatures ( 7°C). Paenibacillus

odorifer, a non-pathogenic psychrotrophic spore-former ubiquitous to the farm

environment, is the dominant bacterial species in pasteurized fluid milk at the end of

shelf life (21 days), often at elevated levels ( 6 log CFU/ml) that can cause noticeable

product defects. Because of the reported spoilage and sporulation potential of P. odorifer,

it is possible that this organism could cause premature spoilage events and degrade

quality in fluid milk products, regardless of the repasteurization step.

 The first objective of this study was to understand rework in the fluid dairy

processing industry including how processors define rework as well as their motivations,

handling practices, and decision-making processes. The second objective of this study

was to evaluate the spoilage potential of P. odorifer in reworked fluid and chocolate

flavored milk and how the growth behavior of this organism would contribute to the

microbial burden of finished product containing rework.

Seven commercial fluid milk processors from the Pacific Northwest were

surveyed regarding their rework handling practices. Processors reported nine typical

reworking motivations with reclaim and packaging problems being the two most

common. The age of the product allowed to be reworked ranged from 3-21 days after

initial pasteurization with typical dilution rates reported to be <20% rework commingled

with 80% fresh product. Some processors had performed in-house studies to assist in

making rework processing decisions while others followed the guidelines outlined

Pasteurized Milk Ordinance (PMO) (United States Public Health Service/Food and Drug,

2017). Quality defects were anecdotally attributed to products containing rework more

often than fresh product. Defects included organoleptic changes, processing issues, and

elevated microbial counts.

To evaluate the spoilage potential of P. odorifer, commercial UHT-pasteurized

fluid milk and chocolate milk were inoculated with four strains of P. odorifer previously

isolated from pasteurized fluid milk (n=2) and chocolate milk (n=2) to achieve an initial

cell density of 1-2 log CFU/ml. Samples were stored at 4°C and 7°C and microbial

populations and organoleptic observations (visual, aroma) were monitored throughout a

31-day shelf life. Maximum growth rates that occurred over the first 7-10 days at both
temperatures were similar between products for all strains (4°C = 0.39-0.53 log

CFU/ml/day; 7°C = 0.83-0.98 log CFU/ml/day). All strains demonstrated a slower

growth rate and lower cell density after 21 days of storage at 4°C in fluid milk (6.83-7.45

log CFU/ml) compared to chocolate milk (7.74-7.83 log CFU/ml).

Thermally-resistant subpopulations of P. odorifer were undetectable for the first

21 days of storage at 4°C in fluid milk and pasteurization conditions were confirmed to

be effective at inactivating P. odorifer vegetative cells. P. odorifer is capable of

developing thermally-resistant cell types including spores; however, it was confirmed

that these subpopulations do not develop in fluid milk unless P. odorifer reaches high cell

density which will not occur within 21 days of processing when milk is stored at 4°C.

These results indicate P. odorifer is unlikely to be the cause for negative quality

implications of fluid milk products containing rework. Collectively, these efforts have

characterized rework practices used in the dairy industry and have clarified that P.

odorifer is unlikely to contribute to increased quality deterioration in products containing

rework.
©Copyright by Casey E. Rush
May 28, 2021
All Rights Reserved
Rework Practices in the Dairy Industry and Their Potential Contribution to Microbial
Spoilage

by
Casey E. Rush

A THESIS

submitted to

Oregon State University

in partial fulfillment of
the requirements for the
degree of

Master of Science

Presented May 28, 2021

Commencement June 2021
Master of Science thesis of Casey E. Rush presented on May 28, 2021.

APPROVED:

Major Professor, representing Food Science and Technology

Head of the Department of Food Science and Technology

Dean of the Graduate School

I understand that my thesis will become part of the permanent collection of Oregon State
University libraries. My signature below authorizes release of my thesis to any reader
upon request.

Casey E. Rush, Author

 ACKNOWLEDGMENTS

This incredible journey would not have been possible without the endless love

and support from an entire army of people that helped celebrate the wins and provide

grace for the learns , that stood for me when I was weak, pushed me when I was

complacent, believed in me when my conviction waivered, and hugged me when I was

afraid.

First and foremost, I want to express my deepest thanks to my mother and father,

Susan and Gary Rush. Mom, you have always believed in me and encouraged me to go

after all of my crazy dreams. You taught me how to put my heart and soul into everything

that I do with love, passion, and integrity. Your vivacious spirit, late-night calls, and

dinner drop-offs were always more than just pick-me-ups to me; you are the most

extraordinary woman, and I am so grateful for you. Dad, you gave me my strong work

ethic and remind me every day to be ambitious and dedicated. You are the inspiration for

my desire to achieve something far greater than myself and you have always believed in

me. I also want to thank my personal cheerleading team: Shana, Brittany, and CJ. The

endless laughs, encouraging text messages, and shoulder for the fears have always given

me the strength see the sunshine even on the cloudiest of days.

I would like to thank my other half, Nathan. You have been there through all the

late nights, self-doubt, personal wins, and rivers of tears. You have always fought the

battle right alongside me, and always knew what I needed even when I didn t. I am

everlastingly grateful for you and your love.

 I would like to extend a sincere thanks to my advisor and mentor, Joy Waite-

Cusic. Joy, I could not have done this without you. You see in me what I am too blind

and afraid to see in myself, and your support gives me the strength to know that even if I

slip, I will never fall alone. Your devotion to wholeheartedly serve the people in your life

inspires me to be the best version of myself for those around me every day. You are truly

a gift to my life and one of the most unforgettable people I have ever met. Thanks will

never be quite adequate.

I would also like to express a special thanks to Lisbeth Goddik and Robin Frojen.

Lisbeth, you have always believed in me and your support and guidance throughout these

last five years has meant the world to me. You saw my potential and provided me a

platform to gain confidence all while modeling how to be a fearless and compassionate

leader. Robin, you were the first person to recognize my passion for dairy, and I will

always remember how you asked me, Who ARE you?! . Thank you for seeing me.

I would like to thank the rest of my committee members, Mike Penner and Tala

Navab-Daneshmand. You both have been instrumental in my education and success in

this project, and I am so grateful for your words of wisdom and encouragement.

I would also like to thank all of my fabulously wonderful friends and lab mates,

Sam Burroughs, Jared Johnson, Brandon Selover, Daria Van De Grift, Carson Davis,

Immanuel Rodriguez, Alejandro Torres, Brandon Riesgaard, Rebecca Bland, and Hussein

Mohamed. You have all played a colossal role in making these years some of the best I

have ever had, as well as the success of this project. I have learned so much from each of

you, and you are all extraordinary people. I am honored to have gotten to know you all,

and I am so excited to see you all flourish in your futures. A special shout-out to Sam
Burroughs and Daria Van De Grift Sam, I will never forget the best hug anyone has

ever given me. Daria, your strength and self-motivation inspire me every day, and I will

always cherish your friendship, resilience, and encouragement.

And finally, I like to express my most sincere gratitude to BUILD Dairy for

funding this project and my dreams, and Oregon State University for providing the

opportunity.
 CONTRIBUTION OF AUTHORS

Joy Waite-Cusic assisted with experimental design, data collection and interpretation,

writing, and editing. Lisbeth Meunier-Goddik assisted with experimental design and

editing. Michael Penner assisted with editing. Jared Johnson, Samantha Burroughs,

Alejandro Torres, Hussein Mohamed, and Brandon Riesgaard assisted in data collection.
 TABLE OF CONTENTS

Page
CHAPTER 1: INTRODUCTION... .. .. 1

CHAPTER 2: REWORK IN THE DAIRY INDUSTRY: AN INDUSTRY SURVEY . .....6

Abstract .. ... ..7

Introduction . . . . ... ..8

Materials and Methods. . ... ... ...11

Results . ... . ... 12

Discussion .... .. ... .30

Conclusion .... ............................ .36

Acknowledgements . . .. 36

CHAPTER 3: EVALUATING PAENIBACILLUS ODORIFER FOR ITS POTENTIAL TO

REDUCE SHELFLIFE IN REWORKED FLUID MILK PRODUCTS .. ................. 42

Abstract . .. .. . .. .43

Introduction . . . . . . . 44

Materials and Methods. . ... . . ....48

Results and Discussion ... .. .52

Conclusion .. .... .. .66

Acknowledgements .... 66

CHAPTER 4: CONCLUSION.. ...... .. 71

BIBLIOGRAPHY ... . .. . ... 73

 LIST OF FIGURES

Figure Page

1. Figure 2.1 Fluid milk products that can serve as the source or recipient of rework .. 24

2. Figure 2.2 Collection, storage and incorporation of milk processing streams that
are used as rework into fluid milk. .. ... . ... 25

3. Figure 3.1 Growth behavior of Paenibacillus odorifer strains (A: FSL R10-2726 (AT35),
B: FSL A6-0363 (AT40), C: FSL E2-0150 (AT2), D: JWC-2503 (AT-novel)) in
chocolate and fluid milk (2%) stored at 4 C and 7 C for 21 days. . ........53

4. Figure 3.2 Thermal inactivation of Paenibacillus odorifer cocktail grown in

pasteurized milk at 25 C for 24 hours prior to treatment. . . .. .....61

5. Figure 3.3 Growth and sporulation of Paenibacillus odorifer strains (A: FSL R10-2726
(AT35), B: FSL A6-0363 (AT40), C: FSL E2-0150 (AT2), D: JWC-2503 (AT-novel))
in fluid milk (2%) stored at 4 C for 31 days. .. ..........63
 LIST OF TABLES

Table Page

1. Table 2.1 An overview of the rework survey questions delivered to all processors ..13

2. Table 2.2 Pasteurization method, product types, and packaging type and size used
by dairy processing facilities (A-G) participating in the rework survey. 14

3. Table 2.3 Rework motivations reported by dairy processing facilities participating

in the rework survey. . .17

4. Table 2.4 Code-date ranges that surveyed dairy processors indicated as standards for
reworking fluid milk products. 18

5. Table 3.1 Quality parameters of fluid milk (2%) inoculated with Paenibacillus odorifer
strains and held at 4 C or 7 C for 21 days. .. ..57
 DEDICATION

Dedicated to the most extraordinary and fearless man I had the opportunity to know and

love. He inspires me every day to be just as resilient and courageous as he was. Fear not,

for I am with you . In loving memory of my brother, Kyle Rush.

 1

CHAPTER 1

INTRODUCTION

Our global population is growing and the food supply will need to support the projected

9-billion occupants in the year 2050 with one strategy concentrating on reducing food waste loss

(Parfitt et al., 2010; Dou et al., 2016). The United States (U.S.) Environmental Protection

Agency (EPA) estimated that in 2018, 103 million tons of food farm-to-fork was wasted with the

manufacturing/processing sector making up almost 39% of that food waste (United States

Environmental Protection Agency, 2020). The food industry has focused their efforts in

optimizing process efficiency to minimize this food waste loss and reduce the incurred economic

and environmental consequences. Each food industry subsidiary is unique in their processing

system including their product loss and waste generation. For some food processors, product loss

and waste generation are inevitable because 100% of the product does not end up in a saleable

finished product container. Examples of inefficiencies include resulting by-product, diluted

beverages discharged from the pipelines to the drain following the cleaning-in-place (CIP) rinse,

excess product from over-production or supply chain disruption, mislabeled products, or blended

flavors between product changeovers. While some food manufacturers have discovered

economically advantageous ways to transform their waste through by-product utilization (i.e.,

whey protein), other processors have had to find creative solutions to capture and utilize these

recovered products that would otherwise be disposed of in the waste stream. This general

practice of recovering and reprocessing product loss or waste into a saleable product is

genericall ermed re ork .

 2

Rework in the dairy industry is a common practice as a way for processors to recover

costs of ingredients while maximizing efficiency and minimizing waste. Reducing the amount of

prod c ha en ers a dair processor s as e s ream is especiall cri ical from an en ironmen al

aspect. Dairy processing is one of the highest contributors of industrial wastewater which impose

a large amount of organic and inorganic substances (solids, phosphorus, nitrogen, fats/oils, sugar,

chemical solutions, etc.) that can harm the environment if left untreated (Milani et al., 2011;

Kolev, 2017; Ahmad et al., 2019; Patra and Duary, 2020). Additionally, the cost of product loss

can quickly add up for a dairy processor. In a Hungarian milk processing facility, the

investigators found that the majority of milk production losses were from technology and

automation, the design of the plant, and overproduction estimating that approximately 1% of a

140,000-liter production run of milk is lost to the drain (Tóth et al., 2014). Furthermore, a case

study assessing a small to medium sized Canadian dairy processing facility valued the yearly

product loss of a chocolate beverage, skim milk, and homogenized milk from just the fillers to be

$163,800 calculated on the raw milk component cost (Aikenhead et al., 2015).

By recovering and reworking these products back into new and fresh products, the

processor could recoup some of their lost capital. However, this could jeopardize the quality of

the finished product by introducing physical, microbiological, and/or chemical hazards (ICMSF,

2005; Center for Food Safety and Applied Nutrition, 2007; Agriculture Marketing Service,

2013). The physical and chemical hazards are largely controlled at the plant level through

stringent operating technology such as metal detectors for metal fragment (physical hazard) and

documentation for traceability and allergen control (chemical hazard) while the microbiological

hazards remain the most critical to consider in any dairy system. Because most products are

reprocessed or repasteurized, there is a low risk for pathogenic concerns; thus, the
 3

microbiological impact is principally the quality implications caused by heat resistant spore

formers and enzyme producers. Quality implications of rework has been previously investigated

for butter (Long, 1939; Dolby, 1965; Chambers, 2002), ice cream (Holm et al., 2002; Cigerdelen,

2011), and processed cheese ( ern ko e al., 2018). To our knowledge, there are no previous

investigations into the quality implications for fluid milk processing.

Quality defects and premature spoilage in pasteurized fluid milk caused by

microorganisms is a frequent challenge in the dairy industry which contributes to the estimated

31% of dairy foods discarded at the retail and consumer level (Buzby et al., 2014; Martin et al.,

2021). There are two primary sources of spoilage organisms in fluid dairy products: low levels of

bacterial spores in raw milk that survive the pasteurization process, and post-pasteurization

contamination (PPC) (Clark e al., 2009; Trm i e al., 2015; Mar in et al., 2018). PPC is

commonly from Gram-negative bacteria present in the dairy processing environment and control

of PPC must be mitigated through sanitation programs (Gopal et al., 2015; Reichler et al., 2018).

Bacterial spores of Bacillus and Paenibacillus species (spp) are readily found in raw milk and

are collected during the milking process from the farm environment (Meer et al., 1991; Coorevits

et al., 2008; Huck et al., 2008; Ivy et al., 2012; Grady et al., 2016). They can survive and

germinate at typical high temperature short time (HTST) pasteurization temperatures (72 C, 15

seconds) and grow under refrigeration conditions (Meer et al., 1991; Chambers, 2002; Huck et

al., 2008; Ranieri et al., 2009; Ivy et al., 2012; Moreno Switt et al., 2014; Gopal et al., 2015;

Beno et al., 2020). Sporogenesis is an adaptive characteristic and is triggered when a vegetative

cell is placed in unfavorable environmental conditions. Sporulation often initiates when nutrients

are scarce, at high cell density, and/or unfavorable storage temperature, low water activity, or

extreme pH; however, when conditions are tolerable or favorable, the spore can then germinate
 4

back into a vegetative cell where replication can proceed (Gauvry et al., 2017). Bacillus spp. and

Paenibacillus spp. can readil gro in milk a refrigera ion empera res ( 7 C) and are repor ed

to be the principal psychrotrophic spore formers isolated from raw and pasteurized milks;

however, Paenibacillus spp. are the predominant genus found in pasteurized milk at the end of

shelf life (Huck et al., 2008; Ranieri and Boor, 2009; Ivy et al., 2012). The majority (59%) of

Paenibacillus species isolated from dairy products are Paenibacillus odorifer (Beno et al., 2020).

Furthermore, P. odorifer is the predominant microorganism found in pasteurized fluid milk

products and can grow to spoilage levels (>6 log CFU/ml) within the shelf life of HTST products

(Ivy et al., 2012; Moreno Switt et al., 2014; Ohkubo et al., 2019; Beno et al., 2020). For HTST

pasteurized products, this typically falls between 14 and 22 days with flavored products having

the shorter shelf life (Ranieri and Boor, 2009). If these products are contaminated with P.

odorifer at the time of packaging, bacterial populations could reach >6.0 log CFU/ml within this

timeframe (Douglas et al., 2000; Buehler et al., 2018; Beno et al., 2020). Discernable

organoleptic deviations have been reported to occur at cell densities around 6.0 log CFU/ml for

fluid milk products contaminated with Paenibacillus spp., including off-flavors (rancid, fruity)

and textural changes resulting from proteolytic and lipolytic activity (Fromm and Boor, 2004;

Clark et al., 2009; Ivy et al., 2012; Beno et al., 2020).

The Pasteurized Milk Ordinance states that processors can reprocess (rework) their

products up to the printed code date which, as previously mentioned, is typically within 14-22

days (United States Public Health Service/Food and Drug, 2017). Because P. odorifer can

produce heat-resistant spores that can survive pasteurization temperatures and then subsequently

grow at refrigeration temperatures to spoilage levels within the time frame that processors are
 5

allowed to rework this product, this organism could have a negative impact on the

microbiological quality of the new product.

The first objective of our study was to use interviews and plant visits to better understand

common rework practices and terminology in the dairy industry, specifically in fluid milk

processing plants (Chapter 2). We used information from the industry survey to consider how

these practices may influence product quality or shelf life of products containing reworked

ingredients. This informed our second objective to evaluate the spoilage potential of P. odorifer

in fluid and flavored milk with a perspective of rework handling and processing (Chapter 3).
 6

CHAPTER 2:

REWORK PRACTICES IN THE DAIRY INDUSTRY: AN INDUSTRY SURVEY

Casey E. Rush, Lisbeth Meunier-Goddik, and Joy G. Waite-Cusic*

Department of Food Science and Technology

Oregon State University

Corvallis, OR 97331, USA

*Author for correspondence: Joy Waite-Cusic

Email: Joy.Waite-Cusic@oregonstate.edu

Prepared for submission to Food Protection Trends

 7

ABSTRACT

Rework is a common practice in the dairy industry as a way for processors to minimize waste

while recovering costs for products that are unsaleable. However, the regulations available for

reworking fluid dairy products are focused on pasteurization as a way to control the quality of

the product, which is an unsterile process. Rework in the fluid milk industry and its implications

to product quality has not been previous investigated. The objective of our research was to

characterize current industry practices for reworking dairy products by surveying dairy

manufacturers with the goal of defining rework (and other terms) and identifying situations

that may contribute to an increased microbial spoilage risk. Seven commercial fluid milk

processors from the Pacific Northwest were interviewed in-person at the respective facility

(n=3), via telephone or internet video (n=2), or in-person away from the facility (n=2) regarding

their current rework handling practices. Processors reported nine typical reworking motivations

with reclaim and packaging problems being the two most common. Products were reworked up

to the code date (21 d) at dilution rates of 20% rework to 80% fresh product. Storage and

processing parameters were also reported. We were able to identify rework conditions that have

the potential to influence product quality or shelf life of products containing reworked

ingredients. This information can guide processors decisions as well as future research that will

help to minimize waste while maintaining quality throughout shelf-life.

 8

INTRODUCTION

Industrial-scale food processing has evolved over the last century to become increasingly

efficient to reduce processing time and labor costs while increasing yield and optimizing product

quality. Despite this emphasis on continuous improvement, product loss and waste generation

during processing are common consequences; thus, 100% of the raw ingredients end up in a

saleable finished product container. A number of factors contribute to processed product waste

such as a resulting by-product, diluted beverages discharged from the pipelines to the drain

following the cleaning-in-place (CIP) rinse, blended flavors between product changeovers,

excess unsold product, or products that don t meet legal standards (e.g., labeling or underfilled).

The United States (U.S.) Environmental Protection Agency (EPA) estimated that in 2018, 103

million tons (93 million metric tons) of food was wasted along the farm-to-fork continuum, with

the manufacturing/processing sector accounting for almost 39% of the waste (United States

Environmental Protection Agency, 2020). Each sector of the food industry is unique in their

processing system and will differ in their product loss and waste generation. While some food

manufacturers have discovered economically advantageous ways to transform their waste

through by-product utilization (i.e., whey protein from cheese processing to acid whey by-

product), other processors have implemented creative solutions to capture potential waste and

utilize these recovered products in their final products. This general practice of recovering and

reprocessing product loss or waste into a saleable product is generically termed rework.

The general strategy of rework in the food industry is common; however, processes and

procedures for handling and processing rework will be unique to each commodity as well to

specific processing facilities. Effective implementation of rework strategies provides a

competitive advantage by reducing waste, but they also present new challenges in traceability
 9

and may have potential impacts on finished product quality. While the traceability challenge may

be solved through documentation and technology, determining the implications of rework

practices on product quality require a processor to consider each step within their processing

system that may contribute to their finish product. Currently, the industry relies on previous

experiences and trial and error to make rework decisions; however, these processors often lack

guidance to ensure confidence that their rework practices do not have a negative impact to their

finished product quality which presents a significant risk to brand reputation.

While the production of dairy milk in the U.S. has increased by 14% over the past decade

(National Agriculture Statistics Service, 2021), fluid milk sales have been trending downward

(Economic Research Service, 2020). However, over the last year alone, the volume of milk sales

has increased by 1.9% while flavored milk has decreased by 0.8% (IRI Custom DMI Market

Advantage Database, 2021). With the increase in production and capricious consumer demands,

processors may also be generating more product loss and waste. The cost of product loss can

quickly add up for a dairy processor. In a Hungarian milk processing facility, the investigators

found that the majority of milk production losses were from technology and automation, the

design of the plant, and overproduction estimating that approximately 1% of a 140,000-liter

production run of milk is lost to the drain (Tóth et al., 2014). Furthermore, TetraPak, a major

equipment manufacturer for the dairy industry, reported that a typical dairy plant can lose 3% of

product to CIP washes alone in a 500,000-gallon (1.9 million liters) processing run (Tobe, 2016).

In a case study assessing a small to medium-sized Canadian dairy processing facility, the yearly

product loss associated with three products (chocolate beverage, skim milk, and homogenized

milk) from the fillers was calculated at $163,800 in raw milk costs alone (Aikenhead et al.,

2015). Reworked fluid milk and flavored milk products are sold at the same price and grade as
 10

non-reworked products. The average price of raw milk fluctuates, but reached a high

$24.07/CWT ($0.47/kg) in 2014 and was down to $18.30/CWT ($0.36/kg) in 2020 (National

Agriculture Statistics Service, 2021). Thus, it can be economically advantageous for a processor

to recover any product that can be reprocessed into a saleable commodity. Additionally, reducing

the amount of product that enters a dairy processor s waste stream is especially critical from an

environmental aspect. Dairy processing is one of the highest contributors of industrial

wastewater which carries a large amount of organic and inorganic substances (solids,

phosphorus, nitrogen, fats/oils, sugar, chemical solutions, etc.) that can harm the environment if

left untreated (Milani et al., 2011; Kolev, 2017; Ahmad et al., 2019; Patra and Duary, 2020).

Dairy processors also have the option of donating their product to a local food distribution center

or for use as animal feed (Dou et al., 2016) which is another solution to diverting product waste

loss from the waste stream if reworking the product is not feasible or economically

advantageous.

Rework in the dairy industry is a way for processors to recover the cost of ingredients

while maximizing efficiency. According to the Dairy Practices Council (DPC), rework can be

defined as, Clean, unadulterated food that has been removed from processing for reasons other

than insanitary conditions or that has been successfully reconditioned by reprocessing and that is

suitable for use as food. Fluid milk and milk products drained from processing equipment at the

end of the run, collected from a defoamer system and milk solids rinsed from equipment,

containers or pipelines, shall be repasteurized only if such milk and milk products are handled in

a sanitary manner and maintained at 7 C (45 F) or less (Dairy Practices Council, 2005). Other

terms used throughout literature and the industry to describe rework or types of rework include

reclaim (Dairy Practices Council, 2005; United States Public Health Service/Food and Drug,
 11

2017), reconditioned (Dairy Practices Council, 2005; United States Code of Federal

Regulations, 2021), and repasteurized (United States Public Health Service/Food and Drug,

2017).

Characterizing rework practices and processing decisions along with targeted product

testing, experimentation, and modeling could help to explain the potential risks of reworking

fluid dairy products. Current regulations for rework are based on reprocessing parameters in

which pasteurization time and storage temperature are the drivers (United States Public Health

Service/Food and Drug, 2017). However, pasteurization is not a sterilization step and the

microbiological quality of products and ingredients can still influence the finished product

quality and shelf life of pasteurized fluid milk (Douglas et al., 2000; Martin et al., 2021). Our

overall objective was to use interviews and plant visits to better understand common rework

practices and terminology in the dairy industry, specifically in fluid milk processing plants. A

secondary objective was to consider how these practices may influence product quality or shelf

life of products containing reworked ingredients. The goal of manuscript was to identify and

characterize the diversity of rework practices used in milk processing facilities to guide future

efforts to support the industry to minimize waste while maintaining quality throughout shelf-life.

MATERIALS AND METHODS

National fluid milk dairy processors (n=15) were individually invited to participate in

interviews about rework practices used in the dairy industry. Between 2019 and 2020,

representatives from seven commercial fluid milk processors from the Pacific Northwest were

interviewed in-person at the respective facility (n=3), via telephone or video call (n=2), or in-

person away from the facility (n=2) about the rework handling procedures at their facilities. Prior
 12

to each interview, the processors were assured of anonymity of their answers and practices.

Interviews were conducted in an open conversational format that included prompts of 16

questions that aimed to understand dairy processing rework practices, including their definition

of rework, frequency of rework and business motivation for reworking product, rework dilution

rates, common noticeable defects, quality testing parameters, and rework management practices

(Table 2.1). During the interview, processors were also asked to also share a general description

of their company and product line. In-person interviews at the facility consisted of a tour of the

production floor and a demonstration of their rework procedures. Concluding each interview, the

processors were provided an opportunity to share any other information pertaining to rework that

they were interested in learning or that was not covered during the survey.

RESULTS

Surveyed Facilities

The average daily production volume of the seven surveyed processing facilities ranged

from 7,000-100,000 gallons per day (26500-38000 liters per day) and collectively manufacture

products for the retail and foodservice markets. A summary of the pasteurization processes as

well as the products, packaging type, and packaging sizes produced by each facility is shown in

Table 2.2. There were six facilities (Facility C, D, E, F, G) that pasteurized products using high

temperature short time (HTST; 72°C, 15 sec (minimum)), one facility (Facility A) that ultra-

pasteurized (UP) their products (280°C, 2 sec), and one facility (Facility B) that processes milk

using both HTST and UP. All facilities receive raw milk for pasteurization and produce

whipping cream and fluid milks (fat-free skim milk, low-fat 1%, reduced-fat 2%, and whole).

Seven facilities produce half-and-half and flavored/blended dairy beverages including chocolate
 13

Table 2.1. An overview of the rework survey questions delivered to all processors.

Survey Question
1. How do you define rework in your facility?
2. What is your motivation for rework?
3. What types of products are reworked? What are the most common reworked products?
4. Which products are reworked into which products?
5. What products can you not rework and why?
6. Can products that leave the facility be reworked? (ex: over-shipment)
7. How are the products reworked?
8. What are your dilution rates for rework and how were these parameters defined?
9. How long can the product typically be stored prior to being reworked? Is there a maximum
time frame the product is stored prior to being reworked (code date range, in storage tanks,
etc.)?
10. Does the product undergo quality testing prior to being reworked or during the rework
process?
11. Are there differences in quality testing for finished reworked product?
12. Is reworked product sold at the same price and grade as non-reworked products?
13. Do customers request/decline products that contain rework?
14. How much rework is completed each day/week/month for your facility?
15. In your experience, what are some issues or defects that you have observed from
reworking products in your facility?
16. What other information would be valuable to you as a processor that could be obtained
during this rework investigation?
 14

Table 2.2. Pasteurization method, product types, and packaging type and size used by dairy
processing facilities (A-G) participating in the rework survey.
Milk Processing Facility
A B C D E F G
Pasteurization Method
HTST X X X X X X
UP X X
Product Types
Pasteurized milk X X X X X X X
Half-and-half X X X X X
Cream X X X X X X X
Flavored milk beverages X X X X X X
Eggnog X X X X X
Lactose-free milk products X X X
Ice cream X X X X
Packaging Type and Size
Carton (waxed, paperboard)
½ pint X X X
Pint X X X X
Quart X X X X
½ gallon X X X
Jug/bottle (plastic)
16 oz. bottle X X
½ gallon X X X X X X
Gallon X X X X X X
Bag-in-box (5 lb) X X X
 15

flavored milk, strawberry flavored milk, and eggnog. Four facilities produce ice cream mixes,

and three facilities produce lactose-free products. Four facilities package their products in

varying sizes of plastic milk jugs, plastic bottles, and cardboard cartons with three facilities

manufacturing 5-lb bag-in-box products while the remaining three facilities package their

products exclusively in plastic milk jugs. All eight dairy milk processors declared they

implement rework practices and subsequently answered all survey questions including a

description of their current operation.

Defining “Rework”

When discussing rework, all eight processors employed rework as a blanket term to

describe a product, an ingredient, and a process. The majority of processors (5 out of 7) broadly

defined rework as a product that failed a quality standard. Some facilities used supplemental

synonyms to name specific types of rework . The lexicon introduced by survey respondents for

rework included reclaim (3 facilities), recovery (1 facility), and rerun (1 facility). Two

facilities utilized reclaim to describe packaged products that were predetermined as rework and

immediately removed from the processing line. Conversely, one facility specifically delineated

reclaim as unpackaged product that had been recovered directly from the fillers whereas

another facility designated this product recovery . Finally, one facility simply employed

rerun as a substitute term for rework. There were two processors that distinguished rework

as the amalgamation of separate lots. Both of these facilities have the capability to recover

product from the production line and immediately add it back into the balance tank during the

same production run for re-pasteurization; thus, the collected and commingled product would
 16

retain the same lot code. Consequentially, these processors only consider a product to contain

rework when two separate lot codes are combined.

Motivations for Rework

Fluid milk processors reported a total of nine typical motivations for rework in their

facilities (Table 2.3). The most common motivations included packaging errors, code date

challenged product, and recovered product loss (reclaim). Other motivations included mitigating

product loss in special situations and making decisions around products that could cause

problematic issues to the rework system.

Packaging problems

Common motivations for rework reported were associated with packaging problems.

Packaging problems included underfilled (short filled) containers, incorrect labeling, and

damaged or defective containers. All processors reported that they rework packaged products

that were underfilled. Packaging flaws included mislabeled products, misshaped or cosmetically

damaged containers, and missing caps. One of the facilities does not rework products with

cosmetic packaging flaws, instead they donate these products to a non-profit food distribution

center.

Code-date challenged product

Code-date challenged product is designated as product stored at the production facility

that is unsold and approaching its sell-by date. This concern was common for processors that

produce milk for the school lunch program (1/2-pint cartons), especially when school breaks

(spring, winter, summer) create disruption in product demand. Each processor established their

own standard for the code range in which a product is eligible for rework (Table 2.4). One
 17

Table 2.3. Rework motivations reported by dairy processing facilities participating in the rework
survey.
Facilities Indicating
Rework Motivations
Motivation
Underfilled containers 7/7 (100%)
Packaging flaw (cosmetic) 6/7 (86%)
Code date (unsold product) 5/7 (71%)
Elevated microbial counts 4/7 (57%)
a
Packaged reclaim 5/7 (71%)
Prepackaged reclaima 3/7 (43%)
b
Temperature-dependent 1/7 (14%)
Packaging leakers 2/7 (29%)
c
Re-sanitized containers 2/7 (29%)
a
Reclaim: product captured following CIP and product changeovers.
bTemperature-dependent motive are products that are reworked due to not be refrigerated

immediately following packaging.

cRe-sanitized containers caused from products stored in milk crates that fell over during storage

or transport.
 18

Table 2.4. Code-date ranges that surveyed dairy processors indicated as standards for reworking
fluid milk products.
Facilities Indicating
Rework Code Date Range
Rework Code Date
Within 3 days of production 1/7
Up to 7 days of code date 2/7
Up to 5 days of code date 2/7
Up to code date 2/7
 19

facility s UP products can have a printed code date between 65-100 days depending on the

product and can be reworked up to 7 days within these code dates. Three facilities that have

products with printed code dates up to 21 days and rework products to within 5 (2 facilities) or 7

days (1 facility) of code date. One facility reported to rework their products up to their printed

code date of 21 days from production. Another facility only reworks product that is within 3 days

of the production date. Their decision to not rework code-date challenged product was based on

in-house studies that showed an increased potential for premature spoilage if they used product

stored beyond 3 days after processing.

Four facilities reported reworking products with elevated microbial counts standard plate

counts (SPC) or thermoduric counts. Processors will only rework these types of products if they

are within the specified bacterial limits for Grade A raw milk outlined in the Pasteurized Milk

Ordinance (PMO) (SPC 300,000 CFU/ml if commingled with other producer milk prior to

pasteurization) (United States Public Health Service/Food and Drug, 2017). Two facilities

indicated that elevated microbial counts were their top motivation for rework. These facilities

only rework microbially challenged product if counts are <3-log CFU/ml; otherwise, these

products are discarded. One facility reported to rework product containing elevated coliform

counts (>10 CFU/ml) as their only microbial motivation. Three processors do not rework

products with elevated microbial counts; instead, these products are either donated for animal

feed or discarded.

Reclaim: Recovering loss from product changeovers

All surveyed facilities reported using reworking strategies to recover off-specification

product associated with product changeovers. Most processors refer to this process as reclaim.

Common transitions for fluid milk processors include: skim milk removed during butterfat
 20

standardization, transitions from lower butterfat products to higher butterfat products, transitions

from dissimilar product changeovers (i.e., fluid to flavored), product recovered from the pipes

prior to and/or after sanitation. Processors typically flush a predetermined volume through the

fillers for immediate disposal prior to collecting and capturing the reusable volume of product.

Surveyed processors described two types of reclaim systems: bulk reclaim or packaged

reclaim. Three of the surveyed facilities use a bulk reclaim system to divert product away from

the fillers to a holding tank or storage vessels where it may be stored for up to 3 days at 7 C

until it is reworked. Two of the facilities have a bulk reclaim system that consists of a conduit

draining system that is installed directly under the fillers. It is programmed to flush the initial

unusable product volume for immediate disposal and then the operator will manually switch the

system to reclaim a predetermined volume of product that is pumped into a holding tank. One

facility manually collects reclaim in 10-gallon (38-liter) stainless steel milk cans and

immediately transfers it into the balance tank for incorporation into the same production lot;

therefore, this facility does not consider this to be rework . If the lot has since changed or

production has stopped for the day, this reclaimed product will be stored in the milk cans at 7 C

until it can be incorporated into another product lot. In this later situation, this reclaim is

considered rework .

Four milk processors reclaim off-specification product in packaged form. In these

facilities, the diluent is flushed through the fillers down the drain, and once milk product begins

to flow through, packaging begins. The first 20-30 filled containers will be filled with product

that does not meet product specification. These containers are immediately removed from the

line and held in cold storage (up to a processor s specified code range) until they are reworked in

a later production lot.

 21

One facility takes a mixed approach to reclaim using a combination of bulk and packaged

strategies. In this facility, the fillers are flushed of the diluent and then a predetermined number

of gallons of product is reclaimed by pumping directly into a holding tank. After this initial bulk

reclaim, packaging begins; however, the first 18-20 containers are immediately removed from

the line and donated to a food distribution warehouse. The number of packaged reclaim removed

from the line is dependent on facility conditions and are established prior to collection to ensure

that the proceeding finished packaged products have hit targeted specifications and are

unadulterated by other ingredients or constituents.

Special situations

While dairy processing is an organized system, operational crises can arise from

equipment failures or human error that make a large volume of product unsaleable. In

unexpected circumstances where product may have been compromised, dairy processors can

employ rework strategies to manage the inherent product loss or waste. The following examples

were shared by facilities in which rework was employed in unique situations:

One facility found a stainless-steel stool at the bottom of a processing tank proceeding a

production run that had been inadvertently left by an operator during a manual deep cleaning.

Although the stool had endured the CIP and sanitation courses, the facility chose to rework the

product lots that had passed through this processing tank lest the stool had cross contaminated

the products.

Only one facility mentioned reworking products due to temperature deviations. They

reported to rework UP products if they were left unrefrigerated (>45 F) for greater than one

hour. This has only occurred in emergency situations when the processing line shut down in the

midst of production and the conveyor belts were unable to transport product to cool storage.
 22

Two facilities reported reworking products that are re-sanitized. This is typically caused

from products stored in milk crates falling over during storage or transport. Two facilities

indicated that they did not rework the re-sanitized products if the container was found to be

damaged or compromised from the incident.

Situations where product is ineligible for rework

According to the PMO, not all products are eligible for repasteurization (reworking).

Mishandled products, contaminated or adulterated product, damaged or leaking containers, out-

of-code containers, and temperature abused products are all disqualified from being reworked

into Grade A dairy products (United States Public Health Service/Food and Drug, 2017). All

surveyed processors echoed many of these reasons that milk products would be classified as

ineligible for reworking in their facilities. Products reported to be rejected for rework by all

processors included temperature-abused or adulterated products, products that have surpassed the

printed code-date on the container, and products that have left control of the facility. Processors

also reported that products containing rework cannot be reworked again. Conversely, two

facilities disclosed that they do rework leaking containers (also known as leakers ). Leakers

are packaged products in which the packaging, usually the carton or jug seal, is damaged enough

to lose product. Processors indicated that products ineligible for rework were either discarded,

donated as animal feed, or donated to a food distribution warehouse depending on the reason the

product could not be reworked.

Products Eligible for Rework

All surveyed processors indicated that fluid milk via reclaim processes from butterfat

standardizations is the most frequent (daily) rework strategy. Other products that are held for
 23

rework include lactose-free fluid milks (3 facilities), whipping creams (6 facilities), half-and-half

(3 facilities), and blended flavored products such as chocolate milk, coffee creamers, and ice

cream mixes (6 facilities) (Figure 2.1). However, not all products are held for reworking. All

four of the surveyed facilities that manufacture strawberry flavored milk reported that strawberry

milk is not typically retained for rework. One processor revealed this decision was founded on

quality defects associated with their strawberry milk products including a noticeable premature

degradation of the red color (FD&C Red No. 40 typically) during storage.

Storing rework prior to processing

Processors described two different options for storing rework prior to processing: 1) bulk

storage: unpackaged product stored in portable containers or a storage tank until they are

reworked (i.e., reclaim); 2) packaged products commingled into portable storage vessels or a

storage tank and held until they are reworked. The storage period varied depending on the system

and storage conditions set by each facility.

Three processors have a bulk storage system for product to be reworked. As described in

the reclaim section above, one facility uses milk cans for bulk short-term storage and two

facilities hold their reclaim product in a refrigerated ( 7 F) recovery or reclaim storage tank.

Two of these facilities typically hold their reclaim product for up to 24 hours whereas one

facility considers their reclaim analogous to raw milk and thus holds it for up to 72 hours prior to

processing.

All seven surveyed processors implement the second storage system where packaged

products are emptied into either a mobile storage vessel (2 facilities), a recovery/reclaim storage

tank (5 facilities), or a rework/processing tank (7 facilities) (Figure 2.2). These products can be
24

Products that are used as rework

Lactose Half Ice
Nonfat 1% 2% Whole Chocolate Strawberry Coffee
Free & Creams Cream
Milk Milk Milk Milk Milk Milk Creamer
Milk Half Mix
Lactose Half Ice
Nonfat 1% 2% Whole Chocolate Strawberry Coffee
Free & Creams Cream
Milk Milk Milk Milk Milk Milk Creamer
Milk Half Mix
Products that accept rework
Figure 2.1 Fluid milk products that can serve as the source or recipient of rework. Increase in color gradient of products
that accept rework signify an increase of product types that are received.
25

Cold Storage
(4○C)
Rework Raw
Processing cream 1-3
tank Separator storage Bullet
Balance days
tote
or
Blending
tank
Raw HTST
pasteurizer Filler
milk Milk Milk 1-3
days
silo Recovery can can
Reclaim im
tank c la
Homo- Re
genizer Dump
Jug Pot
3-21 days
Figure 2.2 Collection, storage and incorporation of milk processing streams that are used as rework into fluid milk. Blue
arrows indicate typical fluid milk processing flow. Yellow arrows indicate rework introduction into the typical processing
flow.
 26

stored in their original packaging up to the processor s set code date range before they are

opened and mixed in these vessels or tanks for storage or processing. Information about the

packaged products is documented and then the product is transported to a rework station within

the facility. Individual containers are opened by unscrewing the cap or with a punchout tool. The

package is then forcefully smashed against the blunt side of the collection receptacle which

quickly empties and crushes the carton or jug. Two facilities use specially designed mobile

rework collection containers ( dump pots ) that are equipped with perforated screens to collect

physical debris (carton glue, cardboard particles, etc.) that may have transferred from the

package to the storage tank during unpackaging. Two facilities collect milk from packaged

products into a jacketed batch tank located on the production floor that does not have

temperature regulation capability. Three facilities open product directly into existing tanks within

the processing stream (reclaim storage, blend, or bulk tanks).

Two facilities open and mix packaged products into mobile storage containers. One

facility empties their packaged products into stainless steel milk cans in the storage cooler

( 7°F). The other facility collects the milk from packaged products into mobile bullet totes that

are stored in the cooler ( 7°F) until they are processed. This facility only utilizes bullet totes

when there is a large volume of product that will be reworked, most commonly a flavored milk

product. Both facilities reported that these mobile storage containers are held in a segregated area

of the cooler and can be stored up to the processor s elected rework code date range; however,

milk cans are typically reworked the same day and bullet totes are reworked within one week.

Five processors deposit or commingle their packaged products into a dedicated

refrigerated ( 45 F) storage tank where they can be stored for up to 72 hours. Two facilities
 27

directly pump their rework into the bulk reclaim tank and store it for up to 24 or 72 hours. The

product is then transferred from the storage tank into a rework or batch/blend tank, staged for

processing, and processed within 12-24 hours. One facility dumps their product into a batch tank

called a rerun tank that has a small filter to capture any foreign debris. It cannot be temperature

regulated and can only connect to the processing line thus products are only placed in this tank

when they are actively reworking these products. One facility commingles their products directly

into a blend tank or bulk tank that is attached to the filler and is processed the same day.

Testing Products Intended for Rework

Each fluid milk processor reported a composition and/or quality analysis of product prior

to approving it for rework. All facilities reported compositional analysis of butterfat content and

solids to facilitate adjustments in the final formulation of the finished product. Six facilities

analyze rework for titratable acidity (TA) measurements and conduct an informal sensory

analysis (aroma and flavor). Two facilities that rework into cream products also perform

functionality tests for whipped cream. Two facilities reported microbial testing, specifically

standard plate count (SPC) and coliform testing, but only if less than 50% of the code date

remained. Samples for these analyses are taken from the bulk tank or from a package prior to

opening and mixing in the storage or rework tank. Results from all tests are received and

reviewed prior to the product being reworked.

Processing Rework and Dilution Rates

All processors dilute their rework with fresh product to minimize potential quality defects

that would be perceived with product that was simply reprocessed. Most facilities rework their
 28

unflavored fluid rework and/or reclaim products into other unflavored products (5 facilities)

and/or flavored or blended dairy products (6 facilities). Two facilities do not rework into

unflavored fluid milk. Processors reported that skim milk and whipping creams typically do not

receive any reclaimed fluid milk because of the difficulty in meeting the target butterfat

requirements. One facility only incorporates rework into their ice cream products.

To successfully incorporate rework into fresh product, the rework should be mixed in uniformly.

The strategy for successfully incorporating of rework is slightly different for unflavored fluid

milk than for flavored or blended dairy products.

To rework fluid dairy products into fresh unflavored fluid dairy products, the rework and

reclaim products are transferred from the storage tanks, storage vessels, or original packages and

commingled into either a rework tank, a batch tank, or a bulk tank as described above. Five

facilities have a specific rework tank in which the transferred products can be stored for up to 24

hours in the production area. Two facilities do not have a separate rework tank and must process

their rework immediately after it enters the production floor. The tank is then joined to the

processing line and the rework tank is discharged into the raw milk line where it mixes with

fresh raw milk and then proceeds through the typical processing stages (i.e., pasteurization and

packaging). All five facilities that rework into unflavored milk products use a dilution rate of

20% (20% rework:80% fresh product).

To rework into flavored or blended milk products, previously flavored or unflavored

rework products can be either pumped from a storage tank or bullet totes or dumped directly

from the container or milk can into a batch or blending tank where they are commingled with

other ingredients (stabilizers, flavor adjuncts, etc.). For large volumes, the composition is tested

to ensure the appropriate ratio of ingredients are added. The batch is then blended with raw milk
 29

and the rest of the ingredients until the target formulation has been obtained. Maximum rework

dilution rates in flavored/blended products varied by facility, ranging from 3-5% in two facilities

to <20% in one facility. The facility that reworks into ice cream mixes does not have a maximum

rework rate set; however, it is not uncommon for rework to be as much as 40% of the finished

product.

Since flavored products contain an amalgamation of ingredients, surveyed facilities

reported that different types and flavors of rework products can be reworked into other flavored

and blended products. Four facilities typically combine like and unlike rework products into

flavored and blended products along with reclaim, fresh milk, flavoring adjuncts, and other

ingredients into the blending tank to reach the composition of the final product they are targeting

(chocolate flavored milk, ice cream mixes, etc.). The reported like products include similar

flavored milks while unlike products can comprise of whipping creams and/or half-and-half

products. Two facilities only rework like products into their finished flavored products along

with reclaim, fresh milk, and other ingredients.

All surveyed processors estimated the typical volume of rework they process within their

facility in one month. Three facilities process 1-2% total production volume as rework per month

and the other four facilities process <1% total production volume as rework per month.

Quality implications associated with rework

Six processors reported that discernable quality implications or a diminishing shelf life

were more commonly observed in products that contain rework when compared to products that

do not contain rework as assessed by in-house analysis. It is important to note that the defects

described are not exclusive to dairy products containing rework. Four facilities noted an
 30

intermittent athletic tape or bandage aroma and flavor in their chocolate and/or lactose free

milks that contained rework. One of these facilities reported other flavor defects in finished

products that contain rework, specifically banana or fruity off flavors in chocolate milk and

a nutty flavor in fluid milk. Two facilities shared when chocolate milk contained reworked

unflavored fluid milk, the products developed flavor defects more commonly than if they

contained reworked flavored milks. Two facilities reported elevated microbiological counts

(standard plate counts (SPC) and thermoduric counts) in finished product containing rework.

Two facilities indicated that their finished flavored products containing >10% rework were more

likely to separate due to air intrusion, stabilizer imbalances, and/or overworking. One facility has

not noticed any discernable flavor defects or diminishing shelf life in products that have been

reworked when compared to non-reworked products.

DISCUSSION

Rework is a way to help minimize waste by reprocessing products that have failed a

quality standard. Quality standards for pasteurized fluid milk are based on the butterfat and

solids percentage as stated in the standard of identity for dairy products listed in

21CFR§131.110, microbiological quality of finished product (coliform counts, SPC, etc.), and

organoleptic qualities (flavor, aroma, visual appearance) (Barbano, 2017; United States Public

Health Service/Food and Drug, 2017; Martin et al., 2021). If a product is classified as rework or

destined for repasteurization, the failed quality standard shall pose no risk of adulteration (Dairy

Practices Council, 2005; United States Public Health Service/Food and Drug, 2017). For fluid

and flavored milk processors, a failed quality standard can include low butterfat from pipeline

collections following sanitation, short fills, and/or diluted flavors from product changeovers.
 31

Excess product that is approaching the sell-by date stored at the production facility can also be

reworked. However, reprocessing products could still jeopardize the quality of the finished

product by introducing other physical, microbiological, and/or chemical hazards (ICMSF, 2005;

Center for Food Safety and Applied Nutrition, 2007; Agriculture Marketing Service, 2013). The

physical and chemical hazards are largely controlled at the plant level through stringent

operating technology such as filtration screens and metal detectors for physical hazards and

documentation for traceability and allergen control for chemical hazards. Because rework dairy

products are required to be repasteurized, there is low risk for pathogenic concerns. Quality

implications of rework has been previously investigated for other dairy products including butter

(Long, 1939; Dolby, 1965; Chambers, 2002), ice cream (Holm et al., 2002; Cigerdelen, 2011),

and processed cheese ( ern kov et al., 2018). To our knowledge, there are no previous

investigations into the quality implications for fluid milk processing. Furthermore, there are

some concerns that rework practices can result in an increase in premature spoilage, flavor

defects, and spore counts in finished products (ICMSF, 2005). Processors are making decisions

in processing their rework without any previous data and must weigh the risks a product or

process could introduce into the operating system.

Rework, particularly reclaim, is a daily reality for dairy processors as a way to

circumvent product loss and reduce waste while recovering some ingredient costs. If the products

are deemed ineligible for rework, or if it is not economically feasible for a processor to spend the

labor and processing costs to rework them, then processors can choose to donate these products

as animal feed or to a food distribution warehouse. Rework is also an option for processors to re-

pasteurize compromised products from unpredictable circumstances into saleable products.

Although each facility employed unique rework practices tailored to their operational
 32

organization and limitations, dairy processors are operating within the requirements of the PMO

with the exception of two processors that reported to rework leakers . It was clear during this

survey that processors would rather reject a product for rework than to rework a potentially

adulterated product. Therefore, one explanation for this deviation can be attributed to the diverse

lexicon used throughout regulatory reports that guide Grade A fluid milk processors in their

rework practices. Dairy processors utilize the term rework as a blanket term to describe rework

and reclaim practices. The PMO uses rework to discuss allergens and the disposal of recalled

milk , and then employs the term repasteurization to discuss rework parameters, including

leakers (United States Public Health Service/Food and Drug, 2017). In the Code of Federal

Regulations (CFR) 21 CFR 117.3, rework is defined as clean, unadulterated food that has

been removed from processing for reasons other than insanitary conditions or that has been

successfully reconditioned by reprocessing and that is suitable for use as food . Furthermore, in

21 CFR 110.80, adulterated food is capable of being reconditioned using a previously proven

and effective method prior to being commingled into food. Conversely, the United States

Department of Agriculture (USDA) outlines specific rework and reclaim guidelines for dairy

products made from Grade B fluid milk (butter, cheese, dry milk, etc.) using similar language

industry processors employ (United States Department of Agriculture, 2009). Aligning the

language in regulatory documents or providing clarification and/or definitions in regard to the

use of rework , repasteurization , and reprocessing in the PMO for Grade A fluid milk

processors may help prevent erroneous processing decisions such as the leaker issue.

Fluid dairy products containing rework and reclaim are at a higher risk for developing

quality defects such as over-processing destabilization, flavor defects, and elevated microbial

counts than products that do not contain rework. Over-processing challenges such as product
 33

separation were typically seen in flavored and blended reworked products that contain other

ingredients such as flavoring adjuncts (i.e. cocoa powder), stabilizers, and sugar in reworking

dilutions >10% which can like be attributed to the stabilizer. Stabilizers require specific ratios

and processing parameters in dairy products for efficient functionality (Tasneem et al., 2014).

Processors also found ways to mitigate issues from reprocessing reworked products. Flavored

and blended products such as flavored milks and ice cream mixes were the most common

commodities to receive reworked products because the supplementary ingredients assist with

palatability (Figure 2.1). The added sugar and flavoring ingredients can mask undesirable flavors

developed during storage or reprocessing such as cooked/scorched flavors derived from

reprocessing fluid milk or cream. Although half and half does not contain these additives, it is

generally consumed as an ingredient (coffee enhancer, bakery applications, etc.) which may also

mask any flavor defects.

There are a multitude of variables in a dairy processing system that could contribute to

spoilage implications; however, there are two main factors that may be more influential in the

rework system: 1) pasteurization, and 2) storage time and temperature. Although rework

products are reprocessed, pasteurization is not a sterilization step and milk acts as a nutritious

growth medium that allows for bacterial growth to occur. Endospore forming psychrotrophic

bacteria, predominantly Bacillus and Paenibacillus species, are passed from the farm

environment to the raw milk as heat-resistant spores (Coorevits et al., 2008; Vissers and

Driehuis, 2008; Martin et al., 2019) which can survive pasteurization and germinate into

vegetative cells (Meer et al., 1991; Ranieri et al., 2009) and subsequently grow under

refrigeration conditions (Ivy et al., 2012; Buehler et al., 2018; Sun et al., 2021). However, all

cells may not germinate after pasteurization and in theory, some species could sporulate during
 34

extended periods of refrigeration of the milk (Buehler et al., 2018; Sun et al., 2021).

Furthermore, the survival of just one spore of Paenibacillus species, the dominant bacteria

isolated at the end of shelf life in pasteurized fluid milk, is capable of growing to spoilage levels

at refrigeration temperatures before the 21 day shelf life (Ranieri et al., 2009; Ivy et al., 2012;

Trm i et al., 2015; Buehler et al., 2018; Beno et al., 2020; Martin et al., 2021). Therefore,

psychrotrophic heat resistant spore-forming bacteria should be of most concern in causing

spoilage events in rework products. All products classified as rework or reclaim are stored for a

varied amount of time at 7 C before they are reprocessed, allowing for increased bacterial

counts and potential sporulation behavior. Multiple processors rework products between 15-21

days of the code date. Additionally, products can be stored for another 24-72 hours during

processing depending on facility practices. Thus, rework products could be up to 24 days old by

the time they are repasteurized which could be increasing the initial cell density of heat-resistant

spoilage bacteria in the finished product. In a study by Buehler et al. (2018), investigators

modeled the growth of spore-forming bacteria commonly isolated from pasteurized milk and

found that the growth rate of the organism and low storage temperatures (4°C) were the two most

important factors in controlling spoilage outcomes. Because spores are likely the most influential

in causing quality defects in reworked products, understanding sporulation behavior of Bacillus

and Paenibacillus species in pasteurized milk over shelf life will be critical information for a

processor. To our knowledge, this has not been previously investigated and should be a

consideration for future studies.

Although there are some guidelines available to processors in the PMO and from the

DPC, the focus is based on food safety, not on quality control. Processors are making their own

decisions to control these parameters at the facility level. Two facilities shared that they had
 35

performed in-house studies to support their rework practices which resulted in processing

practices that were unique from the other surveyed processors. One processor decided to only

rework products within 3 days of the production run due to increased thermoduric counts.

Another processor noted a predictable flavor defect ( band-aid ) in chocolate milk when batches

were reworked over 10% and decided to set this amount as the rework threshold for chocolate

milk. Supplementary analyses were also employed at multiple facilities to capture quality issues

prior to the final testing of the finished pasteurized products with TA tests being the most

common. TA can be used as an indicator test for increased microbial counts in pasteurized milk

due to the change in acidity (Ziyaina et al., 2018); however, one of the processors also used TA

as an indicator of alkali in their reclaim captured after the sanitation step. These analyses are not

required but they do assist processors in making informed decisions for their rework practices.

Processors also have to consider their customers when making decisions around rework. Two

processors shared that they have customers with contracts that prohibit rework in their finished

products.

Although this survey provides a good narrative of a dairy processor s rework practices,

rework is a difficult topic for the industry to openly discuss because outside of reclaim, rework is

fundamentally created through products failing a quality standard. These are sensitive issues that

processors are reluctant to discuss which created some challenges in data collection. Processors

were more comfortable sharing their practices through a live interview (in-person or over the

phone) as compared to a paper or electronic survey. The industry is motivated to understand how

rework can impact finished product quality. Each processor that was surveyed was curious about

the practices employed at other processing facilities to gain confidence in their current system.

One recommendation that could further increase a processor s confidence in rework practices is
 36

aligning the language around rework and reclaim in regulatory documents, particularly the PMO.

Because the driver of product quality failures in reworked products is centered around

psychrotrophic heat resistant spore formers, future investigations to understand the survival and

sporulation behavior of Bacillus and Paenibacillus species commonly isolated from pasteurized

milk could also assist processors in making informed rework decisions.

CONCLUSIONS

While rework is a common practice in the dairy industry, it has not been well described.

This survey provides a detailed summary of a diverse set of rework practices employed at fluid

milk processing facilities ranging in size and production type. Based on these results, we were

able to identify rework conditions that have the potential to influence product quality or shelf life

of products containing reworked ingredients. This information can guide processors decisions as

well as future research that will help to minimize waste while maintaining quality throughout

shelf-life.

ACKNOWLEDGMENTS

We would like to thank BUILD Dairy (Logan, UT) and the Arbuthnot Dairy Center at

Oregon State University (Corvallis, OR) for providing the funding and resources for this survey.

We would also like to thank each of the processors that were willing to participate in this survey.

REFERENCES

Agriculture Marketing Service. 2013. Instructions for Dairy Plant Surveys. Washington, DC.

Ahmad, T., R.M. Aadil, H. Ahmed, U. ur Rahman, B.C.V. Soares, S.L.Q. Souza, T.C. Pimentel,
 37

H. Scudino, J.T. Guimarães, E.A. Esmerino, M.Q. Freitas, R.B. Almada, S.M.R.

Vendramel, M.C. Silva, and A.G. Cruz. 2019. Treatment and utilization of dairy industrial

waste: A review. Trends Food Sci. Technol. 88:361 372. doi:10.1016/j.tifs.2019.04.003.

Aikenhead, G., K. Farahbakhsh, J. Halbe, and J. Adamowski. 2015. Application of process

mapping and causal loop diagramming to enhance engagement in pollution prevention in

small to medium size enterprises: Case study of a dairy processing facility. J. Clean. Prod.

102:275 284. doi:10.1016/j.jclepro.2015.04.069.

Barbano, D.M. 2017. A 100-Year Review: The production of fluid (market) milk. J. Dairy Sci.

100:9894 9902. doi:10.3168/jds.2017-13561.

Beno, S.M., R.A. Cheng, R.H. Orsi, D.R. Duncan, X. Guo, J. Kovac, L.M. Carroll, N.H. Martin,

and M. Wiedmann. 2020. Paenibacillus odorifer, the Predominant Paenibacillus Species

Isolated from Milk in the United States, Demonstrates Genetic and Phenotypic

Conservation of Psychrotolerance but Clade-Associated Differences in Nitrogen Metabolic

Pathways. doi:10.1128/mSphere.00739-19.

Buehler, A.J., N.H. Martin, K.J. Boor, and M. Wiedmann. 2018. Psychrotolerant spore-former

growth characterization for the development of a dairy spoilage predictive model. J. Dairy

Sci. 101:6964 6981. doi:10.3168/jds.2018-14501.

Center for Food Safety and Applied Nutrition. 2007. Hazards and Controls Guide For Dairy

Foods HACCP. Washington, DC.

ern kov , M., J. Nebes ov , R.N. Salek, R. Popkov , and F. Bu ka. 2018. The effect of

rework content addition on the microstructure and viscoelastic properties of processed

cheese. J. Dairy Sci. 101:2956 2962. doi:10.3168/jds.2017-13742.

Chambers, J. V. 2002. The Microbiology of Raw Milk. 3rd ed. R.K. Robinson, ed. Wiley-
 38

Interscience, Inc.

Cigerdelen, E. 2011. Integrated scheduling of ice cream production with rework via mixing.

Eindhoven University of Technology,.

Coorevits, A., V. De Jonghe, J. Vandroemme, R. Reekmans, J. Heyrman, W. Messens, P. De

Vos, and M. Heyndrickx. 2008. Comparative analysis of the diversity of aerobic spore-

forming bacteria in raw milk from organic and conventional dairy farms. Syst. Appl.

Microbiol. 31:126 140. doi:10.1016/j.syapm.2008.03.002.

Dairy Practices Council. 2005. Guidelines for controlling the quality and use of dairy product

rework. Keyport, NJ.

Dolby, R.M. 1965. Changes in moisture distribution caused by partial reworking of butter shortly

after churning. J. Dairy Res. 32:263 267. doi:10.1017/S002202990001863X.

Dou, Z., J.D. Ferguson, D.T. Galligan, A.M. Kelly, S.M. Finn, and R. Giegengack. 2016.

Assessing U.S. food wastage and opportunities for reduction. Glob. Food Sec. 8:19 26.

doi:10.1016/j.gfs.2016.02.001.

Douglas, S.A., M.J. Gray, A.D. Crandall, and K.J. Boor. 2000. Characterization of chocolate

milk spoilage patterns. J. Food Prot. 63:516 521. doi:10.4315/0362-028X-63.4.516.

Economic Research Service. 2020. Dairy Data. Washington, DC.

Holm, S., R.B. Toma, W. Reiboldt, C. Newcomer, and M. Calicchia. 2002. Cleaning frequency

and the microbial load in ice-cream. Int. J. Food Sci. Nutr. 53:337 342.

doi:10.1080/09637480220138133.

ICMSF. 2005. Microbial Ecology of Food Commodities. Second. Kluwer Academic/Plenum

Publishers, New York.

IRI Custom DMI Market Advantage Database. 2021. Total U.S. - MULO Retail YTD Milk
 39

Snapshot.

Ivy, R.A., M.L. Ranieri, N.H. Martin, H.C. den Bakker, B.M. Xavier, M. Wiedmann, and K.J.

Boor. 2012. Identification and characterization of psychrotolerant sporeformers associated

with fluid milk production and processing. Appl. Environ. Microbiol. 78:1853 1864.

doi:10.1128/AEM.06536-11.

Kolev, S. 2017. General Characteristics and Treatment Possibilities of Dairy Wastewater A

Review. Food Technol. Biotechnol. 53:237 242.

Long, H.F. 1939. Bacteriology of butter VII . Effect of reworking butter on growth of bacteria

Bacteriology of Butter 24.

Martin, N.H., D.J. Kent, R.L. Evanowski, T.J. Zuber Hrobuchak, and M. Wiedmann. 2019.

Bacterial spore levels in bulk tank raw milk are influenced by environmental and cow

hygiene factors. J. Dairy Sci. 102:9689 9701. doi:10.3168/jds.2019-16304.

Martin, N.H., P. Torres-Frenzel, and M. Wiedmann. 2021. Invited review: Controlling dairy

product spoilage to reduce food loss and waste. J. Dairy Sci. 104:1251 1261.

doi:10.3168/jds.2020-19130.

Meer, R.R., J. Baker, F.W. Bodyfelt, and M.W. Griffiths. 1991. Psych rotrophic Bacillus spp. in

fluid milk products: A review. J. Food Prot. 54:969 979. doi:10.4315/0362-028X-

54.12.969.

Milani, F.X., D. Nutter, and G. Thoma. 2011. Invited review: Environmental impacts of dairy

processing and products: A review. J. Dairy Sci. 94:4243 4254. doi:10.3168/jds.2010-3955.

National Agriculture Statistics Service. 2021. Milk Production and All Milk Prices. Washington,

DC.

Patra, F., and R.K. Duary. 2020. Waste from Dairy Processing Industries and its Sustainable
 40

Utilization. 1st ed. M. Thakur, V.K. Modi, R. Khedkar, and K. Singh, ed. Springer Nature

Singapore Pte Ltd., Singapore.

Ranieri, M.L., J.R. Huck, M. Sonnen, D.M. Barbano, and K.J. Boor. 2009. High temperature,

short time pasteurization temperatures inversely affect bacterial numbers during refrigerated

storage of pasteurized fluid milk. J. Dairy Sci. 92:4823 4832. doi:10.3168/jds.2009-2144.

Sun, L., K. Atkinson, M. Zhu, and D.J. D Amico. 2021. Antimicrobial effects of a bioactive

glycolipid on spore-forming spoilage bacteria in milk. J. Dairy Sci. 104:4002 4011.

doi:10.3168/jds.2020-19769.

Tasneem, M., F. Siddique, A. Ahmad, and U. Farooq. 2014. Stabilizers: Indispensable

Substances in Dairy Products of High Rheology. Crit. Rev. Food Sci. Nutr. 54:869 879.

doi:10.1080/10408398.2011.614702.

Tobe, F. 2016. Minimizing Production Losses in the Food and Beverage Industry. Accessed.

Tóth, K., C. Borbély, B. Nagy, G. Szabó-Szentgróti, and E. Szabó-Szentgróti. 2014.

Measurement of Food Losses in a Hungarian Dairy Processing Plant. Edible oil fat Refin.

1 10.

Trm i , A., N.H. Martin, K.J. Boor, and M. Wiedmann. 2015. A standard bacterial isolate set for

research on contemporary dairy spoilage. J. Dairy Sci. 98:5806 5817.

doi:10.3168/jds.2015-9490.

United States Code of Federal Regulations. 2021. Title 21 Food and Drugs: CFR Part 117.3,

Definitions.

United States Department of Agriculture. 2009. Instructions for dairy inspection and grading

service. Washington, DC.

United States Environmental Protection Agency. 2020. 2018 Wasted Food Report 42.
 41

United States Public Health Service/Food and Drug. 2017. Grade A Pasteurized Milk

Ordinance (PMO). Washington, DC.

Vissers, M., and F. Driehuis. 2008. On-Farm Hygienic Milk Production. A. Tamime, ed. Wiley-

Blackwell.

Ziyaina, M., B.N. Govindan, B. Rasco, T. Coffey, and S.S. Sablani. 2018. Monitoring Shelf Life

of Pasteurized Whole Milk Under Refrigerated Storage Conditions: Predictive Models for

Quality Loss. J. Food Sci. 83:409 418. doi:10.1111/1750-3841.13981.

 42

CHAPTER 3

EVALUATING PAENIBACILLUS ODORIFER FOR ITS POTENTIAL TO REDUCE

SHELFLIFE IN REWORKED FLUID MILK PRODUCTS

Authors: Casey E. Rush, Jared Johnson, Samantha Burroughs, Brandon Riesgaard, Lisbeth

Meunier-Goddik, and Joy G. Waite-Cusic*

Department of Food Science and Technology

Oregon State University, Corvallis, OR

*Corresponding author:

Joy G. Waite-Cusic, Ph.D.

100 Wiegand Hall

Department of Food Science and Technology

Oregon State University

Corvallis, OR 97331

(541) 737-6825

joy.waite-cusic@oregonstate.edu

Prepared for submission to JDS Communications

 43

ABSTRACT

Rework is a common practice used in the dairy industry as a strategy to help minimize waste

from processing steps or errors that might otherwise render the product unsaleable. It has been

reported that dairy processors may rework their HTST fluid milk products up to code date (21 d)

at a typical dilution rate ≤20% rework into ≥80% fresh raw milk. Heat resistant endospore-

forming bacteria have often contributed to quality defects and premature spoilage in non-

reworked dairy products due to pasteurization temperatures activating spore germination. Our

objective was to determine if rework storage practices would create an increased microbiological

burden on fluid and flavored milk products that could lead to premature spoilage. Paenibacillus

odorifer, a psychrotrophic spore-former and a dominant fluid milk spoilage organism, was

selected as the target to evaluate how rework storage practices might contribute to the microbial

burden of fluid milk. Commercial UHT-pasteurized fluid milk and chocolate milk were

independently inoculated with four strains of P. odorifer (FSL R10-2726, FSL A6-0363, FSL

E2-0150, JWC-2503) at 1-2 log CFU/ml. Samples were stored at 4°C and 7°C, P. odorifer

populations were determined over time. Aroma changes and compositional analysis were used to

monitor product changes during storage. P. odorifer strains grew more quickly and to higher cell

density at 7°C and in chocolate milk when compared to fluid milk. P. odorifer vegetative cells

were extremely sensitive to pasteurization treatments; however, subpopulations of thermally

resistant bacteria, including spores, can develop in cultured milk. While these thermally-resistant

subpopulations would pose an increased spoilage risk to reworked milk products, they do not

develop in fluid milk stored at 4 C by 21 days. The absence of thermally-resistant

subpopulations was verified in retail fluid milks at and beyond their shelf life. Therefore, P.

odorifer is unlikely to contribute to increased spoilage risk in fluid milk containing rework.
 44

INTRODUCTION

Quality defects and premature spoilage in pasteurized fluid milk caused by

microorganisms is a frequent challenge in the dairy industry which contributes to the estimated

31% of dairy foods discarded at the retail and consumer level (Buzby et al., 2014; Martin et al.,

2021). There are two primary sources of spoilage organisms in fluid dairy products: low levels of

heat-resistant bacterial spores in raw milk that survive the pasteurization process, and post-

pasteurization contamination (PPC) (Clark et al., 2009; Trm i et al., 2015; Martin et al., 2018).

PPC is commonly from Gram-negative bacteria present in the dairy processing environment

which must be mitigated through sanitation programs (Gopal et al., 2015; Reichler et al., 2018).

Bacterial spores of Bacillus and Paenibacillus species (spp) are readily found in raw milk and

are collected during the milking process from the farm environment (Meer et al., 1991; Coorevits

et al., 2008; Huck et al., 2008; Ivy et al., 2012; Grady et al., 2016). They can survive and

germinate at typical high temperature short time (HTST) pasteurization temperatures (72°C, 15

seconds) and grow under refrigeration conditions (Meer et al., 1991; Chambers, 2002; Huck et

al., 2008; Ranieri et al., 2009; Ivy et al., 2012; Moreno Switt et al., 2014; Gopal et al., 2015;

Beno et al., 2020). Sporogenesis is an adaptive characteristic and is triggered when a vegetative

cell is placed in unfavorable environmental conditions. This is often seen when nutrients are

scarce, an increase in cell density, and/or unfavorable storage temperatures, water activity, or

pH; however, when conditions are tolerable or favorable, the spore can then germinate back into

a vegetative cell where replication can proceed (Gauvry et al., 2017). Bacillus spp. and

Paenibacillus spp. can readily grow in milk at refrigeration temperatures (≤7°C) and are reported

to be the principal psychrotrophic spore formers isolated from raw and pasteurized milks;

however, Paenibacillus spp. are the predominant genus found in pasteurized milk at the end of
 45

shelf life with cell densities ≥6 log CFU/ml (Huck et al., 2008; Ranieri and Boor, 2009; Ivy et

al., 2012). Discernable organoleptic deviations have been reported to occur at cell densities

around 6.0 log CFU/ml for fluid milk products contaminated with Paenibacillus spp. including

off-flavors (rancid, fruity) and textural changes resulting from proteolytic and lipolytic activity

(Fromm and Boor, 2004; Clark et al., 2009; Ivy et al., 2012; Beno et al., 2020).

Management of the microbial quality of fluid milk is further complicated for dairy

processors that employ reworking practices to minimize waste. According to the Dairy Practices

Council (DPC), rework can be defined as, Clean, unadulterated food that has been removed

from processing for reasons other than insanitary conditions or that has been successfully

reconditioned by reprocessing and that is suitable for use as food. Fluid milk and milk products

drained from processing equipment at the end of the run, collected from a defoamer system and

milk solids rinsed from equipment, containers or pipelines, shall be repasteurized only if such

milk and milk products are handled in a sanitary manner and maintained at 7°C (45°F) or less

(Dairy Practices Council, 2005). The source of the product to be reworked has commonly failed

a quality standard, often due to underfilled containers, cosmetic packaging flaws, unsold product

(products approaching the sell-by date), elevated microbial counts, leakers, and/or diluted

product reclaimed from the fillers between product changeovers and following clean-in-place

(CIP) procedures. Dairy processors often commingle these products with raw milk at various

dilution rates resulting in a fresh product with a new code date. Common dilution rates used by

the industry are 20% rework for (unflavored) fluid milk and 10% rework for chocolate flavored

milk products (Rush et al. 2021). Depending on a processor s decisions regarding their rework

operating procedures, products could be reworked up to the code date printed on the carton. For

HTST pasteurized products, this typically falls between 14 and 22 days with flavored products
 46

having the shorter shelf life (Ranieri and Boor, 2009). If these products are contaminated with

psychrotrophic bacteria at the time of packaging, bacterial populations could reach >6.0 log

CFU/ml within this timeframe (Douglas et al., 2000; Buehler et al., 2018; Beno et al., 2020). The

elevated microbial load of the rework could have a negative impact on the microbiological

quality of the new product. Although there have been no prior investigations into the

microbiological quality implications of reworking fluid and chocolate flavored milks, surveyed

processors reported increased standard plate counts, thermoduric bacteria, and flavor defects

( band-aid , athletic tape , banana , nutty ) more commonly observed in rework products

when compared to products that do not contain rework (Rush et al., 2021). To have a negative

impact on the microbiological quality of the new product, the contaminants of the rework would

have to survive the second pasteurization process.

The majority (59%) of Paenibacillus spp. isolated from dairy products and dairy

environments are Paenibacillus odorifer (Beno et al., 2020). P. odorifer spores have been found

in the dairy environment including the soil, manure, the milking equipment, cow udders, and in

raw milk (Huck et al., 2008; Beno et al., 2020). P. odorifer spores have been reported to

germinate by pasteurization conditions and replicate in pasteurized milk throughout its

refrigerated shelf life (4°C-7°C) (Beno et al., 2020). Sporogenesis is an adaptive characteristic

and is triggered when a vegetative cell is placed in unfavorable environmental conditions. This is

often seen when nutrients are scarce, an increase in cell density, and/or unfavorable storage

temperatures, water activity, or pH; however, when conditions are tolerable or favorable, the

spore can then germinate back into a vegetative cell where replication can proceed (Gauvry et

al., 2017). Milk is nutrient-dense medium that supports the growth of P. odorifer which has also

been reported to grow in refrigeration conditions ≤7°C. These growing conditions have been
 47

verified where P. odorifer was grown in skim milk, whole milk, or skim milk broth (SMB) at

refrigeration temperatures and reached a final cell density of >4 log CFU/ml within 14 days

(Beno et al., 2020) and >6 log CFU/ml within 21 days (Ivy et al., 2012; Moreno Switt et al.,

2014). Sun et al. (2021) and Buehler et al. (2018) reported growth rates of one strain of P.

odorifer (AT15) in whole milk (7°C = 1.01 CFU/ml/day) and skim milk broth (SMB) (4°C = 0.4

CFU/ml/day and 6°C = 0.6 CFU/ml/day), respectively.

There are three stages of P. odorifer growth in fluid milk that could impact whether it

posed an elevated risk when used as rework. The first is the contamination level of P. odorifer

spores in raw milk. Psychrotrophic spore density in fluid milk has been reported by Buehler et al.

(2018) to average approximately 1 spore/6 ml of raw milk (-0.79 log spores/ml). Pasteurization

will shock these spores and they will germinate in the packaging product and grow throughout

the potential on-site storage of the milk destined for rework. The growth behavior of P. odorifer

in skim milk broth at 6°C has been previously reported as an estimate of the growth potential

during distribution, retail, and consumer refrigerator storage. However, dairy processing facilities

typically have more consistent temperature control and fluid and flavored milk products held for

rework are stored at or below at 4 C. Growth of P. odorifer has been predicted at this

temperature; however, it has not been confirmed. This information is necessary to determine the

microbial burden of milk that will be diluted into fresh product and re-pasteurized. While it is

known that P. odorifer spores survive pasteurization, there is a lack of information on the heat

resistance of vegetative cells and no evidence whether P. odorifer would sporulate in milk or

under these storage conditions. Martin et al. (2021) also reported that a spoilage event can occur

in a product container where only 1 spore would be present. If high levels of P. odorifer spores

develop in fluid or flavored milk products during the storage period of prior to rework, then
 48

increased spoilage would be likely for milk products containing rework (Vissers et al., 2007;

Buehler et al., 2018). Understanding time and temperature conditions that lead to P. odorifer

sporulation could guide dairy processors to limit the maximum storage time of milk that might

be reworked in order to minimize premature quality loss.

The overall objective of this research was to determine the potential impact of P. odorifer

on fluid milk products containing rework. This study was designed to characterize the growth

and behavior of the P. odorifer in fluid and chocolate milk stored under conditions (time and

temperature) used by the dairy industry prior to reworking into fresh product. Evaluating growth

rates, maximum cell density, and thermal resistance are critical parameters to determine the

potential impact of P. odorifer on fluid milk products containing rework.

MATERIALS AND METHODS

P. odorifer strains and culture preparation

Four strains of Paenibacillus odorifer were selected to obtain their growth parameters in

both fluid and chocolate flavored milks incubated at refrigeration temperatures. Three of the

strains were requested from the Cornell University Food Safety Laboratory culture collection

using the Food Microbe Tracker database (Ithaca, NY). Requested strains were selected based on

diversity of product type and rpoB allelic types (AT): FSL E2-0150 (AT2; chocolate milk), FSL

R10-2726 (AT35; pasteurized fluid milk), and FSL A6-0363 (AT40; pasteurized fluid milk). The

fourth P. odorifer strain (JWC-2503) was previously isolated and from chocolate milk by Oregon

State University s Food Safety Laboratory (Corvallis, OR). The rpoB allele type was determined

extracting the appropriate sequence from whole genome sequencing data and querying the Food
 49

Microbe Tracker Database. According to the criteria in Ivy et al. (2012), this strain represented a

new rpoB AT with one nucleotide difference from AT74 and two nucleotide differences from

AT07. Sequence alignment files were sent to Cornell University for review and confirmation of

unique rpoB AT and AT number will be assigned pending their review.

To prepare the cultures, the isolates were revived from frozen storage (-80°C) by

transferring an aliquot (3 ml) to Tryptic Soy Broth supplemented with 0.3% Yeast Extract

(TSBYE) (Neogen, Lansing, MI) with incubation at 25°C for 24 h. Cultures were streaked for

isolation on Tryptic Soy Agar plates supplemented with 0.3% Yeast Extract (TSAYE, Neogen)

and incubated at 25°C for 48 h. A single isolated colony of each strain was independently

transferred into TSBYE and incubated at 25°C for 24 h to achieve an approximate cell density of

8 log CFU/ml. The resulting culture was serially diluted in 0.1% peptone water (PW, Neogen) to

achieve a final cell density of 6 log CFU/ml. This served as the inoculum for fluid and chocolate

milk.

P. odorifer growth in milk products

Ultra-High Temperature (UHT) pasteurized reduced fat (2%) fluid milk (Parmalat,

Lactalis American Group, Buffalo, NY) and low fat (1%) chocolate flavored milk (Nesquik,

Nestlé USA Inc., Arlington, VA) products were purchased from a local retailer. Milk (1500 ml)

was aseptically transferred to a sterile Whirl-Pak bag (Nasco, Madison, WI) and stomached for

30 s. Each Whirl-Pak bag was inoculated with 150 µl of the previously prepared P. odorifer

single strain inoculum to achieve a final cell density of 1-2 log CFU/ml. Inoculated milk samples

were aliquoted (200 ml) into sterile 237 ml clear PET wide-mouth plastic jars (Uline, Pleasant

Prairie, WI) in triplicate and incubated at 4°C and 7°C for up to 31 d. Organoleptic (aroma,
 50

visual) observations and microbial counts were monitored throughout the incubation period.

Aerobic plate counts (APC) were determined using standard serial dilution (0.1% PW) with

spread plating on TSAYE and incubation at 25°C for 48 h prior to enumeration. For selected

trials, enumeration of thermotolerant subpopulations were determined in 2% UHT milk stored at

4°C by heat treating milk samples at 72°C for 15 s (thermoduric) or 80°C for 12 m (spore count),

respectively, prior to plating on TSAYE with incubation as described above. After 21 days of

storage at 4 C or 7 C, only the fluid milk samples were analyzed for compositional changes

using the LactoScope Fourier transform infrared spectroscopy (FTIR) FTA-3.4 instrument (Delta

Instruments, Kelvinlaan 3, The Netherlands). Chocolate milk samples were not analyzed using

the LactoScope FT-IR because of the low accuracy for flavored milk products. The LactoScope

FT-IR Samples with obvious contamination were not analyzed using the Lactoscope. An

uninoculated UHT milk sample was analyzed as the control. Each inoculated milk was scanned

three times and the result was averaged before further statistical analysis.

Thermal inactivation of P. odorifer

The thermal resistance of P. odorifer was investigated in inoculated milk. A single

isolated colony of each P. odorifer strain was aseptically transferred into a sterile 50 ml conical

tube with 25 ml of UHT-pasteurized reduced fat (2%) fluid milk and incubated at 25°C for 24 h

(approximately 8 log CFU/ml). The four cultured milk samples were then combined in a

sterilized media bottle and shaken for 30 s. This strain mixture served as the P. odorifer cocktail

and was stored 4°C for 24 h prior to use. Aliquots (900 ml) of UHT-pasteurized reduced fat (2%)

fluid milk (Parmalat, Lactalis American Group, Buffalo, NY) were transferred into 1.5 ml

sterilized centrifuge tubes (Eppendorf, Enfield, CT) and subsequently inoculated with 100 ml of
 51

the P. odorifer cocktail to achieve a target cell density of 7 log CFU/ml. A dry heating block

(Benchmark Dry Bath, Sayreville, NJ) was used to treat milk samples at 63°C (0-30 m), 67°C (0-

20 m), 70°C (0-10 m), and 72°C (0-15 s). All thermal treatments were performed in triplicate.

Spore counts were taken of the cocktail by heat treating 1 ml of the cocktail (undiluted) at 80°C

for 12 m. All samples were stored at 4°C until treated. A thermocouple (EL-USB-TC-LCD,

ThermoWorks, Salt Lake City, UT) was held in an uninoculated milk sample to monitor

temperature to establish come-up time. Once the uninoculated milk sample reached the target

temperature, the time 0 sample was removed from the heat block to quantify the reduction of P.

odorifer during the come-up time. Heat treated samples were immediately transferred to ice and

bacterial survivors were enumerated using serial dilution and spread plating on TSAYE as

described above.

Enumeration of spores in milk products near the end of shelf life

Twenty milk products that varied in code date expiration (0 d-16 d remaining), milk type

(HTST whole, HTST 2%, HTST 1%, UHT whole, chocolate milk), and processor were

purchased from local retailers and stored at 4 C. Each milk product was analyzed for APC as

well as spore count on the day after purchase and again 7 days later. Samples were also analyzed

for survivors after a heat treatment of 72°C, 15 sec to model the estimated survivors if the

product was reworked. All samples were serial diluted (0.1% PW), spread plated on standard

methods agar (SMA) (Neogen) or on TSAYE, and incubated at 25-30°C for 24 h.

 52

Data analysis

Growth rates were obtained by graphing APC data (Log CFU/ml vs. day) in JMP v14

software (SAS, Cary, N.C.) and applying a line of fit to the exponential growth phase for each

strain. For bi-phasic growth patterns, APC data was separated based on first or second

exponential phase and graphed to obtain growth curves for each exponential phase. LactoScope

data was analyzed using a one-way ANOVA to determine if compositional data varied based on

inoculum for each temperature. A Student s t-test was used as a post-hoc analysis to determine

strain differences.

RESULTS AND DISCUSSION

Behavior of P. odorifer strains in milk products under industry-relevant storage conditions

prior to rework

The growth of individual P. odorifer strains over the course of the 21-day shelf life of

fluid milk products at 4 C and 7 C are shown in Figure 3.1. This information is relevant as the

dairy industry may rework packaged product (fluid and chocolate milk) that has been stored

under refrigerated conditions for 3-21 days prior to incorporating as rework into fresh product

(Rush et al. 2021). All four strains were able to grow in both fluid and chocolate milk at both

temperatures. As expected, all P. odorifer strains grew more quickly at 7 C compared to 4 C in

both products. Maximum growth rates over the first 7-10 days at both temperatures were similar

between products for all strains (4 C = 0.39-0.53 log CFU/ml/day; 7 C = 0.83-0.98 log

CFU/ml/day). Previous investigators have monitored the growth of P. odorifer and observed

similar growth rates. Sun et al. (2021) and Buehler et al. (2018) reported similar growth rates of
 53

A) FSL R10-2726 (AT35) (fluid milk source) B) FSL A6-0363 (AT40) (fluid milk source)

P. odorifer (Log CFU/ml)

P. odorifer (Log CFU/ml)

C) FSL E2-0150 (AT2) (chocolate milk source) D) JWC-2503 (AT-novel) (chocolate milk source)
P. odorifer (Log CFU/ml)
P. odorifer (Log CFU/ml)

Figure 3.1. Growth behavior of Paenibacillus odorifer strains (A: FSL R10-2726 (AT35), B: FSL
A6-0363 (AT40), C: FSL E2-0150 (AT2), D: JWC-2503 (AT-novel)) in chocolate and fluid milk
(2%) stored at 4◦C and 7◦C for 21 days. Strains FSL R10-2726 and FSL A6-0363 were originally
isolated from pasteurized fluid milk. Strains FSL E2-0150 and JWC-2503 were originally isolated
from pasteurized chocolate milk. Data points indicate the mean log CFU/ml (n = 3) with error bars
indicating standard error of the mean.
 54

one strain of P. odorifer (AT15) in whole milk (7°C = 1.01 CFU/ml/day) and skim milk broth

(SMB) (6°C = 0.6 CFU/ml/day), respectively. References have reported the minimum growth

temperature for P. odorifer is 5°C (Berge et al., 2002; Priest, 2015); however, Buehler et al.

(2018) used primary data from growth experiments at 6°C to model growth at 4°C using the

approach of Ratkowsky et al. (1983) (Ratkowsky et al., 1983). This approach predicted P.

odorifer AT15 to have a growth rate of 0.4 log CFU/ml/day at 4 C in skim milk broth. Our

results confirm that their predicted growth rate is reasonable; however, it is on the low end of the

range of growth rates presented in Figure 3.1. We believe that this is the first study to report the

actual growth of these P. odorifer strains at 4°C in fluid milk and chocolate milk.

All strains demonstrated a slower growth rate and lower cell density after 21 days of

storage at 4 C in fluid milk (6.83-7.45 log CFU/ml) compared to chocolate milk (7.74-7.83 log

CFU/ml). This difference in growth in fluid and chocolate milk was most obvious for P. odorifer

JWC-2503 (Figure 3.1D). This strain was originally isolated from a spoiled chocolate milk

product and demonstrated improved growth in chocolate milk compared to fluid milk at both

4 C and 7 C. P. odorifer FSL E2-0150 was also originally isolated from chocolate milk;

however, this strain grew equally well in both products (Figure 2.1C). This increased growth rate

and final cell density in chocolate milk may be attributed to the addition of sucrose to this

product as increased growth of P. odorifer has been noted for media supplemented with sucrose

(Priest, 2015). There is little discussion of chocolate milk spoilage in the published literature

with no specific information on the growth of P. odorifer. Douglas et al. (2000) investigated the

spoilage potential of chocolate milk inoculated with raw milk stored at 6°C versus unflavored

milk under the same conditions and reported that the standard plate counts (SPC) reached >6 log

CFU/ml by day 14 while unflavored raw milk remained <3 log CFU/ml, suggesting that
 55

chocolate milk products have a higher spoilage potential than unflavored products (Douglas et

al., 2000). Orleans (2011) also observed faster spoilage rates of commercially available

chocolate milk and reported a lack of a lag phase in comparison to fluid milk growth curves.

They suggested that alanine, an amino acid present in cocoa beans, stimulates the germination of

spores which may explain the rapid initiation of logarithmic growth (Orleans, 2011). The

increased growth rate of some strains of P. odorifer in chocolate milk may contribute to the

reduced shelf life of chocolate milk as compared to fluid milk commonly reported by processors

(personal communication).

Aroma defects in fluid milk were discernable by day 15 (6.55-6.76 log CFU/ml) of

storage at 4°C and by day 6 (6.33-7.12 log CFU/ml) for samples stored at 7°C. There was a

noticeable reduction in the sweet aroma of pasteurized milk and the development of a

solvent across all strains and a fruity aroma present only in samples inoculated with FSL

A6-0363. Changes in the aroma of chocolate milk samples were clear by day 10 (5.8-6.1 log

CFU/ml) for samples stored at 4°C and day 6 (6.4-7.2 log CFU/ml) for samples stored at 7°C and

consisted of an absence of the sweet aroma, and the presence of a burnt chocolate and

solvent aroma across all strains. Samples inoculated with FSL E2-0150 developed a smoky

aroma. These results align with the microbial load typically predicted to lead to observable

spoilage defects in dairy milk (>6 log CFU/ml) (Griffiths et al., 1994; Clark et al., 2009; Martin

et al., 2021). Based on the combination of microbial data and aroma defects, these results

confirm the spoilage potential for P. odorifer in both fluid and chocolate milk.

The chemical quality parameters of fluid milk inoculated with P. odorifer determined

using FT-IR with the LactoScope FT-A revealed interesting differences between incubation

temperatures and individual strains when compared to the uninoculated fluid milk sample (Table
 56

3.1). Lactose showed a statistically significant, but very slight decrease (<0.02%) in inoculated

milk stored at 4°C for 21 days; however, this decrease was more pronounced at 7°C, but still

minor (<0.25%). P. odorifer growth in fluid milk led to a slight decrease in the true protein

(<0.05%) for samples held at 4°C, and an increase (>0.075%) for samples held at 7°C. Non-

protein nitrogen (calculated as urea; NPN/CU) deceased in inoculated milk storage at both 4°C

(>1.5%) and 7°C (2.5-4.9%). Milk urea nitrogen (MUN-2) levels also decreased for both 4°C (2-

3%) and 7° (3-6%). Fatty acid (FA) profiles were also influenced by P. odorifer growth at both

temperatures. Shorter chain FA (C14 or less; de novo) decreased slightly (>0.011%) for all

samples held at 4°C, but significantly increased (>0.093%) for samples held at 7°C. Mixed FA

decreased for all samples held at 4°C and 7°C with the exception of one strain (FSL A6-0363)

held at 7°C that had a large range of variability between samples. Preformed FA increased for all

samples held at 4°C and comparable to slightly decreased for all samples held at 7°C. Generally,

P. odorifer growth in milk at 4 C led to a decrease in the nitrogenous milk components measured

by the LactoScope as well as modifying the absolute and relative concentrations of main fatty

acid groups. For samples stored at 7°C, NPN/CU and MUN-2 decreases were more pronounced,

likely due to higher cell density and more metabolic activity. Relative fatty acid composition

differed between the two incubation temperatures with the concentration of shorter chain fatty

acids being substantially more elevated at the increased temperature.

Genomic research on P. odorifer has revealed that, along with the encoded genes for cold

temperature growth, multiple strains of this organism have been shown to encode for but not

limited to ß-galactosidase (hydrolysis of lactose) (Ivy et al., 2012; Moreno Switt et al., 2014;

Beno et al., 2020), proteolytic systems (hydrolysis of casein) (Moreno Switt et al., 2014),

lipolytic systems (hydrolysis of lipids) (Ivy et al., 2012; Moreno Switt et al., 2014; Priest, 2015),
 57

Table 3.1. Quality parameters of fluid milk (2%) inoculated with Paenibacillus odorifer strains and held at 4 C or 7 C for 21 days. Cell densities
were determined by standard plate count methods on tryptic soy agar + yeast extract (TSAYE) with incubation at 25 C for 48 hours. Chemical
quality parameters were determined using the Lactoscope for Fourier-Transform Infrared Spectroscopy (Lactoscope FT-A, Delta Instruments,
Northvale, NJ). Results are reported as the mean standard error (n = 2-3).
Inoculum Cell density Lactose Nitrogen Fatty Acids
(Log CFU/ml) (anhydrous) True Protein Non-Protein Nitrogen Milk urea nitrogen De novo Mixed Preformed
(Calculated Urea) (MUN-2) (C4:0-C14:0) (Rel%) (C16:0, C16:1, C17:0) (>C18:0)
4 C 7 C 4 C 7 C 4 C 7 C 4 C 7 C 4 C 7 C 4 C 7 C 4 C 7 C 4 C 7 C
Uninoculated N/A 4.641 0.009a 3.298 0.011a 19.205 1.367 a 19.165 1.549 a 0.454 0.013 a 0.719 a
0.776 0.005b
(23.295 0.861 a) (36.900 1.22a) (39.797 0.425c)
FSL R10- 7.4 8.0 4.626 4.404 3.264 3.375 16.907 14.755 16.600 13.895 0.443 0.603 0.669 0.702 0.807 0.770
2726 0.003b 0.003b 0.094 c ab
0.106 c ab b a
0.002 b a
0.003 a ab
c a
(23.080 (29.070 (34.863 (33.825 (42.055 (37.107
0.093a) 0.861ab) 0.109b) 1.22a) b
) 0.425b)
FSL A6-0363 7.5 7.6 4.627 4.466 3.262 3.393 16.960 16.750 16.663 16.145 0.443 0.547 0.670 0.776 0.806 0.755
0.003b 0.003b 0.094 c ab 0.106 c ab 0.002 b b 0.002 b a 0.003 a b
b a
(23.067 (26.350 (34.937 (37.295 (42.008 + (36.353
0.093 a) 0.861b) 0.109b) 1.22a) 0.208b) 0.425b)
FSL E2-0150 7.4 8.1 4.623 4.400 3.262 3.377 17.190 14.330 16.925 13.410 0.443 0.611 0.666 0.706 0.806 0.763
0.003b c b
0.129 c b
0.002 b a
0.003 b a a ab
b
c a
(23.115 (29.385 (34.785 (33.930 (42.100 (36.682
0.114 a) 0.861a) 0.133b) 1.22a) 0.208ab) 0.425b)
JWC-2503 6.8 4.644 4.407 3.247 3.375 17.630 14.517 17.460 13.630 0.428 0.596 0.664 0.692 0.815 0.772
0.003c b b b b 0.002 c a b a 0.004 a a
a c a
(22.445 (28.933 (34.815 (33.593 (42.742 (37.476
0.114b) 0.703ab) 0.133b) 0.997a) 0.208a) 0.345b)
a-c
Data with different superscript lowercase letters within the same column are significantly different (p-value < 0.05).
 58

and nitrogen fixation (Priest, 2015; Beno et al., 2020); thus, the chief metabolism that facilitates

the growth of P. odorifer in milk is not fully understood. Furthermore, Moreno Switt et al.

(2014) suggests that metabolism and enzymatic processes may be dependent on temperature

conditions and that at colder temperatures, P. odorifer is selective in amino acid catabolism. Our

results demonstrated the variance of metabolism for P. odorifer. At 4°C, there was a significant

change in lactose, nitrogen, and FA levels throughout many of the columns; however, there was

most noticeable shift in the nitrogen profiles (NPN/CU and MUN-2). This suggests that P.

odorifer biases nitrogen and other amino acid catabolism at low temperatures around 4°C. We

also found that the Preformed FA increased when compared to the uninoculated sample. At 7°C,

the larger decrease in lactose anhydrase and increase of True Protein suggests stronger ß-

galactosidase and proteolytic activity. Lipolytic activity is also observed through the steep

decrease in Preformed FA with an increase in De Novo FA. Further research is needed to

characterize genomic strain differences that contribute to metabolic processes in refrigerated

milk as well as the development of observable aroma defects.

Final cell densities of P. odorifer in milk products after 21 days of storage at 7 C ranged

from 7.61 to 8.17 log CFU/ml and all strains had been at >7.0 log CFU/ml since day 12 (Figure

3.1). Final cell densities in fluid milk stored at 4 C ranged from 6.83 to 7.83 log CFU/ml after 21

days of storage and appear to just be entering stationary phase at the end of the shelf life. There

was also an observable difference of a lower final cell density at day 21 for two strains. The

mean cell density of JWC-2503 (AT-novel) at 4°C (6.8 0.2 CFU/ml) is lower than all other

strains held at this temperature (7.4-7.5 CFU/ml) while FSL A6-0363 (AT40) is lower at 7°C

(7.6 0.1 CFU/ml) than all the other strains (8.0-8.2 CFU/ml). Previous studies have reported P.

odorifer cell densities of >6 log CFU/ml after 21 days of storage at 6 C (Ivy et al., 2012; Moreno
 59

Switt et al., 2014; Trm i et al., 2015; Buehler et al., 2018). Modeling and Monte Carlo

simulations performed by Buehler et al (2018) predicted significantly less spoilage by P. odorifer

and other psychrotrophic spore formers by reducing the refrigeration temperature from 7 C to

4 C (Buehler et al., 2018). Our results confirm this prediction. Differences in cell density likely

contributed to differences in compositional parameters. For example, JWC-2503 (AT-novel) had

the lowest final cell density (6.8 log CFU/ml) at 4 C which led to smaller changes in NPN and

MUN when compared to strains at higher cell density. Similarly, FSL A6-0363 had the lowest

cell density at 7 C (7.6 log CFU/ml) which led to less extreme changes in NPN and MUN.

A primary goal of these experiments was to determine the maximum cell density of P. odorifer

in fluid and chocolate milk products after 21 days of storage to quantify the potential microbial

burden that these products would contribute when mixed with raw milk for reprocessing. After

21 days of storage P. odorifer cell densities ranged from 6.8 to 7.5 log CFU/ml at 4 C and 7.6 to

8.2 log CFU/ml at 7 C (Figure 1). Previous studies have reported the maximum cell density of P.

odorifer AT15 grown from spores as 6.5 log CFU/ml in skim milk broth incubated at 6 C

(Buehler et al 2018). Ivy et al. (2012) reported that two strains of P. odorifer (AT15 and AT260)

grown from vegetative cells at 6°C in skim milk broth reached >6 log CFU/ml between 8-10

days with a final cell density range between 7-8 log CFU/ml by 21 days. These growth rates may

be lower than our results because of the growth medium used. Skim milk has been reported in

literature to show slower growth of P. odorifer in comparison to whole milk (Sun et al., 2021).

This can partly be explained by referencing our LactoScope data as it appears that lipolytic

metabolism is one nutritional pathway used by P. odorifer at 7°C. These inoculation and storage

studies indicate that fluid and flavored milk held for >5 days at 7 C and >15 days at 4 C could
 60

contribute a high load of bacterial cells (>6 log CFU/ml) when diluted into fresh milk destined

for pasteurization.

Estimating the efficacy of pasteurization to inactivation P. odorifer in milk containing rework

Thermal inactivation experiments were performed to estimate the efficacy of

pasteurization to reduce high cell density populations of P. odorifer that could be present in milk

destined for rework (Figure 3.2). The initial cell density of the milk inoculated with the P.

odorifer cocktail was 7.9 log CFU/ml. As the milk was heated to 63 C, the cocktail was reduced

to 5.4 log CFU/ml (2.5 log reduction) indicating a high sensitivity of P. odorifer to thermal

inactivation when considering HTST pasteurization thermal conditions. After a 5-min holding

time at 63 C, 3.5 log CFU/ml of P. odorifer remained viable; however, no further reduction of P.

odorifer was achieved by increasing the treatment time up to 30 min. Time and temperature

combinations of 63°C for 5-30 min, 70 C for 0-10 min, and 72°C for 15 sec had surviving P.

odorifer cell densities of between 3.0 and 3.1 log CFU/ml. These results indicated the presence

of a subpopulation of thermally resistant cells. The inoculum for these studies was prepared by

incubation of P. odorifer strains in milk at 25 C for 24 hours, so we speculated that the culture

had begun to sporulate. Additional inoculated samples were heat-treated at 80°C for 12 min

which confirmed that 2.0 log CFU/ml of spores were present in the milk sample. Thermal

resistance of Paenibacillus spores has been recorded in literature; however, the emphasis has

been on spores that display excessive thermal resistance. For example, te Giffel et al. (2002)

reported that Paenibacillus spores survived a heat treatments up to 120°C for 30 seconds (te

Giffel et al., 2002). Furthermore, Paenibacillus species have been isolated from UHT-

pasteurized milk, which undergoes a pasteurization temperature of a minimum of 280°C for

P. odorifer Survivors (Log CFU/ml) 61

Control
63◦C 70◦C 72◦C 80◦C
Heat treatment (Temperature (◦C) and Time (min))

Figure 3.2. Thermal inactivation of Paenibacillus odorifer cocktail grown in pasteurized milk at 25◦C for
24 hours prior to treatment. Data are displayed as the mean with error bars representing standard error of
the mean (n = 3). Bars labeled as a treatment time of 0 minutes represent the inactivation due to the
come-up time (CUT) at those temperatures. A treatment of 80◦C for 12 min is the standard thermal
treatment for enumeration of bacterial spores.
 62

2 seconds (Scheldeman et al., 2004; Deeth, 2017). Overall, these results indicate that vegetative

cells of P. odorifer are extremely sensitive to heat and would be easily inactivated by

pasteurization, even at high cell densities, and would not pose an increased spoilage risk for

reworked milk. However, due to the high thermal resistance of P. odorifer spores and perhaps

another thermoduric cell type, it was critical to determine whether these subpopulations exist in

milk held at refrigeration for extended periods of time.

Monitoring the growth rate and development of subpopulations of thermally-resistant P.

odorifer cells in fluid milk during extended refrigerated storage

A second growth study of P. odorifer strains in fluid milk at 4 C was performed to

determine if thermally-resistant subpopulations develop during extended storage (Figure 3.3). All

of the P. odorifer isolates displayed a similar rapid growth rate (0.40-0.52 log CFU/ml/day) in

fluid milk during the first 10 days of storage. However, between days 10 and 15 there was no

growth observed for any of the isolates and the cell density stalled at 5.3-6.2 log CFU/ml. P.

odorifer transitioned into a second slower growth phase (0.09-0.14 log CFU/ml/day) from days

15 to 25 or 30. This growth pattern was not obvious in the previous growth studies at 4 C

reported in Figure 3.1. Further review of the previous data identified a similar pattern for P.

odorifer growth at 7°C in fluid milk (Figure 3.1); however, the stall occurred earlier (between 5

and 9 days) and at a higher cell density (6.3-7.4 log CFU/ml) and the second growth phase was

shorter (3-6 days) and faster (0.15-0.30 log CFU/ml/day) than at 4 C for 21 d and the fluid milk

stored at 4°C for 31 d (Figure 3.2). This growth pattern was not observed in chocolate milk. This

diauxic behavior may be linked to the depletion of a primary carbon or nitrogen source followed
 63

A) FSL R10-2726 (AT35) B) FSL A6-0363 (AT40)

P. odorifer (Log CFU/ml)

P. odorifer (Log CFU/ml)

C) FSL E2-0150 (AT2) D) JWC-2503 (AT-novel)

P. odorifer (Log CFU/ml)

P. odorifer (Log CFU/ml)

Figure 3.3. Growth and sporulation of Paenibacillus odorifer strains (A: FSL R10-2726 (AT35), B: FSL
A6-0363 (AT40), C: FSL E2-0150 (AT2), D: JWC-2503 (AT-novel)) in fluid milk (2%) stored at 4◦C for
31 days. HTST survivors were the subpopulations that survived thermal treatment at 72C for 15 seconds.
Spore count were the subpopulation that survived thermal treatment at 80C for 12 min. Data points
indicate the mean log CFU/ml (n = 3) with error bars indicating standard error of the mean. Dashed line
at 21 days indicates typical end of shelf life and last potential day to use as rework. Detection limit was 0
log CFU/ml.
 64

by the transition and utilization of a secondary nutrition sources (Chu and Barnes, 2016). Based

on our LactoScope results, the nitrogen pathway may actually serve as a valid explanation. For

the 4°C samples, the NPN/CU and MUN-2 decreased the most in comparison to the uninoculated

samples (Table 3.1) which are primarily made up of urea and other nitrogen-based metabolites.

Fromm and Boor (2004) found nitrogen fixation, presented as proteolysis, as a significant change

in Paenibacillus species particularly after 17 days (Fromm and Boor, 2004). However, the

LactoScope data also showed a decrease in protein and fatty acid profiles, and P. odorifer has

been reported to have unique carbohydrate and protein synthesis capabilities (Moreno Switt et

al., 2014); thus, further investigations are recommended on this subject before an confident

explanation can be offered.

For the first 21 days of storage of inoculated milk at 4 C, a thermally-resistant

subpopulation of P. odorifer was not detectable (<0 log CFU/ml) (Figure 3). We continued to

incubate these samples beyond the 21day shelf life to determine if a thermally-resistant

subpopulation could develop under refrigeration conditions. Thermally-resistant subpopulations

(72 C, 15 sec) were first detected on day 26; however, their detection was intermittent across

replicates suggesting a recent change. P. odorifer FSL R10-2726 (AT35) had a resistant

subpopulation of 1.95 log CFU/ml (APC 8.00 log CFU/ml) in one sample replicate, whereas

JWC-2503 (AT-novel) and FSL A6-0363 had low levels of resistant cells (0.3-2.4 log CFU/ml)

in two replicates. With continued storage (day 31), the thermally-resistant subpopulations of P.

odorifer were detected in all milk samples inoculated with FSL R10-2726 (AT35), JWC-2503

(AT-novel), and FSL A6-0363 (AT40) ranging between 0.30-3.09 log CFU/ml, 1.83-3.12 log

CFU/ml, and 1.62-2.44 log CFU/ml, respectively. Only one inoculated milk sample (one

replicate of FSL A6-0363 (AT40)) produced a single colony (0.3 log CFU/ml) on a spore plate
 65

on day 26. These results pose the question if the recommended spore heat treatment thermally

inactivated the spores of P. odorifer. Sun et al. (2021) discussed this possibility and opted to

thermally treat P. odorifer spores at 63°C for 30 minutes. Further research should verify that the

existing spore count temperatures do not underestimate the concentration of P. odorifer in raw

milk and other dairy samples.

P. odorifer FSL E2-1050 (AT2) did not develop a thermally-resistant subpopulation.

Review of the data suggests that three strains that developed resistant subpopulations (FSL R10-

2726 (AT35), JWC-2503 (AT-novel), FSL A6-0363 (AT40)) were approaching their maximum

cell density (stationary phase), whereas FSL E2-1050 (AT2) was still actively replicating (Figure

2.2). Sporulation is an adaptive response that is caused by unfavorable conditions. Chief triggers

for sporulation include nutrient loss, particularly starvation, and environmental factors

(temperature, water activity, etc.) (Gauvry et al., 2017). The sporogenesis has not been

investigated for P. odorifer. It is likely that continued incubation at 4 C would have resulted in

this strain developing thermal resistance and that the thermally-resistant subpopulations of all

strains would have continued to increase.

Spore confirmation in commercially purchased fluid milk

Twenty retail products near, at, or after the end of their shelf life (up to 10 days) were

analyzed for APC, spores, and thermoduric counts. APC counts ranged from 1.7-7.1 log

CFU/ml. Spore counts ranged from 1.0-2.3 spores/ml. Counts were essentially identical between

the spore (80°C for 12 min) and endospore (72°C for 15 sec) treatments. Assuming all

contaminants are Paenibacillus species, the cell density had not yet approached maximum cell

density, so it would be unlikely that thermally resistant subpopulations would have formed in
 66

these products. This further supports the concept that P. odorifer would not develop thermally-

resistant subpopulations within the shelf life of fluid milk under typical storage conditions.

CONCLUSION

The overall objective of our research was to evaluate the potential impact of P. odorifer

as a spoilage organism in milk containing rework. This study demonstrated that P. odorifer is

capable of rapid growth and achieves high cell density in both fluid and chocolate milk stored at

4 C and 7 C. We also demonstrated that vegetative cells of P. odorifer are very sensitive to

pasteurization treatments; however, this species can develop significant levels of thermally-

resistant subpopulations. While these thermally-resistant subpopulations would survive

pasteurization, we demonstrate that these populations do not develop in fluid milk samples stored

at 4 C and therefore are unlikely to pose any risk to reworked product.

ACKNOWLEDGMENTS

This study was funded by the BUILD Dairy Program (Logan, Utah) with additional

support from the Arbuthnot Dairy Center at Oregon State University (Corvallis, OR). The

authors would like to thank Dr. Massimo Bionaz (OSU) for his time and facilitation of using the

LactoScope. We would also like to thank the Cornell University Food Safety Laboratory (Ithaca,

NY), specifically Nicole Martin, Ahmed Gaballa, and Renato Orsi, for assistance with rpoB

allele typing of strain JWC-2503. Finally, we would also like to extend our thanks to Alejandro

Torres for his support and contributions in sampling.

REFERENCES

Beno, S.M., R.A. Cheng, R.H. Orsi, D.R. Duncan, X. Guo, J. Kovac, L.M. Carroll, N.H. Martin,

and M. Wiedmann. 2020. Paenibacillus odorifer, the Predominant Paenibacillus Species

Isolated from Milk in the United States, Demonstrates Genetic and Phenotypic
 67

Conservation of Psychrotolerance but Clade-Associated Differences in Nitrogen Metabolic

Pathways. doi:10.1128/mSphere.00739-19.

Berge, O., M.H. Guinebretière, W. Achouak, P. Normand, and T. Heulin. 2002. Paenibacillus

graminis sp. nov. and Paenibacillus odorifer sp. nov., isolated from plant roots, soil and

food. Int. J. Syst. Evol. Microbiol. 52:607–616. doi:10.1099/00207713-52-2-607.

Buehler, A.J., N.H. Martin, K.J. Boor, and M. Wiedmann. 2018. Psychrotolerant spore-former

growth characterization for the development of a dairy spoilage predictive model. J. Dairy

Sci. 101:6964–6981. doi:10.3168/jds.2018-14501.

Buzby, J.C., H.F. Wells, and J. Hyman. 2014. The Estimated Amount, Value, and Calories of

Postharvest Food Losses at the Retail and Consumer Levels in the United States Jean.

Chambers, J. V. 2002. The Microbiology of Raw Milk. 3rd ed. R.K. Robinson, ed. Wiley-

Interscience, Inc.

Chu, D., and D.J. Barnes. 2016. The lag-phase during diauxic growth is a trade-off between fast

adaptation and high growth rate. Sci. Rep. 6:1–15. doi:10.1038/srep25191.

Clark, S., M. Costello, M. Drake, and F. Bodyfelt. 2009. The Sensory Evaluation of Dairy

Products, 2nd Edition.

Coorevits, A., V. De Jonghe, J. Vandroemme, R. Reekmans, J. Heyrman, W. Messens, P. De

Vos, and M. Heyndrickx. 2008. Comparative analysis of the diversity of aerobic spore-

forming bacteria in raw milk from organic and conventional dairy farms. Syst. Appl.

Microbiol. 31:126–140. doi:10.1016/j.syapm.2008.03.002.

Dairy Practices Council. 2005. Guidelines for controlling the quality and use of dairy product

rework. Keyport, NJ.

Deeth, H. 2017. Optimum Thermal Processing for Extended Shelf-Life (ESL) Milk. Foods
 68

6:102. doi:10.3390/foods6110102.

Douglas, S.A., M.J. Gray, A.D. Crandall, and K.J. Boor. 2000. Characterization of chocolate

milk spoilage patterns. J. Food Prot. 63:516–521. doi:10.4315/0362-028X-63.4.516.

Fromm, H.I., and K.J. Boor. 2004. M : Food Microbiology and Safety Characterization of

Pasteurized Fluid Milk Shelf-life Attributes 69.

Gauvry, E., A.G. Mathot, I. Leguérinel, O. Couvert, F. Postollec, V. Broussolle, and L. Coroller.

2017. Knowledge of the physiology of spore-forming bacteria can explain the origin of

spores in the food environment. Res. Microbiol. 168:369–378.

doi:10.1016/j.resmic.2016.10.006.

te Giffel, M.C., A. Wagendorp, A. Herrewegh, and F. Driehuis. 2002. Bacterial spores in silage

and raw milk. Antonie Van Leeuwenhoek 81:625–630. doi:10.1023/A.

Gopal, N., C. Hill, P.R. Ross, T.P. Beresford, M.A. Fenelon, and P.D. Cotter. 2015. The

prevalence and control of Bacillus and related spore-forming bacteria in the dairy industry.

Front. Microbiol. 6:1–18. doi:10.3389/fmicb.2015.01418.

Grady, E.N., J. MacDonald, L. Liu, A. Richman, and Z.C. Yuan. 2016. Current knowledge and

perspectives of Paenibacillus: A review. Microb. Cell Fact. 15:1–18. doi:10.1186/s12934-

016-0603-7.

Griffiths, M.I.W., R.R. Laing, D. Roy, and A.A. Mafu. 1994. Psychrotrophs in Dairy Products:

Their Effects and Their Control.

Huck, J.R., M. Sonnen, and K.J. Boor. 2008. Tracking heat-resistant, cold-thriving fluid milk

spoilage bacteria from farm to packaged product. J. Dairy Sci. 91:1218–1228.

doi:10.3168/jds.2007-0697.

Ivy, R.A., M.L. Ranieri, N.H. Martin, H.C. den Bakker, B.M. Xavier, M. Wiedmann, and K.J.
 69

Boor. 2012. Identification and characterization of psychrotolerant sporeformers associated

with fluid milk production and processing. Appl. Environ. Microbiol. 78:1853–1864.

doi:10.1128/AEM.06536-11.

Martin, N.H., K.J. Boor, and M. Wiedmann. 2018. Symposium review: Effect of post-

pasteurization contamination on fluid milk quality. J. Dairy Sci. 101:861–870.

doi:10.3168/jds.2017-13339.

Martin, N.H., P. Torres-Frenzel, and M. Wiedmann. 2021. Invited review: Controlling dairy

product spoilage to reduce food loss and waste. J. Dairy Sci. 104:1251–1261.

doi:10.3168/jds.2020-19130.

Meer, R.R., J. Baker, F.W. Bodyfelt, and M.W. Griffiths. 1991. Psych rotrophic Bacillus spp. in

fluid milk products: A review. J. Food Prot. 54:969–979. doi:10.4315/0362-028X-

54.12.969.

Moreno Switt, A.I., A.D. Andrus, M.L. Ranieri, R.H. Orsi, R. Ivy, H.C. den Bakker, N.H.

Martin, M. Wiedmann, and K.J. Boor. 2014. Genomic comparison of sporeforming bacilli

isolated from milk. BMC Genomics 15:7–11. doi:10.1186/1471-2164-15-26.

Orleans, K.A. 2011. Microbiological and Chemical Changes During Shelf-life in Regular and

Chocolate Milk. Ohio State University,.

Priest, F.G. 2015. Paenibacillus. Bergey s Man. Syst. Archaea Bact. 1–40.

doi:10.1002/9781118960608.gbm00553.

Ranieri, M.L., and K.J. Boor. 2009. Short communication: Bacterial ecology of high-

temperature, short-time pasteurized milk processed in the united States. J. Dairy Sci.

92:4833–4840. doi:10.3168/jds.2009-2181.

Ranieri, M.L., J.R. Huck, M. Sonnen, D.M. Barbano, and K.J. Boor. 2009. High temperature,
 70

short time pasteurization temperatures inversely affect bacterial numbers during refrigerated

storage of pasteurized fluid milk. J. Dairy Sci. 92:4823–4832. doi:10.3168/jds.2009-2144.

Ratkowsky, D.A., R.K. Lowry, T.A. McMeekin, A.N. Stokes, and R.E. Chandler. 1983. Model

for bacterial culture growth rate throughout the entire biokinetic temperature range. J.

Bacteriol. 154:1222–1226. doi:10.1128/jb.154.3.1222-1226.1983.

Reichler, S.J., A. Trm i , N.H. Martin, K.J. Boor, and M. Wiedmann. 2018. Pseudomonas

fluorescens group bacterial strains are responsible for repeat and sporadic postpasteurization

contamination and reduced fluid milk shelf life. J. Dairy Sci. 101:7780–7800.

doi:10.3168/jds.2018-14438.

Scheldeman, P., K. Goossens, M. Rodriguez-Diaz, A. Pil, J. Goris, L. Herman, P. De Vos, N.A.

Logan, and M. Heyndrickx. 2004. Paenibacillus lactis sp. nov., isolated from raw and heat-

treated milk. Int. J. Syst. Evol. Microbiol. 54:885–891. doi:10.1099/ijs.0.02822-0.

Sun, L., K. Atkinson, M. Zhu, and D.J. D Amico. 2021. Antimicrobial effects of a bioactive

glycolipid on spore-forming spoilage bacteria in milk. J. Dairy Sci. 104:4002–4011.

doi:10.3168/jds.2020-19769.

Trm i , A., N.H. Martin, K.J. Boor, and M. Wiedmann. 2015. A standard bacterial isolate set for

research on contemporary dairy spoilage. J. Dairy Sci. 98:5806–5817.

doi:10.3168/jds.2015-9490.

Vissers, M., F. Driehuis, M.C. Te Giffel, P. De Jong, and J.M.G. Lankveld. 2007. Short

communication: Quantification of the transmission of microorganisms to milk via dirt

attached to the exterior of teats. J. Dairy Sci. 90:3579–3582. doi:10.3168/jds.2006-633.

 71

CHAPTER 4: CONCLUSION

The first objective of this study was to understand rework in the fluid dairy processing

industry including how processors define rework, their handling practices, and their decision-

making processes. The goal was to then utilize the survey results to design experiments that

would explore the spoilage potential of reworking fluid dairy milk using P. odorifer as a target.

The results of the survey revealed nine major rework motivations, varied storage and processing

practices, and diverse parameters for the eligibility of products that can be used as rework as well

as those that can receive rework. Most importantly, this survey communicated that fluid milk

processors are lacking confidence in their decisions for when to rework or discard these product

streams. In-house studies were performed by a few of the surveyed processors to guide their

current practices including a modification of their rework dilution rates and shortening the

allowable age of the product that can be reworked. Other processors choose to follow the PMO,

which only offers vague guidelines on the eligibility for repasteurizing products to meet safety

expectations. Furthermore, the regulatory information available to dairy processors for reworking

their products is unclear. For example, rework is not defined in the PMO and is discussed as a

repasteurization/reprocess parameter while the CFR defines rework using the term,

“reconditioning”, which refutes some eligibility requirements listed in the PMO based on the

PMO’s language use of “reprocessing”. This may explain why processors are lacking confidence

and are hesitant to share their practices with others in the industry.

The second objective of this study was to evaluate the spoilage potential of P. odorifer in

reworked fluid and chocolate flavored milk. P. odorifer has received a lot of attention in dairy

spoilage investigations over the last ten years because it is a heat-resistant spore former that can
 72

survive pasteurization temperatures and is often found at high cell density at the end of shelf life

in HTST-pasteurized fluid milk. Our results show that P. odorifer vegetative cells are easily

inactivated at HTST-pasteurization temperatures and do not sporulate in fluid milk held at 4°C

within the typical shelf life of HTST-pasteurized fluid milk (21 days). Because processors are

required to rework their products within their printed code date, which is typically 14-21 days for

HTST-pasteurized fluid milk, we have demonstrated that P. odorifer will not cause premature

spoilage in reworked fluid milk products. We also found that P. odorifer spores may be sensitive

to traditional spore heat treatments of 80°C for 12 minutes. Previous studies report that P.

odorifer has low to non-existent spore counts in raw milk; however, we demonstrated that a

subpopulation of P. odorifer cells resisted typical pasteurization thermal treatments (72°C for 15

seconds), but were inactivated at thermal spore treatments. This suggests that the enumeration of

P. odorifer spores in raw milk may have been underestimated when using the standard spore

count method. Further investigations into the sporulation, thermal resistance and behavior of P.

odorifer in milk would be valuable information to the industry.

 73

BIBLIOGRAPHY

Agriculture Marketing Service. 2013. Instructions for Dairy Plant Surveys. Washington, DC.

Ahmad, T., R.M. Aadil, H. Ahmed, U. ur Rahman, B.C.V. Soares, S.L.Q. Souza, T.C. Pimentel,

H. Scudino, J.T. Guimarães, E.A. Esmerino, M.Q. Freitas, R.B. Almada, S.M.R.

Vendramel, M.C. Silva, and A.G. Cruz. 2019. Treatment and utilization of dairy industrial

waste: A review. Trends Food Sci. Technol. 88:361–372. doi:10.1016/j.tifs.2019.04.003.

Aikenhead, G., K. Farahbakhsh, J. Halbe, and J. Adamowski. 2015. Application of process

mapping and causal loop diagramming to enhance engagement in pollution prevention in

small to medium size enterprises: Case study of a dairy processing facility. J. Clean. Prod.

102:275–284. doi:10.1016/j.jclepro.2015.04.069.

Barbano, D.M. 2017. A 100-Year Review: The production of fluid (market) milk. J. Dairy Sci.

100:9894–9902. doi:10.3168/jds.2017-13561.

Beno, S.M., R.A. Cheng, R.H. Orsi, D.R. Duncan, X. Guo, J. Kovac, L.M. Carroll, N.H. Martin,

and M. Wiedmann. 2020. Paenibacillus odorifer, the Predominant Paenibacillus Species

Isolated from Milk in the United States, Demonstrates Genetic and Phenotypic

Conservation of Psychrotolerance but Clade-Associated Differences in Nitrogen Metabolic

Pathways. doi:10.1128/mSphere.00739-19.

Berge, O., M.H. Guinebretière, W. Achouak, P. Normand, and T. Heulin. 2002. Paenibacillus

graminis sp. nov. and Paenibacillus odorifer sp. nov., isolated from plant roots, soil and

food. Int. J. Syst. Evol. Microbiol. 52:607–616. doi:10.1099/00207713-52-2-607.

Buehler, A.J., N.H. Martin, K.J. Boor, and M. Wiedmann. 2018. Psychrotolerant spore-former

growth characterization for the development of a dairy spoilage predictive model. J. Dairy
 74

Sci. 101:6964–6981. doi:10.3168/jds.2018-14501.

Buzby, J.C., H.F. Wells, and J. Hyman. 2014. The Estimated Amount, Value, and Calories of

Postharvest Food Losses at the Retail and Consumer Levels in the United States Jean.

Center for Food Safety and Applied Nutrition. 2007. Hazards and Controls Guide For Dairy

Foods HACCP. Washington, DC.

Černíková, M., J. Nebesá ová, R.N. Salek, R. Popková, and F. Buňka. 2018. The effect of

rework content addition on the microstructure and viscoelastic properties of processed

cheese. J. Dairy Sci. 101:2956–2962. doi:10.3168/jds.2017-13742.

Chambers, J. V. 2002. The Microbiology of Raw Milk. 3rd ed. R.K. Robinson, ed. Wiley-

Interscience, Inc.

Chu, D., and D.J. Barnes. 2016. The lag-phase during diauxic growth is a trade-off between fast

adaptation and high growth rate. Sci. Rep. 6:1–15. doi:10.1038/srep25191.

Cigerdelen, E. 2011. Integrated scheduling of ice cream production with rework via mixing.

Eindhoven University of Technology,.

Clark, S., M. Costello, M. Drake, and F. Bodyfelt. 2009. The Sensory Evaluation of Dairy

Products, 2nd Edition.

Coorevits, A., V. De Jonghe, J. Vandroemme, R. Reekmans, J. Heyrman, W. Messens, P. De

Vos, and M. Heyndrickx. 2008. Comparative analysis of the diversity of aerobic spore-

forming bacteria in raw milk from organic and conventional dairy farms. Syst. Appl.

Microbiol. 31:126–140. doi:10.1016/j.syapm.2008.03.002.

Dairy Practices Council. 2005. Guidelines for controlling the quality and use of dairy product

rework. Keyport, NJ.

Deeth, H. 2017. Optimum Thermal Processing for Extended Shelf-Life (ESL) Milk. Foods
 75

6:102. doi:10.3390/foods6110102.

Dolby, R.M. 1965. Changes in moisture distribution caused by partial reworking of butter shortly

after churning. J. Dairy Res. 32:263–267. doi:10.1017/S002202990001863X.

Dou, Z., J.D. Ferguson, D.T. Galligan, A.M. Kelly, S.M. Finn, and R. Giegengack. 2016.

Assessing U.S. food wastage and opportunities for reduction. Glob. Food Sec. 8:19–26.

doi:10.1016/j.gfs.2016.02.001.

Douglas, S.A., M.J. Gray, A.D. Crandall, and K.J. Boor. 2000. Characterization of chocolate

milk spoilage patterns. J. Food Prot. 63:516–521. doi:10.4315/0362-028X-63.4.516.

Economic Research Service. 2020. Dairy Data. Washington, DC.

Fromm, H.I., and K.J. Boor. 2004. M : Food Microbiology and Safety Characterization of

Pasteurized Fluid Milk Shelf-life Attributes 69.

Gauvry, E., A.G. Mathot, I. Leguérinel, O. Couvert, F. Postollec, V. Broussolle, and L. Coroller.

2017. Knowledge of the physiology of spore-forming bacteria can explain the origin of

spores in the food environment. Res. Microbiol. 168:369–378.

doi:10.1016/j.resmic.2016.10.006.

te Giffel, M.C., A. Wagendorp, A. Herrewegh, and F. Driehuis. 2002. Bacterial spores in silage

and raw milk. Antonie Van Leeuwenhoek 81:625–630. doi:10.1023/A.

Gopal, N., C. Hill, P.R. Ross, T.P. Beresford, M.A. Fenelon, and P.D. Cotter. 2015. The

prevalence and control of Bacillus and related spore-forming bacteria in the dairy industry.

Front. Microbiol. 6:1–18. doi:10.3389/fmicb.2015.01418.

Grady, E.N., J. MacDonald, L. Liu, A. Richman, and Z.C. Yuan. 2016. Current knowledge and

perspectives of Paenibacillus: A review. Microb. Cell Fact. 15:1–18. doi:10.1186/s12934-

016-0603-7.
 76

Griffiths, M.I.W., R.R. Laing, D. Roy, and A.A. Mafu. 1994. Psychrotrophs in Dairy Products:

Their Effects and Their Control.

Holm, S., R.B. Toma, W. Reiboldt, C. Newcomer, and M. Calicchia. 2002. Cleaning frequency

and the microbial load in ice-cream. Int. J. Food Sci. Nutr. 53:337–342.

doi:10.1080/09637480220138133.

Huck, J.R., M. Sonnen, and K.J. Boor. 2008. Tracking heat-resistant, cold-thriving fluid milk

spoilage bacteria from farm to packaged product. J. Dairy Sci. 91:1218–1228.

doi:10.3168/jds.2007-0697.

ICMSF. 2005. Microbial Ecology of Food Commodities. Second. Kluwer Academic/Plenum

Publishers, New York.

IRI Custom DMI Market Advantage Database. 2021. Total U.S. - MULO Retail YTD Milk

Snapshot.

Ivy, R.A., M.L. Ranieri, N.H. Martin, H.C. den Bakker, B.M. Xavier, M. Wiedmann, and K.J.

Boor. 2012. Identification and characterization of psychrotolerant sporeformers associated

with fluid milk production and processing. Appl. Environ. Microbiol. 78:1853–1864.

doi:10.1128/AEM.06536-11.

Kolev, S. 2017. General Characteristics and Treatment Possibilities of Dairy Wastewater – A

Review. Food Technol. Biotechnol. 53:237–242.

Long, H.F. 1939. Bacteriology of butter VII . Effect of reworking butter on growth of bacteria

Bacteriology of Butter 24.

Martin, N.H., K.J. Boor, and M. Wiedmann. 2018. Symposium review: Effect of post-

pasteurization contamination on fluid milk quality. J. Dairy Sci. 101:861–870.

doi:10.3168/jds.2017-13339.
 77

Martin, N.H., D.J. Kent, R.L. Evanowski, T.J. Zuber Hrobuchak, and M. Wiedmann. 2019.

Bacterial spore levels in bulk tank raw milk are influenced by environmental and cow

hygiene factors. J. Dairy Sci. 102:9689–9701. doi:10.3168/jds.2019-16304.

Martin, N.H., P. Torres-Frenzel, and M. Wiedmann. 2021. Invited review: Controlling dairy

product spoilage to reduce food loss and waste. J. Dairy Sci. 104:1251–1261.

doi:10.3168/jds.2020-19130.

Meer, R.R., J. Baker, F.W. Bodyfelt, and M.W. Griffiths. 1991. Psych rotrophic Bacillus spp. in

fluid milk products: A review. J. Food Prot. 54:969–979. doi:10.4315/0362-028X-

54.12.969.

Milani, F.X., D. Nutter, and G. Thoma. 2011. Invited review: Environmental impacts of dairy

processing and products: A review. J. Dairy Sci. 94:4243–4254. doi:10.3168/jds.2010-3955.

Moreno Switt, A.I., A.D. Andrus, M.L. Ranieri, R.H. Orsi, R. Ivy, H.C. den Bakker, N.H.

Martin, M. Wiedmann, and K.J. Boor. 2014. Genomic comparison of sporeforming bacilli

isolated from milk. BMC Genomics 15:7–11. doi:10.1186/1471-2164-15-26.

National Agriculture Statistics Service. 2021. Milk Production and All Milk Prices. Washington,

DC.

Ohkubo, Y., K. Uchida, H. Motoshima, and N. Katano. 2019. Simple membrane filtration

method for estimating numbers of Paenibacillus spp. spores in raw milk, using -

galactosidase activity as a selection criterion. Int. Dairy J. 96:10–14.

doi:10.1016/j.idairyj.2019.04.003.

Orleans, K.A. 2011. Microbiological and Chemical Changes During Shelf-life in Regular and

Chocolate Milk. Ohio State University,.

Parfitt, J., M. Barthel, and S. MacNaughton. 2010. Food waste within food supply chains:
 78

Quantification and potential for change to 2050. Philos. Trans. R. Soc. B Biol. Sci.

365:3065–3081. doi:10.1098/rstb.2010.0126.

Patra, F., and R.K. Duary. 2020. Waste from Dairy Processing Industries and its Sustainable

Utilization. 1st ed. M. Thakur, V.K. Modi, R. Khedkar, and K. Singh, ed. Springer Nature

Singapore Pte Ltd., Singapore.

Priest, F.G. 2015. Paenibacillus. Bergey's Manual of Systematics of Archaea and Bacteria.

Edinburgh.

Ranieri, M.L., and K.J. Boor. 2009. Short communication: Bacterial ecology of high-

temperature, short-time pasteurized milk processed in the united States. J. Dairy Sci.

92:4833–4840. doi:10.3168/jds.2009-2181.

Ranieri, M.L., J.R. Huck, M. Sonnen, D.M. Barbano, and K.J. Boor. 2009a. High temperature,

short time pasteurization temperatures inversely affect bacterial numbers during refrigerated

storage of pasteurized fluid milk. J. Dairy Sci. 92:4823–4832. doi:10.3168/jds.2009-2144.

Ranieri, M.L., J.R. Huck, M. Sonnen, D.M. Barbano, and K.J. Boor. 2009b. High temperature,

short time pasteurization temperatures inversely affect bacterial numbers during refrigerated

storage of pasteurized fluid milk. J. Dairy Sci. 92:4823–4832. doi:10.3168/jds.2009-2144.

Ratkowsky, D.A., R.K. Lowry, T.A. McMeekin, A.N. Stokes, and R.E. Chandler. 1983. Model

for bacterial culture growth rate throughout the entire biokinetic temperature range. J.

Bacteriol. 154:1222–1226. doi:10.1128/jb.154.3.1222-1226.1983.

Reichler, S.J., A. Trmčić, N.H. Martin, K.J. Boor, and M. Wiedmann. 2018. Pseudomonas

fluorescens group bacterial strains are responsible for repeat and sporadic postpasteurization

contamination and reduced fluid milk shelf life. J. Dairy Sci. 101:7780–7800.

doi:10.3168/jds.2018-14438.
 79

Scheldeman, P., K. Goossens, M. Rodriguez-Diaz, A. Pil, J. Goris, L. Herman, P. De Vos, N.A.

Logan, and M. Heyndrickx. 2004. Paenibacillus lactis sp. nov., isolated from raw and heat-

treated milk. Int. J. Syst. Evol. Microbiol. 54:885–891. doi:10.1099/ijs.0.02822-0.

Sun, L., K. Atkinson, M. Zhu, and D.J. D’Amico. 2021. Antimicrobial effects of a bioactive

glycolipid on spore-forming spoilage bacteria in milk. J. Dairy Sci. 104:4002–4011.

doi:10.3168/jds.2020-19769.

Tasneem, M., F. Siddique, A. Ahmad, and U. Farooq. 2014. Stabilizers: Indispensable

Substances in Dairy Products of High Rheology. Crit. Rev. Food Sci. Nutr. 54:869–879.

doi:10.1080/10408398.2011.614702.

Tobe, F. 2016. Minimizing Production Losses in the Food and Beverage Industry. Accessed.

Tóth, K., C. Borbély, B. Nagy, G. Szabó-Szentgróti, and E. Szabó-Szentgróti. 2014.

Measurement of Food Losses in a Hungarian Dairy Processing Plant. Edible oil fat Refin.

1–10.

Trmčić, A., N.H. Martin, K.J. Boor, and M. Wiedmann. 2015. A standard bacterial isolate set for

research on contemporary dairy spoilage. J. Dairy Sci. 98:5806–5817.

doi:10.3168/jds.2015-9490.

United States Code of Federal Regulations. 2021. Title 21 Food and Drugs: CFR Part 117.3,

Definitions.

United States Department of Agriculture. 2009. Instructions for dairy inspection and grading

service. Washington, DC.

United States Environmental Protection Agency. 2020. 2018 Wasted Food Report 42.

United States Public Health Service/Food and Drug. 2017. Grade “A” Pasteurized Milk

Ordinance (PMO). Washington, DC.

 80

Vissers, M., and F. Driehuis. 2008. On-Farm Hygienic Milk Production. A. Tamime, ed. Wiley-

Blackwell.

Vissers, M., F. Driehuis, M.C. Te Giffel, P. De Jong, and J.M.G. Lankveld. 2007. Short

communication: Quantification of the transmission of microorganisms to milk via dirt

attached to the exterior of teats. J. Dairy Sci. 90:3579–3582. doi:10.3168/jds.2006-633.

Ziyaina, M., B.N. Govindan, B. Rasco, T. Coffey, and S.S. Sablani. 2018. Monitoring Shelf Life

of Pasteurized Whole Milk Under Refrigerated Storage Conditions: Predictive Models for

Quality Loss. J. Food Sci. 83:409–418. doi:10.1111/1750-3841.13981.