Cost Savings Realized by Implementation of Routine Microbiological

Identification by Matrix-Assisted Laser Desorption Ionization–Time of

Flight Mass Spectrometry
Anthony Tran,a* Kevin Alby,a* Alan Kerr,a Melissa Jones,a Peter H. Gilligana,b
Clinical Microbiology-Immunology Laboratories, University of North Carolina Health Care,a and Departments of Pathology-Laboratory Medicine and Microbiology-
Immunology,b Chapel Hill, North Carolina, USA

Matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry (MS) is an emerging technology for rapid
identification of bacterial and fungal isolates. In comparison to conventional methods, this technology is much less labor intensive and
can provide accurate and reliable results in minutes from a single isolated colony. We compared the cost of performing the bioMérieux
Vitek MALDI-TOF MS with conventional microbiological methods to determine the amount saved by the laboratory by converting to
the new technology. Identification costs for 21,930 isolates collected between April 1, 2013, and March 31, 2014, were directly compared
for MALDI-TOF MS and conventional methodologies. These isolates were composed of commonly isolated organisms, including com-
monly encountered aerobic and facultative bacteria and yeast but excluding anaerobes and filamentous fungi. Mycobacterium tuber-
culosis complex and rapidly growing mycobacteria were also evaluated for a 5-month period during the study. Reagent costs and
a total cost analysis that included technologist time in addition to reagent expenses and maintenance service agreement costs
were analyzed as part of this study. The use of MALDI-TOF MS equated to a net savings of $69,108.61, or 87.8%, in reagent costs
annually compared to traditional methods. When total costs are calculated to include technologist time and maintenance costs,
traditional identification would have cost $142,532.69, versus $68,886.51 with the MALDI-TOF MS method, resulting in a labo-
ratory savings of $73,646.18, or 51.7%, annually by adopting the new technology. The initial cost of the instrument at our usage
level would be offset in about 3 years. MALDI-TOF MS not only represents an innovative technology for the rapid and accurate
identification of bacterial and fungal isolates, it also provides a significant cost savings for the laboratory.

M ass spectrometry (MS) has been traditionally utilized for

chemical analysis, although its utility was limited to low-
molecular-weight organic compounds (1). Matrix-assisted laser
and hardware maintenance/repair, reducing the time the instru-
ment is available on a given day for identification of isolates. Suf-
ficient protocols need to be in place for times when the MALDI-
desorption ionization–time of flight (MALDI-TOF) MS ex- TOF MS will be unavailable. Although the initial instrumentation
panded the field and allowed for the analysis of biological mole- price is high and maintenance expenses are significant, the cost of
cules with no theoretical upper limit of mass (2). Previously em- identifying an isolate can be very low. Based on this, use of
ployed to determine the mass of peptides and proteins, this MALDI-TOF MS has the potential to improve laboratory effi-
emerging technology has been adapted for rapid identification of ciency, reduce turnaround times, and lower costs (9).
bacterial and fungal isolates in the clinical microbiology labora- We compared the cost of performing the bioMérieux Vitek
tory. In an increasing number of settings, MALDI-TOF MS has MALDI-TOF MS (Durham, NC, USA) with that of conventional
replaced traditional identification methods, including micros- microbiological methods to determine the amount saved by the
copy and determination of phenotypic characteristics, which typ- laboratory after converting to the new technology. This study ex-
ically require multiple steps (3, 4).
The performance of MALDI-TOF MS for the identification of
microorganisms was examined in numerous studies and shown to Received 27 March 2015 Returned for modification 18 April 2015
be accurate and reliable (5–8). In comparison to conventional Accepted 15 May 2015
methods, this technology is much less labor intensive and can Accepted manuscript posted online 20 May 2015
provide accurate and reliable results in minutes from a single iso- Citation Tran A, Alby K, Kerr A, Jones M, Gilligan PH. 2015. Cost savings realized by
lated colony. The only reagents utilized are the target slides that implementation of routine microbiological identification by matrix-assisted laser
contain sample spots used for the identification of microorgan- desorption ionization–time of flight mass spectrometry. J Clin Microbiol
53:2473–2479. doi:10.1128/JCM.00833-15.
isms, an organic matrix solution, formic acid for yeast isolates,
Editor: E. Munson
pipette tips, and disposable loops or toothpicks for sample appli-
Address correspondence to Anthony Tran, atran@health.nyc.gov, or Peter H.
cation. Therefore, reagent rental programs are usually not avail- Gilligan, peter.gilligan@unchealth.unc.edu.
able from the two manufacturers of MALDI-TOF MS, so the in- * Present address: Anthony Tran, Public Health Laboratory, New York City
strument is purchased, which is a large capital cost to the Department of Health and Mental Hygiene, New York, New York, USA; Kevin Alby,
laboratory. In addition, fairly expensive annual maintenance con- Department of Pathology and Lab Medicine, Perelman School of Medicine,
tracts of $25,000 to $30,000 are needed to ensure limited down- University of Pennsylvania, Philadelphia, Pennsylvania, USA.
time of this technically complex equipment. Causes of downtime Copyright © 2015, American Society for Microbiology. All Rights Reserved.
are multifaceted and consist of various issues, including time re- doi:10.1128/JCM.00833-15
quired for remote access to software for manipulation or tuning,

August 2015 Volume 53 Number 8 Journal of Clinical Microbiology jcm.asm.org 2473

Tran et al.

amined the cost savings encountered when reagents and total cost
were calculated.

MATERIALS AND METHODS

Study design. We conducted a 12-month retrospective analysis of poten-
tial cost savings incurred by the Clinical Microbiology and Immunology
Laboratory (CMIL) at the University of North Carolina Hospitals
(UNCH) after implementation of MALDI-TOF MS for routine identifi-
cation of bacteria and yeasts. UNCH is an 803-bed academic medical
center composed of acute care, cancer, children’s, neuroscience, and
women’s hospitals located in Chapel Hill, North Carolina. Identification
costs for 21,930 isolates collected during a 1-year period between April 1,
2013, and March 31, 2014, associated with MALDI-TOF MS and conven-
tional methodologies were directly compared. Specimens consisted of the
most commonly isolated organisms in the clinical microbiology labora-
tory, including Enterobacteriaceae, enterococci, Gram-negative glucose
nonfermenters (GNFs), staphylococci, streptococci, and yeast. A compre-
hensive and structured review of the laboratory information system (SCC
Soft Computer, Clearwater, FL) was conducted to determine the number
of isolates identified during the study period.
Analysis of costs. Three cost calculations were analyzed as part of this
FIG 1 Traditional methods required for identification, by organism category.
study: reagent cost, a direct cost analysis that included technologist time in
GNF, Gram-negative glucose nonfermenters.
addition to reagent expenses, and a total cost analysis that included re-
agent, technologist, and maintenance costs. Reagent expenditures associ-
ated with traditional microbiological methods and MALDI-TOF MS were
directly calculated for each specimen in the study. A standardized formula ing, and Vitek 2 bacterial identification cards, which significantly cut re-
was used to determine total technologist cost: technologist cost ⫽ (iden- agent and quality control costs. The UNCH CMIL mean base technologist
tification time per isolate [minutes]) ⫻ (number of isolates identified) ⫻ salary utilized for this study was $27.00/hour, with an additional 20% for
(mean base technologist salary [per minute]). This was then added to fringe benefits, totaling $0.54/minute.
reagent costs to derive the direct cost: direct cost ⫽ technologist cost ⫹ The UNCH CMIL laboratory staff went through extensive training
reagent cost. Lastly, we calculated the total cost with the addition of the and internal performance monitoring to properly operate the bioMérieux
maintenance service agreement contracts to provide a full picture of the Vitek MS and enhance workflow. This included exercises on how to apply
total savings incurred over the life of the instrument: total cost ⫽ direct the proper amount of colony growth onto the disposable target slides (i.e.,
cost ⫹ maintenance cost/isolate. spotting) to achieve a quality result, troubleshoot identification issues as
Since the maintenance contract was a fixed rather than a variable cost, they arise, and know when to reanalyze (i.e., refire) a spot. The technolo-
we determined the maintenance cost based on our actual maintenance gists were taught to apply one colony per isolate onto a spot on the target
contract amount for fiscal year 2014 and divided that by the total number slide, add 1 ␮l of the matrix solution, and allow it to dry. We urge one
of isolates identified. Our maintenance cost for the instrument and soft- sample spot per isolate, no duplicates. An exception would be if various
ware for fiscal year 2014 was $29,700, which included the annual service colony morphologies exist on a plate and the technologist wants to rule
agreement covering the instrument hardware and Myla software. out another organism, then other morphologies would be sampled. For
Identification time was calculated based on prior experiences with yeast isolates, 0.5 ␮l of formic acid was applied to the sample and allowed
traditional and MALDI-TOF MS methods. Traditional microbiological to dry prior to addition of the organic matrix solution. For workflow, each
methods utilized by the UNCH CMIL include the Vitek 2 microbial iden- technologist evaluated his or her culture, and suspicious colonies were
tification system (bioMérieux, Durham, NC) for identification of Enter- marked and put aside for testing on MALDI-TOF MS. When a sufficient
obacteriaceae. Escherichia coli isolates were an exception, as identification batch of cultures is ready to be analyzed, the technologists each spotted
was based mainly on colony morphology and spot indole (Remel, Lenexa, their own isolates, since they were the most familiar with the culture they
KS). Vitek 2, pyrrolidonyl arylamidase (PYR) disks, 6.5% NaCl broth, and just evaluated. The bioMérieux Vitek MALDI-TOF MS utilizes disposable
bile esculin agar (Remel, Lenexa, KS) were utilized for identification of target slides with three acquisition groups (AGs) per slide. Each acquisi-
enterococci, while oxidase (Remel), Vitek 2, and molecular sequencing tion group has a maximum of 16 spots for identification of isolates, for a
were used for detection of GNFs. Staphylococci and streptococci were total of 48 sample spots per disposable target slide. Staff were encouraged
identified with various agglutination and presumptive (e.g., CAMP, A to utilize as much of the AG as possible to maximize slide usage. Once an
disk, P disk) tests. API strips (bioMérieux), CHROMagar Candida media AG was filled, the slide was inserted into the MALDI-TOF MS for analysis.
(Becton, Dickinson and Company, Franklin Lakes, NJ), and rapid assim- Analysis takes only minutes, and the result is usually an organism
ilation of trehalose (RAT) testing (Remel) for Candida glabrata were used identification that is reported out in our laboratory information system.
for yeast identifications (Fig. 1). However, there are a number of reasons that a reportable identification
For these traditional methods, we provided an average of 6 minutes of may not be generated. The first is due to a failure, where no identification
technologist time for a positive identification of all bacterial organisms is generated. Spots prepared too thickly or too thinly account for the
except staphylococci. Staphylococcus aureus were traditionally identified largest number of identification failures, underscoring the importance of
via BactiStaph (Remel, Lenexa, KS), while other Staphylococcus spp. were proper training. Further reasons for failures include analysis of isolates
detected with a combination of a tube coagulase test and BactiStaph, both that are not included in the current MALDI-TOF MS database and are
simple agglutination methodologies, and were therefore assigned 3 min- ultimately identified by 16S rRNA gene sequencing, isolates that fail to
utes per detection. Yeast isolates were allotted between 3 and 20 minutes meet our own internal standards (UNCH CMIL accepts ⱖ80% probabil-
per positive identification, depending on the genus and species of the yeast ity for valid identification), and isolates that generate identifications that
isolated. Our laboratory has been able to essentially eliminate the routine do not match with colony morphology and are subject to further review.
use of biochemical batteries, API strips for yeast identification, RAT test- Additionally, some identifications reveal other morphologies of organ-

2474 jcm.asm.org Journal of Clinical Microbiology August 2015 Volume 53 Number 8

 MALDI-TOF MS Cost Savings

mulas. Differences in costs between MALDI-TOF MS and conventional

methods are reported in gross and percent savings.
Acid-fast bacilli. As an addendum to the main study, we analyzed
costs over a 5-month period to determine the cost savings of implement-
ing MALDI-TOF MS for the prompt identification of rapidly growing
mycobacteria (RGM) and Mycobacterium tuberculosis complex (MTC).
From May 5 to October 4, 2014, 40 isolates of RGM and MTC were
identified by the CMIL. The process of preparing RGM and MTC isolates
for identification on MALDI-TOF MS is complex and time-consuming,
requiring properly trained laboratory personnel to achieve accurate re-
sults. This procedure involves inactivation of an acid-fast bacillus (AFB)
isolate with 70% ethanol along with mechanical disruption of the organ-
isms with glass beads, centrifugation to concentrate the sample, resuspen-
FIG 2 Diagram of Vitek disposable target slide utilization and isolate identi- sion of the pellet with 70% formic acid and acetonitrile, application of the
fication with an optimal and actual scenario. end product onto the target slide, and covering the specimen with matrix.
These isolates were evaluated using the Saramis database v4.12. This total
process was allotted 23 minutes of technologist time per identification of
RGM and MTC isolates.
isms that have already been identified or were not otherwise reported In contrast, traditional identification involves molecular sequencing
independently (e.g., oropharyngeal flora) and as such do not generate a of the 16S rRNA gene for the determination of most AFB isolates (10). In
reportable identification for that particular sample. One drawback of the our laboratory, where sequencing is done weekly, this typically adds an
instrument is that it cannot distinguish between Escherichia coli and Shi- extra week or more to the identification. Additionally, differentiation be-
gella sp., so testing of these isolates will result in a warning. Isolates that are tween two species of RGM, Mycobacterium abscessus and Mycobacterium
non-lactose fermenting and non-beta-hemolytic require further testing chelonae, requires further molecular sequencing, in our case of the hsp65
on the Vitek 2 to differentiate. region, which adds additional cost and time, as this methodology was also
Costs for MALDI-TOF MS were calculated by tallying the total expen- only performed once a week in our laboratory (11). Each molecular se-
ditures required to perform the test. This included the commercial or- quencing method is allotted 25 min of technologist time to complete.
ganic matrix solution for all organisms plus formic acid for yeasts, Vitek Calculations for RGM and MTC were performed for reagent costs and
MS target slides, pipette tips for application of the matrix and formic acid total cost utilizing a specialized formula. Reagent costs were directly cal-
solutions, and toothpicks for applying the colonies to the target slide. To culated. For 16S molecular sequencing, reagents equated to $61.39 and
properly calculate the cost of the disposable target slides, we estimated the doubled to $122.78 for hsp65 sequencing per test performed. Reagent
total efficiency of reportable MALDI-TOF MS identifications per slide. costs for MALDI-TOF MS were calculated to be $11.33 with all of the
This was estimated at 75%. In other words, three out of every four sample aforementioned required materials and consumables. Total costs were
spots available on the target slide provided an organism identification that then calculated utilizing the direct cost calculation previously mentioned,
was reportable (Fig. 2). This estimate included unused sample spots and where technologist time required for each isolate is added to the direct
the inability to acquire reportable identifications. Technologist time was reagent cost. Of note, the MALDI-TOF MS total cost was calculated with
estimated at 2.5 minutes per isolate for spotting specimens and operating testing and identification of only two isolates performed per AG and per
the instrument, which also included the resampling of failed spots. Addi- run. This reflected our laboratory workflow, as we batch specimens and
tional technologist and reagent costs associated with instrument reanaly- perform MALDI-TOF MS runs for AFB once or twice weekly. Differences
sis were not included because they were negligible, as the technologist only in costs between MALDI-TOF MS and conventional methods are re-
needed to select the sample spot to be reanalyzed and the instrument ported in gross and percent savings.
automatically retested the isolate. No further technologist time was re-
quired. RESULTS
With these assumptions, the total reagent cost per isolate was $0.43.
Over the 12-month study, from April 1, 2013, to March 31, 2014,
This included $0.10 for the organic matrix solution, $0.01 for the tooth-
pick, $0.06 for the pipette tip, and $0.26 for each sample spot. All cost a total of 21,930 isolates were identified in the UNCH CMIL.
estimates were rounded to the nearest cent. The technologist time was Reagent cost analysis. Traditional identification of the isolates
then multiplied by the average technologist salary at UNCH CMIL and would have cost $78,689.62 in reagents alone, compared to
added to the reagent cost to account for direct costs. Calculations were $9,581.01 for identification with MALDI-TOF MS. The use of
performed for reagent and direct costs utilizing the aforementioned for- MALDI-TOF MS equated to a substantial net savings of

TABLE 1 Reagent cost comparison between traditional and MALDI-TOF MS methods

Reagent costs ($) Cost savings
No. of
Organism samples Traditional MALDI-TOF $ %
Enterobacteriaceae 7,503 27,407.46 3,226.29 24,181.17 88.2
Enterococcus spp. 1,454 8,361.20 625.22 7,735.98 92.5
GNFa 3,489 21,154.03 1,501.56 19,652.47 92.9
Staphylococcus spp. 5,790 8,003.24 2,489.70 5,513.54 68.9
Streptococcus spp. 2,332 10,149.14 1,002.76 9,146.38 90.1
Yeast 1,362 3,614.55 735.48 2,879.07 79.7

Total 21,930 78,689.62 9,581.01 69,108.61 87.8

a
Gram-negative glucose nonfermenters.

August 2015 Volume 53 Number 8 Journal of Clinical Microbiology jcm.asm.org 2475

Tran et al.

TABLE 2 Direct cost comparison between traditional and MALDI-TOF MS methods

Total cost ($) Cost savings
No. of
Organism samples Traditional MALDI-TOF $ %
Enterobacteriaceae 7,503 51,717.18 13,355.34 38,361.84 74.2
Enterococcus spp. 1,454 13,072.16 2,588.12 10,484.04 80.2
GNFa 3,489 32,458.39 6,211.71 26,246.68 80.9
Staphylococcus spp. 5,790 17,383.04 10,306.20 7,076.84 40.7
Streptococcus spp. 2,332 17,704.82 4,150.96 13,553.86 76.6
Yeast 1,362 10,197.09 2,574.18 7,622.91 74.8

Total 21,930 142,532.69 39,186.51 103,346.18 72.5

a
Gram-negative glucose nonfermenters.

$69,108.61, or 87.8%, in reagent costs compared to the use of tra- isolate. The direct cost from MALDI-TOF MS was only $1,403.20,
ditional methods (Table 1). When analyzed by organism type, the or $35.08 per isolate, which equates to a 74.0% savings. These
largest savings were realized among Gram-negative GNFs, with a savings were reflected by the reduced cost of reagents required for
92.9% reduction in reagent costs with MALDI-TOF MS. Tradi- MALDI-TOF MS versus molecular sequencing methods, since
tional costs for identifying these organisms mainly involved Vitek technologist times were similar between the two methods (23 and
2 and 16S molecular sequencing. This large percentage of savings 25 min, respectively). Each isolate that required 16S sequencing
was expected, since the use of MALDI-TOF MS for identification cost the UNCH CMIL $61.39 in reagents, while use of hsp65 dou-
of bacteria and yeasts from culture utilizes a minimal amount of bled that cost. Even with the additional reagents required for per-
reagents. forming MALDI-TOF MS, reagent costs accounted for only
Direct cost analysis. If total costs were calculated to include $11.33 per isolate, or 89.7% less than molecular sequencing.
technologist time, traditional identification would have cost Therefore, even during a short period of 5 months, the laboratory
$142,532.69, compared to $39,186.51 for identification with was able to save nearly $4,000, or $100 per identification, by elim-
MALDI-TOF MS (Table 2). Therefore, the implementation of MS inating the need for molecular sequencing.
for identification of bacteria and yeasts netted a savings of Cost per sample. We also calculated the average cost per sam-
$103,346.18, or 72.5%. Again, Gram-negative GNFs yielded the ple identified for the various scenarios described. This simply rep-
largest savings with MALDI-TOF MS, where 80.9% total cost sav- resents the mean of the total costs, including reagent, technologist,
ings was observed. and maintenance costs, for all isolates identified during the study
The above calculations represent an estimation of the annual period. Reagent costs for the traditional methods utilized aver-
savings without the maintenance agreement contracts, which aged $3.59 per isolate and those for MALDI-TOF MS were only
would apply to the first year after purchase of the instrument, $0.43 (Fig. 3). Total costs with traditional methods, including re-
since maintenance is included during the initial year of operation. agent, technologist time, and maintenance agreement contracts,
To properly calculate the cost savings of each subsequent year, were determined to be $6.50 per isolate reported, compared to
beginning with year 2, the annual budget of the maintenance con- $3.14 for with MALDI-TOF MS.
tract needs to be added in. With the additional service contracts
added, the laboratory savings fell to $73,646.18, or 51.7% (Table DISCUSSION
3). Interestingly, there was a loss of 4.4% in savings for the iden- The transition from conventional microbiological methods to
tification of Staphylococcus spp. when maintenance agreement MALDI-TOF MS technology resulted in significant cost savings.
contracts were added to the calculation. There were no other re- The laboratory savings estimate was $73,646.18 (51.7%) in total
markable differences among the other categories. costs during the 12-month study (Table 3). This included reagent,
Acid-fast bacilli. Through traditional methods utilizing mo- technologist, and maintenance agreement costs. Workflow, which
lecular sequencing, it would have cost the CMIL $5,392.08 to affects the amount of time a technologist spends on identifying an
properly identify the 40 RGM and MTC specimens, or $134.80 per isolate, is an important aspect that has direct implications for total

TABLE 3 Total cost comparison between traditional and MALDI-TOF MS methods, including maintenance agreement costs
Total cost ($) Cost savings
No. of
Organism samples Traditional MALDI-TOF $ %
Enterobacteriaceae 7,503 51,717.18 23,516.72 28,200.46 54.5
Enterococcus spp. 1,454 13,072.16 4,557.29 8,514.87 65.1
GNFa 3,489 32,458.39 10,936.89 21,521.50 66.3
Staphylococcus spp. 5,790 17,383.04 18,147.65 ⫺764.61 ⫺4.4
Streptococcus spp. 2,332 17,704.82 7,309.21 10,395.61 58.7
Yeast 1,362 10,197.09 4,418.75 5,778.34 56.7

Total 21,930 142,532.69 68,886.51 73,646.18 51.7

a
Gram-negative glucose nonfermenters.

2476 jcm.asm.org Journal of Clinical Microbiology August 2015 Volume 53 Number 8

 MALDI-TOF MS Cost Savings

FIG 3 Mean reagent and total costs (with and without maintenance costs) per identification of isolates by traditional and MALDI-TOF MS methods.

costs. This includes but is not limited to the number of touches the methods. Although the technologist time and reagent costs were
technologist performs during the process of isolate determina- similar for both methods, addition of the maintenance contract
tion. The UNCH CMIL underwent a Lean assessment that was increased the cost of MALDI-TOF MS identification above that
conducted by bioMérieux as part of the purchase of the MALDI- for conventional identification of Staphylococcus species. How-
TOF MS, which helped the laboratory refine protocols to maxi- ever, traditional methods of identification of Staphylococcus sap-
mize time and efforts when identifying specimens with the rophyticus and Staphylococcus lugdunensis required more technol-
MALDI-TOF MS. ogist time and biochemical tests, which delayed identification of
Laboratories can maximize resources and minimize costs by these organisms for up to 48 h. Additionally, the Vitek MS pro-
batching runs and waiting for all of the sample spots within an vided our laboratory with the ability to identify coagulase-nega-
acquisition group on the target slide to be filled prior to loading tive Staphylococcus isolates to the species level, which was not done
the slides onto the MS; the remaining AGs would be accessible for previously. This is beneficial for blood and sterile body fluid iso-
additional runs at a later time. This helps boost slide usage effi- lates to help differentiate between true infection and contamina-
ciency and provides more savings. This philosophy should be tion of blood and sterile body fluid cultures when multiple colo-
possible for most runs throughout the day without impacting nial morphologies were detected (12–14). For us, these benefits far
turnaround time significantly. An optimal scenario assumes max- outweigh the added costs of MALDI-TOF MS. Additionally, some
imum utilization of reagents and technologist time; i.e., all sample rapid biochemical tests, such as spot indole for Escherichia coli and
spots on the disposable slide are utilized, and 100% of the spots PYR for Streptococcus pyogenes, may be less expensive than MS.
result in a reportable identification (Fig. 2). If this scenario were The inclusion and exclusion of certain organisms in this new
true, then the total cost savings would be $80,894.12, or 56.8%, methodology is something that each laboratory should consider
annually instead of $73,646.18. Realistically, we believe the actual and incorporate into implementation protocols, because MALDI-
cost savings are between this optimal scenario and our study cal- TOF MS reagent costs, workflow, and maintenance contracts vary
culations, since it is not feasible to expect that every spot will be significantly.
utilized on each slide and result in a reportable identification. Verification studies. Verification studies for implementation
Not surprisingly, implementation of MALDI-TOF MS re- of MALDI-TOF MS began in June 2012, and the laboratory went
duced reagent costs by 88% compared to the costs of conventional live with the first set of organisms in October 2012. For this vali-
methods. This is because the only consumables utilized in our dation, we included 731 organisms across 3,411 sample spots.
laboratory for MALDI-TOF MS were the disposable target slides, Through June 2013, the verifications cost an estimated $4,357.78,
matrix, formic acid (mainly for yeasts), pipette tips, and tooth- and more organisms continue to be verified as we encounter them
picks for spotting organisms. in the laboratory.
Interestingly, there was only one category where it was less Acid-fast bacilli. Because of the expense of identifying rapidly
expensive to perform traditional identification than identification growing mycobacteria and Mycobacterium tuberculosis by molec-
with MALDI-TOF MS. Identification of Staphylococcus spp. by ular means, this group of AFB is likely to be more efficiently iden-
MS cost $764.61 more than by the traditional latex agglutination tified by MALDI-TOF MS. Our laboratory cultures AFB through

August 2015 Volume 53 Number 8 Journal of Clinical Microbiology jcm.asm.org 2477

Tran et al.

traditional mycobacterial methods utilizing liquid (Bactec MGIT that were generated. This was not possible for us due to resource
960 mycobacterial detection system) and solid (Lowenstein-Jen- constraints. We were, however, able to audit all of the spots gen-
sen medium; Becton, Dickinson and Company, Franklin Lakes, erated over a 7-day period and found that 79.6% of the spots
NJ) media. Once positive growth is verified via a Kinyoun stain, generated a reportable identification. Similarly, a 2-month audit
the organism is subcultured to 7H11 medium. After sufficient of our slide usage efficiency revealed that 89.8% of sample spots
growth, the organisms are prepared for identification by MALDI- were utilized. These surveys showed that our estimate of 75% total
TOF MS. The rate of growth determines whether the organism is slide efficiency was relatively close to what actually occurs in our
an RGM, and colony morphology helps determine whether MTC laboratory. However, without prospective monitoring of slide
is in the differential. consumption and efficiency throughout the study period, it is im-
In February 2014, RGM and MTC were added to the menu of possible to accurately calculate this value.
approved organisms for identification on MALDI-TOF MS. These It should be noted that this was a cost-savings study that
identifications utilize the research-use-only (RUO) Saramis data- estimated the amount saved by the laboratory after routine
base v4.12 and incur a higher total cost due to a time-consuming implementation of MALDI-TOF MS for identification of the
extraction technique required to process and inactivate the AFB more common organisms encountered in the clinical microbi-
prior to organism identification by MALDI-TOF MS. These costs ology laboratory. It is not a true cost-effectiveness study that
still represent a significant 74% reduction in direct costs after examines costs compared to a defined clinical or natural out-
moving to MALDI-TOF MS from traditional molecular sequenc- come to improve health (15). A cost-effectiveness study would
ing methods while reducing the turnaround time. require evaluation of the turnaround time to results and how it
Capital costs. The main stumbling block with implementing impacts patient care, which is beyond the scope of this paper.
MALDI-TOF MS in the clinical microbiology setting is the capital Further and more in-depth studies, such as that by Tan et al.
cost of acquiring the instrument. With an approximate $270,000 (9), are needed to determine the cost-effectiveness of this new
price tag for the instrument and associated in vitro diagnostics and technology.
RUO software databases, the initial financial hurdle may be too MALDI-TOF MS not only represents an innovative technology
high for some laboratories to overcome. Moreover, laboratories for the rapid and accurate identification of bacterial and fungal
must be mindful of the maintenance that is associated with the isolates, it also provides significant cost savings for the laboratory.
MALDI-TOF MS. The UNCH CMIL pays an additional $29,700 Despite the high capital cost of the instrument, the ease of perfor-
annually for instrument and database/software maintenance. Our mance, the rapid turnaround time to results, and the modest cost
laboratory requires MALDI-TOF MS fine-tuning every 3 to 4 of testing for each sample make this new methodology a paradigm
weeks and has undergone a full laser and three linear detector shift in the field of clinical microbiology.
replacements in the ⬍3 years that the instrument has been in use.
The fine-tuning process, which takes several hours, does not in- ACKNOWLEDGMENTS
terrupt workflow or increase cost in our laboratory, as we batch We thank the laboratory staff at the UNCH CMIL for their contributions,
specimens and test them once the instrument is ready for opera- continued diligence, and dedication to their craft, which helped tremen-
tion again. We estimate that the MALDI-TOF MS was nonfunc- dously with our cost savings.
tional for extended periods (i.e., ⬎18 h) for a total of 4 days during REFERENCES
the study period. During these times, we reverted to traditional 1. Shah HN, Gharbia S. 15 June 2010. Mass spectrometry for microbial
methods of identification. Maintenance is an integral part of proteomics. John Wiley & Sons, Inc., Chichester, United Kingdom. http:
the cost calculation for MALDI-TOF MS and needs to be //dx.doi.org/10.1002/9780470665497.
strongly considered by every laboratory planning on imple- 2. Karas M, Hillenkamp F. 1988. Laser desorption ionization of proteins
with molecular masses exceeding 10,000 daltons. Anal Chem 60:2299 –
menting this technology. It should be noted that maintenance 2301.
costs begin in year 2, since the first year is covered by the man- 3. Buchan BW, Ledeboer NA. 2013. Advances in identification of clinical
ufacturer warranty. yeast isolates by use of matrix-assisted laser desorption ionization-time of
Therefore, understanding the cost savings after implementa- flight mass spectrometry. J Clin Microbiol 51:1359 –1366. http://dx.doi
.org/10.1128/JCM.03105-12.
tion of MALDI-TOF MS is important because of its large initial
4. De Carolis E, Vella A, Vaccaro L, Torelli R, Spanu T, Fiori B, Posteraro
fiscal investment. With an annual savings estimate of $73,646.18, B, Sanguinetti M. 2014. Application of MALDI-TOF mass spectrometry
coupled with the initial year’s savings of $103,346.18, we antici- in clinical diagnostic microbiology. J Infect Dev Ctries 8:1081–1088. http:
pate that the capital cost of the instrument would be offset in just //dx.doi.org/10.3855/jidc.3623.
⬎3 years. The net savings will only increase as we continue to 5. Chao QT, Lee TF, Teng SH, Peng LY, Chen PH, Teng LJ, Hsueh PR.
2014. Comparison of the accuracy of two conventional phenotypic meth-
validate more organisms by MALDI-TOF MS that will ultimately ods and two MALDI-TOF MS systems with that of DNA sequencing anal-
require less molecular sequencing and labor-intensive technolo- ysis for correctly identifying clinically encountered yeasts. PLoS One
gist time for identification. 9:e109376. http://dx.doi.org/10.1371/journal.pone.0109376.
Limitations. There are several limitations to the study. We 6. Cherkaoui A, Hibbs J, Emonet S, Tangomo M, Girard M, Francois P,
Schrenzel J. 2010. Comparison of two matrix-assisted laser desorption
were limited by database inquiries and our financial test cost anal-
ionization-time of flight mass spectrometry methods with conventional
yses. The laboratory information system has limited querying ca- phenotypic identification for routine identification of bacteria to the spe-
pabilities. While we were able to identify the number of isolates cies level. J Clin Microbiol 48:1169 –1175. http://dx.doi.org/10.1128/JCM
reported during the study period, we were unable to identify ex- .01881-09.
actly how many sample spots were utilized for identification of the 7. Eigner U, Holfelder M, Oberdorfer K, Betz-Wild U, Bertsch D, Fahr
AM. 2009. Performance of a matrix-assisted laser desorption ioniza-
organisms that were reported. The only way to track this is to tion-time-of-flight mass spectrometry system for the identification of
prospectively audit the number of spots that were utilized and bacterial isolates in the clinical routine laboratory. Clin Lab 55:289 –
cross-reference that number with the number of isolate reports 296.

2478 jcm.asm.org Journal of Clinical Microbiology August 2015 Volume 53 Number 8

 MALDI-TOF MS Cost Savings

8. Seng P, Drancourt M, Gouriet F, La Scola B, Fournier PE, Rolain JM, taneous infections caused by rapidly growing mycobacteria. Eur J Clin Micro-
Raoult D. 2009. Ongoing revolution in bacteriology: routine identifica- biol Infect Dis 32:161–188. http://dx.doi.org/10.1007/s10096-012-1766-8.
tion of bacteria by matrix-assisted laser desorption ionization time-of- 12. Abdul Rahman Z, Hamzah SH, Hassan SA, Osman S, Md Noor SS.
flight mass spectrometry. Clin Infect Dis 49:543–551. http://dx.doi.org/10 2013. The significance of coagulase-negative staphylococci bacteremia in a
.1086/600885. low resource setting. J Infect Dev Ctries 7:448 – 452. http://dx.doi.org/10
9. Tan KE, Ellis BC, Lee R, Stamper PD, Zhang SX, Carroll KC. 2012. .3855/jidc.2535.
Prospective evaluation of a matrix-assisted laser desorption ionization– 13. Al Wohoush I, Rivera J, Cairo J, Hachem R, Raad I. 2011. Comparing
time of flight mass spectrometry system in a hospital clinical microbiology clinical and microbiological methods for the diagnosis of true bacteraemia
laboratory for identification of bacteria and yeasts: a bench-by-bench among patients with multiple blood cultures positive for coagulase-
study for assessing the impact on time to identification and cost- negative staphylococci. Clin Microbiol Infect 17:569 –571. http://dx.doi
effectiveness. J Clin Microbiol 50:3301–3308. http://dx.doi.org/10.1128 .org/10.1111/j.1469-0691.2010.03372.x.
/JCM.01405-12. 14. Usha MG, Shwetha DC, Vishwanath G. 2013. Speciation of coagulase
10. Janda JM, Abbott SL. 2007. 16S rRNA gene sequencing for bacterial negative Staphylococcal isolates from clinically significant specimens and
identification in the diagnostic laboratory: pluses, perils, and pitfalls. J their antibiogram. Indian J Pathol Microbiol 56:258 –260. http://dx.doi
Clin Microbiol 45:2761–2764. http://dx.doi.org/10.1128/JCM.01228-07. .org/10.4103/0377-4929.120383.
11. Kothavade RJ, Dhurat RS, Mishra SN, Kothavade UR. 2013. Clinical and 15. Gold MR. 1996. Cost-effectiveness in health and medicine. Oxford Uni-
laboratory aspects of the diagnosis and management of cutaneous and subcu- versity Press, New York, NY.

August 2015 Volume 53 Number 8 Journal of Clinical Microbiology jcm.asm.org 2479