Postoperative Pneumonia Prevention

in Pulmonary Resections: A Feasibility

Pilot Study
Tara R. Semenkovich, MD, MPHS, Christine Frederiksen, MS,
Jessica L. Hudson, MD, MPHS, Melanie Subramanian, MD, Marin H. Kollef, MD,
G. Alexander Patterson, MD, Daniel Kreisel, MD, PhD, Bryan F. Meyers, MD, MPH,
Benjamin D. Kozower, MD, MPH, and Varun Puri, MD, MSCI
Division of Cardiothoracic Surgery, Department of Surgery, and Division of Pulmonary and Critical Care Medicine, Department of
Medicine, Washington University, St. Louis, Missouri

Background. Pneumonia after pulmonary resection contemporaneous cohort of 611 patients who met sur-
occurs in 5% to 12% of patients and causes substantial gical inclusion criteria. Preoperative adherence to
morbidity. Oral hygiene regimens lower the incidence of the chlorhexidine toothbrushing regimen was high:
ventilator-associated pneumonias; however, the impact in median 100% (interquartile range: 87% to 100%). Post-
patients undergoing elective pulmonary resection is operatively, 80% of patients continued toothbrushing,
unknown. We conducted a prospective pilot study to whereas 20% declined further participation. Among
assess the feasibility of an oral hygiene intervention in those who participated postoperatively, median adher-
this patient cohort. ence was 86% (interquartile range: 53% to 100%). There
Methods. Patients undergoing elective pulmonary was a trend toward reduction in postoperative pneu-
resection were prospectively enrolled in a single-arm monia: 1.6% (1 of 62) in the intervention cohort versus
interventional study with time-matched controls. Partici- 4.9% (30 of 611) in the time-matched cohort (p [ 0.35).
pants were asked to brush their teeth with 0.12% chlor- The number needed to treat to prevent one case of
hexidine three times daily for 5 days before their pneumonia was 30 patients.
operations and 5 days or until the time of discharge after Conclusions. This pilot study demonstrated patients
their operations. Patients were eligible if they had known can comply with an inexpensive perioperative oral
or suspected lung cancer and were undergoing (1) any hygiene regimen that may be promising for
anatomic lung resection or (2) a wedge resection with reducing morbidity (Clinical Trials Registry:
forced expiratory volume in 1 second or diffusing capacity NCT01446874).
of lung for carbon monoxide less than 50% predicted.
Results. Sixty-two patients were enrolled in the (Ann Thorac Surg 2018;-:-–-)
pilot intervention group and compared with a Ó 2018 by The Society of Thoracic Surgeons

P ostoperative pneumonia is a common complication

after major pulmonary resections, with an incidence
of 5% to 12% [1–4]. It results in significant morbidity, and
infections [8, 9]. Poor dental hygiene and oral colonization
with pathogenic organisms has been associated with
the development of pneumonia in nursing home resi-
can lead to reintubation or tracheostomy, critical illness, dents and hospitalized patients [10–12]. Correspond-
prolonged hospital stay, need for rehabilitation, and ingly, strategies to improve dental care and reduce oral
increased rates of unplanned readmission [5, 6]. Pneu- bacteria have shown promise in decreasing the risk of
monia is also one of the primary causes of postoperative pneumonia: systematic reviews of randomized trials
mortality after thoracic surgery [6, 7]. Consequently, there demonstrated that chlorhexidine-based oral cleansing
is substantial interest in identifying strategies for risk regimens decrease rates of HCAP, with an absolute risk
reduction. reduction of 6% in mechanically ventilated patients [13]
Investigations into the cause of health care–acquired and between 6% and 12% in elderly patients [14]. Based
pneumonia (HCAP) have identified aspiration of oral on the encouraging results with simple and inexpensive
pathogens as a source for bacteria leading to pulmonary oral decontamination in these populations, there may
be a role for dental hygiene in the perioperative period,
Accepted for publication Aug 5, 2018. especially for patients who are at high risk of post-
operative pneumonia.
Address correspondence to Dr Semenkovich, Department of Cardiotho-
racic Surgery, Washington University in St. Louis, 660 S Euclid Ave,
Within the field of thoracic surgery, there are limited
Campus Box 8234, St. Louis, MO 63110; email: semenkovicht@wudosis. data regarding the effectiveness of perioperative oral
wustl.edu. hygiene in reducing the risk of pneumonia after elective

Ó 2018 by The Society of Thoracic Surgeons 0003-4975/$36.00

Published by Elsevier Inc. https://doi.org/10.1016/j.athoracsur.2018.08.008
2 SEMENKOVICH ET AL Ann Thorac Surg
POSTOPERATIVE PNEUMONIA PREVENTION 2018;-:-–-

period of the study. Patients were approached by a study

Abbreviations and Acronyms coordinator and enrolled using convenience sampling.
BMI = body mass index This study was approved by the Institutional Review
CI = confidence interval Board at Washington University in St. Louis, and all
COPD = chronic obstructive pulmonary patients participating in the intervention provided
disease informed consent.
DLCO = diffusing capacity of lungs for
carbon monoxide Clinical Intervention
FEV1 = forced expiratory volume in 1 Participants in the study group (postoperative pneumonia
second
prevention [POPP] cohort) were asked to brush their
HCAP = health care–acquired pneumonia
teeth with 0.12% chlorhexidine solution three times daily
OR = odds ratio
PFT = pulmonary function test [17, 18] in the perioperative period. To encourage
POPP = prevention of postoperative adherence, participants were each given one toothbrush,
pneumonia a supply of chlorhexidine solution for brushing (provided
STS = The Society of Thoracic Surgeons by Sage Products, Cary, IL), and a log sheet for recording
VATS = video-assisted thoracoscopic surgery their perioperative oral hygiene. Patients were asked to
brush for 5 days preoperatively, and until the time of
discharge or 5 days postoperatively if they remained
hospitalized. This facilitated collection of patient logs
operations. One small study of preoperative tooth- before discharge on which their adherence with the
brushing showed pneumonia rates after esophagectomy toothbrushing episodes were recorded in real time.
could be reduced from 32% to 9% by brushing five times Patients were excluded if they did not subsequently
daily, and the effect was greater in a subgroup of patients undergo a qualifying operation, had a preoperative infec-
whose oral flora was positive for pathogenic bacteria tion, or withdrew before participating in the intervention.
before intervention [15]. A larger propensity score The contemporaneous cohort consisted of all patients
matched study of 420 patients showed intensive oral care meeting our inclusion criteria for enrollment who
including professional tooth and tongue cleanings by received their operation during the study. A time-
dentists or hygienists was associated with an odds ratio matched cohort was selected to control for unmeasured
(OR) of 0.37 (95% confidence interval [CI]: 0.20 to 0.65) for institutional factors that could affect pneumonia risk. The
having postoperative pneumonia [16]. The impact of control group was not given any specific instructions
perioperative oral hygiene improvement in patients about oral hygiene.
undergoing elective pulmonary resection, however, is All patients received standard perioperative care at our
unknown. institution including, as indicated, smoking cessation
We conducted a prospective pilot study to assess the counseling, preoperative pulmonary rehabilitation,
feasibility of implementing a short-term perioperative appropriate resection based on PFTs, minimally invasive
oral hygiene intervention in a cohort of patients under- surgery if possible, and early postoperative ambulation,
going pulmonary resection. We hypothesized that our physical therapy, and incentive spirometry. A summary
intervention would have high patient adherence and of patient selection criteria and intervention can be seen
lower the risk of postoperative pneumonia after lung in our Consolidated Standards of Reporting Trials
resection. (CONSORT) diagram in Figure 1.

Outcomes
Patients and Methods
Adherence to the toothbrushing regimen was assessed
Patient Eligibility and Data Sources preoperatively and postoperatively in the study group.
Patients undergoing elective pulmonary resection were Preoperative percent adherence was defined as the
prospectively enrolled in a single-arm interventional number of times the patient brushed out of a possible 15.
study and were compared with time-matched controls. Postoperative percent adherence was defined as the
Adult patients (aged 18 years or more) were eligible for number of times the patient brushed out of the total
inclusion if they had known or suspected lung cancer and number possible (three per inpatient day and on the day
were either undergoing (1) any anatomic lung resection of discharge any episodes before their departure, up to as
or (2) a wedge resection and had either forced expiratory many as 15 if they were hospitalized beyond 5 days).
volume in 1 second (FEV1) or diffusing capacity of lungs Postoperative pneumonia was defined as per The
for carbon monoxide (DLCO) less than 50% predicted. Society of Thoracic Surgeons (STS) guidelines [19] and
Data on patient demographics, comorbidities, pulmonary assessed in both groups. Patients were followed up
function tests (PFTs), operative procedure, and post- according to routine institutional practice, which included
operative outcomes for patients in both the intervention a postoperative clinic visit at approximately 4 weeks.
and control groups were obtained from a prospectively The presence or absence of pneumonia was verified
maintained institutional database. Data on patient by chart review of the initial hospitalization, any
adherence to the intervention were collected during the readmissions, and follow-up appointments through
Ann Thorac Surg SEMENKOVICH ET AL 3
2018;-:-–- POSTOPERATIVE PNEUMONIA PREVENTION

Fig 1. Consolidated Standards of Reporting Trials (CONSORT) flow diagram. (DLCO ¼ diffusing capacity of lungs for carbon monoxide; FEV1 ¼
forced expiratory volume in 1 second; PFTs ¼ pulmonary function tests; POPP ¼ prevention of postoperative pneumonia; Pre-op ¼ preoperative;
Rehab ¼ rehabilitation.)

30 days. This assessment was performed in both the Univariable analyses were then performed to assess the
POPP and contemporaneous cohorts by a trained unadjusted association of each variable with postoperative
research assistant using STS criteria, and individual pneumonia.
surgeons were not involved in classification. Patients
were considered to have postoperative pneumonia if they Results
met three of the following criteria within 30 days after Patient Population
surgery: (1) fever (temperature greater than 38.2
C); (2)
Sixty-two patients were enrolled in the intervention
leukocytosis (white blood cell count greater than 12,000/
mm3); (3) new infiltrate on chest radiograph; (4) positive group (POPP cohort) from October 2012 to November
sputum or bronchial culture; or (5) treatment with anti- 2015. These patients were compared with a contempora-
neous cohort of 611 patients who met the surgical inclu-
biotics. As one of their three criteria, patients were
required to either have an infiltrate or a positive respi- sion criteria during the same time period. There were no
ratory culture. statistically significant differences between the POPP
cohort and the contemporaneous cohort, respectively, in
age (mean 63.5 versus 63.2 years), sex (53.2% versus 45.5%
Statistical Analysis male), race (91.9% versus 90.1% white), FEV1 (79.5%
Statistical analyses were performed using SAS statistical versus 81.0%), DLCO (67.3% versus 69.1%), rates of major
software version 9.3 (SAS Institute, Cary, NC). Descrip- medical comorbidities, pulmonary resection performed
tive statistics were performed comparing patient (71.0% versus 65.8% lobectomy), or surgical approach
demographics, PFTs, comorbidities, operative character- (56.5% versus 56.6% video-assisted thoracoscopic surgery
istics, and outcomes between both cohorts. Continuous [VATS]). That indicates the POPP cohort was represen-
variables were compared between groups using the Wil- tative of the general patient population undergoing pul-
coxon rank sum test. Categoric variables were compared monary resection at our institution. These characteristics
with c2 tests and Fisher’s exact tests, as appropriate. are shown in Table 1.
4 SEMENKOVICH ET AL Ann Thorac Surg
POSTOPERATIVE PNEUMONIA PREVENTION 2018;-:-–-

Table 1. Patient Characteristics

Variable POPP Cohort (n ¼ 62) Contemporaneous Cohort (n ¼ 611) p Value

Patient characteristics
Age, years 63.5  8.6 63.2  11.4 0.943
Male 53.2 (33) 45.5 (278) 0.245
Race
White 91.9 (57) 90.1 (551) 0.686
Black 8.1 (5) 8.7 (53)
Other ... 1.2 (7)
Body mass index, kg/m2 29.1  6.7 27.8  6.2 0.138
Pulmonary function tests
FEV1, % predicted 79.5  18.5 81  19.5 0.853
DLCO, % predicted 67.3  17.7 69.1  20 0.734
Comorbidities
COPD 17.7 (11) 28.3 (173) 0.075
Smoker
Current 22.6 (14) 29 (177) 0.566
Former 58.1 (36) 53.7 (328)
Never 19.4 (12) 17.4 (106)
Lung cancer 90.3 (56) 86.4 (528) 0.387
Pathologic stage
0 ... 2.3 (14) 0.662
I 53.2 (33) 47 (287)
II 22.6 (14) 22.1 (135)
III 11.3 (7) 10.2 (62)
IV ... 2.3 (14)
Unknown 3.2 (2) 2.6 (16)
Preoperative chemotherapy 12.9 (8) 6.4 (39) 0.066
Preoperative radiation 11.3 (7) 9.2 (56) 0.584
Prior lung cancer 4.8 (3) 3 (18) 0.431
Prior cardiothoracic surgery 11.3 (7) 9.8 (60) 0.713
Cardiovascular diseasea 22.6 (14) 19 (116) 0.494
Congestive heart failure ... 0.8 (5) 0.475
Diabetes mellitus 17.7 (11) 13.4 (82) 0.348
Prior other cancer 32.2 (20) 29.3 (179) 0.626
Immunosuppression 4.8 (3) 2.8 (17) 0.418
Operation
Lobectomy 71.0 (44) 65.8 (402) 0.861
Pneumonectomy 6.5 (4) 7.6 (46)
Bilobectomy 4.8 (3) 5.4 (33)
Sleeve 3.2 (2) 2.8 (17)
Segmentectomy 6.5 (4) 9.8 (60)
Wedge 4.8 (3) 7.2 (44)
Lung and chest wall resection 3.2 (2) 1.5 (9)
VATS approach 56.5 (35) 56.6 (346) 0.979
Outcome
Pneumonia 1.6 (1) 4.9 (30) 0.348
a
Cardiovascular disease includes coronary artery disease and peripheral vascular disease.
Values are mean  SD or percentage (n).
COPD ¼ chronic obstructive pulmonary disease; DLCO ¼ diffusing capacity of lung for carbon monoxide; FEV1 ¼ forced expiratory volume in 1
second; POPP ¼ prevention of postoperative pneumonia; VATS ¼ video-assisted thoracoscopic surgery.
Ann Thorac Surg SEMENKOVICH ET AL 5
2018;-:-–- POSTOPERATIVE PNEUMONIA PREVENTION

Table 2. Adherence in the Prevention of Postoperative Pneumonia Cohort

Variable 80% Adherence (n ¼ 25) <80% Adherence (n ¼ 37) p Value

Patient characteristics
Age, years 62.1  7.4 64.4  8.7 0.385
Male 52.0 (13) 48.7 (18) 0.344
Race, white 100 (25) 81.1 (30) 0.070
Body mass index, kg/m2 28.6  6 28.5  6.6 0.460
Pulmonary function tests
FEV1, % predicted 80.3  18.7 78.6  19.2 0.643
DLCO, % predicted 67.8  20.1 67  15.9 0.733
Comorbidities
COPD 16 (4) 18.9 (7) 0.312
Smoker
Current 28 (7) 16.2 (6) 0.274
Former 60 (15) 54.1 (20)
Never 12 (3) 21.6 (8)
Lung cancer 88 (22) 86.5 (32) 0.242
Preoperative chemotherapy 8 (2) 16.2 (6) 0.193
Preoperative radiation 8 (2) 13.5 (5) 0.252
Prior cardiothoracic surgery 8 (2) 13.5 (5) 0.252
Cardiovascular disease 16 (4) 21.6 (8) 0.267
Diabetes mellitus 12 (3) 16.2 (6) 0.288
Prior other cancer 32 (8) 27 (10) 0.337
Operation
Lobectomy 72 (18) 62.2 (23) 0.642
Pneumonectomy 4 (1) 8.1 (3)
Bilobectomy 4 (1) 5.4 (2)
Sleeve ... 5.4 (2)
Segmentectomy 8 (2) 5.4 (2)
Wedge 8 (2) 2.7 (1)
Lung and chest wall resection 4 (1) 2.7 (1)
Approach, %VATS 72 (18) 40.5 (15) 0.034
Outcomes
Pneumonia rate ... 2.7 (1) 0.236

Values are mean  SD or percentage (n).

COPD ¼ chronic obstructive pulmonary disease; DLCO ¼ diffusing capacity of lung for carbon monoxide; FEV1 ¼ forced expiratory volume in 1
second; POPP ¼ prevention of postoperative pneumonia; VATS ¼ video-assisted thoracoscopic surgery.

Adherence multivariable analysis, the only characteristic associated

Within the study group, preoperative adherence to the with excellent adherence to the toothbrushing regimen was
toothbrushing regimen was high: median 100% (inter- VATS approach (OR 3.26, 95% CI: 1.08 to 9.83, p ¼ 0.04).
quartile range: 87% to 100%). Postoperatively, 80%
continued toothbrushing, whereas 20% declined further Postoperative Pneumonia
participation and performed no additional toothbrushing. There was a trend toward reduction in postoperative
Among patients who participated postoperatively, median pneumonia: 1.6% (1 of 62 patients) of the intervention
adherence was 86% (interquartile range: 53% to 100%). cohort versus 4.9% (30 of 611) of the control group. The
We observed that 40% of patients (25 of 62) had excel- difference did not reach statistical significance (p ¼ 0.35)
lent adherence to the protocol, defined as performing in this pilot study (Table 1). Patients who had pneumonia
80% or more of the recommended episodes of tooth- had a postoperative median stay of 7 days (range, 2 to 84)
brushing preoperatively and 80% or more of the recom- compared with 4 days for patients without pneumonia
mended episodes of toothbrushing postoperatively. The (range, 0 to 49 days; p < 0.001). Variables associated with
highly adherent patients were more likely to be white development of postoperative pneumonia included open
(100% versus 81.1%, p ¼ 0.07) and had higher rates of approach (p < 0.001), cardiovascular disease (p ¼ 0.002),
VATS procedures (72.0% versus 40.5%, p ¼ 0.03). Charac- DLCO (p ¼ 0.003), FEV1 (p ¼ 0.01), and prior lung cancer
teristics by adherence rates are shown in Table 2. On (p ¼ 0.04; Table 3). Participation in the intervention
6 SEMENKOVICH ET AL Ann Thorac Surg
POSTOPERATIVE PNEUMONIA PREVENTION 2018;-:-–-

Table 3. Univariable Analyses: Associations With contemporaneous cohort underwent a thoracotomy.

Postoperative Pneumonia There were no significant differences in clinical charac-
Univariable OR p teristics, PFTs, comorbidities, or procedure characteristics
Variable (95% CI) Value between these subgroups (Table 4). The pneumonia rate
was 3.7% (1 of 27) in the POPP cohort and 8.7% (23 of 265)
Participated in POPP cohort 0.32 (0.04–2.37) 0.263 in the contemporaneous cohort, which was not statisti-
Patient characteristics cally significant in this pilot study (p ¼ 0.71). Participation
Age 0.98 (0.95–1.02) 0.306 in the intervention showed an OR of 0.40 (95% CI: 0.05 to
Female versus male 0.53 (0.25–1.10) 0.089 3.12, p ¼ 0.39) for the development of pneumonia on
Nonwhite versus white 0.63 (0.15–2.72) 0.539 univariable analysis (Table 5).
Body mass index 0.96 (0.90–1.02) 0.217
Pulmonary function tests Cost Analysis
FEV1, % predicted 0.98 (0.96–1.00) 0.013 In this study, a 16-ounce bottle of 0.12% chlorhexidine
DLCO, % predicted 0.97 (0.94–0.99) 0.003 solution cost $3.72 and bulk-ordered toothbrushes cost
Comorbidities $0.19 each. With a number needed to treat in the study
COPD 1.73 (0.82–3.63) 0.150 population of 30, the cost of preventing one pneumonia
Smoker was $117. Currently, an inexpensive toothbrush retails for
Current versus never 1.25 (0.37–4.23) 0.725 approximately $1.50 and the outpatient pharmacy at our
Former versus never 1.57 (0.52–4.71) 0.421 institution sells oral chlorhexidine solution for approxi-
Lung cancer 0.50 (0.21–1.20) 0.122 mately $15 without insurance. Using these prices, the
Preoperative chemotherapy 2.06 (0.69–6.17) 0.194 estimated cost to prevent one postoperative pneumonia
Preoperative radiation 1.46 (0.50–4.33) 0.491
would be $495, whereas costs attributable to prolonged
hospitalization from pneumonia range from $5,000 to
Prior lung cancer 3.72 (1.03–13.35) 0.044
$17,000 [20, 21] per patient.
Prior cardiothoracic surgery 1.36 (0.46–4.02) 0.576
Cardiovascular disease 3.21 (1.55–6.80) 0.002
Diabetes mellitus 0.92 (0.32–2.69) 0.880 Comment
Prior other cancer 0.56 (0.23–1.38) 0.208
Pneumonia after pulmonary resection not only contrib-
Immunosuppression 1.09 (0.14–8.44) 0.932 utes to significant morbidity, critical illness, costly pro-
Operation (reference: lobectomy) longed hospital stays, and readmissions, but is also one of
Pneumonectomy 2.64 (0.94–7.46) 0.066 the most lethal complications that a patient can have [6,
Bilobectomy 1.40 (0.31–6.28) 0.662 7]. Evidence-based strategies, including minimally inva-
Sleeve 1.32 (0.17–10.45) 0.792 sive operations and smoking cessation counseling, have
Segmentectomy 0.77 (0.17–3.39) 0.726 been used for risk reduction in the thoracic surgery
Wedge 1.06 (0.24–4.70) 0.942 population, but additional interventions are needed to
Lung and chest wall resection 2.38 (0.29–19.59) 0.421 prevent pneumonia in this high-risk population.
Approach, VATS versus open 0.21 (0.09–0.49) <0.001 In this study, we demonstrated the feasibility of
implementing an oral hygiene regimen for patients un-
CI ¼ confidence interval; COPD ¼ chronic obstructive pulmonary
disease; DLCO ¼ diffusing capacity of lung for carbon monoxide;
dergoing major pulmonary resection. We found
FEV1 ¼ forced expiratory volume in 1 second; OR ¼ odds ratio; extremely high rates of preoperative adherence (median
POPP ¼ prevention of postoperative pneumonia; VATS ¼ video- 100% adherence) and acceptable postoperative adherence
assisted thoracoscopic surgery.
(80% participation), with very few patients withdrawing
after agreeing to participate in the intervention. These
showed an OR of 0.32 (95% CI: 0.04 to 2.37, p ¼ 0.26) for rates are somewhat better than previously reported
development of postoperative pneumonia. Based on the adherence rates with other forms of preoperative decon-
observed incidence, the number needed to treat to pre- tamination [22]. Moreover, this pilot study showed an
vent one case of pneumonia was 30 patients. encouraging trend toward lower rates of pneumonia
The single case of postoperative pneumonia in the overall (1.6% versus 4.9%). Our institutional incidence of
POPP cohort was a 61-year-old white man, a former pneumonia after pulmonary resection was 86% of the
smoker with cardiovascular disease who underwent a National Surgical Quality Improvement Program risk-
thoracotomy and lobectomy with preoperative PFTs; adjusted expected rate during the study period, and this
FEV1 75% of predicted and DLCO 62% of predicted. He intervention may have the potential to lower the inci-
was 73% adherent to his preoperative toothbrushing dence further.
regimen and 64% adherent to his postoperative regimen. Analysis of risk factors for nonadherence showed that
patients who underwent a thoracotomy were less likely to
Open Operation Subgroup Analysis comply with the toothbrushing regimen. Unfortunately,
Because an open operation was strongly associated with these patients generally are at higher risk for post-
postoperative pneumonia, we performed a subgroup operative pneumonia. Our subgroup analysis of open
analysis in this population. Twenty-seven patients operations showed a similar trend toward decreased
within the POPP cohort and 265 patients in the pneumonia rates in the POPP cohort undergoing
Ann Thorac Surg SEMENKOVICH ET AL 7
2018;-:-–- POSTOPERATIVE PNEUMONIA PREVENTION

Table 4. Comparison of Characteristics Between Groups: Open Operation Subgroup Analysis

Variable POPP Cohort (n ¼ 27) Contemporaneous Cohort (n ¼ 265) p Value

Patient characteristics
Age, years 65  6.3 61.9  11.9 0.261
Male 59.3 (16) 51.7 (137) 0.454
Race
White 96.3 (26) 91.3 (242) 0.654
Black 3.7 (1) 7.9 (21)
Body mass index, kg/m2 29.4  7.7 28.3  6.4 0.709
Pulmonary function tests
FEV1, % predicted 75.9  21.5 77.9  17.8 0.716
DLCO, % predicted 64.5  20.2 65.6  18.4 0.991
Comorbidities
COPD 25.9 (7) 32.1 (85) 0.665
Smoker
Current 22.2 (6) 28.7 (76) 0.290
Former 70.4 (19) 55.5 (147)
Never 7.4 (2) 15.9 (42)
Lung cancer 92.6 (25) 87.9 (233) 0.752
Preoperative chemotherapy 22.2 (6) 13.2 (35) 0.240
Preoperative radiation 14.8 (4) 15.1 (40) 1.000
Prior lung cancer 11.1 (3) 4.2 (11) 0.129
Prior cardiothoracic surgery 14.8 (4) 12.1 (32) 0.757
Cardiovascular disease 25.9 (7) 20.0 (53) 0.458
Diabetes mellitus 14.8 (4) 16.2 (43) 1.000
Prior other cancer 25.9 (7) 22.3 (59) 0.635
Operation 0.918
Lobectomy 51.9 (14) 53.6 (142)
Pneumonectomy 14.8 (4) 17.4 (46)
Bilobectomy 11.1 (3) 9.1 (24)
Sleeve 7.4 (2) 6.4 (17)
Segmentectomy 7.4 (2) 7.6 (20)
Wedge ... 2.6 (7)
Lung and chest wall resection 7.4 (2) 3.4 (9)
Outcomes
Pneumonia rate 3.7 (1) 8.7 (23) 0.710

Values are mean  SD or percentage (n).

COPD ¼ chronic obstructive pulmonary disease; DLCO ¼ diffusing capacity of lung for carbon monoxide; FEV1 ¼ forced expiratory volume in 1
second; POPP ¼ prevention of postoperative pneumonia.

thoracotomy (3.7% versus 8.7%), so future studies of oral pneumonia in this pilot study was to establish proof of
hygiene regimens may benefit from targeted in- concept and obtain a point estimate of the potential risk
terventions to increase adherence in this subpopulation, reduction of the intervention, facilitating an appropriate
focused surveys to understand reasons for noncompli- sample size calculation for a future study. A larger trial is
ance, and stratified analysis of these patients. In addition, needed to determine whether the encouraging pre-
although not statistically significant in this pilot study, we liminary results seen here are a true consequence of the
found that less adherent patients were slightly older, oral hygiene intervention. Also, the patients were not
more likely to be of a minority race, and had higher rates randomized in this single-arm study; patients who con-
of comorbidities. This analysis suggests that when sented may have been more willing to perform the
directing resources, vulnerable populations may benefit intervention and that may have introduced bias, partic-
from additional efforts to encourage the intervention. ularly with regard to adherence.
This study’s limitations merit discussion. First, this was These limitations are balanced by several major
a small pilot study that examined patient adherence and strengths. This was a prospective intervention with
was not powered to detect a statistical difference in the clearly defined, easy-to-measure endpoints of adherence
rates of postoperative pneumonia, which is the ultimate and pneumonia. A systematic review demonstrated that
clinical outcome of interest. The purpose of tracking self-reporting questionnaires like the patient diary
8 SEMENKOVICH ET AL Ann Thorac Surg
POSTOPERATIVE PNEUMONIA PREVENTION 2018;-:-–-

Table 5. Univariable Analyses: Open Operation Subgroup resection. Moreover, the encouraging trend of lower rates
Analysis of postoperative pneumonia suggests that this simple and
Univariable OR inexpensive intervention may be promising for reducing
Variable (95% CI) p Value morbidity after pulmonary resection. This pilot study
established proof of concept for a chlorhexidine tooth-
Part of POPP cohort (n ¼ 27) 0.40 (0.05–3.12) 0.386 brushing intervention that merits further investigation in
versus contemporaneous
cohort (n ¼ 265) a larger, multicenter trial, which is being planned.
Patient characteristics
Age 0.99 (0.95–1.02) 0.397 The authors wish to thank Margaret Olsen, PhD, MPH, for
Female versus male 0.64 (0.27–1.51) 0.304 providing statistical advice. Drs Semenkovich and Kollef were
supported by the Barnes Jewish Hospital Foundation. Drs
Nonwhite versus white 1.02 (0.22–4.61) 0.983
Semenkovich and Subramian were supported by National
Body mass index 0.92 (0.85–1.00) 0.051 Institutes of Health (NIH) grant 2T32HL7776-21. Dr Hudson was
Pulmonary function tests supported by NIH grant 5T32CA009621-27. Dr Puri was sup-
FEV1, % predicted 0.99 (0.96–1.01) 0.250 ported by NIH grant K07CA178120. This research was also sup-
ported by the Division of Cardiothoracic Surgery. Sage Products,
DLCO, % predicted 0.98 (0.96–1.01) 0.214 Inc (Cary, Illinois), contributed $10,000 in support, Peridex, and
Comorbidities toothbrushes for the study.
COPD 1.34 (0.56–3.18) 0.511
Smoker
Current versus never 0.51 (0.12–2.16) 0.441
Former versus never 1.07 (0.34–3.37) References
Lung cancer 0.27 (0.10–0.72) 0.008
1. Amar D, Munoz D, Shi W, Zhang H, Thaler HT. A clinical
Preoperative chemotherapy 1.25 (0.40–3.86) 0.700 prediction rule for pulmonary complications after thoracic
Preoperative radiation 1.14 (0.37–3.51) 0.819 surgery for primary lung cancer. Anesth Analg 2010;110:
Prior lung cancer 1.94 (0.41–9.22) 0.405 1343–8.
2. Deslauriers J, Ginsberg RJ, Piantadosi S, Fournier B. Pro-
Prior cardiothoracic surgery 1.02 (0.29–3.60) 0.979 spective assessment of 30-day operative morbidity for sur-
Cardiovascular disease 2.08 (0.84–5.11) 0.112 gical resections in lung cancer. Chest 1994;106(Suppl):329–30.
Diabetes mellitus 0.45 (0.10–1.98) 0.292 3. Allen MS, Blackmon S, Nichols FC, Cassivi SD, Shen KR,
Prior other cancer 0.29 (0.07–1.27) 0.100 Wigle DA. Comparison of two national databases for general
thoracic surgery. Ann Thorac Surg 2015;100:1155–62.
Operation (reference: lobectomy) 0.768 4. Villamizar NR, Darrabie MD, Burfeind WR, et al. Thoraco-
Pneumonectomy 1.22 (0.41–3.62) scopic lobectomy is associated with lower morbidity
Bilobectomy 0.88 (0.19–4.14) compared with thoracotomy. J Thorac Cardiovasc Surg
2009;138:419–25.
Sleeve 0.61 (0.08–4.95)
5. Shapiro M, Swanson SJ, Wright CD, et al. Predictors of major
Wedge 4.40 (0.776–24.95) morbidity and mortality after pneumonectomy utilizing The
Lung and chest wall resection 1.10 (0.13–9.28) Society of Thoracic Surgeons General Thoracic Surgery
Database. Ann Thorac Surg 2010;90:927–35.
CI ¼ confidence interval; COPD ¼ chronic obstructive pulmonary 6. Hu Y, McMurry TL, Isbell JM, Stukenborg GJ, Kozower BD.
disease; DLCO ¼ diffusing capacity of lung for carbon monoxide; Readmission after lung cancer resection is associated with a
FEV1 ¼ forced expiratory volume in 1 second; OR ¼ odds ratio; 6-fold increase in 90-day postoperative mortality. J Thorac
POPP ¼ prevention of postoperative pneumonia; VATS ¼ video- Cardiovasc Surg 2014;148:2261–7.e1.
assisted thoracoscopic surgery.
7. Watanabe S, Asamura H, Suzuki K, Tsuchiya R. Recent
results of postoperative mortality for surgical resections in
lung cancer. Ann Thorac Surg 2004;78:999–1002.
utilized capture adherence with moderate to high reli- 8. Paju S, Scannapieco FA. Oral biofilms, periodontitis, and
ability in the majority of studies [23]. Consequently, the pulmonary infections. Oral Dis 2007;13:508–12.
9. Gomes-Filho IS, Passos JS, Seixas da Cruz S. Respiratory
observed high rates of perioperative toothbrushing sug- disease and the role of oral bacteria. J Oral Microbiol 2010
gest that this is a feasible intervention for thoracic surgery Dec 21; [E-Pub ahead of print].
patients. Pneumonia was defined by STS criteria, and a 10. Bonten MJ, Bergmans DC, Ambergen AW, et al. Risk factors
chart review was performed for all patients during the for pneumonia, and colonization of respiratory tract and
stomach in mechanically ventilated ICU patients. Am J
study, ensuring accuracy of this metric. Given the sub- Respir Crit Care Med 1996;154:1339–46.
stantial morbidity and mortality associated with this 11. Garrouste-Orgeas M, Chevret S, Arlet G, et al. Oropharyn-
common complication, there is certainly a role for iden- geal or gastric colonization and nosocomial pneumonia in
tifying promising preventative interventions. Further- adult intensive care unit patients. Am J Respir Crit Care Med
more, our study shows significantly shorter lengths of 1997;156:1647–55.
12. El-Solh AA. Association between pneumonia and oral care in
stay in patients without pneumonia, and our preliminary nursing home residents. Lung 2011;189:173–80.
cost analysis demonstrates a potential cost savings of 13. Hua F, Xie H, Worthington HV, Furness S, Zhang Q, Li C.
more than $4,500 per pneumonia prevented. Hence, this Oral hygiene care for critically ill patients to prevent
intervention has the potential to be highly cost effective. ventilator-associated pneumonia. Cochrane Database Syst
Rev 2016;10:CD008367.
This pilot study demonstrated that this perioperative 14. Sj€
ogren P, Nilsson E, Forsell M, Johansson O, Hoogstraate J.
oral hygiene regimen can be implemented with high rates A systematic review of the preventive effect of oral hygiene
of adherence for patients undergoing planned lung on pneumonia and respiratory tract infection in elderly
Ann Thorac Surg SEMENKOVICH ET AL 9
2018;-:-–- POSTOPERATIVE PNEUMONIA PREVENTION

people in hospitals and nursing homes: effect estimates and 19. The Society of Thoracic Surgeons. General Thoracic Surgery
methodological quality of randomized controlled trials. J Am Database v.2.3 Training Manual 2018. Available at https://
Geriatr Soc 2008;56:2124–30. www.sts.org/sites/default/files/documents/gtsd_training_
15. Akutsu Y, Matsubara H, Shuto K, et al. Pre-operative dental manual _march2018.pdf. Accessed March 22, 2018.
brushing can reduce the risk of postoperative pneumonia in 20. Avritscher EBC, Cooksley CD, Rolston KV, et al. Serious
esophageal cancer patients. Surgery 2010;147:497–502. postoperative infections following resection of common solid
16. Soutome S, Yanamoto S, Funahara M, et al. Effect of peri- tumors: outcomes, costs, and impact of hospital surgical
operative oral care on prevention of postoperative pneu- volume. Support Care Cancer 2014;22:527–35.
monia associated with esophageal cancer surgery. Medicine 21. Wang J, Olak J, Ultmann RE, Ferguson MK. Assessment of
(Baltimore) 2017;96:e7436. pulmonary complications after lung resection. Ann Thorac
17. Fourrier F, Cau-Pottier E, Boutigny H, Roussel-Delvallez M, Surg 1999;67:1444–7.
Jourdain M, Chopin C. Effects of dental plaque antiseptic 22. Edmiston CE, Bruden B, Rucinski MC, Henen C,
decontamination on bacterial colonization and nosocomial Graham MB, Lewis BL. Reducing the risk of surgical site
infections in critically ill patients. Intensive Care Med infections: does chlorhexidine gluconate provide a risk
2000;26:1239–47. reduction benefit? Am J Infect Control 2013;41(Suppl):49–55.
18. Munro CL, Grap MJ, Jones DJ, McClish DK, Sessler CN. 23. Shi L, Liu J, Koleva Y, Fonseca V, Kalsekar A, Pawaskar M.
Chlorhexidine, toothbrushing, and preventing ventilator- Concordance of adherence measurement using self-reported
associated pneumonia in critically ill adults. Am J Crit Care adherence questionnaires and medication monitoring
2009;18:428–37; quiz 438. devices. Pharmacoeconomics 2010;28:1097–107.