1913

Journal of Food Protection, Vol. 80, No. 11, 2017, Pages 1913–1923
doi:10.4315/0362-028X.JFP-17-113
Copyright Ó, International Association for Food Protection

Research Paper

Evaluation of Third-Party Deep Cleaning as a Listeria

monocytogenes Control Strategy in Retail Delis
SUSAN R. HAMMONS,1 ANDREA J. ETTER,1,2 JINGJIN WANG,1 TONGYU WU,1 THOMAS FORD,3
MICHAEL T. HOWARD,3 AND HALEY F. OLIVER1*

1Department of Food Science (ORCID: http://orcid.org/0000-0001-6756-8790 [T.W.]) and 2Purdue Interdisciplinary Life Sciences Program, Purdue
University, West Lafayette, Indiana 47907; and 3Ecolab, Inc., Greensboro, North Carolina 27409, USA

MS 17-113: Received 11 March 2017/Accepted 27 June 2017/Published Online 20 October 2017

ABSTRACT
The objective of this study was to develop and assess the efficacy of an aggressive deep cleaning sanitation standard
operating procedure (DC-SSOP) in nine retail delicatessens to reduce persistent Listeria monocytogenes environmental
contamination. The DC-SSOP was developed from combined daily SSOPs recommended by the Food Marketing Institute and
input from experts in Listeria control from food manufacturing and sanitation. The DC-SSOP was executed by a trained
professional cleaning service during a single 12-h shutdown period. A modified protocol from the U.S. Food and Drug
Administration Bacteriological Analytical Manual was used to detect L. monocytogenes in samples from 28 food and nonfood
contact surfaces that were collected immediately before and after each cleaning and in samples collected monthly for 3 months.
The DC-SSOP significantly reduced L. monocytogenes prevalence overall during the 3-month follow-up period and produced
variable results for persistent L. monocytogenes isolates. Six delis with historically low to moderate L. monocytogenes prevalence
had no significant changes in the number of samples positive for L. monocytogenes after deep cleaning. Deep cleaning in very
high prevalence delis (20 to 30% prevalence) reduced L. monocytogenes by 25.6% (Padj , 0.0001, n ¼ 294) overall during the
follow-up period. Among delis with extremely high prevalence (.30%), positive samples from nonfood contact surfaces were
reduced by 19.6% (Padj ¼ 0.0002, n ¼ 294) during the follow-up period. The inability of deep cleaning to completely eliminate
persistent L. monocytogenes was likely due to the diverse infrastructures in each deli, which may require more individualized
intervention strategies.
Key words: Deep cleaning; Listeria monocytogenes; Persistence; Retail deli; Sanitation standard operating procedure

Listeria monocytogenes can be found on at least one
food or nonfood contact surface in the majority of retail
delicatessens (delis) in the United States that have been
studied to date (2, 6, 11, 12). In a cross-sectional study of
121 retail establishments in New York State, 63% had at
least one L. monocytogenes–positive environmental sample
(11). In 58% of small establishments and establishments
with a history of failing New York State Agriculture and
Markets inspections, L. monocytogenes was found in at least
one environmental sample (6). In a longitudinal study of 30
delis in large retail stores among three states, we found L.
monocytogenes at least once in 90% of delis (12). Within-
establishment prevalence varies, but contamination was
found on 0 to 40% of surfaces tested (6, 11, 12), and
contamination was more common on nonfood contact
surfaces (NFCS). Many daily sanitation standard operating
procedures (SSOPs) in delis are targeted at control of L.
monocytogenes contamination on food contact surfaces
(FCS) (e.g., counter tops, slicers, and trays) (14). As a
result, L. monocytogenes has been detected on only 2 to 4%
* Author for correspondence. Tel: 765-496-3913; Fax: 765-494-7953;
E-mail: hfoliver@purdue.edu.
of FCS but on 15 to 20% of NFCS (e.g., drains, floors, and
hoses) (6, 11, 12). This difference may be due in part to the
open nature of a retail deli, where L. monocytogenes may
enter the deli on customers’ and workers’ shoes and on cart
wheels, raw meats, fresh produce, food packaging, and
ready-to-eat meats handled in the deli.
We recently tested changes to daily SSOPs developed
by the Food Marketing Institute Listeria Working Group in
30 retail delis among three states and determined that the
intervention were largely ineffective for reducing the
persistence of L. monocytogenes (2). Deli environments
with L. monocytogenes on .10% of FCS and NFCS
cumulatively preintervention did not improve (2, 12). We
hypothesized that these facilities require more aggressive
mitigation strategies to effectively eliminate or reduce L.
monocytogenes contamination from the deli environment.
We then developed and tested a deep cleaning SSOP (DC-
SSOP) aimed at reducing prevalent and ideally eliminating
persistent L. monocytogenes isolates in delis using an
approach based on that for manufacturing and processing
plants. Delis were selected from the group of delis in which
environmental L. monocytogenes monitoring was conducted
1914 HAMMONS ET AL.
for at least 12 months, as described by Simmons et al. (12)
and Etter et al. (2).
MATERIALS AND METHODS
Overall study design. A DC-SSOP was developed and tested
in nine retail full service delis that had participated in our previous
studies (2, 12). Delis were selected (i) to represent historically low
to moderate (0 to 10%) and high (.10%) L. monocytogenes
prevalence and (ii) to be geographically representative among three
U.S. states. Delis in this study were identified by the number
assigned in our previous study of 30 delis (2, 12). A third-party
cleaning service executed the DC-SSOP in all delis. Environmental
sponge samples (n ¼ 28) were collected from FCS, NFCS, and
transfer points (TP) immediately before and after execution of the
deep cleaning to detect L. monocytogenes and other Listeria spp.
Longitudinal follow-up samples were collected from the same sites
during operational hours once monthly for 3 months after the deep
cleaning.
Development process for deep cleaning procedure. Our
group in cooperation with the corporate sanitarians of participating
retailers and sanitarians from large ready-to-eat meat manufacturers
recognized for innovative L. monocytogenes controls developed
the deep cleaning protocol. The Food Marketing Institute guidance
for control of Listeria monocytogenes (3) was consulted as a guide
for the general structure of the DC-SSOP that was designed to
include all regular deli sanitation procedures, including those
conducted less frequently (e.g., monthly or quarterly) in sequence
to minimize cross-contamination of cleaned surfaces. Revisions
were made to the DC-SSOP (see Supplemental Materials) after
executing the first deep cleaning to reflect the division of labor by
teams in zones, which were not captured in the original document.
Preparation and scope of DC-SSOP. A 10-person cleaning
crew hired through a third-party cleaning service was trained on the
DC-SSOP by Purdue University personnel with additional support
from the retailers’ sanitation chemical providers. The same
personnel executed the DC-SSOP in all participating delis. The
DC-SSOP was completed during a single 12-h overnight shutdown
period per deli to minimize business disruptions. Equipment used
strictly for prepared foods or hot products (e.g., proofing cabinets,
ovens, and fryers) were considered out of scope; this equipment
was cleaned and sanitized on the exterior only to reduce potential
surface contamination. When deli department traffic flow inter-
mingled with adjacent departments (e.g., bakery), the floors in both
areas were included in the deep cleaning to control cross-
contamination between NFCS in both departments. Equipment
dedicated to adjacent departments was not addressed in the DC-
SSOP.
Planning responsibilities were divided among four parties:
retailer delis, sanitation provider, third-party cleaning service, and
Purdue University researchers. Each retailer communicated the
planned intervention vertically through the company, scheduled
additional labor to support removal of dry goods and food products
before deep cleaning and to assist deli set up after deep cleaning,
and provided maintenance staff to support each deep cleaning. The
regular sanitation provider obtained and delivered additional
chemicals and special tools (e.g., wet-dry vacuum, foaming
detergent dispenser, and additional hoses) to the deli in advance.
The third-party cleaning service provided personal protective
equipment for their personnel. Purdue University research staff
collaborated with the sanitation provider to train the third-party
cleaning service with the DC-SSOP during a 3-h training session
J. Food Prot., Vol. 80, No. 11
the day before executing the DC-SSOP in the first deli. Research
staff also organized and conducted SSOP execution and verifica-
tion with ATP testing and collected samples to detect L.
monocytogenes and other Listeria spp.
Execution of the DC-SSOP. Execution of tasks throughout
the deep cleaning was organized by (i) responsible party, (ii)
priority, and (iii) estimated timeline for completion (Fig. 1). Deli
associates removed all food products, dry goods, and records
before the third-party cleaning service arrived. Food products were
moved to coolers in the dairy or produce departments to maintain
the cold chain; storage of ready-to-eat products in areas where raw
meat was handled (i.e., meat or seafood departments) was not
allowed. Foot traffic through raw meat areas was strongly
discouraged, and when traffic was unavoidable owing to water
access or other essential supplies, this traffic was restricted to
managers. As opposed to a Listeria recovery process where all
open products would be discarded after detection of Listeria on
food products, this study included no evidence of contaminated
foods, and open products were retained at the retailer’s discretion.
Labor was divided into three teams working in zones across the
deli, with tasks in each zone prioritized to allow a large cleaning
crew to work simultaneously and effectively (Fig. 2). Environ-
mental samples were collected for testing for L. monocytogenes
and Listeria spp. after food products were removed and before
cleaning began. To protect sensitive electronics from potential
water damage, deli scales were removed, cleaned, and sanitized out
of place by a designated team member and reinstalled after the
comprehensive application of sanitizer to the deli department. The
deli case and cold storage room were high priority surfaces and
were designated for extra labor early in the deep cleaning process
to allow these units sufficient time to return to refrigeration
temperature (14). Deli cases were cleaned down to the coils (this
was critical in deli case models where food debris falls through
grates below and behind the coils). Fan motors inside the deli cases
were removed if possible or were protected with plastic to prevent
water damage.
After the deli case and cold storage room were cleaned and
sanitized, focus shifted to the customer service areas and rear
preparation areas of the deli. Rooms were cleaned top to bottom,
back to front; walls were scrubbed from the ceiling down before
FCS were cleaned (e.g., countertops). Drains were cleaned by a
designated crew member with designated gloves, brushes, and a
chlorinated cleaner bucket. After all utensils and small FCS
equipment had been washed, rinsed, and sanitized in the three-
compartment sink, the sink interior and exterior were cleaned a
second time. All deli floors were cleaned concurrently with
chlorinated cleaner agitated by deck brushes and rinsed with new
foam squeegees to control water flow. Sanitizer was applied to all
deli surfaces (including FCS and other areas) that had previously
been cleaned and sanitized individually. Samples were taken after
application of sanitizer but before full reassembly of the deli (e.g.,
before cold storage room racks and scales were replaced in
operational areas).
Cleaning tools and calibration of cleaners and sanitizers.
Chemical dispensing systems were calibrated to higher concentra-
tions within label use specifications by the sanitation provider for
foaming chlorinated cleaner and dispensing quaternary ammonium
sanitizer at disinfectant concentration (400 ppm). New deck
brushes, scrub brushes, and other tools were labeled by contact
surface (i.e., FCS, NFCS, and drain) with a permanent paint marker
then soaked in disinfectant concentration sanitizer for 10 min
before use. Tools were returned to the sanitizer solution when not
J. Food Prot., Vol. 80, No. 11 CONTROL OF L. MONOCYTOGENES AT RETAIL 1915

FIGURE 1. Progressive deep cleaning timeline. Time progresses left to right with tasks color coded by zone (team 1, green; team 2,
orange; team 3, purple). Red and blue are research or administrative tasks. Black vertical line marks the time when cleaning of all FCS
and equipment was completed. Floors and drains were then cleaned in all zones.

in use to minimize the microbial load. Hand sinks and three-
compartment sinks were cleaned first. Sanitizer boot dip stations
(400 ppm of quaternary ammonia) were placed at each entrance to
the deli department and used by all crew members upon entering or
exiting the deli area. Sanitizer in both boot dip stations and tool
storage baths was changed every 2 h or more frequently as needed.
Detection and isolation of L. monocytogenes and Listeria
spp. FCS, NFCS, and TP (n ¼ 28) selected to represent potential
harborage areas for L. monocytogenes and other Listeria species
were sampled immediately before execution of SSOP, after final
application of sanitizer, and then once per month for 3 months after
deep cleaning as previously described (2, 12).
Verification of clean surfaces. The AccuPoint2 ATP
Sanitation Monitoring System (Neogen Corp., Lansing, MI), a
rapid test for ATP, was used to detect organic soils on surfaces
throughout the deep cleaning process (5). In-process ATP swab
samples were collected from a total of 65 FCS, NFCS, and TP sites
(see Supplemental Materials), including the 28 sites sampled for
environmental L. monocytogenes and additional hard-to-clean
areas. ATP results were reported as relative light units (RLU).
Surfaces with RLU above previously determined thresholds (2, 10)
were recleaned until passing values were obtained, i.e., ,150 RLU
for FCS or TP and ,300 RLU for NFCS.
PFGE typing of isolates. One isolate from each L.
monocytogenes–positive site was selected for DNA fingerprinting

FIGURE 2. DC-SSOP division of labor and priorities by zone. Tasks are color coded by zone to match Figure 1 (team 1, green; team 2,
orange; team 3, purple; all teams, white) and prioritized from highest to lowest priority, top to bottom and left to right. An exception was
the deli case, which was highest priority for the customer service area, and deep cleaning of this area was completed before other tasks
began.
1916 HAMMONS ET AL.
by pulsed-field gel electrophoresis (PFGE). Isolates were typed
using the standard protocol provided by the Centers for Disease
Control and Prevention PulseNet protocol (updated March 2013;
http://www.cdc.gov/pulsenet/PDF/listeria-pfge-protocol-508c.pdf)
modified to include extended cell lysis (4 to 16 h) and restriction
digestion (2 to 5.5 h) steps. PFGE patterns were analyzed and
compared with the Cornell University Food Safety Laboratory L.
monocytogenes collection of ~10,000 isolates from retail,
manufacturing, and natural environments using BioNumerics v.
6.0 software (Applied Maths, Austin, TX). Dendrograms assessing
isolate relationships were created using the Dice coefficient and the
unweighted pair group method with arithmetic means (2, 12).
Isolates from the same deli with the same pulsotype (i.e., AscI-ApaI
fingerprint) that were found on at least three sampling dates over
the course of this study were considered persistent (see
Supplemental Materials) (12).
inlA PMSC detection. Evaluation for inlA premature stop
codons (PMSCs) was conducted as previously described (15–17)
for all 107 L. monocytogenes isolates that were PFGE typed (see
Supplemental Materials). Because of the small number of isolates
and low frequency of PMSCs, analyses were limited to descriptive
statistics.
sigB allelic typing. One representative presumptive L.
monocytogenes or Listeria spp. isolate from each sample was
characterized by sigB allelic typing as previously described (7, 12)
to verify species identification (see Supplemental Materials).
Statistical analyses. Environmental monitoring data were
classified into four periods: (i) previous longitudinal testing (once
per month for 6 months) (2), (ii) immediately before deep
cleaning, (iii) immediately after deep cleaning, and (iv) follow-up
testing (once per month for 3 months). Prevalence was calculated
as the number of samples positive for L. monocytogenes divided
by the total number of samples tested. Long-term impacts were
assessed by comparing the mean monthly L. monocytogenes
prevalence after implementation of daily SSOPs as reported by
Etter et al. (2) with the mean monthly prevalence found through
follow-up testing after deep cleaning. We elected to compare
mean monthly L. monocytogenes prevalence because no trends in
monthly sampling prevalence (repeated measures) were observed.
As described by Etter et al. (2), a generalized linear mixed effects
model was constructed using the Proc Glimmix procedure with a
Gaussian distribution in SAS 9.4 (SAS Institute, Cary, NC) to
evaluate the impact of the deep cleaning on L. monocytogenes
prevalence. Preintervention prevalence category (low, ,1%;
moderate, 1 to 10%; high, 10 to 20%; very high, 20 to 30%;
extremely high, .30%) (2), period, surface, and their interactions
were designated as fixed effects, and deli (nested within category)
and period 3 deli(category) were random effects to estimate
prevalence for each monthly sampling event in a deli. Analysis of
the three-variable interaction of category 3 surface 3 period (n ¼
294) and the two-variable interaction of category 3 period (n ¼
98) was also performed. Arcsine square root transformation was
applied to the two-variable analyses to yield better model fit, and
untransformed data were used for the three-variable analysis.
Bonferroni’s adjustment was applied to the a ¼ 0.05 for each least
squares means family of comparisons for variable-level differ-
ences: period aadj ¼ 0.0167; period 3 category aadj ¼ 0.00625;
surface 3 category aadj ¼ 0.0028; surface 3 period 3 category aadj
¼ 0.00208.
J. Food Prot., Vol. 80, No. 11
RESULTS AND DISCUSSION
In this study, we developed and tested a DC-SSOP
aimed at reducing or mitigating persistent L. monocytogenes
contamination in retail delis. Our data indicate that the deep
cleaning (i) had variable effects on L. monocytogenes
prevalence in delis with historically low to moderate
prevalence, (ii) reduced L. monocytogenes prevalence in
delis with .10% historical prevalence, and (iii) had variable
effects on eliminating persistent L. monocytogenes strains.
We also found that ATP testing is a useful tool for verifying
in-process cleaning efficacy in delis. Below we discuss our
findings, the study limitations, challenges, and future
directions.
Deep cleaning interventions had variable effects on
L. monocytogenes prevalence in low and moderate
prevalence delis. We selected six delis (delis 5, 8, 12, 13,
17, and 22; see Supplemental Materials) with historically
low to moderate environmental L. monocytogenes contam-
ination (0 to 10% prevalence) (2, 12) to test the DC-SSOP.
Delis with lower prevalence were selected to determine
whether major disruption of the deli would expose
previously undetected L. monocytogenes niches. Experienc-
es in manufacturing facilities indicate that movement of
equipment, construction, or other major disruptions in the
environment can expose niches harboring L. monocytogenes
and other foodborne pathogens. Detection of more L.
monocytogenes–positive samples after SSOP execution in
these delis would indicate that the execution of the SSOP
increased the risk of cross-contamination in previously well-
controlled environments.
L. monocytogenes results for all environmental samples
collected by our group, including those from previous work
(2, 12), were compiled for each low to moderate prevalence
deli (see Supplemental Materials). No significant changes
were observed among these delis, but decreases in the
number of positive samples were observed in delis 8 and 13
immediately after deep cleaning (Fig. 3 and Supplemental
Materials). In deli 13, 1 of 26 original samples was positive
for L. monocytogenes; after deep cleaning, there was no
detectable L. monocytogenes. Prevalence was reduced from
4 of 26 to 2 of 27 samples in deli 8. L. monocytogenes
prevalence increased in deli 17 from 0 of 25 sites tested
before deep cleaning to 1 of 26 sites (a trash can; Fig. 3 and
Supplemental Materials) after deep cleaning. Detection of L.
monocytogenes on the trash can exterior after the deep
cleaning but not before could indicate that the SSOP
exposed a harborage point. However, because exact sample
collection points were not marked (to prevent targeted
cleaning during the deep cleaning process), after the deep
cleaning a different trash can that still harbored L.
monocytogenes may have been sampled.
The long-term effects of the DC-SSOP were assessed
through monthly operational environmental sample collec-
tion for 3 months after the deep cleaning. No L.
monocytogenes was detected in delis 13 (Fig. 3) and 22
(Supplemental Materials) after deep cleaning. Overall, the
prevalence in deli 13 after deep cleaning was reduced
J. Food Prot., Vol. 80, No. 11
compared with the preintervention prevalence (12) and after
implementation of daily SSOP changes (2) (Fig. 3), and
previously persistent strains (CU-11-282 and CU-314-342)
were not recovered (see Supplemental Materials). One L.
monocytogenes–positive sample was detected in deli 12
during follow-up sampling, but none were recovered
immediately before or after deep cleaning (see Supplemental
Materials). Sporadic L. monocytogenes isolates were
detected on FCS and NFCS in delis 5 and 17 (Fig. 3 and
Supplemental Materials), which is characteristic for retail
delis in which the open environment may allow regular
introduction of contamination during the operational hours
when samples were collected (6, 11, 12). These two delis did
not historically have persistent strains (2, 12).
In deli 8, L. monocytogenes prevalence increased
overall after deep cleaning compared with the prevalence
preintervention (12) and after implementation of the daily
SSOPs (Fig. 3) (2). Although prevalence was reduced from
4 of 26 samples before deep cleaning to 2 of 27 samples
immediately after (Fig. 3), CU-8-340 persisted throughout
the study, from preintervention (12) through postimplemen-
tation of daily SSOPs studies (2) and immediately before,
immediately after, and during deep cleaning follow-up
sampling (see Supplemental Materials). Prevalence of this
persistent strain increased after deep cleaning; 15 isolates
were recovered in this study compared with 8 isolates after
implementation of daily SSOPs (2) and 10 in preintervention
testing (12). This strain was routinely isolated from the
single-basin sink interior, an FCS site (2, 12).
Prevalence was reduced by up to 26% in very high
prevalence delis after deep cleaning. At the three-variable
level (n ¼ 294), surface and category 3 surface significantly
affected L. monocytogenes prevalence (P , 0.0001),
although the effect of surface on L. monocytogenes
prevalence depended upon the preintervention prevalence
level. Period (P ¼ 0.0001), category (P ¼ 0.0029), and
category 3 period were also significant (the effect of period
on L. monocytogenes prevalence depended upon the
preintervention prevalence level, P ¼ 0.0006). The category
3 period 3 surface interaction was not consequential overall
(P ¼ 0.0646). A 10.6% reduction in prevalence was found
between postimplementation of daily SSOPs and immedi-
ately before deep cleaning (Padj ¼ 0.0013). Although the
deep cleaning had no immediate effect, it reduced L.
monocytogenes prevalence by 9.3% over the 3-month
CONTROL OF L. MONOCYTOGENES AT RETAIL 1917
FIGURE 3. Prevalence of L. monocyto-
genes by deli across studies from preinter-
vention (12) through postintervention (2).
follow-up period compared with the prevalence after
implementation of daily SSOPs (2) (Padj , 0.0001).
Prevalence in a very high prevalence deli (deli 2, 20 to
30% average prevalence; Fig. 4) decreased 25.6% (Padj ,
0.0001) overall. Prevalence on FCS in very high prevalence
delis was reduced by 27.3% 6 7.4% (Padj ¼ 0.0003) and
prevalence on TP was reduced by 38.9% 6 7.4% (Padj
, 0.0001) in each monthly sampling event following deep
cleaning compared with immediately after implementation
of daily SSOPs (2). However, four strains (CU-11-320, CU-
55-266, CU-262-79, and CU-8-96) persisted from preinter-
vention (12) through postintervention (2), although none of
these four strains were present immediately before deep
cleaning. This finding may have been due to differences in
sites sampled or viable but nonculturable cells present before
deep cleaning. CU-8-96, CU-55-266, and CU-262-79 were
recovered immediately after deep cleaning or from samples
collected during the follow-up months (Fig. 4). CU-11-320
was repeatedly isolated 1 month after deep cleaning. The
prevalence of CU-11-320 was higher (12.3%, 13 of 106
samples) after deep cleaning than during the preintervention
period (12) or after implementation of daily SSOPs (2.6%, 4
of 154 samples) (2) potentially because of exposure of and
continued harborage in a niche or the elimination of
competing persistent strains. This strain has been found in
multiple delis across several states and is frequently
persistent (2, 12), so it is possible, although not probable
in this instance, that this strain was reintroduced into the deli
after deep cleaning.
L. monocytogenes prevalence on NFCS in delis with
extremely high preintervention prevalence (delis 23 and 28;
.30%) (12) was reduced 19.6% 6 5.3% (Padj ¼ 0.0002).
However, persistent strains were not uniformly reduced or
eliminated. In deli 23, which had .30% prevalence in all
studies (2, 12), four persistent strains (CU-258-322, CU-
258-323, CU-259-322, and CU-262-317) were detected
longitudinally before deep cleaning (2, 12) (Fig. 5).
Prevalence of CU-258-322 was reduced immediately after
deep cleaning but increased over the 3 months after the
initial deep cleaning sampling. CU-258-323 was not isolated
again after deep cleaning. CU-259-322 was persistent in
preintervention testing (12) and transient after implementa-
tion of daily SSOPs (2) but was not recovered in the present
study. CU-262-317 was persistent after implementation of
daily SSOPs (2) but was not found during the present study.
 1918
HAMMONS ET AL.

FIGURE 4. Longitudinal PFGE map of deli 2 showing operational (12), preintervention (12), postintervention (2), and deep cleaning prevalence and persistence of L. monocytogenes.
J. Food Prot., Vol. 80, No. 11
 J. Food Prot., Vol. 80, No. 11

FIGURE 5. Longitudinal PFGE map of deli 23 showing operational (12), preintervention (12), postintervention (2), and deep cleaning prevalence and persistence of L. monocytogenes.
CONTROL OF L. MONOCYTOGENES AT RETAIL
1919
1920 HAMMONS ET AL.
Deli 28, which had .30% L. monocytogenes preva-
lence preintervention (12) and after implementation of daily
SSOPs (2), harbored two persistent strains, CU-262-79 and
CU-262-319 (2, 12). Both were isolated before and after
deep cleaning and were present in two subsequent months
with approximately equal prevalence. However, prevalence
of these strains was reduced compared with preintervention
(12) and after implementation of daily SSOPs (2, 12) (see
Supplemental Materials).
At the two-variable level (n ¼ 98), category (P ¼
0.0155), period (P ¼ 0.0020), and category 3 period (P ¼
0.0058) were all significant. Similar to the three-variable
outcomes, a reduction of 4.0% was observed immediately
before deep cleaning compared with after implementation of
daily SSOPs (2) (Padj ¼ 0.0027). No significant change in L.
monocytogenes prevalence was detected immediately after
deep cleaning (Fig. 3), yet a reduction of 1.5% was
significant in the long term (Padj ¼ 0.0061); no specific
effect was consequential (n ¼ 98).
Further analysis of the category 3 surface interaction,
which had an overall significance of P , 0.0001, suggested
that L. monocytogenes was 36.8% less prevalent on FCS of
low prevalence delis than in their moderate counterparts
(P adj ,0.0001). L. monocytogenes on FCS of low
prevalence delis was also 38.1% less prevalent than on
FCS of extremely high prevalence delis (Padj , 0.0001) (see
Supplementary Materials).
Although a prevalence reduction of 10.6% was detected
immediately before deep cleaning compared with after
implementation of daily SSOPs (2) (P ¼ 0.0013), these data
should be interpreted with caution. Nearly 1 year of
unmonitored deli operations occurred between longitudinal
data collected before deep cleaning (2) and the data collected
immediately before deep cleaning in this study. Apparent
reductions in prevalence may be due to the fact that the
prevalence immediately before deep cleaning was derived
from a single sampling event that may not be representative
of the environment over time. Undocumented changes in
deli environments also may have impacted prevalence
during the sampling hiatus.
Many of the persistent strains we studied were highly
prevalent across multiple surfaces in the deli environment
over almost 2 years (2, 12). Consequently, these strains
likely represent true persistence rather than repeated
reintroduction (9, 13). We hypothesized that persistent
strains found in only one study may be (i) less efficient at
colonizing the deli environment, (ii) outcompeted by an
established highly persistent strain, (iii) located in delis with
more effective deli sanitation, or (iv) a combination of
factors. Conversely, strains that persisted across all three
studies (2, 12) may be highly capable of colonizing and
persisting in the retail deli environment and in response to
increased sanitation challenges within the deli. A study by
our group of persistent and transient retail deli isolates
collected in the preintervention study (12) revealed overall
higher attachment capacity among persistent isolates but no
difference in sanitizer tolerance (17). However, in a follow-
up study using cryo scanning electron microscopy for two
persistent and two transient isolates, a transient strain had
J. Food Prot., Vol. 80, No. 11
greater biofilm forming capacity (1). Further work is needed
to elucidate the factors influencing L. monocytogenes
persistence in the retail deli environment.
There was no change in prevalence of inlA PMSCs
after deep cleaning. Of the 107 representative L.
monocytogenes isolates evaluated for inlA PMSC in this
study, 2 (1.9%) had PMSCs (see Supplemental Materials).
One isolate was found immediately before deep cleaning on
an FCS in deli 13; no other L. monocytogenes isolates were
detected in deli 13 in this study. The other isolate with
PMSCs was found immediately after deep cleaning on an
NFCS in deli 17. The prevalence of inlA PMSCs in this
study was comparable to that in the preintervention study,
where 2.3% of isolates (10 of 442) from the retail deli
environment had inlA PMSCs (17). Prevalence of inlA
PMSCs on FCS, NFCS, and TP was also comparable, with
8.3% (1 of 12), 1.1% (1 of 95), and 0% of isolates,
respectively (17). In contrast, after implementation of daily
SSOPs, we found fewer isolates with inlA PMSCs (4 of 426;
0.94%), and PMSC-containing isolates were primarily found
on NFCS (2 of 4 PMSC-containing isolates) (2).
As in our previous studies (2, 17), the PMSC-containing
isolates in this study represented transient lineage II strains
(4, 8). One was a novel pulsotype, and the other had CU-
200-227, a pattern previously associated with three PMSC-
containing isolates found on FCS in three different delis
from the same geographic region (17). All four isolates
contained the same PMSC mutation (PMSC-4) (15), were
found on FCS, and were located in delis within the same
geographic location. Consequently, the strain may have
originated from shared manufacturers and may represent
cross-contamination from food into the deli environment.
ATP as a verification method and personnel
motivation tool. A total of 1,087 ATP swab samples (99
to 139 per deli) (Table 1) were used to monitor cleaning
efficacy during the deep cleaning process. The wide range of
swabs required was largely because of the varying level of
soils in each deli (i.e. low prevalence delis had less soil
buildup than did high prevalence delis) and differences in
deli size, layout, and equipment. The slicer was a key
variance point among delis. Of the 65 ATP test sites, 5 were
designated slicer test areas, each of which may have been
tested multiple times per slicer. The number of slicers
present had a large impact on the number of swabs used per
deli. Deli 2 had only three slicers, and 99 ATP swabs were
used in total, whereas deli 17 had nine slicers, and 134 ATP
swabs were used (Table 1). Deli 28 was one of the most
heavily soiled delis, as indicated by an average 34.1% L.
monocytogenes prevalence across the previous studies (Fig.
3) (2, 12), and required the most ATP swabs (139) to verify
cleaning (Table 1). Anecdotally, deep cleaning crew
members were motivated by passing ATP scores during
operations. Competition began among employees to achieve
ATP scores below the detection limit. This competition
generated enthusiasm and motivated the crew to clean more
effectively through a 12-h overnight shift.
J. Food Prot., Vol. 80, No. 11
TABLE 1. In-process ATP swab samples collected per deli and
the order in which delis were cleaned
Delia No. of ATP samples collected Order of cleaning
2 99 7 frequently required more than 5 h for a three-person crew to
5 102 1 clean (15 labor hours). Poor design often compounded
8 115 3 issues. In some deli cases, food particles could fall through
13 116 2
23 124 8
12 125 5
22 133 6 were disassembled, which could not be done because of
17 134 4 personnel availability and time constraints. Removal of the
28 139b 9 top plate to completely expose refrigeration coils and
a
Deli number from Simmons et al. (12).
b
On-hand swab supply was exhausted.
Facility design differences challenged the robust-
ness of the DC-SSOP. Facility design and condition were
the most common inhibitors of effective sanitation. Execution
of the DC-SSOP did not open niches of L. monocytogenes in
well-controlled environments (e.g., delis with historically low
prevalence). However, L. monocytogenes prevalence in-
creased after execution of the SSOP in a deli with previously
high prevalence, and efficacy in other high prevalence delis
was mixed. We postulate that these various results are
attributed to the complexity of the deli environment and
physical limitations of the study design. Each participating
deli had a unique structural design and different equipment.
Although the cleaning crew was trained with the DC-SSOP,
every execution had unique challenges. Poorly sloped floors,
partially or completely blocked floor drains, worn grout, and
damaged flooring regularly created challenges for water
control throughout the deep cleaning process in every deli.
Even in deli 12, which had low prevalence and relatively
minimal soils, a clogged cold storage room floor drain
significantly disrupted execution of the DC-SSOP. Standing
water accumulated in the immediate area within 30 min, a
second drain backed up within 90 min, and the crew tracked
heavily soiled gray water throughout the environment for 3.5
h while waiting for the emergency plumber to arrive. In an
ideal situation, cleaning would have been halted while the
drain issue was resolved, but because of the limited allotted
time and business needs of the participating retailers, the crew
continued cleaning to complete the DC-SSOP before the deli
opened for business in the morning. Delis with free-flowing
drains were not free of issues. Overall, 32 (35.6%) of 90
standing water samples tested positive for L. monocytogenes
from preintervention operational sampling (12) through the
sampling after deep cleaning (Table 2).
Deli case and cold storage room require the most
labor hours because of unforeseen challenges. Irregularly
cleaned deli cases and cold storage rooms required more
labor hours and resources than any other area. Prevalence
was 9.3% (5 of 54 samples) on the deli case, 14.3% (1 of 7
samples) on the deli case by raw meat, and 5.6% (3 of 54
samples) on the deli case tray after implementation of daily
SSOPs (Table 2) (2). No L. monocytogenes was found on
the deli case by raw meat or deli case trays immediately after
CONTROL OF L. MONOCYTOGENES AT RETAIL 1921
the deep cleaning and during the follow-up sampling period
(Table 2). One positive sample was found on a deli case
immediately after deep cleaning, but none were found
during the follow-up period. Heavily soiled deli cases
vents into the coil compartment, but the coils were covered
by plating that could be removed only when the cases sides
removal of fan blades and motors often required tools and
knowledge beyond that of deli managers or associates. The
presence of an attentive and knowledgeable maintenance
person greatly facilitates the disassembly of both deli cases
and cold storage room refrigeration units.
The cold storage room was a significant harborage site
for L. monocytogenes even after implementation of daily
SSOPs; 16 (29.6%) of 54 cold room samples, 14 (38.9%) of
36 drain samples, and 4 (7.4%) of 54 wall samples were
positive (Table 2) (2). The deep cleaning did not
significantly reduce prevalence on these sites; 9 (33.3%) of
27 cold room floor samples, 4 (16.7%) of 24 drain samples,
and 1 (3.7%) of 27 wall samples were positive during the
follow-up period (Table 2). The moderately sized (10 by 7-ft
[3 by 2-m]) cold storage room with built-in steel racks in
deli 2 required more than 7 h of cleaning by a two- or three-
person crew using lime remover to eliminate heavy mineral
deposits and soils.
Study design limitations. Resources limited the scope
of the study to nine delis. The power to detect significant
differences was reduced because of the need (i) to maintain
the blindness of the study, (ii) to represent delis with both
high and low L. monocytogenes prevalence, and (iii) to
respect the investment of the retailers in this study.
Significant resources were invested in a professional
cleaning crew, which for consistency had to travel to three
states over 3 weeks. Despite having a well-trained crew of
10 people, additional personnel would have been beneficial.
Significant costs were borne by each deli for labor (to empty
and restock), chemicals, potential lost sales while the deli
was closed, and maintenance. The 12-h shutdown period
was negotiated in advance with participating delis and could
not be violated, even when more time was needed to
effectively clean large and/or heavily soiled delis.
Another design limitation was that multiple surfaces
matched each sample site description (e.g., there was more
than one slicer per deli); without permanently marking
sample sites, the ‘‘before’’ and ‘‘after’’ samples may have
been collected from different places (e.g., different slicers).
Therefore, elimination or appearance (as in deli 2; Fig. 4) of
a positive sample after deep cleaning compared with before
deep cleaning could be the result of a change in test area
rather than the impact of the DC-SSOP. We recognized this
issue at the beginning of the study and accepted this
limitation because risk of site change would present results
more relevant to commercial application of the DC-SSOP
1922 HAMMONS ET AL. J. Food Prot., Vol. 80, No. 11

TABLE 2. Prevalence of L. monocytogenes by site across phases in the nine participating delis
% samples positive for L. monocytogenes (no. positive/no. collected)

Before After Deep cleaning

Sample collection area Preoperationala Operationalb Postinterventionc deep cleaningd deep cleaninge follow-up f

Food contact surfaces

Slicer blade NS 3.7 (2/54) 1.9 (1/54) 0 (0/9) 0 (0/9) 0 (0/27)
Deli case NS 3.8 (2/52) 9.3 (5/54) 0 (0/9) 11.1 (1/9) 0 (0/27)
Deli case by raw meat NS 12.5 (2/16) 14.3 (1/7) 0 (0/1) 0 (0/1) 0 (0/3)
Deli case tray NS 3.7 (2/54) 5.6 (3/54) 0 (0/9) 0 (0/9) 0 (0/27)
3-Basin interior NS 18.5 (10/54) 3.7 (2/54) 0 (0/9) 0 (0/9) 7.4 (2/27)
1-Basin interior NS 9.3 (5/54) 14.8 (8/54) 11.1 (1/9) 11.1 (1/9) 22.2 (6/27)
Cold storage room racks NS 11.1 (6/54) 3.7 (2/54) 11.1 (1/9) 0 (0/9) 0 (0/27)
Cutting board NS 2 (1/51) 0 (0/46) 0 (0/9) 0 (0/9) 0 (0/23)
Rewrap table NS 1.9 (1/53) 1.9 (1/54) 0 (0/9) 0 (0/9) 0 (0/27)
Counter NS 1.9 (1/52) 11.3 (6/53) 0 (0/9) 0 (0/9) 0 (0/27)
Total NS 6.5 (32/494) 6 (29/484) 2.4 (2/82) 2.4 (2/82) 3.3 (8/242)
Touch points
Slicer knob NS 3.7 (2/54) 3.7 (2/54) 0 (0/9) 0 (0/9) 0 (0/27)
Deli case handle NS 5.6 (3/54) 7.4 (4/54) 11.1 (0/9) 11.1 (0/9) 0 (0/26)
Scale NS 7.4 (4/54) 5.6 (3/54) 0 (1/9) 0 (1/9) 0 (0/27)
Total NS 5.6 (9/162) 5.6 (9/162) 3.7 (1/27) 3.7 (1/27) 0 (0/80)
Nonfood contact surfaces
3-Basin exterior NS 11.1 (6/54) 0 (0/54) 0 (0/9) 0 (0/9) 0 (0/27)
3-Basin FTW 20 (3/15) 15.1 (8/53) 24.1 (13/54) 22.2 (2/9) 0 (0/9) 14.8 (4/27)
1-Basin exterior NS 27.8 (15/54) 5.6 (3/54) 0 (0/9) 0 (0/9) 0 (0/27)
1-Basin FTW NS 52.8 (19/36) 55.9 (19/34) 22.2 (2/9) 11.1 (1/9) 33.3 (6/18)
Drain, deli area NS 33.3 (18/54) 30.2 (16/53) 0 (0/9) 11.1 (1/9) 29.6 (8/27)
Floor adjacent to deli drain 33.3 (5/15) 32.7 (17/52) 32.1 (17/53) 22.2 (2/9) 22.2 (2/9) 25.9 (7/27)
Floor deli area NS 14.8 (8/54) 27.8 (15/54) 11.1 (1/9) 11.1 (1/9) 25.9 (7/27)
Cold storage room floor NS 20.4 (11/54) 29.6 (16/54) 22.2 (2/9) 11.1 (1/9) 33.3 (9/27)
Cold storage room wall NS 18.5 (10/54) 7.4 (4/54) 0 (0/9) 0 (0/9) 3.7 (1/27)
Cold storage room drain NS 33.3 (12/36) 38.9 (14/36) 37.5 (3/8) 12.5 (1/8) 16.7 (4/24)
Standing water NS 40.6 (13/32) 30 (9/30) 100 (2/2) 22.2 (2/9) 35.3 (6/17)
Squeegee NS 11.4 (5/44) 34 (16/47) 12.5 (1/8) 0 (0/8) 38.1 (8/21)
Wheeled carts NS 9.3 (5/54) 24.5 (13/53) 11.1 (1/9) 22.2 (2/9) 18.5 (5/27)
Hoses 0 (0/1) 10.8 (4/37) 14.7 (5/34) 0 (0/6) 0 (0/6) 11.1 (2/18)
Trash can in deli area 0 (0/14) 5.6 (3/54) 5.7 (3/53) 11.1 (1/9) 11.1 (1/9) 0 (0/27)
Total 17.8 (8/45) 21.3 (154/722) 22.7 (163/717) 13.8 (17/123) 9.2 (12/130) 18.2 (67/368)
Total overall 17.8 (8/45) 14.2 (195/1378) 14.7 (201/1363) 9.9 (20/232) 7.1 (15/239) 12.5 (75/690)
a
Three months preoperational sampling in retail delis (12).
b
Six months operational sampling in retail delis (12).
c
Six months operational sampling after implementation of daily interventions in retail delis (2).
d
Sampling immediately before deep cleaning.
e
Sampling immediately after deep cleaning.
f
Monthly sampling for 3 months starting 1 month after deep cleaning.

than if test sites were identifiable and therefore potentially
target cleaned by crew members. Collection of follow-up
samples every month, particularly in the first month after
deep cleaning, provided only a low-resolution evaluation of
sustained efficacy. Monthly follow-up sampling would
appear the same whether L. monocytogenes was absent for
1 day or 1 week or 29 days, as long as L. monocytogenes
returned before the 1-month postcleaning sample date.
This DC-SSOP could be used to reduce L. monocyto-
genes prevalence in retail delis, but consistent, effective
implementation requires further work. Planning, preparation,
and communication before the day of cleaning in each deli
were major factors affecting execution. Highly engaged
corporate sanitarians and food safety directors who advo-
cated for additional resources within their organizations
were key to planning success. Well-prepared delis that had
all food products and dry goods removed before cleaning
crews arrived facilitated a more efficient start time with
minimal disruption to customers and regular business
activities. Conversely, in deli departments that were not
well prepared, staff struggled to remove inventory, often
resulting in delayed cleaning start times.
Overall, the DC-SSOP had a variable effect on L.
monocytogenes prevalence in delis with historically low or
J. Food Prot., Vol. 80, No. 11
moderate prevalence. In very high prevalence delis, deep
cleaning reduced prevalence 25.6% as determined during
follow-up sampling and eliminated or reduced 44% of
strains that persisted through preintervention testing (12) and
after implementation of daily SSOPs (2). It did not affect the
prevalence of virulence-reducing inlA mutations in L.
monocytogenes isolates. ATP testing was a useful tool for
in-process verification of cleaning efficacy and supporting
personnel morale; however, its results must be interpreted
with caution because a passing ATP reading may not
indicate absence of L. monocytogenes. The results of this
study indicate that deep cleaning alone cannot eliminate
persistent and transient L. monocytogenes contamination.
Facility maintenance, equipment design, and employee
attitudes are outside the scope of a cleaning procedure yet
have a major impact on the presence of foodborne pathogens
in the retail environment. Generally, more time, labor, and
resources were needed to support effective cleaning. Key
recommendations for delis wishing to incorporate this
strategy include (i) splitting the deep cleaning procedure
into two overnight events for large delis to give more hours
for cleaning with less personnel fatigue, (ii) a greater focus
on NFCS (e.g., floors, drains, and squeegees), and (iii)
incorporating local employees via training, daily monitoring
of sanitation efficacy (e.g., ATP testing), and/or managerial
investments to emphasize food safety practices. This study
did not incorporate local employee engagement, and we
predict that reductions in L. monocytogenes prevalence
could be better maintained if there were increased
engagement of local deli employees. Future directions for
research could include testing new cleaner and sanitizer
formulations designed specifically for soils common in the
deli environment.
ACKNOWLEDGMENTS
This project was supported by the Food Marketing Institute
Foundation. We thank Neogen Corp. for supplying the ATP testing
systems. We appreciate assistance provided by the Purdue University
Statistical Consulting Service, and we thank our staff members E. Wright,
B. Ziegler, S. Chambers, E. Christian, P. Cook, A. Pleitner, and C.
Wickware and the retail delis that volunteered for this study.
SUPPLEMENTAL MATERIAL
Supplemental material associated with this article can be
found online at: https://doi.org/10.4315/0362-028X.JFP-17-113.s1.
REFERENCES
1. Assisi, C. 2016. Genomic and transcriptomic analysis of biofilm
formation in persistent and transient L. monocytogenes. M.Sc. thesis.
Purdue University, West Lafayette, IN.
2. Etter, A. J., S. R. Hammons, S. Roof, C. Simmons, T. Wu, P. W.
Cook, A. Katubig, M. J. Stasiewicz, E. Wright, S. Warchocki, J.
Hollingworth, H. S. Thesmar, S. A. Ibrahim, M. Wiedmann, and H. F.
Oliver. 2017. Enhanced sanitation standard operating procedures have
limited impact on Listeria monocytogenes prevalence in retail delis. J.
Food Prot. 80:1903–1912.
CONTROL OF L. MONOCYTOGENES AT RETAIL 1923
3. Food Marketing Institute. 2006. Guidance for the control of Listeria
monocytogenes risks in retail food stores. Food Marketing Institute,
Arlington, VA.
4. Gray, M. J., R. N. Zadoks, E. D. Fortes, B. Dogan, S. Cai, Y. Chen,
V. N. Scott, D. E. Gombas, K. J. Boor, and M. Wiedmann. 2004.
Listeria monocytogenes isolates from foods and humans form distinct
but overlapping populations. Appl. Environ. Microbiol. 70:5833–
5841.
5. Hammons, S. R., M. J. Stasiewicz, S. Roof, and H. F. Oliver. 2015.
Aerobic plate counts and ATP levels correlate with Listeria
monocytogenes detection in retail delis. J. Food Prot. 78:825–830.
6. Hoelzer, K., B. D. Sauders, M. D. Sanchez, P. T. Olsen, M. M.
Pickett, K. J. Mangione, D. H. Rice, J. Corby, S. Stich, E. D. Fortes,
S. E. Roof, Y. T. Grohn, M. Wiedmann, and H. F. Oliver. 2011.
Prevalence, distribution, and diversity of Listeria monocytogenes in
retail environments, focusing on small establishments and establish-
ments with a history of failed inspections. J. Food Prot. 74:1083–
1095.
7. Nightingale, K., L. Bovell, A. Grajczyk, and M. Wiedmann. 2007.
Combined sigB allelic typing and multiplex PCR provide improved
discriminatory power and reliability for Listeria monocytogenes
molecular serotyping. J. Microbiol. Methods 68:52–59.
8. Orsi, R. H., H. C. den Bakker, and M. Wiedmann. 2011. Listeria
monocytogenes lineages: genomics, evolution, ecology, and pheno-
typic characteristics. Int. J. Med. Microbiol. 301:79–96.
9. Pan, Y., F. Breidt, and S. Kathariou. 2006. Resistance of Listeria
monocytogenes biofilms to sanitizing agents in a simulated food
processing environment. Appl. Environ. Microbiol. 72:7711–7717.
10. Pleitner, A. M., T. K. Chapin, S. R. Hammons, A. Van Stelten, K. K.
Nightingale, M. Wiedmann, L. M. Johnston, and H. F. Oliver. 2015.
Development and evaluation of a multi-institutional case studies–
based course in food safety. J. Food Sci. Educ. 14:76–85.
11. Sauders, B. D., M. D. Sanchez, D. H. Rice, J. Corby, S. Stich, E. D.
Fortes, S. E. Roof, and M. Wiedmann. 2009. Prevalence and
molecular diversity of Listeria monocytogenes in retail establish-
ments. J. Food Prot. 72:2337–2349.
12. Simmons, C., M. J. Stasiewicz, E. Wright, S. Warchocki, S. Roof, J.
R. Kause, N. Bauer, S. Ibrahim, M. Wiedmann, and H. F. Oliver.
2014. Listeria monocytogenes and Listeria spp. contamination
patterns in retail delicatessen establishments in three U.S. states. J.
Food Prot. 77:1929–1939.
13. Stasiewicz, M. J., H. F. Oliver, M. Wiedmann, and H. C. den Bakker.
2015. Whole-genome sequencing allows for improved identification
of persistent Listeria monocytogenes in food-associated environments.
Appl. Environ. Microbiol. 81:6024–6037.
14. U.S. Food and Drug Administration. 2009. Food Code 2009.
Available at: http://www.fda.gov/downloads/Food/GuidanceR
egulation/UCM189448.pdf. Accessed 9 July 2016.
15. Van Stelten, A., and K. K. Nightingale. 2008. Development and
implementation of a multiplex single-nucleotide polymorphism
genotyping assay for detection of virulence-attenuating mutations in
the Listeria monocytogenes virulence-associated gene inlA. Appl.
Environ. Microbiol. 74:7365–7375.
16. Van Stelten, A., J. M. Simpson, T. J. Ward, and K. K. Nightingale.
2010. Revelation by single-nucleotide polymorphism genotyping
that mutations leading to a premature stop codon in inlA are common
among Listeria monocytogenes isolates from ready-to-eat foods but
not human listeriosis cases. Appl. Environ. Microbiol. 76:2783–
2790.
17. Wang, J., A. J. Ray, S. R. Hammons, and H. F. Oliver. 2015.
Persistent and transient Listeria monocytogenes strains from retail deli
environments vary in their ability to adhere and form biofilms and
rarely have inlA premature stop codons. Foodborne Pathog. Dis.
12:151–158.