Food Protection Trends - September/October 2022 - 379

Microbial enumeration
E. coli and Salmonella in water samples were enumerated
before and after UV treatment using the methodology
reported by Usaga et al. (15) and Jones et al. (9) for UVtreated
liquid foods and unfiltered surface irrigation water,
respectively. Appropriate serial dilutions in sterile 0.1%
peptone water were aseptically plated in duplicate in petri
dishes, and 15 mL of Trypticase soy agar (Difco, BD)
was pour plated. After solidification, petri dishes were
incubated at 35 ± 2°C for 20 ± 2 h. The differences between
the log-transformed microbial counts before and after UV
exposure were calculated. Although high microbial loads
were expected due to background microbiota in surface
water, a nonselective growth medium was used, as described
previously (9). Because UV exposure may sublethally
damage bacterial cells and affect their growth in selective
media, use of a nonselective nutrient medium prevents
overestimation of the log reductions, which represents a
safety concern. The inoculated microorganisms (Salmonella
and E. coli) were not differentiated from the background
microbiota during enumeration on nonselective medium,
but due to the high inoculum level, the target microorganism
levels significantly surpassed the background microbial
populations. The total log reduction was calculated.
Statistical analysis
Effects of testing dates, initial bacterial counts, and water
pH and turbidity on the difference in log-transformed
microbial counts before and after treatment were explored
with the Wilcoxon rank sum test (for categorical variables)
and an analysis of variance (for numerical variables). Multivariate
regression analyses were conducted including variables
with a univariate P < 0.20, using a backward stepwise
elimination procedure.
RESULTS
The overall median difference in bacterial counts before
and after the UV treatment was 6.3 log CFU/mL (Table 1).
When E. coli and Salmonella were evaluated separately, the
median difference before and after UV treatment was slightly
lower for E. coli (6.2 log CFU/mL) than for Salmonella (6.4
log CFU/mL) (multivariate P < 0.001). The date of testing
(six samples per date) was a significant predictor of the
reductions obtained (multivariate P < 0.001). Within the
ranges tested, neither water pH (6.35 to 8.19) nor turbidity
(11.9 to 58.9 NTU) (Table 2) were significantly associated
with the reduction in the multivariate models (P > 0.05).
However, inoculation levels (initial counts) also differed by
testing date and were slightly higher for Salmonella (median,
7.8 log CFU/mL) than for E. coli (median, 7.7 log CFU/
mL), which could explain the difference in log reductions
obtained. Although a nonselective medium was used for
microbial enumeration, the results indicated significant
inactivation of the total microbial load. Differences in initial
microbial populations may be influenced by the presence of
uneven background microbiota in the water samples because
with the selected approach the total microbial load was
enumerated.
DISCUSSION
UV light is a nonthermal and environmentally friendly
approach for inactivation of pathogens in surface agricultural
water. In the present study, UV light treatment effectively
reduced a high load of two inoculated vegetative microorganisms
of food safety relevance, regardless of the normal and
expected variability of surface water properties over time.
That variability, over the period tested, may explain why
testing date significantly influenced the microbial reductions
obtained. These results are in agreement with previously
reported findings for unfiltered surface irrigation water from
different geographic locations with lower turbidity ranges
(≤ 20 NTU) collected over a shorter sampling period (9).
Because UV systems are most effective when water
is clear and free of suspended particles, most technical
recommendations for surface water indicate the need to
couple filtration with UV to guarantee treatment efficacy
(3). However, this suggestion may need to be revised based
on recent findings. Jones et al. (9) described 3-log microbial
inactivation in UV-treated irrigation water at relatively high
turbidity (20 NTU), and in the present investigation the
inactivation was > 6 log CFU/mL with the same technology
and water samples with more than two times higher turbidity
values of up to 58.9 NTU. These results should be replicated
with other commercially available thin-film UV devices
that ensure exposure to a constant UV radiation dose using
a turbulent flow regime and an adjusted flow rate based on
the sample absorptivity profile. Nevertheless, microbial
validation of each UV device is necessary before use, and the
performance of UV units must be monitored periodically to
confirm the efficacy of the technology as required by the U.S.
Environmental Protection Agency for the UV treatment of
drinking water (17).
Water disinfection by UV radiation is dependent on multiple
variables in addition to water properties, such as UV treatment
duration and intensity. The relationship between the required
UV dose and the UV intensity (measured by UV sensors), flow
rate, and transmittance must be established and monitored
to ensure sufficient disinfection of microbial pathogens (18).
For example, UV water sterilization systems for greenhouse
irrigation water are designed for exposures of 80 to 250 mJ/
cm2
(12). The considerably lower UV dose (14 mJ/cm2
) used
in the present study, albeit with a turbulent flow regime, may
represent a treatment alternative for turbid water sources.
Because any material that absorbs or reflects UV light,
such as dissolved solids in water (e.g., iron), can decrease
UV transmittance and therefore reduce the germicidal
effect, a detailed physicochemical characterization of
each water source is necessary before implementing UV
September/October Food Protection Trends 379
Food Protection Trends - September/October 2022

Table of Contents for the Digital Edition of Food Protection Trends - September/October 2022
