Towards microbial safety of fresh vegetables in Rwanda

Global consumption of vegetables and the associated reported foodborne illnesses have been increasing in tandem. This study was commissioned to analyze the microbial risk from “farm to fork” along the fresh vegetable supply chain in Rwanda to explore microbial safety options that can contribute to an integrated system to detect, control and prevent foodborne infections. Specific study objectives were developed, the first being to estimate the burden of foodborne infectious illnesses using the available data to obtain insight into the general plight of food safety issues and to also develop a framework for future investigations. Second, to investigate the microbial safety status, handling practices and risk exposure factors along the vegetable supply chain.

To understand the overall impact of foodborne illnesses on human health in Rwanda the burden of food related illnesses in year 2013 was estimated using the DALY (disability adjusted life year) metric, as encouraged by the World Health Organization. DALY is a measure that combines years of life lost due to premature death and healthy years lost due to disability during sickness. Study findings indicate that for the year 2013, watery diarrhea occurred all year round as by the surveillance system data, resulting to an estimated 672 (95% credible interval [CrI] 424 ‒ 932) DALY per million inhabitants, bloody diarrhea was seasonal coinciding with the rainy months and caused an estimated 213 (95% CrI 50 ‒  475) DALY per million, typhoid and cholera manifested as outbreaks with an estimated 73 (95% CrI 57 ‒  91) and 1 (95% CrI 0 ‒ 2) DALY per million respectively. These data show that the health burden is high and we suspect that a large proportion is caused by consumption of contaminated food.

Also investigated in this study, was the prevalence of foodborne pathogens in farm vegetables and agricultural water (used for irrigation and on-farm washing of vegetables) in Rwanda. In agricultural water from rivers, lakes, lagoons, ground and marshlands, traces in the form of DNA from a wide variety of virulent pathogenic organisms were detected, including enteroinvasive, enteroaggregative, and enterotoxigenic E. coli, Vibrio cholera, Yersinia pestis and the parasite Cyclospora. DNA from thermo-tolerant Campylobacter spp. was found in 87% of the samples. Although this does not mean that all these pathogens were still alive by the time of detection; presence of DNA is an indication that the pressure of pathogens in agricultural water and the environment is high. Indeed from 99 samples of farm vegetables, different viable foodborne pathogens were isolated viz. Listeria monocytogenes (1%), Campylobacter spp. (3%), Salmonella spp. (5%) and pathogenic E. coli (6%).

In addition to tracing the pathogens at farm level, selected indicator microorganisms were investigated, to get an impression of their survival, growth and/or inactivation along the vegetable supply chain. The latter was complemented with a detailed observation of handling practices along the supply chain. The mean count of Enterobacteriaceae in 11 types of vegetables increased slightly from farm to markets to reach an average of 6.0 log cfu/g upon arrival at FSEs. During food preparation microbial counts were significantly reduced by washing with or without sanitizers, trimming/peeling, with an average of 2.1 log cfu/g from start to end of salad preparation. Ready-to-eat salads prepared by FSEs met the guidelines by 91% and 22% for coagulase-positive staphylococci (104 cfu/g) and presumptive Listeria spp. (102 cfu/g) respectively. Because washing and sanitization procedures differ from one FSE to another, a laboratory study was designed to mimic the practices at FSEs with the aim to select optimal washing and sanitization procedures. Findings in the field study with FSEs revealed that about 61% of the visited FSEs used sanitizers during washing of fresh vegetables, in particular, potassium permanganate (KMnO4) in 39 % of FSEs, sanitizing powder (a mixture of polyphosphate, sodium hydrogen carbonate and active chlorine), 13%; sodium hypochlorite (NaClO), 7 %; and sodium dichloroisocyanurate (NaDCC) in 2%. Average reduction ranged from 1.0 log (for KMnO4) to 3.1 log ( for NaDCC). In the laboratory study, average inactivation observed with indicator microorganisms ranged from 0.7 log (for water alone) to 3.0 log (for NaDCC). Out of the 8 sanitizers that were evaluated, 5 sanitizers (NaDCC [90 ppm], NaClO [200 ppm], lemon juice [98%], acetic acid [2 %] and sanitizing powder [4 g/L]) resulted in significantly higher inactivation compared to water alone. A contact time of 5 minutes and a salad-sanitizer ratio of 1: 20 were considered optimal for kitchen based washing of the studied leafy vegetables with NaDCC and NaClO sanitizers. This study also reveals that the most widely used sanitizer (25ppm KMnO4) was not more effective than washing with only water and an indication that a policy, guideline or regulation on kitchen based washing and sanitization of salad vegetables should be enacted.

The high prevalence (15%) of foodborne pathogens associated vegetables at farm level and increasing trends in levels of indicator microorganisms from farm to FSEs, raises concern about the potential presence of foodborne pathogens in ready-to-eat salads. By using @risk software (Palisade Corporation, Ithaca, NY, USA) and Monte Carlo simulation (100,000 iterations), the number of cases due to serving leafy vegetables contaminated pathogenic E. coli have been estimated with a mode of 12 million cases of illness per year and 0.1 probability of illness per serving. These estimates can be considered high compared to previous risk assessments in other countries. However it should be acknowledged that these estimates also include mild illness from less harmful pathotypes of pathogenic E. coli. viz. enterotoxigenic E. coli (ETEC) and enteropathogenic E. coli. To further advise risk managers, seven “what if scenarios” were simulated to compare with the baseline model. The scenario of improving washing and sanitization (3.0 log reduction) at FSEs resulted to a less than 2 fold change in the predicted microbial risk. While a 2 fold change was observed for the scenario of channelling all vegetables through supermarkets instead of traditional markets. Farm interventions reducing the prevalence and levels of pathogenic E. coli in the base line model by 90%, the introduction of a cold chain and skipping the market step, all resulted to a 10 fold reduction in predicted microbial risk. The scenario of reducing or avoiding contamination and cross contamination along the supply chain led to 1000 fold reduction in the predicted microbial risk. Lastly, farm to fork measures combining three different scenarios ( avoiding contamination from farm to fork, farm interventions (90% reduction) and improving sanitization) were predicted to reduce microbial risk by a factor of 1 million.

Risk estimates should be analyzed and weighted in order to prioritize risk management measures and interventions, not only for their effectiveness in improving public health, but also for their feasibility, acceptability and affordability. From the predicted microbial risk in this study, approaches to change the routes or temperatures of supply chain in Rwanda may not be given priority by risk managers except for the scenario of skipping the market step so that FSEs get vegetables straight from farms. It can be observed from this study that changing the supply chain by introducing a cold chain or refrigeration at market (supermarkets) may not address the microbial risk associated with fresh vegetables. Instead, risk managers can focus on addressing the factors leading to contamination and cross contamination from “farm to fork”. It has been demonstrated in this study that washing and sanitization at FSEs alone (current practise) is not enough to address microbial safety concerns. Consequently, risk assessors and risk manager should identify all the possible sources of microbial hazards along the entire supply chain and devise measures and interventions to address them and spearhead risk communication strategies among stakeholders. This study recommends embracing the concepts of “Global One Health” in order to move towards sustainable microbial safety of fresh vegetables in Rwanda and the Globe.

We hope that the results from this study will be helpful for policy makers and risk managers, not only in approaching the microbial safety concerns of vegetables along the supply chain but also in developing national integrated food chain systems.