First Nation-Wide Analysis of Food Safety and Acceptability Data in Lebanon

The challenges to food safety in Lebanon are numerous and have coalesced to pose a serious public health concern. This is evident in well-documented food poisoning outbreaks and adulteration cases. In response, the Lebanese government initiated an unprecedented food safety campaign (2015–2017) that aimed to test food samples that were randomly collected from foodservices and industries across the country. The data were made available publicly, but they were never analyzed to prioritize and determine high risk foods and most prevalent contaminants nationally or across governorates. To answer these questions, we performed an in-depth statistical analysis of the data, which included 11,625 individual food samples. Our analysis showed that water (55% of tested water samples), spices (49.3%), red meat (34.4%), poultry (30.9%) and dairy (28.3%) were the main foods associated with the highest rejection rates. The most common biological contaminants detected in rejected foods were sulfate-reducing bacteria (34.7%), Escherichia coli (32.1%), coliforms (19.6%), Staphylococcus aureus (12.8%), and Salmonella (11.6%). We conclude that Lebanon needs rigorous and sustainable programs to monitor the quality and safety of foods. Given the lack of resources, we recommend putting emphasis on extensive outreach programs that aim at enhancing food safety knowledge from farm to fork.


Introduction
Access to sufficient quantities of safe and nutritious food is a prerequisite for sustaining life and good health. However, chemical and microbiological contamination of food can lead to a plethora of human diseases that range from diarrhea to cancer [1,2]. The majority of foodborne diseases have been attributed to microbial pathogens, such as Escherichia coli, norovirus, Campylobacter and non-typhoidal Salmonella [3]. Furthermore, it is well documented that these foodborne pathogens exert a significant toll on public health and economies in both developed and developing countries [4]. Global estimates show that 600 million individuals (~1 in 10 people) are affected by foodborne illnesses every year, which leads to the death of 420,000 individuals [3]. Subsequently, outbreaks of foodborne illnesses adversely impact health care systems as well as national food brands, trade, tourism, and socioeconomic growth, especially in developing countries [5]. In these countries, foodborne diseases perpetuate the vicious cycle of disease, malnutrition, and poverty, which significantly affect the most vulnerable populations, including infants, young children, the elderly, and nutritionally deprived or immunocompromised individuals [5]. In addition, children under the age of 5 years are at a notably high risk, comprising awareness, and provide plans to focus limited resources to address food safety priorities. Consequently, the aim of our study was to analyze the data collected during the MoPH campaign between 2015-2017 in order to better understand microbial food safety, identify the most frequently contaminated foods and contaminants, and assess the risk associated with food in the Lebanese market.

Data Description
The data were collected during a campaign launched by the Lebanese MoPH from 2015 to 2017. The campaign involved random and uninformed inspections of various facilities across Lebanon, which included restaurants, grocery stores and supermarkets, butcheries, fisheries, roasteries, and bakeries. The data were shared publicly in 110 separate reports on the MoPH website (https: //www.moph.gov.lb/en/Pages/4/126/food-safety) [34]. The reports were qualitative and included a description of the sample (food/drink/ingredient type), sample location, sampled foodservice or food-production facility, identity of detected contaminant(s), and the food unacceptability according to the criteria adopted by the Lebanese Standards Institution (LIBNOR) [34]. However, the reports did not always explicitly specify whether the food was cooked, raw, and/or ready-to-eat [34]. The last report was posted on 17 July 2017 [34].

Statistical Analysis
Due to the diversity of the tested samples (cooked, raw, frozen, fresh, etc.), we categorized the samples based on food type to facilitate statistical analysis. Therefore, the categories analyzed were red meat, poultry meat, dairy, bakery, fish, nuts, desserts, spices, water, and other (which included various items such as vinegar, beans, jam, vegetables, coffee, and tea among others). The data in the published reports were then checked for completeness and were coded and logged into the Statistical Package for Social Sciences (SPSS Inc., Chicago, IL, USA) software version 18 for Windows, which was later used for statistical analyses. Descriptive statistics were performed and presented as frequencies and percentages for categorical variables. In addition, the associations between food category and type of microbiological contaminant as well as the prevalence of unacceptable (rejected) food categories were assessed using Chi-square test. Comparisons between governorates (the administrative divisions of Lebanon) were also performed using Pearson's Chi-square test. Additionally, the association between the microbial acceptability of food categories and location of sampling (governorates) was explored using simple logistic regression, with microbiological acceptability as the dependent variable. To evaluate the correlates of microbial acceptability risk, a multiple logistic regression model was utilized. In this model, variables were included if they were significantly associated with the dependent variable in the univariate analysis. Odds ratios (OR) and their respective 95% confidence intervals were computed. All reported p-values were based on two-sided tests and were compared with a significance level of 5%.

Results
Analysis of the data from 2015-2017 (n = 11,625) showed that 71.3% (n = 8291) of all tested samples were labelled as microbially acceptable, whereas 28.7% (n = 3334) of the samples were unacceptable (Table 1). Based on the number of unacceptable samples within each food category, five food categories (red meat, poultry meat, dairy, spices, and water) were identified as high risk. These categories also contained the highest percentages of unacceptable samples per category, where water, spices, red meat, poultry meat, and dairy harbored 55.0%, 49.3%, 34.4%, 30.9%, and 28.3% of unacceptable samples per category, respectively (Table 1). While assessing the distribution of unacceptable samples according to the governorate (geographical location), it was observed that 31.7% of all samples from the North governorate of Lebanon were rejected, while 29.9% and 27.3% of the samples were unacceptable in South Lebanon and Mount Lebanon, respectively (Table 2). Beirut, the capital of Lebanon, had the lowest percentage (23.8%) of unacceptable samples ( Table 2). Statistical analysis revealed a significant variation in the prevalence of unacceptable food samples across governorates (p-value < 0.001; Table 2). Furthermore, odds ratio analysis indicated that it was almost 1.4 times more likely to encounter unacceptable food products in the North and South governorates compared to the governorate with the lowest food rejection percentage (Beirut) ( Table 2). Further analysis of the data revealed that the food categories with the highest percentages of unacceptable samples included red meat, poultry meat, dairy, and spices in the North, South, Mount Lebanon, and the Bekaa governorates (Table 3). However, red meat, spices, and other foods were the three highest risk categories in the Beirut governorate (Table 3).  Evaluation of the microbial criterion that led to the rejection (unacceptability) of the food samples showed that sulfate-reducing bacteria (SRB), E. coli, aerobic bacteria, and coliforms were the most common contaminates (Table 4). The levels of these bacteria were the criteria for rejecting 34.7%, 32.1%, 21.7%, and 19.6% of the unacceptable samples, respectively (Table 4). In addition, unacceptable levels of Staphylococcus aureus, Salmonella spp., and yeast and fungi resulted in the rejection of 11.6%-12.8% of the samples (Table 4). Notably, red meat samples constituted 65.2% of all rejected samples contaminated with unacceptable levels of S. aureus, 46.2% with SRB, 55.3% with E. coli, 52.5% with L. monocytogenes, and 32.6% with Salmonella spp., respectively (Table 4). Additionally, poultry meat samples constituted 62.0% of all rejected samples contaminated with unacceptable levels of Salmonella spp., 33.8% with SRB, 23.2% with aerobic bacteria, and 22% with L. monocytogenes, respectively (Table 4). Dairy samples constituted 38.7% of all rejected samples contaminated with unacceptable levels of coliforms, 37.1% with yeast/fungi, 23.7% with L. monocytogenes, and 19.7% with E. coli, respectively (Table 4). A notable finding was that spices constituted 74.1% and 35% of the rejected samples due to aflatoxins and yeast/fungi, respectively (Table 4). Further analysis revealed that there was a statistically significant association between food group and the microbial contaminants (p-value < 0.001) ( Table 4).
Although the MoPH reports did not always explicitly designate whether the tested food was ready-to-eat, screening the data showed that L. monocytogenes occurred in 14 cheese samples and 13 minced red meat samples (with or without herbs and onions), which are traditionally consumed without further cooking in Lebanon. Furthermore, Salmonella spp. were detected in six cheese and 54 minced red meat samples. Unacceptable levels of fecal indicators, E. coli and coliforms, occurred in 319 cheese samples, while E. coli was detected in 306 minced red meat samples. Additionally, the occurrence of a variety of microorganisms and toxins in spices demonstrates the contamination of ingredients used in different foods (Table 4).

Discussion
In response to increasing public concern, the ministry of public health (MoPH) in Lebanon launched a food safety campaign (2015-2017) [34]. Although the number of collected samples was modest and testing focused on a limited list of microbial contaminants, the campaign presented a chance to shed light on food safety and major foodborne contaminants in Lebanon using experimental data. Here, we downloaded the data from the public website of MoPH and performed in-depth statistical analysis. Our overarching objective was to provide awareness to stakeholders on food safety as well as prioritize the risks for efficient allocation of resources and control efforts.
It was notable that 28.7% of all tested samples were found to be microbially unacceptable (Table 1). Accordingly, these samples were rejected by the MoPH, because they pose a potential food safety and quality risk. The number of rejected samples showed that more than one in four of the tested samples were contaminated. This is important, because food contaminated with pathogens like Salmonella and Listeria and toxins secreted by fungi (aflatoxins), S. aureus, and C. botulinum is known to cause life-threatening infections and acute and chronic illnesses in humans [35][36][37][38][39]. This includes diarrhea, nausea, vomiting, stomach cramps, sepsis, renal failure, meningitis, respiratory and muscular paralysis, miscarriage, stillbirth, and cancer among others [35][36][37][38][39]. Additionally, SRB are associated with ulcerative colitis and are thought to trigger inflammatory bowel diseases (IBD) [40]. However, when considering the microbial parameters (Table 4), it can be immediately observed that the testing was not comprehensive. For example, foodborne viruses and parasites, such as norovirus, hepatitis A, and Cryptosporidium, were not tested. This is important, because norovirus is the leading cause of foodborne gastroenteritis worldwide [41]. Additionally, 10,400 cases of hepatitis A were reported in Lebanon between 2005 and 2017 [42]. Similarly, there are recent reports of Cryptosporidium and other parasitic infections in the Lebanese population [43][44][45]. The MoPH reports did not also include testing of potential chemical contaminants in the food. Taken together, this suggests that there is a possibility that the rejection rate might increase if other microbial pathogens/indicators/chemicals were to be tested. Consequently, the Lebanese consumers might be at a notably high risk of contracting foodborne diseases, which has a significant impact on public health and highlights the pre-emptory need to address food safety in Lebanon.
When analyzing the distribution of unacceptable food samples across governates, we noted that the highest number (31.7%) of rejected samples were collected in the North governate, while the lowest (23.8%) was in Beirut, the capital of Lebanon (Table 2). This perhaps reflected a possible socioeconomic impact on food safety in Lebanon, because the North is generally known to be facing more challenges in poverty, income distribution, and infrastructure in comparison to the rest of the country [46,47]. However, red meat, poultry meat, dairy, and spices samples were the most problematic in every governorate (Table 3). Regardless, the high rate of microbially unacceptable samples in essential food categories highlights the potential risks of food contamination to the Lebanese population, especially in the poorest governorates.
Several important foodborne pathogenic bacteria, such as Salmonella, L. monocytogenes, S. aureus, and C. botulinum, were detected in the tested food samples in Lebanon (Table 4). A notable observation was also the presence of aflatoxins in spices (Table 4). Therefore, the data highlight gaps in the hygienic and sanitation systems deployed in the food chain in Lebanon. For example, indicators of fecal contamination (E. coli and coliforms) and poor sanitation (S. aureus) [48] were among the major microbial contaminants in rejected food samples ( Table 4). The presence of the aforementioned pathogens and toxins in foods might have severe health repercussions, especially in children, the elderly, and immunocompromised and disenfranchised populations. Given the status of the economy and the current increase in the cost of medical care in Lebanon, these food contaminants might inflict higher mortalities and morbidities in susceptible populations. This scenario is even more plausible because of the proliferation of antimicrobial resistance in food systems in Lebanon [14,21,22,28,29,32,49], which can lead to recalcitrant and hard-to-treat infections [50].
Red meat, poultry, and dairy were among the most contaminated food categories in Lebanon. The contamination of red meat is notable, because it can be consumed raw as part of popular Lebanese dishes, which increases the risk to consumers. Poultry and dairy are also popular in the Lebanese diet and all three food categories constitute an essential component of the economy [51][52][53]. Here, it should be noted that the observed results in this study largely depend on the methodologies deployed by the MoPH personnel, their skills, and budget. Regardless, food safety initiatives in Lebanon might benefit from focusing on these food categories to identify and mitigate contamination effectively. In turn, this will also considerably reduce the burden of foodborne infections in Lebanon and allow food safety stakeholders to operate sustainably.
Lebanon lacks a well-established food safety infrastructure to aid in identifying and tracking foodborne outbreaks [33]. Therefore, the burden of foodborne diseases in Lebanon is underestimated [9]. This, along with deficient resources and infrastructure, will affect the actions and policies employed to control foodborne disease outbreaks [54]. However, based on our analysis, it can be readily concluded that food safety must be a central concern for Lebanon, because evidence points to widespread contamination of foods that can have detrimental effect on health and economy. This is predicted to become more severe, because of the current political and economic crises in Lebanon. Our analysis can help stakeholders to focus and prioritize their engagement in food safety in Lebanon by addressing the foods that pose the highest risk.
Interventions which, if implemented, would substantially improve food safety in Lebanon must include (1) prioritizing the evaluation of the major food systems and the anecdotal sources of foodborne contamination (local and imported), (2) establishing a baseline for a national surveillance system for foodborne pathogens (bacteria, viruses and parasites) and antimicrobial resistance, (3) recruiting competent staff and using computerized systems and upgraded laboratories, (4) defining roles and tasks of various official and private agencies involved in food safety, (5) educating, training and raising awareness among the public and business owners, and (6) devising control measures to reduce the risk for consumers. Funding: This work was funded by a University Research Board (URB) grant from the American University of Beirut (AUB) and seed funding from the Center for Food Safety, the University of Georgia (UGA).