Power Law for Estimating Underdetection of Foodborne Disease Outbreaks, United States.

We fit a power law distribution to US foodborne disease outbreaks to assess underdetection and underreporting. We predicted that 788 fewer than expected small outbreaks were identified annually during 1998-2017 and 365 fewer during 2018-2019, after whole-genome sequencing was implemented. Power law can help assess effectiveness of public health interventions.

Confocal laser microscopy for assessment of surgical margins during radical prostatectomy.

OBJECTIVE: To evaluate the feasibility of confocal laser microscopy (CLM) for intraoperative margin assessment as faster alternative to neurovascular structure-adjacent frozen-section examination (NeuroSAFE) during robot-assisted radical prostatectomy (RARP). PATIENTS AND METHODS: Surgical margins were assessed during 50 RARP procedures in patients scheduled for NeuroSAFE. Posterolateral sections were cut and imaged with CLM and further processed to conform with the NeuroSAFE protocol. Secondary resection (SR) was performed in case a positive surgical margin (PSM) was observed with NeuroSAFE. Afterwards, the CLM images were non-blinded assessed for the presence of PSMs. The accuracy of both NeuroSAFE and CLM was compared with conventional histopathology. Agreement for detection of PSMs between NeuroSAFE and CLM was evaluated with Cohen's kappa coefficient. Procedure times were compared with a Wilcoxon signed-ranks test. RESULTS: In total, 96 posterolateral sections of RP specimens were evaluated for the presence of PSMs. CLM identified 15 (16%) PSMs and NeuroSAFE identified 14 (15%) PSMs. CLM had a calculated sensitivity, specificity, positive predictive value and negative predictive value of 86%, 96%, 80% and 98% respectively for the detection of PSMs compared to definite pathology. After SR, residual tumour was found in six of 13 cases (46%), which were all identified by both techniques. There was a substantial level of agreement between CLM and NeuroSAFE (κ = 0.80). The median procedure time for CLM was significantly shorter compared to NeuroSAFE (8 vs 50 min, P < 0.001). The main limitation of this study was the non-blinded assessment of the CLM images. CONCLUSIONS: Compared to NeuroSAFE, CLM is a promising technique for intraoperative margin assessment and is able to reduce the time of intraoperative margin assessment.

Salmonella Serotypes Associated with Illnesses after Thanksgiving Holiday, United States, 1998-2018.

We sought to determine which Salmonella serotypes cause illness related to the Thanksgiving holiday in the United States and to foods disproportionately eaten then (e.g., turkey). Using routine surveillance for 1998-2018 and a case-crossover design, we found serotype Reading to be most strongly associated with Thanksgiving.

Nearest-Neighbors Matching for Case-Control Study Analyses: Better Risk Factor Identification From a Study of Sporadic Campylobacteriosis in the United States.

BACKGROUND: Case-control studies are commonly used to explore factors associated with enteric bacterial diseases. Control of confounding is challenging due to a large number of exposures of interest and the low frequencies of many of them. METHODS: We evaluated nearest-neighbors matching in a case-control study (originally 1:1 matched, published in 2004) of sporadic Campylobacter infections that included information on 433 exposures in 2632 subjects during 1998-1999. We performed multiple imputations of missing data (m = 100) and calculated Gower distances between cases and controls using all possible confounders for each exposure in each dataset. We matched each case with ≤20 controls within a data-determined distance. We calculated odds ratios and population attributable fractions (PAFs). RESULTS: Examination of pairwise correlation between exposures found very strong associations for 1046 pairs of exposures. More than 100 exposures were associated with campylobacteriosis, including nearly all risk factors identified using the previously published approach that included only 16 exposures and some less studied, rare exposures such as consumption of chicken liver and raw clams. Consumption of chicken and nonpoultry meat had the highest PAFs (62% and 59%, respectively). CONCLUSIONS: Nearest-neighbors matching appear to provide an improved ability to examine rare exposures and better control for numerous highly associated confounders.

Increased Incidence of Antimicrobial-Resistant Nontyphoidal Salmonella Infections, United States, 2004-2016.

Salmonella is a major cause of foodborne illness in the United States, and antimicrobial-resistant strains pose a serious threat to public health. We used Bayesian hierarchical models of culture-confirmed infections during 2004-2016 from 2 Centers for Disease Control and Prevention surveillance systems to estimate changes in the national incidence of resistant nontyphoidal Salmonella infections. Extrapolating to the United States population and accounting for unreported infections, we estimated a 40% increase in the annual incidence of infections with clinically important resistance (resistance to ampicillin or ceftriaxone or nonsusceptibility to ciprofloxacin) during 2015-2016 (≈222,000 infections) compared with 2004-2008 (≈159,000 infections). Changes in the incidence of resistance varied by serotype. Serotypes I 4,[5],12:i:- and Enteritidis were responsible for two thirds of the increased incidence of clinically important resistance during 2015-2016. Ciprofloxacin-nonsusceptible infections accounted for more than half of the increase. These estimates can help in setting targets and priorities for prevention.

Estimate of Burden and Direct Healthcare Cost of Infectious Waterborne Disease in the United States.

Provision of safe drinking water in the United States is a great public health achievement. However, new waterborne disease challenges have emerged (e.g., aging infrastructure, chlorine-tolerant and biofilm-related pathogens, increased recreational water use). Comprehensive estimates of the health burden for all water exposure routes (ingestion, contact, inhalation) and sources (drinking, recreational, environmental) are needed. We estimated total illnesses, emergency department (ED) visits, hospitalizations, deaths, and direct healthcare costs for 17 waterborne infectious diseases. About 7.15 million waterborne illnesses occur annually (95% credible interval [CrI] 3.88 million-12.0 million), results in 601,000 ED visits (95% CrI 364,000-866,000), 118,000 hospitalizations (95% CrI 86,800-150,000), and 6,630 deaths (95% CrI 4,520-8,870) and incurring US $3.33 billion (95% CrI 1.37 billion-8.77 billion) in direct healthcare costs. Otitis externa and norovirus infection were the most common illnesses. Most hospitalizations and deaths were caused by biofilm-associated pathogens (nontuberculous mycobacteria, Pseudomonas, Legionella), costing US $2.39 billion annually.

Estimating the Number of Illnesses Caused by Agents Transmitted Commonly Through Food: A Scoping Review.

Estimates of the overall human health impact of agents transmitted commonly through food complement surveillance and help guide food safety interventions and regulatory initiatives. The purpose of this scoping review was to summarize the methods and reporting practices used in studies that estimate the total number of illnesses caused by these agents. We identified and included 43 studies published from January 1, 1995, to December 31, 2019, by searching PubMed and screening selected articles for other relevant publications. Selected articles presented original estimates of the number of illnesses caused by ≥1 agent transmitted commonly through food. The number of agents (species or subspecies for pathogens) included in each study ranged from 1 to 31 (median: 4.5; mean: 9.2). Of the 40 agents assessed across the 43 studies, the most common agent was Salmonella (36; 84% of studies), followed by Campylobacter (33; 77%), Shiga toxin-producing Escherichia coli (25; 58%), and norovirus (20; 47%). Investigators used a variety of data sources and methods that could be grouped into four distinct estimation approaches-direct, surveillance data scaled-up, syndrome or population scaled-down, and inferred. Based on our review, we propose four recommendations to improve the interpretability, comparability, and reproducibility of studies that estimate the number of illnesses caused by agents transmitted commonly through food. These include providing an assessment of statistical and nonstatistical uncertainty, providing a ranking of estimates by agent, including uncertainties; describing the rationale used to select agents and data sources; and publishing raw data and models, along with clear, detailed methods. These recommendations could lead to better decision-making about food safety policies. Although these recommendations have been made in the context of illness estimation for agents transmitted commonly through food, they also apply to estimates of other health outcomes and conditions.

Salmonella Serotypes: A Novel Measure of Association with Foodborne Transmission.

Most nontyphoidal Salmonella (NTS) illnesses in the United States are thought to be foodborne. However, transmission routes likely vary among the different serotypes. We developed a relative ranking of NTS serotypes according to the strength of their association with foodborne transmission. We used Laboratory-based Enteric Disease Surveillance data to estimate the proportion of infections for each Salmonella serotype reported from 1998 to 2015 and Foodborne Disease Outbreak Surveillance System data to calculate the proportion of foodborne outbreak-associated Salmonella illnesses caused by each serotype. We calculated the ratios of these proportions to create a foodborne relatedness (FBR) measure for each serotype. Of the top 20 serotypes, Saintpaul (2.14), Heidelberg (1.61), and Berta (1.48) had the highest FBR measures; Mississippi (0.01), Bareilly (0.13), and Paratyphi B variant L(+) tartrate(+) (0.20) had the lowest. The FBRs for the three most prevalent serotypes were 1.22 for Enteritidis, 0.77 for Typhimurium, and 1.16 for Newport. This method provides a quantitative approach to estimating the relative differences in the likelihood that an illness caused by a particular serotype was transmitted by food, which may aid in tailoring strategies to prevent Salmonella illnesses and guide future research into serotype-specific source attribution.

Penile intraepithelial neoplasia: Nomenclature, incidence and progression to malignancy in the Netherlands.

OBJECTIVE: To determine the incidence of penile intraepithelial neoplasia in the Netherlands using a nationwide histopathology registry and to discuss the nomenclature of premalignant penile lesions. METHODS: Data from patients in the Netherlands diagnosed with a premalignant penile lesion between January 1998 and December 2007 were collected from the nationwide histopathology registry (PALGA); this database covers all pathology reports of inhabitants in the Netherlands. The premalignant lesions included were erythroplasia of Queyrat; Bowen's disease; bowenoid papulosis; mild, moderate and severe dysplasia; and carcinoma in situ of the penis. The terminology used in the pathological reports was translated to penile intraepithelial neoplasia. The grading was made analogous to that of vulvar premalignant lesions. RESULTS: The PALGA database enrolled 380 patients with premalignant penile lesions. Severe premalignant lesions, penile intraepithelial neoplasia III, were found in 254 patients (67%), penile intraepithelial neoplasia II in 84 (22%) and penile intraepithelial neoplasia I in 42 patients (11%). Most lesions were located on the prepuce (45%), followed by glans (38%) and shaft (3%). The median age of patients with penile intraepithelial neoplasia was 58 years. Progression to malignant disease occurred (2% for penile intraepithelial neoplasia I vs 7% for penile intraepithelial neoplasia III) in 26 patients. CONCLUSIONS: Penile intraepithelial neoplasia is a rarely diagnosed condition. Because of the wide variation of terms used for premalignant intraepithelial neoplasia of the penis, we recommend restricting this nomenclature to penile intraepithelial neoplasia.

Risk Factors for Acanthamoeba Keratitis-A Multistate Case-Control Study, 2008-2011.

OBJECTIVE: To identify modifiable risk factors contributing to Acanthamoeba keratitis (AK) infection. METHODS: A case-control investigation was conducted. Case patients were soft contact lens wearers with laboratory-confirmed AK. Control were soft contact lens wearers ≥12 years of age, with no history of AK. Case patients were recruited from 14 ophthalmology referral centers and a clinical laboratory. Control were matched on state of residence and type of primary eye care provider (ophthalmologist or optometrist). Participants were interviewed using a standardized questionnaire. Univariable and multivariable conditional logistic regression analyses were conducted. Matched odds ratios (mORs) were calculated. RESULTS: Participants included 88 case patients and 151 matched control. Case patients were more likely to be aged <25 years (unadjusted mOR 2.7, 95% confidence interval 1.3-5.5) or aged >53 years (mOR 2.5, 1.1-5.7), and more likely to be men (mOR 2.6, 1.4-4.8). Unadjusted analyses identified multiple risk factors: rinsing (mOR 6.3, 1.3-29.9) and storing lenses in tap water (mOR 3.9, 1.2-12.3), topping off solution in the lens case (mOR 4.0, 2.0-8.0), having worn lenses ≤5 years (mOR 2.4, 1.3-4.4), rinsing the case with tap water before storing lenses (mOR 2.1, 1.1-4.1), and using hydrogen peroxide (mOR 3.6, 1.1-11.7) versus multipurpose solution. Significant risk factors in multivariable modeling included age >53 years, male sex, topping off, and using saline solution. CONCLUSIONS: Numerous modifiable risk factors for AK were identified, mostly involving hygiene practices. To reduce the risk of AK, lens wearers should observe recommended lens care practices.

Evaluation of the Use of Zero-Augmented Regression Techniques to Model Incidence of Campylobacter Infections in FoodNet.

The Foodborne Diseases Active Surveillance Network (FoodNet) is currently using a negative binomial (NB) regression model to estimate temporal changes in the incidence of Campylobacter infection. FoodNet active surveillance in 483 counties collected data on 40,212 Campylobacter cases between years 2004 and 2011. We explored models that disaggregated these data to allow us to account for demographic, geographic, and seasonal factors when examining changes in incidence of Campylobacter infection. We hypothesized that modeling structural zeros and including demographic variables would increase the fit of FoodNet's Campylobacter incidence regression models. Five different models were compared: NB without demographic covariates, NB with demographic covariates, hurdle NB with covariates in the count component only, hurdle NB with covariates in both zero and count components, and zero-inflated NB with covariates in the count component only. Of the models evaluated, the nonzero-augmented NB model with demographic variables provided the best fit. Results suggest that even though zero inflation was not present at this level, individualizing the level of aggregation and using different model structures and predictors per site might be required to correctly distinguish between structural and observational zeros and account for risk factors that vary geographically.

Ceftriaxone-Resistant Nontyphoidal Salmonella from Humans, Retail Meats, and Food Animals in the United States, 1996-2013.

BACKGROUND: Ceftriaxone resistance in Salmonella is a serious public health threat. Ceftriaxone is commonly used to treat severe Salmonella infections, especially in children. Identifying the sources and drivers of ceftriaxone resistance among nontyphoidal Salmonella is crucial. MATERIALS AND METHODS: The National Antimicrobial Resistance Monitoring System (NARMS) tracks antimicrobial resistance in foodborne and other enteric bacteria from humans, retail meats, and food animals. We examined NARMS data reported during 1996-2013 to characterize ceftriaxone-resistant Salmonella infections in humans. We used Spearman rank correlation to examine the relationships between the annual percentage of ceftriaxone resistance among Salmonella isolates from humans with isolates from retail meats and food animals. RESULTS: A total of 978 (2.9%) of 34,100 nontyphoidal Salmonella isolates from humans were resistant to ceftriaxone. Many (40%) ceftriaxone-resistant isolates were from children younger than 18 years. Most ceftriaxone-resistant isolates were one of three serotypes: Newport (40%), Typhimurium (26%), or Heidelberg (12%). All were resistant to other antimicrobials, and resistance varied by serotype. We found statistically significant correlations in ceftriaxone resistance between human and ground beef Newport isolates (r = 0.83), between human and cattle Typhimurium isolates (r = 0.57), between human and chicken Heidelberg isolates (r = 0.65), and between human and turkey Heidelberg isolates (r = 0.67). CONCLUSIONS: Ceftriaxone resistance among Salmonella Newport, Typhimurium, and Heidelberg isolates from humans strongly correlates with ceftriaxone resistance in isolates from ground beef, cattle, and poultry, respectively. These findings support other lines of evidence that food animals are important reservoirs of ceftriaxone-resistant Salmonella that cause human illness in the United States.

Antimicrobial Resistance Among Nontyphoidal Salmonella Isolated From Blood in the United States, 2003-2013.

BACKGROUND: Salmonella causes an estimated 100 000 antimicrobial-resistant infections annually in the United States. Salmonella antimicrobial resistance may result in bacteremia and poor outcomes. We describe antimicrobial resistance among nontyphoidal Salmonella blood isolates, using data from the National Antimicrobial Resistance Monitoring System. METHODS: Human nontyphoidal Salmonella isolates from 2003 to 2013 were classified as fully susceptible, resistant to ≥1 antimicrobial agent, or resistant to a first-line agent. Logistic regression was used to compare resistance patterns, serotypes, and patient characteristics for Salmonella isolated from blood versus stool and to determine resistance trends over time. RESULTS: Approximately 20% of blood isolates had antimicrobial resistance to a first-line treatment agent. Bacteremia was associated with male sex, age ≥65 years, and specific serotypes. Blood isolates were more likely to be resistant to ≥1 agent for serotypes Enteritidis, Javiana, Panama, and Typhimurium. Blood isolates were most commonly resistant to tetracycline (19%), and more likely resistant to a first-line agent (odds ratio, 1.81; 95% confidence interval, 1.56-2.11) than stool isolates. Ceftriaxone resistance increased in blood isolates from 2003 to 2013 (odd ratio, 1.12; 95% confidence interval, 1.02-1.22). CONCLUSIONS: Resistance to first-line treatment agents in patients with Salmonella bacteremia is a concern for public health and for informing clinical decisions. Judicious antimicrobial use is crucial to limit resistance.

Assessment of the Incubation Period for Invasive Listeriosis.

We characterized incubation periods among outbreak-associated listeriosis cases, using a simulation model to account for patients with multiple exposure dates. The median was 11 days; 90% of cases occurred within 28 days, and incubation periods varied by clinical manifestation.

Estimated Incidence of Antimicrobial Drug-Resistant Nontyphoidal Salmonella Infections, United States, 2004-2012.

Salmonella infections are a major cause of illness in the United States. The antimicrobial agents used to treat severe infections include ceftriaxone, ciprofloxacin, and ampicillin. Antimicrobial drug resistance has been associated with adverse clinical outcomes. To estimate the incidence of resistant culture-confirmed nontyphoidal Salmonella infections, we used Bayesian hierarchical models of 2004-2012 data from the Centers for Disease Control and Prevention National Antimicrobial Resistance Monitoring System and Laboratory-based Enteric Disease Surveillance. We based 3 mutually exclusive resistance categories on susceptibility testing: ceftriaxone and ampicillin resistant, ciprofloxacin nonsusceptible but ceftriaxone susceptible, and ampicillin resistant but ceftriaxone and ciprofloxacin susceptible. We estimated the overall incidence of resistant infections as 1.07/100,000 person-years for ampicillin-only resistance, 0.51/100,000 person-years for ceftriaxone and ampicillin resistance, and 0.35/100,000 person-years for ciprofloxacin nonsusceptibility, or ≈6,200 resistant culture-confirmed infections annually. These national estimates help define the magnitude of the resistance problem so that control measures can be appropriately targeted.

Comparing Characteristics of Sporadic and Outbreak-Associated Foodborne Illnesses, United States, 2004-2011.

Outbreak data have been used to estimate the proportion of illnesses attributable to different foods. Applying outbreak-based attribution estimates to nonoutbreak foodborne illnesses requires an assumption of similar exposure pathways for outbreak and sporadic illnesses. This assumption cannot be tested, but other comparisons can assess its veracity. Our study compares demographic, clinical, temporal, and geographic characteristics of outbreak and sporadic illnesses from Campylobacter, Escherichia coli O157, Listeria, and Salmonella bacteria ascertained by the Foodborne Diseases Active Surveillance Network (FoodNet). Differences among FoodNet sites in outbreak and sporadic illnesses might reflect differences in surveillance practices. For Campylobacter, Listeria, and Escherichia coli O157, outbreak and sporadic illnesses are similar for severity, sex, and age. For Salmonella, outbreak and sporadic illnesses are similar for severity and sex. Nevertheless, the percentage of outbreak illnesses in the youngest age category was lower. Therefore, we do not reject the assumption that outbreak and sporadic illnesses are similar.

Postdiarrheal hemolytic uremic syndrome in United States children: clinical spectrum and predictors of in-hospital death.

OBJECTIVE: To assess the clinical spectrum of postdiarrheal hemolytic uremic syndrome (D(+)HUS) hospitalizations and sought predictors of in-hospital death to help identify children at risk of poor outcomes. STUDY DESIGN: We assessed clinical variables collected through population-based surveillance of D(+)HUS in children <18 years old hospitalized in 10 states during 1997-2012 as predictors of in-hospital death by using tree modeling. RESULTS: We identified 770 cases. Of children with information available, 56.5% (430 of 761) required dialysis, 92.6% (698 of 754) required a transfusion, and 2.9% (22 of 770) died; few had a persistent dialysis requirement (52 [7.3%] of 716) at discharge. The tree model partitioned children into 5 groups on the basis of 3 predictors (highest leukocyte count and lowest hematocrit value during the 7 days before to 3 days after the diagnosis of hemolytic uremic syndrome, and presence of respiratory tract infection [RTI] within 3 weeks before diagnosis). Patients with greater leukocyte or hematocrit values or a recent RTI had a greater probability of in-hospital death. The largest group identified (n = 533) had none of these factors and had the lowest odds of death. Many children with RTI had recent antibiotic treatment for nondiarrheal indications. CONCLUSION: Most children with D(+)HUS have good hospitalization outcomes. Our findings support previous reports of increased leukocyte count and hematocrit as predictors of death. Recent RTI could be an additional predictor, or a marker of other factors such as antibiotic exposure, that may warrant further study.

A rapid assessment of drinking water quality in informal settlements after a cholera outbreak in Nairobi, Kenya.

Populations living in informal settlements with inadequate water and sanitation infrastructure are at risk of epidemic disease. In 2010, we conducted 398 household surveys in two informal settlements in Nairobi, Kenya with isolated cholera cases. We tested source and household water for free chlorine residual (FCR) and Escherichia coli in approximately 200 households. International guidelines are ≥0.5 mg/L FCR at source, ≥0.2 mg/L at household, and <1 E. coli/100 mL. In these two settlements, 82% and 38% of water sources met FCR guidelines; and 7% and 8% were contaminated with E. coli, respectively. In household stored water, 82% and 35% met FCR guidelines and 11% and 32% were contaminated with E. coli, respectively. Source water FCR≥0.5 mg/L (p=0.003) and reported purchase of a household water treatment product (p=0.002) were associated with increases in likelihood that household stored water had ≥0.2 mg/L FCR, which was associated with a lower likelihood of E. coli contamination (p<0.001). These results challenge the assumption that water quality in informal settlements is universally poor and the route of disease transmission, and highlight that providing centralized water with ≥0.5 mg/L FCR or (if not feasible) household water treatment technologies reduces the risk of waterborne cholera transmission in informal settlements.

Bacterial Enteric Infections Among Older Adults in the United States: Foodborne Diseases Active Surveillance Network, 1996-2012.

BACKGROUND: A growing segment of the population-adults aged ≥65 years-is more susceptible than younger adults to certain enteric (including foodborne) infections and experience more severe disease. MATERIALS AND METHODS: Using data on laboratory-confirmed infections from the Foodborne Diseases Active Surveillance Network (FoodNet), we describe trends in the incidence of Campylobacter spp., Escherichia coli O157, Listeria monocytogenes, and nontyphoidal Salmonella infections in adults aged ≥65 years over time and by age group and sex. We used data from FoodNet and other sources to estimate the total number of illnesses, hospitalizations, and deaths in the United States caused by these infections each year using a statistical model to adjust for underdiagnosis (taking into account medical care-seeking, stool sample submission, laboratory practices, and test sensitivity). RESULTS: From 1996 to 2012, 4 pathogens caused 21,405 laboratory-confirmed infections among older adults residing in the FoodNet surveillance area; 49.3% were hospitalized, and 2.6% died. The average annual rate of infection was highest for Salmonella (12.8/100,000) and Campylobacter (12.1/100,000). Salmonella and Listeria led as causes of death. Among older adults, rates of laboratory-confirmed infection and the percentage of patients who were hospitalized and who died generally increased with age. A notable exception was the rate of Campylobacter infections, which decreased with increasing age. Adjusting for underdiagnosis, we estimated that these pathogens caused about 226,000 illnesses (≈600/100,000) annually among U.S. adults aged ≥65 years, resulting in ≈9700 hospitalizations and ≈500 deaths. CONCLUSIONS: Campylobacter, E. coli O157, Listeria, and Salmonella are major contributors to illness in older adults, highlighting the value of effective and targeted intervention.

2008 outbreak of Salmonella Saintpaul infections associated with raw produce.

BACKGROUND: Raw produce is an increasingly recognized vehicle for salmonellosis. We investigated a nationwide outbreak that occurred in the United States in 2008. METHODS: We defined a case as diarrhea in a person with laboratory-confirmed infection with the outbreak strain of Salmonella enterica serotype Saintpaul. Epidemiologic, traceback, and environmental studies were conducted. RESULTS: Among the 1500 case subjects, 21% were hospitalized, and 2 died. In three case-control studies of cases not linked to restaurant clusters, illness was significantly associated with eating raw tomatoes (matched odds ratio, 5.6; 95% confidence interval [CI], 1.6 to 30.3); eating at a Mexican-style restaurant (matched odds ratio, 4.6; 95% CI, 2.1 to ∞) and eating pico de gallo salsa (matched odds ratio, 4.0; 95% CI, 1.5 to 17.8), corn tortillas (matched odds ratio, 2.3; 95% CI, 1.2 to 5.0), or salsa (matched odds ratio, 2.1; 95% CI, 1.1 to 3.9); and having a raw jalapeño pepper in the household (matched odds ratio, 2.9; 95% CI, 1.2 to 7.6). In nine analyses of clusters associated with restaurants or events, jalapeño peppers were implicated in all three clusters with implicated ingredients, and jalapeño or serrano peppers were an ingredient in an implicated item in the other three clusters. Raw tomatoes were an ingredient in an implicated item in three clusters. The outbreak strain was identified in jalapeño peppers collected in Texas and in agricultural water and serrano peppers on a Mexican farm. Tomato tracebacks did not converge on a source. CONCLUSIONS: Although an epidemiologic association with raw tomatoes was identified early in this investigation, subsequent epidemiologic and microbiologic evidence implicated jalapeño and serrano peppers. This outbreak highlights the importance of preventing raw-produce contamination.