Outbreak of Listeriosis Associated With Homemade Mexican-StyleCheese—NorthCarolina,October2000–January2001

On November 13, 2000, health-care providers at a hospital in Winston-Salem, North Carolina, contacted the local health department about three cases of listeriosis within a 2-week period in recent Mexican immigrants. The North Carolina General Communicable Disease Control Branch, in collaboration with the Forsyth County Health Department, the North Carolina Departments of Agriculture and Consumer Services (NCDA&CS) and Environment and Natural Resources, the Food and Drug Administration (FDA), and CDC investigated this outbreak of Listeria monocytogenes infections. This report summarizes the results of the investigation, which implicated noncommercial, homemade, Mexican-style fresh soft cheese produced from contaminated raw milk sold by a local dairy farm as the causative agent. Culturally appropriate education efforts are important to reduce the risk for L. monocytogenes transmission through Mexican-style fresh soft cheese.

providers at a hospital in Winston-Salem, North Carolina, contacted the local health department about three cases of listeriosis within a 2-week period in recent Mexican immigrants. The North Carolina General Communicable Disease Control Branch, in collaboration with the Forsyth County Health Department, the North Carolina Departments of Agriculture and Consumer Services (NCDA&CS) and Environment and Natural Resources, the Food and Drug Administration (FDA), and CDC investigated this outbreak of Listeria monocytogenes infections. This report summarizes the results of the investigation, which implicated noncommercial, homemade, Mexican-style fresh soft cheese produced from contaminated raw milk sold by a local dairy farm as the causative agent. Culturally appropriate education efforts are important to reduce the risk for L. monocytogenes transmission through Mexican-style fresh soft cheese.
A case was defined as L. monocytogenes (isolated from a normally sterile site or with placental tissue staining positive using immunohistochemical techniques) in a mother of a stillborn or premature infant (Ͻ37 weeks' gestation), or a mother with a febrile illness, who was a Winston-Salem resident during October 24, 2000-January 1, 2001. Through active case finding, 12 cases were identified. On initial interview, most patients reported eating unlabeled Mexican-style fresh soft cheese bought at local markets or from doorto-door vendors. A case-control study was conducted to determine risk factors for illness; the questionnaire addressed symptoms, diet, and groceryshopping histories during the month preceding illness. L. monocytogenes isolates from patients, raw milk, and cheese were tested using pulsed-field gel electrophoresis (PFGE). Environmental inspections of homes, local markets, and dairy farms were conducted.
All 12 patients were Hispanic; 11 were women with a median age of 21 years (range: 18-38 years), and one was a 70year-old immunocompromised man. All but one infection were laboratory confirmed. The 11 women did not speak English, were born in Mexico, and had resided in the United States for a median of 2 years (range: 0-5 years). One had traveled outside Forsyth County during the month preceding illness. Ten women were pregnant, and infection with L. monocytogenes resulted in five stillbirths, three premature deliveries, and two infected newborns. The 11th woman was 5 months postpartum when she presented to a local hospital with meningitis caused by L. monocytogenes. She had no preexisting medical conditions. The male patient, who presented with a brain abscess, was receiving corticosteroid therapy after brain tumor surgery. On hospital admission, the 11 women reported symptoms that included fever (nine), chills (nine), headache (nine), abdominal cramps (five), stiff neck (five), vomiting (three), and photophobia (two).
The male patient was excluded from the case-control study because of difficulty finding suitable controls. In the case-control study, a mother and her fetus or newborn were counted as one case-patient. Controls were identified at a Women, Infants, and Children program office and through the county's record of women enrolled in the state's Baby Love Program, which provides outreach and prenatal-care home visits. A median of four controls (range: three to six controls) per case was selected. Controls were restricted to female Hispanic Winston-Salem residents and matched to patients by age and pregnancy status.
Various members of the Hispanic immigrant community made the Mexicanstyle fresh soft cheese from raw milk in their homes. Inspectors found unlabeled homemade cheese in all three of the small local Latino grocery stores they visited in Winston-Salem. In addition, many persons regularly sold the cheese in parking lots and by going door-todoor. Owners of two local dairies reported selling raw milk. Milk samples were obtained from these two Forsyth County dairies and from three dairies in neighboring counties. L. monocytogenes isolates were obtained from nine patients, three cheese samples from two stores, one cheese sample from the home of a patient, and one raw milk sample from a manufacturing grade dairy. All 14 isolates had indistinguishable PFGE patterns, indicating a common link.
NCDA&CS conducted an investigation at a manufacturing grade dairy farm to determine the potential source of L. monocytogenes contamination. NCDA&CS collected milk samples from all 49 cows in the herd and samples from the bulk milk storage tanks. Milk from each cow was tested for somatic cell FROM THE CENTERS FOR DISEASE CONTROL AND PREVENTION count to identify mastitic cows. Milk from each cow also was tested for presence of L. monocytogenes. Repeated testing did not identify any cow with milk confirmed positive for L. monocytogenes, suggesting that the cows were not infected and that L. monocytogenes may have originated from environmental contamination.
As a result of this outbreak, North Carolina health authorities stopped the sale of raw milk by the dairy farm to noncommercial processors and educated store owners that it is illegal to sell unregulated dairy products. Officials cited the outbreak as sufficient reason to strengthen laws prohibiting the sale of raw milk except to regulated processors. Using already established programs (e.g., Baby Love Program), North Carolina officials recommended reinforcing and expanding the community awareness of the hazards of eating unpasteurized fresh cheese while pregnant. Finally, steps were taken to add listeriosis to the list of reportable diseases in North Carolina.

CDC Editorial Note: The investigation of this outbreak implicated Mexi-
can-style fresh soft cheese made from unpasteurized milk and hotdogs, two vehicles commonly identified as causes of L. monocytogenes outbreaks. The laboratory investigation resulted in isolation of L. monocytogenes from patients, cheese, and raw milk at a dairy farm. Molecular subtyping identified indistinguishable PFGE patterns, establishing the link between human disease, the cheese, and the source of the raw milk used to make the cheese.
Because of the health risks associated with the consumption of raw milk and raw milk products, FDA requires pasteurization of all dairy products sold across state lines except cheese made from raw milk that has to be aged a minimum of 60 days. 1,2 Despite North Carolina laws prohibiting the sale and consumption of raw milk and raw milk products, such practices persist in some communities as a result of consumers' taste preferences and for cultural reasons. The popularity of queso fresco, a Mexican-style fresh soft cheese made from unpasteurized milk, has resulted in several outbreaks in Hispanic communities since the 1980s. In 1985, an outbreak of septic abortions attributed to L. monocytogenes occurred among Hispanics in Los Angeles and Orange counties, California. 3 In 1997, three outbreaks of multidrug resistant Salmonella serotype Typhimurium DT104 complex strains occurred in Hispanic communities in northern California and Washington. 4,5 Because queso fresco in these communities is produced in private homes, food safety regulations are difficult to enforce. Education of milk and cheese producers and consumers about the increased risk for acquiring infections, particularly L. monocytogenes, from consuming unpasteurized milk or fresh soft cheese made from unpasteurized milk, complemented by regulatory action, are the keys to making cheese safe. Successful communication of public health messages to the Hispanic community about the risk for eating Mexican-style fresh soft cheese made from raw milk can be challenging because of language and other social barriers.
The findings in this report are subject to at least four limitations. First, interviewers were not blinded to the status of the persons they were interviewing. Second, efforts were made to select controls from the same population as case-patients; however, controls were selected on the basis of use of public health service programs. Most controls were selected from a county registry for a free prenatal care program that does not require documentation to obtain service. Third, during the study, rumors spread in the community that the suspected vehicle of infection was homemade Mexican-style fresh soft cheese. Fi-nally, patients may have had better recall of potential exposures than controls.
Following a listeriosis outbreak in Yakima County, Washington, an education program to train grandmothers, the primary cheese producers in that community, in the safe production of soft cheeses was introduced and was well received. A licensing requirement for commercial cheese makers and appropriate regulatory action also may curtail the sale of fresh soft cheese made from unpasteurized milk. Twenty-eight states permit the sale of raw milk directly from farmers to consumers. 6 Until all states prohibit such sales, outbreaks associated with eating queso fresco and other unpasteurized dairy products may continue despite efforts to educate consumers, especially those who do not speak or read English and whose cultural dietary habits favor such products. patient and two additional inmates in the dormitory (dorm Y). On May 15, 2000, the state's department of health and department of corrections and CDC initiated an investigation to identify additional cases and determine risk factors for HBV infection. This report summarizes the results of the investigation, which identified additional cases of HBV infection in this correctional facility and underscores the need to implement hepatitis B vaccination in correctional facilities.

REFERENCES
Current inmates who had resided in dorm Y at any time since October 1, 1999, were offered serologic testing for HBV infection and were interviewed about exposures during the preceding 6 months, including sexual activity, being tattooed, sustaining a cut or injury, being exposed to another inmate's blood, sharing a razor, and injection drug use. Acute HBV infection was defined as the presence of IgM antibody to hepatitis B core antigen (IgM anti-HBc) with or without the presence of hepatitis B surface antigen (HBsAg). Chronic HBV infection was defined as the presence of HBsAg and total (IgG and IgM) anti-HBc, and absence of IgM anti-HBc. Resolved infection was defined as the presence of total anti-HBc, but absence of IgM anti-HBc and HBsAg. Persons testing negative for anti-HBc and HBsAg were considered susceptible to HBV infection.
Of 103 eligible inmates, 97 (94%), including the sexual contacts of the inmate with chronic infection, consented to serologic testing. Of these 97 inmates, six (6%) had acute HBV infection, one (1%) had chronic infection, and 16 (16%) had resolved infection. The acute HBV infection rate among susceptible dorm Y inmates was 8%. Two inmates reported nonspecific symptoms (e.g., influenza-like illness) during the preceding 6 months. In addition to the index patient, one of the two other sexual contacts of the inmate with chronic infection had acute infection.
The six inmates with acute infection and 70 (95%) of 74 susceptible inmates were interviewed. Having sex with another man was the only risk factor as-sociated with acute HBV infection (risk ratio = 12.2; 95% confidence inter-val=3.5-42.2) and accounted for two of six acute infections.
The correctional facility is comprised of 14 dormitories housing 96 inmates each; it operates at 99% capacity. Inmates move within the facility to participate in daily scheduled activities and frequently move among dormitories during their incarceration. Condoms are not available to inmates. Because of the HBV transmission in dorm Y, on June 6, 2000, serologic testing was offered to inmates who resided in the remainder of the facility to determine if further HBV transmission had occurred.
Of 1247 inmates in the remainder of the facility, 1026 (82%) consented to serologic testing and completed a selfadministered questionnaire, which collected information on demographic characteristics and history of behaviors or characteristics that may have placed them at risk for HBV infection both during incarceration and during their lifetime. Of the 1026 inmates, 10 (1%) had chronic HBV infection and 178 (17%) had resolved infection. Of 838 susceptible inmates, five (Ͻ1%) were identified with previously undiagnosed acute HBV infection, resulting in an acute infection rate of 0.6% among inmates who did not reside in dorm Y, and an overall infection rate of 1.2% (11 of 918). Of the inmates with acute infection who did not reside in dorm Y, two were housed in one dormitory and the remainder resided in three other dormitories. None reported risk factors for HBV infection during the preceding 6 months.
Risk behaviors were evaluated to determine the potential for susceptible inmates to acquire HBV infection. Among the 907 susceptible inmates who completed the questionnaire, 473 (52%) reported at least one exposure while incarcerated that could have resulted in HBV transmission. These included injecting drugs ( To control the outbreak, the state's department of corrections offered hepatitis B vaccination to all susceptible inmates in dorm Y. In addition, acutely and chronically infected inmates were notified of their infection status, received a clinical assessment, and postexposure prophylaxis was provided to their contacts. The state's department of health and department of corrections are collaborating to implement routine hepatitis B vaccination for all inmates in the correctional system.

CDC Editorial Note:
The findings in this report document HBV transmission in a correctional facility, including a cluster of cases of acute infection in one dormitory and additional cases distributed throughout the facility. Most persons with acute HBV infection in the correctional facility were asymptomatic, and serologic surveys were needed to determine the extent of HBV transmission. The overall infection rate of 1% reflected infections acquired during the preceding 6 months and was higher than the estimated incidence of 1% per year in previous studies. 1,2 This serologic survey also indicated that 1% of inmates had chronic infection and that none were aware of their infection status.
HBV is transmitted primarily by percutaneous or permucosal exposures to an infected person. Risk factors associated with HBV infection include having multiple sex partners, having had an STD, being a man who has sex with men, injection drug use, and being a sexual or nonsexual household contact of a person with chronic HBV infection. 3 Receiv-FROM THE CENTERS FOR DISEASE CONTROL AND PREVENTION ing a tattoo has not been associated with community acquired HBV infections among nonincarcerated populations in the United States 4 ; however, transmission could occur if the tattoo is applied using contaminated equipment.
Sex with another man accounted for only 20% of new infections in this investigation. However, this and other behaviors prohibited by the correctional facility (e.g., injecting drugs) probably are underreported by inmates. Inmates with previously unrecognized chronic HBV infection may have served as a source for infection, similar to household contacts of persons with chronic infection. 5 Housing data were not available to determine if persons with acute HBV infection were more likely to have been a cellmate of a chronically infected inmate.
The findings in this report are consistent with previous reports of HBV transmission in prison settings. 1,2 Since 1982, the Advisory Committee on Immunization Practices has recommended hepatitis B vaccination of longterm inmates with a history of risk factors for infection. 3 Although a large proportion of inmates in this prison reported current or previous risk factors for HBV infection, none of the susceptible inmates had been vaccinated.
In the state correctional system in this report, approximately one third of inmates are released each year (Department of Corrections, unpublished data, 2000). Previously incarcerated persons represent a population at risk for HBV infection. Approximately 30% of persons with acute hepatitis B report a history of incarceration. 6 Hepatitis B vaccination of prisoners would prevent ongoing HBV transmission among inmates in prison facilities and after they have been released into the community. Because of the high proportion of inmates with previous risk factors for HBV infection and the difficulty in ascertaining current risk factors, experts in correctional health recommend vaccination of all inmates. 7 Some states have implemented successfully routine hepatitis B vaccination of prisoners. However, identifying resources to purchase and administer vaccine remains the major barrier to national implementation of this strategy. Partnerships between state health and corrections departments can help to implement hepatitis B vaccination and promote effective strategies for prevention of other STDs and infections in correctional facilities. 8 VITAMIN A, A FAT-SOLUBLE, HEATstable nutrient (retinol) derived from animal sources and certain fruits and vegetables, forms the basic component of retinal pigments and plays a vital role in optimal health, growth, and development. Vitamin A deficiency (VAD) (serum retinol Յ20 µg/dL [Յ0.7 µmol/L] for subclinical VAD) can substantially increase the risk for childhood mortality from infectious and noninfectious causes. 1-3 VAD impairs the mobilization and transport of iron and is usually associated with anemia and reduced growth. 4,5 VAD is a major public health problem in parts of Africa, Asia, Latin America, and the Western Pacific. 1,6 In Chuuk and Pohnpei, two of the four Federated States of Micronesia (FSM) (2000 population: 107,008), nutrition surveys during the early 1990s documented VAD prevalences among the highest in the world (CDC, unpublished data, 1991;U.S. Public Health Service, unpublished data, 1994). In response to these findings, FSM health authorities, with support of the United Nations Children's Fund (UNICEF), began distributing vitamin A supplements in 1993 and 1998 in Chuuk and Pohnpei, respectively. In November 1999, FSM requested assistance from CDC in VAD assessment surveys of children in Kosrae and Yap, the other two FSM states. This report summarizes levels of serum retinol and prevalence of VAD and other indicators of nutritional status among children aged 24-59 months in Kosrae and Yap. The findings indicated low serum retinol levels and high VAD prevalences but no substantial stunting or wasting. A comprehensive, long-term national strategy is needed in FSM to promote sustained improvement in vitamin A status.

REFERENCES
FSM is an island nation in the western Pacific Ocean. Kosrae state is a single island divided into 21 enumeration districts. Yap comprises four large islands and 134 small islands, primarily atolls, and is divided into 93 villages. For logistic reasons, only the three large islands connected by bridges (Yap proper) were included in the survey. These islands represent approximately 62% of the Yap population.
During January-February 2000, FSM health authorities, UNICEF, and CDC surveyed children aged 24-59 months and their mothers or reproductiveaged female caregivers in Kosrae and Yap. A separate cluster survey was performed in each state. The sample size for each state was calculated to yield a prevalence estimate with 5% error assuming 50% VAD prevalence. Because of uneven village sizes (range: 157-537 residents per village in Kosrae and one-580 in Yap), clusters were selected using the proportionate-to-population size sampling method. Investigators selected 13 villages in Kosrae and 29 villages in Yap. In each village, all children aged 24-59 months identified from a comprehensive list of vaccination records were eligible for the survey. Children were excluded who had moved into the village during the 6 months preceding the survey or had experienced fever or diarrhea during the preceding 24 hours or cough for Ն4 weeks. If more than one eligible child lived in a household, investigators randomly selected one for the survey.