Page images

In California, intrastate sale of raw milk and raw colostrum is legal and regulated. This report summarizes the investigation of these cases by CDPH, the California Department of Food and Agriculture (CDFA), and four local health departments and subsequent actions to prevent illnesses. As a result of this and other outbreaks, California enacted legislation (AB 1735), which took effect January 1, 2008, setting a limit of 10 coliforms/mL for raw milk sold to consumers. Raw milk in several forms, including colostrum, remains a vehicle of serious enteric infections, even if the sale of raw milk is regulated.

In mid-September 2006, the parent of one of the two children hospitalized with HUS notified CDFA that both children had consumed raw skim milk from dairy A in the days before illness onset. CDFA notified CDPH and the local health departments of the reports. Dairy A, a licensed raw milk dairy, sells raw milk, raw cream, raw butter, raw cheese, raw colostrum,* and kefir throughout California at retail stores and nationwide via Internet sales, all under a single brand (brand A).

On September 21, 2006, based on the reports from CDPH, CDFA issued a recall and quarantine order for all raw milk, raw cream, and raw colostrum produced by dairy A. The order was extended on September 22 to include all raw products from dairy A, except for cheeses aged at least 60 days according to California and Food and Drug Administration (FDA) standards.† Dairy A also was placed under a separate restriction by CDFA during September 21-29 that prevented it from bottling fluid milk and cream because of persistent high standard plate counts.

For this investigation, a case was defined as illness with an onset date of August 1, 2006, or later in a California resident with 1) culture-confirmed E. coli O157:H7 infection with the outbreak strain or 2) HUS with or without culture confirmation, and exposure to raw milk. Case finding was conducted by notifying all California local health departments and infection-control practitioners and reviewing molecular subtyping results from the CDPH Microbial Diseases Laboratory. The 61 health jurisdictions in California were notified on September 20, 2006, to be alert for cases of E. coli O157:H7 and other Shiga toxin-producing E. coli associated with consumption of raw milk. They were asked to report immediately to CDPH any enteric illnesses associated with raw milk or colostrum consumption.

Raw colostrum is secreted during the first few days after giving birth. It contains higher amounts of protein and antibodies than regular raw milk, but is processed in the same way as raw milk.

* The 60-day curing process has historically been considered sufficient to eliminate or reduce pathogens that were in the milk; however, its efficacy has been questioned, and FDA is reviewing the safety of raw milk cheeses.

Six cases were identified; four persons had culture-confir infections, one had a culture-confirmed infection and H and one had HUS only. The median age of patients w years (range: 6–18 years), and four of the patients (67%) ** boys. The six cases identified during this investigation w geographically dispersed throughout California. All six patien reported bloody diarrhea; three (50%) were hospitalized. " ness onset occurred during September 6-24, 2006. I from the five patients with culture-confirmed infections ha indistinguishable pulsed-field gel electrophoresis (PFGE) p terns. The PFGE pattern was new to the PulseNet National Molecular Subtyping Network for Foodborne D ease) database and differed markedly from the pattern of th E. coli O157:H7 strain associated with a concurrent multistr outbreak linked to spinach consumption (1). Four of the fo E. coli O157:H7 isolates were subtyped by multiple-locus varable-number tandem repeat analysis (MLVA) according to a protocol used by CDPH laboratory and were found to han closely related MLVA patterns (2).

Five of six patients reported they had consumed brand raw dairy products in the week before their illness onset: the sixth patient denied drinking brand A raw milk, although family routinely purchased it. Among the five patients w consumed brand A dairy products, two consumed raw whole milk, two consumed raw skim milk, and one consumed raw chocolate-flavored colostrum. Four of the five patients retinely drank raw milk from dairy A. One patient was exposes to brand A dairy product only once; he was served raw choco late colostrum as a snack when visiting a friend. No othe food item was commonly consumed by all six patients. N other illness was reported among household members wh consumed brand A dairy products.

To assess the level of exposure to raw dairy products amcnf patients with E. coli O157:H7 infection, CDPH epidemisgists reviewed exposure histories for the 50 most recent E... O157:H7 cases reported to CDPH during 2004-2001 Among patients who had been asked about exposure to milk on the case report, only one of 47 (2%) had consund raw milk in the week before illness onset. Exposure to milk was similarly low (3%) among Californians who responded to a population survey (3).

Environmental Investigation

Using purchase information supplied by the patients' fane lies, investigators determined that the patients consumed r milk from lots produced at dairy A during September 3-13 2006. Milk samples from these production dates were no available for testing. Fifty-six product samples from sever lots with code dates of September 17, 2006, or later we retrieved from retails stores and dairy A and were tested

LE. Microbial testing results for dairy A raw milk product samples with code dates of September 17 through October 9, 2006 — California

[blocks in formation]

pic microflora, total coliform, fecal coliform, and E. coli 7:H7. The outbreak strain of E. coli O157:H7 was not d in any product samples. However, standard aerobic plate its and coliform counts of collected samples with code › of September 17 through October 9, 2006, were indicaof contamination (Table). Colostrum samples had high -dard plate counts and total coliform counts, and fecal orm counts of 210-46,000 MPN/g. California standards --- standard plate counts for raw and pasteurized milk to :00 CFU/mL and total coliform counts for pasteurized to 10 coliform bacteria/mL. At the time of this outbreak, ornia did not have a coliform standard for milk sold raw nsumers. California also classifies colostrum as a dietary lement, for which it has no microbiologic standards, rather a milk product.

OFA and CDPH conducted an initial inspection and onmental investigation of the milk plant and dairy on ember 26. E. coli O157:H7 was not isolated from any of ·· environmental samples. Samples from three heifers yielded li O157:H7, but the PFGE and MLVA patterns of these li O157:H7 isolates differed from the outbreak pattern. rted by: J Schneider, MPH, J Mohle-Boetani, MD, D Vugia, MD, ornia Dept of Public Health. M Menon, MD, EIS officer, CDC. orial Note: Raw cow milk and raw milk products have .... implicated in the transmission of multiple bacterial patho, including Campylobacter spp., Brucella, Listeria cytogenes, Salmonella spp., and E. coli. In a recent review coli O157 infections, raw milk products accounted for of outbreaks during a 20-year period (4). E. coli O157:H7 ponsible for an estimated 73,000 cases of illness annuand serious sequelae, including HUS and death (5). Chilolder adults, and persons with low levels of gastric acid articularly vulnerable (6).

[ocr errors]

w milk products tested from dairy A were not produced ng the same time as the products consumed by the nts in this outbreak. Although the outbreak strain of iO157 was not isolated from dairy A products, the tested acts did have high standard plate counts, many exceed

ing California standards for raw milk, and total coliform counts that exceeded California standards for pasteurized milk. Nonoutbreak strains of E. coli O157 also were isolated from samples from dairy A cows, indicating shedding of this pathogen in the herd. Raw milk from dairy A was the likely vehicle of transmission, but the exact mode of milk contamination in this outbreak was not determined. Asymptomatic cows can harbor pathogens and cause human illness by shedding pathogens in untreated milk or milk products. These findings suggest that if raw milk had been subject to the same coliform standard as pasteurized milk in California, milk from dairy A might have been excluded from sale and this outbreak might have been averted.

FDA mandates that all milk and milk products for direct human consumption be pasteurized in final package form if they are to be shipped for interstate sale (7). States regulate milk shipped within their state. Currently, 21 states require pasteurization of all milk products for sale. However, 25 states, including California, allow raw milk to be sold in some form to the public. Those states that permit the sale and consumption of raw milk report more outbreaks of foodborne disease attributed to raw milk than those states that have stricter regulations. During 1973-1992, raw milk was implicated in 46 reported outbreaks. Nearly 90% of these outbreaks (40 out of 46) occurred in states that allow the sale of raw milk, suggesting that even the regulated sale of raw milk might not be adequate to prevent associated illnesses (8).

This is the first outbreak reported to CDC in which colostrum has been an implicated food vehicle. This outbreak represents the first time colostrum has been reported to CDC as a form of raw milk consumed by any patients in raw milkassociated outbreak, although information on the type of raw milk is reported inconsistently in outbreak surveillance. Colostrum is purported to have increased concentrations of nutrients and protective antibodies and is marketed as a dietary supplement in California; consequently, it is regulated by the CDPH Food and Drug Branch. The colostrum products tested in this investigation were nearly as contaminated

as other forms of raw milk tested; therefore, in this outbreak, the risk for human illness from consuming either product was probably similar. Exemption of colostrum from state dairy regulations is not supported by the findings in this outbreak investigation.

From 1998 to May 2005, raw milk or raw milk products have been implicated in 45 foodborne illness outbreaks in the United States, accounting for more than 1,000 cases of illness (CDC, unpublished data, 2007). Because illnesses associated with raw milk continue to occur, additional efforts are needed to educate consumers and dairy farmers about illnesses associated with raw milk and raw colostrum. To reduce the risk for E. coli O157 and other infections, consumers should not drink raw milk or raw milk products.


This report is based, in part, on data contributed by D Buglino, MPH, K Smith-Sayer, Nevada County Public Health Dept; S Fortino, Riverside County Dept of Public Health; J Van Meter, MPH, County of San Diego Health and Human Svcs Administration; E Frykman, MD, San Bernardino County Dept of Public Health; L Crawford-Miksza, PhD, C Myers, S Himathongkham, DVM, M Palumbo, PhD, J Atwell, T Chang, D Csuti, S Fontanoz, Y Gerbremichael, J Glover, DVM, J O'Connell, B Sun, DVM, C Wheeler, MD, Y Zhao, California Dept of Public Health; California Dept of Food and Agriculture; and M Lynch, MD, Div of Foodborne, Bacterial, and Mycotic Diseases, CDC.


1. CDC. Ongoing multistate outbreak of Escherichia coli serotype O157:H7 infections associated with consumption of fresh spinach-United States, September 2006. MMWR 2006;55:1045–6.

2. Hyytiä-Trees E, Smole SC, Fields PA, Swaminathan B, Ribot EM. Second generation subtyping: a proposed PulseNet protocol for multiple-locus variable-number tandem repeat analysis of Shiga toxinproducing Escherichia coli O157 (STEC O157). Foodborne Pathog Dis 2006;3:118-31.

3. CDC. Foodborne diseases active surveillance network (FoodNet): population survey atlas of exposures, 2002. Atlanta, GA: US Department of Health and Human Services, CDC; 2004:205. Available at http://www.

4. Rangel JM, Sparling PH, Crowe C, Griffin PM, Swerdlow DL. Epidemiology of Escherichia coli O157:H7 outbreaks, United States, 1982– 2002. Emerg Infect Dis 2005;11:603–9.

5. Mead PS, Slutsker L, Dietz V, et al. Food-related illness and death in the United States. Emerg Infect Dis 1999;5:607–25.

6. Dundas S, Todd WT, Stewart AI, et al. The central Scotland Escherichia coli O157:H7 outbreak: risk factors for the hemolytic uremic syndrome and death among hospitalized patients. Clin Infect Dis 2001;33:923–31. 7. Food and Drug Administration. Grade “A” pasteurized milk ordinance: 2003 revision. Rockville, MD: US Department of Health and Human Services; 2004. Available at html.

8. Headrick ML, Korangy S, Bean NH, et al. The epidemiology of raw milk-associated foodborne disease outbreaks reported in the United States, 1973 through 1992. Am J Public Health 1998;88:1219–21.

[blocks in formation]

On August 29, 2007, the Connecticut Department of Pi lic Health was notified by a physician of suspect cutaneo anthrax involving a drum maker and one of his three chi dren. The drum maker had been working with untreated gar hides from Guinea in West Africa. This report summari results of the joint epidemiologic and environmental inve gation conducted by public health officials, environment agencies, and law enforcement authorities. The investigation revealed that the drum maker was exposed while working wit a contaminated goat hide from Guinea and that his won place and home were contaminated with anthrax. His child, was most likely exposed from cross-contamination of the home. The findings underscore the potential hazard of wor ing with untreated animal hides from areas with epizootic anthrax and the potential for secondary cases from envirormental contamination.

On July 22, while sanding a newly assembled goat-hide drur in his backyard shed, the drum maker felt a sting on his right forearm. He then proceeded to an upstairs bathroom in his house to wash his arm. Two days later, a painless 2 cm papular lesion with surrounding edema developed at the site. The man sought medical attention and was prescribed cephalexin and then clindamycin for a presumptive infected spider bite. On August 28, after the skin lesion progressed to an eschar with lymphangitic spread, the man consulted an infectious disease practitioner, who sent a biopsy specimen of the lesion to the Connecticut State Laboratory. Culture was negative, bu Bacillus anthracis was detected by polymerase chain reaction (PCR). The patient received ciprofloxacin for suspect cutane ous anthrax.

On August 31, the Connecticut Department of Pub Health was notified of a second suspect case of cutaneos anthrax in the drum maker's child aged 8 years, who de oped a painless, 1 cm ulcer of 3 days' duration over the scap that did not improve under treatment with amoxicilla clavulanate. Culture of the lesion was negative, but biegs specimens tested positive for B. anthracis by PCR at the Co necticut State Laboratory and by PCR and immunohistochem istry assay at CDC. The patient was treated with penicillir.

Also on August 31, an epidemiologic investigation was in tiated to identify the primary source of exposure and the extent of dissemination of B. anthracis spores. The investig tion included interviews with the index patient and his family and environmental testing. The family had moved into the

se in December 2006. The index patient made traditional t African drums (known as djembe drums) by soaking nal hides in water, stretching them over the drum body, scraping and sanding them. At the end of June, a contact New York City told the index patient that he had some goat hides from Guinea. Shortly thereafter, the index ent purchased 10 of them, making the transaction on a et corner in New York City. Whether these goat hides were orted legally is unknown. The index patient used three of e hides to make drums during the time he developed


1 animal hides and drums in progress were stored in a yard shed. Drum making usually occurred at the shed ance. The affected child never participated in any drum ing and had no known exposure to animal hides. He played ors on carpeted floors and was prohibited from entering shed.

nce childhood, the drum maker had been taught by his er, who also made djembe drums, to routinely use latex es and wear tight-fitting goggles when drum making. He was taught to use designated work clothes with long es, which were laundered periodically. In addition, the n maker wore disposable facemasks to avoid the strong -associated with animal hides. He always removed his clothes and shoes before entering the house. One excepto these practices occurred on July 22, when the drum er wore short sleeves and went indoors to an upstairs bathwithout removing his work attire. Although he kept all On making equipment in the shed, the drum maker somes brought other items from the shed into the house. n September 5 and 6, targeted environmental sampling conducted collaboratively by the Federal Bureau of Invesion (FBI), the Environmental Protection Agency, and the necticut Department of Environmental Protection. The chose to participate because anthrax is a select bioterrorism t.* On the basis of initial positive results for B. anthracis everal areas of the house, extensive testing was performed tek later to guide decontamination efforts. Specimens ided swabs of all hides and drum heads (Figure) after transto the state laboratory, seven of which underwent addial wipes and punch biopsies; 16 wipe samples of the shed, ading table surfaces and coat hooks 5 feet above the nd; and a swab sample of the car used for transporting ecently purchased hides. House testing included vacuum les from carpeted areas and composite wipe samples from ted hard surfaces in all regularly used areas.

mation on selected agents and toxins available from the CDC Select Agent ram at

1os and additional information available at es/danbury.

[merged small][graphic][merged small]

The following were culture positive for B. anthracis: six (24%) of 25 drum heads, including the recently sanded drum; 15 (42%) of 35 hides, some of which were exposed to ambient dust in the shed; all 16 shed samples, many indicating heavy growth; the car trunk; and 18 (26%) of 72 house specimens, including vacuum samples from the upstairs hallway and both affected patients' bedrooms and swab and wipe samples from the laundry room and upstairs bathroom. DNA from all environmental isolates of B. anthracis and the cutaneous biopsy specimens were sent to CDC for genotyping using multiple-locus variable-number tandem repeat analysis (MLVA) (1). All isolates were MLVA genotype 1, as was the B. anthracis DNA detected in the child's biopsy specimen.

Federal, state, and local officials completed a comprehensive remediation process that included fumigation of the house with chlorine dioxide. The house and shed were cleared for occupancy on December 22, 2007, after all post-remediation samples had tested negative for anthrax. Because of exposure to aerosolized spores in the shed from drum making, the drum maker was continued on ciprofloxacin for a total of 60 days from the date of last presumed exposure based on recommendations established by CDC for postexposure prophylaxis against inhalation anthrax (2). No other contacts were identified with potential inhalation exposure. With the exception of lymphangitic scarring of the drum maker's arm, the illnesses in both patients resolved without sequelae.

Reported by: J Stratidis, MD, Danbury Hospital; S LeRoy, MPH, Danbury Health Dept; D Barden, MT (HHS), K Kelley, PhD, J Fontana, PhD, K Purviance, MPH, M Cartter, MD, J Hadler, MD, Connecticut Dept of Public Health. K Glynn, DVM, A Hoffmaster, PhD, M Guerra, DVM, S Shadomy, DVM, T Smith, MD, C Marston, National Center for Zoonotic, Vector-Borne, and Enteric Diseases; K Martinez, MSEE, National Institute for Occupational Safety and Health; A Guh, MD, EIS Officer, CDC.

Editorial Note: This report highlights the individual and environmental risks for anthrax from using contaminated goat hides brought from West Africa for drum making. It also describes the first case in the United States of naturally acquired cutaneous anthrax in a personal contact caused by cross-contamination from drum making.

Since 2006, three unrelated cases of anthrax, including the first case described in this report, have been reported from direct occupational association with djembe drums made from untreated animal hides from West Africa. The first two cases were inhalation anthrax. One occurred in a New York City drum maker exposed while making a djembe drum from contaminated hides, and the other occurred in a man in Scotland who died of anthrax septicemia after playing or handling djembe drums newly made from contaminated hides (3,4). The Connecticut cases and the New York City case were caused by B. anthracis of MLVA genotype 1, a different genotype than the Ames strain used in the 2001 mail-related anthrax attacks (1). Although MLVA genotypes from West Africa have not been systematically studied, the widespread nature of genotype 1 (1) and its presence in the West African hides implicated in the New York City and Connecticut cases suggest that genotype 1 might be commonly found in West Africa.

The drum making process of stretching, scraping, and sanding animal hides could have released and potentially aerosolized any B. anthracis spores present on untreated hides, exposing the drum maker and contaminating the surrounding environment. However, despite direct exposure, the drum maker described in this report did not develop inhalation anthrax. He developed cutaneous anthrax only after wearing short sleeves and experiencing a penetrating injury or insect bite, which could have served to inoculate spores into the skin.

The Connecticut drum maker routinely wore personal protective equipment (PPE). His wearing a facemask might have reduced the amount of inhalation exposure. However, even if he had worked with all recommended precautions (3), such as working in a well-ventilated area using PPE that included a N95 respirator, his risk for cutaneous and inhalation exposure would have been lessened but not necessarily eliminated, and environmental contamination would still have occurred and required remediation.

In this investigation, environmental sampling indices tracking of spores into the house by the drum maker, et through his work clothes or objects brought from the s leading to exposure and subsequent development of cut ous anthrax in his child. Few cases of anthrax have reported in children in the United States because most ex sures are acquired occupationally. However, household me bers can be exposed through cross-contamination of liv areas. In 1978, dust samples from vacuum cleaners in houses of textile mill workers tested positive for B. anthra suggesting that workers carried spores into their homes 5 case series of cutaneous anthrax in a Pennsylvania mill to indicated that 4% of all cases during a 22-year perie occurred in household members of mill workers, includi their children (6).

Decontamination of affected areas to minimize the risk f secondary cases of anthrax can be time-consuming and expensive. The cost of environmental cleanups on Capitol H in the District of Columbia and in postal facilities affected s the 2001 anthrax attacks ranged from $464,000 to $200, million (7).

To eliminate individual and environmental risks for anthra in drum making, public health agencies have long advised that animal hides of unknown origin or from areas of epzootic anthrax should not be used. However, imported an mal hides from West Africa, particularly goat hides, remain in demand because they are prized by drum makers for ther acoustical quality. Because anthrax outbreaks in livestock frequently occur in West Africa, hides brought into the United States might contain B. anthracis spores. The Animal and Plan Health Inspection Service (APHIS) of the U.S. Department of Agriculture has the authority to regulate importation of animal hides, mainly to prevent the introduction of foreign animal diseases of agricultural importance into the United States. However, APHIS does not mandate screening of imported hides for B. anthracis (8), and potentially contar nated hides might continue to be imported. In additi importation can bypass legal channels (3). Currently, the We Health Organization recommends the use of sporicidal tre ments to disinfect all contaminated animal hides, including, ethylene oxide fumigation, gamma irradiation, preservara in a 5% formaldehyde solution, or chemical treatment wi hydrochloric acid or salt in appropriate concentrations d durations (9,10).

Although safer practice in djembe drum making is needed to protect drum makers and others who might be exposed inadvertently, the best preventive measure is to use anima hides known to be free of anthrax spores. The use of PPE not considered a safe alternative to the use of anthrax-fre

« PreviousContinue »