Wound Botulism Associated With Black Tar Heroin Among Injecting Drug Users

Context.— Wound botulism (WB) is a potentially lethal, descending, flaccid, paralysis that results when spores of Clostridium botulinum germinate in a wound and elaborate neurotoxin. Since 1988, California has experienced a dramatic increase in WB associated with injecting "black tar" heroin (BTH), a dark, tarry form of the drug.

Objective.— To identify risk factors for WB among injecting drug users (IDUs).

Design.— Case-control study based on data from in-person and telephone interviews.

Participants.— Case patients (n=26) were IDUs who developed WB from January 1994 through February 1996. Controls (n=110) were IDUs newly enrolled in methadone detoxification programs in 4 counties.

Main Outcome Measures.— Factors associated with the development of WB.

Results.— Among the 26 patients, the median age was 41.5 years, 15 (58%) were women, 14 (54%) were non-Hispanic white, 11 (42%) were Hispanic, and none were positive for the human immunodeficiency virus. Nearly all participants (96% of patients and 97% of controls) injected BTH, and the mean cumulative dose of BTH used per month was similar for patients and controls (27 g and 31 g, respectively; P=.6). Patients were more likely than controls to inject drugs subcutaneously or intramuscularly (92% vs 44%, P<.001) and used this route of drug administration more times per month (mean, 67 vs 24, P<.001), with a greater cumulative monthly dose of BTH (22.3 g vs 6.3 g, P<.001). A dose-response relationship was observed between the monthly cumulative dose of BTH injected subcutaneously or intramuscularly and the development of WB (χ² for linear trend, 26.5; P<.001). In the final regression model, subcutaneous or intramuscular injection of BTH was the only behavior associated with WB among IDUs (odds ratio, 13.7; 95% confidence interval, 3.0-63.0). The risk for development of WB was not affected by cleaning the skin, cleaning injection paraphernalia, or sharing needles.

Conclusions.— Injection of BTH intramuscularly or subcutaneously is the primary risk factor for the development of WB. Physicians in the western United States, where BTH is widely used, should be aware of the potential for WB to occur among IDUs.

AN ESTIMATED 80000 Americans, including 18000 Californians, inject illicit drugs.¹ Compared with the general population, injecting drug users (IDUs) are at increased risk for disease and death^2-6 and use a disproportionate amount of medical resources.^7,8 Soft tissue infections are a particular problem.^9-11

In 1994, the California Department of Health Services (CDHS) noted an increasing number of cases of wound botulism (WB), an unusual soft tissue infection. Botulism is best known as a potentially lethal paralytic disease that results from ingestion of preserved food containing preformed botulinum toxin. Wound botulism is a clinically similar syndrome of flaccid, symmetric, descending paralysis that results when spores of Clostridium botulinum, an obligate anaerobe, are inoculated into a wound or other devitalized tissue.¹² After gaining access to this relatively anaerobic environment, the spores germinate and elaborate the most potent toxin known.¹³ Historically, the implicated wound has been a crush injury or other gross trauma to an extremity.¹⁴ The first reported case of WB associated with injecting drug use occurred in 1982 in New York City.¹⁵

All forms of botulism are reportable diseases in California. Since 1988, the year of California's first reported WB case associated with injecting drug use, the number of WB cases has increased dramatically, totaling 49 from 1988 through 1995; 46 of these cases occurred in IDUs, nearly all of whom injected primarily heroin. In each case, the same type of heroin was used: "black tar" heroin (BTH), a black, gummy form of the drug that usually is synthesized in makeshift factories adjacent to opium poppy fields in several Mexican states.¹⁶

We performed a case-control study to compare IDUs who developed WB with other heroin users to identify risk factors associated with the disease.

Case patients (patients) were IDUs who developed laboratory-confirmed WB from January 1, 1994, through March 1, 1996. All CDHS case records since 1994 were reviewed, and current telephone numbers and addresses of patients were obtained, generally through hospital records or by contact tracing. Telephone or in-person interviews were conducted directly with each patient. Patients unable to speak because of ongoing mechanical ventilation provided written responses to in-person interviews conducted by nursing staff. Patients without access to a telephone were interviewed in person by staff of the Local Assistance Branch, CDHS, Sacramento, Calif.

Controls were persons newly enrolled in 1 of 4 methadone detoxification programs in geographically and ethnically distinct cities in California (Oakland, Los Angeles, San Jose, and Stockton) during March through May 1996. Recruitment strategies for controls varied by clinic. At the Los Angeles clinic, 25 consecutive patients were enrolled in the study by the clinic intake supervisor. At the Oakland, Stockton, and San Jose clinics, study investigators visited each of these 3 on 3 nonconsecutive mornings. Each detoxification program patient who received methadone was asked to participate in the study and was offered a meal voucher to encourage participation. Before interviews, educational flyers about the study and about WB associated with injecting drug use (ie, "shooter's botulism") were distributed at the clinics; these flyers warned users of the local WB epidemic among IDUs and explained warning symptoms but did not discuss hypotheses about the causes of shooter's botulism. After each interview, participants were given verbal and written information about shooter's botulism.

Because our selection of methadone clinic study sites was not random and to examine whether our control group was representative of California methadone clinic attendees, we also compared baseline patient characteristics of our control group with those of a reference group. The reference group consisted of all enrollees in methadone detoxification or maintenance programs during 1994 or 1995 from the 17 California counties that have reported WB among IDUs.

Patients were questioned about baseline personal characteristics and 34 drug purchasing, storing, and using practices in the month before developing WB; controls were asked identical questions about drug-related practices in the month before starting detoxification, including frequency and quantity of all drugs injected; frequency and method of cleaning needles and syringes; frequency of needle sharing; source of water or other solvent used for dissolving heroin; type of apparatus used for heating the heroin-water mixture; whether cotton balls or cigarette filters (eg, "cottons"), through which the heroin mixture is drawn into the syringe, were stored and reused; frequency and type of skin cleansing before injection; which body sites were used for injection; and injection technique (intravenous vs intramuscular vs subcutaneous). Participants were also asked about their recent medical history and how frequently they developed soft tissue abscesses. Each participant estimated the quantity of drugs he or she used by stated drug weight at point of purchase. When this was not possible, quantities were estimated by dollar amounts, which were converted to weights using the costs most frequently cited by study participants: $80 per gram of heroin and $100 per gram of cocaine.

Bivariate analyses were performed by using the Fisher exact test or the Mantel-Haenszel χ² test (with the Yates correction) for discrete variables and the Student t test or the Wilcoxon 2-sample test for continuous variables. Since 34 practices and characteristics were assessed in the bivariate analyses, the Bonferroni correction was used to provide a stringent test of significance. Therefore, bivariate associations were judged significant only if P<.002 (ie, 0.05/34). Dose-response relationships were assessed using the χ² test for linear trend (Epi Info 6.02, Centers for Disease Control and Prevention, Atlanta, Ga). Multivariable analyses were performed by applying backwards-elimination logistic regression to all variables associated with a significance level of P<.2 in the bivariate analysis; likelihood ratio χ² statistics were compared to assess the goodness of fit of increasingly parsimonious multivariable models.¹⁷ Because controls were clustered by county, robust (Huber/White/sandwich) estimators of variance were used to calculate SEs in the multivariable analyses (Stata 5.0, Stata Corp, College Station, Tex). Analyses involving needle-exchange programs were restricted to patients and controls from Alameda, Los Angeles, San Joaquin, and Santa Clara counties. All comparisons were 2-tailed.

A total of 35 cases of laboratory-confirmed WB associated with injection drug use occurred in California during the 26-month study period. Of these 35 patients, all but 2 required lengthy hospitalization; all but 5 required mechanical ventilation. Of the 34 cases in which botulinum toxin typing was performed, 30 (88%) were caused by botulinum toxin type A and 4 (12%) by type B. The case participation rate was 74%; 5 (45%) of 11 patients diagnosed in 1994 and 21 (87.5%) of 24 patients diagnosed after January 1, 1995, were interviewed. Seven patients could not be located despite repeated attempts, and 2 refused to participate. Control participation rates varied by study site between 50% and 80% but were not precisely determined because not all potential controls could be enumerated at all sites.

No blacks and no persons infected with the human immunodeficiency virus (HIV) developed WB during the study period; the HIV status of controls was not documented. Patients were less likely than controls and the reference group to be non-Hispanic black and more likely to be non-Hispanic white and to be female, although these demographic trends did not reach statistical significance. Baseline characteristics were otherwise similar between groups (Table 1).

Of 34 drug purchasing, storing, and using practices analyzed, the behavior most strongly associated with the development of WB was injecting BTH subcutaneously or intramuscularly (skin-popping) rather than intravenously (Table 2). A total of 33 of 35 patients reported this route of BTH administration at least occasionally. One of the 2 exceptions was an intravenous amphetamine user who insisted that she had never skin-popped, never used BTH, and never shared paraphernalia with BTH users. The other exception was a patient who reported injecting BTH but only intravenously.

The total quantity of BTH used monthly by patients and controls was similar, but the amount that was injected subcutaneously or intramuscularly was greater among patients than among controls (mean, 22.3 vs 6.3 grams; P<.001). A dose-response relationship was observed between the monthly dosage of BTH injected subcutaneously or intramuscularly by quartile and the risk of developing WB (χ² for linear trend=26.5; P<.001). This dose-response relationship remained when analysis was limited to persons who reported injecting BTH subcutaneously or intramuscularly (χ² for linear trend=4.1; P=.04).

Patients reported having more abscesses that received medical treatment in the previous year than controls (Table 2). Although cleaning the skin before injection may have protected against developing soft tissue abscesses (P=.07, data not shown), cleaning the skin before injection or cleaning syringes between injections did not protect against developing WB. Using needle-exchange programs did not protect against WB, and sharing injection paraphernalia was not associated with the disease.

In the final (parsimonious) multivariable model, injecting BTH subcutaneously or intramuscularly (odds ratio [OR], 13.7; 95% confidence interval [CI], 3.0-63.0; P=.001), sex, and race were the only factors associated with WB. Hispanic and non-Hispanic whites had an increased risk of WB (OR, undefined; P<.001) and women had a nonsignificant increased risk of WB (OR, 2.2; 95% CI, 0.8-5.9; P=.13). To better quantify the relationship between BTH dose and risk for illness, race- and sex-adjusted ORs were calculated for quartiles of cumulative BTH dose injected subcutaneously or intramuscularly per month. When compared with IDUs who denied injecting BTH subcutaneously or intramuscularly, the odds of developing WB were 6-fold higher among "occasional" subcutaneous or intramuscular injection users (whose dosage was in the lowest quartile, $20-$480 of BTH per month) and 25-fold higher among "heavy" users (whose dosage of drug injected subcutaneously or intramuscularly was in the highest quartile, $2000-$6300 of BTH per month) (Figure 1).

From 1951 through 1995, 68 cases of WB were reported to CDHS. An average of 0.49 WB cases per year were reported from 1951 through 1987; 2.25 cases per year were reported in 1988 through 1991, 3 cases in 1992, 4 in 1993, 11 in 1994, and 23 in 1995.¹⁸ From 1988 through 1995, only 2 WB cases among IDUs were reported from outside California, and they occurred in Arizona (Foodborne and Diarrheal Disease Branch, Centers for Disease Control and Prevention, unpublished data, 1996).

This study confirms that the ongoing epidemic of WB in California is strongly associated with subcutaneous or intramuscular injection (skin-popping) of BTH and provides evidence to suggest that C botulinum contamination of BTH is not the result of specific drug storage or other using behaviors among IDUs. Although intravenous injection of heroin provides a stronger initial "high," "skin popping" is favored by users reluctant to inject intravenously, who desire to avoid telltale "track marks," or for whom venous access is difficult because of obesity or the scarring of veins from repeated use. Black tar heroin was the only drug that patients with WB reported skin-popping.

Black tar heroin was introduced to US drug users in the 1970s and slowly gained market share from traditional white heroin because Central and South American suppliers were able to develop dominant distribution networks and because initially BTH was cheaper and more potent (up to 50% by weight diacetylmorphine). Since the late 1980s, BTH has become the predominant form of heroin in the United States west of the Mississippi River. In 1993, 20 of 21 samples of California heroin purchased by undercover agents of the US Drug Enforcement Agency were BTH.¹⁹ At present, heroin is highly impure and contains contaminants (by-products of the manufacturing process), adulterants (chemicals such as methamphetamine, strychnine, or xylocaine), and diluents (inert materials, such as dextrose, that are used to "cut" heroin, providing bulk and weight and increasing the distributors' profit margin). The tarry color and consistency of BTH has led to the use of unusual diluents, including ground paper fiber soaked in black shoe polish and, according to anecdotal reports, dirt.

Our study was not designed to determine at which step in heroin production and distribution contamination was most likely to occur. However, there are several reasons to suspect that contamination occurs during "cutting." First, the last step in the conversion of opium to BTH involves boiling the product with a strong acid at 150°C for several hours, which should destroy even heat-resistant C botulinum spores. Therefore, contamination before or during this step is unlikely to result in a contaminated final product. Second, during detailed interviews, all 26 patients denied adding other substances or solvents or using unusual water sources when preparing drugs for injection (2 controls reported occasionally using beer or wine as a solvent). Therefore, inadvertent contamination of heroin by individual users seems unlikely to be the source of C botulinum spores. Third, frequency of skin cleansing before injection and the type of cleanser used were not associated with WB, and the quantity of BTH used was a more important factor than the frequency of use. Therefore, skin contamination with C botulinum appears unlikely to be a major source of this epidemic. Fourth, the frequency of using new paraphernalia, the frequency and method of cleaning old paraphernalia, and the sharing of paraphernalia were not associated with WB. Therefore, spread of WB via fomites or blood is unlikely to be a factor in this epidemic. For these reasons, we suspect that BTH is most likely being contaminated when diluted (eg, possibly with soil) after manufacture or during distribution.

Fundamental questions about this outbreak remain unanswered and probably reflect incomplete understanding of heroin distribution. For example, the reason that the location of cases (Figure 2) was spread throughout California yet essentially spared the rest of the western United States (where BTH is also distributed) is unexplained, although this pattern is consistent with contamination of BTH during in-state distribution. The lack of cases of WB occurring among black IDUs is also unclear. Drug enforcement officials hypothesize that distrust between black IDUs and suppliers of BTH manufactured by smaller, less experienced producers and distributors (which might be more highly contaminated) has minimized the use of this type of BTH among black IDUs.

Reported cases of WB may represent only a small fraction of this epidemic. Nearly all patients diagnosed as having botulism in California are described, because botulinum antitoxin is available to California physicians only through CDHS. However, botulism is a rapidly progressive disease and persons with limited access to care or who delay seeking health care may be dying outside the hospital. In these circumstances, the diagnosis of WB could be missed. For example, if postmortem examination revealed detectable serum opiate levels, an IDU might be presumed to have died from an overdose. In addition, there have been several instances in which diagnosis of WB has been delayed despite consultation with neurologists and infectious disease specialists; a subset of persons with WB may have been diagnosed as having another neuromuscular disorder.¹⁹

Unlike the more widely publicized infectious complications of drug injection eg, HIV infection and viral hepatitis B, WB is not contagious. In our study, risk of disease was not associated with markers of exposure to other IDUs (eg, frequency of sharing needles). Although 3 clusters of WB involving pairs of "shooting partners" (persons that use drugs together) have been reported, in all 3 episodes the partners had used the same heroin (S.B.W., unpublished data, 1996).

The CDHS has been unable to procure BTH samples large enough to adequately test for C botulinum. Furthermore, the samples obtained have not been closely linked to BTH samples thought to have caused illness. Accordingly, our suspicion that BTH contains botulism spores remains unproven. However, we have cultured C botulinum from the internal surface of a syringe used by a patient with WB, and we have also cultured related Clostridia species from BTH samples belonging to other patients with WB.¹⁹

We do not know the source of C botulinum spores contaminating BTH. However, our findings suggest that BTH is contaminated before sale to IDUs and that simple measures, such as cleaning the skin before injection and cleaning syringes, are unlikely to prevent WB.

The CDHS has taken several steps to make drug users, public health officials, and physicians aware of this growing problem. In October 1995, informational packets containing both technical and lay fact sheets were sent to every public health jurisdiction in California for distribution to emergency departments, needle-exchange programs, and methadone clinics. A report of the outbreak has been published.¹⁶ Education of IDUs and clinic staff was also provided to the methadone clinics participating in this study.

Subcutaneous or intramuscular injection of BTH is the primary risk factor for shooter's botulism. In addition to counseling IDUs to stop using BTH or, at least, to minimize the amount of BTH that is injected, additional efforts are needed to increase awareness of WB among IDUs and health care workers who serve them and to increase access of heroin users to methadone detoxification and maintenance programs. Physicians in the western United States should be alerted to the potential for WB occuring among IDUs.