Two independent laboratory databases, maintained by the Botulism Reference Service at Health Canada, Ottawa, Ontario, and the British Columbia Public Health Reference Microbiology Laboratory, Vancouver, British Columbia, were examined for cases of foodborne botulism confirmed during 1985–2005. Information regarding the number of clinical cases, age and sex of patients, implicated food, case history, laboratory analysis, date, and location of the outbreak were extracted.

To ensure consistency in data recording and analysis throughout the study period, we used the 2009 national case definition for confirmed cases of foodborne botulism (13). A case of foodborne botulism is confirmed on the basis of clinical evidence and a positive laboratory specimen (i.e., detection of botulinum neurotoxin in serum, feces, gastric aspirate, or food or isolation of C. botulinum from feces or gastric aspirate). In addition to cases with laboratory confirmation, persons with botulism who were epidemiologically linked to a laboratory-confirmed case were considered to have confirmed, and thus reportable, cases. Detection of botulinum neurotoxin and isolation of viable C. botulinum from foods and clinical specimens were performed according to Health Canada method MFHPB-16 (14). Data on length of hospitalization were retrieved from the Hospital Morbidity Database (HMDB) of the Canadian Institute for Health Information (www.cihi.ca); all records that listed botulism in the first 3 suspected diagnostic codes were extracted for the years 1994–2005 (all years currently available). These records were then matched to the laboratory records by age, sex, date of admission, date of sample, and province of residence. Only cases with laboratory confirmation were included in the analysis of the HMDB data.

To provide a descriptive epidemiology of outbreaks of foodborne botulism in Canada during 1985–2005, we analyzed data from laboratory-confirmed outbreaks (in Canada, because of the urgency of the disease, 1 case of botulism constitutes an outbreak) in terms of case numbers and rates, demographic characteristics of patients, length of hospitalization, food types, outbreak settings, serotype, and circumstances of occurrence. The rates of disease were calculated by using census data from Statistics Canada and other sources (15,16). Statistical analysis of the HMDB data was done in SPSS version 19 (IBM, Armonk, NY, USA). The Mann-Whitney test was done to compare central tendency of the data; relationships were considered significant at p<0.05.

Of 91 laboratory-confirmed outbreaks (a total of 205 cases), C. botulinum type E was implicated in 75 (86.2%), followed by types A (7, 8.1%) and B (5, 5.7%). Median patient age was 45 years (range 3–80 years); 93 (48.4%) were male. Three cases in the Nunavik region were recorded as repeated episodes of type E botulism, with the second episodes occurring 10–20 years after the initial intoxication. Overall, the number of outbreaks of foodborne botulism did not decrease during the study period, with a mean of 4.3 outbreaks/year. Outbreaks involved 1–37 cases, and 78% of outbreaks involved 1 or 2 cases. Mean annual incidence was 0.03 cases/100,000 population. The annual number of cases was marked by peaks associated with large outbreaks (Figure 1). The number of cases appeared to decrease during the last 5 years of the study period; lower numbers of outbreaks with multiple cases were reported (Table 1).

A total of 11 deaths were reported (case-fatality rate 5.4%). Within the 21-year study period, the case-fatality rate declined from 8.8% in 1985–1989 to 2.4% in 1990–1994 and then remained below 3.6% during the remaining 5-year intervals (Table 1). However, because mortality data were derived from the laboratory databases that do not officially keep mortality records, these data should be interpreted with caution.

Three of the reported cases occurred in pregnant women. Case history revealed that 2 of these women were exposed to type E botulinum neurotoxin and 1 to type B botulinum neurotoxin. No apparent effect on pregnancy was reported.

Severity of illness was assessed by using HMDB data for 1994–2005. The length of hospitalization was obtained for 70 patients with type E botulism (Table 2); the low number of type A and type B botulism cases recorded in HMDB did not enable valid data reporting. Although HMDB does not list deaths, the authors are aware of 1 fatal type E case from among these patients. Of botulism type E case-patients with known length of hospitalization, 37 (52.9%) were female and 33 male, which was not significantly different (p = 0.633). The median length of hospitalization was 7 days for female (mean 14.8 days) and male (mean 8.7 days) patients; these findings were not significantly different (p>0.05). The mean for the female group was influenced by 1 fatal case in which the patient was hospitalized for 225 days. When type E case-patients were subdivided based on age, age groups differed in length of hospitalization, but a trend was not apparent (Table 2).

Hospital procedure codes were extracted from HMDB for 54 patients with type E botulism. These 54 cases represent only a subset of the total type E cases, and all were laboratory confirmed and matched with laboratory records. Intubation alone was recorded for 16 (30%) patients, antitoxin alone for 21 (39%) patients, and both for 5 (9%) patients; 12 (22%) patients received neither antitoxin nor ventilation.

For the group comprising all patients that received antitoxin (n = 26), the median length of hospital stay was 5 days (mean 15.1 days, SD 43.2), significantly (p = 0.003) shorter than for the group that did not receive antitoxin (n = 28), which had a median length of hospital stay of 11 days (mean 13.3 days, SD 8.32). For the group comprising all patients that were intubated (n = 21), the median length of hospital stay was 13 days (mean 24.5 days, SD 46.6), significantly (p<0.001) longer than for the group that did not undergo intubation (n = 33), which had a median length of hospital stay of 5 days (mean 7.58 days, SD 6.24). Intubation is likely a marker of severity of symptoms and is not thought to be the cause of increased length of hospitalization. No determination for the elapsed time period between symptom onset and the administration of antitoxin or ventilation was available.

Most (85, 93.4%) confirmed botulism outbreaks originated in Quebec, British Columbia, Nunavut, and the Northwest Territories. Reported cases predominantly occurred in Quebec and British Columbia, with 89 and 71 cases, respectively; these cases accounted for 78% of the total number of cases (Table 3). Of 51 outbreaks in Quebec, 45 (88%) occurred in the Nunavik region of northern Quebec; 91% of these were clustered in 3 villages of southern Ungava Bay (Kuujjuaq, Kangiqsualujjuaq, and Tasiujaq), which are inhabited by an Inuit population of 2,587 (Figure 2). In British Columbia, 9 (64%) of 14 outbreaks (20 cases) occurred in First Nations communities located along the Pacific Coast. The high number of cases recorded in the province was primarily because of 2 large restaurant-associated outbreaks in Vancouver that affected 37 persons (17) and 11 persons (18). In Ontario, 3 outbreaks were recorded, with 1 affecting 3 persons and causing 1 death. No botulism cases were reported in the provinces of Alberta, Saskatchewan, Manitoba, New Brunswick, Nova Scotia, and Prince Edward Island during the study period.

Other than a single outbreak in Nunavut caused by consumption of fermented fish heads, all type E outbreaks with known food sources in northern Canada were linked to marine mammals, including beluga whales, seals, and walruses (Table 4). Of the 41 botulism outbreaks in Nunavik, 32 (78%) were caused by food products of seal origin; aged meat (igunaq) and aged flippers (utjaq) were most frequently implicated foods in Nunavik, accounting for 15 (47%) seal-related outbreaks. Beluga whale was involved in 16 (80%) of 20 outbreaks in Nunavut and the Northwest Territories. Eleven of these outbreaks were caused by muktuk, which is aged pieces of skin (with fat and meat) of the beluga whale.

In the First Nations communities of British Columbia, 9 outbreaks of type E botulism were associated with the consumption of aged fish products. Fermented salmon eggs (stink eggs) have been the primary source of botulism, accounting for 8 of 9 (89%) outbreaks in British Columbia coastal communities. Because no carbohydrates are available in these eggs for a fermentative transformation into organic acids, the aging process involves putrefaction rather than fermentation.

The 1985 type B outbreak involving garlic-in-oil at a Vancouver restaurant affected 37 persons; 24 were hospitalized, and none died (17) (Table 4). The outbreak occurred in 2 clusters, with 11 cases in the first cluster during July 28–August 2, 1985, and 26 cases in the second cluster during August 29–September 5, 1985. In this incident, garlic-in-oil used in sandwiches was kept at room temperature for 8 months. In a 1987 botulism outbreak at a hotel restaurant in Vancouver, 11 persons became ill after consuming bottled chanterelle mushrooms contaminated with type A botulinum neurotoxin (18). The mushrooms were grown on Vancouver Island and bottled at the hotel using an in-house heating process. One jar recovered during the investigation was found to contain 4,000 minimal lethal dose/mL of type A neurotoxin in the liquid phase.

A third outbreak associated with a restaurant occurred in Ontario in 2002 and involved a single case of type A botulism linked to a baked potato (19). This patient began to show gastro-intestinal and neurologic symptoms 12 hours after eating the potato, the remains of which had been discarded and was unavailable for testing. The patient was hospitalized for >6 months and released with long-term sequelae; the patient continued to experience weakness and vision problems 4 years after the incident.

Two unrelated incidents involving commercial ready-to-eat meat products were reported in Quebec in 1995 and 2001 (Table 4). In both instances, products that were intended to be refrigerated were stored at room temperature, which enabled growth of C. botulinum and toxin production. In the 1995 incident, C. botulinum type B was isolated from fecal samples of 2 persons and from a commercial country-style pâté. In the 2001 incident, C. botulinum type A was isolated from a cooked boneless pork product and from fecal and gastric liquid samples from a patient. The cooked boneless pork product was recalled as a precautionary measure.

Several incidents during the 21-year study period involved non-Native home-prepared foods (Table 4). Home-canned mushrooms were linked to a type B botulism incident that affected 1 person in Quebec. Home-canned asparagus was involved in 1 type A botulism incident in Ontario; 3 persons required intensive care, and 1 died. In this incident, 8 of 10 glass jars of whole-stalk asparagus showed evidence of odorous gas production during the investigation. One additional type A case was linked to the consumption of home-canned vegetable and beef soup.

A variety of noncanned home-prepared foods, including bean soup, spaghetti sauce, sausages, and fish, were also implicated in botulism outbreaks, although most of the incidents could not be laboratory confirmed because no remaining food was available. Home-smoked fish was responsible for 3 cases of type E botulism, with 1 fatality, in the Yukon in 1995. Diagnosis and treatment with antitoxin was delayed until 3 days after the onset of symptoms. Another incident of fish consumption affected 1 tourist traveling in Quebec in 2001, but the mode of preparation was not recorded. Home-prepared spaghetti sauce made with home-canned sausage caused botulism in 1 person from Quebec in 2000. Another type A case was associated with the consumption of homemade bean soup, but its preparation was not documented.

Of 91 total confirmed outbreaks in Canada during the study period, 77 (84.6%) were confirmed by detection of botulinum neurotoxin only or by detection of neurotoxin and isolation of C. botulinum from laboratory specimens. The remaining 14 outbreaks were confirmed by detection of C. botulinum in gastric contents and/or fecal samples only, without detection of neurotoxin in clinical samples. Food was recovered in 69 (75.8%) of the outbreak investigations. The detection rate for botulinum neurotoxin in foods (78.3%) was found to be higher than for serum (34.9%), feces (34.5%), or gastric contents (11.0%) (Table 5). C. botulinum was also more frequently detected in foods (83.3%), followed by gastric contents (51.3%) and feces (40.3%). Two botulism cases were confirmed by the detection of C. botulinum in liver tissue during autopsies.