Descriptive epidemiology of verotoxin-producing E. coli reported in Ontario, 1996-2005

1 April 2007

Volume 33

Number 07

Z Liu, MD (1), S Calvin, BSc, DVM, MSc (1),DMiddleton, BSc, DVM, MSc (1)*

1 Ministry of Health and Long-Term Care, Toronto, Ontario
* Corresponding author. Dean.middleton@moh.gov.on.ca

Introduction

Verotoxin-producing Escherichia coli (VTEC) can cause enteric illness among humans, which may include symptoms such as diarrhea, bloody diarrhea, and abdominal cramps. Approximately 8% of persons with VTEC progress to a more serious condition called Hemolytic Uremic Syndrome (HUS), which can result in renal failure and death⁽¹⁾. VTEC, including HUS, has been a Reportable Disease under Ontario's Health Protection and Promotion Act since 1987. E. coli subtype O157:H7 is the most well recognized VTEC. The purpose of this paper is to provide descriptive epidemiological findings related to VTEC reported in Ontario for the period 1996 to 2005.

Method

The Reportable Disease Information System (RDIS) database was accessed on 17 October, 2006, to obtain records of VTEC cases reported to the Ontario Ministry of Health and Long-Term Care (MOHLTC) for the years 1996 to 2004. Cases were included in the year corresponding to their episode date variable “Epidate”.

In 2005, the Integrated Public Health Information System (iPHIS) database replaced RDIS as the primary method of reporting diseases in Ontario. iPHIS was accessed on 18 September, 2006 to obtain records of VTEC cases reported to the MOHLTC for the year 2005. All cases whose “Episode Date” was in the year 2005 were analyzed.

The case definition for VTEC was: (1) clinically compatible signs and symptoms with (a) identification of verocytotoxin in a stool specimen or (b) isolation of one or more strains of verocytotoxigenic E. coli from stool or blood or (c) an epidemiologic link to one or more laboratory-confirmed cases or (2) HUS diagnosed by a physician and not caused by defects in serum complement, chemotherapy, immunosuppressants in organ transplants, pregnancy or oral contraceptives, or known infections other than E. coli.

For the purposes of analysis, the data from the two databases were merged when the fields could be matched (e.g., age, gender, postal code, reporting date, onset date, and responsible health unit). The analysis was done separately on RDIS and iPHIS data for variables that could not be matched (e.g., hospitalization, deaths, and travel).

A large, waterborne outbreak of E. coli O157:H7 occurred in Walkerton, Ontario in 2000. The large number of cases associated with this outbreak markedly skewed the findings in 2000. To control for the influential effect of this outbreak, the cases associated with it were removed from most of the analyses except where noted.

Table 1. Incidence rate of VTEC cases in Ontario by Health Unit, 1996-2004*

The geographic distribution of VTEC cases was displayed in Table 1 by the incidence rate by health unit. The incidence rate was calculated by dividing the average annual number of cases by the average annual population in each health unit.

The case-hospitalization rate was calculated by dividing the number of hospitalized cases by the total number of cases for a particular year.

The 2005 data pertaining to “exposure risk settings” and “transmission mode” were excluded due to the large proportion of missing data (94% and 87% missing, respectively). The 2005 hospitalization data were not included because of inconsistent definitions within RDIS and iPHIS.

“Risk settings” that accounted for < 1% of the total cases were combined into the category 'other'. Similarly, “modes of transmission” that accounted for < 0.4% of the total cases were combined into the category 'other'.

Travel-associated cases were defined as cases that were identified as having travelled outside of Ontario in the period 1 to 8 days (i.e., the incubation period for E. coli) prior to onset of illness, unless it was identified that the case obtained their illness in Ontario.

Results

There was a total number of 5,171 VTEC cases reported from 1996 to 2005. The annual average was 517 cases and the median was 397 cases. The number of cases ranged from a high of 1,712 in 2000 to a low of 275 cases in 2005. The mean and median percent of outbreak-associated cases was 19% and 10%, respectively (Table 2).

Table 2. E. coli cases by year, 1996-2005, Ontario (n = 5,171)

There were 1,270 cases linked to the Walkerton outbreak in 2000. These cases were removed from subsequent analysis, except where noted. There was also a total of 3,901 cases included in the subsequent findings. The mean and median annual number of cases were 390 and 397, respectively. The number of cases ranged from a high of 467 in 1996 to a low of 275 cases in 2005. The annual number of cases, by year, is shown in Figure 1.

Figure 1. VTEC cases by year, 1996-2005, Ontario* (n = 3,901)

*Excluding 1,270 Walkerton outbreak-associated cases in the year 2000.

Figure 2. VTEC distribution by age group and gender, 1996-2005, Ontario* (n = 3,901)

*Excluding 1,270 Walkerton outbreak-associated cases in the year 2000.

The health units with the highest rates of VTEC were in southern Ontario (Table 1). The mean and median age for the 10-year period was 27 and 18 years, respectively. The ages ranged from < 1 to 97 years. The '0 to 9' and '10 to 19' year age-groups accounted for > 50% of the cases (Figure 2). The gender distribution for the entire period was 55% female versus 45% male.

Over the 10-year period, 58% of the cases occurred in summer months from June to September (Figure 3).

Figure 3. Average VTEC cases bymonth, 1996-2004, Ontario* (n = 3,901)

*Excluding 1,270 Walkerton outbreak-associated cases in the year 2000.

The number of hospitalized cases ranged from 82 in 1996 to 51 in 2004. The case-hospitalization rates ranged from 13% to 18%. A total of 25 deaths occurred during the time-period (Table 3). VTEC was identified as either an underlying factor or a direct cause of 74% of the deaths. The remaining 26% of responses were 'missing' or 'unknown'. The mean and median of the 25 deaths was 58 and 68 years, respectively. The age range was 2 to 94 years of age. 'Home' was the most frequent “risk setting” (Table 4). 'Foodborne' transmission was identified as the most frequent “mode of transmission” (Table 5).

Table 3. Hospitalization and deaths of E. coli Cases, 1996-2005, Ontario (n = 3,901)

Table 4. Risk settings for VTEC Cases, 1996-2004, Ontario (n = 3,626)*

Table 5. Mode of transmission for VTEC cases, 1996-2004, Ontario (n = 3,626)*

Discussion

It is recognized that passive surveillance processes capture only a proportion of VTEC cases occurring in the community. According to Michel et al., VTEC cases are likely underreported by 78% to 88% in Ontario⁽²⁾. Furthermore, the potential for bias should be considered when interpreting the results presented in this report because a number of variables in the database had a large percentage of missing or unknown responses.

There appears to be a decreasing trend in the annual number of VTEC cases from 467 cases in 1996 to 275 cases in 2005. This trend does not include the large number of cases that were associated with the Walkerton outbreak in 2000. The decreasing trend in the number of cases over time is further supported by the findings from 1987 to 1995 in which the number of cases decreased from a high of 674 in 1989 to a low of 392 cases in 1993⁽³⁾.While there appears to be a decrease in the average annual number of cases since 1989, the threat of a large scale outbreak remains, as evidenced by the Walkerton outbreak.

Analysis of the number of outbreak-associated cases revealed that the median percent of cases associated with outbreaks was 10%. In 2000, of the 1,712 cases that Ontario experienced, 74% (1,270 cases) consisted of Walkerton outbreak-associated cases. Of the 454 cases in 2003, several small outbreaks occurred with outbreak- associated cases comprising 38% (172 cases) of the total cases in the year. It is likely that a number of sporadic cases are linked to a common source but are not identified as such.

The distribution of health units with the highest rates of VTEC over the time period were in southern Ontario (Table 1). These findings were consistent with the findings from a report by Michel et al. on VTEC cases for the years 1990 to 1995⁽²⁾. This paper associates VTEC incidence with cattle density, and suggests that living in an agricultural region where cattle are raised could be an important risk factor for VTEC.

The reason that there were more females (55%) than males affected over the period was not known. The age distribution was predominantly (> 50%) persons < 20 years of age. Fifty-eight percent of the cases occurred between June to September. All of these findings exhibited little variation over the years.

There were 0 to three deaths experienced annually, except in 2000 when there were a total of 10 deaths, seven of which were associated with the Walkerton outbreak.While responses revealed that VTEC was an underlying factor or a direct cause of 74% of the deaths, it is hypothesized that the true percent is likely higher.

'Home' (36%) was the most frequently reported “risk setting”, followed by 'restaurants' (11%). It should be noted that 35% of the responses were 'unknown' or missing. 'Food' (35%) was the most frequent “mode of transmission”, followed by 'person- to-person' (5%) and 'water' (3%) transmission. It should be noted that 51% of the responses were 'unknown' or missing. All of these findings exhibited little variation over the study period.

Conclusion

It appears that the trend in the annual number of VTEC cases over the period 1996 to 2005 is decreasing. The findings for the annual number of VTEC cases for the period 1987 to 1995 further support the trend of decreasing VTEC cases⁽²⁾. The highest number of cases (674) was identified in 1989 and the lowest number (275) was identified in 2005. While there appears to be a decrease in the average annual number of cases since 1989, the threat of a large scale outbreak remains, as evidenced by the Walkerton outbreak in 2000.

References

1. Heymann DL. Control of communicable diseases manual. 18^th ed., American Public Health Association, 2004:160-4.

2. Michel P, Wilson JB, Martin SWet al. Temporal and geographic distribution of reported cases of Escherichia coli O157:H7 infections in Ontario. Public Health and Epidemiology Report Ontario 2001, 12;2:164-71.

3. LeBer CA. Reported cases/outbreaks of verotoxin-producing E. coli in Ontario, 1987 to 1995. Public Health and Epidemiology Report Ontario 1996, 7;12:448-9.