Shiga toxin-producing Escherichia coli, Idaho.

Abstract

To the Editor: Data collected from expanded surveillance study suggest that more than half of Idaho Shiga toxin–producing Escherichia coli (STEC) illnesses are caused by non-O157 serotypes. Using data from a regional medical center whose stool culture protocol included Shiga toxin testing, we predicted Idaho’s STEC incidence to be significantly higher if non-O157 STEC were routinely detected by immunoassay. Recent findings suggest that the prediction was accurate in an expanded surveillance area. Several studies have shown an increased incidence of non-O157 STEC infections in the United States. For example, a community hospital in Virginia detected non-O157 serotypes in 31% of patients with STEC from 1995–2002 (1). A 1998 Nebraska study that analyzed 30,000 diarrheal stool samples found that non-O157 and O157:H7 STEC were equally prevalent (2). Additionally, findings from a Connecticut study of laboratory-confirmed cases (3), STEC surveillance results from Montana (4), and a recent study from Michigan (5) indicate that non-O157 serotypes comprise a substantial percentage of STEC cases. In other countries, nonculture-based methods are routinely used for STEC detection (6). However, E. coli O157:H7 culture methods remain the focus in the United Kingdom, Canada, and the United States (6). Reliance on culture methods can result in misleading interpretations of STEC prevalence. For example, 93% of STEC infections in Canada are reported to be E. coli O157:H7, yet a Manitoba 1992 study showed that when toxin assays were used, 35% of the recovered STEC isolates were non-O157 serotypes (6). Analysis of reported non-O157 STEC cases in Idaho showed a similar trend. From 2002–2004, 66% of Idaho’s non-O157 cases originated in Health District 7, where >70% of stool cultures are screened by enzyme immunoassay (EIA) for Shiga toxin (Premier EHEC, Meridian Bioscience, Cincinnati, OH, USA). This rate was disproportionately higher than that of the remaining 6 health districts, which primarily use culture methods to screen for E. coli O157:H7. We hypothesized that this disproportion was due to differences in stool culture protocol. To test this premise, we conducted enhanced surveillance for 16 months in a “low” STEC incidence area, Health District 5. A total of 2,065 stools submitted for culture were screened for Shiga toxin by EIA. With this approach, reported non-O157 STEC incidence rose from <1 case/year/100,000 population to 11 cases/year/100,000 population. Additionally, 56% of recovered STEC isolates were non-O157 serotypes, mirroring the proportion of non-O157 detected in District 7. Notably, this appears to be the endemic rate for District 5 because no non-O157 STEC outbreaks or matching pulsed-field gel electrophoresis patterns were detected during the surveillance period. Although our study captured only a portion of stool cultures in Idaho, our findings demonstrated an increased prevalence of non-O157 STEC in the region when nonculture methods were used. Two barriers cited for not routinely screening diarrheal stools for Shiga toxin are cost and perception of low non-O157 STEC incidence. While toxin testing is more expensive than culture testing, the potential effects of misdiagnosis may outweigh cost concerns. A study estimating the financial repercussions of E. coli O157 infections in the United States suggested that annual cost associated with this pathogen is $405 million, with the cost per case varying from $26 for those who do not seek medical care to $6.2 million for a patient with fatal hemolytic uremic syndrome (HUS) (7). Non-O157 STEC infections have been an important cause of HUS in many countries. For example, a 3-year prospective study in Germany and Austria reported that non-O157 serotypes comprised 90 (43%) of 207 STEC isolates from stools of 394 pediatric patients with HUS (8). Further, a 6-year Danish study of 343 registered STEC patients found that 76% of STEC and 48% of HUS cases were attributable to non-O157 serotypes (9). In the United States, continued reliance on O157 STEC culturing hinders our ability to determine the financial effects and the proportion of HUS cases attributable to non-O157 STEC. Some evidence suggests that the testing focus may be changing in the United States. We used US Census Bureau population statistics to translate reported O157:H7 and non-O157 STEC cases for each state into incidence data. Despite widespread variation in STEC testing and incidence among states, there has been a significant statistical decline in the proportion of E. coli O157:H7 among total STEC cases every year since 2001 (Figure; p<0.001) (10). Consistent with this trend, the incidence of non-O157 STEC in the United States has increased (10). This may indicate that more laboratories are adopting Shiga toxin testing protocols, as we are advocating in Idaho. Our findings suggest that perceptions of low non-O157 STEC incidence in Idaho are probably artifactual and due to overemphasis on culture methods for O157 STEC. Our ongoing EIA-based surveillance highlights the need for continued investigation of the epidemiology of non-O157 STEC disease. We conclude that O157 STEC culturing has limited usefulness in areas like the Idaho health districts investigated, where non-O157 serotypes accounted for 55% of STEC illnesses. The true involvement of non-O157 in STEC disease will remain obscured as long as screening methods focus on traditional culture methods. Figure Shiga toxin–producing Escherichia coli (STEC) incidence trends, United States, 2002–2005.