METHODS

Routine testing of all community faecal specimens in Salford involves selective culture for Salmonella spp (DCA agar and XLD agar; LabM, Bury, UK), Shigella spp (DCA agar), Campylobacter spp (Campylobacter selective agar; Oxoid, Basingstoke, UK), and Escherichia coli 0157 (cefixime tellurite sorbitol MaConkey agar; Oxoid), auramine staining for cryptosporidium oocysts, and a direct wet preparation examination for ova, cysts, and parasites. Rotavirus antigen detection by latex agglutination (Rotagen; Biokit, Barcelona, Spain) is performed on samples from patients under 5 years of age. There were no changes in laboratory protocols during the two study periods.

The microbiology computer system database (TelePath CDS; TelePath Ltd, Birmingham, UK) was retrospectively searched for details of faecal examination requests from Salford residents for the period 26 April to 6 June in 1998 and in 1999. Data were analysed on a patient basis rather than on a specimen basis (that is, multiple samples from the same patients, of which there were 21 and three during the period 26 April to 6 June in 1999 and in 1998, respectively, were treated as a single request). Statistical analysis included calculation of the 95% confidence intervals for the percentage difference in positive isolates between the two time periods. Illness onset dates for cryptosporidium positive individuals were obtained from the local consultant in communicable disease control (CCDC; RM).

Some general practitioners serving Salford patients use microbiology laboratory services other than the Hope Hospital microbiology laboratory. Of all Salford residents notified to the CCDC with laboratory confirmed cryptosporidiosis between 26 April and 6 June 1999, 77% had been tested at Hope Hospital. Our data therefore refer to most, but not all, of the Salford population.

RESULTS

Figure 1 shows illness onset dates for cryptosporidium positive individuals between 26 April and 6 June 1999. Two peaks can be seen, one around the 24 April and the other around the 7 May. Onset dates were not available for six patients. Figure 2 shows the number of general practice patients from whom faeces specimens were received during seven day intervals spanning the period 5 April to 13 June, both for 1998 and 1999. The number of samples increased during the period 26 April to 16 May in 1999 compared with the same period in 1998. Patients with cryptosporidiosis submitted faeces for testing between 0 and 23 days (median, six) after the onset of symptoms.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Illness onset dates for cryptosporidium positive individuals between 26 April and 6 June 1999.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Number of community faecal specimens received between 5 April and 13 June, 1998 and in 1999 (weekly intervals).

Table 1 shows the percentage change in the number of individuals positive for enteric pathogens between 1999 and 1998 for the period 26 April to 6 June. The total number of patients tested more than doubled and the number of patients positive for cryptosporidium increased more than eightfold. There was no appreciable change in the number of patients with Campylobacter spp, salmonella, shigella, giardia, or rotavirus. Because only one isolate of shigella and one isolate of giardia were obtained over the two time periods, statistical analysis was not undertaken. Calculation of 95% confidence intervals demonstrates that only the increase in cryptosporidium was significant. Table 2 shows the analysis, with age stratification, of microbiology results on community patients from whom faeces specimens were received during the period 26 April to 6 June in 1998 and in 1999. The observed increase in cryptosporidium positive patients in the 1999 period was weighted towards the younger ages, particularly those under 10 years, whereas the increase in patients testing negative for the routinely tested enteric pathogens was more evenly distributed.

DISCUSSION

This outbreak of cryptosporidiosis was confirmed by both the absolute and proportionate (relative to the total number of patients tested for enteric pathogens) rises in the numbers of cryptosporidium positive patients, compared with the same period the previous year. Despite a twofold increase in the total number of patients tested during the outbreak period, there was no proportionate increase in the numbers of patients positive for other routinely tested enteric pathogens. The rise in the number of patients tested during the outbreak period is more than can be accounted for by all the patients testing positive for the routinely tested pathogens. These excess patients may represent patients with cryptosporidiosis (or other recognised pathogens) testing laboratory negative, patients with non-infective causes of diarrhoea, or patients with viral gastroenteritis (for example, Norwalk-like virus, adenovirus, calicivirus). However, following the enhanced surveillance request and press publicity, the rise was probably the result of a lower threshold for individuals with diarrhoea (regardless of cause) seeking medical advice and a lower threshold for requesting faeces examination by general practitioners.

“The rise in the number of patients tested during the outbreak period is more than can be accounted for by all the patients testing positive for the routinely tested pathogens”

General practitioners were not given explicit advice for obtaining a faecal specimen but were asked to sample patients with suspected infective diarrhoea. It is accepted that there is wide variation in faecal sampling between general practitioners. It was not possible to determine the number of patients presenting with symptoms of gastroenteritis to all general practitioners during the two periods. This would have required a separate study and manual retrieval of data in many practices. One possible explanation for the absence of an increase in the isolation of other pathogens is that local general practitioners are correctly sampling patients who have bacterial pathogens. However, this hypothesis is not borne out by other studies, although they did demonstrate that patients with illness caused by common bacterial pathogens, such as Salmonella sp and Campylobacter sp, are more likely to see their general practitioners, to have faeces examined, and to be reported to local and national surveillance networks than individuals with viral gastroenteritis.^4,5 In one of these studies the ratio of cases in the community to cases reported to national surveillance was calculated.⁴ This ratio was lower for bacteria (salmonella, 3.2 : 1) than for viruses (rotavirus, 35 : 1).

However, our study is not directly comparable with these because our objective was to determine whether enhanced surveillance (as a result of an outbreak) picked up additional cases of gastrointestinal infection and not to determine the background incidence of gastrointestinal infection in the community. We acknowledge, however, that our study has limitations because of the small sample size.

A coincidental community wide outbreak of diarrhoea caused by pathogens other than those routinely tested (such as Norwalk-like viruses) cannot be ruled out, but seems unlikely.