Environmental isolates

Environmental samples were examined for the possibility of cross-contamination and transmission. Of the 380 samples were taken, only 18 samples were culture-positive; ⁶ from ICU-1, none from ICU-2, 3 from ICU-3 and 9 from ICU-4.

3.4 Comments

In spite of the growing number of studies devoted to C. difficile-related diseases in the western countries, studies on C. difficile-associated diseases in the Middle East especially Kuwait are almost non-existent with a few exceptions (Rotimi et al., 2002; Jamal et al., 2002). This may be explained, in part, by the lack o f expertise (until recently), technology and facilities for culturing anaerobic pathogens. Besides, the high proportion of asymptomatic carriers of C. difficile encountered in our hospitals (Prof VO Rotimi; personal communication) makes the interpretation of a positive culture equivocal. However, in a previous study reported by Akhter et al. (Akhter et al., 1994) in nearby Riyadh, Saudi Arabia, C. difficile cytotoxin was detected in 9.5% of patients with acute gastroenteritis, suggesting that this organism might be an important enteric pathogen

in the Gulf and Middle East. To support this conjecture, is the fact that C. difficile has been isolated from the stool of symptomatic patients in Turkey (Soyletir et al., 1996) and Israel (Rudensky et al., 1993; Rivilin et al., 1998; Boaz et al., 2000).

The overall 10.3% prevalence of acquisition of C. difficile amongst patients admitted to the 4 ICUs, whose initial cultures were negative, reported in this study, is similar to the prevalence rate of 9.7% reported by Barbut et al., (Barbut et al., 1996) in various intensive care units in French hospitals but lower than the 30% reported elsewhere by Samore and colleagues in 1994 (Samore et al., 1994b), although in a relatively smaller population of patients than the number investigated in this study. The acquisition rate of C. difficile patients in our bum unit ICU was 6.5% and none of them had C. difficile-associated diseases, unlike the study by Gmbe and colleagues, who found a prevalence of 9.8% CDI among 112 critically ill burned patients in their ICU (Gmbe et al., 1987). Still and his colleagues (Still et al., 2002) evaluated 1753 patients in their bum unit over 3-year period and reported 18 cases (1%) in whom C. difficile toxin was detected. A relatively high acquisition rate of 18.2% occurred in our haematology wards of KCCC (ICU-3), an experience much higher than the 2.6% rate reported by Tabaqchali and Wilks in 1992 (Tabaqchali and Wilks, 1992) in the absence of outbreak, and 15% reported by Wroblewska et al. (2005), similar to our setting, but lower than the 39.2% reported by Heard et al. (1986) in the presence of an outbreak. However, our finding was similar to the 16.6%, reported by Delmee et al. (1987).

A finding of toxin A/B positivity in the stool was significantly higher among infected patients with diarrhoea than among asymptomatic carriers. However, there were two cases of CDI from whom non-toxigenic C. difficile strains were isolated and who had no toxins in their stools. This may be explained by the relatively lack of sensitivity of the ELISA TOX A/B kits (O’Connor et

al., 2001). This is buttressed by the isolation of a toxin A-negative/toxin B-positive strain from a symptomatic patient without evidence of neither the organism nor the toxin in Amiri hospital.

This was the only strain of such description that was isolated in this study. Anecdotal reports of previous studies have described such strains in asymptomatic children (Delmee et al., 1988;

Depitre et al., 1993). However, in recent years, toxin-variable strains, such as those producing cytotoxin but not enterotoxin, have been described in symptomatic adults and appear to be more prevalent in Japan (Kato et al., 1997), USA (Johnson et al., 2001) and UK (Brazier et al., 1999).

In addition, an outbreak involving 16 patients, due to toxin variable C. difficile, has been described in Canada (Al-Barrak et al., 1999). At the present time, from an epidemiological perspective, these toxin-variable strains do not constitute a problem in our hospitals.

Environmental contamination and carriage on the hands of the healthcare workers have been documented as playing a significant role in the transfer of C. difficile from patient to patient (Mulligan et al., 1979; Kim et al., 1981; Malamou-Ladas et al., 1983; McFarland et al., 1989).

McFarland et al. (1989) found that carriage of C. difficile on the hands of healthcare providers nearly always involved the same type as the patients strain and sometimes occurred even after patient care practices thought ordinarily to pose little risk, e.g. daily patient assessment, physical examination or charting. It is noteworthy that in this present study, we did not recover any C.

difficile from the hands of healthcare providers. This may reflect the compliance of our nursing and medical staff with hand washing hygiene which is enhanced by continuous eduction, demonstration of hand washing techniques and displaying posters at the clinical areas. Our finding is concordant with the reports of other workers in Israel (Rudensky et al., 1993), in the UK (Malamou-Ladas et al., 1983) and in the USA (Gerding et al., 1986) who did not recover C.

difficile from the hands of staff members caring for their C. difficile culture-positive patients.

The environment around some of the patients in 3 of the 4 ICUs was contaminated with C.

difficile. The frequency of such contamination correlated with the patient's clinical status. It was highest around patients with diarrhoea, an observation supported by earlier reports (Mulligan et al., 1979; Kim et al., 1981; McFarland et al., 1989). This contamination of the environment from asymptomatic carrier is of special concern because it involves patients whose carrier state usually remain undetected. Thus, these asymptomatic carriers, who are nursed without enteric precautions, constitute credible potential sources of cross-infection and perhaps outbreak of infection. The environmental prevalence of C. difficile was low in our ICUs. This may be related to the method for isolation and culturing of the organisms as we used enrichment broth. Other studies (Wilcox et al., 2000b) showed that incorporation of lysozyme (5mg/l) into selective agar significantly increased the recovery of C. difficile from the environmental samples probably due to increased germination of dormant spores. In addition, Verity and colleages found that C.

difficile was recovered significantly more frequently from swabs plated directly onto C. difficile selective media containing lysozyme than from enrichment broth (Verity et al., 2001).

None of the conventional enteric pathogens was isolated in this study and our data showed that the main cause of nosocomial diarrhoea among ICU patients in Kuwait is C. difficile. This finding is in agreement with several earlier reports from elsewhere (Fan et al., 1993; Rohner et al., 1997). It therefore stands to reason to suggest that, in the western/developed countries, stool culture for common enteric pathogens need not be done for hospitalized patients who had stayed in the hospital for more than three days unless clinically or epidemiologically indicated.

CHAPTER 4