Hong Kong Journal of Emergency Medicine. HT Wong, TL Que, PL Ho

Transcription

1 Hong Kong Journal of Emergency Medicine What have we learnt from bacterial stool culture results? A retrospective study of hospitalised gastroenteritis cases in a regional hospital in Hong Kong HT Wong, TL Que, PL Ho Background: Stool culture is one of the common investigations done for patients with gastroenteritis, and fluoroquinolones have been used frequently for suspected bacterial gastroenteritis. Objective: To study the yield of bacterial pathogens in stool culture in a local regional hospital for in-patients with gastroenteritis and the individual pathogens identified. Also, the value of stool culture and the prescription of fluoroquinolones as empirical antibiotics were reviewed. Methods: This was a retrospective study. All inpatients with the principal diagnosis of "gastroenteritis" in the year 2007 were reviewed. We excluded pregnant patients and patients under 18 years of age. Patients were divided into two age groups and data were analysed including demographics, clinical findings, admission time and culture results. Antibiotics prescription behaviour was also analysed. Results: A total of 837 adult patients fulfilled the criteria. Among them, 562 cases had their stool saved and sent for bacterial culture. Eighty-nine cases were found to have their stool cultures positive for bacteria. The identified pathogens were namely Vibrio parahaemolyticus (38.2%), Salmonella species (34.8%), Campylobacter species (11.2%), Plesiomonas species (10.1%) and Aeromonas species (3.4%) respectively. The yield of positive stool culture was 15.8%. Conclusion: Vibrio parahaemolyticus and Salmonella species were still the most common pathogens for bacterial gastroenteritis among hospitalized patients in Hong Kong, especially in the young and middle aged. Their occurrence followed a seasonal pattern. Physicians had a tendency to prescribe fluoroquinolones for cases with fever and travel history. Also, the problem of antibiotic resistance did exist. It is important for practitioners to follow local guidelines before ordering microbiological investigations on stool and prescribing empirical antibiotic treatment. (Hong Kong j.emerg.med. 2010;17:27-33) Correspondence to: Wong Ho Tung, MBBS, MRCSEd Tuen Mun Hospital, Accident and Emergency Department, Tsing Chung Koon Road, Tuen Mun, N.T., Hong Kong ho_tung2001@hotmail.com Tuen Mun Hospital, Department of Clinical Pathology, Tsing Chung Koon Road, Tuen Mun, N.T., Hong Kong Que Tak Lun, FRCPath, FRCPA, FHKAM(Pathology) The University of Hong Kong, Department of Microbiology and Centre of Infection, Pokfulam Road, Hong Kong Ho Pak Leung, MD, FACP, FRCPath

2 28 Hong Kong j. emerg. med. Vol. 17(1) Jan % 34.8% 11.2% 10.1% 3.4% 15.8% Keywords: Anti-bacterial agents, fluoroquinolones, diarrhea, drug resistance Introduction Most acute diarrhoeal illnesses are self-limited or viral in origin, and nearly half of them last for less than one day. 1 However, the associated symptoms such as abdominal cramps, vomiting and even dehydration may impede an individual from performing usual daily activities. It is estimated that acute diarrhoea itself accounts for more than two million deaths annually, despite worldwide reductions in mortality. 2 In the United States, an estimated 211 to 375 million episodes of acute diarrhoea occur each year. 3 Physicians would commonly order stool for culture and sensitivity tests. Literature reviews suggest that the prevalence of getting a positive stool culture in adult patients have been highly variable, from 2% to 51%. 4-7 One may doubt the effectiveness of performing a stool culture to guide treatment. Moreover, it is not surprising that the use of fluoroquinolones has become a commonplace amongst gastroenteritis patients, as it has been shown to reduce the duration of diarrhoea in patients with both culture-positive and culture-negative gastroenteritis. 4 However, like other antibiotics, drug resistance problems can arise. We aimed to study the positive yield of stool cultures in a local regional hospital, and the prevalence of specific pathogens found in the positive stool cultures. The value of stool culture and the practice of giving empirical fluoroquinolones would also be assessed. Materials and methods This was a retrospective study. All adult patients with age equal to or more than 18 years who were hospitalised in the Tuen Mun Hospital with gastroenteritis symptoms during 1st January to 31st December 2007 were included. Eligible patients were identified and retrieved from the computerised Clinical Data Analysis Retrieving System, which was accessible in all hospitals under the Hospital Authority. "Gastroenteritis" as the principal diagnosis was used as the main search criteria. Thus, all ICD codes with the term "gastroenteritis" were retrieved. They included patients who were admitted to the medical and geriatrics, surgical, gynaecological and emergency medicine wards. Pregnant patients or those who developed diarrhoea after admission were excluded. The hospital's electronic medical record system was accessed for patient details. The recruited patients were divided into two age groups (18-64 and 65 years old). The length of hospital stay, patient demographics, admission time, and stool culture results were compared between these two groups. Factors that could potentially influence the antibiotic prescription behaviour of physicians were also analysed including age, fever, travelling history and admission time. The susceptibility of the bacterial pathogens towards different antibiotics was also analysed. The frequency

3 Wong et al./bacterial gastroenteritis 29 of occurrence of gastroenteritis due to Vibrio and Salmonella species was studied with reference to the monthly temperature. In our hospital, the following selective media were routinely used for processing stool samples: selenite-f broth (for Salmonella), alkaline peptone water broth (for Vibrio species), cycloserinecefoxitin-fructose agar and neomycin agar (for Clostridium difficile). Stool assays for Clostridium difficile toxin detection were performed upon request. Analysis of categorical data was done by using the Chisquare test and Fisher's exact test with Yates' correction. Continuous data like the length of stay in hospital was analysed by unpaired student t-test. The work was approved by the Ethics Committee of the New Territories West Cluster, Hospital Authority. Results We identified 891 patients who were hospitalised for acute gastroenteritis. Fifty-four patients were excluded from analysis because they were pregnant or having wrong entry of the principal diagnosis. Hence, the final study population included 837 patients of which 488 were female and 349 patients were male. The mean age of patient was 59.9 (SD±21.7) years. The demographical data, clinical and microbiological findings are shown in Table 1. The patients were divided into two age groups, years and 65 years respectively. Overall, 43.5% (364/837) of the patients were hospitalised for one day only and 78.3% (655/837) were discharged within three days of hospitalisation. The duration of hospitalisation for patients aged years were significantly shorter than those aged 65 years (mean±sd=2.0±1.5 days versus 4.7±2.2 days, p<0.001). In both groups, approximately two-thirds of patients had their stools sent for bacterial culture, but the older age group had more stools sent for viral investigation. Regarding the bacteriology results, 89 patients had a bacterial pathogen found, which means a 15.8% (89/562) yielding rate. The organisms found in descending order were: Vibrio parahaemolyticus (38.2%, 34/89), Salmonella species (34.8%, 31/89), Campylobacter species (11.2%, 10/89), Plesiomonas species (10.1%, 9/89), Aeromonas species (3.4%, 3/89) and Clostridium difficile (2.2%, 2/89). The isolation frequencies of Salmonella and V. parahaemolyticus from stool varied seasonally. The monthly number of isolates ranged from 0-7 (peak in July) for Salmonella and 0-8 (peak in August) for V. parahaemolyticus. Most were found in the hotter months (May to September) of the year, with 58.1% (18/31) for Salmonella and 76.5% (26/34) for V. parahaemolyticus (Figure 1). Susceptibility test was performed for a total of 75 isolates, comprising of V. parahaemolyticus, Salmonella and Campylobacter species. Susceptibility to Table 1. Patient demographics, clinical and microbiological findings No yr (n=453) 65 yr (n=384) P value Male sex 185 (40.8%) 164 (42.7%) Duration of hospitalisation, Mean ± SD (days) 2.0± ±2.2 <0.001 Stool culture performed: Bacteriology 305 (67.3%) 257 (66.9%) Virology 106 (23.4%) 128 (33.3%) Pathogen found: Total no. of bacteria <0.001 Vibrio parahaemolyticus 31 3 <0.001 Salmonella species 29 2 <0.001 Campylobacter species Plesiomonas species Aeromonas species Clostridium difficile

4 30 Hong Kong j. emerg. med. Vol. 17(1) Jan 2010 Figure 1. Relationship between mean air temperature and monthly cases of Vibrio parahaemolyticus and Salmonella gastroenteritis in a regional hospital in Hong Kong, (Data for the mean air temperature were obtained from the Hong Kong Observatory). ciprofloxacin was highest for V. parahaemolyticus (100%, 34/34), followed by Salmonella (93.5%, 29/31) and lowest for Campylobacter (0%, 0/10). For tetracycline, the susceptibility rates were 100% (34/34) for V. parahaemolyticus isolates, 71.0% (22/31) for Salmonella isolates and 10.0% (1/10) for Campylobacter isolates. All Campylobacter isolates were sensitive to erythromycin. We studied the factors potentially influencing the prescription behaviour of the physicians and they were shown in Table 2. About 28.6% (239/837) patients were prescribed fluoroquinolones. The proportions of patients who were given fluoroquinolones with fever and travel history (46.4% and 5.9%) outnumbered those without antibiotics given (9.0% and 1.3%) significantly (p<0.01). There was no association between the time the patients were admitted and the prescription of prophylactic antibiotics. The yielding rate of a bacterial pathogen was significantly higher in the antibiotics group. Among all the 89 patients who had their stool culture positive for bacterial pathogens, 81 (91.0%) were discharged before the stool culture results came back. The mean duration of stay of the patients with a positive stool culture was 3.4 days, but it took on average 3.9 days for the stool culture results to come back. Discussion The positive yield rate of a bacterial pathogen was 15.8% in our study, and this is much lower than some of the other studies around the world. 4-6 If the ordering of culture took into account specific symptoms, the yield could be higher. For example, a study in Thailand showed that for patients with positive dysentery symptoms, i.e., three or more loose stools containing blood or mucus associated with at least one of the

5 Wong et al./bacterial gastroenteritis 31 Table 2. Analysis of factors potentially influencing the decision to give empirical antimicrobial Empirical treatment with fluoroquinolone Yes (n=239) No (n=598) P value Age 65 years 99 (41.4%) 285 (47.7%) Fever (temperature >38ºC) 111 (46.4%) 54 (9.0%) <0.001 Travel history 14 (5.9%) 8 (1.3%) Bloody diarrhoea 8 (3.3%) 6 (1.0%) Admission time: Night shift, 00:00-08:59 48 (20.1%) 131 (21.9%) Day shift, 09:00-17:00 93 (38.9%) 244 (40.8%) Evening shift, 17:01-23:59 98 (41.0%) 223 (37.3%) Stool culture results: Overall 47 (19.7%) 42 (7.0%) <0.001 Salmonella species 18 (7.5%) 13 (2.2%) Vibrio parahaemolyticus 17 (7.1%) 17 (2.8%) Campylobacter species 4 (1.7%) 6 (1.0%) following symptoms: abdominal cramp, nausea, vomiting, or temperature greater than 38ºC, the positive yield would be as high as 90.9%. 8 A local study showed that bloody diarrhoea occurred in up to 10. 8% of stool culture positive patients. 9 Therefore it is important to focus on specific symptoms when sending stool cultures in order to improve the yield. In our study, most patients (91.0%) were already discharged before the stool culture results were available. The mean duration of stay of our patients with bacterial pathogens (3.4 days) was shorter than the mean duration of time needed before the stool culture results became available (3.9 days). This is compatible with what Manatsathit et al mentioned in their paper. 10 Hence, stool culture has a limited value for the acute management of most patients presenting with diarrhoea. Although performing stool culture takes time, it represents an important public health concern especially on statistics and epidemiology. It is still an important microbiological investigation, but we probably need a policy for better utilisation in our locality so as to increase the yield. Different guidelines have been raised due to differences in epidemiology. In general, bloody diarrhoea, fever, dehydration, diarrhoea that does not subside after a few days, 7,10-12 outbreaks of gastroenteritis or food poisoning, 13 recent travel to high risk areas, and immunocompromised states 14 are commonly quoted indications for doing stool culture. There is no clear guideline in Hong Kong currently as to under what circumstances should stool culture be sent. The necessity to perform such investigation largely depends on local data and epidemiology. From Table 1, we can see that the number of patients under age 65 with Salmonella and Vibrio infections significantly outnumbered the elderly group. Also, both of these organisms are mostly found in the warmer months (May to September) in Hong Kong. The occurrence of V. parahaemolyticus has been reported in many Asian countries including Japan, probably due to the fact that many Asians love to eat raw fish, raw oyster, and inadequately cooked seafood. 15,16 The Japanese eating style adopted by the younger generation in Hong Kong could be partly contributing to the high incidence of V. parahaemolyticus infection here. In order to maximise the yield of stool culture, we suggest that stool should be sent for culture only if one of the following features exists: severe diarrhoea with dehydration, a temperature of more than 38.5ºC, passage of bloody stools, recent travel to high-risk countries, group diarrhoea or food poisoning for those aged less than 65 years old, and those who are immunocompromised. Fluoroquinolones such as ciprofloxacin is a common type of empirical antibiotic for bacterial gastroenteritis. Fluoroquinolones have been shown to reduce the

6 32 Hong Kong j. emerg. med. Vol. 17(1) Jan 2010 duration of diarrhoea in patients with both culturepositive and culture-negative gastroenteritis. 4 Although ciprofloxacin appears to be an ideal antibiotic against V. parahaemolyticus and Salmonella infections here, it is less effective for Campylobacter. The number of isolates tested in this study was too small to justify any conclusion on resistance pattern. There were other studies in Hong Kong which also supported a similar finding. 9,17 Unfortunately, with the increased usage of empirical antibiotics, resistance to fluoroquinolones is now on the rising trend all around the world. 18,19 However, the truth is that most types of gastroenteritis do not require antibiotics. Fluoroquinolones would be beneficial in Shigella and enterotoxigenic and enteroinvasive E. coli infections, and the recommended dose of treatment is ciprofloxacin 500 mg twice daily for 1 to 5 days. 11,12 For Salmonella, V. parahaemolyticus, Aeromonas, Plesiomonas, Campylobacter and Yersinia infections, antibiotics are usually not required, 11,12 except in special situations, such as extremes of age and immunocompromised states. As mentioned above, stool culture results usually take time and it would be difficult to predict the type of bacterial pathogen beforehand. Notably, the numbers of positive isolates in stool culture in this study were higher among the antibiotics group (Table 2). It could be related to the signs and symptoms that the patients were having, in particular fever and travel history (both p <0.01), and physicians had relied on these two parameters to predict whether a patient was having bacterial gastroenteritis and the subsequent decision to prescribe fluoroquinolones. Thielman and Guerrant suggested that for patients with evidence of inflammatory diarrhoea, empirical treatment with fluoroquinolones, pending faecal testing, was reasonable. 11 The Coordinating Committee in Accident and Emergency Services in Hong Kong proposed in 2007 that empirical antibiotics be used under some situations: fever >38.5ºC, immunocompromised state, bloody stool, profuse diarrhoea, and duration of diarrhoea more than 3 days. 20 We therefore encourage frontline physicians to follow this set of local guidelines and to use antibiotics accordingly, so as to prevent overuse of fluoroquinolones and the emergence of drug resistance; at the same time preventing the side effects of drugs. One last point to add here is the duration of fluoroquinolone to be used. As mentioned, many articles suggest the use of ciprofloxacin for a short duration of 3 to 5 days, 10-13,20 but these are mostly for shigellosis or infection with E. coli. Before stool culture results are available, it is not easy to predict the type of bacterial pathogen that the patient is suffering from. We therefore suggest following the above regime when bacterial gastroenteritis is suspected. Limitations Data in this retrospective review were mainly collected from computer records, and some data might have been lost in case the physicians did not key in the necessary information, e.g. travel history, bloody diarrhoea etc. Only hospitalised patients were included, which might not be applicable to the general population. Specific gastrointestinal tract infections such as shigellosis or cholera might be missed when using "gastroenteritis" as the search term. Different rationales for giving prophylactic antibiotics among different physicians would be a bias to the results as well. Conclusions The commonest bacterial pathogens in Hong Kong are Vibrio parahaemolyticus and Salmonella species. Although the bacterial yield for stool culture remained low, it might give important health information to the public and contribute to epidemiology data and so provide guidance to our treatment and prevention plans. Local and departmental guidelines should be referred to before ordering investigations and prescriptions. Declaration This study has been submitted by the first author to the University of Hong Kong for partial fulfilment of the requirement for the Postgraduate Diploma in Infectious Disease.

7 Wong et al./bacterial gastroenteritis 33 References 1. Goodman L, Segreti J. Infectious diarrhea. Dis Mon 1999;45(7): Kosek M, Bern C, Guerrant RL. The global burden of diarrheal disease, as estimated from studies published between 1992 and Bulletin of the World Health Organization [serial online] 2003;81(3): [cited 2009 Jul 28]. Available from: bulletin/volumes/81/3/kosek0303.pdf. 3. Mead PS, Slutsker L, Dietz V, McCaig LF, Bresee JS, Shapiro C, et al. Food-related illness and death in the United States. Emerg Infect Dis 1999;5(5): Goodman LJ, Trenholme GM, Kaplan RL, Segreti J, Hines D, Petrak R, et al. Empiric antimicrobial therapy of domestically acquired acute diarrhea in urban adults. Arch Intern Med 1990;150(3): Fox R, Taylor A, Beeching NJ, Nye FJ. Clinical audit of ciprofloxacin use in adults admitted to hospital with gastroenteritis. J Infect 1996;33(1): Kaminski N, Bogomolski V, Stalnikowicz R. Acute bacterial diarrhea in the emergency room: therapeutic implications of stool culture results. J Accid Emerg Med 1994;11(3): Singapore Ministry of Health. Use of antibiotics in adults. Singapore Ministry of Health guidelines [online] [cited 2009 Jul 28]. Available from: moh.gov.sg/mohcorp/publications.aspx?id= Murphy GS, Bodhidatta L, Echeverria P, Tansuphaswadikul S, Hoge CW, Imlarp S, et al. Ciprofloxacin and loperamide in the treatment of bacillary dysentery. Ann Intern Med 1993;118(8): Chan SS, Ng KC, Lyon DJ, Cheung WL, Cheng AF, Rainer TH. Acute bacterial gastroenteritis: a study of adult patients with positive stool cultures treated in the emergency department. Emerg Med J 2003;20(4): Manatsathit S, Dupont HL, Farthing M, Kositchaiwat C, Leelakusolvong S, Ramakrishna BS, et al. Guideline for the management of acute diarrhea in adults. J Gastroenterol Hepatol 2002;17 Suppl:S Thielman NM, Guerrant RL. Acute infectious diarrhea. N Engl J Med 2004;350(1): DuPont HL. Guidelines on acute infectious diarrhea in adults. The Practice Parameters Committee of the American College of Gastroenterology. Am J Gastroenterol 1997;92(11): Farthing M, Feldman R, Finch R, Fox R, Leen C, Mandal B, et al. The management of infective gastroenteritis in adults. A consensus statement by an expert panel convened by the British Society for the Study of Infection. J Infect 1996;33(3): Aranda-Michel J, Giannella RA. Clinical practice. Acute diarrhea: a practical review. Am J Med 1999;106 (6): Nair GB, Ramamurthy T, Bhattacharya SK, Dutta B, Takeda Y, Sack DA. Global dissemination of Vibrio parahaemolyticus serotype O3:K6 and its serovariants. Clin Microbiol Rev 2007;20(1): Bitterman RA. Acute gastroenteritis and constipation. In: Rosen P, editor. Emergency medicine: concepts and clinical practice, 4th ed. St. Louis, MO: Mosby Year Books; p Chan SSW, Ng KC, Lyon DJ, Cheung WL, Rainer TH. Empiric antibiotics for acute infectious diarrhea. Hong Kong Pract 2001;23(10): Smith KE, Besser JM, Hedberg CW, Leano FT, Bender JB, Wicklund JH, et al. Quinolone-resistant Campylobacter jejuni infections in Minnesota, Investigation Team. N Engl J Med 1999;340(20): Ruiz J, Marco F, Oliveira I, Vila J, Gascón J. Trends in antimicrobial resistance in Campylobacter spp. causing traveler's diarrhea. APMIS 2007;115(3): Coordinating Committee in A&E Services, Hospital Authority, Hong Kong. COC guideline on antibiotic use in A&E [A&E Clinical Guideline No. 23]. [online] 2007 Jul. [cited 2009 Jul 28]. Available from: ae_clin_guidelines/ae_antibiotic_guideline.pdf.