Interventions Trade off between benefits and harms Empirical antibiotic treatment in travellers' diarrhoea Empirical antibiotic treatment in community acquired diarrhoea

Summary points RCTs have found that empirically treating travellers' diarrhoea with antibiotics reduces the length of illness by one to two days. In community acquired diarrhoea, RCTs have found that ciprofloxacin reduces the duration of diarrhoea by one to two days. Trials of other antibiotics have found no evidence of benefit or have not reported on time to cure. Some RCTs found that treatment prolonged excretion of organisms and was associated with the development of resistant organisms.

Background

Definition

Diarrhoea is defined as watery or liquid stools, usually with an increase in stool weight above 200 g per day and an increase in daily stool frequency.

Incidence/prevalence An estimated 4000 million cases of diarrhoea occurred worldwide in 1996, resulting in 2.5 million deaths.¹ In developing countries, diarrhoea is reported to cause more deaths in children aged under five years than any other condition.¹ In the United States, which has a low incidence, the estimated incidence for infectious intestinal disease is 0.44 episodes per person per year, or one episode per person every 2.3 years, resulting in about one consultation with a doctor per person every 28 years.² The epidemiology of travellers' diarrhoea (people who have crossed a national boundary) is not well known. Incidence is higher in travellers going to developing countries but varies widely by location and season of travel. ³

Aetiology

The cause depends on geographic location, standards of food hygiene, sanitation, water supply, and season. The commonly identified causes of sporadic diarrhoea in adults in developed countries include Campylobacter, Salmonella, Shigella, Escherichia coli, Yersinia, protozoa, and viruses, but no pathogens are identified in over half of patients. In returning travellers, about 80% of cases are caused by bacteria, such as enterotoxigenic E coli, Salmonella, Shigella, Campylobacter, Vibrio, enteroadherent E coli, Yersinia, and Aeromonas.

Prognosis

In developed countries, death from infectious diarrhoea is rare although serious complications causing admission to hospital sometimes occur, such as severe dehydration and renal failure. Elderly people and those in long term care have an increased risk of death.⁴

Aims

To reduce the infectious period, length of illness, risk of dehydration, risk of transmission to others, and rates of severe illness and to prevent complications and death.

Outcomes

Time from start of treatment to last loose stool; number of loose stools per day; relief of cramps, nausea, and vomiting; rate of hospital admission; incidence of severe illness; duration of excretion of organisms; and presence of bacterial resistance.

Methods

Clinical Evidence search and appraisal in June 1999. Trial quality was assessed on allocation concealment and inclusion of all randomised participants. Trials were excluded if they did not meet epidemiological quality criteria. Most trial participants had moderate to severe diarrhoea, usually defined as acute diarrhoea lasting less than a week, more than three loose stools in 24 hours or more than two in eight hours, and symptoms of an enteric illness such as nausea, vomiting, and cramps.

Question: What are the effects of empirical antibiotic treatment in travellers' diarrhoea?
Empirical treatment with antibiotics shortened illness duration in adults with diarrhoea acquired overseas. Treatment was associated in some people with prolonged presence of bacterial pathogens in the stool and development of resistant strains.

Benefits

We found no systematic review. We found 15 RCTs in a total of 2251 travellers 5-19 comparing empiric use of one or more antibiotics versus placebo. Eight trials evaluated quinolones,^5-12 two evaluated cotrimoxazole,^{6, 13-15} and one evaluated each of trimethoprim,¹⁴ aztreonam,¹⁶ bicozamycin,¹⁷ pivmecillinam,¹⁸ and rifaximin.¹⁹ Seven trials studied US students aged >18 years visiting Guadalajara, Mexico, during summer months. The other eight were in different locations. Entry criteria varied among trials, and treatment duration ranged from a single dose to five days. All trials found reduced duration of diarrhoea ranging from 1 to 2.5 days, but confidence intervals were not available from published data in seven of the trials. The largest trial, in which 70% of the 598 participants had a history of recent travel, reported a one day improvement in the median duration of diarrhoea, from four to three days (no confidence interval available).⁵

Harms

Adverse effects varied by agent, with inci- dence in the trials ranging from 1.7%⁷ to 18%.¹¹ Common reported harms were gastrointestinal symptoms (cramps, nausea, anorexia), dermatological symptoms (rash), and respiratory symptoms (cough, sore throat). In the largest trial, people with salmonella infection treated with norfloxacin had significantly prolonged excretion of Salmonella in stool compared with those given placebo (median time to clearance of Salmonella from stool 50 days in the norfloxacin group compared with 23 days in the placebo group).⁵ In addition, six of nine Campylobacter isolates obtained after treatment had developed resistance to norfloxacin. One small trial reported that four of eight participants treated with ciprofloxacin developed resistant isolates at 48 hours (difference from placebo group 50%, 95% CI 15% to 85%).⁷ One trial reported three cases of continued excretion of Shigella in people treated with trimethoprim-sulphamethoxazole. Two of these isolates became resistant to the drug. The participants were clinically well. Other trials did not find post-treatment resistance or did not report it.⁸

Comment

Studies were generally well conducted. All but one⁸ were double blinded. Participant blinding through use of identical placebo was used and well described in 10 of the studies, and probably adequate in the remaining five although not as clearly stated. However, only one study reported using an appropriate statistical method for analysing time to event outcomes.¹⁵ Several trials reported surrogate end points, such as change in faecal consistency,¹⁹ rather than the primary outcome of interest.^12,18,19

Question: What are the effects of empiric antibiotic treatment in community acquired diarrhoea?

RCTs have found that ciprofloxacin reduces duration of diarrhoea developed in the community by one to two days. Trials of other empiric treatments with antibiotics either found no effect or did not report data on time to cure.

Benefits

We found no systematic review. We found nine RCTs in eight reports^20-27 (1760 participants) comparing one or more antibiotics with placebo. Trials were conducted in 12 sites in 11 countries. Four trials were conducted in developed countries, and the others took place in developing countries. The largest study included 332 adult inpatients in a multicentre trial of fleroxacin.²⁰ Eight trials evaluated quinolones,^20-27 four evaluated cotrimoxazole, ^21,22,25 and one evaluated cloquinol.²¹ Entry criteria varied between trials, and treatment duration ranged from a single dose to five days. Three trials found that antibiotics reduced illness duration^24,27 or decreased number of liquid stools by 48 hours,²⁰ while five found no benefit in reducing illness duration.^21-23,26 One trial found reduced duration for ciprofloxacin but not for cotrimoxazole.²⁵

Harms

Adverse effects varied by agent. In one RCT of lomefloxacin, 33% of treated participants reported adverse effects compared with 2.7% in the placebo group (ARI 31%, 95% CI 17% to 46%). Two were withdrawn from the trial after developing anaphylactoid reactions.²³ In the same trial, 18% of treated participants developed isolates resistant to lomefloxacin.²³ In the multicentre trial of ciprofloxacin and cotrimoxazole, five people with Campylobacter isolated from stool (two treated with ciprofloxacin and three treated with cotrimoxazole) developed isolates resistant to the respective agents.²⁵ In the largest trial, three deaths occurred - two people treated with fleroxacin and one person who received placebo. Two of the deaths occurred from hypovolaemic shock (one with fleroxacin, one with placebo).²⁰

Comment

The main pathogenic organisms found in each study varied and may partly explain variations in effect. Reported outcomes varied between trials, precluding direct comparisons or summary of treatment effect on the basis of published reports.

Clinical Evidence Infectious Diseases adviser is Paul Garner, Liverpool, UK.
Clinical Evidence is published by BMJ Publishing Group. It is a compendium of the best available evidence for effective healthcare. Student subscription rate £45/$66, issues 2 and 3. For more information including how to subscribe, please visit our website at www.evidence.org

Pictures from Toilets of the World website by Bob Cromwell, Lafayette, Indiana.

1. The World Health Report 1997. Geneva: World Health Organization, 1997:14-22.
2. Garthwright WE, Archer DL, Kvenberg JE. Estimates of incidence and costs of intestinal infectious diseases in the United States. Public Health Rep 1988; 103:107-115.
3. Cartwright RY, Chahed M. Foodborne diseases in travellers. World Health Stat Q 1997;50:102-110.
4. Lew JF, Glass RI, Gangarosa RE, et al. Diarrheal deaths in the United States 1979 through 1987. JAMA 1991;265:3280-3284.
5. Wistrom J, Jertborn M, Ekwall E, et al. Empiric treatment of acute diarrheal disease with norfloxacin: a randomized, placebo-controlled study. Swedish Study Group. Ann Intern Med 1992;117:202-208.
6. Ericsson CD, Johnson PC, DuPont HL, et al. Ciprofloxacin or trimethoprim-sulfamethoxazole as initial therapy for travelers' diarrhea: a placebo-controlled, randomized trial. Ann Intern Med 1987; 106:216-220.
7. Wistrom J, Gentry LO, Palmgren AC, et al. Ecological effects of short-term ciprofloxacin treatment of travellers' diarrhoea. J Antimicrob Chemother 1992;30:693-706.
8. Wistrom J, Jertborn M, Hedstrom SA, et al. Short-term self-treatment of travellers' diarrhoea with norfloxacin: a placebo-controlled study. J Antimicrob Chemother 1989;23:905-913.
9. DuPont HL, Ericsson CD, Mathewson JJ, et al. Five versus three days of ofloxacin for traveler's diarrhea: a placebo-controlled study. Antimicrob Agents Chemother 1992;36:87-91.
10. Mattila L, Peltola H, Siitonen A, et al. Short-term treatment of traveler's diarrhea with norfloxacin: a double-blind, placebo-controlled study during two seasons. Clin Infect Dis 1993;17:779-782.
11. Salam I, Katelaris P, Leigh-Smith S, et al. Randomised trial of single-dose ciprofloxacin for travellers' diarrhoea. Lancet 1994;344:1537-1539.
12. Steffen R, Jori R, DuPont HL, et al. Efficacy and toxicity of fleroxacin in the treatment of traveler's diarrhea. Am J Med 1993;94(3A):S182-186.
13. DuPont HL, Reves RR, Galindo E, et al. Treatment of travelers' diarrhea with trimethoprim/sulfamethoxazole and with trimethoprim alone. N Engl J Med 1982;307:841-844.
14. Ericsson CD, Johnson PC, DuPont HL, et al. Role of a novel antidiarrheal agent, BW942C, alone or in combination with trimethoprim-sulfamethoxazole in the treatment of traveler's diarrhea. Antimicrob Agents Chemother 1986;29:1040-1046.
15. Ericsson CD, DuPont HL, Mathewson JJ, et al. Treatment of traveler's diarrhea with sulfamethoxazole and trimethoprim and loperamide. JAMA 1990; 263:257-261.
16. DuPont HL, Ericsson CD, Mathewson JJ, et al. Oral aztreonam, a poorly absorbed yet effective therapy for bacterial diarrhea in US travellers to Mexico. JAMA 1992;267:1932-1935.
17. Ericsson CD, DuPont HL, Sullivan P, et al. Bicozamycin, a poorly absorbable antibiotic, effectively treats travellers' diarrhoea. Ann Intern Med 1983;98:20-25.
18. Christensen OE, Tuxen KK, Menday P. Treatment of travellers' diarrhoea with pivmecillinam [letter]. J Antimicrob Chemother 1988;22:570-571.
19. Vinci M, Gatto A, Giglio A, et al. Double-blind clinical trial on infectious diarrhoea therapy: rifaximin versus placebo. Curr Ther Res 1984;36:92-99.
20. Butler T, Lolekha S, Rasidi C, et al. Treatment of acute bacterial diarrhea: a multicenter international trial comparing placebo with fleroxacin given as a single dose or once daily for 3 days. Am J Med 1993;94(3A):187-194.
21. De la Cabada FJ, DuPont HL, Gyr K, et al. Antimicrobial therapy of bacterial diarrhea in adult residents of Mexico-lack of an effect. Digestion 1992;53:134-141.
22. Lolekha S, Patanachareon S, Thanangkul B, et al. Norfloxacin versus co-trimoxazole in the treatment of acute bacterial diarrhoea: a placebo controlled study. Scand J Infect Dis 1988;56(suppl):35-45.
23. Ellis-Pegler RB, Hyman LK, Ingram RJ, et al. A placebo controlled evaluation of lomefloxacin in the treatment of bacterial diarrhoea in the community. J Antimicrob Chemother 1995;36:259-263.
24. Pichler HE, Diridl G, Stickler K, et al. Clinical efficacy of ciprofloxacin compared with placebo in bacterial diarrhea. Am J Med 1987;82(suppl 4A):329-332.
25. Goodman LJ, Trenholme GM, Kaplan RL, et al. Empiric antimicrobial therapy of domestically acquired acute diarrhea in urban adults. Arch Intern Med 1990;150:541-546.
26. Noguerado A, Garcia-Polo I, Isasia T, et al. Early single dose therapy with ofloxacin for empirical treatment of acute gastroenteritis: a randomised, placebo-controlled double-blind clinical trial. J Antimicrob Chemother 1995;36:665-672.
27. Dryden MS, Gabb RJ, Wright SK. Empirical treatment of severe acute community-acquired gastroenteritis with ciprofloxacin. Clin Infect Dis 1996;22:1019-1025.