Statement on Travellers' Diarrhea

Introduction

Diarrhea is a common medical problem affecting travellers, especially those who travel to low and middle income countries where there is a higher risk of encountering suboptimal sanitation and hygiene conditions ^{Reference 1}. Travellers' diarrhea (TD) can adversely affect travel plans and incur financial costs to the traveller, especially if medical care is required while travelling. Several factors, both travel-related (i.e. destination and type of travel) and traveller-related (i.e. country of origin, age) affect the risk of acquiring diarrhea ^{Reference 2} and the severity of symptoms.

The purpose of this statement is to provide health care professionals with information on risk factors, and recommendations for the prevention and treatment of TD. Information specifically addressing persistent diarrhea in the returned traveller is included in a separate statement ^{Reference 3}.

Background

Clinical and epidemiological features

Symptoms of TD range from mild to severe. Classical TD is defined as the passage of three or more unformed stools in a 24 hour period with at least one accompanying symptom including: nausea, vomiting, abdominal cramps or pain, fever or blood in stools (dysentery) ^{Reference 4}.

Symptoms of TD in adults tend to occur early during the trip, with onset dates reported on average during the third or fourth day of travel^{Reference 5}, ^{Reference 6}. However, children and youth under the age of 20 have been reported to experience a later average onset at eight days ^{Reference 7}. The duration of TD generally averaged between three to four days among adult travellers ^{Reference 5}, ^{Reference 6}, although average duration of incapacitation (i.e. unable to pursue planned activities) due to TD did not exceed 30 hours ^{Reference 5}, ^{Reference 6}, ^{Reference 8}. Longer durations of TD were observed in children, particularly those two years of age and younger ^{Reference 7}. Between 2 to10% of travellers may develop persistent diarrhea (i.e. lasting two weeks or longer) ^{Reference 9}.

Although TD is usually a mild and self-limiting disease, between 5% to 20% of travellers sought professional help (i.e. consulted a physician, nurse or pharmacist), between 30% to 60% used some form of medication, and some individuals required hospitalization ^{Reference 5-7}. In addition, between 12% to 50% of travellers were incapacitated for part of the trip due to TD ^{Reference 5}, ^{Reference 6} and 5% to10% may develop post-infectious irritable bowel syndrome (PI-IBS) ^{Reference 9}.

Etiological agents

TD is mainly acquired through the ingestion of food and beverages contaminated with pathogens that cause diarrhea. The most common etiologic agents for TD are bacterial, viral and parasitic. Bacterial pathogens, particularly enterotoxigenic and enteroaggregative Escherichia coli (ETEC and EAEC respectively) and Campylobacter, are the most common. A review of 51 studies on TD ^{Reference 10} found that roughly one third of TD cases from Latin America, the Caribbean, Africa, and South Asia were due to ETEC and one third of the cases from Southeast Asia were due to Campylobacter. Other bacterial pathogens such as Shigella and Salmonella accounted for a combined 10% to 15% of TD cases in those regions. Aeromonas and Plesiomonas species were more commonly reported in Asia and Africa and accounted for 5% to 8% of TD cases in those regions, while Vibrios accounted for 9% of the cases in Southeast Asia.

The most common viral pathogens that cause TD are noroviruses and rotaviruses which accounted for 19% to 25% of TD cases in Latin America, the Caribbean and Africa and between 3% to 5% of TD cases in Asia ^{Reference 10}. Noroviruses were also implicated in many gastrointestinal outbreaks on cruise ships ^{Reference 11}.

Parasitic pathogens such as Giardia, Cryptosporidium and Entamoeba histolytica, accounted for 2% or less of TD cases from Latin America, the Caribbean and Africa and between 8% to12% of cases in Asia ^{Reference 10}. Cyclospora cayetanensis has also caused diarrhea in travellers returning from Latin America, the Indian sub-continent and Southeast Asia ^{Reference 12}. Although parasitic pathogens are responsible for a smaller proportion of TD cases globally, diarrhea due to parasites tends to be more protracted and, consequently, requires health care intervention more frequently upon return from travel. Among travellers who visited a GeoSentinel travel clinic ^{Reference ii} to seek post-travel medical care, parasitic pathogens represents the most frequently identified cause of acute diarrhea ^{Reference 13}. Diarrhea due to Giardia and other gastrointestinal parasites were also reported more frequently in long-term travellers seeking post-travel medical care than short-term travellers ^{Reference 14}.

In the review article noted above ^{Reference 10}, no etiologic agent could be identified for approximately 40% to 50% of TD cases despite thorough microbiological evaluation. However, there is evidence to suggest that bacterial pathogens are responsible for many of these pathogen-negative TD cases as there are many documented cases of symptoms being reduced through use of antibacterials. Two studies using more sensitive laboratory methods for detecting pathogens such as PCR demonstrated that ETEC, EAEC and diffusely adherent Escherichia coli (DAEC) accounted for 26% to 30% of TD cases originally characterized as pathogen negative in travellers to Guatemala, Mexico, Jamaica and India ^{Reference 15}, ^{Reference 16}.

Epidemiology

A review of data from observational studies of diarrhea rates among travellers originating from high income countries found that incidence rates for two-week stays ranged from 20% to 90% for travel to high-risk regions (low and middle income countries) ^{Reference 4}. Between 55% to 59% of ill returned travellers who visited one of the travel clinics associated with GeoSentinel were diagnosed with acute diarrhea ^{Reference 1}. A study of travellers visiting four high-risk countries conducted in the late 1970s and again in the late 1990s found that TD rates remained similar over this 20 year time period ^{Reference 6}, ^{Reference 17}. However, a subsequent study in one of these four countries has shown a decline in rates of TD since the late 1990s which is thought to be due to efforts for improved hygiene in tourist facilities ^{Reference 18}.

Preliminary findings from C-EnterNet, an integrated enteric pathogen surveillance system with two sentinel sites in Canada (Ontario and British Columbia), for the 2011 surveillance year indicated that 30% of all cases of reportable enteric disease in Canada were associated with international travel ^{Reference 19}. There have been regional variations reported, with 25% in Waterloo, Ontario (from June 2005 to May 2009) ^{Reference 20} and 40% reported in British Columbia (in 2008) ^{Reference 21}. In the study from Waterloo, the travel-related cases accounted for 18% of the hospitalizations for enteric illness reported during that study period ^{Reference 20}. Note that the studies conducted through C-EnterNet only target reportable illnesses such as campylobacter enteritis, salmonellosis and giardiasis and therefore do not include other more notable travel-related etiologies for diarrhea such as enterotoxigenic Escherichia coli. Hence, the findings above only represent a fraction of all travel-related cases of diarrhea in Canada.

Methods

This statement was developed by a working group comprised of volunteers from the CATMAT committee, none of whom declared a relevant conflict of interest. All working group members were approved by the CATMAT secretariat and chair. The working group, with support from the secretariat, was responsible for: literature retrieval, synthesis and analysis; and the development of key questions and draft recommendations. The final statement was approved by the full CATMAT membership.

Recommendations in this statement on interventions to prevent and treat TD were developed using the GRADE methodology, wherever relevant and feasible. This approach has been increasingly adopted by guideline developers Reference 22, Reference 23. It stresses transparency and provides an explicit framework in which the following factors are considered and weighed when making a recommendation(s): confidence in the estimate of effect (quality of data, see Box 1: FAQ on pg. 6 for details); balance of benefits and harms; and values and preferences. Resulting recommendations are expressed as strong or conditional (see Box 1: FAQ on pg. 6 for details).

Various recommendations for preventive and treatment interventions provided within this statement include off-label use of medications. Product monographs or other similarly approved standards or instructions for use should be reviewed prior to use.

The following summarizes the process used to develop this statement:

For the GRADE recommendations:

1. The following key "PICO" (population of interest, intervention, comparison and outcome) questions were identified:

1. Among Canadian travellers, does the administration of the inactivated oral cholera vaccine (Dukoral^®) decrease the risk of acquiring TD as compared to no vaccine (placebo)?
2. Among Canadian travellers, does the administration of a relevant chemoprophylactic agent (i.e., antisecretory or antibiotic) decrease the risk of acquiring TD as compared to no chemoprophylaxis (placebo)?
3. Among Canadians having acquired TD during travel, does the administration of a relevant therapeutic agent (i.e., antisecretory, antimotility, or antibiotic) decrease the duration and/or severity of TD as compared to no therapy (placebo)?
4. Among Canadians having acquired TD during travel, does the administration of a relevant therapeutic agent (i.e., antisecretory, antimotility, or antibiotic) decrease the duration and/or severity of TD as compared to an alternative therapy (e.g., addition of antimotility to antibiotic, different class of antibiotic)?

2. Key questions to define the magnitude of benefits and harms were also identified:

1. What harms are associated with TD chemoprophylactic and therapeutic agents, as well as with vaccination?
2. What are the important risk factors for TD among travellers (e.g., destination, duration of travel, age, comorbidities such as infection with HIV or acid suppression/achlorhydria)?
3. What are the values and preferences of travellers regarding the magnitude of risk reduction in TD that would make use of the relevant intervention worthwhile given the associated cost and inconvenience?

3. With the aid of a reference librarian, a strategy was developed to identify relevant literature. Several electronic databases (Ovid MEDLINE, Embase, Global Health and Scopus) and the Cochrane Review Database were searched using variations on the term "travellers' diarrhea" and the relevant search term or terms for each intervention of interest. The search spanned the initial date for each database up to June 1, 2013. For all searches, only articles in English and/or French were retained. See Appendix 1 for an example of a search strategy used. Reference lists from relevant studies were also scanned to identify any studies not captured by the database searches.

4. From these searches, literature addressing the population of interest, intervention, comparison and outcome ("PICO") and other questions was identified. Systematic reviews that addressed the efficacy and safety of the TD interventions were specifically sought out.

5. Although some studies evaluated prevention and treatment of mild or moderate TD, our recommendations only addressed outcomes using the classical definition of TD: three or more unformed stools with at least one enteric symptom within a 24 hour period. For studies evaluating antibiotics and vaccine, those conducted in a non-traveller population were also excluded. For antisecretory and antimotility studies, non-traveller populations were considered in situations where traveller data were scarce, but their inclusion in the analysis led to a rating down in the overall quality of evidence. Finally, several studies were excluded that evaluated antibiotics which are either no longer available in Canada or are no longer prescribed for TD due to widespread antibiotic resistance.

6. A quality assessment of studies evaluating the efficacy of each of the TD interventions was performed, and results were collated into evidence profiles and summary of findings tables (see Appendix 2) as per the GRADE methodology^{Reference 24-26}.

7. Recommendations were developed for use of each TD intervention, taking into consideration:

1. our confidence in the estimates of the efficacy and harms of each intervention,
2. the balance of harms and benefits, and
3. the values and preferences of travellers.

The cost of each intervention, normally borne by the traveller, was not explicitly considered as there were no data available on willingness to pay (WTP) for TD in travellers.

For the evaluation of interventions not given a GRADE recommendation:

8. The evaluation of certain interventions is not amenable to the GRADE approach, either due to lack of credible alternatives to the intervention in question (i.e., best practices considered as "common sense" approaches) or an insufficient evidence base. As such, CATMAT provided recommendations for these interventions based on a review of the literature, as well as expert opinion. The additional non-GRADE questions that were considered are the following:

1. What are the documented antimicrobial resistance patterns for each of the antibiotics recommended for use in prevention and treatment of TD; and does pathogen susceptibility to antimicrobial treatment vary by destination?
2. Do hygiene and/or food and water precautions reduce the risk of acquiring TD among travellers?
3. Does use of probiotics, prebiotics, or a combination of the two (synbiotics) reduce the risk of acquiring TD among travellers?
4. What are the best practices associated with managing TD-related dehydration among travellers?

Results

Risk factors

The following travel-related and traveller-related factors have been shown to affect the risk of acquiring TD or influence the type or severity of symptoms of TD.

Travel-related

Travel destination has a large influence on the risk of acquiring TD Reference 1, Reference 6. In a global retrospective observational analysis of gastrointestinal infection among illreturned travellers who visited a GeoSentinel clinic, travel to sub-Saharan Africa, South America or South Asia was associated with the highest reporting rate ratios (RRRs), which ranged from 203 to 890 (reference group: Northern and Western Europe); travel to Oceania, the Middle East, North Africa, Central America, the Caribbean or Southeast Asia was associated with lower reporting rate ratios (RRRs from 41 to 104); and travel to south/central/eastern Europe, North America, Northeast Asia or Australasia was associated with the lowest reporting rate ratios (RRRs from 2 to 17) Reference 1. Rates also varied between countries within the same region: for example, within North America, reporting rates were highest in Mexico and within Southeast Asia, reporting rates were lowest in Malaysia and Singapore Reference 1. Among the Canadian GeoSentinel sites, approximately 30% of TD cases acquired their illness in the Caribbean and Central America (including Mexico), followed by South Central Asia and sub-Saharan Africa (18% each), Southeast Asia and South America (7% each), and 20% from other regions Reference 27.

The type of travel and accommodation can also influence the risk of acquiring TD. The TD attack rates were higher among backpackers or travellers participating in adventure tours (i.e. staying in private accommodations, camps, cheap hotels or similar places) compared to those on a beach vacation (i.e. stay in one hotel along the shore) ^{Reference 17}, ^{Reference 28}. In addition, business travellers had a lower risk for TD compared to tourists and honeymooners ^{Reference 5}, ^{Reference 6}. A few studies have shown that staying in luxury accommodations compared to standard hotels does not necessarily reduce the risk of acquiring TD ^{Reference 17},^{Reference 28}.

The incidence of TD has been reported to increase with increased duration of stay of up to two to three weeks ^{Reference 5}, ^{Reference 28}, ^{Reference 29}. Long-term travellers (trip duration more than 6 months) were more likely to have chronic diarrhea, giardiasis and post-infectious irritable bowel syndrome than short-term travellers (trip duration less than one month), while acute diarrhea and acute bacterial diarrhea were more common in short-term travellers ^{Reference 14}.

Seasonality of travel appears to influence the risk of TD. In one study, Austrians who travelled to various regions in Africa, Asia, and in Central and South America had lower incidence rates for TD overall when they travelled during the months of December to March (typically colder months) than those who travelled during the months of June to September (typically warmer months) ^{Reference 28}.

Traveller-related

Age of the traveller may affect the risk of TD. Younger adults (aged 30 years or younger) have been shown to be at highest risk for TD compared to older adults ^{Reference 5}, ^{Reference 6}, ^{Reference 28}. In one observational study, children (< 18 years of age) had a non-significant morbidity risk increase due to acute diarrhea compared to adults (adjusted for gender, travel region, reason for travel and travel duration), however younger children (≤ 11 years) had a significantly higher rate compared to older children ^{Reference 30}. Another study also showed that small children (≤ 2 years) tended to have a more severe and prolonged clinical course for TD compared to other pediatric age groups ^{Reference 7}.

Travellers who originated from high or intermediate risk regions for TD (i.e. South Asia and sub-Saharan Africa) had lower rates for TD compared to travellers who originated from low risk regions (i.e. North America and Australasia) ^{Reference 5}, ^{Reference 31}. Lower rates were also observed among travellers who reported recent ^{Reference 5}, ^{Reference 6} or prior ^{Reference 17}, ^{Reference 32} travel to the tropics or other low and middle income countries. Lower rates of TD were also reported in individuals who had experienced TD in the preceding year ^{Reference 33}.

Some studies have shown that individuals may have a genetic susceptibility to certain etiologic agents for TD Reference 34-36. Other pathophysiologic factors were also shown in several studies to influence the risk of diarrhea (i.e. diarrhea in HIV-infected individuals was found to be strongly associated with low CD4 cell counts Reference 37; use of medication that reduces gastric acid secretion, such as proton pump inhibitors Reference 38 and histamine 2 antagonists Reference 39, increases susceptibility to bacterial infections such as Campylobacter and Salmonella). In the case of proton pump inhibitors however, the magnitude of the enhanced susceptibility to acute diarrhea from chronic use is not clear Reference 38.

Prevention - Best practices

Evidence regarding various "common sense" interventions thought to prevent TD was reviewed, including hand hygiene, food and beverage selection and water purification. As these are non-invasive interventions with broad applicability, they were not subject to a GRADE evaluation.

Hand Hygiene

The evidence for the effectiveness of hand hygiene (i.e. washing hands with soap and water or disinfection through the use of alcohol-based hand sanitizers) in preventing diarrhea in travellers is limited. Furthermore, hand hygiene would not be expected to prevent illness related to the consumption of contaminated food and water. Nevertheless, the importance of hand hygiene in reducing the risk of diarrheal illness among non-travel-related cases in both low to middle income ^{Reference 40-42} and high income ^{Reference 43}, ^{Reference 44} countries has been well-documented. Findings from a systematic review conducted on the benefits of hand washing found that interventions that promote hand washing can reduce diarrheal episodes by about one-third ^{Reference 45}. Therefore, hand washing with soap and water is recommended before preparing meals, before eating meals, and after urination or defecation.

Alcohol-based hand sanitizers are also becoming a more commonly used source of hand hygiene. A few studies found that hand rubbing with an alcohol-based solution was comparable to ^{Reference 46}, or better than ^{Reference 47-49}, hand washing with an antiseptic soap at reducing bacterial hand contamination. Therefore in the absence of ready access to soap and water, alcohol-based hand sanitizers may aid in reducing the risk of diarrheal illness among travellers.

Food and Beverage Selection

The ingestion of contaminated foods and beverages is an important risk factor for acquiring enteric pathogens associated with TD. The presence of these pathogens in food and beverage samples taken from higher risk travel destinations ^{Reference 50-52}, as well as association of their consumption with travel-related enteric outbreaks ^{Reference 53} has been identified. An informal review of the literature ^{Reference 54} failed to find a correlation between practicing standard dietary precautions and the risk of acquiring TD. These studies, however, were mostly based on retrospective surveys prone to recall bias, as well as low response rates, and failed to take into account important modifying factors such as host immunity, age and location where meals are prepared. For example, preparing one's own food likely improves the level of food hygiene and has been shown to significantly lower the risk of developing TD ^{Reference 50}.

Notwithstanding this lack of evidence, and in spite of some studies showing low compliance among travellers in following recommended dietary precautions ^{Reference 55}, ^{Reference 56}, travellers may still benefit from exercising caution in the choice of food and beverage consumption while travelling to higher risk areas as outlined below:

Adequately heated or pasteurized foods and beverages

Temperatures above 65°C have been found to reliably kill all bacterial pathogens, thereby making consumption of foods and beverages that are served steaming hot ^{Reference 57} a lower risk option. Consumption of undercooked or raw meats and seafood ^{Reference 53}, ^{Reference 58} and unpasteurized eggs and dairy products ^{Reference 53} have been implicated in the risk for TD and are best avoided. Foods cooked earlier in the day and not sufficiently reheated are also best avoided ^{Reference 59}.

Foods that are thoroughly cleaned and stored in hygienic conditions

Foods, particularly fruits and vegetables, which are not cooked or heated should be washed thoroughly in clean water or peeled prior to consumption to remove enteropathogens from the food surface. For travellers unable to prepare their own food, it is best to avoid fruits and vegetables that are difficult to clean (e.g., broad leafed vegetables) or peel ^{Reference 60}, or foods that are prepared, stored or served in unsanitary conditions ^{Reference 61}.

Soaking fruits and vegetables in disinfectants such as dilute bleach or permanganate solutions has been shown to reduce contamination. However, concentrations and contact time have not been well studied, and protozoal cysts will generally be resistant to relatively brief and incomplete contact. In addition, bleach loses its disinfectant properties in the presence of many organic compounds ^{Reference 62}.

Foods with low water and high sugar content

Bacteria need moisture for growth therefore moist food items served at room temperature are best avoided ^{Reference 63}. Dry items such as bread and rolls would be safer to consume ^{Reference 64}. However, the high sugar content in certain moist foods such as syrups, jellies, jams and honey inhibit the growth of bacteria and are assumed to be safe ^{Reference 63}.

Bottled carbonated and alcoholic drinks

An in vitro study of survival of several TD-related enteric pathogens in beverages ^{Reference 65} found that these pathogens were killed most quickly in wine, followed by carbonated drinks and beer. Greatest pathogen growth was observed in non-chlorinated drinking water and milk. Therefore bottled carbonated and alcoholic drinks may be relatively safe to drink while travelling.

Ice and bottled water

Ice made from purified water should be safe to consume; however, ice served at restaurants or by vendors may have been made from contaminated water sources and thus may not be safe Reference 66. Several studies found the bacteriological quality of various brands of bottled water sold in several international destinations to be highly variable, and some judged to be unsatisfactory by accepted health standards Reference 51, Reference 67-69. Studies conducted in two higher risk countries found that all imported brands of bottled water tested were within the World Health Organization (WHO) standards for purity while some domestic brands were not Reference 70, Reference 71. However, non-carbonated bottled water with intact seals can generally be assumed to be safe to drink.

Water Purification

Purified water is safe to drink. Water purification while travelling may be achieved through heat, chemical disinfection (combined with filtration if possible) or through ultraviolet (UV) radiation. Bringing water to a boil is the most effective way of producing potable water because all common enteric pathogens are readily inactivated or killed by heat upon boiling, even at moderately high altitudes ^{Reference 72}, ^{Reference 73}. Water should be boiled for one minute or kept covered once boiled for slow cooling ^{Reference 73}. Small portable heating coils or a kettle with an electrical outlet and current flexibility are inexpensive ways to ensure a constant supply of purified water.

Filters that trap particles of at least 0.2µm in size are effective against most bacteria and parasites; however most filters are not reliable for removal of viruses Reference 73. Therefore if possible, water filtration should be followed by chemical disinfection Reference 74.

Chemical disinfection may be achieved through the addition of a halogen such as iodine or chlorine to the water. Iodine, available in tablet form and in low concentrations, is effective in killing bacterial, viral and protozoal (except Cryptosporidium) pathogens ^{Reference 73}, ^{Reference 75}. Iodine is contraindicated in pregnant women and those with thyroid disease ^{Reference 74} and its use should be limited to periods of one month or less ^{Reference 73}.

Chlorine is available in a variety of formulations, both tablet and liquid, including simple household bleach and commercial preparations of chlorine dioxide. Each form releases free chlorine in water that effectively kills many enteric pathogens depending on concentration and contact time ^{Reference 72}. Halogen treated water may often be unpalatable. However, the taste can be improved by reducing the halogen concentration and increasing contact time proportionately. Alternatively, halogen treated water can be run through a filter that contains activated carbon or ascorbic acid crystals can be added after the required contact time has been achieved ^{Reference 62}. Water purified by methods which do not have residual disinfecting activity can become recontaminated during storage. Halogens have prolonged activity, and in higher concentrations, allows water to be stored safely for prolonged periods.

UV pens emit rays that can kill bacteria, viruses, protozoa and other parasites in clear water; however, they can be costly compared to other water purification methods and they do not work in turbid conditions (cloudy water) Reference 74. Solar water disinfection (SODIS) combines the effect of thermal heating of solar light with UV radiation to eliminate pathogens Reference 76. However, this method may not be practical for most travellers due to the time required to disinfect the water: up to 48 hours of solar exposure, depending on the intensity of sunlight available as well as the sensitivity of the pathogens.

The choice of water purification method will vary according to the traveller's itinerary and personal preferences. Long-term travellers may prefer to boil their water throughout their travels as filters have finite life-spans and chemically-treated water is often unpalatable. Most short-term travellers on business trips or resort holidays may prefer limiting themselves to commercially bottled beverages while campers may prefer portable water filters possibly combined with halogen treatment.

Prevention - Interventions

Probiotics, prebiotics and synbiotics

Due to various limitations explained below, CATMAT was unable to make a GRADE recommendation on the use of probiotics, prebiotics and synbiotics in the prevention of TD. An informal evaluation of the evidence is provided below.

Probiotics are live microbial food ingredients (i.e. certain types of living bacteria or yeast) that, when ingested in sufficient quantities, provide health benefits to the consumer ^{Reference 77}. Prebiotics are non-digestable food ingredients (i.e., certain types of dietary fiber) which provide health benefits by selectively stimulating growth of certain bacteria in the colon ^{Reference 78}. Synbiotics are products containing both probiotics and prebiotics. In Canada, probiotics, prebiotics and synbiotics are classified as natural health products (NHP) and, if reviewed by Health Canada, will be assigned a Natural Product Number (NPN) which is displayed on the product container. The Canadian regulations governing NHP are separate from the regulations governing prescription drugs. The more serious the health claim being made on the label, the higher the required level of evidence.

Several meta-analyses and reviews have evaluated the clinical effectiveness of probiotics in the prevention or treatment of symptoms for a variety of gastrointestinal diseases, such as lactose intolerance, irritable bowel syndrome, antibiotic-associated diarrhea, and those due to Helicobacter pylori and Clostridium difficile infections Reference 77, Reference 79-81. However, only a few randomized controlled studies have examined the use of probiotics in the prevention of TD with an outcome showing significant effects Reference 82-86; and only one of four meta-analyses showed a significant pooled effect Reference 79, Reference 87-89. Furthermore, it is difficult to interpret the findings because of the differing probiotic species, formulations and dosages used in the studies, and due to methodological problems within the studies themselves (i.e. poor compliance, recall bias). Of the various probiotic species studied, Saccharomyces boulardii Reference 84, Reference 85 and Lactobacillus rhamnosus GG Reference 82, Reference 83, both of which are stable at room temperature when lyophilized Reference 87, appeared to be the most promising for prevention of TD with no significant side effects.

Evidence related to the preventive effects of prebiotics and synbiotics for TD is also limited ^{Reference 90-92}. Variability in the study designs prevented comparisons as each study evaluated a different compound and used differing dosages and duration of treatment.

Prevention - Vaccination and Chemoprophylaxis

Vaccination - Oral cholera vaccine (Dukoral^®)

GRADE recommendation:

CATMAT suggests that the oral cholera vaccine (killed whole cells plus recombinant B-subunit, WC-rBS, licenced in Canada as Dukoral^{® Reference i}) not be routinely administered to Canadian travellers as a means of preventing TD; Conditional recommendation, moderate confidence in estimate of effect versus placebo.

Efficacy

The pooled results from three randomized controlled trials (RCTs) ^{Reference 94-96}, as well as results from one Cochrane review ^{Reference 97}, found no increased benefit for the oral cholera vaccine for preventing an episode of TD during travel when compared to those vaccinated with a placebo ^{Reference iii}. Overall, 35% of vaccinated subjects in the three studies developed TD versus 37% of non-vaccinated subjects, for a pooled RR of 0.94 and 95% confidence interval (95% CI) of 0.82 - 1.09. Additionally, these studies found no difference in effect for prevention of TD related to ETEC when compared to placebo. Lower confidence in estimate of effect can primarily be attributed to an ambiguous, potentially non-standard definition of TD in one study ^{Reference 94}, as well as a non-standard immunization protocol in one of the other studies ^{Reference 95}. Three observational studies ^{Reference 98-100} evaluating use of Dukoral^® for prevention of diarrhea in returned travellers have found a beneficial effect for those who had been vaccinated when compared to travellers visiting the same clinic who had not been vaccinated. Two other observational studies found no difference in effect ^{Reference 101}, ^{Reference 102}. However, these five observational studies were not included in the assessment due to serious limitations with the selection of the comparison group: those who were not vaccinated had either refused vaccination or were not referred for vaccination because they were judged to be engaging in types of travel at lower risk for cholera (and thus by extrapolation also potentially at lower risk for TD). In both cases, this presents important differences in risk profile between the vaccinated and non-vaccinated groups which quite probably biased the results.

Dukoral^® is licensed in Canada for prevention of and protection against TD and/or cholera in adults and children 2 years of age and older who will be visiting areas where there is a high risk of contracting TD caused by ETEC or cholera caused by V. cholerae. This indication is largely based on a field study conducted in an endemic population with a primary outcome of ETEC diarrhea ^{Reference 103}. This study was considered in the review of the evidence but was excluded from the analysis given that it was not conducted in a traveller population that is potentially exposed to a broad spectrum of TD-causing bacteria.

Although routine use of Dukoral^® for TD prevention is not recommended by CATMAT, certain selected short-term travellers at high risk for health complications or serious inconvenience from TD may find that the potential benefits of the vaccine based on their personal values and preferences, coupled with a low likelihood of adverse events (see section below), outweigh the burden of their risk. As such, the following travellers may still be considered for Dukoral^® vaccination:

• those for whom a brief illness cannot be tolerated (i.e., elite athletes, some business or political travellers);
• those with increased susceptibility to TD (e.g., due to achlorhydia, gastrectomy, history of repeated severe TD, young children > 2 years);
• those who are immunosuppressed due to HIV infection with depressed CD4 count or other immunodeficiency states;
• those with chronic illnesses for whom there is an increased risk of serious consequences from TD (e.g., chronic renal failure, congestive heart failure, insulin dependent diabetes mellitus, inflammatory bowel disease).

It should be noted that consideration of these groups is based on expert opinion and that there are no published data on Dukoral^® use in these specific groups.

Harms

We were unable to assess with GRADE the safety of the inactivated oral cholera vaccines due to insufficient detail provided on adverse events. No serious adverse reactions were recorded and no differences were observed between vaccine and placebo groups in each of the three RCTs, except for a slightly higher number of "gastrointestinal symptoms" in the placebo group of one study ^{Reference 94}.

Other vaccines

Some other vaccination interventions currently being developed include an oral ETEC-specific vaccine (killed ETEC whole cells plus recombinant cholera B-subunit), an ETEC LT subunit vaccine delivered by transcutaneous patch (LT patch), and a live attenuated oral cholera vaccine (CVD 103 HgR). A GRADE assessment of these interventions was not conducted since they are still in varying stages of clinical development and are not currently licenced in Canada. Two RCTs ^{Reference 96}, ^{Reference 104} found no increased benefit for the oral ETEC vaccine for preventing either an episode of all-cause TD or ETEC-associated TD during travel when compared to those vaccinated with a placebo. The LT patch ^{Reference 105}, ^{Reference 106} and the live oral cholera vaccine ^{Reference 107} were also evaluated in traveller populations: neither vaccine was found to increase benefit for prevention of TD as compared to placebo. A vaccine to target another pathogen responsible for TD, Shigella, is in early-stage human clinical trials, but cannot be evaluated at this time ^{Reference 108}.

Viral agents such as rotavirus can also cause TD in children. A live oral rotavirus vaccine is recommended by the National Advisory Committee on Immunization (NACI) for infants starting at 6 to15 weeks of age (see NACI guidelines for details and exceptions) ^{Reference 109}.

Antisecretory agents - Bismuth subsalicylate (BSS)

GRADE recommendation:

• CATMAT recommends that bismuth subsalicylate (BSS) be considered as an option for preventing TD for adults at significant risk, and who are willing to accept multiple doses per day (2.1-4.2g/day, divided in four doses per day); Strong recommendation, high confidence in estimate of effect versus placebo.
• CATMAT suggests that a lower dosage (1.05g/day) of BSS could be used to prevent TD in situations where a higher dosage is not feasible; Conditional recommendation, low confidence in estimate of effect versus placebo, low confidence there is no difference in effect between high and low dosage.

Efficacy

Four RCTs investigating the use of BSS versus placebo for the prevention of TD were evaluated with a GRADE assessment, of which three had an adequate definition for TD ^{Reference 110-112}. Overall, a strong protective effect after three to four weeks of follow-up was observed for BSS: RR = 0.55 (95% CI: 0.44 - 0.67), resulting in 250 fewer cases of TD per 1000 travellers treated. This strong effect was similarly found when restricted to those receiving a high or low dosage of BSS: RR = 0.51 (95% CI: 0.39 - 0.65) and RR = 0.65 (95% CI: 0.50 - 0.86) respectively. Similarly, there was no difference in effect found when comparing high to low dosage: RR = 0.87 (95% CI: 0.63 - 1.22). However, the results for low dosage only, as well as those for comparing high to low dosage, are of lower quality since they rely more heavily on the results of one study where there were low levels of compliance to therapy, and are limited in their ability to detect a true effect due to a lower number of subjects (imprecision). Although we were unable to assess with GRADE any differences in efficacy between liquid and tablet forms of BSS, the results do not appear to differ between the two delivery mechanisms.

Harms

We were unable to assess with GRADE the risk of developing side effects when using BSS versus placebo due to inconsistencies in reporting. The evidence does not appear to indicate any serious harm associated with BSS use. There is a probable increased risk for experiencing black tongue and black stool, although these side effects are not harmful ^{Reference 112}. There are also reports of increases in constipation in those taking BSS ^{Reference 111}, although this is not reported consistently across studies ^{Reference 110}. There did not appear to be a difference in risk of side effects between high and low dosages. Bismuth subsalicylate should be avoided by those allergic to aspirin and during pregnancy. Those taking other concurrent medications should check for possible interactions with BSS. Prophylactic BSS at these doses has not been studied for periods longer than four weeks. Prolonged use of BSS in children carries a risk of salicylate intoxication and bismuth encephalopathy, as well as a theoretical risk of Reye's syndrome ^{Reference 113}. Use of BSS is permitted in the case of certain children aged two years and older, based on an individual assessment of risks and benefits. BSS use is not recommended in children younger than two years old.

Antibiotics

In general, TD is a self-limited disorder, and routine use of antibiotics for prophylaxis may expose the traveller to risks which exceed those of the illness. These risks are often not well documented in studies, but theoretically would include an increased risk for carriage and infection with antibiotic resistant pathogens, antibiotic associated diarrhea and infection with Clostridium difficile, and other adverse reactions including hypersensitivity reactions, photosensitivity reactions, tendinopathy and cardiac arrhythmia. These types of adverse events are well documented when these antibiotics are used for other indications, although discussion of their frequency and severity is beyond the scope of this review. Therefore the risks and benefits of the use of antibiotics for diarrhea prophylaxis need to be carefully considered for each individual, and their use would not be warranted routinely. This is particularly true for children (1 to 17 years of age), since, in addition to the risks outlined above, there is a potential risk of cartilage damage and arthropathies associated with use of fluoroquinolones (see details below in discussion of harms associated with fluoroquinolones) ^{Reference 114}, ^{Reference 115}. Antibiotic chemoprophylaxis for TD in children should be limited to specific situations such as children with immunoglobulin A deficiency or other conditions known to significantly increase the risk and/or severity of TD. It should also be noted that TD studies in children are non-existent and that all discussion of antibiotic use in children assumes an efficacy similar to that observed in the adult study populations. For a summary of optimal doses for each antibiotic, please see Table 2.

Antibiotics - fluoroquinolones

GRADE recommendation:

• CATMAT suggests that fluoroquinolones be considered as an option in the prevention of TD in select high-risk short-term traveller populations where chemoprophylaxis is considered essential (see below for definition of this population); Conditional recommendation, high confidence in estimate of effect versus placebo. Balance of benefits and harms based on available evidence on adverse events and antimicrobial resistance patterns.

Efficacy

The results from four RCTs ^{Reference iv}, ^{Reference 118-121} demonstrate that use of fluoroquinolones over a period of five to 21 days provides a significantly and substantially decreased risk of developing TD: RR = 0.12 (95% CI: 0.07 - 0.21), resulting in 293 fewer cases per 1000 travellers treated. When individual fluoroquinolones were assessed (ciprofloxacin and norfloxacin), this strong effect persisted.

Harms

We were unable to assess with GRADE the evidence on adverse reactions with fluoroquinolones due to inconsistencies in reporting. However, the studies do not indicate any significant increase in serious adverse reactions in the fluoroquinolone group as compared to the placebo group. Two studies presented the possibility of adverse skin reactions from treatment: one participant reported a case of generalized skin rash Reference 118, while two other participants reported sunburn causing blisters Reference 116. It is unclear if these reactions were related to treatment. While these studies may appear inconclusive, there is evidence from fluoroquinolone use in non-traveller populations suggesting that adverse reactions present a potential risk to travellers. Although no studies have been done on the risks of Clostridium difficile to travellers using fluoroquinolones, their use in a clinical setting has been shown to significantly increase risk for C. difficile-associated diarrhea Reference 122, while there is also a rising concern about greater numbers of cases of C. difficile infection acquired in non-health care settings during travel Reference 123. Additionally, safety data collected from children (6 months to 16 years old; n=2,523) participating in one of three clinical trials evaluating the efficacy of levofloxacin Reference v for treating pneumonia or acute otitis media demonstrated a significantly increased risk of musculoskeletal adverse events (primarily due to reports of arthralgia) in children receiving levofloxacin as compared to non-fluoroquinolone antibiotics Reference 115. Finally, an increased risk of tendonitis and tendon rupture, particularly the Achilles tendon, has been observed for patients taking fluoroquinolones. Although this is a rare event, risk is greater for those 60 years of age or older, those using concomitant steroid therapy, as well as in kidney, heart and lung transplant recipients Reference 114. Due to the risk observed in non-traveller populations for adverse effects, which generally increases with duration of treatment, fluoroquinolones should only be considered as a prevention option in selected high-risk short-term travellers:

• those for whom a brief illness cannot be tolerated (i.e., elite athletes, some business or political travellers);
• those with increased susceptibility to TD (e.g., due to achlorhydia, gastrectomy, history of repeated severe TD);
• those who are immunosuppressed due to HIV infection with depressed CD4 count or other immunodeficiency states;
• those with chronic illnesses for whom there is an increased risk of serious consequences from TD (e.g., chronic renal failure, congestive heart failure, insulin dependent diabetes mellitus, inflammatory bowel disease).

Antimicrobial resistance

It is important to note that in the 20 to 25 years since studies were published evaluating the efficacy of fluoroquinolones to prevent TD, in vitro studies of returning travellers have found increased resistance of various TD-related pathogens to fluoroquinolones. Resistance levels range from one to 10% for E. coli pathogens (ETEC and EAEC) ^{Reference 124-127} and are much higher (71% to 84%) for Campylobacter strains tested from military personnel stationed in Thailand and travellers visiting Nepal ^{Reference 128}, ^{Reference 129}. One other study found no resistance to Salmonella for ciprofloxacin and norfloxacin ^{Reference 130}. Several studies also found elevated resistance to nalidixic acid treatment for several key pathogens ^{Reference 124}, ^{Reference 127}, ^{Reference 129}, ^{Reference 131}, with rates being particularly high in travellers to the Indian subcontinent (64% in one study, ^{Reference 131}). This is of particular concern since it has been demonstrated that nalidixic acid-resistant strains of pathogens are associated with treatment failure when using fluoroquinolones to treat TD ^{Reference 130}. The above-mentioned studies evaluated resistance in vitro and results may not necessarily correlate with actual clinical response to treatment. The antimicrobial resistance evidence therefore cannot be assessed using GRADE. However, the extent to which this in vitro fluoroquinolone antimicrobial resistance has been documented, specifically for the geographic regions of the Indian subcontinent (India and Nepal in particular) and Southeast Asia (Thailand in particular), leads CATMAT to suggest that alternative preventive measures be explored for high-risk travellers visiting these regions.

Antibiotics - azithromycin

We did not find any evidence to evaluate the use of azithromycin in the prevention of TD. Studies in several low and middle income countries evaluating the efficacy of campaigns to prevent trachoma through mass treatment with azithromycin noted a reduced risk for acute diarrhea in children 14 years of age and under when evaluated up to one month after treatment ^{Reference 132-134}. Azithromycin may be an acceptable choice for prevention in pediatric patients or in patients for whom fluoroquinolones are contra-indicated, when antibiotic prophylaxis is justifiable.

Antibiotics - rifaximin

GRADE recommendation:

• CATMAT suggests that rifaximin be considered as an option in the prevention of TD; Conditional recommendation, moderate confidence in estimate of effect versus placebo. Balance of benefits and harms based on available evidence on antimicrobial resistance patterns.

Efficacy

An assessment of five RCTs ^{Reference 135-139} found a strong protective effect against TD when rifaximin was administered for two to three weeks during travel as compared to placebo:^{Reference vi} RR = 0.42 (95% CI: 0.33 - 0.53), resulting in 213 fewer cases per 1000 travellers treated. The quality of the evidence was downgraded for potential publication bias due to the fact that results were unavailable for one large study (n=660) registered on the U.S. government's clinical trials database and completed in 2008 ^{Reference 141}.

Antimicrobial resistance

Evidence of antimicrobial resistance patterns associated with rifaximin use in travellers was difficult to assess. Contrary to fluoroquinolones, there are no established thresholds for resistance based on minimum inhibitory concentration (MIC). Evidence on in vitro activity in samples taken from travellers is mixed although generally favourable. One study ^{Reference 126} demonstrated that rifaximin showed intermediate activity against all pathogens evaluated versus high activity for the quinolones, whereas a more recent study ^{Reference 127} found rifaximin exhibited good activity against all pathogens. Another study ^{Reference 131} showed that rifaximin MIC levels for strains of ETEC and EAEC did not change between 1997 and 2008 whereas they did increase substantially for quinolones and azithromycin. However, in vitro testing on mechanisms of rifaximin resistance has revealed that high level resistance can be conferred on strains of ETEC and EAEC with a single step mutation. It appears that selection of resistance to rifaximin is easier than for other commonly used antibiotics ^{Reference 142}. More definitive study of antimicrobial resistance to rifaximin will need to be conducted before a more conclusive assessment of its long-term efficacy can be given.

Harms

We were unable to assess with GRADE the safety of rifaximin for prevention of TD due to inconsistencies in reporting on adverse events. However, all studies stated that there were no serious adverse events and no difference in number of adverse events between rifaximin and placebo groups. There are no data on the use of this agent in children (≤ 12 years old) and CATMAT therefore does not recommend the use of rifaximin in this age group.

Treatment

Antisecretory agents - Bismuth subsalicylate (BSS)

We were unable to make a GRADE assessment of BSS use for treatment. Nevertheless, there are some data which suggests that it is safe and moderately effective, and could be considered as an alternative to the other treatment agents discussed in this section. Of the four RCTs evaluating the efficacy of BSS compared to placebo in a traveller population, two had an inadequate definition of TD ^{Reference 143}. Of the remaining two studies ^{Reference 144}, ^{Reference 145} there was a lack of commonality in the outcomes assessed. The available evidence appears to indicate a beneficial effect: the two studies using a valid definition of TD found an increased association with absence of diarrhea after 24 hours ^{Reference 144} and cure from all TD symptoms after 48 hours ^{Reference 145}, as well as significant albeit modest reductions in duration of diarrhea and mean number of stools passed. We were unable to formally assess the safety of BSS. Two of the three studies made mention of adverse events, one of which indicated that black tongue and black stool were seen in high numbers ^{Reference 145}. However, there were no serious adverse reactions reported in any of these studies and no other significant difference in occurrence of events between treatment groups. Treatment with BSS is not recommended in children younger than 2 years old and is contraindicated in acetylsalicylic acid allergy ^{Reference 113}.

Antimotility agents - Loperamide

GRADE recommendation:

• CATMAT suggests that the antimotility agent loperamide be considered as an option in the treatment of TD; Conditional recommendation, low to moderate confidence in estimate of effect compared to placebo.

Efficacy

Three RCTs were identified which had an adequate definition of TD and evaluated the efficacy of loperamide compared to placebo in traveller populations generally using a three-day treatment regimen ^{Reference 146-148}. However, due to inconsistency across these studies in outcomes being assessed, we decided to increase the evidence base by including studies with non-traveller populations in our assessment ^{Reference 149-151}, although this requires us to downgrade the quality of the evidence for indirectness. Confidence in the estimate of effect was also lowered for three of the four outcomes assessed with GRADE due to an insufficient number of study subjects. Two studies ^{Reference 150}, ^{Reference 151} in the non-traveller population found that loperamide was associated with a significant increase in first relief from acute diarrhea after 4, 12 and 24 hours of treatment when compared to placebo: RR for first relief from acute diarrhea after 4 hours = 1.69 (95% CI: 1.17 - 2.45), resulting in 145 more cases of first relief after 4 hours per 1000 travellers treated. Similarly, two studies (of which one was in the traveller population) also observed this beneficial effect for complete relief of acute diarrhea after 24 hours ^{Reference 150} and 48 hours ^{Reference 148}. The evidence on reduction of duration of diarrhea, however, is mixed with two studies ^{Reference 147}, ^{Reference 150} showing a significantly reduced time to complete relief from diarrhea of approximately 18 to 24 hours as compared to placebo, while two other studies ^{Reference 148}, ^{Reference 149} found non-significant reductions in duration for loperamide. Finally, there was evidence to support a small but significant reduction in the intensity of diarrhea: an average of 1.6 fewer stools during the first 24 hours of treatment ^{Reference 146}, ^{Reference 149} and an average of 2.3 fewer unformed stools after 48 hours of treatment ^{Reference 148}.

We were unable to assess with GRADE the efficacy of loperamide as compared to BSS due to a limited number of studies, combined with a lack of commonality in outcomes assessed. However, two studies ^{Reference 152}, ^{Reference 153} with an adequate definition of TD did compare these two therapies directly in the traveller population and the results appear to indicate an advantage for loperamide. One of the studies ^{Reference 153} evaluated duration of diarrhea and found that loperamide significantly reduced the mean time to last unformed stool by approximately 10 hours when compared to BSS. The same study also found that loperamide treatment was significantly associated with no further dose needed after 24 hours. Finally, both studies evaluated intensity of diarrhea and found persons receiving loperamide experienced a small but significant reduction in average number of stools during various time periods within the first 24 hours of treatment when compared to those receiving BSS.

Loperamide use in travelling children has not been studied. However, one RCT conducted in children aged two to 11 with acute diarrhea ^{Reference 154} found that loperamide treatment significantly reduced duration and severity with no difference between loperamide and placebo treatment groups with respect to drug-related adverse events. Dosages differ by age group (see Table 2) and treatment should not exceed two days. Loperamide should not be administered to children under two years of age ^{Reference 113}.

Harms

We were unable to formally assess the safety of loperamide use due to insufficient detail provided on adverse events. However, all of the relevant studies mention that there were no significant differences in adverse events between study populations and that no serious adverse events were recorded.

A small study suggests an increase in adverse events with the use of diphenoxylate (Lomotil, an agent related to loperamide) for treatment of shigella infection ^{Reference 155}. Lomotil has a less favourable side effect profile, and it has not been studied in the treatment of TD.

Loperamide in combination with antibiotics

GRADE recommendation:

• CATMAT suggests that the use of the antimotility agent loperamide in conjunction with antibiotic therapy be considered as an option in the treatment of TD; Conditional recommendation, moderate to high confidence in estimate of effect compared to antibiotic use alone.

Efficacy

Six RCTs^{Reference vii} assessed various outcomes related to this intervention, of which five used an adequate definition of TD ^{Reference 147}, ^{Reference 156-159}. One meta-analysis evaluating this intervention was also consulted ^{Reference 160}. Results from four studies ^{Reference 147}, ^{Reference 156}, ^{Reference 158}, ^{Reference 159} evaluating cure rates found that loperamide used in combination with an antibiotic was significantly associated with a greater cure after 24 hours and 48 hours of therapy when compared to antibiotic use alone: RR for complete relief from TD after 24 hours=1.55 (95% CI: 1.28 - 1.86), resulting in 200 more cases of complete relief after 24 hours per 1000 travellers treated. These same four studies also evaluated treatment failures and found that adding loperamide to antibiotic therapy significantly reduced the risk of a failure. Estimates of effect for two of the four outcomes were rated down due to substantial variation between studies in the observed direction of effect (inconsistency). Given the relatively mild nature of most episodes of TD, and the acceptable efficacy of antibiotics or loperamide alone, it is reasonable to reserve the combination of the two for treatment of severe diarrhea and/or when treatment with antimotility or antibiotic alone is unsuccessful.

Harms

We were unable to formally assess safety of adjunctive loperamide use due to insufficient detail provided on adverse events. However, all of the relevant studies mention that there were no significant differences in adverse events between study populations and that no serious adverse events were recorded.

Antibiotics - fluoroquinolones

GRADE recommendation:

• CATMAT suggests that fluoroquinolones be considered as an option in the treatment of TD; Conditional recommendation, moderate confidence in estimate of effect versus placebo. Balance of benefits and harms based on available evidence on adverse events and antimicrobial resistance patterns.

Efficacy

Nine RCTs assessed fluoroquinolone for treatment of TD versus placebo, of which six used an adequate definition of TD ^{Reference 161-166}. One Cochrane review evaluating this intervention was also consulted ^{Reference 167}. Three of these studies evaluated cure rate ^{Reference 162}, ^{Reference 163}, ^{Reference 165}, of which two ^{Reference 162}, ^{Reference 165} found an overall increased cure after 72 hours for norfloxacin when compared to placebo: RR = 1.81 (95% CI: 1.39 - 2.37), resulting in 322 more cases of cure after 72 hours per 1000 travellers treated. The other study found ofloxacin use increased cure after 48 hours and after five days using either three-day or five-day treatment regimens, while the three-day regimen also increased cure after 24 hours when compared to placebo ^{Reference 163}. The estimate of effect was rated down due to imprecision. Although we were unable to formally assess reductions in diarrhea duration, four studies ^{Reference 161}, ^{Reference 163}, ^{Reference 165}, ^{Reference 166} found significant reductions in time to last unformed stool in the treatment group, ranging from 28 to 52 hours faster than those taking placebo.

Harms

The evidence from the studies evaluated does not appear to indicate any serious harm associated with fluoroquinolone use. Three of the nine studies provided sufficient results to assess the safety of fluoroquinolone use ^{Reference 163}, ^{Reference 166}, ^{Reference 168} and the results indicate that there is a possible increased risk of adverse event for those taking fluoroquinolones compared to placebo: RR = 1.39 (95% CI: 1.05 - 1.83), resulting in 80 more travellers with some sort of adverse event per 1000 travellers treated. Some of the more common adverse events reported include headaches, constipation, nausea and fatigue, although there is no clear evidence of greater risk of developing any individual symptom for fluoroquinolone users. Indeed, the quality of this evidence is very low, most notably due to the lack of standardized reporting of adverse events across studies and inconsistency in estimates of effect among the three studies. Please also refer to the Prevention section of this statement for theoretical risks of fluoroquinolone use and C. difficile infection, as well as risks for cartilage damage and arthropathies.

Antimicrobial resistance

Fluoroquinolones should be used with caution in situations where elevated levels of resistance have been documented (see section on chemoprophylaxis).

Antibiotics - azithromycin

GRADE recommendation:

• CATMAT suggests that azithromycin be considered as an option in the treatment of TD; Conditional recommendation, low confidence in estimate of effect versus fluoroquinolone use. Balance of benefits and harms based on available evidence on antimicrobial resistance patterns and adverse events.

Efficacy

No studies were found which evaluated the efficacy of azithromycin as compared to placebo in traveller populations. There were five RCTs which directly compared the efficacy of azithromycin to fluoroquinolones ^{Reference viii} in the treatment of TD, of which four had an adequate definition of TD Reference 169-172. For three of the outcomes of interest, no difference in efficacy was found between the two treatment groups: recovery from TD after 24 hours (RR = 0.79, 95% CI: 0.61 - 1.01); recovery after 72 hours (RR = 1.16, 95% CI: 1.00 - 1.33); and treatment failure (RR = 1.02, 95% CI: 0.45 - 2.32). However, for rapid or immediate cure from TD, azithromycin was associated with a reduced effect compared to fluoroquinolones: RR = 0.46 (95% CI: 0.25 - 0.84). Conversely, for cure from TD after 48 hours of treatment, azithromycin was associated with a greater effect compared to fluoroquinolones: RR = 1.34 (95% CI: 1.08 - 1.66). Although we were unable to assess duration of diarrhea using GRADE, three of the four studies providing some information on time to last unformed stool Reference 170-172 demonstrated no difference in duration between azithromycin and fluoroquinolones. Taken together, these results suggest that azithromycin's ability to provide relief from TD is equivalent to that of fluoroquinolones. However, confidence in the estimate of effect is low due to various factors including: insufficient number of events for certain outcomes (imprecision); variability in results between each study (inconsistency); and differences between studies in terms of dosages and use of loperamide as an adjunct therapy (indirectness).

Antimicrobial resistance

A search of the literature on antimicrobial resistance patterns in azithromycin use for TD was inconclusive. One study demonstrated that azithromycin exhibited high activity against all TD-related pathogens ^{Reference 126}, while another demonstrated that the concentrations of this antibiotic needed to inhibit travel-related ETEC and EAEC have been increasing since the late 1990s ^{Reference 173}. Although azithromycin is recommended as an alternative to fluoroquinolones in Southeast Asia due to resistance patterns observed in that region, two studies have demonstrated relatively elevated levels of resistance in Campylobacter in both travellers to Nepal ^{Reference 129} and U.S. military stationed in Thailand ^{Reference 128}. On the other hand, results from in vitro studies have not been proven to predict a failed clinical outcome.

Harms

The evidence does not appear to indicate any serious harm associated with use of azithromycin. Two studies assessed specific adverse events (nausea and vomiting) in these two treatment populations ^{Reference 171}, ^{Reference 172} and found a greater risk of nausea immediately following first treatment with azithromycin: RR = 6.23 (95% CI: 1.48 - 26.26), resulting in 68 more travellers with nausea in the first 30 minutes of treatment per 1000 treated with azithromycin as compared to those treated with fluoroquinolones. Otherwise, there were no differences in other measures of nausea and vomiting. The quality of this evidence was low, primarily due to imprecision related to small study population and number of events, as well as the fact that one of the four studies was less comparable since it included loperamide as an adjunct to therapy.

Antibiotics - rifaximin

GRADE recommendation:

• CATMAT suggests that rifaximin be considered as an option in the treatment of TD; Conditional recommendation, high confidence in estimate of effect versus placebo, moderate to high confidence in estimate of effect versus ciprofloxacin. Balance of benefits and harms based on available evidence on antimicrobial resistance patterns.

Efficacy

Two relatively recent RCTs ^{Reference 166}, ^{Reference 174} evaluated the efficacy of rifaximin as compared to placebo in traveller populations. Rifaximin was associated with a higher percentage of travellers cured of TD (RR = 1.29, 95% CI: 1.15 - 1.45, resulting in 177 more travellers cured of TD at the end of follow-up per 1000 treated) and reduced the risk of treatment failure (RR = 0.50, 95% CI: 0.38 - 0.67) compared to placebo. Two RCTs also directly compared the efficacy of rifaximin to that of the fluoroquinolone ciprofloxacin ^{Reference 166}, ^{Reference 175}. There was no significant difference between rifaximin and fluoroquinolones with respect to proportion cured of TD (RR = 0.98, 95% CI: 0.90 - 1.07), or treatment failure (RR = 1.81, 95% CI: 0.96 - 3.43). The estimate of effect for treatment failure was rated down due to imprecision. These results suggest that treatment of TD with rifaximin has the same efficacy as treatment with fluoroquinolones.

Antimicrobial resistance

Antimicrobial resistance issues were still difficult to assess at the time of this writing and will need to be closely monitored. Please see the section on chemoprophylaxis with rifaximin for discussion of this subject.

Harms

Rifaximin appeared to be safe, with no difference reported in the proportion of adverse events between treatment and placebo groups (RR = 0.96, 95% CI: 0.83 - 1.11), nor between the rifaximin and fluoroquinolone treatment groups (RR = 1.01, 95% CI: 0.76 - 1.35). There are no data on the use of this agent in children (≤ 12 years old) and CATMAT therefore does not recommend the use of rifaximin in this age group.

Antibiotic treatment: conclusions and other considerations

As a general rule, antibiotic use for treatment should be limited as much as possible due to the adverse events and antimicrobial resistance patterns discussed above. If symptoms from TD are mild, the preferred mode of treatment should be oral rehydration and loperamide (or BSS). However, should this line of treatment fail or more serious symptoms be present in the traveller, antibiotic use may be justified. For more severe cases of TD and whenever feasible, it is also advisable to obtain culture and antibiotic sensitivity for known pathogens in order to facilitate optimal choice of treatment regimen.

One limitation of assessing efficacy and harm of antibiotic use for treatment of TD in children is the lack of studies evaluating this age group. As such, all discussion of antibiotic use in children assumes an efficacy similar to that observed in the adult study populations. Some specific clinical observations on children, however, may be made. Children under the age of 18 should not be administered fluoroquinolones for treating TD unless the benefits are felt to outweigh the potential risks and other alternatives are not feasible. Otherwise, azithromycin should be used in this age group or cefixime if azithromycin is contraindicated ^{Reference 113}. There is evidence that cefixime is efficacious against several of the pathogens which cause TD. However, an increase in cases of antibiotic-associated colitis has been noted in at least one pediatric population after treatment for profuse diarrhea with cefixime ^{Reference 176}. Finally, while reports exist of an increased risk for developing hemolytic-uremic syndrome (HUS) in children given antibiotics for treatment of enterohemorrhagic Escherichia coli (EHEC) ^{Reference 177}, this should not be a major consideration in the decision to treat TD empirically with antibiotics. EHEC is a pathogen primarily found in children in higher income countries and is rarely seen in TD studies ^{Reference 10}. Similarly, there is an absence of HUS cases in the TD literature, suggesting that this is a very rare complication amongst travellers.

Managing TD symptoms - rehydration

Oral replacement of fluid levels and electrolytes is of primary importance in managing any case of TD, and most existing recommendations are based on treatment of acute diarrhea. Children, particularly those two years old and younger, are at high risk for dehydration. Many oral rehydration solutions (ORS) formulas are available at pharmacies both in Canada and overseas, while solutions can also be concocted at home (see Table 3 for more information on preparing ORS). The traditional ORS, however, does not reduce diarrhea duration or severity. Since 2002, the WHO has been recommending the use of a reduced osmolarity ORS which does reduce diarrhea symptoms ^{Reference 178}. Reduced osmolarity ORS can be approximated by diluting two parts standard ORS with one part boiled or treated water ^{Reference 179}. Commercial or WHO-type ORS products are commonly used at a dose of 50 mL/kg for mild dehydration (3-5% body weight loss). However, parents often have difficulty properly assessing a child's hydration status, and should be strongly advised to seek medical attention when any significant dehydration is suspected ^{Reference viv}, ^{Reference 179}. Evidence from clinical studies evaluating a variety of dietary options demonstrate that an unrestricted diet initiated early on in the rehydration process has no negative effects on the course or symptoms of diarrhea when compared to ORT alone ^{Reference 180}. Since early feeding does not appear to be detrimental and is beneficial in terms of improved nutrition, a regular age-appropriate diet should be re-established at an early stage in rehydration. This includes any type of milk (full-strength, human, animal, containing lactose, etc.) and most foods. However, although evidence is lacking, it is suggested that foods high in fat and simple sugars should be avoided due to a tendency of fats to delay gastric emptying and the ability of simple sugars to exacerbate diarrhea through osmotic effects ^{Reference 180}. Fluids should be consumed at a rate to allay thirst and maintain pale-coloured urine.

Treatment of TD upon return from travel

Any febrile traveller with diarrhea who has visited a malaria endemic area must have blood films performed immediately to rule out malaria. Patients with severe TD not responding to empiric therapy and those with severe underlying medical conditions, immunosuppression, or grossly bloody stools should be referred to a specialist for further evaluation. Travellers with persistent diarrhea lasting more than 14 days, despite therapy, should be managed according to the CATMAT statement of persistent diarrhea in the returned traveller ^{Reference 3}.

Values and Preferences

Recommendations made using GRADE need to take into account the values and preferences of the patient for each of the treatment and prevention options. Unfortunately, there is limited information available for these preferences among the travelling population. One study did assess travellers' willingness to take antibiotic chemoprophylaxis as well as various treatment regimens for TD ^{Reference 181}. A relatively high percentage of travellers indicated they preferred not taking antibiotics for prevention of TD. There was also high variability between North American and European respondents with respect to their preferences for antidiarrheal treatment as well as their ability to correctly assess their level of TD risk for their chosen country of travel. These factors serve to reinforce the conditional nature of CATMAT recommendations for these preventive and treatment interventions.

Conclusions and Research Needs

With the exception of BSS for prevention of TD (strong recommendation for use), CATMAT conditionally recommends the use of each of the other GRADE-evaluated preventive and therapeutic products assessed in this statement. These recommendations are conditional due to: demonstrated weak effects, weakness in the evidence base for a given intervention and/or the uncertain weight which should be accorded to potential harms of the intervention. For this latter point, one of the potential harms lies in the use of antibiotics which may select for carriage of resistant pathogens by the host. This in turn could lead to an ill traveller being treated for TD (or another infection) with ineffective antibiotics. Although this risk has been well-demonstrated in other domains, we have no reliable evidence on the presence or magnitude of the risk in the case of TD. As such, CATMAT recommends that more systematic surveillance and research be undertaken on resistance patterns of pathogens in the returned traveller who has taken a course of antibiotics to prevent or treat TD. This information will serve to improve assessment of baseline risk for resistance based on destination and type of travel. Although CATMAT had moderate confidence in the available evidence to conditionally recommend against routine use of the oral cholera vaccine (Dukoral^®) for prevention of TD, further research evaluating the efficacy of this vaccine to prevent TD would be necessary in order to make a more definitive recommendation for or against its use in specific populations. Of equal importance will be the systematic review of studies evaluating the efficacy of other vaccines currently in development which target TD-related pathogens, most notably ETEC vaccines. Finally, CATMAT also recommends further research on specific species of probiotics, compounds of prebiotics or combinations of the two (i.e., synbiotics) in order to better evaluate their efficacy in preventing TD.

Acknowledgements

This statement was developed by the Travellers' Diarrhea Working Group: Libman M (Chair), Bui Y, McDonald P, Steensma C.

CATMAT Members: McCarthy A (Chair), Boggild A, Brophy J, Bui Y, Crockett M, Ghesquiere W, Greenaway C, Henteleff A, Libman M, Teitelbaum P.

Liaison members: Hui C (Canadian Paediatric Society), Gershman M (US Centers for Disease Control and Prevention).

Ex-officio members: McDonald P (Division of Anti-Infective Drugs, Health Canada), Tepper M (Directorate of Force Health Protection, Department of National Defence), Charlebois P (Canadian Forces Health Services Centre, Department of National Defence), Schofield S (Pest Management Entomology, Department of National Defence).

Member Emeritus: Jeane C.W.L.

CATMAT gracefully acknowledges the contribution of Francesca Reyes-Domingo for her assistance in developing this statement

Tables

Table 1 - List of studies^{Table 1 - Footnote *} considered for inclusion in analysis of efficacy of TD prevention and treatment

Author / Year	Study Population	Treatment and Dosage / Comparison	Follow-up	Outcome Relative Risk (95%CI)	Comments
Oral cholera vaccine (Dukoral®) for prevention of TD
Peltola 1991	Finnish travellers to Morocco, age ≥15yrs n=615	WC-BS vaccine, 2 doses given 3 to 21 days before departure vs. placebo	Not indicated	Occurrence of TD: 0.77 (0.59, 1.00)	Included in analysis Definition of TD: unclear if quantity of loose stools assessed.
Scerpella 1995	U.S. college students in Mexico, age ≥18yrs n=502	WC-rBS vaccine, 2 doses given upon arrival and 10 days later vs. placebo	35 days from first dose	Occurrence of TD: 1.04 (0.87, 1.24)	Included in analysis Vaccine administered post-arrival.
Wiedermann 2000	Travellers to tropical and subtropical destinations, adults and children n=125	WC-rBS, 2 doses given 7 to 30 days before departure vs. placebo	Upon return (range 7-23 days)	Occurrence of TD: 1.28 (0.68, 2.39)	Included in analysis Definition of TD: did not include an enteric symptom.
Clemens 1988	Endemic population of women and children in rural Bangladesh n=49,612	WC-BS vaccine, 3 doses given 6 weeks apart vs. WC vaccine only and placebo	14-365 days	Occurrence of TD: 0.33 (0.13, 0.84)	Excluded from analysis Study conducted in non-traveller population.
Lopez-Gigosos 2007 (observational study)	Spanish travellers to high-risk cholera regions, adults (mean age=35yrs) n=237	WC-rBS, doses and schedule not given vs. travellers at same clinic prior to vaccine availability or refused vaccine	Upon return (range 7-134 days)	Occurrence of TD: 0.57 (0.38, 0.85)	Excluded from analysis Retrospective cohort study. Risk profile of comparison group differs from that of vaccinated group.
Lopez-Gigosos 2009 (observational study)	Spanish travellers to high-risk cholera regions, adults n=362	WC-rBS, doses and schedule not given vs. travellers at same clinic prior to vaccine availability or refused vaccine	Upon return (range 14-154 days)	Occurrence of TD: 0.57 (0.41, 0.81)	Excluded from analysis Retrospective cohort study. Risk profile of comparison group differs from that of vaccinated group.
Torrell 2009 (observational study)	Spanish adventure travellers at high risk for TD, age 18-35yrs n=658	WC-rBS, 2 doses given minimum 7 to 14 days before departure vs. travellers at same clinic prior to vaccine availability or refused vaccine	30-90 days after return	Occurrence of TD: 0.44 (0.33, 0.58)	Excluded from analysis Retrospective cohort study. Risk profile of comparison group differs from that of vaccinated group.
Gabutti 2012 (observational study)	Italian travellers to areas at risk for cholera and TD, adults (mean age=38yrs) n=296	WC-rBS, 2 doses given minimum 7 to 14 days before departure vs. unvaccinated (no further details)	Upon return (no further details)	Occurrence of TD: 0.68 (0.41, 1.15)	Excluded from analysis Retrospective cohort study. Risk profile of comparison group differs from that of vaccinated group.
Lopez-Gigosos 2013 (observational study)	Spanish travellers to high-risk cholera regions, adults (mean age=35yrs) n=1074	WC-rBS, 2 doses given minimum 7 to 14 days before departure vs. travellers at same clinic judged as low-risk for cholera and travelling to same regions as vaccinated group	7-14 days after return	Occurrence of TD: 0.91 (0.78, 1.07)	Excluded from analysis Prospective cohort study. Risk profile of comparison group differs from that of vaccinated group.
Bismuth subsalicylate (BSS) for the prevention of TD
DuPont 1980	U.S. college students in Mexico, age not given n=128	4.2g BSS/day (liquid), four doses of 1.05g vs. placebo	21 days	Occurrence of TD: 0.37 (0.23, 0.61)	Included in analysis
Steffen 1986	Swiss travellers to tropical countries, age 16- 70yrs a) n=231 b) n=160 c) n=143 d) n=156	2.1g or 1.05g BSS/day (tablet) vs. placebo 2.1g BSS/day (tablet), two doses of 1.05g vs. placebo 1.05g BSS/day (tablet), two doses of 525mg vs. placebo 2.1g BSS/day vs. 1.05g BSS/day	30 days	a) Occurrence of TD: 0.68 (0.53, 0.87) b) Occurrence of TD: 0.68 (0.50, 0.91) c) Occurrence of TD: 0.68 (0.50, 0.94) d) Occurrence of TD: 0.99 (0.61, 1.41)	Included in analysis 21% loss to follow up 27% poor compliance to treatment.
DuPont 1987	U.S. college students in Mexico, age ≥19yrs a) n=172 b) n=109 c) n=121 d) n=114	2.1g or 1.05g BSS/day (tablet) vs. placebo 2.1g BSS/day (tablet), four doses of 524mg vs. placebo 1.05g BSS/day (tablet), four doses of 262mg vs. placebo 2.1g BSS/day vs. 1.05g BSS/day	21 days	a) Occurrence of TD: 0.49 (0.30, 0.80) b) Occurrence of TD: 0.35 (0.16, 0.74) c) Occurrence of TD: 0.60 (0.35, 1.03) d) Occurrence of TD: 0.58 (0.25, 1.31)	Included in analysis
Graham 1983	In-patient, healthy young adults, age not given n=31	2.1g BSS/day (tablet), four doses of 600mg vs. placebo	4 days	Occurrence of TD: 0.24 (0.06, 0.92)	Excluded from analysis Number of unformed stools less than classic TD definition.
Fluoroquinolones for the prevention of TD
Johnson 1986	U.S. college students in Mexico, age not given n=120	Norfloxacin 400mg once daily for 14 days vs. placebo	21 days	Occurrence of TD: 0.12 (0.04, 0.31)	Included in analysis
Wiström 1987	Swedes travelling outside of Northern Europe, mean age=45yrs n=127	Norfloxacin 200mg twice daily for 5-21 days vs. placebo	Up to 23 days	Occurrence of TD: 0.16 (0.04, 0.69)	Included in analysis Association persists when stratified by area of travel (high vs. low risk).
Scott 1990	U.S. military personnel in Egypt, mean age=26yrs n=262	Norfloxacin 400mg once daily for 7 days vs. placebo	11 days	Occurrence of TD: 0.07 (0.02, 0.30)	Included in analysis
Heck 1994	U.S. volunteers in Honduras, age 18-70yrs n=230	Ciprofloxacin 500mg once daily for 15 days vs. placebo	20 days	Occurrence of TD: 0.15 (0.06, 0.38)	Included in analysis
Rademaker 1989	Dutch travellers to Tunisia, mean age=28yrs n=54	Ciprofloxacin 500mg once daily for 7 days vs. placebo	10 days	Occurrence of TD: 0.06 (0.01, 0.42)	Excluded from analysis Number of unformed stools less than classic TD definition.
Rifaximin for prevention of TD
DuPont 2005	U.S. college students in Mexico, age not given n=219	Rifaximin 200mg once daily for 14 days vs. placebo Rifaximin 200mg twice daily for 14 days vs. placebo Rifaximin 200mg three times daily for 14 days vs. placebo	14 days	Occurrence of TD (all treatment arms combined): 0.27 (0.17, 0.43) Occurrence of TD (600mg arm only): 0.24 (0.12, 0.50)	Included in analysis
Armstrong 2010	U.S. military personnel in Turkey, median age=36yrs n=100	Rifaximin 1100mg once daily for 14 days vs. placebo	14 days	Occurrence of TD: 0.33 (0.09, 1.13)	Included in analysis Possible lack of concealment of treatment allocation.
Martinez-Sandoval 2010	U.S. college students in Mexico, age not given n=210	Rifaximin 600mg once daily for 14 days vs. placebo	14 days	Occurrence of TD: 0.32 (0.19, 0.54)	Included in analysis
Flores 2011	U.S. college students in Mexico, mean age=25yrs n=101	Rifaximin 550mg once daily for 14 days vs. placebo	14 days	Occurrence of TD: 0.72 (0.27, 1.92)	Included in analysis
Zanger 2013	German travellers to south and southeast Asia, 18-64yrs n=258	Rifaximin 200mg twice daily for 6- 28 days (duration of travel) vs. placebo	6-28 days	Occurrence of TD: 0.63 (0.42, 0.96)	Included in analysis
Bismuth subsalicylate (BSS) for treatment of TD
DuPont 1977	U.S. college students in Mexico, age not given n=137	4.2g BSS (liquid), 525mg every half-hour for 3.5 hours or 8.4g BSS (liquid), 1.05g every half-hour for 3.5 hours vs. placebo	24 hours	Presence of diarrhea after 24hrs (treatment arms combined): 0.39 (0.21, 0.76)	Included in review, excluded from analysis Some Latin American students included in study group.
Steffen 1988a	U.S. college students in Mexico, age not given n=112	4.2g BSS/day (liquid), 525mg every half-hour max. 8 doses, for two days vs. placebo	72 hours	Complete cure after 48hrs: 1.33 (0.86, 2.04)	Included in review, excluded from analysis Some Latin American students included in study group.
Steffen 1988b	European travellers to West Africa, age not given n=133	4.2g BSS/day (liquid), 1.05g every hour max. 4 doses, for two days vs. placebo	48 hours	Complete cure after 48hrs: 1.82 (1.11, 2.99)	Excluded from analysis Number of unformed stools less than classic TD definition.
Steffen 1988c	European travellers to developing countries, mean age=36yrs n=830	4.2g BSS (liquid), 525mg every half-hour for 3.5 hours, for two days vs. placebo	48 hours	Complete cure after 48hrs: 1.38 (1.09, 1.75)	Excluded from analysis Number of unformed stools less than classic TD definition. Large loss to follow-up. Poor compliance to treatment.
Loperamide for treatment of TD
Hughes 1995	British patients in general practice, aged 18-75yrs n=202	2mg loperamide plus 2mg after each loose stool, max. 16mg, for three days vs. placebo	72 hours	First relief of diarrhea after 4hrs: 1.55 (0.98, 2.45) First relief of diarrhea after 12hrs: 1.51 (1.05, 2.18) First relief of diarrhea after 24hrs: 1.38 (1.06, 1.79) Complete relief of diarrhea after 24hrs: 1.48 (1.05, 2.10)	Included in analysis Not a traveller population.
Steffen 1988c	European travellers to developing countries, mean age=36yrs n=800	4mg loperamide plus 2mg after each loose stool, max. 8mg, for two days vs. placebo	48 hours	Complete relief of diarrhea after 24hrs: 1.74 (1.31, 2.31)	Excluded from analysis Number of unformed stools less than classic TD definition. Large loss to follow-up. Poor compliance to treatment.
Van Loon 1989	Expatriates living in Bangladesh, mean age=36yrs n=50	4mg loperamide plus 2mg after each loose stool, max. 16mg, for two days vs. placebo	5 days	Mean number of stools after 24hrs (mean difference): -1.40 (-3.38, 0.58)	Included in analysis
Bergström 1986	Swedish outpatients, median age=32yrs n=112	4mg loperamide plus 2mg after each loose stool, max. 16mg, for two days vs. placebo	5 days	Mean number of stools after 24hrs (mean difference): -1.75 (-3.55, 0.05)	Included in analysis Not a traveller population.
DuPont 1990	U.S. college students in Mexico, mean age=26yrs n=203	4mg loperamide plus 2mg after each loose stool, max. 8mg, for two days vs. 4.9g BSS/day, 612.5mg every 30-60min max. 8 doses, for two days	48 hours	No additional doses needed after 24hrs: 1.75 (1.35, 2.28)	Included in review, excluded from analysis No blinding (open label study).
Johnson 1986	U.S. students in Latin America, age not given n=156	4mg loperamide plus 2mg after each loose stool, max. 16mg, for two days vs. 30mL BSS each half-hour for 3.5 hours, for two days	48 hours	Median number of stools after 4 hours (median difference): -0.5 (cannot calculate 95%CI: p<0.004)	Included in review, excluded from analysis No mean and standard deviation provided.
Steffen 1988c	European travellers to developing countries, mean age=36yrs n=800	4mg loperamide plus 2mg after each loose stool, max. 8mg, for two days vs. placebo	48 hours	Complete relief of diarrhea after 24hrs: 1.74 (1.31, 2.31)	Excluded from analysis Number of unformed stools less than classic TD definition. Large loss to follow-up. Poor compliance to treatment.
Loperamide combined with antibiotics for the treatment of TD
Ericsson 1990	U.S. college students in Mexico, mean age=24yrs n=92	TMP/SMX 160mg/800mg, twice daily and loperamide 4mg plus 2mg after each loose stool max. 16mg vs. loperamide only (same dosage) plus placebo, 3 days	5 days	Clinical cure at 24hrs: 1.97 (1.34, 2.90) Clinical cure at 48hrs: 1.24 (0.98, 1.55) Clinical cure at 72hrs: 1.18 (1.03, 1.35) Treatment failure: 0.33 (0.04, 3.04)	Included in analysis
Petruccelli 1992	U.S. military personnel in Thailand, age not given n=97	Ciprofloxacin 750mg, single dose and loperamide 4mg plus 2mg after each loose stool max. 16mg vs. loperamide only (same dosage) plus placebo, 3 days	72 hours	Clinical cure at 24hrs: 1.06 (0.84, 1.34) Clinical cure at 48hrs: 1.10 (0.83, 1.45) Clinical cure at 72hrs: 0.86 (0.75, 1.00)	Included in analysis
Ericsson 1997	U.S. college students in Mexico, mean age=27yrs n=110	Ofloxacin 400mg, single dose and loperamide 4mg plus 2mg after each loose stool max. 16mg vs. loperamide only (same dosage) plus placebo, 3 days	5 days	Treatment failure: 0.12 (0.01, 2.09)	Included in analysis Single blind study.
DuPont 2007	U.S. college students in Mexico, mean age=26yrs n=206	Rifaximin 200mg three times daily and loperamide 4mg plus 2mg after each loose stool max. 16mg vs. loperamide only (same dosage) plus placebo, 2 days	5 days	Clinical cure at 24hrs: 1.79 (1.17, 2.75) Clinical cure at 48hrs: 1.30 (0.99, 1.70) Clinical cure at 72hrs: 1.10 (0.91, 1.34)	Included in analysis
Ericsson 2007	U.S. college students in Mexico, mean age=23yrs n=112	Azithromycin 500mg, single dose and loperamide 4mg plus 2mg after each loose stool max. 16mg vs. loperamide only (same dosage) plus placebo, 2 days	4 days	Clinical cure at 24hrs: 1.42 (1.11, 1.82) Clinical cure at 48hrs: 1.27 (1.07, 1.51) Clinical cure at 72hrs: 1.23 (1.06, 1.42) Treatment failure: 0.17 (0.04, 0.71)	Included in analysis
Taylor 1991	U.S. military personnel in Egypt, age not given n=97	Ciprofloxacin 500mg, twice daily and loperamide 4mg plus 2mg after each loose stool max. 16mg vs. loperamide only (same dosage) plus placebo, 3 days	72 hours	Clinical cure at 24hrs: 1.23 (0.98, 1.55) Clinical cure at 48hrs: (0.81, 1.16)	Excluded from analysis Number of unformed stools less than classic TD definition.
Fluoroquinolones for the treatment of TD
Wiström 1989	Swedish travellers to developing countries, mean age=38yrs n=106	Norfloxacin 400mg twice daily for three days vs. placebo	72 hours	Cured of TD at 72hrs: 1.97 (1.32, 2.95)	Included in analysis
Mattila 1993	Finnish travellers to Morocco, mean age=42yrs n=106	Norfloxacin 400mg twice daily for three days vs. placebo	72 hours	Cured of TD at 72hrs: 1.69 (1.18, 2.42)	Included in analysis
DuPont 1992	U.S. college students in Mexico, mean age=28yrs n=232	Ofloxacin 300mg twice daily for three days or five days vs. placebo	5 days	Adverse events 7.84 (0.45, 135.58) Duration of diarrhea N/A	Adverse events: included in analysis Other outcomes: included in review, excluded from analysis
Steffen 1993	Guests in a hotel in The Gambia, mean age=37yrs n=195	Fleroxacin 400mg, single dose or for two days vs. placebo	72 hours	Adverse events: 1.71 (1.25, 2.34) Cured of TD at 72hrs: 1.71 (1.25, 2.34)	Adverse events: included in analysis Other outcomes: Excluded from analysis Number of unformed stools less than classic TD definition.
Taylor 2006	Patients consulting travel health clinics in Mexico, Guatemala, India, or Peru, mean age=33yrs n=202	Ciprofloxacin 500mg twice daily for three days vs. placebo	5 days	Adverse events 0.96 (0.59, 1.56) Median time to last unformed stool: 1.89 (1.34, 2.65)	Adverse events: included in analysis Other outcomes: included in review, excluded from analysis
Ericsson 1987	U.S. college students in Mexico, mean age=28yrs n=127	Ciprofloxacin 500mg twice daily for five days vs. placebo	5 days	Duration of diarrhea N/A	Included in review, excluded from analysis
Wistrom 1992	U.S. volunteers in Mexico, mean age=34yrs n=42	Ciprofloxacin 250mg twice daily for three days vs. placebo	72 hours	Cure after 48hrs N/A Duration of diarrhea, N/A Mean number of stools N/A	Excluded from analysis Very small study population and large exclusion (only 15 subjects evaluated for efficacy).
Salam 1994	British military personnel in Belize, age not given n=88	Ciprofloxacin 500mg, single dose vs. placebo	72 hours	Cured of TD at 72hrs: 1.21 (1.02, 1.44)	Excluded from analysis Number of unformed stools less than classic TD definition.
Azithromycin for the treatment of TD
Kuschner 1995	U.S. military personnel in Thailand, mean age=30yrs n=79	Azithromycin 500mg once daily for three days vs. Ciprofloxacin 500mg once daily for three days	72 hours	Recovered by 24hrs: 0.70 (0.37, 1.35) Recovered by 48hrs: 1.02 (0.76, 1.37) Recovered by 72hrs: 1.06 (0.96, 1.16) Treatment failure: 1.26 (0.52, 3.07)	Included in analysis
Adachi 2003	U.S. college students in Mexico, mean age=25yrs n=217	Azithromycin 1000mg, single dose vs. Levofloxacin 500mg, single dose	4 days	Immediate cure: 0.37 (0.17, 0.79) Treatment failure: 1.26 (0.52, 3.07)	Included in analysis
Sanders 2007	U.S. military and their beneficiaries in Turkey, median age=31yrs n=207	Azithromycin 1000mg, single dose plus loperamide vs. Levofloxacin 500mg, single dose plus loperamide	72 hours	Immediate cure: 0.76 (0.26, 1.99) Recovered by 24hrs: 0.82 (0.61, 1.11) Recovered by 48hrs: 1.41 (0.96, 2.07) Recovered by 72hrs: 0.95 (0.50, 1.80) Nausea immediately after first dose: 7.62 (0.97, 59.86) Vomiting immediately after first dose: 2.86 (0.12, 69.40) Nausea during remainder of follow-up: 0.94 (0.72, 1.31) Vomiting during remainder of follow-up: 0.70 (0.34, 1.45)	Included in analysis
Tribble 2007	U.S. military personnel in Thailand, median age=26yrs n=156	Azithromycin 500mg once daily for three days or single dose vs. Levofloxacin 500mg once daily for three days	72 hours	Recovered by 24hrs: 0.70 (0.37, 1.35) Recovered by 48hrs: 1.02 (0.76, 1.37) Recovered by 72hrs: 1.06 (0.96, 1.16) Nausea immediately after first dose: 5.15 (0.68, 39.13) Vomiting immediately after first dose: 1.56 (0.06, 37.60) Nausea during remainder of follow-up: 2.02 (0.60, 6.84) Vomiting during remainder of follow-up: 1.29 (0.26, 6.41)	Included in analysis
Rifaximin for the treatment of TD
Steffen 2003	Patients consulting travel health clinics in Mexico, Guatemala, or Kenya, mean age=29yrs n=380	Rifaximin 600mg, 200mg three times daily or 1200mg, 400mg three times daily for three days vs. placebo	5 days	Clinical cure after 5 days: 1.29 (1.15, 1.45) Treatment failure: 0.47 (0.33, 0.68) Adverse events: 0.93 (0.80, 1.07)	Included in analysis
Taylor 2006	Patients consulting travel health clinics in Mexico, Guatemala, India, or Peru, mean age=33yrs n=298	Rifaximin 600mg, 200mg three times daily for three days vs. placebo Rifaximin 600mg, 200mg three times daily for three days vs. Ciprofloxacin 1000mg, 500mg twice daily for three days	5 days	Clinical cure after 5 days: 1.25 (1.05, 1.48) Treatment failure: 0.55 (0.35, 0.88) Adverse events: 1.07 (0.71, 1.61) Clinical cure after 5 days: 0.98 (0.86, 1.11) Treatment failure: 1.80 (0.81, 4.02) Adverse events: 1.11 (0.73, 1.69)	Included in analysis
DuPont 2001	U.S. college students in Mexico and travellers to Jamaica, mean age=26yrs n=187	Rifaximin 800mg, 400mg twice daily for three days vs. Ciprofloxacin 1000mg, 500mg twice daily for three days	5 days	Clinical cure after 5 days: 0.99 (0.89, 1.10) Treatment failure: 1.82 (0.63, 5.22) Adverse events: 0.92 (0.62, 1.37)	Included in analysis
Footnote 1 *studies are randomized, double-blind, placebo-controlled trials unless otherwise specified

Table 2 - Dosages for various agents to prevent and treat travellers' diarrhea

Agent	Dosage – chemoprophylaxisFootnote *	Dosage – treatmentFootnote *	Comments
Antimotility agents
Loperamide	N/A	Adults: 4 mg initially + 2 mg after each loose stool (16 mg daily maximum) Children: First 24 hours: 2-6 years (13-20kg): 1 mg, 3 times daily 6-8 years (20-30kg): 2 mg, twice daily 8-12 years (>30kg): 2 mg, 3 times daily From 24-48 hours, after each loose stool: 0.1 mg/kg (not to exceed initial dose)	Contraindicated in infants < 2 years of age.
Antisecretory agents
Bismuth subsalicylate	Two 262 mg tablets (524 mg), 4 times daily (4.2 maximum dose/24hrs) Higher doses (4.2g/24hrs or 4 tabs, 4 times daily) appear to be effective, but represent a high pill burden, and are not clearly shown to be superior to one tablet 4 times daily (2.1g/day). Halving of the dose for chemoprophylaxis to 1.05g/24hrs is a possible alternative if higher dosage is not feasible.	Adults: 524-1048 mg every 30-60 minutes as needed (4.2 g maximum dose / 24hrs) Children: Dose every 30-60 minutes as needed 2-4 years: 88-176 mg (0.7 g maximum dose /24hrs) 5-9 years: 131-262 mg (1.05 g maximum dose / 24hrs) 10-14 years: 262-524 mg (2.1 g maximum dose / 24hrs)	Available evidence does not allow a GRADE-based recommendation for the use of BSS to treat TD. Chemoprophylaxis not recommended in children and treatment not recommended in infants < 2 years of age. Chemoprophylaxis should not exceed three weeks in adults. Treatment should not exceed 2 days in adults and children. Efficacy between liquid and tablet forms of BSS does not appear to differ, though this has not been formally evaluated. Treatment contraindicated in ASA allergy.
Antibiotic agents
Norfloxacin	400 mg, once daily	800 mg single dose 400 mg, twice daily for 3 days	Fluoroquinolones: Treatment contraindicated in pregnant women. Caution recommended for travellers to the Indian sub-continent and southeast Asia due to presence of antibiotic resistant pathogens. Risks for adverse events may be increased in children under the age of 18, and these risks should be balanced against the potential benefits.
Ciprofloxacin	500 mg, once daily	Adults: 500-1000 mg single dose 500 mg, twice daily for 3 days Treatment course need not be completed if symptoms resolve, and lower doses may be sufficient. Children: 20-30 mg/kg/day, divided, twice daily for 3 days (max 1.5 g/24 hours)
Ofloxacin	300 mg, once daily	400 mg single dose 200 mg, twice daily for 3 days
Levofloxacin	500 mg, once daily	1000 mg single dose 500 mg, once daily for 3 days
Azithromycin	Children:	Adults: 1000 mg single dose 500 mg, once daily for 3 days Children: 5-10 mg/kg, once daily for 3 days (max 500 mg/24 hours)	Alternative for treatment for those travelling to regions with high quinolone resistance. Prevention of TD in children with azithromycin has not been studied, and should be used with caution.
Rifaximin	600mg, once daily Lower doses, such as 200mg once or twice daily, have been used with variable effectiveness.	200mg, 3 times daily for 3 days	For prevention: only the 550mg tablet for Rifaximin (Zaxine) is market-approved in Canada. However, CATMAT considers this difference in dosage to not be clinically important. For treatment: 200mg tablets are not available in Canada. The manufacturer does not recommend splitting the 550mg tablets. 200mg tablets are approved for use in some other countries. Rifaximin has not been studied nor approved for use in children under the age of 12, nor has dosing in children been established.
Cefixime	N/A	400 mg single dose 8 mg/kg, once daily for 3 days	Alternative for children if quinolones and macrolides contraindicated. There are no clinical data for the use of cefixime in TD.
Footnote * *Various recommendations for preventive and treatment interventions provided within this statement include off-label use of medications. Product monographs or other similarly approved standards or instructions for use should be reviewed prior to use.

Table 3 - Preparing oral rehydration solutions at home

Homemade oral rehydration solution
Ingredients	Amount
Purified water	1 L (4¼ cups)
Salt	2.5 mL (½ teaspoon)
Sugar	30 mL (6 teaspoons)

Oral inactivated cholera vaccine (WC-BS and WC-rCTB) versus non-vaccinated for prevention of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Cholera vaccine (WC-BS and WC-rCTB)	non-vaccinated (RCT only)	Relative (95% CI)	Absolute
Episode of travellers' diarrhea (follow-up 0-35 days; assessed with: Patient self-assessment and clinical follow-up)
3Table 2 - Footnote 1	randomised trials	no serious risk of biasTable 2 - Footnote 2	no serious inconsistencyTable 2 - Footnote 3	seriousTable 2 - Footnote 4	no serious imprecision	none	216/616 (35.1%)	232/626 (37.1%)	RRTable 2 - Footnote * 0.94 (0.82 to 1.09)	22 fewer per 1000 (from 67 fewer to 33 more)	MODERATE	CRITICAL
Footnote * *RR = relative risk Studies: Peltola 1991, Scerpella 1995, Wiedermann 2000 Footnote 1 1Five observational studies (Lopez-Gigosos 2007, 2009, 2013, Torrell 2009, Gabutti 2012) were excluded due to low quality of study design. One RCT study (Clemens 1988) was excluded as it was not conducted in a traveller population. Footnote 2 2Although one study did not clearly state loss to follow-up and one other had 21% loss to follow-up, these were not deemed sufficient to rate down for quality. Footnote 3 3Although I 2 value is high, all three RCTs include the null, all 95%CIs overlap and test for heterogeneity is non-significant. Footnote 4 4Due to heterogeneity in type of vaccine, vaccine administration & definitions used for TD.

Bismuth subsalicylate (BSS) versus Placebo for prevention of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Bismuth subsalicylate (BSS)	Placebo	Relative (95% CI)	Absolute
Episode of travellers' diarrhea during follow-up (follow-up 0-30 days; assessed with: Patient self-assessment)
3Table 3 - Footnote 1	randomised trials	no serious risk of biasTable 3 - Footnote 2	no serious inconsistencyTable 3 - Footnote 3	no serious indirectness	no serious imprecision	none	105/335 (31.3%)	109/196 (55.6%)	RR 0.55 (0.44 to 0.67)	250 fewer per 1000 (from 184 fewer to 311 fewer)	HIGH	CRITICAL
Studies: DuPont 1980, Steffen 1986, DuPont 1987. Footnote 1 1One study (Graham 1983) was excluded due to inadequate definition of TD. Footnote 2 2Although one of the studies had a large loss to follow-up, when it is removed, the effect remains strong and becomes more consistent. Footnote 3 3Although the I 2 values are high, the studies all show a strong protective effect and 95%CIs all overlap with each other.

High dose BSS (2.1 g - 4.2 g per day) versus Placebo for prevention of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	High dose BSS (2.1-4.2g per day)	Placebo	Relative (95% CI)	Absolute
Episode of travellers' diarrhea during follow-up (follow-up 0-30 days; assessed with: Patient self-assessment)
3Table 4 - Footnote 1	randomised trials	no serious risk of biasTable 4 - Footnote 2	no serious inconsistencyTable 4 - Footnote 3	no serious indirectness	no serious imprecision	none	59/201 (29.4%)	109/196 (55.6%)	RR 0.51 (0.39 to 0.65)	272 fewer per 1000 (from 195 fewer to 339 fewer)	HIGH	CRITICAL
Studies: DuPont 1980, Steffen 1986, DuPont 1987. Footnote 1 1One study (Graham 1983) was excluded due to inadequate definition of TD. Footnote 2 2Although one of the studies had a large loss to follow-up, when it is removed, the effect remains strong and becomes more consistent. Footnote 3 3Although the I 2 values are high, the studies all show a strong protective effect and 95%CIs all overlap with each other.

Low dose BSS (1.05 g per day) versus Placebo for prevention of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Low dose BSS (1.05g per day)	Placebo	Relative (95% CI)	Absolute
Episode of travellers' diarrhea during follow-up (follow-up 0-30 days; assessed with: Patient self-assessment)
2	randomised trials	seriousTable 5 - Footnote 1	no serious inconsistency	no serious indirectness	seriousTable 5 - Footnote 2	none	46/134 (34.3%)	69/130 (53.1%)	RR 0.65 (0.50 to 0.86)	186 fewer per 1000 (from 74 fewer to 265 fewer)	LOW	CRITICAL
Studies: Steffen 1986, DuPont 1987. Footnote 1 1One of the two studies had a large loss to follow-up. Footnote 2 2Insufficient sample size and number of events.

High dose BSS (2.1 g - 4.2 g per day) versus Low dose BSS (1.05 g per day) for prevention of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	High dose BSS (2.1-4.2 g per day)	Low dose BSS (1.05 g per day)	Relative (95% CI)	Absolute
Episode of travellers' diarrhea during follow-up (follow-up 0-30 days; assessed with: Patient self-assessment)
2	randomised trials	serious Table 6 - Footnote 1	no serious inconsistency	no serious indirectness	serious Table 6 - Footnote 2	none	45/139 (32.4%)	46/134 (34.3%)	RR 0.87 (0.63 to 1.22)	45 fewer per 1000 (from 127 fewer to 76 more)	LOW	CRITICAL
Studies: Steffen 1986, DuPont 1987. Footnote 1 1One of the two studies had a large loss to follow-up. Footnote 2 2Insufficient sample size and number of events.

Fluoroquinolones versus Placebo for prevention of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Fluoroquinolones	Placebo	Relative (95% CI)	Absolute
Episode of travellers' diarrhea during follow-up (follow-up 5-21 days; assessed with: Patient self-assessment)
4Table 7 - Footnote 1	randomised trials	no serious risk of bias	no serious inconsistency	no serious indirectness	no serious imprecisionTable 7 - Footnote 2	none	13/316 (4.1%)	109/327 (33.3%)	RR 0.12 (0.07 to 0.21)	293 fewer per 1000 (from 263 fewer to 310 fewer)	HIGH	CRITICAL
Studies: 1) Norfloxacin (n=3): Johnson 1986, Wiström 1987, Scott 1990; 2) Ciprofloxacin (n=1): Heck 1994 Footnote 1 1One study (Rademaker 1989) was excluded due to inadequate definition of TD. Footnote 2 2Insufficient Optimal Information Size (OIS): number of events recommended by GRADE authors to achieve a RR reduction of 25%, given alpha=0.05 & beta=0.2 is 250 vs. 122 observed. However, all point estimates show a large, significant protective effect.

Rifaximin versus placebo for prevention of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Rifaximin	placebo	Relative (95% CI)	Absolute
Occurrence of travellers' diarrhea (combined tx arms DuPont 2005) (follow-up 0-26 days; assessed with: Patient self-assessment and clincial follow-up)
5	randomised trials	no serious risk of bias	no serious inconsistencyTable 8 - Footnote 1	no serious indirectnessTable 8 - Footnote 2	no serious imprecision	reporting biasTable 8 - Footnote 3	74/475 (15.6%)	135/368 (36.7%)	RR 0.42 (0.33 to 0.53)	213 fewer per 1000 (from 172 fewer to 246 fewer)	MODERATE	CRITICAL
Studies: DuPont 2005 (all 3 treatment arms combined), Armstrong 2010, Martinez-Sandoval 2010, Flores 2011, Zanger 2013 Footnote 1 1Although the I 2 values are high, all studies show a consistent protective effect for rifaximin and all 95%CIs overlap. Footnote 2 2Dosage varies between studies. However, estimates change very little when only 550mg-600mg dosage results are considered. Footnote 3 3One large (n=660) unpublished study completed in 2008 was found on ClinicalTrials.gov database: no results reported. Studies are all of small or moderate size.

Loperamide versus placebo for treatment of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Loperamide	placebo	Relative (95% CI)	Absolute
Number of loose stools during first day of treatment (follow-up 0-5 days; measured with: Patient diary and clinical follow-up; Better indicated by lower values)
2	randomised trials	no serious risk of bias	no serious inconsistency	seriousTable 9 - Footnote 1	seriousTable 9 - Footnote 2	none	65	67	-	MD 1.59 lower (2.92 to 0.26 lower)	LOW	CRITICAL
First relief of acute diarrhea after 4 hours of treatment (follow-up 0-3 days; assessed with: Patient diary)
2	randomised trials	no serious risk of bias	no serious inconsistency	seriousTable 9 - Footnote 3	seriousTable 9 - Footnote 4	none	55/155 (35.5%)	33/157 (21%)	RR 1.69 (1.17 to 2.45)	145 more per 1000 (from 36 more to 305 more)	LOW	CRITICAL
First relief of acute diarrhea after 12 hours of treatment (follow-up 0-3 days; assessed with: Patient diary)
2	randomised trials	no serious risk of bias	no serious inconsistency	seriousTable 9 - Footnote 3	seriousTable 9 - Footnote 4	none	74/155 (47.7%)	50/157 (31.8%)	RR 1.50 (1.13 to 1.99)	159 more per 1000 (from 41 more to 315 more)	LOW	CRITICAL
First relief of acute diarrhea after 24 hours of treatment (follow-up 0-3 days; assessed with: Patient diary)
2	randomised trials	no serious risk of bias	no serious inconsistency	seriousTable 9 - Footnote 3	no serious imprecisionTable 9 - Footnote 5	none	108/155 (69.7%)	79/157 (50.3%)	RR 1.38 (1.15 to 1.66)	191 more per 1000 (from 75 more to 332 more)	MODERATE	CRITICAL
*MD = mean difference Studies: 1) Number of loose stools: Bergstrom 1986, van Loon 1989; 2) First relief of acute diarrhea: Hughes 1995, van den Eynden 1995 Footnote 1 1One of the two studies does not use a traveller study population. One of the studies did not specify if stools were unformed. Footnote 2 2Standard errors used to calculate 95%CIs were approximated from graphical representations (van Loon) or were derived from median and range data (Bergstrom) and as such, may not be precise. Insufficient sample size and number of events. Both studies have 95%CIs which include the null. Footnote 3 3Both studies did not use a traveller study population. Footnote 4 4Insufficient sample size and number of events (OIS). Footnote 5 5Number of events and sample size is borderline acceptable. However, both studies have 95%CIs which do not include the null.

Loperamide and antibiotic versus antibiotic alone for treatment of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Loperamide and antibiotic	antibiotic alone	Relative (95% CI)	Absolute
Complete relief of TD after 24hrs (follow-up 0-5 days; assessed with: Patient assessment with clinical follow-up)
4Table 10 - Footnote 1	randomised trials	no serious risk of bias	no serious inconsistencyTable 10 - Footnote 2	no serious indirectnessTable 10 - Footnote 3	no serious imprecision	none	143/254 (56.3%)	92/253 (36.4%)	RR 1.55 (1.28 to 1.86)	200 more per 1000 (from 102 more to 313 more)	HIGH	CRITICAL
Complete relief of TD after 48hrs (follow-up 0-5 days; assessed with: Patient assessment with clinical follow-up)
4Table 10 - Footnote 1	randomised trials	no serious risk of bias	no serious inconsistencyTable 10 - Footnote 2	no serious indirectnessTable 10 - Footnote 3	no serious imprecision	none	186/254 (73.2%)	150/253 (59.3%)	RR 1.24 (1.09 to 1.40)	142 more per 1000 (from 53 more to 237 more)	HIGH	CRITICAL
Complete relief of TD after 72hrs (follow-up 0-5 days; assessed with: Patient assessment with clinical follow-up)
4	randomised trials	no serious risk of bias	seriousTable 10 - Footnote 4	no serious indirectnessTable 10 - Footnote 3	no serious imprecision	none	212/254 (83.5%)	196/255 (76.9%)	RR 1.09 (1.00 to 1.19)	69 more per 1000 (from 0 more to 146 more)	MODERATE	CRITICAL
Treatment failure (follow-up 0-5 days; assessed with: Patient assessment with clinical follow-up)
3	randomised trials	no serious risk of bias	no serious inconsistency	no serious indirectnessTable 10 - Footnote 5	seriousTable 10 - Footnote 6	none	3/142 (2.1%)	19/144 (13.2%)	RR 0.18 (0.06 to 0.55)	108 fewer per 1000 (from 59 fewer to 124 fewer)	MODERATE	CRITICAL
Studies: 1) Complete relief at 24hrs, 48hrs and 72hrs: Petruccelli 1992, Ericsson 1990, 2007, DuPont 2007; 2) Treatment failure: Ericsson 1990, 1997, 2007 Footnote 1 1One study (Taylor 1991) was excluded due to inadequate TD definition. Footnote 2 2Although the I 2 values are high, 95%CIs from all studies overlap and magnitude of effect is relatively consistent. Footnote 3 3Although one study had a different definition of outcome (included improvement), this was not judged sufficient to downgrade. Footnote 4 4Substantial variation in direction of effect, 95%CIs do not all overlap, high I 2. Footnote 5 5Although all the studies were conducted in the same population, this was not judged as sufficient to downgrade. Footnote 6 6Insufficient number of subjects and events to detect effect (OIS). Two of three studies have extremely wide 95%CIs which include null.

Fluoroquinolones versus placebo for treatment of travellers' diarrhea

Quality assessment							No of patients		>Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Fluoroquinolones versus placebo	Control	Relative (95% CI)	Absolute
Cured of TD at 72 hours (follow-up 0-3 days; assessed with: Patient assessment and some clinical follow-up)
2Table 11 - Footnote 1	randomised trials	no serious risk of bias	no serious inconsistency	no serious indirectness	seriousTable 11 - Footnote 2	none	70/97 (72.2%)	41/103 (39.8%)	RR 1.81 (1.39 to 2.37)	322 more per 1000 (from 155 more to 545 more)	MODERATE	CRITICAL
Adverse events (follow-up 0-5 days; assessed with: Patient assessment and some clinical follow-up)
3	randomised trials	no serious risk of bias	seriousTable 11 - Footnote 3	seriousTable 11 - Footnote 4	seriousTable 11 - Footnote 2	none	109/378 (28.8%)	50/243 (20.6%)	RR 1.39 (1.05 to 1.83)	80 more per 1000 (from 10 more to 171 more)	VERY LOW	CRITICAL
Studies: 1) Cured at 72hrs: Wiström 1989 (norfloxacin), Mattila 1993 (norfloxacin); 2) Adverse events: DuPont 1992 (ofloxacin), Steffen 1993 (fleroxacin), Taylor 2006 (ciprofloxacin) Footnote 1 1Two studies (Salam 1994 and Steffen 1993) were excluded from analysis since they had an inadequate definition of traveller's diarrhea (1+ unformed stools). Footnote 2 2Number of events and sample size probably insufficient to detect effect (according to Optimal Information Size criteria). Footnote 3 3Differing direction of effect across studies. Large I 2 value observed. Footnote 4 4Lack of standardized criteria across studies for definition of what constitutes an adverse effect.

Azithromycin versus fluoroquinolones for treatment of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Azithromycin	Fluoroquinolones	Relative (95% CI)	Absolute
Rapid / immediate cure (follow-up 0-4 days; assessed with: Patient self-assessment and some clinical follow-up)
2	randomised trials	no serious risk of bias	no serious inconsistency	seriousTable 12 - Footnote 1	seriousTable 12 - Footnote 2	none	14/211 (6.6%)	30/207 (14.5%)	RR 0.46 (0.25 to 0.84)	78 fewer per 1000 (from 23 fewer to 109 fewer)	LOW	CRITICAL
Treatment failure (follow-up 0-4 days; assessed with: Patient self-assessment and clinical follow-up)
2	randomised trials	no serious risk of bias	no serious inconsistencyTable 12 - Footnote 3	seriousTable 12 - Footnote 4	seriousTable 12 - Footnote 5	none	10/140 (7.1%)	10/143 (7%)	RR 1.02 (0.45 to 2.32)	1 more per 1000 (from 38 fewer to 92 more)	LOW	CRITICAL
Recovered by 24hr (follow-up 0-3 days; assessed with: Patient self-assessment and some clinical follow-up)
3	randomised trials	no serious risk of bias	no serious inconsistency	seriousTable 12 - Footnote 6	seriousTable 12 - Footnote 7	none	73/244 (29.9%)	79/191 (41.4%)	RR 0.79 (0.61 to 1.01)	87 fewer per 1000 (from 161 fewer to 4 more)	LOW	CRITICAL
Recovered by 48hr (follow-up 0-3 days; assessed with: Patient self-assessment and some clinical follow-up)
3	randomised trials	no serious risk of bias	seriousTable 12 - Footnote 8	seriousTable 12 - Footnote 6	no serious imprecision	none	129/244 (52.9%)	75/191 (39.3%)	RR 1.34 (1.08 to 1.66)	134 more per 1000 (from 31 more to 259 more)	LOW	CRITICAL
Recovered by 72hr (follow-up 0-3 days; assessed with: Patient self-assessment and some clinical follow-up)
3	randomised trials	no serious risk of bias	seriousTable 12 - Footnote 8	seriousTable 12 - Footnote 6	no serious imprecision	none	144/244 (59%)	88/191 (46.1%)	RR 1.16 (1.00 to 1.33)	74 more per 1000 (from 0 more to 152 more)	LOW	CRITICAL
Nausea immediately after first dose (30 min) (follow-up 0-3 days; assessed with: Clinical observation)
2	randomised trials	no serious risk of bias	no serious inconsistency	seriousTable 12 - Footnote 6	seriousTable 12 - Footnote 7	none	18/209 (8.6%)	2/154 (1.3%)	RR 6.23 (1.48 to 26.26)	68 more per 1000 (from 6 more to 328 more)	LOW	IMPORTANT
Vomiting immediately after first dose (30 min) (follow-up 0-3 days; assessed with: Patient self-assessment and some clinical follow-up)
2	randomised trials	no serious risk of bias	no serious inconsistency	seriousTable 12 - Footnote 6	seriousTable 12 - Footnote 7	none	2/209 (0.96%)	0/154 (0%)	RR 2.13 (0.23 to 19.75)	-	LOW	IMPORTANT
Nausea (starting after tx) during remainder of 3-day observation period (follow-up 0-3 days; assessed with: Patient self-assessment and some clinical follow-up)
2	randomised trials	no serious risk of bias	no serious inconsistency	seriousTable 12 - Footnote 6	no serious imprecision	none	65/209 (31.1%)	57/153 (37.3%)	RR 1.01 (0.77 to 1.31)	4 more per 1000 (from 86 fewer to 115 more)	MODERATE	IMPORTANT
Vomiting (starting after tx) during remainder of 3-day observation period (follow-up 0-3 days; assessed with: Patient self-assessment and some clinical follow-up)
2	randomised trials	no serious risk of bias	no serious inconsistency	seriousTable 12 - Footnote 6	seriousTable 12 - Footnote 9	none	16/209 (7.7%)	17/154 (11%)	RR 0.78 (0.41 to 1.52)	24 fewer per 1000 (from 65 fewer to 57 more)	LOW	NOT IMPORTANT
Studies: 1) Rapid cure: Adachi 2003, Sanders 2007; 2) Treatment failure: Kuschner 1995, Adachi 2003; 3) Recovered by 24hrs, 48hrs, 72hrs: Adachi 2003, Sanders 2007, Tribble 2007; 4) Adverse events (nausea and vomiting): Sanders 2007, Tribble 2007 Footnote 1 1One study includes loperamide use in its dosage whereas the other does not. Footnote 2 2One of the two studies has a 95%CI which includes large benefit and large risk. Number of events and sample size most likely insufficient to detect effect (OIS). Footnote 3 3Although the estimates show opposing effects, the 95%CIs both include the null and I 2 is low. Footnote 4 4Study locations differ with respect to Campylobacter prevalence. Also differing dosages for treatment. Footnote 5 595%CIs for both studies include large benefit and large risk. Insufficient sample size and number of events to detect effect (OIS). Footnote 6 6One study includes loperamide use in its dosage whereas the other does not. Study locations differ with respect to Campylobacter prevalence. Also differing dosages for treatment. Footnote 7 7All studies have a 95%CI which includes large benefit and the null. Number of events and sample size most likely insufficient to detect effect (OIS). Footnote 8 8Not all studies have 95%CIs overlapping null. Relatively high I 2. Footnote 9 9All studies have 95%CIs which include substantial benefit and harm. Number of events and sample size most likely insufficient to detect effect (OIS).

Rifaximin versus placebo for treatment of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Rifaximin	placebo	Relative (95% CI)	Absolute
Wellness after 120 hours of follow-up (follow-up 0-5 days; assessed with: Patient self-assessment with clinical follow-upTable 13 - Footnote 1)
2	randomised trials	no serious risk of bias	no serious inconsistency	no serious indirectnessTable 13 - Footnote 2	no serious imprecision	none	335/427 (78.5%)	134/220 (60.9%)	RR 1.29 (1.15 to 1.45)	177 more per 1000 (from 91 more to 274 more)	HIGH	CRITICAL
Treatment failure (follow-up 0-5 days; assessed with: Patient self-assessment with clinical follow-upTable 13 - Footnote 3)
2	randomised trials	no serious risk of bias	no serious inconsistency	no serious indirectnessTable 13 - Footnote 2	no serious imprecision	none	70/427 (16.4%)	72/220 (32.7%)	RR 0.50 (0.38 to 0.67)	164 fewer per 1000 (from 108 fewer to 203 fewer)	HIGH	CRITICAL
Adverse events (follow-up 0-5 days; assessed with: Patient self-assessment with clinical follow-up)
2	randomised trials	no serious risk of bias	no serious inconsistency	no serious indirectnessTable 13 - Footnote 2	no serious imprecision	none	215/450 (47.8%)	115/229 (50.2%)	RR 0.96 (0.83 to 1.11)	20 fewer per 1000 (from 85 fewer to 55 more)	HIGH	CRITICAL
Studies: Steffen 2003, Taylor 2006 Footnote 1 1Wellness is defined as 48hrs with no unformed stool and no fever, or 24hrs without watery stools, maximum two soft stools, and no clinical symptoms after 120hrs of follow-up from first dose. Footnote 2 2Although there were differences in dosage for treatment, effect did not appreciably change when limited to 600mg dosage. Footnote 3 3Treatment failure is defined as clinical deterioration or worsening of symptoms after at least 24hrs of therapy or illness continuing after 120 hrs of treatment with study medication or after at least 24hrs of therapy.

Rifaximin versus ciprofloxacin for treatment of travellers' diarrhea

Quality assessment							No of patients		Effect		Quality	Importance
No of studies	Design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Rifaximin	ciprofloxacin	Relative (95% CI)	Absolute
Wellness after 120 hours of follow-up (follow-up 0-5 days; assessed with: Patient self-assessment with clinical follow-upTable 14 - Footnote 1)
2	randomised trials	no serious risk of bias	no serious inconsistency	no serious indirectnessTable 14 - Footnote 2	no serious imprecision	none	232/290 (80%)	162/195 (83.1%)	RR 0.98 (0.90 to 1.07)	17 fewer per 1000 (from 83 fewer to 58 more)	HIGH	CRITICAL
Treatment failure (follow-up 0-5 days; assessed with: Patient self-assessment with clinical follow-upTable 14 - Footnote 3)
2	randomised trials	no serious risk of bias	no serious inconsistency	no serious indirectnessTable 14 - Footnote 2	seriousTable 14 - Footnote 4	none	34/279 (12.2%)	12/188 (6.4%)	RR 1.81 (0.96 to 3.43)	52 more per 1000 (from 3 fewer to 155 more)	MODERATE	CRITICAL
Adverse events (follow-up 0-5 days; assessed with: Patient self-assessment with clinical follow-up)
2	randomised trials	no serious risk of bias	no serious inconsistency	no serious indirectnessTable 14 - Footnote 2	no serious imprecision	none	84/292 (28.8%)	58/194 (29.9%)	RR 1.01 (0.76 to 1.35)	3 more per 1000 (from 72 fewer to 105 more)	HIGH	CRITICAL
Studies: DuPont 2001, Taylor 2006 Footnote 1 1Wellness is defined as 48hrs with no unformed stool and no fever, or 24hrs without watery stools, maximum two soft stools, and no clinical symptoms after 120hrs of follow-up from first dose. Footnote 2 2Although treatment dosages differ (rifaximin 600mg vs. 800mg), these differences were considered to be minimal. Footnote 3 3Treatment failure is defined as clinical deterioration or worsening of symptoms after at least 24hrs of therapy or illness continuing after 120hrs of treatment with study medication or after at least 24hrs of therapy. Footnote 4 4Both studies have wide 95%CIs which include large harm and benefit. There are most likely an insufficient sample size and number of events to detect effect.