Figure

Figure. Study protocol for investigating risk for contracting ESBL-producing Enterobacteriaceae among travelers from Finland with TD. LO–AMD–, not treated with medication; LO+AMD–, treated with LO alone; LO–AMD+, treated with AMDs alone; LO+AMD+,...

We reviewed our recent data on ESBL-E acquisition among 430 travelers from Finland (6), selecting those with TD for separate analysis (Figure). All the volunteers provided fecal samples and completed questionnaires before and after travel. Symptoms of TD and use of medications, such as loperamide and AMDs, were included in the posttravel questionnaires. The countries visited were grouped as described (Table 1; 6); processing of fecal specimens and identification of ESBL-E were detailed in our previous study (6). TD was defined by the World Health Organization criteria: passing >3 loose/liquid stools per 24 hours, or more frequently than normal (14).

Study participants were divided into 4 groups by treatment of TD: those taking no loperamide or AMDs (LO–AMD–); only loperamide (LO+AMD–); only AMDs (LO–AMD+); or loperamide plus AMDs (LO+AMD+). Those having taken AMDs for non-TD indications were categorized in groups with those with TD who took AMDs.

We used a multivariable binary logistic regression model to test our main hypotheses. Loperamide, AMDs, and their interaction (effect modification) were included in the model, along with risk factors that showed a p value <0.2 in univariate analysis in our previous study (6): sex, travel destination, use of AMDs, meals with residents of the location, contact with local healthcare, sites of meals, accommodations, duration of travel, age, and use of alcohol. Variables were eliminated to the final model by using backward selection of factors by Akaike Information Criteria, except for loperamide and the use of AMDs and their interaction, which were forced to the final model. Missing values were taken into account by multiple imputations, to reduce possible biases and efficiency loss, assuming that data were missing at random. We analyzed statistics using SPSS statistical software version 21 (IBM Corporation, Armonk, NY, USA).

Of all travelers in the previous study (6), a total of 288 of 430 (67%) who reported TD constituted the final study group (Table 1). ESBL-E was contracted by 26% of the subjects: 21% in the LO–AMD– group; 20% in the LO+AMD– group (adjusted odds ratio [aOR] 0.8, 95% CI 0.4–1.7); 40% in the AMD+LO– group (aOR 2.9, 95% CI 1.2–7.4); and 71% in the LO+AMD+ group (aOR 7.4, 95% CI 1.7–32.6) (Table 2). aOR for the interaction term of loperamide and AMDs was 3.1 (95% CI 0.6–16.6). Travel destination remained an independent risk factor, and sharing meals with locals appeared protective (Table 2).

Studies showing AMD treatment of patients with TD to be an independent risk factor for contracting ESBL-E (5–7) have evoked the question of less harmful treatments. The recommendation to restrict AMDs to severe cases (5,6,15) seems reasonable, as TD generally remains mild or moderate and resolves spontaneously (12,15). If symptoms require medical treatment, loperamide appears to be a sensible alternative for travelers who have no fever or bloody stools. However, because of its antimotility effect, its safety against contracting resistant intestinal bacteria has been questioned. Among studies that explored risk factors for ESBL carriage, we found none that showed data on the use of loperamide alone or in combination with AMDs.

Consistent with our previous analysis (6), we found AMD treatment of TD was an independent risk factor for colonization with ESBL-E; the rate increased from 21% (LO–AB–) to 40% (LO–AB+) (aOR 2.9, 95% CI 1.2–7.4). When used alone, loperamide did not add to the risk (20% colonized in the LO+AB– group).

In the group taking both loperamide and AMDs, the colonization rate was strikingly high, increasing from 21% (LO–AB–) to 71% (LO+AB+). The rate also appeared to exceed the risk for using AMDs alone (40%), yet the interaction term of loperamide and AMDs did not reach statistical significance (aOR 3.1, 95% CI 0.6–16.6), and the small subject number resulted in wide CIs. Theoretically, an additional increase in the risk seen in the combination group could be brought about by loperamide: because of its antimotility effect, contact time of the AMD to the gut lumen is increased, and the selection pressure posed by the AMD may be prolonged, thus intensifying its unfavorable effects.

Our study design was limited by not including a randomized allocation of therapy and the varied use of loperamide according to symptoms. However, no association was seen between the severity of symptoms and acquisition of ESBL-E (data not shown).

Studies are needed to compare the relative risk posed by various AMD classes. Exploration of the influence of nonantimicrobial antidiarrheal agents with only antisecretory effect, such as racecadotril, as monotherapy and in combination with AMDs would also be beneficial.

Our results show that loperamide alone offers a safe alternative to AMDs for TD treatment because it does not add to the resulting in a low risk for acquiring drug-resistant intestinal bacteria. In contrast, combining loperamide with AMDs predisposes to ESBL-E colonization and may add to the substantial risk posed by AMDs alone. Our data dispute the safety of this combination.