Loperamide Plus Azithromycin More Effectively Treats Travelers’ Diarrhea in Mexico than Azithromycin Alone

Charles D. Ericsson MD, Herbert L. DuPont MD, Pablo C. Okhuysen MD, Zhi‐Dong Jiang MD, PhD, MPH, Margaret W. DuPont MS
DOI: http://dx.doi.org/10.1111/j.1708-8305.2007.00144.x 312-319 First published online: 1 September 2007

Abstract

Background Because the combination of loperamide and some antimicrobials has proven to be more efficacious than the antimicrobial agent alone in the treatment of travelers’ diarrhea, we set out to prove loperamide plus azithromycin was more efficacious than azithromycin alone.

Methods During the summers of 2002 to 2003, 176 US adults recently arrived in Guadalajara, Mexico were enrolled in a prospective, double‐blinded, randomized trial of the treatment of acute diarrhea. Subjects received single doses (1,000 or 500 mg) of azithromycin or a single 500 mg dose of azithromycin plus loperamide. Subjects gave a pre‐ and post‐treatment stool sample for analysis and maintained daily diaries of symptoms and passage of stools.

Results The duration of diarrhea was significantly (p= 0.0002) shorter following treatment with azithromycin plus loperamide (11 h) than with either dose of azithromycin alone (34 h). In the first 24 hours, the average number of unformed stools passed was 3.4 (azithromycin alone) and 1.2 (combination) for a significant (p < 0.0001) difference of 2.2 unformed stools. This difference equated with 20% of azithromycin‐treated subjects continuing to pass six or more unformed stools in the first 24 hours post‐treatment compared with only 1.7% of combination‐treated subjects.

Conclusions For the treatment of travelers’ diarrhea in an Escherichia coli predominant region of the world, a single 500 mg dose of azithromycin appeared as effective as a 1,000 mg dose. Loperamide plus 500 mg of azithromycin was safe and more effective than either dose of azithromycin. To realize the substantial clinical benefit that accrues to a subset of subjects, we feel loperamide should routinely be used in combination with an antimicrobial agent to treat travelers’ diarrhea.

For the treatment of travelers’ diarrhea, the combination of loperamide and an antimicrobial agent has proven beneficial in a number of studies. 1–4 Because of rising resistance among enteric pathogens in most of the world to trimethoprim/sulfamethoxazole (TMP/SMX) and in South and Southeast Asia to fluoroquinolones, the need for additional effective antimicrobial agents developed. 5–9 This need has been partially met by the approval of rifaximin to treat travelers’ diarrhea, but rifaximin was not approved for the treatment of invasive pathogens. 10–13 Azithromycin has also proven effective in the treatment of travelers’ diarrhea, shigellosis, and enteric fever and remains substantially effective against enteric pathogens. 14–18

The aim of the present study was to demonstrate the benefits of adding loperamide to azithromycin in the treatment of travelers’ diarrhea in Mexico. Previous studies of loperamide have been criticized because the benefit of adding loperamide to an antimicrobial agent, compared with use of an antimicrobial agent alone, was passage of only an average of one to two fewer unformed stools in the first 24 hours post‐treatment. In the present study, we set out to show how the outcome variable, average number of unformed stools, masks a substantial benefit of treatment to a subset of the treated subjects.

Methods

This prospective, double‐blind, randomized, clinical trial was conducted during the summer months of 2002 and 2003 in Guadalajara, Mexico, among 176 US students attending sessions in the University of Arizona, University of San Diego, and Universidad Autonoma de Guadalajara summer programs. The study was approved by the Committee for the Protection of Human Subjects at the University of Texas Health Science Center at Houston as well as the Institutional Review Boards at the Universities of San Diego, Arizona, and Autonoma de Guadalajara.

Eligible subjects included men or women at least 18 years of age who developed acute diarrhea, which was defined as passage of three or more unformed stools in the preceding 24 hours accompanied by one or more signs or symptoms of enteric infection (eg, nausea, vomiting, abdominal cramps, tenesmus, passage of grossly bloody stools, or fecal urgency) with a duration of illness of less than or equal to 72 hours. Exclusion criteria included pregnancy, breast feeding, an unstable medical condition, taking two or more doses of an antidiarrheal medication in the 24 hours before enrollment or any number of doses of symptomatic therapy within 2 hours of enrollment, or receiving an antimicrobial drug with expected activity against enteric bacterial pathogens within 7 days prior to enrollment.

After giving written consent, subjects submitted an unformed stool sample for analysis and were randomized to one of three oral treatment groups: azithromycin, 1,000 mg; azithromycin, 500 mg; or azithromycin, 500 mg, plus loperamide as a 4 mg loading dose, followed by 2 mg after each loose stool not to exceed a total of 16 mg/d for 2 days. Double dummies were used to blind the study. Medications were randomized in blocks of 12 subjects.

Use of additional antidiarrheal compounds and antipyretics other than acetaminophen were prohibited during the study. Subjects were seen daily for 4 days in our clinics in Guadalajara. Subjects completed daily diaries of clinical symptoms and recorded the time and form of all stools passed during the 4‐day observation period. Stool form was divided into three categories: “formed” (stool retains its shape), “soft” (stool takes the shape of a container), and “watery” (stool can be poured). Both soft and watery stools were considered unformed. Safety was assessed by means of physical examination and review of subjects’ recorded symptoms on daily diaries. A second stool was collected on day 5 for microbiologic assessment.

Bacterial pathogens sought in the laboratory in Mexico included Shigella, Salmonella, Campylobacter jejuni, Aeromonas, Plesiomonas, Vibrio spp, and Yersinia enterocolitica. Protozoal pathogens, including Entamoeba histolytica, Cryptosporidium, and Giardia, were identified by enzyme‐linked immunosorbent assay (Alexon, Sunnydale, CA, USA). Five Escherichia coli–like colonies from each stool sample were inoculated into individual peptone stabs and transported to Houston for further identification. Enterotoxigenic E coli was identified by the capacity of isolates to produce heat labile (LT) and heat stable (ST) enterotoxin as determined by DNA hybridization/probe technique. 19 Bacterial isolates from enrollment and day 5 stools were tested for susceptibility to azithromycin (and for comparison with ciprofloxacin) by standard tube dilution assay. 8

The primary clinical efficacy end point was duration of post‐treatment diarrhea, which was defined as the time from initiation of therapy until passage of the last unformed stool after which subjects met criteria to be declared well. By definition wellness occurred when there was (1) passage of no unformed stools in a 48‐hour interval with no fever, with or without other clinical symptoms or (2) passage of no watery stools and no more than two soft stools in a 24‐hour interval without fever or other clinical symptoms of enteric infection. Mild excessive gas and mild flatulence were not considered symptoms of continuing illness. Subjects who passed no additional unformed stools after initiation of study therapy were defined as having duration of diarrhea of 0 hour. For analysis purposes, subjects who could not be defined as well during the 4‐day observation period were assigned a time to last unformed stool of 97 hours, as were subjects declared treatment failures at any time during the observation period.

Secondary end points included the number of unformed stools passed and degree of clinical symptoms experienced in 24‐hour intervals after initiation of treatment. The numbers of subjects who failed treatment were compared. Treatment failure was defined as (1) worsening of, or failure to improve, clinical symptoms after at least 24 hours of therapy or (2) illness continuing after 72 hours. All subjects declared treatment failure were analyzed as if they had continuing illness longer than 96 hours. Bacterial eradication was defined as absence of the pretreatment etiologic agent in the post‐treatment stool sample. A potential adverse event was any clinical finding or patient complaint recorded on the daily diary or voiced at clinic visit that was not present in the 24‐hour period before enrollment in the trial.

Statistical analyses were based on the intent‐to‐treat principle. All subjects with the exception of dropouts were included in the analysis of treatment groups to which they were randomized. Sample size determination was based on experience with previous studies comparing antibiotic treatment with antibiotic plus loperamide in which sample sizes permitted detection of superiority of the loperamide containing arm in decreasing the average duration of diarrhea from over 24 hours to less than 12 hours in a three‐way comparison at a power of 0.8. Significance was defined as a p value of less than or equal to 0.05. The study was not powered to detect superiority of the 1,000 mg of azithromycin dose or noninferiority of the 500 mg dose for which the sample size would have needed to be nearly doubled. 20

Duration of post‐treatment diarrhea was analyzed by the Gehan–Wilcoxon test. For analysis of clinical signs and symptoms of enteric illness, and proportions with treatment failure, bacteriologic cure, and adverse events, the chi‐square test or the Fisher’s exact test was used. Average number of unformed stools was compared using the Student’s t‐test.

Results

Table 1 shows comparative demographics, patients’ presenting clinical features, and results of fecal testing and bacteriology at baseline for the three treatment groups. No differences in baseline characteristics were documented except that mild nausea was more common in the combination treatment group (p= 0.035) and blood in stools was more common in the 1,000 mg azithromycin group (p= 0.032). Most students who dropped out of the study simply did not return to clinic and did not hand in any diaries; otherwise, all were doing well clinically when they were questioned by telephone. One student in the 1,000 mg azithromycin group vomited 5 minutes after taking the medicine and decided not to participate further. One student in the 500 mg azithromycin group complained of bloating on day 1 and constipation on days 1 and 2 and decided to drop out of the study on day 2; diarrhea was not worsening.

View this table:
Table 1

Comparison of demographics, clinical features of illness at enrollment and fecal testing, and bacteriology results in the three treatment groups

Treatment group
Comparative featuresAzithromycin (1,000 mg)Azithromycin (500 mg)Azithromycin (500 mg) plus loperamide
No. of enrolled576059
No. of dropouts743
No. of analyzed505656
Percent females505461
Average age (y)242323
Clinical features 24 h prior to enrollment
 Average hours ill514635
 Average no. of unformed stools5.55.94.9
No. of with no/mild/moderate/severe symptoms
 Nausea25/14/8/330/14/8/420/26*/4/6
 Vomiting45/2/1/249/5/2/048/4/1/3
 Abdominal cramps5/16/23/611/18/19/87/22/19/8
 Gas23/14/11/220/21/13/218/25/9/4
 Urgency14/17/11/214/23/11/89/22/17/8
 Tenesmus32/12/5/141/10/5/042/6/8/0
 Fecal incontinence49/1/0/054/2/0/052/3/1/0
 Temperature elevation49/1/0/051/3/1/156/0/0/0
Percent of enrollment stools positive for
 Fecal leukocytes434638
 Mucus565553
 Occult blood19*711
No. of enrollment stools with
Shigella363
Enterotoxigenic E coli
  ST567
  LT797
  ST/LT toxins5712
 EAEC4813
Providencia042
Cryptosporidium536
 No pathogen isolated262217
 Mixed organisms588
ST/LT, EAECST/LT, EAEC (2)ST/LT, EAEC (3)
ST/LT, CryptosporidiumST, EAECST, EAEC
ST, CryptosporidiumLT, EAEC (2)LT, EAEC, Cryptosporidium (2)
LT, EAEC (2)ST/LT, ProvidenciaST/LT, Cryptosporidium
Shigella, LT, EAECShigella, LT
Shigella, LT
  • ST = heat stable toxin; LT = heat labile toxin; EAEC = enteroaggregative E coli.

  • * All comparisons were not significantly different except for more patients in the azithromycin plus loperamide group with mild nausea (p= 0.035) and more patients in the azithromycin 1,000 mg group with blood in stools (p= 0.032).

Bacterial enteropathogens tested for susceptibility to azithromycin and ciprofloxacin included enteroaggregative E coli (n= 22); enterotoxigenic Ecoli producing heat labile (LT) toxin (n= 37), heat stable (ST) toxin (n= 26), or both ST and LT toxins (n= 26); and Shigella (n= 7). Azithromycin was highly active against all E coli enteropathogens with a minimal inhibitory concentration (MIC90) for 90% of isolates of 0.03 μg/mL. For Shigella, the MIC90 was 4 μg/mL. The MIC90 of ciprofloxacin for all enteropathogens was 0.16 μg/mL.

Table 2 shows the comparative outcomes for the three treatment groups. The average number of hours after beginning therapy to passage of the last unformed stool was significantly fewer for the combination treatment group (11 h) than for either antibiotic treatment group (34 h; p= 0.0002). Likewise, the duration of diarrhea was shorter in the combination‐treated group (8 h) than either antibiotic‐treated group (20 and 16 h; p= 0.006) among those who became well within 72 hours. Similar significant differences favored combination treatment in the subset analyses of those with any bacterial enteropathogen (p= 0.0001) or no pathogen isolated (p= 0.0001). Combination therapy appeared to be more effective against Shigella and Cryptosporidium; however, the numbers were small and the outcomes were not statistically different.

View this table:
Table 2

Comparative efficacy and tolerability of single 1,000 or 500 mg doses of azithromycin and the combination of loperamide and a single 500 mg dose of azithromycin and in the treatment of acute travelers’ diarrhea

Treatment group
Azithromycin (1,000 mg)Azithromycin (500 mg)Azithromycin (500 mg) plus loperamide
N505656p Value
Average duration (h) of diarrhea after beginning treatment*
 All cases3434110.0002
  All bacterial causes (n)33.6 (21)38.1 (31)9.2 (36)0.0001
  No pathogen isolated (n)33.4 (26)27.6 (22)14.5 (17)0.008
  Shigella (n) 19.2 (3)26.3 (6)11.8 (3)NS
  Cryptosporidium (n) 46.8 (5)32.5 (3)24.6 (6)NS
Average duration (h) of diarrhea among those becoming well within 72 h 20.416.48.20.006
Percent well§ at hours of study
 0818500.0002
 6183466
 12384679
 18405079
 24545984
 48787393
 72807996
Average no. of unformed stool passed (h)
 0–244.32.51.2<0.0001
 25–481.51.20.30.0008
 49–720.50.60.20.07
Treatment failures defined as
 No. of persisting ill after 72 h682
 No. of declared treatment failures during 72 h440
 Total (%)10 (20)12 (21)2 (4)0.01
Average no. of 2 mg doses of loperamide taken after the 4 mg loading dose3.02.10.90.0001
No. with possible adverse event
 Rash011NS
 Headache254NS
 Dizziness010NS
  • NS = not significant; LT = heat labile toxin; EAEC = enteroaggregative E coli.

  • * Subjects continuing to pass loose stools within the 73‐ to 96‐hour period of observation were assigned 97 hours for their duration of diarrhea.

  • Includes two cases coinfected with LT and LT/EAEC (group 2) and one case coinfected with LT (group 3); see Table 1.

  • Seventy‐two hours is the longest period during which subjects can have an accurate duration of diarrhea calculated given a 4‐day period of observation.

  • § Wellness is defined as having passed the last unformed stool. See text for detailed definition.

Table 2 also shows that 50% of subjects treated with azithromycin and loperamide passed no unformed stools after beginning therapy compared with 8 and 18% of subjects in the azithromycin treatment groups. Differences favoring combination treatment persisted at 72 hours. The number of treatment failures was significantly fewer in the combination‐treated group than the antibiotic‐treated groups (Table 2).

The average number of unformed stools passed per 24‐hour period after beginning treatment (Table 2) was significantly fewer in the combination‐treated group compared with the antibiotic‐treated groups in both the 0‐ to 24‐hour and the 25‐ to 48‐hour periods. Comparing the combination‐treated group and the antibiotic‐treated groups combined, the difference in number of unformed stools passed per 24 hour averaged 2.2 during the 0‐ to 24‐hour period and 1.0 during the 25‐ to 48‐hour period.

Table 2 also shows the number of 2 mg doses taken after the 4 mg loading dose. On average, those taking active loperamide took significantly fewer doses (0.9; p= 0.0001) than did those taking placebo in the antibiotic‐alone treatment groups (3.0 and 2.1).

Analysis of symptoms that were present at enrollment showed a benefit of loperamide in relief of cramps. At baseline, 53% (56/106) and 48% (27/56) of subjects treated with antibiotic alone and combination therapy, respectively, manifested moderate to severe cramps. After 1 day of therapy, the percentages had dropped to 45% (antibiotic alone) and 27% (combination therapy; p < 0.02). Symptoms of enteric disease (eg, nausea, vomiting, abdominal cramps, gas, urgency, or tenesmus) that were not present at enrollment appeared in 42% of subjects (44/106) treated with antibiotic alone compared with only 18% of subjects (10/56) who also received loperamide (p < 0.003). The occurrence of such symptoms despite antibiotic treatment was considered likely manifestations of disease rather than adverse events. The occurrence of 57% (42 − 18/42 × 100) fewer such symptoms in the loperamide‐treated subjects was felt to represent a benefit of loperamide treatment.

Therapy was well tolerated (Table 2). Two subjects developed rash that was self‐limiting. Eleven subjects complained of headache that did not necessitate specific treatment. One subject complained of dizziness.

Table 3 shows the percent of eradication at day 5 of the initial enteropathogen from stool for all treatment groups. Shigella was eradicated from all stools in all treatment groups. About 94% of all E coli were eradicated from stools of subjects treated with 1,000 mg of azithromycin compared with 85 and 89% in the 500 mg and combination groups, respectively; however, these differences were not significantly different. Failure of eradication of E coli from the day 5 stool did not correlate with adverse clinical outcome, in keeping with previous study showing the lack of correlation of clinical outcome with results of test‐of‐cure cultures.

View this table:
Table 3

Eradication of enteropathogens from stool

All treatment groups
OrganismNo. of culture pairs*Percent eradicated
Enterotoxigenic E coli
 ST1587
 LT1794
 ST/LT2488
EAEC2286
All E coli7888
Shigella8100
Providencia367
Cryptosporidium1050
  • ST = heat stable toxin; LT = heat labile toxin; EAEC = enteroaggregative E coli.

  • * Enrollment isolate accompanied by a day 5 test‐of‐cure stool sample.

Because the benefit of loperamide treatment—expressed as passage of an average of 2.2 fewer unformed stools in the first 24 hours and 1.0 fewer unformed stools in the second 24 hours—is intuitively not perceived as a clinically important benefit, we decided to do a subset analysis to better understand the benefits of loperamide. Table 4 shows analysis of the number of patients over time, who continue to pass three or more (enrollment criteria) or six or more (moderate to severe disease) unformed stools in the antibiotic‐alone and combination treatment groups. After the first day of treatment with antibiotic alone, 50% of subjects still met study entry criteria and were passing an average of 5.6 stools. Only 20% of those treated with loperamide continued to meet study criteria. The differences were particularly impressive among subjects continuing to have moderate to severe diarrhea in the first 24 hours, with 20% in the antibiotic‐alone group continuing to pass an average of 8.7 unformed stools compared with only 1.7% of those in the loperamide treatment group continuing to pass an average of 8 unformed stools. Significant differences persisted into the second 24‐hour period after treatment (Table 4).

View this table:
Table 4

The effect of addition of loperamide to azithromycin on the number of patients continuing to pass unformed stools, and the average number of unformed stools passed, during the first and second day of therapy

Period of studyAzithromycin alone*Azithromycin plus loperamidep Value
Baseline
N10656
 Average no. of unformed stools passed5.74.9NS
0–24 h
N10556
 Average no. of unformed stools passed3.41.2<0.0001
 Percent of patients passing three or more unformed stools5020<0.001
 Average no. of unformed stools/patient passing three or more5.64.0
 Percent of patients passing six or more unformed stools201.70.001
 Average no. of unformed stools/patient passing six or more8.78.0
25–48 h
N10456
 Average no. of unformed stools passed1.30.30.0008
 Percent of patients passing three or more unformed stools1640.018
 Average no. of unformed stools/patient passing three or more5.23.0
 Percent of patients passing six or more unformed stools700.048
 Average no. of unformed stools/patient passing six or more7.60
  • NS = not significant.

  • * Because no significant differences existed in the clinical outcomes of subjects treated with 1,000 or 500 mg of azithromycin, subjects in these two groups were lumped to form the “azithromycin‐alone” group for purposes of this comparison with azithromycin plus loperamide.

Discussion

In a previous study of travelers’ diarrhea conducted in Mexico, a region of the world with disease caused predominately by E coli, a single 1,000 mg dose of azithromycin proved to be as effective as single 500 mg dose of levofloxacin. 12 In the present study, a single 500 mg dose of azithromycin appeared to be equivalent to a single 1,000 mg dose with an average duration of diarrhea of 34 hours following either dose. While this study was not powered to prove statistically either equivalence or noninferiority, the likelihood of a true difference is remote given the equal outcomes. Compared with either dose of azithromycin, the addition of loperamide to a single 500 mg dose of azithromycin proved most effective in limiting the average duration of diarrhea to 11 hours compared with 34 hours. Similar benefits were seen in the subsets of all bacterial illness and no pathogen isolated. While differences favored combination therapy in the groups with Shigella and Cryptosporidium disease, the numbers were too small to be statistically significant. Compared with treatment with azithromycin alone, the addition of loperamide relieved abdominal cramps better. Use of loperamide also substantially prevented the onset of enteric symptoms, which were not present at enrollment but appeared over the next 24 hours despite beginning antibiotic treatment. Other than self‐limited rash attributed to the azithromycin in two subjects, other potential adverse effects were likely attributable to underlying disease.

The benefits of the combination of an antibiotic and loperamide have been documented in earlier published studies of TMP/SMX 1,2 and the fluoroquinolone ofloxacin 3,4 in the same population in Mexico. Loperamide also afforded added clinical benefit in combination with rifaximin. 21 In two studies, the benefits of loperamide combined with ciprofloxacin were less apparent, but the use of loperamide was considered safe. 22,23

In a region of the world like Mexico, where E coli are the prevalent organisms causing travelers’ diarrhea and where resistance to TMP/SMX has become a clinically important problem, 9 excellent empiric therapy of travelers’ diarrhea can be achieved with the combination of loperamide with rifaximin or a fluoroquinolone. The present study shows that loperamide plus 500 mg of azithromycin is an acceptable option in this setting.

In some regions of the world like Southeast Asia, invasive pathogens are relatively common, and Campylobacter is both prevalent and resistant to TMP/SMX and fluoroquinolones. 5,8 Although rifaximin is active against invasive pathogens, including Campylobacter, it is not approved to treat invasive pathogens. Therefore, the drug of choice in regions like Southeast Asia is azithromycin. The addition of loperamide to an antibiotic in the treatment of invasive pathogens is likely beneficial as evidenced by the benefit of adding loperamide to ciprofloxacin to the treatment of Shigella or enteroinvasive E coli disease. 24 Previous study 15 documenting the benefits of azithromycin in regions like Southeast Asia, however, used a total dose of azithromycin greater than 500 mg, so the benefits of 500 mg of azithromycin plus loperamide in the present study should not be extrapolated as beneficial in Southeast Asia until additional research is done in this region. Likely, however, higher doses of azithromycin, as previously studied, plus loperamide would be more efficacious that azithromycin alone in the treatment of travelers’ diarrhea in Southeast Asia.

Criticism of the addition of loperamide to an antimicrobial agent has been that the benefit is, on average, only one or two loose stools per 24‐hour period. For this reason, some authorities suggest adding loperamide to an antimicrobial agent only under special circumstances such as when toilet facilities are not readily available or when participation in an important meeting should not be interrupted by the urgent need to have a bowel movement. The problem with such criticism is that the calculation of an average benefit of one to two loose stools fails to indicate how much benefit might actually be realized by a subset of the population. In the present study, 50% of those receiving azithromycin alone continued to meet study enrollment criteria in the 24 hours after beginning treatment and 20% had moderate to severe diarrhea (averaging 8.7 unformed stools). During this same period, 20% of loperamide‐treated subjects continued to meet enrollment criteria and only 1.7% continued to have moderate to severe disease. Although the differences were smaller, the relative benefits of loperamide continued into the second day post‐treatment. To realize fully these substantial clinical benefits, we feel the data support the routine addition of loperamide to an antimicrobial agent in the empiric treatment of patients with travelers’ diarrhea as defined in the present study.

Declaration of Interests

C. D. E., H. L. D., and P. C. O. are consultants for Salix Pharmaceuticals and have received honoraria for speaking for Salix Pharmaceuticals (Morrisville, NC, USA). C. D. E., H. L. D., P. C. O., Z.‐D. J., and M. W. D. have received grants from Salix Pharmaceuticals, Pfizer (New York, NY, USA), Optimer (SanDiego, CA, USA), and Ioma (New York, NY, USA) through the University of Texas, Houston, Texas.

Acknowledgments

The study is registered at ClinicalTrials.gov (No. 2006‐08‐01). The study was funded by the investigators. The authors wish to acknowledge Dr F. Javier de la Cabada and Dr Francisco Martinez‐Sandoval for their continued support of our research in Mexico.

Footnotes

  • Presented at the 9th Conference of the International Society of Travel Medicine, Lisbon, Portugal, May 1 to 5, 2005.

References

View Abstract