Probiotic Escherichia coli Nissle 1917 Versus Placebo for Treating Diarrhea of Greater Than 4 Days Duration in Infants and Toddlers

Abstract

Background: 

Administering probiotics can prevent or cure some forms of diarrhea. The efficacy of probiotic Escherichia coli Nissle 1917 (EcN) in infants and toddlers with diarrhea >4 days was tested by a double-blind trial.

Methods: 

One hundred fifty-one children aged 1–47 months with nonspecific diarrhea were randomized to receive either EcN suspension (N = 75) or placebo (N = 76). Diarrhea had to meet the following definition: >3 watery or loose nonbloody stools in 24 hours of a diarrheal episode persisting for >4 consecutive days but ≤14 days. All children were well nourished or only moderately malnourished, mildly dehydrated, and received oral rehydration at study commencement. They were treated orally with 1–3 mL EcN suspension (1 mL contains 10⁸ viable cells) or placebo daily for 21 days. Primary objective was to confirm a better response rate (reduction of daily stool frequency to ≤3 watery or loose stools over ≥4 days) with EcN.

Results: 

The 7-day response was higher for the EcN group than placebo (EcN 78.7%, placebo 59.2%). Significant differences were observed on days 14 (EcN 93.3%, placebo 65.8%, P = 0.0017) and 21 (EcN 98.7%, placebo 71.1%, P < 0.001). Kaplan-Meier survival analysis resulted in a significant difference of 3.3 days between the groups (P < 0.0001); median time to response for EcN was 2.4 and 5.7 for placebo. EcN was safe and well tolerated.

Conclusions: 

In the conditions of this trial EcN was a suitable remedy for diarrhea >4 days in young children.

There is accumulating evidence that administering probiotics can prevent or cure some forms of diarrhea. This has been shown in several clinical studies in children and adults as summarized in recently published meta-analyses.^1–5 Different probiotics such as lactobacilli and bifidobacteria as well as Saccharomyces boulardii have been used. The circumstances differ when children, for example, with acute gastroenteritis are treated or when adults try to prevent diarrhea after antibiotic therapy or when traveling. The therapeutic goal is to shorten the time of illness and to relieve the symptoms.

The nonpathogenic Escherichia coli strain Nissle 1917 (EcN) has been used in Europe as a licensed probiotic drug for chronic inflammatory, functional, and infectious bowel diseases for almost 90 years. In 1917, the strain was isolated by A. Nissle from the feces of a soldier in the First World War who did not develop enterocolitis in a highly infested surrounding, and it was subsequently shown to inhibit the growth of enteropathogens.⁶ The strain was introduced into medical practice by Nissle and was initially used to treat diarrhea. Since then EcN has shown therapeutic success in treating diarrhea in clinical practice.⁷ Some open-label trials and studies according to good clinical practice (GCP) were done to confirm this.^8–10 One such study focused on the treatment of acute diarrhea directly after onset of disease.⁹ In addition, EcN was evaluated in fullterm and preterm children with regard to tolerance, immunomodulation, and the prevention of colonization with pathogenic microbes.^11–13

EcN is available as enteric-coated capsules and as a suspension. The present trial on EcN suspension was conducted in infants and toddlers with diarrhea lasting for >4 days. The aim was to evaluate the response rate and time to response using EcN compared with placebo.

MATERIALS AND METHODS

This was a multicenter, prospective, randomized, double-blinded placebo-controlled, phase III clinical trial. The study was performed in accordance with the requirements of GCP and the Revised Declaration of Helsinki. The study was approved by the Independent Ethics Committee of the Federal Agency of Drugs Quality Control, Russia, and by the Independent Ethics Committee of the State Enterprise Center of Immunobiological Medicines at the Ministry of Health of Ukraine.

Eligible participants were recruited in 11 university clinics in Moscow, Russia, Kiev, and Dnepropetrovsk, Ukraine, between February and April 2005. All were treated as outpatients. Before admission to the trial, each patient's legal guardian was informed by the investigator about the nature, significance, and possible consequences of the trial and its procedure as well as about the efficacy and adverse drug effects of the trial medication, and written consent was obtained.

One hundred fifty-one infants and toddlers between 1 and 47 months of age with nonspecific diarrhea were admitted to the study. Only those patients were selected whose diarrheal symptoms met the following definition: more than 3 watery and loose stools in 24 hours of a nonbloody diarrheal episode which persisted for more than 4 consecutive days but no longer than 14 days. All patients were white, either well nourished or only moderately malnourished (midarm circumference according to Pirlich et al¹⁴), and did not show signs of any other organic or infectious disease. Seventy-five patients were allocated to the EcN group, and 76 to the placebo group (Fig. 1). All study personnel and participants were blinded to treatment assignment for the duration of the study. Only the independent statisticians and an Independent Data Monitoring Committee (IDMC) saw unblinded data, but none had any contact with study participants. Inclusion criteria were age <4 years at the time of enrolment, diarrhea per definition given above, moderate dehydration (5–10% loss of body weight), and written informed consent by each patient's legal guardian. Exclusion criteria comprised participation in another clinical trial, intake of EcN within the past 3 months before enrolment, of food supplements, or drugs which contain living microorganisms or their metabolic products within 7 days before enrolment or during the trial, and of antibiotics or other drugs for treatment of the current diarrheal episode. In addition, severe dehydration (>10% loss of body weight), weight or growth below the 5th percentile or a shift downward in the weight or growth percentiles that crosses 2 or more percentiles, breast-feeding, premature birth, as well as severe or chronic disease of the bowel or severe concomitant diseases were excluded by the protocol.

FIGURE 1.: Flow diagram of participants in the study.

The randomization schedule was generated using the method of randomly permuted blocks (block size 4, software SAS, version 9.1).

Baseline Data.

No relevant differences were observed between the 2 treatment groups regarding demographic and other baseline characteristics such as gender, age, height, weight, and the state of nutrition of the infants (Table 1). There was also no difference in systolic and diastolic blood pressure as well as heart rate and body temperature between the treatment groups at baseline. Standard microbiologic analyses based on PCR¹⁵ were employed to check for various pathogens (eg, Noroviruses, Adenoviruses, Rotaviruses, Campylobacter, Salmonella, Yersinia, Shigella, and pathogenic E. coli strains). With regard to the case history and to standard microbiology in the participating hospitals at baseline, possible reasons for the current diarrheal episode are listed in the Table 1.

TABLE 1: Demographic and Other Baseline Characteristics

Study Medication.

The investigational drug was a commercially available suspension for oral use containing nonpathogenic Escherichia coli strain Nissle 1917 (Mutaflor Suspension, Ardeypharm, Herdecke, Germany; 10⁸ viable microorganisms per mL). As placebo an identical preparation, a suspension devoid of the active substance, was used. Depending on age, daily doses of the study medication (EcN or placebo) were:

• Infants <1 year → 1 mL once daily
• Toddlers ≥1 to ≤3 years → 1 mL twice daily
• Toddlers >3 to <4 years → 1 mL 3 times daily

Doses were applied for 21 days.

Study Design.

Each patient stayed in hospital for the replenishment of fluid and electrolytes using oral rehydration treatment for 4 hours before therapy commencement. Rehydration treatment was administered according to the WHO manual “Treatment of Diarrhea” (2001). After admission examination by the investigators, data with regard to drinking behavior, stool frequency, stool consistency, admixture of blood and mucus, frequency of abdominal pain and cramps, as well as vomiting were entered by the parents at home using a diary. With the next follow-up visit to the hospital (visits due on days 3, 5, 7, 10, 14, and 21) these data were checked and, together with additional data generated at the visit (eg, body weight, body temperature, general state of health), transferred to an electronic case report form by the investigators. Empty medication containers were returned to the centers to check for patient compliance. If more than 15% of the trial medication was not administered to the patient, this was regarded as a protocol violation. Also with the control visits, adverse events (AEs) were recorded by the investigators. All study-related procedures strictly followed the current International Conference on Harmonization guidelines regarding GCP.

Evaluation.

The primary efficacy variable of this study was the treatment response defined as a reduction of stool frequency to ≤3 watery or loose stools within each 24 hours for a period of at least 4 consecutive days. The primary end point was the response rate. Secondary endpoints included the time to response, improvement of the clinical symptoms (stool frequency, stool quality with regard to admixture of blood and mucus, frequency of abdominal pain/cramps and vomiting, body temperature), general state of health (“very good,” “good,” “moderate,” or “poor”), occurrence of AEs, and tolerance to the study medication (“very good,” “good,” “moderate,” or “poor”).

Statistical Analyses.

The study was conducted according to a 3-stage group sequential test design (O'Brien/Fleming type) with possible sample size adaptation¹⁶ after the 2 planned interim analyses (first and second interim analysis after 50 and 100 patients, respectively, final analysis after 150 patients). The 2 primary efficacy variables were the response rates after 7 and 14 days of treatment, respectively. To show the superiority of EcN versus placebo treatment the response rates after 7 and 14 days of treatment were compared (overall type I error rate α = 0.025 one-sided). Because of testing 2 primary efficacy variables, the nominal type I error rate for each variable was set equal to 0.0125 (one-sided). The statistical comparison of EcN versus placebo of the response rates after 7 days as well as after 14 days was performed using the 2-sample χ² test at the interim analyses and final analysis (inverse normal method for combining the P values of the stages).

Further, the following exploratory analyses were conducted. The response rates after 21 days of treatment were computed and compared between treatment groups by means of 2-sample χ² test for rates (one-sided). In addition, a time-to-response analysis was performed (Kaplan-Meier method). An exploratory comparison of the time-to-response curves was performed using the log-rank test (one-sided). Data from all assessments were tabulated using descriptive statistical methods. The primary analysis was performed on the full analysis set (based on the intention-to-treat principle, ITT) including all randomized patients who took at least one dose of study medication. In the per-protocol analysis (PP) patients with major protocol violations were excluded. The analysis sets were defined in a blind review of the data.

The sample size was estimated prospectively as follows: assuming response rates of EcN of 45% and 50% after 7 and 14 days, respectively, and assuming a 25% difference to placebo for both endpoints, the statistical power of the test procedure was expected to be at least 80%, if each of the 3 test stages consisted of 25 patients per treatment group (α = 0.0125 one-sided, interim analyses after 1/3 and 2/3 of the planned number of patients). IDMC was responsible for reviewing unblinded results of the interim analyses and for giving recommendations to continue the study with the preplanned or a reassessed sample size or to stop the study. The IDMC operated under a written charter that included well-defined specific Standard Operating Procedures. Two interim analyses were performed. No sample size adaptation was required. IDMC recommended to continue the study until the preplanned sample size of 150 patients was reached.

RESULTS

All efficacy analyses were performed on the ITT (N = 151) and PP (N = 149) data sets. As the ITT and PP analyses were almost identical (2 placebo-treated patients were excluded from PP analysis because of an irregular duration—3 and 19 days, respectively—of the current diarrheal episode), all results supplied here relate to the ITT data set.

Of 151 patients recruited for the study, 84.8% showed a response to treatment within the 21-day period of treatment and observation. The number of infants showing response to treatment on day 7 was higher in the EcN group than in the placebo group [EcN 59 of 75 patients (78.7%), placebo 45 of 76 (59.2%), P = 0.0758]. Significant differences were observed on days 14 [EcN 70 of 75 (93.3%), placebo 50 of 76 (65.8%), P = 0.0017; 97.5% CI: 0.05–0.38] and 21 [EcN 74 of 75 (98.7%), placebo 54 of 76 (71.1%), P < 0.001, exploratory].

Kaplan-Meier survival analysis resulted in a significant difference of 3.3 days between the groups (P < 0.0001); median time to response for EcN was 2.4 days (95% CI: 1.53–3.99) and 5.7 days (95% CI: 3.79–7.24) for placebo (Fig. 2).

FIGURE 2.: Time to response—Kaplan-Meier analysis (ITT).

At the end of study, the efficacy of the study medication was better in the EcN group than in the placebo group. The number of stools showed an almost identical decrease from baseline to the end of study in both treatment groups (EcN −3.9 ± 1.1 stools, placebo −3.7 ± 1.2 stools) but the decrease proceeded faster in EcN patients than in placebo patients as indicated by a median number of stools ≤3 per day, which occurred for the first time on day 5 in the EcN group and on day 7 in the placebo group.

No patient had blood in the stool during the study in either treatment group. The number of patients with mucus, abdominal pain, or cramps decreased from baseline to the end of study (Table 2). In both groups, only a few patients had vomiting at baseline. Body temperature showed an almost identical decrease over the course of the study in both treatment groups (EcN −0.4 ± 0.4°C; placebo −0.3 ± 0.4°C). From baseline to the end of study, the body weight increased and the status of dehydration improved in both treatment groups. The increase in body weight was slightly greater in the EcN group (0.7 ± 0.4 kg) than in the placebo group (0.5 ± 0.4 kg), and at the end of study there were more patients with moderate dehydration in the placebo group [11 of 76 patients (14.5%)] than in the EcN group [1 of 75 patients (1.3%)]. The general state of health markedly improved in the EcN group during the course of the study. In 97.0% of EcN patients initially assessed as having a “moderate” state of health, this assessment improved to “good” or “very good” in the course of the treatment. In the placebo group the respective figure was 62.3%.

TABLE 2: Abdominal Symptoms Before and After Therapy for Diarrhea

Three of 75 patients (4.0%) in the EcN group and 4 of 76 patients (5.3%) in the placebo group experienced a total of 11 AEs (EcN: limb injury, traumatic hematoma, hypersensitivity; placebo: influenza, vomiting, abdominal pain, cough, rhinorrhea, dermatitis, viral upper respiratory tract infection, nasopharyngitis). None of these AEs were rated serious or severe but 4 of them were assessed as being possibly related to the study medication (EcN: hypersensitivity; placebo: vomiting, abdominal pain, and dermatitis). One patient in the placebo group was withdrawn because of influenza. Three AEs (EcN: hypersensitivity; placebo: influenza, dermatitis) still existed at study termination. In the other 8 AEs complete recovery was achieved.

EcN was found safe and well tolerated. There was no difference between the ratings by the investigators and the parents (Fig. 3).

FIGURE 3.: Assessment of tolerance to therapies by investigator and parents.

DISCUSSION

In adults it may not always be necessary to consider antidiarrheal treatment because of the self-limiting character of the disease. For infants and toddlers an early, evidence-based treatment is important to improve the quality of life (both of the infant and the parents) and may even save lives. The aim of this multicenter, prospective, randomized, double-blinded placebo-controled phase III study was to investigate the therapeutic efficacy of orally administered E. coli strain Nissle 1917 (EcN) in treating nonspecific diarrhea >4 days of duration in infants and toddlers. EcN showed superiority to placebo after 14 and 21 days of treatment. In our study, only patients with a history of diarrhea of at least 4 consecutive days were included to exclude patients who were likely to develop spontaneous healing. Within the first 3 days of treatment there was a difference in the response rates. EcN therapy improved symptoms markedly and shortened the duration of diarrhea by 3.3 days, which was statistically significant and clinically relevant. Diarrhea of >4 days has not been thoroughly studied in clinical trials to date. However, there is a chance for the untreated patient to develop chronic disease. After 21 days of EcN therapy 74 of 75 patients responded to the treatment while diarrhea persisted in almost one-third of the placebo-treated patients. As a result, EcN will possibly have a preventive effect on chronification of disease.

The efficacy of orally administered EcN in treating diarrhea was confirmed by our group in a recent multicenter, prospective, randomized, double-blinded placebo-controled phase III study.⁹ In that study, children with acute diarrhea also showed onset of treatment response earlier in the EcN group (median of 2.5 days) than in the placebo group (median of 4.8 days). The time saved by EcN in comparison with placebo was 2.3 days (P = 0.0007, Kaplan-Meier analysis). The response rate was higher in the EcN group (94.5%) than in the placebo group (67.2%) (P < 0.001). In contrast, recent data obtained with various probiotic preparations in acute diarrhea reveal limited efficacy for most of the strains tested.¹⁷ With regard to shortening the diarrheal duration none of these preparations reached 2 days.

The data presented here confirm the experience of other researchers using different probiotics for the treatment of gastroenteritis. A remarkable number of clinical trials have been published in this field (eg, using lactobacilli, bifidobacteria, and Saccharomyces boulardii^1–5). Most showed improvement of the clinical picture and a reduction in the duration of diarrheal symptoms by 0.5–1.5 days. However, it is impossible to directly compare the efficacy results from different clinical trials because of differences in the study populations, baseline nutritional status, and duration of disease before enrolment.

We are well aware that the assessment of the state of health by the investigators and parents is, to some extent, subjective, and also the rates of abdominal pain and cramps are difficult to determine in very young children. However, they were evaluated as secondary endpoints and help to better understand the full clinical picture of the disease with therapy. They underscore the significance of the primary findings. The same applies to the assessment of tolerance to treatment that was done in addition to the recording of AEs. In this context, the dosage of EcN used in this trial bases on the manufacturer's recommendation that was backed up by data from literature.^9,11–13

The E. coli strain Nissle 1917 is one of the best characterized strains used as a probiotic drug. Different strain-specific characteristics are described (eg, 6 iron-acquisition systems), secretion of 2 microcins, formation of biofilms under various conditions, and a unique structure of the lipopolysaccharide.^18–21 EcN has been shown to protect human intestinal epithelial cells from invasion by different enteroinvasive bacterial pathogens.²² It has also been shown that EcN modulates the immune response (eg, by the stimulation of IgA secretion by lymphocytes)¹³ and induction of human β-defensin-2 in enterocytes.²³ Most recently, EcN was able to prevent acute secretory diarrhea in a pig model of intestinal infection.²⁴ It seems reasonable to further explore the antagonistic activity of EcN against different pathogenic microbes in vivo in patients.