J Antimicrob Chemother 2017; 72: 3177–3180

doi:10.1093/jac/dkx254 Advance Access publication 23 August 2017

Downloaded from https://academic.oup.com/jac/article-abstract/72/11/3177/4091566 by Medical Library, Kaoshiung Veterans General Hospital user on 11 December 2018
A randomized controlled trial of probiotics for Clostridium difficile
infection in adults (PICO)
Anna K. Barker1, Megan Duster2, Susan Valentine2, Timothy Hess2, Laurie Archbald-Pannone3,4, Richard Guerrant4
and Nasia Safdar2,5,6*

1
Department of Population Health Sciences, School of Medicine and Public Health, University of Wisconsin, Madison, WI, USA; 2Division
of Infectious Disease, Department of Medicine, School of Medicine and Public Health, University of Wisconsin-Madison, Madison, WI,
USA; 3Division of General, Geriatric, Palliative and Hospital Medicine, Department of Medicine, School of Medicine, University of Virginia,
Charlottesville, VA, USA; 4Division of Infectious Diseases and International Health, Department of Medicine, School of Medicine,
University of Virginia, Charlottesville, VA, USA; 5William S. Middleton Memorial Veterans Affairs Hospital, Madison, WI, USA;
6
Department of Infection Control, University of Wisconsin Hospital and Clinics, Madison, WI, USA

*Corresponding author. Tel: !1-608-213-4075; Fax: !1-608-263-4464; E-mail: [email protected]

Received 22 May 2017; returned 23 June 2017; revised 29 June 2017; accepted 29 June 2017

Background: Clostridium difficile is the most common cause of hospital-acquired infections, responsible for
.450000 infections annually in the USA. Probiotics provide a promising, well-tolerated adjunct therapy to stand-
ard C. difficile infection (CDI) treatment regimens, but there is a paucity of data regarding their effectiveness for
the treatment of an initial CDI.
Objectives: We conducted a pilot randomized controlled trial of 33 participants from February 2013 to February
2015 to determine the feasibility and health outcomes of adjunct probiotic use in patients with an initial mild to
moderate CDI.
Methods: The intervention was a 28 day, once-daily course of a four-strain oral probiotic capsule containing
Lactobacillus acidophilus NCFM, Lactobacillus paracasei Lpc-37, Bifidobacterium lactis Bi-07 and B. lactis Bl-04.
The control placebo was identical in taste and appearance. Registered at clinicaltrials.gov: trial registration
number " NCT01680874.
Results: Probiotic adjunct therapy was associated with a significant improvement in diarrhoea outcomes. The
primary duration of diarrhoea outcome (0.0 versus 1.0 days; P " 0.039) and two exploratory outcomes, total
diarrhoea days (3.5 versus 12.0 days; P " 0.005) and rate of diarrhoea (0.1 versus 0.3 days of diarrhoea/stool
diary days submitted; P " 0.009), all decreased in participants with probiotic use compared with placebo. There
was no significant difference in the rate of CDI recurrence or functional improvement over time between treat-
ment groups.
Conclusions: Probiotics are a promising adjunct therapy for treatment of an initial CDI and should be further
explored in a larger randomized controlled trial.

Introduction Meta-analyses of randomized controlled trials (RCTs) have

Clostridium difficile is implicated in over 450000 infections and demonstrated that ingestion of specific probiotic strains during
15000 deaths annually in the USA.1 Treatment of C. difficile infec- antibiotic therapy significantly reduces the risk of developing a
tion (CDI) is a major challenge. Approximately 25%–30% of pa- CDI.11–14 However, studies evaluating the effectiveness of pro-
tients treated for a primary CDI experience recurrence.2–4 biotics as treatment during an initial CDI episode are lacking. Thus,
The opportunistic nature of C. difficile allows it to proliferate and we undertook a pilot RCT to determine the feasibility and health
cause infection in the gastrointestinal (GI) tract during microbial outcomes of adding daily probiotic as an adjunct therapy to stand-
disruption. Probiotics are live microorganisms that confer a health ard antibiotic treatment for participants with an initial CDI. We
benefit on the host when administered in adequate amounts.5,6 hypothesized that, compared with placebo, adjunct probiotics
Probiotic therapies pass through the host GI tract while harbouring would lead to a reduction in the duration of C. difficile diarrhoea
a spectrum of potentially protective benefits.7–10 and reduce the likelihood of CDI recurrence.

C The Author 2017. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.
V
For Permissions, please email: [email protected].
3177
Barker et al.

Methods between the treatment arms were conducted using the Wilcoxon rank-

Downloaded from https://academic.oup.com/jac/article-abstract/72/11/3177/4091566 by Medical Library, Kaoshiung Veterans General Hospital user on 11 December 2018
sum test, the v2 test or Fisher’s exact test with a significance level of 0.05.
This Phase 2, pilot, RCT is registered with clinicaltrials.gov (NCT01680874) and
approved by the Health Sciences Institutional Review Board at the University
of Wisconsin (UW) Madison. Study enrolment took place at UW Hospital, Results
a 536 bed tertiary care academic medical centre, from February 2013
to February 2015. The study protocol for this trial was published previously.15 Of the 590 potential participants assessed for eligibility, 43 met in-
clusion and exclusion criteria and 33 enrolled (76.7% response
rate). Two participants, one from each treatment group, were lost
Study participants
to follow-up before they submitted any stool diary data. Thus, the
In accordance with inclusion criteria, all study participants were adults
sample size was 31 for all primary and exploratory diarrhoea out-
experiencing an initial episode of mild to moderate CDI. Exclusion criteria
comes except for CDI recurrence. Among the 31 participants who
included severe CDI, prior history of CDI, other known aetiology of diar-
rhoea, history of chronic intestinal disease, abdominal surgery in the past submitted stool diary data, 10 reported ,7 weeks of data, with an
3 months, presence of ileus, colostomy, gastric or nasogastric tube, a se- average among these participants of 31.5 non-missing days.
verely immunocompromised state, pregnancy, unavailability for follow-up, Recurrence was evaluated in 28 participants who recovered from
unwillingness to terminate other probiotic use, or enrolment in another in- their CDI during the study and submitted follow-up stool samples.
vestigational drug trial.15 Patients were randomized using a random- Baseline characteristics of probiotic and placebo groups were simi-
number generator in a 1:1 ratio in permuted blocks of four. All patients pro- lar (Table 1).
vided informed consent prior to enrolment. The median duration of diarrhoea was 1.0 day longer in the pla-
cebo group than the probiotic group (P " 0.039; Table 2). The differ-
Intervention ence in recurrence was not significant (Table 2). Both exploratory
diarrhoeal outcomes, total diarrhoea days and rate of diarrhoea,
All participants received a placebo or single multi-strain oral probiotic cap-
sule (Lactobacillus acidophilus NCFM, ATCC 700396; Lactobacillus paracasei were significantly worse for participants treated with placebo com-
Lpc-37, ATCC SD5275; Bifidobacterium lactis Bi-07, ATCC SC5220; B. lactis pared with participants treated with probiotics (Table 2).
Bl-04, ATCC SD5219; 1.70%1010 cfu per capsule) daily for 4 weeks. The pills There was no significant difference between the groups con-
were identical in appearance and taste. cerning participants’ functional ability at week 4 or 8 (Table 2) or
total number of adverse events. Almost all (96.8%) participants
Timeline experienced at least one adverse event. GI disorders were the
most common (75% probiotic, 80% placebo).
During the trial, all participants continued standard CDI antibiotic treatment.
Clinic visits occurred at weeks 0, 4 and 8. Participants submitted a fresh stool
sample and daily stool diary, modified from the Bristol Stool Consistency
Scale,16 at those timepoints. Weekly telephone calls were conducted to
Discussion
evaluate treatment adherence, adverse effects, ongoing symptoms and Combination probiotic treatment was associated with significant
functional status, assessed by the Barthel Index of Activities of Daily Living.17 improvement in diarrhoea outcomes for participants, compared
with placebo. Shortening the duration of an initial CDI could allow
Outcomes patients to stop antibiotic therapy sooner, having considerable
Two primary outcomes were duration of diarrhoea and CDI recurrence.
downstream implications for reducing antibiotic resistance.
Diarrhoea was defined as three or more loose stools in 24 h. Duration of To our knowledge, no other studies have investigated probiotics
diarrhoea was the number of consecutive 24 h periods with diarrhoea that as adjunct therapy for initial CDIs. Existing studies focused on pro-
occurred after study enrolment, beginning on the first stool diary day. biotics for the prevention of initial and recurrent CDIs and recurrent
Recurrence was defined as the presence of toxigenic C. difficile in the stool CDI treatment. Transient probiotic GI colonization is a key compo-
of a symptomatic participant, who had previously cleared their CDI. There nent of the biological mechanism reducing CDI in all these clinical
were two secondary outcomes: functional status (weeks 4 and 8) and total situations. Our findings trend in the same direction as recent
adverse events (week 8). meta-analyses evaluating probiotic RCTs in the other C. difficile
Two exploratory diarrhoea outcomes were defined after completion of contexts.11,13,14 However, the effects of probiotics are known to be
data collection, but prior to unblinding the treatment allocation scheme. strain-specific and current guidelines from the IDSA classify pro-
These include the total number of days that participants experienced diar-
biotic therapy as an unresolved issue requiring further study.18
rhoea and the rate of diarrhoea. These measurements were introduced
due to high rates of incomplete stool diary data, which made it desirable to
Future evaluations of the use of this combination probiotic are
conduct an analysis of diarrhoea events using the denominator of person- warranted. Recruitment proved difficult at our single study site.
time. Missing data was defined as three of more consecutive days without The primary barrier was stringent inclusion and exclusion criteria.
a recorded stool diary entry. If 1 or 2 days passed without a stool diary Small sample size is a limitation of this study. Type I error is particu-
entry, we assumed that the participant did not have a bowel movement larly concerning for low powered studies and it is possible that our
during this time and it was not considered missing. The rate of diarrhoea findings do not reflect a true difference between treatment
was calculated as the total number of diarrhoea days divided by the total groups. However, it is reassuring that the results are consistent
number of non-missing stool diary days. across outcome measures. In future studies, utilization of multiple
study sites or a longer study duration would allow for the recruit-
Statistical analysis ment of a larger sample size.
All statistical analyses of main outcomes were performed with the SAS This pilot study revealed feasibility concerns regarding stool
software (version 9.4; Cary, NC, USA) using ITT methodology. Comparisons diary recording. We added a diarrhoea rate outcome to

3178
An RCT of probiotics for C. difficile infection JAC
Table 1. Patient demographics and health status at enrolment

Downloaded from https://academic.oup.com/jac/article-abstract/72/11/3177/4091566 by Medical Library, Kaoshiung Veterans General Hospital user on 11 December 2018
Variable Placebo, N " 15 Probiotic, N " 16 P

Demographics
age (years), median (IQR) 57.0 (42.0–73.0) 65.0 (42.5–70.5) 0.51
race, n (%) 0.37
white 13 (87) 15 (94)
other 2 (13) 1 (6)
female, n (%) 11 (73) 11 (69) 0.78
C. difficile treatment
primary antibiotic treatment, n (%) 0.72
vancomycin 9 (60) 8 (50)
metronidazole 6 (40) 8 (50)
days of antibiotic treatment prior to enrolment, median (IQR) 3.0 (1.0–5.5) 2.0 (1.0–3.3) 0.16
Baseline health status, n (%)
shortness of breath 3 (20) 2 (13) 0.57
angina 1 (7) 1 (6) 0.96
muscle aches 3 (20) 4 (25) 0.67
prior GI surgery 9 (60) 7 (44) 0.37
cough 2 (13) 7 (44) 0.06
diarrhoea 15 (100) 15 (94)
nausea 5 (33) 4 (25) 0.60
frequent urination 0 (0) 0 (0)
painful joints 5 (33) 4 (25) 0.69

Table 2. Study outcomes regarding diarrhoeal symptoms, CDI recurrence and Barthel Index functional score

Variable Placebo, N " 15 Probiotic, N " 16 P

Duration of diarrhoea (cumulative days), median (IQR) 1.0 (0.0–13.0) 0.0 (0.0–2.0) 0.039
Diarrhoea rate (days of diarrhoea/stool diary days submitted), median (IQR) 0.3 (0.1–0.5) 0.1 (0.0–0.2) 0.009
Total number of days with diarrhoea (days), median (IQR) 12.0 (6.0–25.0) 3.5 (1.0–8.0) 0.005
Stool diary number of days, median (IQR) 55.0 (38.0–56.0) 57.0 (37.0–59.0) 0.55
CDI recurrencea, n (%) 0.96
0 12 (92) 14 (93)
1 1 (8) 1 (7)
Total Barthel Index score, mean (SD)
week 4 18.3 (2.7) 18.2 (3.4) 0.86
week 8 19.1 (3.2) 18.8 (2.5) 0.23

a
N " 13 for placebo and N " 15 for probiotic.

account for the differences in person-time reported between at discharge. There was no systematic difference between groups
participants. To minimize reporting bias, we employed a 2 day concerning when diary recording started. This measurement
cut-off to differentiate between gaps in stool diary entries due should be standardized in future studies and include an assess-
to a lack of bowel movements versus incomplete reporting. ment of inpatient hospital records.
Future studies could better engage participants using an elec- Finally, by definition, recurrence could only be assessed
tronic stool diary tool or linking study compensation to diary among the subset of participants who cleared their initial CDI
completion. Both were utilized effectively in a recent study of during the study period and provided follow-up stool samples.
patients with irritable bowel that tracked 90 day stool Bias due to disruption in the randomization scheme is minimized
histories.19 in this study, as only five participants were excluded from the
The duration of diarrhoea measurement was another limita- analysis of CDI recurrence (three placebo, two probiotic). Future
tion, as it could not capture diarrhoea that occurred before the first studies focused on recurrence should evaluate the effect of pro-
stool diary was recorded. The stool diary start date varied, with biotics for prevention in a randomized population of patients, all of
some participants recording in the hospital and others beginning whom have already recovered from a CDI.

3179
Barker et al.

In this pilot study, a combination probiotic was found to de- 3 Pepin J, Alary M-E, Valiquette L et al. Increasing risk of relapse after treat-

Downloaded from https://academic.oup.com/jac/article-abstract/72/11/3177/4091566 by Medical Library, Kaoshiung Veterans General Hospital user on 11 December 2018
crease significantly the duration of CDI diarrhoea compared with a ment of Clostridium difficile colitis in Quebec, Canada. Clin Infect Dis 2005; 40:
placebo control. Additional studies are needed to investigate this 1591–7.
finding in a larger patient population, but these results are promis- 4 Norén T, Akerlund T, Ba € ck E et al. Molecular epidemiology of hospital-
ing. Given the current burden of CDIs in hospitals, even a small de- associated and community-acquired Clostridium difficile infection in a
Swedish county. J Clin Microbiol 2004; 42: 3635–43.
crease in diarrhoea duration for patients is poised to have a
considerable impact on C. difficile transmission and antibiotic 5 Hill C, Guarner F, Reid G et al. Expert consensus document. The
International Scientific Association for Probiotics and Prebiotics consensus
usage and warrants rigorous assessment in efficacy trials.
statement on the scope and appropriate use of the term probiotic. Nat Rev
Gastroenterol Hepatol 2014; 11: 506–14.
6 WHO, Food and Agriculture Organization of the United Nations. Probiotics
Acknowledgements in Food: Health and Nutritional Properties and Guidelines for Evaluation. 2006.
A subset of these results was presented in poster form at the Society for ftp://ftp.fao.org/docrep/fao/009/a0512e/a0512e00.pdf.
Healthcare Epidemiology Conference, St Louis, MO, USA, 2017. 7 Lee CH, Steiner T, Petrof EO et al. Frozen vs fresh fecal microbiota trans-
The study protocol was published previously in the journal plantation and clinical resolution of diarrhea in patients with recurrent
Contemporary Clinical Trials.15 Clostridium difficile infection: a randomized clinical trial. JAMA 2016; 315:
142–9.
8 Forestier C, De Champs C, Vatoux C et al. Probiotic activities of Lactobacillus
casei rhamnosus: in vitro adherence to intestinal cells and antimicrobial prop-
Funding erties. Res Microbiol 2001; 152: 167–73.
This worked was supported by a grant from the National Institute
9 Schneider S-M, Girard-Pipau F, Filippi J et al. Effects of Saccharomyces bou-
on Aging at the National Institutes of Health (grant number
lardii on fecal short-chain fatty acids and microflora in patients on long-term
5R03AG040669-02) to N. S. Pre-doctoral training support was provided
total enteral nutrition. World J Gastroenterol 2005; 11: 6165–9.
to A. K. B. by the National Institutes of Health (grant numbers
UL1TR000427 and TL1TR000429). The funders had no role in study de- 10 Ohland CL, Macnaughton WK. Probiotic bacteria and intestinal epithelial
sign, data collection and interpretation, or the decision to submit the barrier function. Am J Physiol Gastrointest Liver Physiol 2010; 298: G807–19.
work for publication. 11 Goldenberg JZ, Ma SSY, Saxton JD et al. Probiotics for the prevention of
Clostridium difficile-associated diarrhea in adults and children. Cochrane
Database Syst Rev 2013; issue 5: CD006095.
12 McFarland LV. Meta-analysis of probiotics for the prevention of antibiotic
Transparency declarations associated diarrhea and the treatment of Clostridium difficile disease. Am J
None to declare. Gastroenterol 2006; 101: 812–22.
13 Hempel S, Newberry SJ, Maher AR et al. Probiotics for the prevention and
Author contributions treatment of antibiotic-associated diarrhea: a systematic review and meta-
analysis. JAMA 2012; 307: 1959–69.
A. K. B. drafted and edited the manuscript. M. D. performed non-
microbiome laboratory analyses and drafted and edited the microbiol- 14 Ritchie ML, Romanuk TN. A meta-analysis of probiotic efficacy for gastro-
ogy methods. S. V. developed study inclusion and exclusion criteria and intestinal diseases. PLoS One 2012; 7: e34938.
edited the original study protocol. T. H. performed data analyses and 15 Barker A, Duster M, Valentine S et al. Probiotics for Clostridium difficile in-
drafted the statistical analysis methods. L. A.-P. and R. G. participated in fection in adults (PICO): study protocol for a double-blind, randomized con-
study design and edited the manuscript. N. S. participated in study de- trolled trial. Contemp Clin Trials 2015; 44: 26–32.
sign, drafted and edited the original study protocol and edited the 16 Riegler G, Esposito I. Bristol scale stool form. A still valid help in medical
manuscript. practice and clinical research. Tech Coloproctol 2001; 5: 163–4.
17 Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Md
State Med J 1965; 14: 61–5.
References 18 Dubberke ER, Carling P, Carrico R et al. Strategies to prevent Clostridium
1 Lessa FC, Mu Y, Bamberg WM et al. Burden of Clostridium difficile infection difficile infections in acute care hospitals: 2014 update. Infect Control Hosp
in the United States. N Engl J Med 2015; 372: 825–34. Epidemiol 2014; 35 Suppl 2: S48–65.
2 Musher DM, Aslam S, Logan N et al. Relatively poor outcome after 19 Palsson OS, Baggish JS, Turner MJ et al. IBS patients show frequent fluctu-
treatment of Clostridium difficile colitis with metronidazole. Clin Infect ations between loose/watery and hard/lumpy stools: implications for treat-
Dis 2005; 40: 1586–90. ment. Am J Gastroenterol 2012; 107: 286–95.

3180