Diarrhoea Antibiotic Management Using Over-the-Counter Nutraceuticals in Daily Practice (DIAMOND): A Feasibility RCT on Alternative Therapy to Reduce Antibiotic Use

Background: Although rarely indicated, antibiotics are commonly used for acute diarrhoea in China. We conducted a randomized, double blind exploratory clinical trial of loperamide, berberine and turmeric for treatment of acute diarrhoea. Methods: Adults with acute uncomplicated diarrhoea were randomized to 4 groups: (A) loperamide; (B) loperamide and berberine; (C) loperamide and turmeric; (D) loperamide, berberine and turmeric. All participants were given rescue ciprofloxacin for use after 48 hours if symptoms worsened or were unimproved. Primary endpoints were feasibility and ciprofloxacin use during the 2 week follow-up period. Semi-structured interviews were conducted following recruitment. Results: Only 21.5% (278/1295) of patients screened were deemed eligible, and 49% (136/278) of these consented and entered into the final analysis. Most participants had mild symptoms, because most patients with moderate or severe symptoms wanted to be given antibiotics. Follow-up was good (94% at 2 weeks). Only two participants used rescue antibiotics compared to 65% of acute diarrhoea patients in the hospital during the recruitment period. The median symptom duration was: 14 hours in group B (IQR 10-22), 16 hours in group D (IQR 10-22), 18 hours in group A (IQR 10-33), 20 hours in group C (IQR 16-54). Re-consultation rates were low. There were no serious treatment-related adverse events. Most interviewed participants said the treatment was effective. Conclusion: Although recruitment was challenging because of widespread expectations for antibiotics, patients with mild diarrhoea accepted to try an alternative. This therapy requires further evaluation in a fully powered, randomised controlled trial among a broader sample.


Introduction
Acute diarrheal disease is still a major global burden.1-3 Despite the limited benefit and potential harm from antibiotic use, acute diarrhoea is commonly treated with antibiotics in low and middle income countries (LMICs). 4 Recent audits demonstrate that antibiotics are currently the standard treatment in many places in China.5 There is a clear relationship between antibiotic prescribing and antibiotic resistance,5, 6 which is a major threat to public health as new antibiotic drug development takes a long time and very significant investment.7 China is one of the highest antibiotic consuming countries, with high rates of antibiotic resistance reported. 8 An effective non-antibiotic treatment for acute diarrhoea could relieve the pressure on Chinese doctors to prescribe antibiotics. Patients in primary care are often frustrated with the lack of effective interventions for self-limiting illnesses and a majority (72%) report that they expect to receive a prescription for something to help their symptoms,9 especially when they are poor and have spent precious resources to travel to see a doctor. Telling them that their problem will resolve spontaneously may lead to conflicts.10 There is thus a significant incentive for Chinese doctors to prescribe something to meet patients' satisfaction.
Berberine is a natural product contained in many herbal medicines which are traditionally used for the treatment of diarrhoea. 11 It is likely to be effective in acute diarrhoea,12, 13 probably working through both antibacterial and anti-inflammatory effects, protecting against lipopolysaccharide induced intestinal injury by binding to TLR4/MD-2 receptors.14, 15 Berberine use has been associated with changes in gut microbiota and an antidiarrheal effect. 16 Doses up to 400mg daily were found to be effective in two trials in acute diarrhoea. 12,13 Curcumin (from Turmeric) affects a range of cell modulating pathways and has antibacterial, antiviral, antioxidant and anti-inflammatory effects 17,18,19. It is likely to be effective in diarrhoea of either infectious or non-infectious origin.18, 20-22 One study of curcumin showed a rapid resolution of HIV-related diarrhoea 22 at a dose of 500mg three times daily and doses up to 12000 mg per day have been used safely.20 Loperamide is a widely used anti-motility agent but also acts synergistically with antibacterial agents.23 A combination of berberine, turmeric and loperamide addresses a wide range of mechanistic pathways which are potential causes of acute diarrhoea.

Preliminary data and experience
Anecdotal experience with the combination of turmeric, berberine and loperamide for traveller's diarrhoea has yielded dramatic results in 6 participants (DB, personal communication), which is faster than would be expected by loperamide alone.24 All these agents are widely available over the counter, so this combination could provide an accessible treatment for acute diarrhoea globally and reduce inappropriate antibiotic use. However, although it is plausible that all three components are effective, it has not been clearly shown whether all three components are really needed for effective treatment. Before a fully powered trial can be justified, a smaller exploratory trial to document the feasibility and acceptability of using alternatives to antibiotics is warranted.

Recruitment and feasibility outcomes
In total 1296 patients were screened, 1160 were excluded, and 136 were recruited. The majority of exclusions were either the doctor's assessment that the patient was too severely unwell (57%) or the patients were already on antibiotics (9%). (Figure 1) Figure 1 Consort report of patients' pathway and the reasons for declined patients.
(ISD= established ischaemic heart disease) Figure 2. Proportion of excluded patients by reason.

Baseline characteristics
Baseline characteristics are presented in Table 1. Overall, gender was fairly balanced across randomised groups. Most patients were young or middle-aged (mean 33 years, SD 12 years), and 39% were originally from Shanghai, while the rest had migrated there for work. The median number of loose stools in the last 24 hours before treatment was 4 (IQR 4 -5.5). Symptom severity was mild with an overall mean of 1.1 (sd 0.7) and was balanced across randomised groups. The response rate for baseline diarrhoea severity was high: among 136 recruited participants, 125 (92%) recorded diarrhoea severity. Response rates for other baseline symptom items were lower: 69 (51%) reported abdominal cramping severity, 29 (21%) reported nausea severity, and few patients (<10%) reported the remaining symptom items (vomiting, generally unwell, fever, muscle ache, headache, disturbed sleep, interference with normal activities, interference with social activities). *Symptoms (nausea, vomiting, abdominal cramping, generally unwell, fever, muscle ache, headache, disturbed sleep, interference with normal activities, interference with social activities, diarrhoea) rated on a scale from 0 to 6, where 0 indicates no problem, 1 indicates 'very little problem', 2 indicates 'slight problem', 3 indicates 'moderately bad', 4 indicates 'bad', 5 indicates 'very bad' and 6 indicates 'as bad as it could be'

Follow-up outcomes
Follow-up rates were high, with 133 people (98%) followed up at 24 hours, 131 (96%) at 48 hours, 128 (94%) at 1 week, and 128 (94%) at 2 weeks. In total there were 8 patients who withdrew or were lost to follow up. Among those withdrawing, three were concerned with side effects from the trial medicines and 5 could not be reached after 24h or 48h ( Figure 3).

Antibiotic use
Overall only 3 participants took antibiotics during the study period. No patients in groups B (berberine + loperamide), C (turmeric + loperamide) or D (berberine & turmeric + loperamide) were Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 13 October 2020 doi:10.20944/preprints202010.0280.v1 advised by a recruitment doctor to take rescue antibiotics, and 2 patients in group A (loperamide only) were advised to take rescue antibiotics. One person took rescue antibiotics in group C (loperamide & turmeric) without being advised to do so.

Effect of diarrhoea treatment on antibiotic use, duration and severity of symptoms (Table2)
Crude outcomes along with adjusted estimates for antibiotic use, duration of diarrhoea, proportion of patients with diarrhoea resolved at 24 hours and mean symptom severity during the first 48 hours are presented in Table 2. There was no statistically significant difference in duration of diarrhoea between groups, nor in the proportion of patients reporting diarrhoea resolved at 24 hours. The mean symptom severity on the 0-6 Likert scale during the first 48 hours was very mild overall and was similar across all groups.  *Analyses adjusted for number of loose stools at baseline, mean severity and prior duration. All mean differences, odds ratios (ORs), hazard ratios (HRs) reported in comparison with reference group A

Side-effects and re-consultations (Table 3)
No constipation was reported for groups B or C at 24 hours or 48 hours. Constipation affected a small number of patients in groups A and D, and the mean severity of constipation at 24 and 48 hours was mild. No rashes were reported at 24 or 48 hours. There were very small numbers of unscheduled re-consultations and re-contacts.

Usual care for acute diarrhoea
We extracted data from the hospital information system for all patients diagnosed with only acute diarrhoea without any comorbidities over the same time period. 1367 patients in the same age range were analysed. There were 757 male (55%), and 610 female (45%) patients. Mean age was 41 years old. The antibiotic use rate overall during the study period was 914 (67%) and the injectable antibiotic use rate was 64% (581/914). More than 60% were given 1 antibiotic and 38% received 2 antibiotics. 63% were given levofloxacin and 32% a 3rd generation cephalosporin.

Interview study
We attempted to contact 30 randomly selected participants for interviews via phone call. 7 participants did not answer, 6 interviews could not be completed because of connectivity problems, 3 were too busy to be interviewed. Fourteen participants were interviewed. We also interviewed six recruiting doctors (face to face) and six on-site researchers (3 by phone calls and 3 face to face, Supplemental table). All interviews lasted about 30 minutes (15-55 minutes). Thematic analysis 34 was employed for the transcripts. We identified three key themes in this study ( Figure 4). Table 4 showed the detailed quotes from interview. Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 13 October 2020 doi:10.20944/preprints202010.0280.v1 Figure 4 Three themes from interviews.
Factors facilitating use of nutraceuticals to manage acute diarrhoea and recruitment included the expectation for doctors to prescribe "something" to relieve symptoms. Both doctors and patients had very positive views on the effectiveness of the nutraceuticals. All interviewed patients and doctors expressed satisfaction with their treatment effects. Some patients even asked what those medications were and where to purchase them. Most interviewed patients felt that this trial had a flexible follow-up schedule and that the research assistants and doctors were friendly and provided very detailed explanations. Most of the doctors and RAs considered this trial was carefully designed and the concept was new.
However, there were also significant barriers to use of nutraceuticals to manage acute diarrhoea and to recruitment. Most patients believed antibiotics were "effective" for acute diarrhoea and most doctors would use antibiotics if patients "demand" them, as they worried about patients' satisfaction and subsequent confrontation. All doctors and most RAs mentioned the doctor-patient relationship was a big barrier for trust. There were two arguments during the recruitment period. Half of the doctors would consider antibiotics for acute diarrhoea even without patients' demands as they believed antibiotics are effective and shorten patients' symptom duration. Doctors reported that they were initially suspicious of the anticipated effect as berberine and loperamide were commonly used previously and they did not feel they worked, while curcumin was familiar to them but they had not used it. However, positive patient feedback increased their confidence in this trial. One doctor reported that he started to recruit more patients with higher symptom scores after he observed that patients could accept the therapy and it was seen to be effective. Only two patients complained that turmeric capsules were too big to swallow. As there had not been a randomised trial like this before, doctors agreed that it was hard to tell whether these medicines would really work until they saw the results.
There were several suggestions for improving the trial's processes for future studies. Although nine interviewed patients found that the diary was easy to use, four mentioned they could not carefully read and fully understand the diary and one suggested some wording changes. The stool samples were challenging to collect as it was difficult for patients to produce a stool on demand. Communication skills training was suggested by several doctors for on-site research assistants to allow them to communicate effectively with patients. All research assistants and doctors commented on the need for public awareness campaigns to reduce the demand for antibiotics and relieve the pressure on doctors. Five patients recommended better provision of information of the adverse effects of unnecessary antibiotic use. Ten patients reported that they wanted the public to benefit as they felt they had a very quick recovery compared with other therapies they used to use. All participants expressed their willingness to recommend this therapy to others and felt this was a meaningful trial and would support using the trial medication.

Summary of findings
Only 21% of patients screened for inclusion into this study were deemed to be eligible, and half of those consented for study enrolment. The follow up was excellent throughout, there was a very low rate of withdrawal and no safety concerns. Both doctors and patients were happy with the treatments and were willing to scale up or recommend to others. Most patients were happy with the diary and considered it easy to use, although some older patients found it more difficult. Blinding was successful.
The main barrier to recruitment is the fraught doctor patient relationship which is common in China. Doctors fear patient dissatisfaction and find it difficult to avoid the deeply entrenched routine of using injections and injectable antibiotics. During consultations many patients requested injections and antibiotics, increasing pressure on doctors to comply, and one patient complaint was only resolved by involving a local policeman.
Although this feasibility study was not powered to detect a difference between groups in the time to recovery, this study does provide support for a larger trial to document the effective use of alternatives to antibiotics for the treatment of acute diarrhoea in China. Only 2% (3/136) of participants used antibiotics, in contrast to the 67% of the patients in routine care at the same hospital during the study period and more than 60% of the prescribed antibiotics were given by injection, although the symptom severity is difficult to compare without a randomly assigned usual care control group. However, all usual care data came from patients diagnosed with acute diarrhoea without co-morbidity within the same age range during the same recruitment period. Some were eligible for inclusion but refused to participate in the trial. A proper designated usual care group will be useful to determine the effective comparison for future scale up.

Comparison with the existing literature
There have been frequent reports of violence and threats of violence against doctors.10 One study identified 459 criminal cases recorded in the Chinese Judgment online system from 2013-2016 involving crimes against healthcare workers in China, 1.6% of cases resulted in death, more than 70% involved physical abuse.35 In 2019 a new law was passed to protect healthcare workers, which is effective from June 1, 2020.36 Sadly, four days before this law a doctor was killed during an emergency consultation.36 In light of this, most doctors are afraid to take any risks, which might explain why half of the diarrhoea patients were excluded by doctors who considered the patient's condition too serious.
There have been few well-designed studies on traditional Chinese medicines (TCM) as alternatives to antibiotics for relief of symptoms for acute diarrhoea.37,38 Trial insurance requirements were only recently launched in China 39,40 as one of the doctors mentioned that it was the first time he had seen a trial covered by insurance.
This feasibility study has provided evidence that it is possible to implement a trial like this in China in spite of the recruitment challenges observed. Other studies have also shown similar antibiotic use rates in China 41 despite the expansion of antibiotic stewardship in hospitals and the National Action Plan on containing antimicrobial resistance in 2016.42 One national survey on antimicrobial stewardship program (ASP) mentioned that even more than 65% doctors were familiar with ASP only 46% of them had correct answers with ASP test.43 This indicates the urgent need for further training for doctors.

Strengths and Limitations
This is the first feasibility trial investigating the use of nutraceuticals to treat acute diarrhoea in China. A strength of this study is that we used both quantitative and qualitative methods to explore several feasibility challenges. Although the study was under-powered to compare the effectiveness of each nutraceutical, each treatment resulted in symptoms resolving rapidly, and most patients and doctors felt the treatments were effective.
This study has limitations. First, most of the recruited patients experienced very mild symptoms.26 As discussed above, recruitment may have been influenced by the doctors' complex relationship with patients in China.10 Many of the recruited patients were young with a college education level, who might be expected to have lower expectations for antibiotics. We also did not systematically collect stool samples to determine whether bacterial, parasitic or viral infections were present; this was originally planned, but proved logistically difficult in the outpatient setting as patients were mostly unable to provide a stool specimen at the time of consultation.

Implications for policy, practice and further research
Our results suggest that progression to a full randomised trial is feasible. Adequate onsite support will be needed as doctors are very busy and many lack research experience. Minor modifications will be needed to the diary to make it clearer and easier to complete. The evidence that patients recovered within 48 hours without antibiotics can be used to help doctors and patients accept using alternatives to antibiotics.
From this study we can see that antibiotic stewardship programs are urgently needed in many hospitals and clinics to address over-use of antibiotics especially for those with acute diarrhoea in China. Effective alternatives to antibiotics are likely to be an important intervention for these patients. A large fully powered trial is needed to define the effectiveness by comparing placebo, loperamide, berberine and loperamide + berberine in a wider range of patients with acute diarrhoea in China.

Setting and participants
Recruitment took place from 10th January to 30th September 2019 in a tertiary care hospital outpatient setting in China, where up to 75% of patients with acute diarrhoea are given antibiotics. 25 We recruited adults aged 18 to 70 presenting with acute diarrhoea, defined as at least 3 unformed stools in the previous 24 hours, and with a duration of less than 7 days without complications. 26 We excluded participants with vomiting as the most prominent symptom, visible blood in the stool, temperature greater than 39 degrees, suspected to have acute cholera or pseudomembranous colitis, who were immunocompromised, who had allergy to any of the proposed agents, symptom duration more than 7 days, pregnant women, known chronic bowel disease, established ischaemic heart disease or a history of cardiac arrhythmias, and prolonged QT interval.

Randomisation
The doctors provided participants with a slip with a unique computer-generated random ID number together with a coded intervention group number prepared by the study statistician. After recruitment the participant went to the pharmacist who dispensed the corresponding medication pack. Onsite research assistants (RAs) were available to further answer questions for participants. Recruiting doctors, delivery pharmacists and research assistants were blinded to treatment allocation.

Intervention
Each participant pack contained a 3 day supply of one of the following combinations, and each combination was to be used after each loose stool up to 3 times per day, to stop when the diarrhoea stops: Group A) loperamide (Xi'an Janssen Pharmaceutical Co.Ltd. 2mg tablets) 4mg initially then 2mg following each loose stool, up to three times per day.
Group B) loperamide used as in A) and berberine (ReYoung Pharmaceutical Co.Ltd.100mg tablets) 100mg up to four times daily following each loose stool.
Group C) loperamide used as in A) and turmeric (contains 6-8% curcumin, Nu U Nutrition, York, UK, 500mg capsules) 500mg up to three times daily following each loose stool.
Group D) the combination of all 3 (loperamide, berberine, and turmeric all as above).
A 'rescue' antibiotic (ciprofloxacin 750mg stat) was given to be used only if symptoms were not starting to settle within 24 hours, after telephone or face-to-face assessment by the designated doctor.
We included non-active nutraceuticals in the medication packs (amino acids) 27 (Zhejiang CONBA Pharmaceutical Co.Ltd.CONBA G20120506 0.8g/tablet) for those individuals not receiving triple therapy so that every individual had three medication containers to use.
All participants were instructed to seek medical assistance again in the event that symptoms progressed. Follow-up was scheduled at 24 hours (telephone, clinic visit if needed), 48 hours (clinic visit), day 7 (clinic visit) and day 14 (telephone).

Outcome measures
Feasibility outcomes included recruitment rate, exclusion rate and reasons, rejection rate and reasons, withdrawal rate and follow-up rate. Each participant was asked to complete a daily symptom diary which has been used successfully in trials of medicines for diarrhoea.24 Exploratory outcomes included use of rescue antibiotics in 24 hours, the duration of symptoms, the proportion with diarrhoea resolved at 24 hours, and the severity of symptoms.

Severity of symptoms
The severity score was documented in the first 48 hours because this is the time when symptoms are the most severe and when nutraceuticals might make an important difference. The diary recorded the number of stools during the previous 24 hours, the consistency of the last stool and the time since the last loose stool. The diary also recorded the severity of symptoms: diarrhoea, vomiting, nausea, abdominal pain, anal burning, fever, disturbed sleep, feeling generally unwell, and interference with normal activities. Each symptom was scored on Likert scales which have been shown to be valid and sensitive to change for a variety of infections. 28-31 Scores ranged from 0 to 6: 0=no problem, to 6=as bad as it could be. Also documented was the duration of symptoms rated at least moderately bad (3) and the time taken for all symptoms to be rated as very little or no problem (1 or 0).

Statistical analysis
In addition to feasibility outcomes, we aimed to recruit a sufficient sample to detect a difference between 50% using rescue antibiotics in the loperamide group and 15% in any other intervention group (for alpha 0.05 and 80% power), since 50% are likely to have resolved in 24 hours.32,33 We therefore aimed to enrol 30 participants per group with a slight over-enrolment to account for loss to follow up. All analyses were conducted following the intention-to-treat principle. Only complete cases were included, there was no imputation of missing data. Descriptive statistics were used to report feasibility and clinical outcomes, and regression analyses adjusting for number of loose stools at baseline, mean severity and prior duration were conducted for exploratory outcomes where possible. All analyses were performed using STATA 16 (Stata, College Station, TX, USA).

In-depth Interviews
We conducted semi-structured qualitative interviews of patients who participated in this study, and interviewed recruiting doctors and on-site research assistants after trial recruitment was completed. The interviews were conducted by telephone or face to face. Participants were invited to provide verbal consent. We aimed to understand what did and did not work well, whether they were willing to recommend this therapy and suggestions for a future scale up study. Patients and doctors were also asked about what treatment they normally use for acute diarrhoea and whether they would be willing to recommend the trial therapy to others.

Conclusions
Recruitment of patients with anything other than mild diarrhoea was very challenging in the current clinical environment in China. Use of loperamide and berberine may relieve symptoms as a viable alternative to antibiotics in China. This approach should be scaled up for further evaluation in a randomised controlled trial to investigate its effectiveness.

Conflicts of Interest:
We declare there are no competing interests.

Availability of data and materials:
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Ethics approval and consent to participate: Ethical approval was first obtained by the WHO China Clinical Trials Central Medical Ethics Committee. Reference No: ChiECRCT-2017098. We also obtained the ethical approval from the medical ethics committee of the implementing hospital Shanghai Jiading district centre hospital. Reference No: 201802.

Supplemental table
Supplement