Donor, Patient age and Exposure to Antibiotics Are Associated With the Outcome of Faecal Microbiota Transplantation for Recurrent Clostridioides Difficile Infection

A Prospective Cohort Study

Simon M. D. Baunwall; Mette M. Hansen; Sara E. Andreasen; Marcel K. Eriksen; Nina Rågård; Jens Kelsen; Anne K. Grosen; Susan Mikkelsen; Christian Erikstrup; Jens F. Dahlerup; Christian L. Hvas

Disclosures

Aliment Pharmacol Ther. 2023;58(5):503-515.

Abstract and Introduction

Abstract

Background: Faecal microbiota transplantation (FMT) is effective for recurrent Clostridioides difficile infection (rCDI), but its effect varies inexplicably.

Aims: To optimise the effectiveness of FMT for rCDI and validate determinants for effect

Methods: We conducted a cohort study, including all patients treated with FMT for rCDI between October 2018 and June 2020. Statistical process control was used to evaluate the impact of prospective quality improvement on the effect of single FMT treatments per 10–11 patients. Targeting an 80% effect, optimisations included changes to processing procedures, preparation and clinical application of FMT. The primary outcome was the resolution of Clostridioides difficile-associated diarrhoea at week 8. If CDI recurred, FMT was repeated. All patients were followed for 8 weeks after their latest FMT.

Results: 183 patients with rCDI received 290 FMT treatments. A single FMT achieved resolution at week 8 in 127 (69%, 95% CI: 62%–76%), while repeated FMT cumulatively achieved resolution in 167/183 (91%, 95% CI: 86%–95%). The single FMT effect varied between 36% and 100% over time. In a mixed-effect model, patient age above 65 years, non-rCDI antibiotics at week 1 post-FMT, and donor were associated with effect. Neither increasing the dosages of faecal microbes nor standardising the processing improved outcomes.

Conclusion: FMT has a high cumulative effectiveness in patients with rCDI following multiple administrations, but the single FMT effect is variable and may be optimised using statistical process control. Optimising FMT by considering patient age, post-FMT antibiotics, donor and multiple administrations may improve the treatment outcomes.

ClinicalTrials.gov: (Study identifier: NCT03712722).

Introduction

Faecal microbiota transplantation (FMT) is an effective microbiota-based therapy for managing recurrent Clostridioides difficile infection (CDI), a common and serious health threat.^[1,2] CDI often occurs after antibiotic treatment disrupts the intestinal microbiota and manifests as diarrhoea that may progress to pseudomembranous colitis, toxic megacolon or death. Comorbid patients over 65 years are particularly susceptible to CDI, which carries an excess in health care cost and mortality.^[3] Despite appropriate standard antibiotic treatment, CDI often recur, and approximately 12% of all patients with CDI develop multiple, recurrent CDI.^[4–7] For these patients, microbiota-restoration with FMT may be the only effective treatment for breaking the infectious cycle.

In Europe, fewer than 10% of the patients with an indication for FMT, three CDI episodes or more, receive FMT.^[7] Practical barriers, safety concerns and unexplainable variations in effect limit its use and hinder adoption into clinical practice.^[2] Encapsulated formulations have improved the practicality, but 12%–50% of patients treated with a single FMT still experience recurrence.^[1]

Previous studies identified several factors associated with treatment response to FMT for recurrent CDI, some of which are potentially modifiable.^[1,8–14] These factors broadly fall into three categories: (i) patient-related (such as patient age and comorbidity), (ii) FMT procedure-related (such as delivery method and dose) and (iii) CDI-related (such as disease severity).^[8] These factors need validation to be clinically applicable, and no previous prospective studies have actively intervened on the potentially modifiable factors to optimise the clinical effect of FMT.

Statistical process control (SPC) is a quality improvement technique for monitoring data and making prospective improvements.^[15,16] Unlike conventional statistics, SPC assesses data on an ongoing basis, enabling prospective data-driven decisions. SPC has successfully been used in healthcare settings to implement procedures, monitor the effect and optimise for better outcomes.^[15,17]

In the present study, we aimed to optimise and monitor the effectiveness of FMT for recurrent CDI using SPC to validate and prospectively intervene on factors previously identified as associated with lower treatment responses.

1 2 3 4

Next Section

Abstract and Introduction
Methods
Results
Discussion
References

Table 1. Baseline characteristics of the included 183 patients with recurrent *Clostridioides difficile* infection (CDI) treated with faecal microbiota transplantation (FMT).
	Overall (N = 183)
Patient characteristics
Female sex n (%)	106 (58)
Age (years), median (IQR)	69 (53–79)
BMI (kg/m²), median (IQR)	24 (20–27)^a
Charlson Comorbidity Index, median (IQR)	2 (1–4)
WHO performance status, median (IQR)	2 (1–3)^b
Immunosuppressed n (%)	25 (14)
Inflammatory bowel disease
Active disease n (%)	22 (12)
In remission n (%)	14 (8)
PEG tube n (%)	12 (7)
Proton pump inhibitor treatment n (%)	61 (33)
Patients treated at FMT centre n (%)	167 (91)
Patients treated at FMT facility n (%)	16 (9)
CDI characteristics
Number of CDI episodes, median (IQR)	3 (2–3)
CDI severity
Mild/moderate n (%)	59 (32)
Severe n (%)	59 (32)
Severe, complicated n (%)	65 (36)
Hospitalised patients during infection n (%)	58 (32)
Admitted patients to intensive during infection n (%)	7 (4)
CDI toxin/subtype
Toxin A n (%)	166 (91)
Toxin B n (%)	178 (97)
Binary toxin n (%)	34 (19)
CD027 n (%)	5 (3)
Historic treatment(s) for CDI
Vancomycin n (%)	182 (99)
Metronidazole n (%)	55 (30)
Fidaxomicin n (%)	10 (5)

Abbreviations: BMI, body mass index; CDI, Clostridioides difficile infection; IQR, interquartile range; PEG, percutaneous endoscopic gastrostomy; WHO, World Health Organisation.
^aData missing for three patients.
^bData missing for four patients.

Table 2. Bivariate analysis of all identified factors associated with treatment response of a single faecal microbiota transplantation treatment.
Factors	Unit	OR (95% CI)	p-value
FMT-related factors
Capsule protocol	Version 1	Ref
Capsule protocol	Version 2	0.97 (0.48–1.98)	0.94
Crude faeces weight	Per 1 g	0.99 (0.97–1.01)	0.23
Antibiotics during follow-up	No	Ref
Antibiotics during follow-up	Yes	0.38 (0.18–0.81)	0.012
Antibiotics week 1 after FMT	No	Ref
Antibiotics week 1 after FMT	Yes	0.26 (0.09–0.77)	0.015
Treatment site	FMT service	Ref
Treatment site	FMT centre	1.40 (0.48–4.07)	0.530
CDI characteristics
CDI severity	Mild/moderate	Ref
	Severe	1.00 (0.45–2.22)	1.000
	Severe, complicated	0.79 (0.37–1.70)	0.547
FMT episode, per 1 additional CDI		0.96 (0.75–1.22)	0.729
Hospitalisation during CDI	No	Ref
Hospitalisation during CDI	Yes	0.49 (0.25–0.94)	0.032
Intensive care during CDI	No	Ref
Intensive care during CDI	Yes	0.57 (0.12–2.66)	0.478
PPI during CDI	No	Ref
PPI during CDI	Yes	0.77 (0.40–1.48)	0.428
Patient-related factors
Sex	Male	Ref
Sex	Female	2.43 (1.28–4.62)	0.007
Age above 65	No	Ref
Age above 65	Yes	0.37 (0.18–0.76)	0.006
WHO performance status, score	0	Ref
	1	1.10 (0.30–4.05)	0.888
	2	0.29 (0.08–0.97)	0.045
	3	0.56 (0.15–2.04)	0.375
	4	0.32 (0.08–1.34)	0.119
Charlson Comorbidity Inventory	1 unit increase	0.86 (0.75–0.99)	0.039
Immunosuppression	No	Ref
Immunosuppression	Yes	0.51 (0.21–1.20)	0.122
Inflammatory bowel disease	No	Ref
	Yes, in remission	2.05 (0.66–6.40)	0.217
	Yes, active	0.61 (0.20–1.85)	0.380

Abbreviations: CDI, Clostridioides difficile infection; CI, confidence interval; FMT, faecal microbiota transplantation; OR, odds ratio; WHO, World Health Organisation.

Table 3. Multivariate analysis of factors associated with resolution of *Clostridioides difficile*-associated diarrhoea (CDAD) at week 8 after faecal microbiota transplantation (FMT) treatment.
Factors	Single FMT treatment resolution of CDAD, week 8 (logistic regression)		All FMT treatments resolution of CDAD, week 8 (mixed-effect)		Donor validation cohort^a resolution of CDAD, week 8 (mixed-effect)
Factors	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p
Intercept, baseline	6.44 (3.38–13.34)	<0.001	5.40 (2.53–11.50)	<0.001	7.38 (3.08–17.68)	<0.001
Male sex	0.45 (0.23–0.87)	0.019	0.86 (0.45–1.63)	0.642	–	–
Age above 65 years	0.44 (0.21–0.91)	0.030	0.41 (0.21–0.82)	0.012	0.38 (0.18–0.79)	0.009
Antibiotic week 1 after FMT	0.25 (0.08–0.78)	0.018	0.29 (0.09–0.98)	0.045	0.25 (0.07–0.88)	0.031
Donor factor^b	–	–	–	–	0.45 (0.23–0.88)	0.019
Random effects^c
N	–		183 patients		145 patients
Observations	183		290		211
R ² (marginal/conditional)	0.104		0.070/0.236		0.132/0.183

Abbreviations: CI, confidence interval; OR, odds ratio.
^aThe complete dataset (290 FMT procedures in 183 patients) was split into two cohorts: a prediction cohort, which included the first five treatments from each of the 20 donors to evaluate the donor factor, and the validation cohort, which consisted of the remaining 211 procedures. The analysis is based on the validation cohort holding 211 FMT treatments administered to 145 patients.
^bThe donor factor is a predictive binary measure judging whether the donor used was associated with the target of 80% or higher effect rate based on their performance in the first five FMT treatments in the prediction cohort of 79 FMT treatments. This factor evaluates whether the past performance of a donor is predictive of future outcomes.
^cIn the mixed-effect analysis, random effect parameters account for the multi-level structure of the data by considering the variability between patients (N) relative to the total FMT treatments (observations) they received. This allows for the estimation of individual factor effect while accounting for paired nature of the data.