Key Points

• Low-strength evidence suggests that dietary fiber supplementation with psyllium decreases FI frequency by 2.5 occurrences per week after 1 month of use; clonidine has no effect; and PFMT-BF with electrostimulation is no more effective than PFMT-BF on FI severity and the FI Quality of Life Scale (FIQL) scores over 2 to 3 months.
• Low-strength evidence at 6 months suggests that dextranomer tissue-bulking injections are more effective than sham injections on the FIQL, number of FI-free days, and percent of adults with at least 50 percent reduction from baseline in FI episodes; no more effective than PFMT-BF with or without electrostimulation on FI severity (PFMT-BF -5.4 versus dextranomer -4.6 point Vaizey improvements) and FIQL; and no more effective than sham injection on FI severity (-2.5 versus -1.7 point sham CCFIS improvement) or FI episode frequency. (See Appendix E for FI outcome measures)
• Moderate-strength evidence suggests that Durasphere® (off-label) bulking injections reduce FI severity (-4 to -5 points in CCFIS) up to 6 months, but gains diminish thereafter.
• Evidence is insufficient for PFMT-BF versus standard care (such as dietary fiber and stool-modifying drugs); all other PFMT studies assessed refinements in treatment delivery by comparing PFMT to another variation of PFMT.
• Evidence was insufficient for all other nonsurgical interventions.
• In most cases, short-term outcomes improvements in both treated and active controls met minimum clinically important differences (MID) when those values were known (usually 2 to 6 points, various scales, Appendix E); studies that claimed greater improvements typically excluded nonresponders, noncompleters, or those not fully compliant with study protocols.
• The wide range of outcome measures limited comparability across studies.
• Most nonsurgical RCTs (84%) had moderate or high risk of bias.
• Incomplete reporting of baseline patient information and FI etiologies was common.
• Most evidence was short term (Appendix F, Table F2).

Dietary Fiber

The evidence for dietary fiber and fiber supplementation in FI is exclusively short term (up to 3 months (Table 2). Two RCTs^36,37 assessed the 31 day and 38 day effects of various dietary fiber supplements on FI frequency, and stool frequency and consistency. Low-strength evidence³⁶ suggests that dietary fiber supplementation with psyllium reduces FI frequency by 2.5 occurrences per week and has no effect on FI quality of life as measured with the FIQL.³⁶

Table 2

KQ 1. Randomized controlled trial evidence for dietary fiber and dietary fiber supplementation for fecal incontinence.

Evidence was insufficient for other outcomes, including one moderate risk of bias RCT that found no added benefit of dietary fiber in addition to loperamide on FI severity and the FIQL over 3 months³⁸ (Table 2). Evidence was insufficient for methylcellulose plus loperamide versus no treatment³⁹ (Appendix F, Table F5).

Pharmacological Treatments

Drug studies were exclusively short term (1 to 6 weeks) and most were 1 month in duration. The effectiveness of oral and topical medications for FI was examined in 11 RCTs: three of topical phenylephrine versus placebo^40-42 (Table 3), four of antidiarrheal medications^43-46 (three versus placebo, one with active comparators, Table 4), and four studies of other medications^47-50 (all versus placebo, Table 5). Low-strength evidence suggests that oral clonidine has no effect on FI severity as measured with the FI and Constipation Assessment (FICA).⁴⁷ Evidence was insufficient for loperamide,^43-46 topical phenylephrine (10%^41,42 and 30%⁴⁰), zinc-aluminum ointment,⁴⁸ estrogen cream,⁴⁹ and valproate sodium.⁵⁰

Table 3

KQ 1. Randomized controlled trial evidence for topical phenylephrine (sphincter function enhancement drug) for fecal incontinence.

Table 4

KQ 1. Randomized controlled trial evidence for antidiarrheal drugs for fecal incontinence.

Table 5

KQ 1. Randomized controlled trial evidence for other drugs for fecal incontinence.

Pelvic Floor Muscle Training and Adjunctive Modalities

Pelvic floor muscle training using biofeedback (PFMT-BF) was the most frequently studied intervention in the literature we reviewed; 16 studies (13 RCTs and 3 OBS^51-53) assessed the effects of PFMT-BF with or without electrostimulation on the outcomes of FI frequency and severity, quality of life (general and FI-specific, the FIQL), and perceived improvement (Tables 6-12 and Appendix F, Table F5).

We found insufficient evidence for PFMT-BF versus standard care (such as dietary fiber, stool-modifying drugs, and/or advice, Table 6 and Appendix F, Table F5). The definition of standard care varied across studies. Only two RCTs^54,55 (Table 6) with moderate⁵⁵ and high⁵⁴ risk of bias assessed the benefit of PFMT-BF versus standard care, and one high risk of bias observational study⁵³ (Appendix F, Table F5) examined PFMT-BF plus standard care versus standard care alone. Most of the literature focused on ways to improve or prolong the purported benefits of PFMT for FI by comparing PFMT to another variation of PFMT,^56-66 rather than to establish the benefits of it. Only two RCTs used PFMT alone as a control^56,57 (Table 7); all other studies (Tables 8-12) assessed refinements in PFMT delivery by testing one form of PFMT against another, including PFMT plus FI education⁵⁸ (Table 8), biofeedback sensor comparisons^59,60 (Table 9), exercise comparisons⁶¹ (Table 10), electrostimulation frequency comparisons^62,63 (Table 11), electrostimulation to augment PFMT-BF^64-66 (Table 12), or examined the mode of training delivery (by phone or in-person) on outcomes⁵¹ (Appendix F, Table F5). Risk of bias was moderate to high in all PFMT studies.

Table 6

KQ 1. Randomized controlled trial evidence for pelvic floor muscle training with biofeedback (PFMT-BF) versus standard care.

Table 7

KQ 1. Randomized controlled trial evidence for pelvic floor muscle training with biofeedback (PFMT-BF) versus PFMT alone.

Table 8

KQ 1. Randomized controlled trial evidence for PFMT-BF versus PFMT plus education.

Table 9

KQ 1. Randomized controlled trial evidence for PFMT-BF versus PFMT with digital rectal feedback.

Table 10

KQ 1. Randomized controlled trial evidence for types of exercise used for PFMT-BF for fecal incontinence.

Table 11

KQ 1. Randomized controlled trial evidence for PFMT-BF with electrostimulation for fecal incontinence: comparison of frequencies.

Table 12

KQ 1. Randomized controlled trial evidence for PFMT-BF with electrostimulation versus PFMT-BF for fecal incontinence.

We found low-strength evidence that PFMT-BF with electrostimulation is no more effective than PFMT-BF on FI severity and FI quality of life (FIQL).^65,66 Evidence was insufficient for all other PFMT comparisons.^51,52,56-64

PFMT-BF was associated with improvements in FI outcomes (usually 2 to 6 points, various scales) in most studies, but improvements did not differ significantly from those of the comparison group. Most PFMT RCTs reported 3 to 6 month outcomes (Appendix F, Table F2); only four studies reported outcomes for randomized patients beyond 6 months.^55,57,61,63

Anal Electrostimulation

Evidence was insufficient for home-based anal electrostimulation without PFMT versus home-based sham stimulation on symptoms and FI severity,⁶⁷ and for home-based electrostimulation versus hospital-based therapy⁶⁸ in the short term (Table 13). The extremely low compliance with home-based electrostimulation in one RCT⁶⁷ (only 25 percent of the treatment group used the stimulator at least 20 of the 34 protocol-recommended hours) suggests that home-based stimulator use for FI may not be an acceptable option to patients, even if it worked.

Table 13

KQ 1. Randomized controlled trial evidence for electrostimulation (without PFMT) for fecal incontinence.

Mixed Nonsurgical Interventions

Mixed interventions were primarily assessed for two groups of adults: older adults residing in nursing homes and adults with spinal cord injuries (SCI). Both groups may deal with FI, constipation, or both. The goal of bowel management is to minimize extremes and maintain bowel regularity.

Two RCTs focused on bowel management interventions for adults with SCI^70,71(Table 14). Females comprised 31 percent of enrolled adults; the overall median age was 48 years. One moderate risk of bias study found that transanal irrigation improved bowel and FI outcomes more than supportive, guidelines-based care over 10 weeks.⁷⁰ One high risk of bias study reported that a 6-week step-wise, increasing intensity bowel management program worsened FI outcome.⁷¹

Table 14

KQ 1. Randomized controlled trial evidence for interventions to manage fecal incontinence in adults with spinal cord injury.

In contrast, two high risk of bias RCTs assessed staff-directed interventions for FI and bowel issues in nursing home residents with mixed results (Table 15). Females comprised 83 percent of enrolled residents; the overall mean age was 87 years. Both interventions focused on multiple factors affecting bowel regularity, including aspects of diet, fluids, activity, and care. One RCT found significant reductions in FI frequency with prompted toileting four times per day, exercise and increased fluid offering 5 days per week.⁷² The other RCT was a multicomponent intervention for UI and FI, which did not affect FI frequency.⁷³

Table 15

KQ 1. Randomized controlled trial evidence for interventions to manage fecal incontinence in older adults in nursing homes.

Posterior Tibial Nerve Stimulation

Percutaneous posterior tibial nerve stimulation (PTNS) is not FDA approved for FI but is currently being studied as a nonoperative off-label treatment option, especially prior to considering permanent SNS (www.clinicaltrials.gov). One small, moderate risk of bias RCT (Table 16) examined the effects of PTNS versus SNS on FI episodes and the CCFIS.⁷⁴ The evidence for PTNS is insufficient.

Table 16

KQ 1. Randomized controlled trial evidence for percutaneous tibial nerve versus sacral nerve stimulation.

Anal Sphincter Tissue-Bulking Injections

Four low risk of bias RCTs (Table 17) examined anal sphincter tissue-bulking injections: two RCTs of dextranomer,^75,76 which is FDA-approved for FI, and two of an off-label injectable, Durasphere®^77,78 (FDA-approved for urethral bulking for urinary incontinence).

Table 17

KQ 1. Randomized controlled trial evidence for injectable tissue bulking agents for fecal incontinence.

Low-strength evidence at 6 months post-treatment suggests that dextranomer tissue-bulking injections are no more effective than PFMT-BF with or without electrostimulation on FI severity and FI-related quality of life as measured by the FIQL.⁷⁵

Low-strength evidence at 6 months post-treatment suggests that dextranomer tissue-bulking injections are more effective than sham injections on FI quality of life (FIQL scale), the number of FI-free days, and in reducing FI episodes 50 percent or more from baseline over 6 months, but no more effective than sham injection on FI severity (CCFIS) and FI episode frequency.⁷⁶

Durasphere® (off-label) anal sphincter injections improved FI severity (CCFIS) by several points shortly after injections, but gains diminished slightly between 6 months and 1 year.^77,78 Both studies used a non-FDA approved comparator (PTQ™).

Key Points

• Evidence is insufficient for all surgical interventions for FI. Few surgical treatments were examined in RCTs or OBS studies; aims and treatments were highly varied and all surgical studies had moderate to high risk of bias.
• The overwhelming majority of surgical studies are case series (low quality evidence).
• Evidence is insufficient for SNS versus supportive care for FI up to 1 year; for turning the stimulator on versus having it off on FI severity and frequency in newly-implanted patients; for stimulation at 75 percent or 50 percent of sensory threshold versus stimulation at sensory threshold; for high-frequency, low prolonged pulse width stimulation to regain efficacy in persons with sustained loss of efficacy after chronic stimulation; and for turning stimulation off versus leaving it on during the night on FI outcomes.
• Surgical outcomes, in general, were reported for longer term followup than for nonsurgical interventions.
• In half of the RCTs where MID was known, outcomes improvements with treatment and sometimes control interventions met or exceeded MID in intermediate-term outcomes (CCFIS, Appendix E), although adults with complications or those lost to follow-up were omitted from half of those analyses.
• Articles commonly lacked important baseline information (such as patient characteristics, FI etiologies and outcomes at enrollment). In particular, SNS studies included minimal nonphysiologic patient information.

Sacral Nerve Stimulation (SNS)

Surgically placed SNS is used when conservative measures have failed to afford the desired level of fecal continence. There are two main limitations of SNS: (1) the stimulator battery has a limited lifetime and needs to be surgically replaced within the stimulator approximately every 5 years and (2) the nervous system adapts to stimulation over time which may result in the loss of efficacy on FI in some adults. Only one RCT⁸⁰ assessed the effectiveness of SNS with the stimulator on versus off in newly implanted patients; more recent studies focused on the maintenance of SNS battery life while maintaining continence effects,^81,82 measures to regain SNS efficacy that was dwindling,⁸³ and the comparison of SNS to best supportive care.⁸⁴ All SNS RCTs were crossover studies (Appendix F, Table F4) that almost exclusively enrolled adult females approximately 60 years old with mixed FI etiologies.

The evidence for SNS is insufficient because all five studies had moderate or high risk of bias, and none assessed the same treatment-outcome combination. Evidence is insufficient to compare the effectiveness of SNS versus supportive care on FI outcomes up to 1 year;⁸⁴ the effects of turning the stimulator on versus off on FI severity and frequency in newly-implanted patients;⁸⁰ stimulation at 75 percent or 50 percent of sensory threshold versus stimulation at sensory threshold;⁸¹ high-frequency, low prolonged pulse width stimulation to regain efficacy in persons with sustained loss of efficacy after chronic stimulation;⁸³ and turning stimulation off versus on at night on FI outcomes.⁸²

Observational studies provided insufficient evidence for SNS versus sphincteroplasty⁸⁵ (Appendix F, Table F5) and open versus percutaneous lead placement⁸⁶ (Appendix F, Table F5).

Anal Sphincter Repair (Sphincteroplasty)

Surgical repair of the anal sphincter is performed for adults with FI resulting from anal sphincter tears that have accompanying moderate to severe FI and have suboptimal resolution with conservative treatment. Only two RCTs^87,88 (Appendix F, Table F4) and five observational studies^85,89-92 (Appendix F, Table F5) examined sphincteroplasty.

Observational studies (Appendix F, Table F5) provided insufficient evidence to compare the effectiveness of sphincter repair with stoma (fecal diversion) versus sphincter repair alone;⁸⁷ adjuvant biofeedback following anal sphincter repair versus sphincter repair alone;⁸⁸ a perineal versus a posterior forchette incision in overlapping anal sphincter repair;⁸⁹ sphincteroplasty with pelvic floor repair versus sphincteroplasty;⁹⁰ anal sphincter repair versus SNS;⁸⁵ sphincteroplasty versus anterior levatorplasty;⁹¹ and direct versus anterior sphincter repair.⁹²

Anal Sphincter Replacement

Evidence was insufficient (Appendix F, Table F5) to compare the effectiveness of the artificial bowel sphincter (ABS) versus conservative medical management from one RCT of 14 patients with severe FI,⁹³ and for the ABS versus a non-FDA approved magnetic anal sphincter⁹⁴).

Other Surgeries and Mixed Treatment Comparisons

Appendix F, Tables F4 and F5 include other surgical studies and mixed treatment comparisons. Evidence was insufficient to compare the effectiveness of total pelvic floor repair versus gluteus maximus transposition without electrical stimulation for postobstetric neuropathic FI;⁹⁵ postanal repair versus total pelvic floor repair for neurogenic FI;⁹⁶ and total pelvic floor repair versus anterior levatorplasty versus postanal repair for neurogenic FI.⁹⁷

Evidence was insufficient for levatorplasty surgery versus nonsurgical anal plug electrostimulation⁹⁸ and for SNS versus a non-FDA approved surgery (magnetic sphincter).⁹⁹

Evidence was insufficient for recommendations after failed sphincteroplasty.¹⁰⁰ Only one high risk of bias observational study compared the outcomes of three surgical treatments used in adults who had at least one prior sphincteroplasty with unsatisfactory outcomes.¹⁰⁰ (Appendix F, Table F5).