American Gastroenterological Association Technical Review on the Management of Mild to Moderate Ulcerative Colitis

Key Message:

Low- (<2g/d), standard- (2–3g/d) and high-dose (>3g/d) mesalamine is more effective than placebo for induction and maintenance of remission in patients with extensive mild-moderate UC (moderate to high quality evidence). High- and standard-dose mesalamine is more effective than low-dose mesalamine for inducing and maintaining remission in patients with mild-moderate UC (moderate quality evidence). High-dose mesalamine may have a small benefit over standard-dose mesalamine for induction and maintenance of remission (moderate quality evidence), particularly in a subset of patients with moderate UC.

Effect estimate:

Potential Harms of Intervention:

All doses of mesalamine were well tolerated in RCTs for induction of remission, with lower rates of treatment discontinuation with active intervention as compared to placebo (eFigure 7). Among patients in remission, low- (RR, 0.83 [0.52–1.32]) and standard-dose mesalamine (RR, 0.62 [0.51–0.76]) were at least as well or better tolerated than placebo for maintenance of remission (Figure 8); standard-dose mesalamine was better tolerated than low-dose mesalamine, with lower rates of drug discontinuation (4 trials; RR, 0.82 [0.67–1.00]). Overall, mesalamine is a safe medication. Rare cases of interstitial nephritis have been observed, usually within 12 months of starting therapy. In an epidemiologic study of 19,000 patients from the UK, the incidence of any renal disease in patients with IBD taking 5-ASA was estimated at 0.17 per 100 patients/year, which was similar to the incidence in IBD patients not taking 5-ASAs (0.25), but higher than controls (0.08).⁶⁶ Periodic renal function testing is suggested by the FDA, though no specific studies have evaluated the frequency of testing. Rare case reports of pancreatitis, pericarditis, myocarditis, and pneumonitis have also been observed. In addition, a small subset of patients (<5% in adults, 12% in pediatric patients) may develop paradoxical worsening of colitis-type symptoms due to a hypersensitivity reaction.^{67, 68}

When treating patients with suboptimal response to standard-dose mesalamine, the potential small benefit of escalating to high-dose mesalamine should be weighed against the potential risk of delaying alternative, more effective therapy, such as corticosteroids, immunosuppressive and/or biologic therapy.

Quality of Evidence:

The overall body of evidence regarding mesalamine for mild-moderate UC was deemed to be at low risk of bias, without considerable inconsistency or indirectness; most of the studies used validated disease activity metrics for patient enrollment and outcome ascertainment. High quality evidence supports the use of standard- or high-dose of mesalamine over placebo for induction and maintenance of remission; evidence supporting low-dose mesalamine over placebo was rated down due to serious imprecision due to low event rate (<200 events), resulting in moderate confidence in estimates (Table 3). There is moderate to high confidence in estimates supporting the use of standard- or high-dose mesalamine over low-dose mesalamine for achieving remission. There was no trial comparing high-dose mesalamine to low-dose mesalamine for maintenance of remission, but indirect evidence indicates it is probably more efficacious (Table 4). There is moderate quality evidence suggesting a small benefit of high-dose mesalamine over standard-dose mesalamine, particularly in a subset of patients with moderate UC; low quality evidence suggests a small benefit for maintenance of remission, due to very imprecise estimates (Table 5).

Discussion:

While all doses of mesalamine are effective in patients with mild-moderate UC, standard- (U.S. Food and Drug Administration approved) and high-dose mesalamine are more effective than low-dose mesalamine for induction and maintenance of remission. Given absence of dose-dependent toxicity and potential risk of suboptimal disease control with low-dose mesalamine, standard-dose mesalamine may be the preferred treatment strategy when using these medications. The magnitude of benefit with high-dose mesalamine over standard-dose mesalamine was small, and it is associated with higher costs. High-dose mesalamine may be considered for induction of remission in a subset of patients at high risk of treatment failure before escalating to immunosuppressive therapy, such as treatment-naïve patients with moderate disease activity, patients with suboptimal response to standard-dose mesalamine, and patients who require corticosteroids for achieving remission. This is based on indirect evidence suggesting dose-dependent efficacy of mesalamine; no clinical trials have specifically been conducted to address the benefit of dose escalation in patients with suboptimal response to standard-dose mesalamine. A British cost-utility analysis suggested that high-dose mesalamine is more cost-effective than standard-dose mesalamine over 12-weeks in patients with moderate UC.⁶⁹ However, the small incremental benefit with high-dose mesalamine should be discussed with patients in the context of potential risks associated with delaying more effective, albeit immunosuppressive therapy. The benefit of continuing high-dose mesalamine for maintenance of remission is unclear, with moderate confidence in estimates supporting a small benefit, over standard-dose mesalamine. The two included maintenance trials did not report whether patients enrolled in this study required standard- vs. high-dose mesalamine for induction of remission. Since increasing time in remission is associated with lower subsequent risk of relapse, de-escalation from high-dose to standard-dose mesalamine (in a subset of patients who achieved remission with high-dose mesalamine) may be considered after 12 months of high-dose mesalamine therapy, through shared decision-making. A cost-effectiveness analysis comparing two maintenance strategies – no maintenance mesalamine, with mesalamine 4.8 g/day given for flares vs. maintenance mesalamine 2.4 g/day, escalated and maintained at 4.8 g/day after the first flare – failed to demonstrate cost-effectiveness of a long-term maintenance strategy.⁷⁰

Key Message:

In patients with extensive mild-moderate UC, diazo-bonded 5-aminosalicylates (balsalazide, olsalazine) are effective for induction (moderate quality evidence) and maintenance of remission (low quality evidence). Diazo-bonded 5-aminosalicylates may be slightly more effective than standard-dose mesalamine for induction and maintenance of remission, with comparable tolerability (low quality evidence). Diazo-bonded 5-aminosalicylates are probably more effective and better tolerated than sulfasalazine for induction of remission (moderate quality evidence), with comparable efficacy to sulfasalazine for maintenance of remission (low quality evidence).

Effect estimate and Quality of Evidence:

Diazo-bonded 5-ASA vs. mesalamine:

Based on 5 trials comparing diazo-bonded 5-ASA (1 trial of olsalazine 3g/day,⁷⁹ 4 trials of balsalazide 4.5–6.75 g/day^80–83) with standard-dose mesalamine (2.4–3g/d) for induction of remission, rates of failing to achieve remission was numerically lower with diazo-bonded 5-ASA as compared to standard dose mesalamine (RR, 0.81 [0.60–1.08]) (eFigure 12). When analysis was limited to only balsalazide, results were similar (RR, 0.73 [0.52–1.02]). There was no significant difference in tolerability of diazo-bonded 5-ASA and mesalamine (RR of treatment discontinuation, 0.74 [0.39–1.39]); on limiting analysis only to balsalazide, there was a trend favoring superior tolerability of balsalazide for induction of remission (RR, 0.41 [0.15–1.11]) (eFigure 13). Based on 4 trials of maintenance therapy, risk of clinical relapse was significantly lower with diazo-bonded 5-ASA as compared to mesalamine (RR, 0.69 [0.51–0.98]);^84–87 of note, all mesalamine trials used low-dose mesalamine (1–1.5g/day) for maintenance of remission (eFigure 14). There was no difference in tolerability of diazo-bonded 5-ASA and low-dose mesalamine for maintenance therapy (RR of treatment discontinuation, 0.99 [0.40–2.50]) (eFigure 15).

Diazo-bonded 5-ASA vs. sulfasalazine:

Eight trials comparing diazo-bonded 5-ASA (olsalazine 1–3g/day, balsalazide 6.75g/day) with sulfasalazine (3–6g/day) for induction of remission demonstrated a lower risk of failing to achieve clinical remission with diazo-bonded 5-ASA (RR, 0.77 [0.61–0.96]) (eFigure 16);^88–95 results were stable on sensitivity analysis after excluding studies using sub-optimal doses of diazo-bonded 5-ASA. Diazo-bonded 5-ASA was better tolerated than sulfasalazine with significantly lower rates of treatment discontinuation (RR, 0.31 [0.14–0.70]) (eFigure 17). Based on 6 trials of maintenance therapy, there was no significant difference in rates of relapse between diazo-bonded 5-ASA-treated and sulfasalazine-treated patients (RR, 1.07 [0.98–1.16]) (eFigure 18).^{94, 96–100} All trials used sulfasalazine 2g/day; 5 trials compared it with olsalazine (3 trials at dose of 1g/day and two trials used 2g/day) and one trial compared it with low-dose balsalazide (2g/day). Among patients in remission, there was no significant difference in tolerability of diazo-bonded 5-ASA and sulfasalazine (RR of treatment discontinuation, 1.02 [0.99–1.06]); as noted above, five trials used olsalazine (eFigure 19).

Potential Harms of Intervention:

Diazo-bonded 5-ASAs are safe medications, with very low rates of idiosyncratic serious or life-threatening complications. However, approximately 20% patients may develop watery diarrhea with olsalazine; this effect is not observed with balsalazide.¹⁰¹ They are better tolerated than sulfasalazine, and approximately 80% patients intolerant or allergic to sulfasalazine tolerate 5-ASA medications. Renal dysfunction is very rare with 5-ASA, but periodic renal function tests are suggested.

Discussion:

Based on the evidence presented above, diazo-bonded 5-ASA is an effective and safe alternative to mesalamine for treating the majority of patients with mild-moderate UC. Given the higher costs of mesalamine than diazo-bonded 5-ASA for some patients, diazo-bonded 5-ASA may be an alternative. The observed potential superiority of diazo-bonded 5-ASA over mesalamine should be interpreted with caution – olsalsazine, but not balsalazide, has a dose-dependent side effect of secretory diarrhea, in up to 20% patients, which may limit its tolerability especially for maintenance therapy.¹⁰¹ Due to a high pill burden associated with balsalazide, compliance and adherence may be lower. Moreover, these studies compared diazo-bonded 5-ASA with low- or standard-dose mesalamine, both of which are inferior to high-dose mesalamine. Hence, the potential benefits of diazo-bonded 5-ASA may be offset by limited ability to dose-escalate to doses equivalent to high-dose mesalamine. Moreover, maintenance studies compared diazo-bonded 5-ASA to low-dose mesalamine, which itself is inferior to standard-dose mesalamine. As noted above, diazo-bonded 5-ASA (and mesalamine) are better tolerated than sulfalsazine, particularly during induction therapy in patients with active UC-related symptoms.^{102, 103} Rates of treatment discontinuation during maintenance therapy with diazo-bonded 5-ASA were not different than for sulfasalazine which may be due to self-selection of sulfasalazine-tolerant patients during maintenance trials.

Key Message:

In patients with extensive mild-moderate ulcerative colitis, sulfasalazine is effective for induction (moderate quality evidence) and maintenance of remission (low quality evidence). Mesalamine may be slightly more effective than sulfasalazine for induction (moderate quality evidence) and maintenance of remission (low quality evidence) with lower rate of adverse events.

Effect estimate and Quality of Evidence:

Sulfasalazine vs. mesalamine:

Based on 7 trials of induction therapy, there was no significant difference in efficacy of sulfasalazine (3g/d) and mesalamine (1.5–4g/d) for induction of remission (RR, 1.07 [0.91–1.26]), without heterogeneity (I²=0%) (eFigure 24).^{26, 47, 50, 110–113} On restricting comparison to 4 trials comparing 3g/d sulfasalazine with standard-dose (2–3g/d) mesalamine, the effect estimate was more pronounced favoring superiority of mesalamine (RR, 1.27 [0.94–1.73]);^{47, 50, 111, 112} in contrast, there was no difference in efficacy of sulfasalazine and low-dose mesalamine (<2g/d) (RR, 1.00 [0.83–1.21]) for induction of remission.^{26, 110, 113} Rate of adverse events was significantly higher in sulfasalazine-treated patients as compared to mesalamine-treated patients (RR, 1.69 [1.12–2.55]), though rates of treatment discontinuation due to adverse events did not reach statistical significance (RR, 1.55 [0.64–3.77]) (eFigure 25). Based on 6 trials, sulfasalazine (1–4g/d, most common dose, 2g/d) was numerically but not statistically inferior to mesalamine for maintenance of remission (RR, 1.13 [0.91–1.40]) without heterogeneity (eFigure 26).^{110, 114–118} Of note, in five trials evaluating maintenance of remission, low-dose mesalamine was used (0.75–1.5g/d),^{110, 114, 116–118} and only one small trial compared sulfasalazine with standard-dose mesalamine.¹¹⁵ There was no difference in tolerability (RR, 1.11 [0.51–2.44]) or risk of adverse events (RR, 1.05 [0.43–2.57]) between sulfasalazine- and mesalamine-treated patients during maintenance therapy (eFigure 27).

Potential Harms of Intervention:

As noted above, sulfasalazine is not as well tolerated as mesalamine, with higher rate of treatment discontinuation due to adverse events. Between 10–45% patients may develop dose-related adverse effects, including nausea, dyspepsia, headache and fatigue with sulfasalazine.^{102, 119} Moreover, sulfasalazine has also been associated with serious cutaneous reactions such as toxic epidermal necrolysis and Stevens Johnson syndrome, pancreatitis, hepatotoxicity, drug-induced connective tissue disease, bone marrow suppression, interstitial nephritis, and hemolytic anemia or megaloblastic anemia, which are primarily attributed to the sulfapyridine moiety. Periodic monitoring of complete blood count and liver function tests is suggested in patients being treated with sulfasalazine. Sulfasalazine has also been associated with reversible quantitative and qualitative changes in spermatogenesis,¹²⁰ but has not been associated with birth defects, still births, miscarriages, low birth weights or pre-term delivery as compared to the general population or with untreated patients with IBD.¹²¹ However, patients who tolerate induction therapy with sulfasalazine are likely to tolerate maintenance therapy well. Mesalamine is better tolerated than sulfasalazine, and approximately 80% patients intolerant or allergic to sulfasalazine would tolerate 5-ASA medications.^{102, 103}. Renal dysfunction is very rare with 5-ASA.

Discussion:

Sulfasalazine is metabolized by gut flora into sulfapyridine and 5-ASA, which act as anti-inflammatory agents. It is an effective option for induction and maintenance of remission, though it may not be as effective as standard-dose mesalamine. Sulfasalazine, especially at higher doses, may not be well tolerated, which limits dose escalation in patients with a sub-optimal response, whereas the dose of mesalamine may be escalated usually without any adverse events. Hence, mesalamine may be a preferred treatment option for the majority of patients with mild-moderate UC. However, patients who have tolerated and achieved remission with sulfasalazine tolerate it well long-term, and may be maintained on the same therapy. Sulfasalazine is very inexpensive, and may be considered when alternatives are cost-prohibitive. Additionally, sulfasalazine may be beneficial in patients with colitis-associated arthralgias, given its efficacy in management of mild rheumatoid arthritis.¹²²

Key Message:

In patients with extensive mild-moderate UC, combined oral and rectal 5-ASA therapy is probably more effective than oral 5-ASA therapy alone for induction of remission (moderate quality evidence), and may be more effective for maintenance of remission (low quality evidence).

Effect estimate:

We identified 4 trials comparing oral+rectal 5-ASA (either sulfasalazine 3g/d or standard-dose oral mesalamine combined with 1–4g rectal sulfasalazine or mesalamine enemas) with oral sulfasalazine or standard-dose mesalamine for induction of remission in patients with either left-sided colitis and/or extensive colitis (~94% patients in 3 trials with left-sided colitis; 100% patients in one trial with extensive colitis).^123–126 Based on meta-analysis, risk of induction of remission was significantly higher with combination therapy compared with oral 5-ASA monotherapy (RR, 0.68 [0.49–0.94]) without significant heterogeneity (I²=31%) (eFigure 28). Two trials compared oral+rectal 5-ASA (low-dose mesalamine combined with 1–4g mesalamine enemas twice/week) with oral 5-ASA alone (low-dose mesalamine) for maintenance of remission over 52 weeks.^{127, 128} Outcomes was defined based on CAI defined remission or need for rescue medications in one trial, and maintenance of endoscopic remission with or without symptoms in the other trial. Based on these two trials, combination therapy with 5-ASA was superior to oral 5-ASA therapy for maintenance of remission (RR, 0.45 [0.20–0.97]) (eFigure 29).

Quality of Evidence:

The overall body of evidence suggesting superiority of combination oral+rectal 5-ASA therapy for induction of remission was rated as moderate quality, due to low event rate leading to imprecision (Table 8). Evidence supporting combined 5-ASA therapy over oral 5-ASA monotherapy for maintenance of remission was rated as low quality, due to indirectness (comparator group received low-dose mesalamine, whereas combined mesalamine amount in intervention group exceeded 2g) and imprecision (low event rate).

Potential Harms of Intervention:

Both oral and rectal 5-ASA were well tolerated in trials.

Discussion:

Optimization of 5-ASA therapy is a critical step before escalating to immunosuppressive therapy in patients with mild-moderate UC. With moderate quality evidence supporting a benefit of combining oral+rectal 5-ASA over standard-dose 5-ASA, this combination regimen is potentially an effective strategy for all patients with extensive mild-moderate UC, particularly those who are sub-optimally controlled with oral 5-ASA therapy alone, or patients with moderate disease activity. Two studies suggested a quicker onset of action by combining oral and rectal 5-ASA,^{124, 125} whereas Vecchi et al did not observe any significant difference in time to achieving remission.¹²⁶ In a survey of 100 patients with UC, rapidity of onset of relief was highly valued by patients over pill burden or once-daily dosing.¹²⁹ This beneficial effect of combined therapy may be related either to higher effective 5-ASA dose delivered, or to an independent topical anti-inflammatory effect of adding rectal 5-ASA therapy. While induction studies used only standard-dose oral 5-ASA for both intervention and comparator arm, it is conceivable that adding 5-ASA enemas may increase efficacy of high-dose oral 5-ASA.

Maintenance therapy with combined oral+rectal 5-ASA may also be an effective strategy, especially in patients sub-optimally controlled with oral 5-ASA alone and who value avoiding potentially avoid immunosuppressive therapy. In a focus-group when presented with a scenario of achieving steroid-induced remission after failure of 5-ASA therapy, patients valued avoiding immunosuppressive therapy, and were willing to use higher doses of 5-ASA to avoid side effects.¹³⁰ However, studies have also demonstrated low compliance with rectal therapy, and most patients wanted to reserve rectal therapy as adjunct to be used during acute flares.

Key Message:

In patients with mild-moderate UC treated with oral mesalamine, there is probably no difference between equivalent doses of mesalamine administered once daily vs. multiple-times per day for inducing and maintaining remission (moderate quality evidence).

Effect estimate:

We identified 4 trials comparing equivalent doses of mesalamine administered once daily vs. multiple times per day (2 or more divided doses) for induction of remission,^{37, 40, 131, 132} and 11 trials comparing different mesalamine dosing schema for maintenance of remission, in patients with mild-moderate UC.^{60, 62, 133–141} All studies used standard-dose mesalamine; no trials of once daily administration of diazo-bonded 5-ASA or sulfasalazine were identified. There was no significant difference in rates of induction of remission between once daily and multiple times daily mesalamine (4 trials, 944 patients; RR, 0.96 [0.85–1.08]), with minimal heterogeneity (eFigure 30). In trials of maintenance therapy, there was no significant difference in rates of maintaining remission when equivalent doses of mesalamine was administered once daily or multiple times per day (11 trials, 4465 patients; RR, 0.96 [0.85–1.07]) with minimal heterogeneity (eFigure 31). Within clinical trials, there was no difference in adherence to therapy (>80% of recommended doses taken) between once daily (pooled adherence, 92.4%; range, 41.7–98.3) and multiple times per day administration (93.6%; range, 37.5–99.6) (11 trials, 3305 patients; RR for failure to adhere to therapy, 1.09 [0.78–1.50] (eFigure 32).

Potential Harms of Intervention:

There was no difference in rates of treatment discontinuation between patients treated with once daily or multiple times daily mesalamine (10 trials, 4081 patients; RR, 0.98 [0.90–1.06]) (eFigure 33). In contrast to clinical trials, adherence to multiple times per day regimens may be lower as compared to once daily regimen in clinical practice which may lead to higher rates of relapse.

Quality of Evidence:

The overall body of evidence suggesting comparability of once daily vs. multiple times per day administration of equivalent doses of mesalamine for both induction and maintenance of remission was rated as moderate quality (Table 9). Evidence was rated down for imprecision due to wide confidence intervals on either side of unity.

Discussion:

In contrast with clinical trials which are closely regulated with high adherence rates, real-world observational studies in patients with chronic diseases have consistently demonstrated significantly lower adherence to complex dosing regimens, involving multiple divided doses, over simplified once daily dosing regimens. In a meta-analysis of 51 prospective studies utilizing an electronic monitoring device to measure adherence, compared with once daily regimen, ‘taking’ adherence was 6.7%, 13.5%, and 19.2% lower in twice-, 3-times, and 4-times (n = 57) daily regimens and ‘regimen’ adherence was 13.1%, 24.9%, and 23.1% lower in twice-, 3-times, and 4-times daily regimen, respectively.¹⁴² In clinical practice, adherence to mesalamine is low (~40–60%) particularly during maintenance phase.^{143, 144} Non-adherence is associated with significantly higher risk of disease relapse and burden and costs of hospitalization.^{144, 145} Overall, non-adherence in patients with mild-moderate UC can be attributed to patient-related factors (young age, male sex, being single, full-time employment), disease-related factors (symptomatic remission, recent diagnosis) and treatment-related factors (complexity of dosing regimen, heavy pill burden, perception of lack of benefit of medications, side effects).^{143, 144, 146} One key modifiable factor to improving compliance is simplifying dosing regimens without compromising efficacy.¹⁴⁶ Based on clinical trials, we observed moderate confidence in estimates supporting that once daily mesalamine is as effective and safe as multiple times per day mesalamine, both for induction and maintenance of remission. Studies on patient preferences have demonstrated that both patients and physicians value ease of once daily administration over multiple times per day administration, though it was not as highly valued by patients as perceived by physicians.^{129, 147} A British cost-effectiveness analysis suggested superiority of once daily mesalamine over multiple times per day administration, at equivalent doses.¹⁴⁸ None of the trials evaluated once daily administration of sulfasalazine and diazo-bonded 5-ASA, and these have conventionally been administered in multiple divided doses, in part due to high pill burden and pill size.

Key Message:

In patients with mild-moderate UC, oral budesonide MMX is probably effective in inducing remission (moderate quality evidence), and controlled ileal-release budesonide may be effective in inducing remission (low quality evidence); these medications are unsuitable for maintenance of remission. Oral budesonide MMX is not more effective than oral mesalamine or controlled ileal-release budesonide for inducing clinical remission (low quality evidence). Controlled ileal-release budesonide is probably inferior to oral mesalamine for induction of remission (moderate quality evidence). Budesonide may not be inferior to prednisone for induction of remission, and has lower risk of adverse events (low quality evidence). In mesalamine-refractory patients with mild-moderate UC, addition of oral budesonide MMX to mesalamine may be modestly effective in inducing remission (moderate quality evidence).

Effect estimate:

Key Message:

In patients with mild-moderate ulcerative proctosigmoiditis or ulcerative proctitis, rectal 5-ASA therapy may be superior to oral 5-ASA therapy for induction and maintenance of remission (very low quality evidence).

Effect estimate:

We identified 4 trials comparing 4–8 weeks of rectal vs. oral 5-ASA for induction of remission, of which 3 were conducted exclusively in patients with ulcerative proctosigmoiditis (up to 50cms from anal verge) or ulcerative proctitis.^{125, 158–160} Rectal 5-ASA intervention included mesalamine enemas 4g/d (3 trials) or mesalamine suppository 400mg three times/day (1 trial), whereas oral 5-ASA intervention included standard-dose mesalamine (2 trials), high-dose mesalamine (1 trial) or sulfasalazine 4g/d (1 trial). Outcomes were variably defined based on standardized disease activity index or physician global assessment. On meta-analysis, there was a trend favoring efficacy of rectal 5-ASA over oral 5-ASA (RR, 0.43 [0.14–1.31]), with considerable heterogeneity (I²=80%) (eFigure 38). After excluding one trial comparing high-dose MMX mesalamine with mesalamine enemas in patients with left-sided colitis,¹⁶⁰ overall effect favoring rectal 5-ASA was significant (RR, 0.28 [0.14–0.56]) without any heterogeneity (I²=0%). We identified 3 trials comparing intermittent rectal 5-ASA (mesalamine enemas 4g 2–3 times/week or 1 week/month) and oral 5-ASA (sulfasalazine 2g/day in 2 trials, and low-dose mesalamine in 1 trial) for maintaining 5-ASA induced remission in patients with mild-moderate ulcerative proctosigmoiditis.^161–163 All trials were single-blinded, with unclear randomization and allocation concealment schemes. On meta-analysis, there was a trend favoring rectal 5-ASA over oral 5-ASA (RR, 0.69 [0.41–1.17]), without heterogeneity (I²=0%) (eFigure 39).

Potential harms of intervention:

Both oral and rectal 5-ASA were well tolerated in included trials and in clinical practice.

Discussion:

For patients with distal colitis, evidence suggests that rectal 5-ASA therapy may be more effective than oral 5-ASA for induction and maintenance of remission although there is very low confidence in these estimates. This beneficial effect may be related to topical delivery of higher dose 5-ASA at site of most active disease. However, decisions on preferred route of treatment administration are patient-sensitive, with inter-individual variability. Rectal 5-ASA may be inconvenient for some patients, and in patients with active disease, retaining enemas may be difficult. In survey studies and focused group-based qualitative studies, effectiveness and route of administration were rated as most important attributes by patients with mild-moderate UC.^{129, 164} Patients generally preferred oral administration over rectal therapy, and wanted to reserve rectal therapy as adjunct to be used during acute flares.^{130, 147}

Key Message:

In adults with mild-moderate ulcerative proctosigmoiditis, mesalamine enemas are probably more effective than placebo for induction and maintenance of remission (moderate quality evidence). Rectal corticosteroid therapy (foam or enema) is more effective than placebo for induction of remission (high quality evidence). Rectal corticosteroids have not been studied for maintenance of remission. Rectal mesalamine enemas are probably more effective than rectal corticosteroids for induction of remission (moderate quality evidence).

Effect estimates:

Key Message:

In adults with mild-moderate ulcerative proctitis, mesalamine suppositories are probably more effective than placebo for induction of remission (moderate quality evidence) and maintenance of remission (low quality evidence). No randomized trials of corticosteroid suppositories were identified in patients with ulcerative proctitis; indirect evidence derived from efficacy of corticosteroid enemas in patients with ulcerative proctosigmoiditis suggests benefit of corticosteroid suppositories for induction of remission in patients with ulcerative proctitis (low quality evidence).

Effect estimates:

Key Message:

In patients with mild-moderate UC, the benefit of probiotics over placebo, or over mesalamine for induction and maintenance of remission is uncertain (low to very low quality evidence).

Effect Estimates:

Key Message:

In patients with mild-moderate UC despite 5-ASA therapy, the benefit of adding oral curcumin for induction of remission is unclear (very low quality evidence), but it may be beneficial for maintenance of remission (low quality evidence).

Effect Estimates:

Based on 3 RCTs with 169 patients with mild-moderate UC despite standard-dose mesalamine, there was a trend towards benefit with addition of oral curcumin over placebo for induction of clinical remission, though it did not reach statistical significance (RR, 0.70 [0.48–1.03]) (eFigure 48).^203–205 One trial of curcumin enema also showed comparable effect estimates.²⁰⁶ There was considerable heterogeneity in effect estimates, with one trial using curcumin 3g/d (containing 95% curcuminoid) showing markedly beneficial effect and another trial using low dose curcumin 150mg/d, showing no benefit. In a single trial of maintenance therapy in mesalamine-treated patients with quiescent UC, addition of oral curcumin was more effective than placebo in maintaining remission over 6 months (failure to maintain remission: 4/45 [8.9%] vs. 13/44 [29.5%]; RR, 0.30 [0.11–0.85]).²⁰⁷

Quality of evidence:

Overall, the body of evidence favoring oral curcumin over placebo for induction of remission with active mild-moderate UC despite 5-ASA, was rated as very low quality due to high risk of bias in included trials (inadequate blinding, allocation concealment, protocol violation, and an unexpectedly low placebo remission rate), inconsistency (statistically, in summary estimate, and clinically, with studies using widely variable curcumin dosage), and imprecision (Table 20). In contrast, evidence supporting the use curcumin 2g/d for maintenance of remission in 5-ASA treated patients with quiescent UC was rated as low quality due to imprecision and risk of bias.

Potential harms of intervention:

Curcumin is well tolerated without any significant treatment-related side effects. However, the potential benefit of using curcumin for induction of remission in patients sub-optimally controlled on 5-ASA should be weighed against risks of delaying more effective therapy. Similarly, when curcumin is added to 5-ASA for maintenance of remission, it’s potential benefit should be weighed in the context of risk of relapse – for patients at low risk of relapse maintained on 5-ASA, adding curcumin increases pill burden and may contribute to inconvenience; for patients at high risk of relapse on 5-ASA, adding curcumin may potentially delay initially of a more effective, albeit probably immunosuppressive therapy.

Discussion:

Similar to the general population, patients with UC also frequently use complimentary and alternative therapies with the hope of controlling disease using a ‘natural’, non-toxic approach.²⁰⁸ While several herbal and dietary supplements have been studied albeit in poorly designed studies, curcumin, a naturally occurring phenol, active yellow pigment of turmeric, belonging to the ginger family is one of the best studied. Biologically, curcumin has immunomodulating, proapoptotic, and antiangiogenic properties that make it potentially beneficial in patients with immune-mediated diseases.²⁰⁹ However, literature on use of curcumin in IBD has been limited by lack of dose-finding efficacy studies, inability to develop a true placebo indistinguishable from curcumin (due to taste and color of curcumin), and small study size. In the study by Lang et al, which was the only strongly positive study of curcumin for induction of remission, there was exceptionally low placebo remission (0%) and response (12.5%) rate, unlike that seen in other trials in patients with mild-moderate UC.²⁰⁵ Additionally, the majority of patients in this trial had normal serum inflammatory markers and may have had mild as opposed to moderate disease activity. Large, well-designed, phase II or III studies of curcumin are warranted in patients with mild-moderate UC, who do not adequately respond to induction therapy with optimized 5-ASA therapy, to adequately inform its role in these patients. Similarly, larger studies confirming relative and absolute efficacy of add-on curcumin for maintaining remission in 5-ASA-treated patients are warranted, to inform its role in maintenance of remission.

Key Message:

In patients with mild-moderate UC, fecal microbiota transplantation may be effective for induction of remission (low quality evidence), but there is very limited data to inform its role for maintenance of remission.

Effect Estimates:

Four RCTs in 281 patients with active UC despite usual care comparing FMT with placebo were identified; most patients had mild-moderate disease activity at time of enrollment.^210–213 The studies were heterogeneous in design with inclusion criteria, route of FMT administration and follow-up periods varying between studies. FMT was delivered via colonoscopy (x1) followed by enemas (x2 in study by Costello et al;²¹⁰ x39 administered 5 days/week for 8 weeks in study by Paramsothy et al²¹²) in the two Australian studies, via enemas (weekly, for 6 weeks)²¹¹ or nasoduodenal tube (x2, every 3 weeks)²¹³ in one each of the other studies. Patient follow-up periods varied from 7 to 12 weeks. In two studies, stool was obtained from a single donor, whereas in two Australian studies, stool was pooled from 3–7 donors, and included a mix of fresh and/or frozen stool; varying quantities of stool was instilled ranging from 8.3g to 120g. Comparator was either autologous stool in two trials and water in two trials. Based on meta-analysis of these 4 trials, FMT was more effective than placebo at inducing clinical remission over a 6–12 week period (RR, 0.80 [0.71–0.89]) (eFigure 49) and endoscopic remission (RR, 0.77 [0.63–0.93]) without significant heterogeneity. FMT was well tolerated with similar rates of adverse events in the FMT and placebo groups (RR, 1.41 [0.55–3.60]) (eFigure 50). No trials evaluating the efficacy of FMT for maintenance of remission were identified. From a recent meta-analysis of FMT in ulcerative colitis, 5 non-comparative cohort studies on long-term effectiveness of FMT were identified.²¹⁴ In these studies, of 44 patients who underwent 1–5 FMT treatments at time of active disease and were followed from 4–72 months, 22 patients had clinical response, 16 patients were unchanged and 3 patients deteriorated.

Quality of evidence:

The overall body of evidence supporting benefit of FMT over placebo for induction of remission was rated as low quality. Evidence was rated down for inconsistency (variability in intervention, with differences in route and frequency of administration, amount and source of stool) and imprecision (small sample size with low event rate) (Table 21). Though patients in these trials were along the entire spectrum of UC severity with a substantial proportion on immunomodulators and/or biologic agents, we did not rate down for indirectness since the majority of patients had mild-moderately active disease at enrollment, and the anticipated benefits are likely to be exaggerated in a population with milder disease, as compared to patients with severe disease (who are inherently more treatment resistance).

Potential harms of intervention:

In RCTs, FMT was well tolerated with very low rates of serious adverse events (worsening colitis seen in 3/140 patients receiving FMT vs. 4/137 patients receiving placebo). In a meta-analysis of 50 studies of FMT for UC or alternative reasons (mainly recurrent Clostridium difficile infection), FMT was well tolerated, with the most common FMT-attributable adverse event being abdominal discomfort.²¹⁵ The incidence of FMT-attributable adverse events was 43.6% (89/204) and 17.7% (76/430) for FMT delivered through upper and lower gastrointestinal routes, respectively. A total of 44 kinds of serious adverse events were reported in 9.2% patients, including death (3.5%, 38/1089), infection (2.5%, 27/1089), relapse of IBD (0.6%, 7/1089) and Clostridium difficile infection (0.9%, 10/1089). Besides direct FMT-attributable risks, risks of potentially ineffective therapy and delay in initiation of proven effective therapies are conceivable.

Discussion:

With increasing recognition of the role of dysbiosis in the pathogenesis of IBD, FMT is an attractive intervention. While FMT has established itself as highly effective in the management of recurrent C.difficile infection, FMT for the treatment of UC is still experimental. As noted above in 4 small RCTs, there is low quality evidence supporting a beneficial effect of FMT. Each trial used different stool processing and delivery protocols, with varying intensity and variable donor pool, which makes it difficult to recommend a particular protocol or technique over the other. The mechanism of action of FMT in UC is unclear. Besides generally increasing bacterial diversity, certain patterns of microbial changes have been associated with response to FMT in patients with UC. For example, an increase in Clostridium clusters IV and XVIII was observed in those who responded in two RCTs and Bacteroidetes including Sutterela spp and Fusobacterium spp have been associated with non-response to FMT.^{211, 212} How these changes in microbial composition result in a beneficial effect in some patients with UC is unclear, and may be related to functional alterations in luminal and epithelial metabolic and biochemical processes as well as mucosal immune responses to the microbiota. Overall, this new treatment approach warrants in-depth mechanistic as well as large-scale clinical evaluation.