Systematic Review

Efficacy of Escalated Maintenance Anti-tumour Necrosis Factor Therapy in Crohn's Disease

Vandita Y. Mattoo; Chamara Basnayake; William R. Connell; Nik Ding; Michael A. Kamm; Mark Lust; Ola Niewiadomski; Alexander Thompson; Emily K. Wright


Aliment Pharmacol Ther. 2021;54(3):249-266. 

In This Article


Search Results and Study Characteristics

The search results are summarised in Figure 1. All eligibility criteria were met in the 68 articles included. Of the included studies, nine were RCTs, nine were post hoc analyses of the RCTs, and five were single-arm trials. We included two reports of the Steenholdt RCT, with a 12-week follow-up[20] and a 20-week follow-up.[21] In addition, 44 observational studies were reviewed, of which only 10 were prospective studies. Of the retrospective studies, one was a follow-up study of the Trough Concentration Adapted Infliximab Treatment trial (TAXIT).[22]

Figure 1.

PRISMA flow diagram of study selection

Adalimumab was dose escalated to 40 mg weekly in most studies (n = 27), although a small number of studies escalated patients to 80 mg every other week (n = 4), 80 mg weekly (n = 1) or 40 mg every 10 days (n = 2). For infliximab, dose escalation was defined as a shortening of treatment interval to 4–7 weeks (n = 7), an increase in dose up to or exceeding 10 mg/kg (n = 9), or a combination of both (n = 21). Across included studies, secondary loss of response to standard dosing was defined by global physician assessment and/or objective evidence of increased disease activity in patients with previous response to therapy. Clinical response and remission were generally reported using standardised clinical indices (Crohn's Disease Activity Index [CDAI] or Harvey Bradshaw Index [HBI]). Across the included studies, clinical response was defined as a decrease in CDAI score of 50–70 points or more, or a decrease in HBI score of 3 or more points. A CDAI score less than 150 or a HBI score below five indicated clinical remission.

Risk of Bias Assessment

The risk of bias assessment for RCTs is presented in Figure S6. All but one study[23] described the randomisation process and allocation concealment sufficiently. None of the studies were assessed to be at risk of attrition bias due to incomplete outcome data. The CALM trial[3] and the PRECISION trial[12] were at a high risk of bias due to lack of participant or investigator blinding. The outcome assessors were not blinded to TDM data in the Steenholdt RCT[20] leading to a high risk of bias. Overall, five of the RCTs were found to be at low risk of bias,[9–11] three studies were at a high risk of bias,[3,12,20,21] and one study[23] was rated as having some concerns around risk of bias.

The majority of observational studies (n = 34) received 4–6 stars out of 9 on the NOS scale indicating a moderate risk of bias, the remainder of the studies (n = 23) were at a low risk of bias (Table S7). As data from observational studies were largely retrospective, outcomes of interest were often present at the outset increasing the potential for bias. Important confounders were uncontrolled in some cases with differences in care delivery and nonstandardised protocols for anti-TNF optimisation.

Standard Maintenance Dosing of Infliximab and Adalimumab

Our literature search identified three RCTs that validated anti-TNF dosing schedules in large prospective cohorts.[9–11] The Crohn's Trial of the Fully Human Antibody Adalimumab for Remission Maintenance (CHARM)[10] was a double-blind maintenance study in patients with moderately to severely active CD (n = 854). All patients received open-label induction with adalimumab 80 mg at week 0 and 40 mg at week 2. Following this, patients were randomised to one of three maintenance arms: adalimumab 40 mg every other week, 40 mg weekly or placebo. The coprimary endpoints were clinical remission (CDAI < 150) at weeks 26 and 56. Both adalimumab dosing regimens were found to be superior to placebo and equally effective in maintaining remission at both week 26 (39.5% 40 mg every other week vs 46.5% 40 mg weekly, P = 0.22) and week 56 (36% vs. 41.4%, P = 0.34).[10]

In ACCENT I,[11] all patients received a 5 mg/kg infliximab infusion at baseline (n = 580). Patients were then randomised to receive placebo (group I), infliximab 5 mg/kg (group II) or infliximab 10 mg/kg (group III) with infusions at week 2, week 6 and every 8 weeks thereafter. Infliximab maintenance treatment was superior to placebo, with infliximab-treated patients more likely to be in clinical remission (CDAI < 150) at week 30 compared to patients receiving placebo (OR 2.7, 95% CI 1.6–4.6). However, there was no clinically significant difference in remission rates between the two infliximab dosing strategies at week 30 (OR 1.3, 95% CI 0.74–2.20) or week 54 (OR 1.58, 95% CI 0.90–2.80).[11]

Efficacy of Anti-TNF Dose Escalation

Clinical Response and Remission. Clinical outcomes following anti-TNF dose escalation are shown in Table 1 and Table 2. The majority of clinical efficacy data were from retrospective studies with subcohorts that underwent empiric dose escalation (without the use of TDM) for loss of response to standard treatment. Twenty studies reported clinical outcomes following empiric escalation of maintenance adalimumab.[10,23–41] Response rates ranged from 33% to 100% (Figure 2), and remission was achieved in 15% to 83% of patients at any point after treatment escalation.[10,23–41] In 15 of these studies, more than half of the patients regained clinical response after dose escalation.[23–25,27–32,35–39,42] In the study by Narula et al,[31] only one (7.1%) of the 14 anti-TNF naïve patients escalated to 40 mg weekly achieved clinical remission at the end of follow-up. For infliximab, rates of clinical response and remission following empiric escalation for treatment failure ranged from 34% to 90% and 26% to 81%, respectively, across 12 studies (Figure 3).[28,31–33,43–50] In eight of these studies, intensified therapy recaptured response in over 50% of the patients.[28,36,45–50] In one retrospective study (n = 265), a switch to adalimumab achieved numerically higher rates of clinical response (83.3%) than infliximab dose escalation (42.3%).[32]

Figure 2.

The rate of clinical response (%) following empiric dose escalation for patients with secondary loss of response to standard maintenance adalimumab in the included studies. Bubble size is proportional to the number of dose-escalated patients in each study. The number within each bubble corresponds to the study reference

Figure 3.

The rate of clinical response (%) following empiric dose escalation for patients with secondary loss of response to standard maintenance infliximab in the included studies. Bubble size is proportional to the number of dose-escalated patients in each study. The number within each bubble corresponds to the study reference

Sustained clinical response and remission beyond 12 months from the time of dose escalation were reported in 14 studies and ranged from 33% to 90% for response and 28%-81% for remission.[27,31,33–36,41,42,45–48,50,51] Similar rates of loss of response to dose-escalated treatment (tertiary loss of response) were reported across five studies.[27,30,44,45,52] In the study by Ma et al,[30] over half (56.8%) of those that responded clinically to escalated adalimumab subsequently lost response (median time to tertiary loss of response 47.9 weeks [IQR 24.7–80.3]). In another study[45] of 82 patients receiving high dose infliximab therapy, 33% discontinued treatment after a median of 31.3 weeks.

Biochemical Response and Remission. Twelve studies analysed the biomarker response to anti-TNF dose escalation.[3,25,27,34,39,45,49,51,53–56] Of these, six showed a significant reduction in CRP with anti-TNF dose escalation.[45,49,51,54–56] A post hoc analysis[53] of the TAILORIX trial showed a significant decrease in faecal calprotectin (P = 0.049) but not in CRP (P = 0.193) following escalation of infliximab dosing.

Optimisation Strategies for Anti-TNF Therapy

Interval Shortening Versus Dose Doubling. In five retrospective studies comparing dose escalation strategies, doubling the infliximab dose versus shortening the treatment interval led to similar clinical and biochemical outcomes.[47,48,51,53,54] In the study by Dreesen et al,[54] dose escalation failed to achieve therapeutic infliximab levels (≥3 μg/mL) when high titre antidrug antibodies (>282 ng/L) were present, regardless of dosing strategy. However, when antibodies were quantifiable but low (<282 ng/mL), a combination of increased dose and reduced interval was most effective at restoring drug levels, followed by dose doubling alone, which was superior to interval shortening alone (P = 0.0005). In the absence of anti-infliximab antibodies, all three strategies were equally effective at restoring serum drug levels in patients with loss of response.[54] For adalimumab, Duveau et al[27] showed that 40 mg weekly dosing was associated with higher rates of clinical response at 12 months (P = 0.03) and lower rates of tertiary loss of response (P = 0.01) compared to 80 mg every other week.

Dose Escalation Versus Anti-TNF Reinduction. One retrospective study[57] (n = 80) compared anti-TNF reinduction to shortening the dosing interval in a cohort with complex CD (91% stricturing or penetrating) and secondary loss of response. There was no significant difference in the rates of treatment failure within 12 months of reinduction compared to dose escalation (24% vs 15%, P = 0.27).[57] However, beyond 24 months, rates of treatment failure were higher in the reinduction group compared to the dose escalation group, leading to a higher overall rate of treatment failure after reinduction over the course of the study (48% vs 24%, P = 0.02).[57]

Therapeutic Drug Monitoring. Three RCTs compared TDM-guided dosing to empiric dose escalation.[20,21,58,59] In a single-blind noninferiority trial, 69 CD patients with loss of response to standard infliximab therapy were randomised to empiric escalation (5 mg/kg every 4 weeks) or optimised using a TDM algorithm.[20,21] The primary objective of this study was to demonstrate that the use of reactive TDM had comparable clinical efficacy to empiric dose escalation, but was significantly more cost effective. Clinical response rates were similar between the two groups at 12 weeks (57.6% TDM group vs 52.8% empiric group, P = 0.810)[20] and 20 weeks of follow-up (75.8% vs 55.6%, P = 0.128).[21] This was also observed for perianal disease, with a similar decrease in Perianal Disease Activity Index (PDAI) scores for patients in the reactive TDM group versus the empiric treatment group at 12 weeks (P = 0.421)[20] and 20 weeks (P = 0.495).[21]

Restellini et al.[60] compared reactive TDM to empiric dosing in a retrospective cohort of CD patients with loss of response to adalimumab (n = 104). The primary outcome was the rate of composite remission, defined as HBI<5 with biomarker normalisation and Simple Endoscopic Activity Score (SES-CD) <3. Similar proportions of patients achieved the primary endpoint in the TDM group versus the empiric escalation group at 3 months (42.3% vs 16%, P = 0.619), 6 months (30.8% vs 45.5%, P = 0.328) and 12 months (39.4% vs 27.3%, P = 0.978).[60] Conversely, Kelly et al[61] compared endoscopic outcomes after TDM based dose adjustment versus standard clinical management in CD and UC patients with documented endoscopic disease at baseline. While the baseline SES-CD was comparable between the two groups, the TDM group had lower SES-CD scores compared to the clinically dosed group at median 6 months after dose optimisation (P = 0.054).[61]

In the TAXIT trial,[58] all patients were initially dose adjusted to achieve therapeutic infliximab trough levels between 3 and 7 μg/mL. Following this proactive optimisation phase, the proportion of CD patients in clinical remission increased significantly compared to baseline (88.4% vs 65.1%, P = 0.020). However, at 12 months after randomisation, there was no significant difference in clinical remission rates between the proactive TDM group versus patients dosed on symptoms alone (62.6% vs 54.9%, P = 0.353). In terms of secondary endpoints, the proactive TDM group had fewer relapses (7% vs 17%, P = 0.018) and longer times to relapse (P = 0.017) compared to those who received standard care based on clinical symptoms. Pouillon et al.[22] conducted a retrospective follow-up study of the TAXIT cohort examining mucosal healing and the durability of infliximab treatment. The rates of mucosal healing were high at the conclusion of TAXIT, and similar between the proactive TDM group and the standard care group (91% vs 90%, P = 1). Overall, the rate of infliximab discontinuation was similar, however, within 1 year, more patients in the clinically based dosing group discontinued treatment compared to the proactive TDM group (37% vs 9.5%, P = 0.04). Similarly, a multicenter retrospective study of 264 IBD patients demonstrated a significantly lower cumulative probability of treatment failure in infliximab-treated CD patients where dosing was optimised by proactive versus reactive TDM over a median follow-up of 2.4 years (log-rank P < 0.001).[62]

More recently, the TAILORIX trial[59] evaluated proactive TDM in a cohort of biologic naïve patients with active luminal CD (n = 122). The primary endpoint was corticosteroid-free remission (CDAI < 150) at all visits between week 22 and week 52 of the study, with no ulcers at week 54, no surgery for bowel resection or abscess, and no new fistula. Again, incorporating proactive TDM data into clinical decision-making showed no significant benefit, with comparable proportions of patients achieving the primary endpoint across two dosing arms that utilised infliximab trough levels and one dosing arm that relied on symptoms and biomarkers alone (P = 0.50).

Fernandes et al. conducted a prospective study of proactive TDM, with a historical control group that received conventional treatment.[63] IBD patients completing infliximab induction were prospectively enrolled in the proactive TDM group (n = 135) and were subsequently optimised to maintain infliximab trough levels between 5 and 10 μg/mL, with the primary outcome of faecal calprotectin remission (<250 μg/g) at 2 years of follow-up. The majority of patients were biologic naïve (85.6%) and were receiving concomitant immunomodulators (82.3%). At the end of follow-up, rates of faecal calprotectin remission were higher in the proactive TDM group compared to the control group (72.0% vs 53.8%, P = 0.018). Proactive TDM was also associated with higher rates of clinical remission (87.8% vs 72.5%, P = 0.014).[63] In an earlier study of the same cohort, CD patients (n = 153) that were optimised by proactive TDM had higher rates of endoscopic remission (75.8% vs 41.7%, P < 0.001) and fewer adverse disease outcomes (48.5% vs 70.0%, P = 0.019) when compared to the historical control group that received standard care.[64]

Individualised Bayes-based Dosing. The literature search identified two studies of Bayesian-based anti-TNF dosing in prospective cohorts of IBD patients receiving maintenance infliximab therapy.[12,65] The PRECISION trial (n = 80) evaluated proactive dashboard dosing for adult IBD patients in stable clinical remission.[12] Patients were randomised to either a 'precision dosing' group where infliximab maintenance doses were individualised to achieve and maintain trough levels at a target of 3 μg/mL using a dashboard system or a control group without dose adjustments. The dashboard system utilised Bayesian software and individual patient data, including drug levels, antidrug antibody titres, inflammatory markers, and other patient-specific variables such as body weight and gender, to proactively determine the optimal dose and infusion interval. A higher proportion of patients optimised by dashboard dosing remained in clinical remission, with fewer relapses, during 12 months of treatment compared to those who received standard care (90.9%% vs 63.6%, P = 0.008). Patients in the precision dosing group also had lower faecal calprotectin levels (47 μg/g [IQR 6–120] vs 144 μg/g [IQR 59–666], P = 0.031) but this was not observed for CRP levels. Similarly, Juncosa et al found that clinical remission rates increased significantly from baseline in the 34 CD patients where treatment was proactively intensified with guidance from a dashboard system to maintain infliximab trough levels >3 μg/mL (23.5%-38.2%, P = 0.007).[65]

Tight Control. The CALM trial[3] compared early treatment escalation using a tight control strategy to standard clinical management in a prospective cohort of 244 adult patients with moderate to severe CD. The tight control group had treatment escalations according to strict clinical and biochemical criteria (CDAI ≥150, CRP≥5 mg/L, faecal calprotectin ≥250 μg/g or prednisolone use). The standard care group was escalated based on clinical disease activity alone (CDAI decrease of <100 points compared with baseline or CDAI ≥200 or prednisolone use). The primary endpoint of mucosal healing (CDEIS <4) with the absence of deep ulcers at 48 weeks was achieved by 46% of the tight control group compared to 30% of the clinical management group (P = 0.010). Overall, more patients in the tight control group were escalated to adalimumab 40 mg weekly compared to those in the standard care group. A retrospective follow-up study of the CALM cohort found that patients who achieved endoscopic or deep remission after 1 year of intensified treatment were significantly less likely to have disease progression, including new fistulas, strictures and hospitalisation or surgery for CD, over a median of three years.[4] In a 'real world' context, a retrospective study of dose-escalated CD patients (n = 149) found that management via a virtual clinic, where there is a tendency towards a 'tight control' approach with regular biomarker analysis, is associated with higher rates of treatment escalation success compared to standard outpatient care (P < 0.001).[66]

Relationship Between Anti-TNF Dosing Schedule and Serum Drug Level

Thirteen studies reported on serum drug levels following dose escalation.[22,23,29,45,53,54,67–74] Of these, seven showed significantly higher drug levels in patients on escalated anti-TNF therapy compared to standard dosing.[29,45,53,54,68–70] In a prospective study[29] of 168 CD patients on maintenance adalimumab therapy, over half (65.4%) required escalation to weekly dosing. Adalimumab trough levels increased after dose escalation compared to baseline (4.8 μg/mL [2.3–8.9] to 9.4 μg/mL [1.2–16.4], P = 0.001). The increase in drug level was significantly higher for those that responded clinically to dose-escalated therapy versus nonresponders (5.9 μg/mL [1.9–7.3] vs 0.0 μg/mL [0.0–1.7], P < 0.0001). In a post hoc analysis[69] of the Steenholdt RCT, infliximab trough levels were found to be increased across all patients after the treatment interval was shortened to 4 weeks. However, the increase in drug level from the point of standard treatment failure was significantly higher in those that regained clinical response during the 12 week follow-up after dose escalation compared to nonresponders (9.9 μg/mL [7.4–12.9] vs. 4.7 μg/mL [1.7–10.6), P = 0.040).

These findings are supported by the Paul et al. study,[71] where an increase in infliximab trough level >0.5 μg/mL after dose escalation was strongly associated with clinical remission (P = 0.044) and mucosal healing (P = 0.048). In a post hoc analysis of the TAILORIX study by Dreesen et al.,[53] infliximab trough levels after dose escalation were found to be significantly higher in patients without endoscopic ulceration at 54 weeks compared to those with endoscopic ulceration (8.0 μg/mL [6.7–9.3] vs 5.1 μg/mL [4.5–5.8], P = 0.002). Similarly, an infliximab trough level >7 μg/mL at all time points following dose escalation was associated with radiologic response (P = 0.039) and remission (P = 0.019) in a sub-group analysis of the TAILORIX cohort.[75] In five studies, however, an exposure-response relationship following treatment intensification was not observed.[23,68,72–74,76]

Factors Affecting Response to Dose Escalation

Twenty-two studies analysed factors which affect response to escalated anti-TNF therapy.

Clinical Disease Activity. In a small prospective study (n = 42) examining the efficacy of adalimumab 80 mg weekly in patients with active luminal CD, CDAI<260 at baseline predicted short term clinical remission following dose escalation (P = 0.0016).[25] Motoya et al[39] found that baseline CDAI was significantly lower in patients who achieved clinical remission at week 24 after escalation to adalimumab 80 mg every other week compared to those not achieving remission (OR 0.961, 95% CI 0.9307–0.9920). In the Suzuki et al. trial,[49] baseline CDAI <220 did not predict clinical remission at 40 weeks after dose escalation (P = 0.116). The absence of prolonged steroid use was associated with sustained steroid-free remission after dose escalation in one study,[28] while the need for concurrent steroids at anti-TNF induction predicted nonresponse to dose escalation in another study.[30]

Disease Phenotype. In the majority of literature reviewed, disease phenotype did not predict clinical response to dose escalation.[25,30,39,45,51,56,59] In two retrospective studies, patients with complicated disease were more likely to respond to escalated therapy than patients with an inflammatory phenotype.[27,47] However, in two other retrospective studies, nonstricturing nonpenetrating disease was associated with successful dose escalation.[66,77]

Disease Duration. Disease duration did not affect clinical response to escalated anti-TNF treatment in most cases.[25,27,39,45] Older age at diagnosis was associated with a better response to dose escalation in a small number of studies.[47,60,66]

Prior Anti-TNF Exposure. Of 10 studies, six showed no difference in response to escalated therapy according to previous anti-TNF exposure.[15,25,27,45,47,57] In the Restellini et al. study,[60] patients without prior infliximab exposure were more likely to achieve composite remission at 12 months following adalimumab escalation compared to anti-TNF experienced patients (41.7% vs 20%, P = 0.025). Ma et al.[30] found that previous anti-TNF exposure predicted earlier time to tertiary loss of response (P = 0.048).

Concomitant Immunomodulators. Of 11 studies evaluating the role of combination therapy in patients with loss of response to anti-TNF agents,[27,30,39,44,45,47–49,51,57,60] only one study[57] showed a therapeutic benefit. Srinivasan et al[57] found that immunomodulator co-therapy was associated with a significantly longer time to treatment failure following dose escalation compared to anti-TNF monotherapy (P < 0.03).

Inflammatory Biomarkers. There were conflicting data about the effect of biochemical inflammation at baseline on the response to dose escalation. In five retrospective studies, lower CRP and faecal calprotectin at standard treatment failure were associated with a durable clinical response to dose escalation.[16,27,47,57,60] However, five other studies did not find any association between biomarker levels at loss of response to standard dosing and the success of treatment escalation.[25,30,39,51,54] Higher serum albumin,[49,57,60,71] lower platelet count,[60] and higher haemoglobin levels[60] at baseline predicted a better response to anti-TNF escalation.

Serum Drug Level and Antidrug Antibodies at Baseline. In three retrospective studies, serum drug levels at the time of treatment failure did not distinguish responders to dose escalation from nonresponders.[45,57,60] However, in a prospective single-arm trial[49] (n = 39) of intensified infliximab therapy, clinical remission rates at 40 weeks after dose escalation were higher in patients with a trough infliximab level ≥1 μg/mL at baseline compared to those with a level <1 μg/mL (P = 0.033). In a prospective study of therapeutic drug monitoring, Paul et al[71] showed that an infliximab trough level <2 μg/mL along with an antidrug antibody titre <200 ng/mL at baseline were strongly predictive of subsequent mucosal healing after dose escalation (P = 0.004). Dreesen et al[54] found that an antibody titre <282 ng/mL at baseline independently predicted therapeutic drug levels (P = 0.003) and clinical response (P = 0.034) after dose intensification compared to patients with an antibody titre >282 ng/mL.

Efficacy of Escalated Anti-TNF Dosing for Perianal Crohn's Disease

Four studies reported perianal disease outcomes in dose-escalated patients.[12,20,21,78,79] In a cross-sectional study of 117 CD patients with fistulising perianal disease, patients that achieved the primary outcome of fistula healing were more likely to be on escalated dosing regimens when compared to those who did not (OR 8.3, 95% CI 3.5–19.9).[78] Serum infliximab levels were also higher in patients that achieved fistula healing compared to those with active fistulas (15.8 μg/mL [IQR 9.9–27] vs 4.4 μg/mL [IQR 0–9.8], P < 0.0001). Similarly, in a retrospective study by Strik et al patients on intensified infliximab dosing tended to have higher rates of fistula closure (80%) compared to those on standard dosing (59%).[79] Infliximab trough levels were again higher amongst those that achieved fistula closure (6.0 μg/mL [IQR 5.4–6.9] vs 2.3 μg/mL [IQR 1.1–4.0], P < 0.001) compared to those who did not. This was also seen in the adalimumab-treated patients, with higher drug levels in patients with fistula closure compared to active fistulising disease (7.4 μg/mL [IQR 6.5–10.8] vs 4.8 μg/mL [IQR 1.7–6.2], P = 0.003). However, the rate of fistula closure was numerically higher in the standard dosing group (81.8%) compared to the dose intensified group (50%).[79] In the PRECISION trial, treatment was de-escalated for three patients with infliximab trough levels above 3 μg/mL, which was the set therapeutic target.[12] However, this resulted in a flare of perianal disease, with re-opening of a previously closed perianal fistula in the three patients.[12]