Crohn’s disease
We identified 889 eligible patients with CD (Fig. 1). The mean (SD) age was 36.9 (14.2) years, 44.2% were female, 73.0% were treatment-naive (Table 1), and 1185 treatment episodes were used during follow-up. Because only 10 patients used CTZ, they were excluded from all analyses. Therefore, a total of 1175 treatment episodes were included.
Table 1
Baseline characteristics of patients with CD and UC
Characteristics | Patients with CD n = 889 | Patients with UC n = 255 |
Age, mean (SD), y | 36.9 (14.2) | 41.3 (15.6) |
Sex, n (%) | | |
Male | 496 (55.8) | 120 (47.1) |
Female | 393 (44.2) | 135 (52.9) |
Socio-economic status, n (%) | | |
Low | 140 (15.7) | 32 (12.5) |
Medium | 461 (51.9) | 130 (51.0) |
High | 288 (32.4) | 93 (36.5) |
Treatment episode at baseline, n (%) | | |
First | 649 (73.0) | 183 (71.8) |
Second | 178 (20.0) | 56 (22.0) |
Third | 59 (6.6) | 15 (5.9) |
Fourth or later | 3 (0.3) | 1 (0.4) |
Switches during follow-up, n (%) | | |
None | 648 (72.9) | 169 (66.3) |
1 | 191 (21.5) | 61 (23.9) |
2 | 44 (4.9) | 24 (9.4) |
≥ 3 | 6 (0.7) | 1 (0.4) |
Disease duration, mo | 66.0 (17.0, 143.0) | 36.0 (1.0, 121.5) |
Oral corticosteroids courses, median (IQR) | 0.0 (0.0, 1.0) | 1.0 (0.0, 2.0) |
Visits to a primary care physician, median (IQR) | 9.0 (5.0, 13.0) | 11.0 (6.0, 14.0) |
At least one hospitalization, n (%) | 232 (26.1) | 74 (29.0) |
Comorbidities, n (%) | | |
Cardiovascular disease | 39 (4.4) | 17 (6.7) |
Diabetes | 32 (3.6) | 23 (9.0) |
Hypertension | 93 (10.5) | 33 (12.9) |
Obesity | 102 (11.5) | 42 (16.5) |
Chronic kidney disease | 71 (8.0) | 26 (10.2) |
Osteoporosis | 92 (10.3) | 33 (12.9) |
Cancer | 42 (4.7) | 19 (7.5) |
Depression and anxiety | 137 (15.4) | 56 (22.0) |
Charlson Comorbidity Index, median (IQR) | 0.0 (0.0, 1.0) | 0.0 (0.0, 1.0) |
CD, Crohn’s disease; COPD, chronic obstructive pulmonary disease; IQR, interquartile range; UC, ulcerative colitis. |
Figure 2 shows the disposition of patients according to treatment episode and the year of treatment initiation. Among treatment-naive patients (Fig. 2A), the use of ADA increased between 2015 and 2016 (from 53.9–62.7%) and remained stable in 2017 (61.8%). The use of IFX decreased (from 44.5–26.1%) while the use of VDZ increased, reaching 11.3% in 2017. UST was first introduced for treatment-naive patients with CD in 2018, and only two patients used it in 2017. Among patients initiating their second treatment episode (Fig. 2B), there was a decrease in the use of both ADA (from 52.9–34.3%) and IFX (from 27.1–16.4%) during 2015–2018. The use of VDZ increased from 20.0% in 2015 to 39.3% in 2017 and then decreased to 29.9% in 2018, following the introduction of UST as a treatment option for the second or later biologic. Among those initiating their third or later treatment episode (Fig. 2C), the annual use of ADA and IFX was low (< 10% each). VDZ was the most prevalent treatment in 2015 and 2016 (96.4% and 84.2%, respectively) but use decreased, reaching 18.6% in 2018. UST use increased from 7.9% in 2016 to 74.6% in 2018.
Across all treatment episodes, the proportion of highly adherent patients (PDC ≥ 80%) ranged from 75.6–77.8% (Fig. 3). The proportion of high adherence was significantly higher at 79.6% (P < 0.001) for drugs with intravenous (IV) administration (ie, VDZ, IFX) compared with 72.4% for drugs with subcutaneous (SC) administration (ie, UST, ADA). Similarly, adherence to drugs administered every 4 to 11 weeks (ie, UST, VDZ, IFX) was significantly higher (79.6%) than for drugs administered every 1 to 2 weeks (ie, ADA, 72.4%). High adherence was most prevalent for VDZ (83.7%), followed by IFX (80.0%), ADA (72.4%), and UST (68.9%; P < 0.001).
In treatment episodes in which patients had ≥ 12 months of follow-up (n = 1047 episodes), discontinuation occurred in 32.1% of all episodes (Table 2). Termination versus switching of any biologic therapy occurred in 13.8% versus 13.9% of patients, respectively, and restarting occurred in 4.3%. Discontinuation of IFX and VDZ was mostly the result of switching, whereas most patients discontinued ADA and UST due to termination of treatment.
Table 2
Discontinuation rates by type, treatment episode, administration mode, dosing schedule, and drug for patients with CD and UC during the first 12 months of follow-up
| Discontinuation type, % | Total |
| Restarting the same drug after a gap | Switching to a new treatment episode | Treatment termination | Rate, % | P value |
Treatment | | | | | |
CD | | | | | |
Episode | | | | | |
First (n = 649) | 4.8 | 12.6 | 14.2 | 31.6 | 0.811 |
Second (n = 258) | 3.1 | 17.4 | 13.2 | 33.7 |
Third or later (n = 140) | 4.3 | 13.6 | 13.6 | 31.4 |
Administration mode | | | | | |
IV (n = 526) | 3.8 | 18.3 | 13.3 | 35.4 | 0.023 |
SC (n = 521) | 4.8 | 9.6 | 14.4 | 28.8 |
Dosing schedule | | | | | |
Every 1–2 wk (n = 513) | 4.7 | 9.4 | 14.4 | 28.5 | 0.014 |
Every 4–11 wk (n = 534) | 3.9 | 18.4 | 13.3 | 35.6 |
Drug | | | | | |
ADA (n = 513) | 4.7 | 9.4 | 14.4 | 28.5 | 0.062 |
IFX (n = 287) | 4.9 | 18.8 | 12.5 | 36.2 |
VDZ (n = 201) | 2.0 | 19.9 | 14.4 | 36.3 |
UST (n = 46) | 6.5 | 8.7 | 13.0 | 28.3 |
Overall (n = 1047) | 4.3 | 13.9 | 13.8 | 32.1 | NA |
UC | | | | | |
Episode | | | | | |
First (n = 183) | 2.7 | 25.7 | 13.7 | 42.1 | 0.594 |
Second or later (n = 144) | 4.2 | 15.3 | 15.3 | 34.7 |
Administration mode | | | | | |
IV (n = 234) | 3.8 | 18.4 | 11.1 | 33.3 | 0.008 |
SC (n = 93) | 2.2 | 28.0 | 22.6 | 52.7 |
Dosing schedule | | | | | |
Every 1–2 wk (n = 93) | 2.2 | 28.0 | 22.6 | 52.7 | 0.008 |
Every 4–11 wk (n = 234) | 3.8 | 18.4 | 11.1 | 33.3 |
Drug | | | | | |
ADA (n = 93) | 2.2 | 28.0 | 22.6 | 52.7 | 0.03 |
IFX (n = 82) | 4.9 | 40.2 | 6.1 | 51.2 |
VDZ (n = 152) | 3.3 | 6.6 | 13.8 | 23.7 |
Overall (n = 327) | 3.4 | 21.1 | 14.4 | 38.8 | NA |
ADA, adalimumab; CD, Crohn’s disease; IFX, infliximab; IV, intravenous; NA, not available; SC, subcutaneous; UC, ulcerative colitis; UST, ustekinumab; VDZ, vedolizumab. |
P values < 0.05 are bolded. |
Discontinuation rates were similar across all treatment episodes (31.4–33.7%, P = 0.811; Table 2) and across the different drugs (P = 0.062). Discontinuation was higher (P = 0.023) for IV therapies compared with SC therapies and for patients using drugs administered every 4 to 11 weeks compared with every 1 to 2 weeks.
Significant differences (P = 0.017) were recorded in persistence with therapy by drug (Fig. 4). At 12 months of follow-up, the highest discontinuation percentages were calculated for those using IFX (36.5% [95% CI, 30.8%, 41.8%]) and VDZ (36.2 [29.6%, 42.4%]) and the lowest for those using UST (25.6% [14.6%, 35.4%]) and ADA (27.9% [24.0%, 31.6%]). Differences in discontinuation rates did not reach statistical significance when the analysis was limited to treatment-naive (P = 0.11) or treatment-experienced (P = 0.17) patients.
The multivariable analyses (Table 3) showed that, compared with ADA, IFX increased the risk for discontinuation among treatment-naive patients (hazard ratio [HR], 1.38 [95% CI, 1.03, 1.84]) but not among treatment-experienced patients (HR, 1.21 [0.70, 2.08]).
Table 3
Adjusted HRs and 95% CIs for treatment discontinuation by treatment experience status in the first year among patients with CD and UC with ≥ 12 months of follow-up
| Patients with CD | Patients with UC |
| Treatment-naive n = 647* | Treatment-experienced n = 398 | Treatment-naive n = 183 | Treatment-experienced n = 135† |
| HR | 95% CI | P value | HR | 95% CI | P value | HR | 95% CI | P value | HR | 95% CI | P value |
Age | 1.01 | 1.00, 1.02 | 0.039 | 0.99 | 0.98, 1.00 | 0.2 | 1.01 | 1.00, 1.03 | 0.14 | 1 | 0.98, 1.03 | > 0.9 |
Sex | | | | | | | | | | | | |
Male | — | — | | — | — | | — | — | | — | — | |
Female | 0.89 | 0.67, 1.18 | 0.4 | 1.08 | 0.76, 1.52 | 0.7 | 1.15 | 0.73, 1.83 | 0.5 | 0.69 | 0.36, 1.30 | 0.2 |
SES | | | | | | | | | | | | |
Low | — | — | | — | — | | — | — | | — | — | |
Medium | 0.77 | 0.54, 1.11 | 0.2 | 0.92 | 0.57, 1.48 | 0.7 | 0.98 | 0.48, 1.98 | > 0.9 | 0.6 | 0.25, 1.40 | 0.2 |
High | 0.63 | 0.42, 0.94 | 0.025 | 0.89 | 0.52, 1.54 | 0.7 | 0.87 | 0.41, 1.82 | 0.7 | 0.62 | 0.25, 1.53 | 0.3 |
Disease duration, y | 1.01 | 0.99, 1.03 | 0.4 | 0.99 | 0.96, 1.02 | 0.6 | 0.99 | 0.95, 1.02 | 0.5 | 1.01 | 0.96, 1.06 | 0.8 |
Drug | | | | | | | | | | | | |
ADA | — | — | | — | — | | — | — | | — | — | |
IFX | 1.38 | 1.03, 1.84 | 0.03 | 1.21 | 0.70, 2.08 | 0.5 | 0.82 | 0.49, 1.37 | 0.5 | NA |
VDZ | 1.15 | 0.63, 2.09 | 0.7 | 1.39 | 0.91, 2.13 | 0.13 | 0.28 | 0.14, 0.54 | < 0.001 | 0.38 | 0.21, 0.70 | 0.002 |
UST | NA | 1.05 | 0.55, 2.02 | 0.9 | NA | NA |
Charlson Comorbidity Index | 0.82 | 0.69, 0.98 | 0.03 | 1.02 | 0.86, 1.20 | 0.9 | 0.98 | 0.79, 1.20 | 0.8 | 0.98 | 0.73, 1.30 | 0.9 |
ADA, adalimumab; CD, Crohn’s disease; HR = hazard ratio; IFX, infliximab; NA, not available; SES, socioeconomic status; UC, ulcerative colitis; UST, ustekinumab; VDZ, vedolizumab. |
*2 patients receiving UST were excluded. |
†9 patients receiving IFX were excluded. |
P values < 0.05 and corresponding HR and 95% CIs are bolded |
Ulcerative colitis
A total of 255 patients with UC were included. The mean (SD) age was 41.3 (15.6) years, 52.9% were female, and 71.8% were treatment-naive (Table 1), and 366 treatment episodes were used during follow-up. Treatment episodes that included UST (n = 1), GOL (n = 8), and TOF (n = 4) were excluded from the analyses because of the small number of patients. Therefore, a total of 354 treatment episodes were included.
Figure 5 shows the disposition of patients according to treatment episode and the year of treatment initiation. Among the 183 treatment-naive patients during 2015–2017, the use of ADA ranged from 18.6–34.1%; the use of IFX decreased from 54.5–27.5%; and the use of VDZ increased from 11.4–47.5% (Fig. 5A). Among patients starting their second or later treatment episode during 2015–2018, use of ADA varied between 22.2% and 38.5%, as did the use of IFX (between 2.6% and 29.6%). The use of VDZ decreased, from 66.7% in 2015 to 48.1% in 2018 (Fig. 5B).
Across all treatment episodes, the proportion of highly adherent patients (PDC ≥ 80%) ranged from 80.3–86.0% (Fig. 6). The proportion of high adherence was higher (88.6%; P < 0.001) among those using IV-administered drugs (ie, IFX and VDZ) versus the SC drug, ADA (63.6%). Adherence to drugs administered every 4 to 11 weeks (ie, IFX, VDZ) was higher (88.6%; P < 0.001) than adherence to drugs administered every 1 to 2 weeks (63.3%; ie, ADA). In all treatment episodes, the highest adherence rates were for VDZ (89.1%), IFX (87.8%), and ADA (63.6%).
In patients with UC, discontinuation at 12 months was analyzed only for treatment episodes in which patients had ≥ 12 months of follow-up (n = 327 episodes; Table 2). The overall discontinuation rate was 38.8%, with no significant difference observed between treatment-naive patients (42.1%) and treatment-experienced patients (34.7%; P = 0.594). The 52.7% discontinuation rate for ADA (SC administration, every 1–2 weeks) was higher (P = 0.008) than for IFX and VDZ (33.3%; IV administration every 4 to 11 weeks). The most common type of discontinuation was switching, except for VDZ, for which discontinuation was mostly due to treatment termination.
The Kaplan-Meier analysis (Fig. 4) showed that VDZ had the lowest discontinuation rate in the first 12 months (23.2% [95% CI, 16.4%, 29.5%]). Higher discontinuation rates were observed for IFX (49.5% [37.8%, 59.0%]) and ADA (52.5% [42.2%, 61.5%]). In treatment-naive patients, VDZ also had a lower (P = 0.0023) discontinuation rate (21.9% [11.1%, 31.4%]) than ADA (56.5% [39.6%, 68.7%]) and IFX (50.7% [37.8%, 60.9%]; P = 0.0023). Similarly, in the treatment-experienced group, VDZ had a lower discontinuation rate (24.3% [15.1%, 32.5%]) than ADA (48.9% [33.1%, 61.1%]) and IFX (47.9% [6.8%, 70.8%]; P = 0.0023). The multivariable analyses (Table 3) showed that compared with ADA, both treatment-naive and treatment-experienced patients using VDZ had decreased risk for discontinuation (HR, 0.28 [95% CI, 0.14, 0.54] and 0.38 [0.21, 0.70], respectively).