TCZ-IV Clinical Trial Database Population
The TCZ-IV all-exposure RA clinical trial population includes data from 5185 patients who received ≥1 dose of TCZ-IV and includes all data from the time of first TCZ dose, representing 17,905.9 PYs of exposure (Table 1). From the SMQ “GIP—narrow” search, 70 events were reported in 53 patients. Of these, 34 events in 31 patients were adjudicated as GIPs, for an incidence rate (95% CI) of 1.9 (1.3–2.7) events per 1000 PYs. Of these 34 events, 9 occurred in 9 patients (29%) with a history of diverticular disease, gastritis, or ulcer, and 16 occurred in 16 patients (52%) with a diagnosis of diverticular disease at the time of surgery or during the course of the clinical trial. There was no increase in number of GIPs with increased TCZ exposure or duration of study (Table 1). The overall incidence rate observed during greatest PY exposure period (>36 months, with 6872.9 PYs of exposure) was comparable with rates observed at earlier time points. The majority of GIPs (85%) in the TCZ-IV all-exposure population occurred in the lower GI tract.
Table 1 Incidence rates and number of GIPs in TCZ-IV-exposed patients with rheumatoid arthritis in clinical trials by 6-month periods
In the all-exposure TCZ-IV clinical trial population, baseline demographic data were comparable between the overall population and patients who experienced a GIP (Table 2). The mean (standard deviation) age was greater for patients who experienced an adjudicated GIP than for the overall population: 58.5 (10.6) years compared with 51.7 (12.8) years—a trend with age being a risk factor for GIP [17]. The proportion of patients with prior exposure to aTNF agents was 22.6% for those who experienced an adjudicated GIP compared with 17.0% for the overall population. Of patients with an adjudicated GIP, 17.9% had a history of smoking compared with 18.0% of patients in the overall population.
Table 2 Baseline characteristics of TCZ-IV-exposed patients with rheumatoid arthritis in clinical trials
SC-administered TCZ has been studied in 2 RA clinical trials, with a total of 2039 PYs of exposure (n = 1374). The incidence rate of GIP events in the TCZ-SC all-exposure population was slightly lower (1.5 per 1000 PYs) than in the TCZ-IV all-exposure population. Given the limited exposure and lack of increased risk, data from the TCZ-SC clinical trials are not included in this analysis.
TCZ Global Postmarketing Safety Database
The SMQ “GIP—narrow” search retrieved 449 events in 437 patients from the global postmarketing safety database for which a causal or temporal association between GIPs and TCZ could not be excluded, for an overall reporting rate (95% CI) of 1.2 (1.1–1.3) GIP events per 1000 PYs over 382,621 PYs of postmarketing TCZ exposure (Table 3).
Table 3 GIPs in patients treated with TCZ reported to the global postmarketing safety database
Of 295 patients with available information for a comprehensive medical analysis, perforation occurred in the upper GI tract in 36 patients, the lower GI tract in 202 patients, and an unspecified location in 60 patients; 3 patients experienced >1 GIP (Table 3). Of the 202 patients with a lower GIP, 152 (75.2%) were receiving glucocorticoids, 96 (47.5%) were receiving NSAIDs, 78 (38.6%) were receiving both glucocorticoids and NSAIDs, and 17 (8.4%) were receiving aspirin. Also of the 202 patients with a lower GIP, 135 (66.8%) had reported diverticular problems, including a medical history in 23 patients (11.4%; 14 with diverticulosis, 7 with diverticulitis, and 2 with both diverticulosis and diverticulitis), concurrent diverticular disease in 22 patients (10.9%; 14 with diverticulosis, 4 with diverticulitis, and 4 with both diverticulosis and diverticulitis; 1 patient had both concurrent and medical history of concurrent diverticular disease), and 91 patients with no medical history of concurrent condition of diverticular disease for whom diverticulosis, diverticulitis and/or diverticular perforation was reported at the time of GIP diagnosis (7 with diverticulosis, 17 with diverticulitis, 24 with diverticular perforation, 5 with diverticular perforation accompanied by diverticulosis, 36 with diverticular perforation accompanied by diverticulitis, and 2 with diverticular perforation accompanied by both diverticulosis and diverticulitis).
Of 437 patients with GIP retrieved from the postmarketing safety database, 21 (4.8%) had fatal outcomes. In 18 of these 21 patients, GIP was reported as the underlying event of death. Of these 18 patients, perforation occurred in the lower GI tract in 14 patients, the upper GI tract in 2 patients, and unspecified locations in 2 patients. In the remaining 3 patients, the causes of death were peritonitis secondary to perforation of colonic diverticulitis, perforative peritonitis and disseminated intravascular coagulation, and unrelated causes.
US Insurance Claims Database
A total of 58,334 patients with RA treated with an aTNF, abatacept, or TCZ who met the eligibility criteria for analysis were identified in the healthcare claims. Patients were followed for a mean of 535 days. Overall, approximately 50% of patients had no evidence of prior exposure to a qualifying biologic DMARD. However, of the patients who initiated TCZ, most (88%) had prior biologic exposure; therefore, the safety analysis reported in detail here focused only on the population of patients who had received ≥1 prior biologic DMARD.
Of the patients with prior biologic exposure, the mean age was similar for those who received an aTNF, abatacept, or TCZ (52.2, 54.4, and 54.1 years, respectively; Table 4). Patients who received TCZ tended to have received more prior biologics; 27.3% of patients in the TCZ group received ≥3 prior biologics compared with 11.4% in the abatacept group and 7.8% in the aTNF group. Similar proportions of patients across the groups had exposure to prescription NSAIDs (43.6%, 41.0%, and 41.9% for the aTNF, abatacept, and TCZ groups, respectively). Mean exposure to oral glucocorticoids was also similar across groups.
Table 4 Baseline demographic and clinical characteristics of patients with rheumatoid arthritis and prior biologic exposure from a healthcare insurance claims database
Of the patients with prior biologic exposure, there were 21 GIPs (10 from the aTNF group, 5 from the abatacept group, and 6 from the TCZ group) per the specific definition and 31 (14 from the aTNF group, 8 from the abatacept group, and 9 from the TCZ group) per the sensitive definition (Table 5). The GIP incidence rate (95% CI) per the specific definition for the TCZ group was 1.8 (0.7–4.0) per 1000 PYs compared with 0.6 (0.3–1.2) and 0.8 (0.3–2.0) per 1000 PYs for the aTNF and abatacept groups, respectively. According to the specific definition, the incidence rates (95% CI) for the individual aTNF agents ranged from 0.4 (0.1–1.4) to 1.3 (0.3–3.8) per 1000 PYs with widely overlapping CIs. According to the sensitive definition, the incidence rate (95% CI) per 1000 PYs for the TCZ group was 2.8 (1.3–5.2) compared with 0.9 (0.5–1.5) and 1.4 (0.6–2.7) for the aTNF (with widely overlapping CIs) and abatacept groups, respectively. The majority of GIPs (71.4%) were reported in the lower GI tract, regardless of treatment or definition. The lower GI tract perforations were mostly associated with diverticular diagnoses, including three cases in the aTNF group, four in the abatacept group, and two in the TCZ group.
Table 5 Incidence ratesa of GIPs using sensitive and specific definitions in patients with rheumatoid arthritis and prior biologic exposure from a healthcare insurance claims database
The unadjusted absolute rate differences (95% CIs) between the TCZ and aTNF groups for all GIPs were 1.2 (−0.3 to 2.5) and 1.9 (0.0–3.7) per 1000 PYs per the specific and sensitive definitions, respectively. For lower GIPs, the unadjusted absolute rate differences were 1.2 (−0.2 to 2.5) and 1.8 (0.1–3.6) per 1000 PYs per the specific and sensitive definitions, respectively.
Multivariate analysis compared adjusted GIP incidence rates between aTNF and TCZ cohorts with prior exposure to ≥1 biologic. The adjusted IRRs (95% CI) for all GIPs in patients who received TCZ versus those who received an aTNF were 2.2 (0.7–6.6; specific definition) and 2.2 (0.9–5.4; sensitive definition). For lower GIP events, the adjusted IRRs in patients who received TCZ versus those who received an aTNF were 4.0 (1.1–14.1; specific definition) and 3.1 (1.1–8.4; sensitive definition). Both the number of prior biologics and the comorbidities index were significantly associated with the risk of GIP events [IRR: 1.5 (95% CI, 1.0–2.2), P = 0.036 and 1.2 (95% CI, 1.0–1.4), P = 0.018, respectively].
In the biologic-naïve populations, no patients treated with TCZ (12% of the overall TCZ exposure) experienced a GIP; in patients treated with aTNFs (75-times greater patient exposure than for TCZ), the GIP incidence rate was similar to the overall aTNF-exposed population.