SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Objective

To compare the incidence of cancer following tumor necrosis factor α (TNFα) inhibitor therapy to that with commonly used alternative therapies across multiple immune-mediated diseases.

Methods

The Safety Assessment of Biological Therapeutics study used data from 4 sources: national Medicaid and Medicare databases, Tennessee Medicaid, pharmacy benefits plans for Medicare beneficiaries in New Jersey and Pennsylvania, and Kaiser Permanente Northern California. Propensity score–adjusted hazard ratios (HRs) and 95% confidence intervals (95% CIs) were computed to estimate the relative rates of cancer, comparing those treated with TNFα inhibitors to those treated with alternative disease-modifying therapies. The cancer-finding algorithm had a positive predictive value ranging from 31% for any leukemia to 89% for female breast cancer.

Results

We included 29,555 patients with rheumatoid arthritis (RA) (13,102 person-years), 6,357 patients with inflammatory bowel disease (1,508 person-years), 1,298 patients with psoriasis (371 person-years), and 2,498 patients with psoriatic arthritis (618 person-years). The incidence of any solid cancer was not elevated in RA (HR 0.80 [95% CI 0.59–1.08]), inflammatory bowel disease (HR 1.42 [95% CI 0.47–4.26]), psoriasis (HR 0.58 [95% CI 0.10–3.31]), or psoriatic arthritis (HR 0.74 [95% CI 0.20–2.76]) during TNFα inhibitor therapy compared to disease-specific alternative therapy. Among RA patients, the incidence of any of the 10 most common cancers in the US and of nonmelanoma skin cancer was not increased with TNFα inhibitor therapy compared to treatment with comparator drugs.

Conclusion

Short-term cancer risk was not elevated among patients treated with TNFα inhibitor therapy relative to commonly used therapies for immune- mediated chronic inflammatory diseases in this study.

Medications designed to inhibit the effects of tumor necrosis factor α (TNFα inhibitors) have become important components of the treatment of multiple chronic inflammatory disorders, including rheumatoid arthritis (RA), inflammatory bowel disease (Crohn's disease and ulcerative colitis), psoriasis, psoriatic arthritis, and ankylosing spondylitis. During the premarketing trials of infliximab, adalimumab, and etanercept, the incidence of lymphoma was higher in patients treated with these TNFα inhibitors than was expected based on rates in the general population. However, interpretation of these data was complicated by higher reported risk of lymphoma among patients with RA, particularly those with more severe disease (1–3).

In 2006, a systematic review and meta-analysis by Bongartz et al reported a higher incidence of cancer, including lymphoma and nonmelanoma skin cancer, among patients treated with infliximab or adalimumab in placebo-controlled trials (4). A subsequent meta-analysis demonstrated a higher, but not statistically significant, incidence of malignancy with etanercept versus comparison therapies (5). Thus, these studies suggested a possible increased risk of cancer among TNFα inhibitor–treated patients with short-term therapy. In contrast, a meta-analysis of randomized trials of TNFα inhibitors for Crohn's disease found no increased risk of cancer with TNFα inhibitor therapy (6). Similarly, a recent meta-analysis of 74 randomized trials of infliximab, adalimumab, or etanercept for a variety of conditions demonstrated no significant increased risk of cancers other than nonmelanoma skin cancer, for which there was a 2-fold higher incidence with TNFα inhibitors. In that study, there was some evidence that relative risks might vary between specific TNFα inhibitor medications (7). Finally, a meta-analysis of patients with early RA treated with TNFα inhibitors or placebo found no increased risk of cancer with TNFα inhibitor therapy (8).

Because clinical care generally entails a choice between alternative active treatments, we designed the Safety Assessment of Biological Therapeutics (SABER) study as a comparative safety analysis of TNFα inhibitors and disease-specific alternative treatment strategies (9). One component of the SABER study assessed whether the risk of the most common cancers in the US was higher among patients treated with TNFα inhibitors than among those receiving disease-specific alternative therapies. Here we report that the incidence of these common cancers was not significantly higher in patients receiving TNFα inhibitors.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

The SABER study is a retrospective cohort study that combines data from 4 sources: 1) national Medicaid and Medicare databases (Medicaid Analytic eXtract, 2000–2005, excluding Tennessee; Medicare, 2000–2006; and Medicare Part D, 2006); 2) Tennessee Medicaid (TennCare, 1998–2005); 3) New Jersey's Pharmaceutical Assistance to the Aged and Disabled and Pennsylvania's Pharmaceutical Assistance Contract for the Elderly (1998–2006); and 4) Kaiser Permanente Northern California (1998–2007). A common programming algorithm was used to identify patients with autoimmune diseases who were initiating TNFα inhibitors and comparator drugs.

Exposure definitions.

The SABER methods of cohort assembly and definitions of new users of TNFα inhibitors and comparator therapies have been previously reported (9). Briefly, we first identified patients with RA, inflammatory bowel disease, psoriasis, psoriatic arthritis, or ankylosing spondylitis on the basis of International Classification of Diseases, ninth revision (ICD-9) diagnostic codes and medical therapies. We limited the cohort to new users of TNFα inhibitors and/or the comparator therapy, where new use required that patients have 1 full year of data prior to the first prescription that defined a new course of therapy and no use of TNFα inhibitor therapy in all available data within the database. The comparator therapies differed according to the disease being treated: RA—initiation of hydroxychloroquine, sulfasalazine, or leflunomide following therapy with methotrexate; inflammatory bowel disease—initiation of azathioprine or mercaptopurine; psoriasis—initiation of retinoids, high-potency topical steroids, or phototherapy following treatment with methotrexate; psoriatic arthritis and ankylosing spondylitis—initiation of methotrexate or sulfasalazine.

Inclusion and exclusion criteria.

We identified all new users of either TNFα inhibitors or comparator therapies in the 4 data sets. We sought to exclude patients with a history of cancer defined as any code for cancer other than nonmelanoma skin cancer by excluding those with at least 1 ICD-9 diagnosis code recorded in the year prior to the initiation of therapy. We also excluded patients with a history of organ transplant, human immunodeficiency virus infection, or treatment with tacrolimus or cyclosporine during the 1-year look-back period. These latter conditions were used as censoring events if they occurred after the start of followup.

We excluded patients who used another biologic medication from outside the TNFα inhibitor class in the 365-day period prior to exposure and censored patients who initiated biologic agents from outside the TNFα inhibitor class after cohort entry. This was particularly important for rituximab, which can be used to treat lymphoma.

Outcome definitions.

We identified incident cancers for patients in Kaiser Permanente using the Kaiser Permanente Northern California cancer registry. For each of the other data sources, incident cancers were identified using an adaptation of the algorithm developed and validated by Setoguchi et al using Medicare data (10) as we previously employed in assessing rates of malignancy in patients with juvenile idiopathic arthritis (11). For all disease groups, we examined the following outcomes: any lymphoma, any leukemia, any solid cancer, and nonmelanoma skin cancer. For RA patients, we also studied the 10 most common cancers in the US.

Because the Setoguchi algorithm was developed in an older population and for a limited number of cancers, we determined the sensitivity, specificity, and the positive predictive value (PPV) of our adaptation of Setoguchi's algorithm to identify incident cancers for each of the 10 most common cancers in the US. We tested our adaptation of the Setoguchi algorithm as applied to the electronic health record data in Kaiser Permanente using the Kaiser Permanente Northern California cancer registry as the gold standard. This cancer registry is one of several sites that submit data to the Surveillance, Epidemiology, and End Results (SEER) program, the largest cancer registry in the US. SEER case ascertainment rates are documented to be >98%. Nonmelanoma skin cancer is not routinely captured in SEER and therefore was not evaluated (12). (Further details of the validation study are available from the corresponding author.) The sensitivity of the algorithm exceeded 60% for all cancers other than melanoma (56%) and leukemia (28%) (further information is available from the corresponding author). After employing a 1-year period prior to therapy initiation (look back) to exclude patients with cancer diagnosed prior to the start of therapy, the PPV of the algorithm ranged from 31% (any leukemia) to 89% (female breast cancer). The low predictive values were due in part to cancer diagnoses that represented prevalent cancers not detected using the 1-year look-back period, and ranged from 26% (any leukemia) to 61% (prostate cancer) (further information is available from the corresponding author).

Baseline covariates.

As described previously, the list of covariates was developed to capture variables believed to be potentially associated with the exposure of interest and the various outcomes included in the SABER study (9). Baseline covariates included demographics (age, sex, race, residence [urban/rural], nursing home/community dwelling, area income, and calendar year), generic markers of comorbidity (number of hospitalizations, outpatient and emergency room visits, and number of different medication classes filled), and surrogate markers of disease severity (extraarticular disease manifestations, number of intraarticular and orthopedic procedures, number of laboratory tests ordered for inflammatory markers, chronic obstructive pulmonary disease, diabetes, and use of cancer screening tests [prostate-specific antigen testing and mammography]).

Statistical analysis.

For each period of new use of a TNFα inhibitor or comparator drug, we calculated a propensity score to summarize the covariates recorded during the look-back period. Propensity scores were computed using unconditional logistic regression with use of a TNFα inhibitor as the dependent variable. Propensity scores were computed separately in each data set. We then examined the distribution of the propensity scores across each comparator-treated group and excluded TNFα inhibitor–treated patients with propensity scores in areas with no overlap with the comparator-treated group and vice versa (13).

Medication exposure was treated as a time-updated variable such that patients could accrue followup time in one or both of the treatment arms sequentially and not simultaneously. However, once patients were treated with a TNFα inhibitor, they could not accrue followup time in the comparator cohort. In the primary as-treated analysis, followup continued until the earliest of the outcome of interest—the end of the available data or discontinuation of exposure to TNFα inhibitors or comparator drugs. Our definition of discontinuation of therapy allowed for a grace period of 30 days between the expected end of a dispensing of drug and the next prescription. For patients treated with the comparator drugs, initiation of a TNFα inhibitor also marked the end of followup in the comparator cohort and the beginning of followup in the TNFα inhibitor cohort. We also conducted a first exposure carried forward analysis in which patients who discontinued the comparator treatment continued to contribute followup time to that cohort until they initiated TNFα inhibitor therapy or experienced the outcome of interest; patients treated with a TNFα inhibitor contributed followup time to that cohort until the end of the available data or the date that they experienced the outcome of interest. Thus, the first analysis examined only outcomes that occurred while patients were receiving therapy, and the secondary analysis allowed for an indefinite lag between discontinuation of therapy and the onset of cancer. Medians and interquartile ranges (IQRs) are reported.

For each outcome, we computed hazard ratios (HRs) using Cox regression stratified by data set. Because patients could accrue followup time in both the TNFα inhibitor and comparator cohorts, we used the Huber-White “sandwich” variance estimator to calculate robust standard errors for all estimates (14). The only independent variables in the model were the exposure group (TNFα inhibitor versus comparator drugs), use of corticosteroids at baseline, and the propensity score, which was categorized in quintiles. Separate analyses were performed for the outcomes of any lymphoma, any hematologic cancer, any solid cancer other than nonmelanoma skin cancer, and nonmelanoma skin cancer, defined as squamous cell or basal cell, across diseases for which a TNFα inhibitor was indicated. Because the number of cancer diagnoses among patients with diseases other than RA were relatively small, analysis of specific cancers was limited to RA patients. We elected to report HRs only when there were at least 5 patients with the outcome of interest and at least 1 patient with the outcome in each of the treatment arms.

A sensitivity analysis was conducted to estimate the potential effect of prevalent cancers being misdiagnosed as incident cancers. Under the assumption of an observed null association between TNFα inhibitor use and cancer risk, we varied the distribution of observed cancers that were prevalent rather than incident and the distribution of these cancers across the treatment groups to estimate the potential magnitude of bias that may have resulted from this misclassification (further information is available from the corresponding author).

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

From the entire SABER cohort, we included 29,555 RA patients (19,750 taking a TNFα inhibitor, 9,805 taking a comparator drug), 6,357 patients with inflammatory bowel disease (2,657 taking a TNFα inhibitor, 3,700 taking a comparator drug), 1,298 patients with psoriasis (563 taking a TNFα inhibitor, 735 taking a comparator drug), 2,498 patients with psoriatic arthritis (1,036 taking a TNFα inhibitor, 1,462 taking a comparator drug), and 1,486 patients with ankylosing spondylitis (783 taking a TNFα inhibitor, 703 taking a comparator drug). The characteristics of the TNFα inhibitor–treated patients and the comparator-treated groups were generally similar (Table 1).

Table 1. Characteristics of the patients in the cohorts with immune-mediated disease*
 RAIBDPsoriasisPsAAS
TNFα inhibitor treated (n = 19,750)Comparator treated (n = 9,805)TNFα inhibitor treated (n = 2,657)Comparator treated (n = 3,700)TNFα inhibitor treated (n = 563)Comparator treated (n = 735)TNFα inhibitor treated (n = 1,036)Comparator treated (n = 1,462)TNFα inhibitor treated (n = 783)Comparator treated (n = 703)
  • *

    RA = rheumatoid arthritis; TNFα = tumor necrosis factor α; IBD = inflammatory bowel disease; PsA = psoriatic arthritis; AS = ankylosing spondylitis; KPNC = Kaiser Permanente Northern California; TennCare = Tennessee Medicaid; PAAD/PACE = Pharmaceutical Assistance to the Aged and Disabled (New Jersey)/Pharmaceutical Assistance Contract for the Elderly (Pennsylvania); COPD = chronic obstructive pulmonary disease; NA = not applicable; UV = ultraviolet.

Person-years of followup, primary analysis8,242.64,859.7813.22694.9227.1143.5398.1219.9313.3119.8
Person-years of followup, secondary analysis17,850.615,352.12,865.32,290.2485.3481.1779.0642.5657.9462.3
Female, %85.185.866.965.965.259.663.363.030.839.0
Race, %          
 White62.361.678.776.169.668.481.977.766.862.3
 Black16.315.915.715.08.78.63.194.57.710.5
 Other21.422.55.58.921.723.015.017.925.527.2
Age, %          
 <35 years6.15.036.539.116.311.611.111.119.520.9
 35–65 years60.160.455.050.967.360.372.569.472.567.9
 >65 years33.934.68.69.916.428.216.419.57.811.3
Health plan, %          
 Medicaid74.062.873.758.767.557.865.650.670.259.3
 KPNC13.621.514.126.722.229.026.139.122.928.6
 TennCare7.211.411.113.09.211.05.06.26.510.5
 PAAD/PACE5.24.31.11.61.12.23.34.10.41.6
Charlson comorbidity score, %          
 01.82.968.170.960.455.248.859.263.265.0
 156.756.722.221.026.327.931.926.526.823.9
 2+41.440.59.78.013.316.919.414.310.011.1
COPD, %20.020.718.116.617.417.714.914.011.412.7
Household income, median (range) $36,230 (0–200,001)36,763 (0–186,998)37,426 (7,236–154,906)39,150 (8,292–161,452)37,063 (12,351–127,565)38,565 (4,732–142,704)40,208 (9,487–200,001)43,256 (0–200,001)40,105 (10,974–144,406)40,694 (0–142,459)
Hospitalizations in the prior year, median (range)0 (0–25)0 (0–28)1 (0–20)0 (0–16)0 (0–16)0 (0–10)0 (0–13)0 (0–14)0 (0–10)0 (0–14)
Index therapy, %          
 Infliximab34.7NA96.8NA3.6NA21.7NA25.9NA
 Adalimumab22.7NA3.2NA2.5NA11.6NA9.5NA
 Etanercept42.6NANANA94.0NA66.7NA64.6NA
 HydroxychloroquineNA48.9NANANANANANANANA
 LeflunomideNA37.4NANANANANANANANA
 SulfasalazineNA13.7NANANANANA26.5NA55.1
 MethotrexateNANANANANANANA73.5NA45.0
 Azathioprine/mercaptopurineNANANA100.0NANANANANANA
 UV light therapyNANANANANA36.6NANANANA
 Topical steroidsNANANANANA20.3NANANANA
 RetinoidNANANANANA43.1NANANANA

In the RA patient cohort, the followup time was 13,102 person-years (median 0.5 years [IQR 0.2–1.1 years] for those taking a TNFα inhibitor and median 0.3 years [IQR 0.2–0.6 years] for those in the comparator-treated group) in the analysis limited to followup time while patients were receiving the drug. This increased to 33,203 person-years (median 1.5 years [IQR 0.7–2.8 years] for those taking a TNFα inhibitor and median 1.4 years [IQR 0.6–3.0 years] for those in the comparator-treated group) in the first exposure carried forward analysis for RA (Table 1). Among the 9,805 RA patients in the comparator-treated group, 2,400 (24%) were subsequently treated with a TNFα inhibitor. The median followup for this subgroup in the first exposure carried forward analysis was 0.9 years (IQR 0.4–1.9 years) with the comparator drug and 1.4 years (IQR 0.6–2.6 years) with a TNFα inhibitor. For the propensity score–adjusted analysis, the proportion of patients with nonoverlapping propensity scores in the RA solid tumor analysis was 1.5%.

Consistent with the size of the cohorts, the greatest number of cancer diagnoses occurred in the RA patient cohort (Table 2). The cancer incidence rates were similar in the primary and secondary analyses, although the number of events was 2–3-fold higher in the secondary analysis.

Table 2. Cancer incidence and new diagnoses in the study cohorts*
 RAIBDPsoriasisPsAAS
  • *

    Values are cancer incidence rates per 100 person-years in the study cohorts (number of new cancer diagnoses). Rates and counts of cases were computed prior to exclusion of patients at the extremes of the propensity score distributions. See Table 1 for definitions.

  • Followup is censored when the patient discontinues therapy.

  • Followup continues even if the patient discontinues therapy.

Primary analysis     
 Any lymphoma     
  TNFα inhibitor0.12 (22)0.00 (0)0.00 (0)0.00 (0)0.00 (0)
  Comparator0.13 (8)0.05 (<5)0.00 (0)0.00 (0)0.00 (0)
 Any lymphoma or leukemia     
  TNFα inhibitor0.13 (23)0.06 (<5)0.00 (0)0.00 (0)0.00 (0)
  Comparator0.17 (10)0.05 (<5)0.00 (0)0.00 (0)0.00 (0)
 Any solid cancer     
  TNFα inhibitor0.91 (160)0.82 (13)0.78 (<5)0.77 (8)0.44 (<5)
  Comparator1.10 (66)0.41 (9)1.21 (<5)0.81 (7)1.47 (<5)
 Nonmelanoma skin cancer     
  TNFα inhibitor0.31 (54)0.25 (<5)0.52 (<5)0.29 (<5)0.15 (<5)
  Comparator0.35 (21)0.65 (14)1.22 (<5)0.34 (<5)0.73 (<5)
Secondary analysis     
 Any lymphoma     
  TNFα inhibitor0.14 (52)0.00 (0)0.00 (0)0.05 (<5)0.00 (0)
  Comparator0.11 (22)0.05 (<5)0.00 (0)0.06 (<5)0.00 (0)
 Any lymphoma or leukemia     
  TNFα inhibitor0.16 (61)0.08 (<5)0.00 (0)0.05 (<5)0.00 (0)
  Comparator0.16 (31)0.06 (6)0.00 (0)0.06 (<5)0.14 (<5)
 Any solid cancer     
  TNFα inhibitor1.05 (394)0.80 (43)0.61 (5)0.84 (17)0.45 (6)
  Comparator1.14 (225)0.63 (47)1.07 (17)0.84 (30)0.79 (11)
 Nonmelanoma skin cancer     
  TNFα inhibitor0.35 (134)0.26 (14)0.49 (<5)0.39 (8)0.15 (<5)
  Comparator0.39 (77)0.40 (30)0.37 (6)0.11 (<5)0.43 (6)

In the primary analysis, rates of solid cancers were not significantly higher in the TNFα inhibitor–treated group than in the comparator-treated group across all immune-mediated diseases (Table 3) and as shown in Kaplan-Meier curves for RA (Figure 1). For all diseases and cancer types where there was a sufficient number of events to estimate HRs, no significant increase in risk was observed for any lymphoma, any leukemia or lymphoma, or nonmelanoma skin cancer in both the primary and secondary analyses (Table 3).

Table 3. Hazard ratios of incident cancer for TNFα inhibitors versus comparator therapies*
 RAIBDPsAPsoriasisAS
  • *

    Values (except for cancers with <5 events) are the hazard ratio (95% confidence interval). See Table 1 for definitions.

  • Followup is censored when the patient discontinues therapy.

  • Only 2 events remain after propensity score matching.

  • §

    Followup continues even if the patient discontinues therapy.

Primary analysis     
 Any lymphoma0.83 (0.33–2.05)<5 events<5 events<5 events<5 events
 Any lymphoma or leukemia0.65 (0.28–1.53)<5 events<5 events<5 events<5 events
 Any solid cancer0.80 (0.59–1.08)1.42 (0.47–4.26)0.74 (0.20–2.76)0.58 (0.10–3.31)0.03 (0.002–0.45)
 Nonmelanoma skin cancer0.83 (0.49–1.42)0.08 (0.01–0.82)0.74 (0.06–8.72)<5 events<5 events
Secondary analysis§     
 Any lymphoma1.25 (0.71–2.20)<5 events<5 events<5 events<5 events
 Any lymphoma or leukemia0.99 (0.60–1.61)0.41 (0.07–2.35)<5 events<5 events<5 events
 Any solid cancer0.94 (0.79–1.12)1.18 (0.66–2.09)0.88 (0.39–1.98)0.83 (0.23–3.02)0.15 (0.03–0.76)
 Nonmelanoma skin cancer1.07 (0.79–1.46)0.37 (0.13–1.07)2.65 (0.33–21.07)0.35 (0.04–3.43)0.57 (0.13–2.58)
thumbnail image

Figure 1. Kaplan-Meier curves for patients with rheumatoid arthritis (RA), comparing rates of solid cancers in the tumor necrosis factor inhibitor (TNF-I)–treated group with rates of solid cancers in the comparator-treated group. A, Results from the primary analysis. B, Results from the secondary analysis. The primary analysis examined only outcomes that occurred while patients were receiving therapy, and the secondary analysis allowed for an indefinite lag between discontinuation of therapy and the onset of cancer (see Patients and Methods). The Kaplan-Meier curves did not suggest an increasing relative risk with longer followup time.

Download figure to PowerPoint

We examined the relative hazard of the 10 most common cancers in the US in the RA patient cohort (Table 4). Patients treated with a TNFα inhibitor did not have a significantly increased risk of any of these cancers. The largest HR was observed for colorectal cancer in the primary analysis (HR 1.75 [95% confidence interval 0.65–4.68]).

Table 4. Relative hazards of the 10 most common cancers in the US among RA patients treated with TNFα inhibitors or comparator therapies*
 Primary analysisSecondary analysis
TNFα inhibitor, no. of cancersComparator, no. of cancersHR (95% CI)§TNFα inhibitor, no. of cancersComparator, no. of cancersHR (95% CI)§
  • *

    In the primary analysis, there were 6,000 nonbiologic disease-modifying antirheumatic drug (DMARD) person-years and 17,582 TNFα inhibitor person-years. In the secondary analysis, there were 19,699 nonbiologic DMARD person-years and 37,559 TNFα inhibitor person-years. Both nonbiologic DMARD person-years and TNFα inhibitor person-years were based on analysis of any solid cancer as an outcome. HR = hazard ratio; 95% CI = 95% confidence interval (see Table 1 for other definitions).

  • Followup is censored when the patient discontinues therapy.

  • Followup continues even if the patient discontinues therapy. In the secondary analysis, exposure continues until start of the alternative therapy even if the patient discontinues therapy.

  • §

    Adjusted for propensity score and average steroid dose at baseline.

Any lymphoma1380.83 (0.33–2.05)31191.25 (0.71–2.20)
Any leukemia<5 events670.57 (0.21–1.54)
Prostate<5<50.49 (0.14–1.71)11130.61 (0.31–1.22)
Female breast10110.60 (0.31–1.15)30330.90 (0.60–1.36)
Lung13140.77 (0.40–1.49)46491.03 (0.73–1.46)
Colorectal1561.75 (0.65–4.68)29251.06 (0.62–1.82)
Uterine<5 events<5<50.80 (0.29–2.23)
Kidney/renal pelvis<5 events780.77 (0.33–1.77)
Bladder<5 events9<51.30 (0.40–4.23)
Malignant melanoma<5 events<560.84 (0.32–2.23)

To assess the effect of prevalent cancers being identified as incident cancers, we computed the potential magnitude of bias that could result across a range of assumptions for the following variables: the proportion of false-positive cancer diagnoses that were prevalent cancers and the distribution of these false-positive prevalent cancers between the TNFα inhibitor–treated and comparator-treated groups. For an observed relative risk of 1.0, the true relative risk could be elevated as high as 2.0 if the proportion of falsely identified incident cancers that were prevalent cancers was 50% and 100% of these were in the comparator-treated group (Table 5).

Table 5. Results of a sensitivity analysis to assess the effect of the proportion of falsely identified incident cancers that were truly prevalent cancers and the distribution of these prevalent cancers between the tumor necrosis factor α inhibitor–treated and comparator-treated groups*
Falsely identified incident cancers that were truly prevalent cancers, %False-positive prevalent cancers that were in the comparator-treated group, %True relative risk
  • *

    The analysis assumes an observed relative risk of 1.0, a sensitivity of 76%, a positive predictive value of the algorithm of 60%, and that falsely identified incident cancers that were not prevalent cancers were evenly distributed between the medication exposure groups.

351001.61
35751.26
35671.17
35501.00
501002.00
50751.40
50671.26
50501.00
601002.33
60751.50
60671.31
60501.00

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Previous evaluations of the relative risk of cancer associated with TNFα inhibitors have come to differing conclusions, with some but not all meta-analyses of clinical trial data suggesting that TNFα inhibitors may increase the risk of cancer, particularly among RA patients. Observational studies are potentially able to overcome the sample size and short followup limitations of clinical trials but can be biased by channeling and residual confounding. Several observational studies have come to differing conclusions about the risk of cancer with TNFα inhibitor therapy (15–21). For example, within a Crohn's disease center of excellence, patients treated with TNFα inhibitors appeared to have an increased incidence of cancer (15). Several studies suggest a possible increased risk of skin cancers among RA patients treated with TNFα inhibitors, particularly when used with concomitant immunosuppressant medications (17, 21). Previous studies (17–20) did not show a higher incidence of any solid cancer or lymphoma among RA patients treated with TNFα inhibitors. In this large retrospective cohort study that included a broad spectrum of patients from multiple health plans throughout the US, we did not observe evidence of an increased risk of cancer associated with TNFα inhibitor therapy, either in an analysis limited to the period of current therapy or in an analysis that continued followup after the patient discontinued therapy.

Medications can increase the incidence of cancer by initiating the cancer process, by promoting progression of precancerous states to invasive cancer, or both. Relatively limited data are available on time since initiation of therapy and the risk of cancer. Askling et al recently reported that there was no increased risk of cancer among a cohort of RA patients newly starting TNFα inhibitor therapy in Sweden and that risk did not increase with greater time since initiation (22). In our study, we used 2 different definitions of exposure. In the primary analysis, followup was censored when the drug was discontinued, whereas the secondary analysis continued to follow up these patients in their original exposure group, regardless of changes in treatment. The 2 analyses produced similar HRs. Furthermore, the Kaplan-Meier curves did not suggest an increasing relative risk with longer followup time, although the number of patients with long duration of followup was relatively small. Of note, the absolute incidence rates in our cohort were remarkably similar to that in the cohort of Askling et al (22).

A unique aspect of the SABER study was the ability to address the same question in different disease states using the same methodology. RA patients have an increased risk of lymphoma that is associated with the underlying inflammation (1–3), while there is little if any increased risk of lymphoma among patients with inflammatory bowel disease in the absence of immunosuppression (23, 24). The HRs for any solid cancer were similar in the cohorts of RA patients and inflammatory bowel disease patients. In our analysis of any lymphoma or leukemia, there was no significantly increased risk in either population. In the inflammatory bowel disease patient cohort, the TNFα inhibitor therapy was nonsignificantly associated with a lower incidence of lymphoma or leukemia (HR 0.41), whereas in the RA patient cohort the HR was 0.99. This apparent difference may be due to the choice of comparator therapy for the analysis of inflammatory bowel disease, since thiopurines are associated with an increased incidence of lymphoma (24, 25). Similar patterns were seen for nonmelanoma skin cancer, which has also been associated with thiopurine therapy (26, 27).

Major strengths of the SABER initiative include the large sample sizes, the diverse patient population, the comparison of TNFα inhibitor therapy to the relevant alternative therapy, and the ability to examine the association of TNFα inhibitor therapy with various adverse events across several different diseases for which TNFα inhibitor therapy is indicated and widely used. Our study findings are broadly generalizable to the US population, given that the data resources included low-income Medicaid beneficiaries, those with private health insurance in Northern California, and a low-income elderly population from the Northeast. However, there are several limitations to this study that should be considered when interpreting the results.

We had limited power to assess the association among some of the disease subgroups and for rare outcomes. This may have been due in part to the timing of US Food and Drug Administration (FDA) approval of TNFα inhibitor therapy for these indications. Adalimumab, the second TNFα inhibitor to be approved by the FDA for use in Crohn's disease, received formal approval in February 2007. Likewise, infliximab and adalimumab did not receive FDA approval for use in psoriasis until 2006 and 2008, respectively.

Studies that rely on previously collected electronic data must consider the accuracy of the methods used to identify the exposures and the outcomes. For medications, we used data on filled prescriptions. For our outcome assessment, we used 2 different methods. We were able to use cancer registry information to assign outcomes for the patients from Kaiser Permanente Northern California, and we relied on a validated algorithm to identify cancer outcomes in claims data (10). We replicated the work of Setoguchi et al (10), who tested a claims data–based definition for selected cancers in Medicare beneficiaries. Our results were similar with respect to sensitivity and PPV in a health plan with computerized data, Kaiser Permanente Northern California. We extended the validation work to other solid cancers and report generally similar test characteristics across most of the cancers. Given the similar results in Medicare and Kaiser Permanente, it is likely that the algorithm performs comparably in Medicaid and TennCare data.

As is typical of observational studies using administrative and electronic health record data, the algorithm for identifying incident cancers was imperfect. Setoguchi et al previously demonstrated that the magnitude of bias toward the null association expected based on the performance characteristics of the cancer-finding algorithm is small when the misclassification is nondifferential (10). We further estimated that ∼40–70% of the false-positive incident cancers were actually prevalent cancers, but that approximately half of these were identified as prevalent and excluded from our cohort study by using the 1-year look-back period.

Therapy with TNFα inhibitors may be selectively avoided in patients with a history of cancer, and such channeling may bias observational studies against finding an association between TNFα inhibitor therapy and cancer risk. Our sensitivity analysis (further information is available from the corresponding author) suggests that our observed relative risk estimates of ∼1 could reflect true relative risk estimates of up to 2.0 due to such misclassification if one assumed that all prevalent cancers were in the comparator-treated group. However, we further explored the distribution of prevalent cancer among the false-positive cancer diagnoses in our validation study and found no difference between the exposure groups (data not shown). Thus, the magnitude of such bias in our study is likely smaller than we estimate in our sensitivity analysis. Similarly, patients with a history of cancer may be more likely to develop a second cancer, but in general this represents a small proportion of all patients with an incident cancer (28) and as such should have a much smaller effect on the observed results. Because of the imperfect nature of the algorithm and the potential for channeling TNFα inhibitors to patients without a history of cancer, the lack of association observed in this study should be interpreted cautiously. While the results suggest that a strong association between TNFα inhibitors and cancer risk is unlikely, our sensitivity analysis demonstrates that we cannot rule out small-to-modest associations. Whether such increased risk is clinically important depends on the clinical situation (29, 30).

We elected to combine multiple cancers together in our outcomes of any solid cancer or any lymphoma or leukemia. One advantage of this is that the larger number of events increases statistical precision. The composite measure also accounts for the possibility that the therapy could increase the incidence of one cancer while decreasing the incidence of another, thus helping physicians and patients to make decisions that account for each of these effects. However, a disadvantage is that the effect of specific cancer incidence rates can be obscured in analysis of the composite measure. For this reason we performed additional analyses of the 10 most common cancers in the US. While these results were generally consistent with the analysis of the composite measure, it is possible that TNFα inhibitors could increase or decrease the incidence of less common cancers. Extremely large studies would be required to assess this.

Despite combining data from 4 different sources, given that cancer is rare, the sample sizes and followup time were still relatively small for some of the outcome measures and resulted in the imprecision of the risk assessment in non–RA patient cohorts. However, for several of these non–RA disease cohorts of patients, our study is among the largest to date. Furthermore, the duration of TNFα inhibitor therapy was short and limits our ability to detect cancers that may occur later in the course of therapy; however, our use patterns were consistent with known rates of loss of response during maintenance therapy (31–35). However, the HRs measured in this study do not suggest an important increase in the risk of any of the cancers under study during the first several years of initiating treatment.

We elected to compare TNFα inhibitor–treated patients to other patients with the same underlying disease process but who were treated with other nonbiologic therapies. While this generally reflects the treatment decision facing clinicians and patients, this does not directly test the hypothesis of whether TNFα inhibitor therapies increase the risk of cancer (such as compared to no therapy or to placebo). Some of the comparator drugs that we studied have been strongly associated with selected cancers. Particularly well established is the association between thiopurine therapy and the risk of lymphoma (24, 25). Such therapies may also increase the risk of skin cancer and cervical cancer (26, 27, 36–38). To the extent that the comparator drugs increase the risk of the cancer, our new user comparative safety design may obscure a true biologic effect of TNFα inhibitors. In contrast, previous or concurrent thiopurine treatment among the TNFα inhibitor–treated cohort could potentially result in attribution of cancer risk to the TNFα inhibitor when in fact it was the thiopurine that led to the cancer. Additionally, our choice of comparators may still be subject to some unadjusted confounding, such as being related to duration of the underlying disease, despite the propensity score methods. While there is no perfect study design to disentangle this issue, our design closely reflects the treatment decisions made every day in clinical practice.

We did not address cancer-related mortality in this study, as cause of death can be difficult to determine from administrative data. Results of the association of TNFα inhibitors with all-cause mortality will be reported elsewhere.

In conclusion, we did not observe an increased incidence of cancer early in the course of treatment among patients treated with TNFα inhibitors in this large cohort study across multiple patient populations. Similarly, among RA patients, the incidence of the 10 most common cancers was not higher among patients treated with TNFα inhibitor therapy. However, the potential for channeling of comparator therapies to patients with a history of cancer or other risk factors for cancer could have biased the results in a manner that would obscure a small-to-modest association between TNFα inhibitors and cancer. Similarly, the outcomes of long-term therapy will require further study in these or other cohorts.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Haynes had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Haynes, Beukelman, Curtis, Herrinton, Graham, Solomon, Chen, Liu, Saag, Lewis.

Acquisition of data. Curtis, Herrinton, Solomon, Griffin, Liu, Saag.

Analysis and interpretation of data. Haynes, Beukelman, Curtis, Newcomb, Herrinton, Graham, Chen, Liu, Saag, Lewis.

Acknowledgements

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

We thank Hopiy Kim for assistance with the malignancy identification algorithm. On behalf of the SABER collaboration, we thank Parivash Nourjah (AHRQ); Robert Glynn, Mary Kowal, Joyce Lii, Jeremy Rassen, and Sebastian Schneeweiss (Brigham and Women's Hospital); Rita Ouellet-Hellstrom, Jane Gilbert, Carolyn McCloskey, Kristin Phucas, and James William (FDA); Leslie Harrold and Marcia Raebel (Kaiser Permanente); and Carlos Grijalva and Ed Mitchel (Vanderbilt University). We acknowledge the Tennessee Bureau of TennCare of the Department of Finance and Administration and the Tennessee Department of Health, Office of Health Statistics, which provided the TennCare data.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  • 1
    Baecklund E, Ekbom A, Sparen P, Feltelius N, Klareskog L. Disease activity and risk of lymphoma in patients with rheumatoid arthritis: nested case-control study. BMJ 1998; 317: 1801.
  • 2
    Baecklund E, Iliadou A, Askling J, Ekbom A, Backlin C, Granath F, et al. Association of chronic inflammation, not its treatment, with increased lymphoma risk in rheumatoid arthritis. Arthritis Rheum 2006; 54: 692701.
  • 3
    Zintzaras E, Voulgarelis M, Moutsopoulos HM. The risk of lymphoma development in autoimmune diseases: a meta-analysis. Arch Intern Med 2005; 165: 233744.
  • 4
    Bongartz T, Sutton AJ, Sweeting MJ, Buchan I, Matteson EL, Montori V. Anti-TNF antibody therapy in rheumatoid arthritis and the risk of serious infections and malignancies: systematic review and meta-analysis of rare harmful effects in randomized controlled trials [published erratum appears in JAMA 2006;295:2482]. JAMA 2006; 295: 227585.
  • 5
    Bongartz T, Warren FC, Mines D, Matteson EL, Abrams KR, Sutton AJ. Etanercept therapy in rheumatoid arthritis and the risk of malignancies: a systematic review and individual patient data meta-analysis of randomised controlled trials. Ann Rheum Dis 2009; 68: 117783.
  • 6
    Peyrin-Biroulet L, Deltenre P, de Suray N, Branche J, Sandborn WJ, Colombel JF. Efficacy and safety of tumor necrosis factor antagonists in Crohn's disease: meta-analysis of placebo-controlled trials. Clin Gastroenterol Hepatol 2008; 6: 64453.
  • 7
    Askling J, Fahrbach K, Nordstrom B, Ross S, Schmid CH, Symmons D. Cancer risk with tumor necrosis factor α (TNF) inhibitors: meta-analysis of randomized controlled trials of adalimumab, etanercept, and infliximab using patient level data. Pharmacoepidemiol Drug Saf 2011; 20: 11930.
  • 8
    Thompson AE, Rieder SW, Pope JE. Tumor necrosis factor therapy and the risk of serious infection and malignancy in patients with early rheumatoid arthritis: a meta-analysis of randomized controlled trials. Arthritis Rheum 2011; 63: 147985.
  • 9
    Herrinton LJ, Curtis JR, Chen L, Liu L, Delzell E, Lewis JD, et al. Study design for a comprehensive assessment of biologic safety using multiple healthcare data systems. Pharmacoepidemiol Drug Saf 2011; 20: 1199209.
  • 10
    Setoguchi S, Solomon DH, Glynn RJ, Cook EF, Levin R, Schneeweiss S. Agreement of diagnosis and its date for hematologic malignancies and solid tumors between medicare claims and cancer registry data. Cancer Causes Control 2007; 18: 5619.
  • 11
    Beukelman T, Haynes K, Curtis JR, Xie F, Chen L, Bemrich-Stolz CJ, et al, on behalf of the Safety Assessment of Biological Therapeutics Collaboration. Rates of malignancy associated with juvenile idiopathic arthritis and its treatment. Arthritis Rheum 2012; 64: 126371.
  • 12
    Warren JL, Klabunde CN, Schrag D, Bach PB, Riley GF. Overview of the SEER-Medicare data: content, research applications, and generalizability to the United States elderly population. Med Care 2002; 40: IV-3-18.
  • 13
    Glynn RJ, Schneeweiss S, Sturmer T. Indications for propensity scores and review of their use in pharmacoepidemiology. Basic Clin Pharmacol Toxicol 2006; 98: 2539.
  • 14
    Lin DY, Wei LJ. The robust inference for the Cox proportional hazards model. J Am Stat Assoc 1989; 84: 10748.
  • 15
    Lewis JD. Immortal time bias in estimates of mortality among infliximab-treated patients with Crohn's disease. Gut 2010; 59: 15867.
  • 16
    Fidder H, Schnitzler F, Ferrante M, Noman M, Katsanos K, Segaert S, et al. Long-term safety of infliximab for the treatment of inflammatory bowel disease: a single-centre cohort study. Gut 2009; 58: 5018.
  • 17
    Wolfe F, Michaud K. Biologic treatment of rheumatoid arthritis and the risk of malignancy: analyses from a large US observational study. Arthritis Rheum 2007; 56: 288695.
  • 18
    Wolfe F, Michaud K. The effect of methotrexate and anti–tumor necrosis factor therapy on the risk of lymphoma in rheumatoid arthritis in 19,562 patients during 89,710 person-years of observation. Arthritis Rheum 2007; 56: 14339.
  • 19
    Askling J, Fored CM, Baecklund E, Brandt L, Backlin C, Ekbom A, et al. Haematopoietic malignancies in rheumatoid arthritis: lymphoma risk and characteristics after exposure to tumour necrosis factor antagonists. Ann Rheum Dis 2005; 64: 141420.
  • 20
    Askling J, Fored CM, Brandt L, Baecklund E, Bertilsson L, Feltelius N, et al. Risks of solid cancers in patients with rheumatoid arthritis and after treatment with tumour necrosis factor antagonists. Ann Rheum Dis 2005; 64: 14216.
  • 21
    Chakravarty EF, Michaud K, Wolfe F. Skin cancer, rheumatoid arthritis, and tumor necrosis factor inhibitors. J Rheumatol 2005; 32: 21305.
  • 22
    Askling J, van Vollenhoven RF, Granath F, Raaschou P, Fored CM, Baecklund E, et al. Cancer risk in patients with rheumatoid arthritis treated with anti–tumor necrosis factor α therapies: does the risk change with the time since start of treatment? Arthritis Rheum 2009; 60: 31809.
  • 23
    Lewis JD, Bilker WB, Brensinger C, Deren JJ, Vaughn DJ, Strom BL. Inflammatory bowel disease is not associated with an increased risk of lymphoma. Gastroenterology 2001; 121: 10807.
  • 24
    Beaugerie L, Brousse N, Bouvier AM, Colombel JF, Lemann M, Cosnes J, et al. Lymphoproliferative disorders in patients receiving thiopurines for inflammatory bowel disease: a prospective observational cohort study. Lancet 2009; 374: 161725.
  • 25
    Kandiel A, Fraser AG, Korelitz BI, Brensinger C, Lewis JD. Increased risk of lymphoma among inflammatory bowel disease patients treated with azathioprine and 6-mercaptopurine. Gut 2005; 54: 11215.
  • 26
    Singh H, Nugent Z, Demers AA, Bernstein CN. Increased risk of nonmelanoma skin cancers among individuals with inflammatory bowel disease. Gastroenterology 2011; 141: 161220.
  • 27
    Peyrin-Biroulet L, Khosrotehrani K, Carrat F, Bouvier AM, Chevaux JB, Simon T, et al, for the CESAME Study Group. Increased risk for nonmelanoma skin cancers in patients who receive thiopurines for inflammatory bowel disease. Gastroenterology 2011; 141: 16218.e5.
  • 28
    Fraumeni JF Jr, Curtis RE, Edwards BK, Tucker MA. Introduction. In: Curtis RE, Freedman DM, Ron E, Ries LA, Hacker DG, Edwards BK, et al, editors. New malignancies among cancer survivors: SEER Cancer Registries, 1973-2000. NIH Publication No.: 05-5302. Bethesda, MD: National Cancer Institute; 2006. p. 17.
  • 29
    Lewis JD, Schwartz JS, Lichtenstein GR. Azathioprine for maintenance of remission in Crohn's disease: benefits outweigh the risk of lymphoma. Gastroenterology 2000; 118: 101824.
  • 30
    Johnson FR, Ozdemir S, Mansfield C, Hass S, Miller DW, Siegel CA, et al. Crohn's disease patients' risk-benefit preferences: serious adverse event risks versus treatment efficacy. Gastroenterology 2007; 133: 76979.
  • 31
    Agarwal SK, Maier AL, Chibnik LB, Coblyn JS, Fossel A, Lee R, et al. Pattern of infliximab utilization in rheumatoid arthritis patients at an academic medical center. Arthritis Rheum 2005; 53: 8728.
  • 32
    Buch MH, Bingham SJ, Bryer D, Emery P. Long-term infliximab treatment in rheumatoid arthritis: subsequent outcome of initial responders. Rheumatology (Oxford) 2007; 46: 11536.
  • 33
    Ogale S, Hitraya E, Henk HJ. Patterns of biologic agent utilization among patients with rheumatoid arthritis: a retrospective cohort study. BMC Musculoskelet Disord 2011; 12: 204.
  • 34
    Colombel JF, Sandborn WJ, Rutgeerts P, Enns R, Hanauer SB, Panaccione R, et al. Adalimumab for maintenance of clinical response and remission in patients with Crohn's disease: the CHARM trial. Gastroenterology 2007; 132: 5265.
  • 35
    Hanauer SB, Feagan BG, Lichtenstein GR, Mayer LF, Schreiber S, Colombel JF, et al. Maintenance infliximab for Crohn's disease: the ACCENT I randomised trial. Lancet 2002; 359: 15419.
  • 36
    Long MD, Herfarth HH, Pipkin CA, Porter CQ, Sandler RS, Kappelman MD. Increased risk for non-melanoma skin cancer in patients with inflammatory bowel disease. Clin Gastroenterol Hepatol 2010; 8: 26874.
  • 37
    Long MD, Kappelman MD, Pipkin CA. Nonmelanoma skin cancer in inflammatory bowel disease: a review. Inflamm Bowel Dis 2011; 17: 14237.
  • 38
    Hutfless S, Fireman B, Kane S, Herrinton LJ. Screening differences and risk of cervical cancer in inflammatory bowel disease. Aliment Pharmacol Ther 2008; 28: 598605.