Table 3.

Sensitivity analyses for association between statin use and cancer-specific mortality in lung cancer patients

User vs. nonuser1-11 prescriptions vs. none≥ 12 prescriptions vs. none
Cancer-specific mortalityAll patientsPerson yearsAdjusted HRa (95% CI)PAdjusted HRa (95% CI)PAdjusted HRa (95% CI)P
Main analysis: Statin user versus nonuser after diagnosis1,7352,7925,2270.89 (0.78–1.02)0.090.94 (0.81–1.09)0.390.81 (0.67–0.98)0.03
Subgroup analyses: Statin user versus nonuser after diagnosis, restricted to:
 Males1,0071,6202,9280.94 (0.79–1.12)0.620.99 (0.82–1.21)0.960.85 (0.66–1.09)0.19
 Females7281,1722,2990.83 (0.67–1.03)0.090.86 (0.68–1.10)0.230.78 (0.58–1.06)0.11
 Prediagnosis statin nonusersb1,2221,9123,6910.78 (0.60–1.02)0.070.64 (0.45–0.91)0.011.06 (0.72–1.57)0.77
 Prediagnosis statin usersb4026971,1320.76 (0.54–1.07)0.120.83 (0.59–1.17)0.290.63 (0.42–0.93)0.02
 Small-cell lung cancer2683384890.66 (0.44–0.99)0.050.64 (0.41–0.99)0.050.74 (0.41–1.36)0.33
 Non–small cell lung cancer1,1631,9954,0900.93 (0.79–1.09)0.381.01 (0.84–1.21)0.900.80 (0.64–1.01)0.06
Sensitivity analyses: Statin user versus nonuser after diagnosis
 Increasing lag to 1 year1,7352,7925,2270.92 (0.80–1.05)0.230.92 (0.80–1.07)0.300.90 (0.72–1.14)0.39
 Increasing lag to 1 year (simvastatin only)1,7352,7925,2270.86 (0.74–1.00)0.060.84 (0.71–1.00)0.050.91 (0.70–1.19)0.50
 Additionally adjusting for smoking before diagnosis1,5842,5644,7600.90 (0.78–1.03)0.120.95 (0.82–1.11)0.520.82 (0.66–0.97)0.03
 Additionally adjusting for BMI before diagnosis1,3402,2064,1540.93 (0.80–1.08)0.340.98 (0.83–1.15)0.800.85 (0.68–1.04)0.12
 Additionally adjusting for small cell/non–small cell1,4312,3334,5790.86 (0.75–1.00)0.050.92 (0.78–1.09)0.320.77 (0.62–0.95)0.02
 Additionally adjusting for bisphosphonate usec1,7352,7925,2270.89 (0.78–1.02)0.090.94 (0.81–1.09)0.400.81 (0.67–0.98)0.03
 Based upon first year after diagnosisd1,7352,7925,2270.90 (0.78–1.03)0.130.97 (0.83–1.12)0.660.75 (0.60–0.93)0.01
 Nested case–control analysise1,7050.92 (0.79–1.07)0.270.98 (0.82–1.15)0.770.81 (0.65–1.01)0.06
  • aExcept where otherwise stated, all analyses of post-diagnostic statin use adjusted for year of diagnosis, age at diagnosis, gender, surgery within 6 months of diagnosis, radiotherapy within 6 months, chemotherapy within 6 months, comorbidities (before diagnosis, including cerebrovascular disease, chronic pulmonary disease, congestive heart disease, diabetes, myocardial infarction, peptic ulcer disease, peripheral vascular disease, and renal disease), other medication use (after diagnosis, as time varying covariates, specifically low-dose aspirin and beta-blockers), and deprivation (in fifths).

  • bBased upon statin use in the year before diagnosis, restricted to individuals with 1 year of records before lung cancer diagnosis.

  • cContains all confounders in a along with bisphosphonate use after diagnosis (as a time-varying covariate).

  • dSimplified analysis, not requiring time-varying covariate use, comparing statin users with statin nonusers in the first year after diagnosis in individuals living more than 1 year after cancer diagnosis, adjusted for all confounders in a but other medication use also restricted to first year after diagnosis.

  • eEstimate and 95% CIs are for adjusted OR, 25% (433/1,705) of cancer-specific deaths used statins compared with 29% (2,388/8,150) of risk-set controls (not dying from cancer), matched on age at diagnosis, year of diagnosis, gender, and adjusted for all other confounders in a.