Skip to main content
. 2018 Oct 4;4(12):1683–1690. doi: 10.1001/jamaoncol.2018.4154

Table 2. Cumulative Mean Aspirin Dose and Risk of Hepatocellular Carcinoma in 13 371 Women (1980-2012) and Men (1986-2012) in the Pooled NHS and HPFS Cohorts.

Characteristic No. of Weekly Aspirin Tablets (Cumulative Mean) P Value for Trenda
<0.5 0.5 to <1.5 1.5 to <5 ≥5
Cases per person-years 46/1 425 726 23/867 338 24/1 207 350 15/731 776
Model 1, HR (95% CI)b 1 [Reference] 0.83 (0.49-1.40) 0.50 (0.30-0.84) 0.55 (0.30-1.01) .008
Model 2, HR (95% CI)c 1 [Reference] 0.86 (0.50-1.46) 0.48 (0.28-0.82) 0.48 (0.26-0.89) .003
Model 3, HR (95% CI)d 1 [Reference] 0.87 (0.51-1.48) 0.51 (0.30-0.86) 0.49 (0.28-0.96) .006
Model 4, HR (95% CI)e 1 [Reference] 1.20 (0.65-2.20) 0.56 (0.32-0.99) 0.37 (0.17-0.81) .005

Abbreviations: HR, hazard ratio; HPFS, Health Professionals Follow-up Study; NHS, Nurses’ Health Study.

a

Calculated using cumulative mean aspirin dose (continuous), among regular aspirin users.

b

Model 1 was conditioned on age (continuous years), year of questionnaire return, and sex (ie, cohort).

c

Model 2 includes model 1 covariates plus race/ethnicity (white vs nonwhite), body mass index (continuous measure), alcohol intake (0-4.9, 5.0-14.9, ≥15.0 g/d), smoking status (current vs prior vs never), physical activity (<3.0, 3.0-8.9, ≥9 metabolic equivalent task–hours per week), diabetes (yes vs no), hypertension (yes vs no), dyslipidemia (yes vs no), regular multivitamin use (≥2 multivitamin tablets per week vs no), regular use of oral antidiabetic medications (yes vs no), and regular use of statins (yes vs no). All relevant covariates were updated over time.

d

Model 3 includes model 2 regular use of nonaspirin nonsteroidal anti-inflammatory drugs (≥2 nonaspirin nonsteroidal anti-inflammatory drug tablets per week vs no), assessed as a time-varying covariate.

e

Model 4 includes model 3 covariates plus duration of regular aspirin use (continuous, years).