Table 3.
Multivariable Risk of Incident HCC according to Weight Change During Overall Adulthood1 in the NHS and HPFS Cohorts
| Adulthood Weight Change 1 | Loss ≥ −5kg | Stable Weight − 5 to < 5kg | Gain 5 to <10kg | Gain 10 to <20 kg | Gain ≥20 kg | Continuous HR (95% CI) | P for Trend |
|---|---|---|---|---|---|---|---|
|
| |||||||
| Women (NHS) | |||||||
| Cases, N. | 28 | 14 | 9 | 20 | 27 | -- | -- |
| Person-years | 641,893 | 437,150 | 354,306 | 601,258 | 541,923 | -- | -- |
| MV adjusted HR‡; (95%CI) | 1.68 (0.56–4.77) | 1 (ref.) | 1.19 (0.43–2.82) | 1.35 (0.66–3.96) | 2.02 (0.90–5.78) | 1.02 (1.00–1.10) | 0.029 |
| MV adjusted HR‡ + young adult BMI* (95%CI) | 1.70 (0.59–4.90) | 1 (ref.) | 1.15 (0.42–2.72) | 1.33 (0.65–3.65) | 1.95 (0.90–5.59) | 1.02 (1.00–1.10) | 0.036 |
| Men (HPFS) | |||||||
| Cases, N. | 5 | 16 | 12 | 11 | 16 | -- | -- |
| Person-years | 98,253 | 319,404 | 234,249 | 302,190 | 145,924 | -- | -- |
| MV adjusted HR‡; (95%CI) | 2.94 (0.65–7.29) | 1 (ref.) | 1.72 (0.38–7.78) | 2.71 (0.83–11.60) | 4.77 (1.12–12.09) | 1.08 (1.04–1.13) | 0.001 |
| MV adjusted HR‡ + young adult BMI* (95%CI) | 3.04 (0.67–7.72) | 1 (ref.) | 1.92 (0.42–8.76) | 2.86 (0.88–12.29) | 4.30 (1.07–11.12) | 1.06 (1.02–1.12) | 0.004 |
| Combined cohort | |||||||
| Cases, N. | 33 | 30 | 21 | 31 | 43 | -- | -- |
| Person-years | 740,146 | 756,554 | 588,555 | 903,448 | 687,847 | -- | -- |
| MV adjusted HR‡; (95%CI) | 1.81 (0.86–3.02) | 1 (ref.) | 1.38 (0.66–2.11) | 1.74 (1.02–2.62) | 3.04 (1.61–5.75) | 1.03 (1.01–1.08) | 0.010 |
| MV adjusted HR‡ + young adult BMI* (95%CI) | 1.83 (0.86–3.08) | 1 (ref.) | 1.40 (0.67–2.16) | 2.09 (1.11–3.95) | 2.61 (1.42–5.22) | 1.03 (1.01–1.09) | 0.010 |
Abbreviations: NHS, Nurses’ Health Study; HPFS, Health Professionals Follow-up Study; No., number; kg, kilogram; HR, hazard ratio; CI, confidence interval; MET, metabolic equivalent task; MV, multivariable model; BMI, body mass index
Overall adulthood weight change was defined as the difference in weight in kilograms between early adulthood (i.e. age 18 years [women] or age 21 years [men]) and present, using biennially updated data. For details, see Methods
Young adult BMI was defined as BMI at age 18 years (women) or 21 years (men).
Multivariable Cox proportional hazard regression models were stratified by age (years), and year of questionnaire return, with further adjustment for race (white vs. non-white), alcohol intake (0–4.9, 5–14.9, ≥15 g/day), smoking status (current vs. prior vs. never), type 2 diabetes (yes vs. no), hypertension (yes vs. no), dyslipidemia (yes vs. no), regular aspirin use (≥2 tablets per week vs. no), adherence to a healthy diet, defined by the alternative healthy eating index 2010 (AHEI) without alcohol, and physical activity, in continuous metabolic equivalent task (MET)-hours per week. Additional models further adjusted for young-adult BMI. All covariates (with the exception of race/ethnicity and young-adult BMI) were updated over each questionnaire cycle and modeled as time-varying covariates. For the combined cohort, results were meta-analyzed.