RESUMO
OBJECTIVE: Various approaches have been used in the literature to handle missing vital status data in cancer registries. We aimed to compare these approaches to determine which led to the least biased estimates in typical analytic tasks of cancer registries. METHODS: A simulation study was performed using data from the Swiss National Agency for Cancer Registration for six tumor types. First, 5%, 10% and 15% missingness in the vital status were introduced artificially in the complete data. Second, missing vital status data were handled by applying no, single or multiple imputations. Five-year overall survival estimates, relative survival or standardized incidence ratio were computed. Estimates were compared with the true value. RESULTS: Standardized incidence ratio estimates for colorectal cancer obtained with multiple imputation yielded least biased results (-0.06 to -0.04), but the widest confidence intervals. Single imputation was more biased (-0.32) than using no imputation at all (-0.21). A similar pattern was observed for overall survival and relative survival. CONCLUSION: This simulation study indicated that often used single imputation (sometimes referred to as simulating follow-up times) techniques to fill in missing vital status data are likely too biased to be useful in practice. Multiple imputation approaches yielded standardized incidence ratio, overall and relative survival estimates with the least bias, indicating reasonable performance that is likely to generalize to other settings.
RESUMO
BACKGROUND: Reference intervals of thyroid-stimulating hormone (TSH) and free thyroxine (FT4) are statistically defined by the 2·5-97·5th percentiles, without accounting for potential risk of clinical outcomes. We aimed to define the optimal healthy ranges of TSH and FT4 based on the risk of cardiovascular disease and mortality. METHODS: This systematic review and individual participant data (IPD) meta-analysis identified eligible prospective cohorts through the Thyroid Studies Collaboration, supplemented with a systematic search via Embase, MEDLINE (Ovid), Web of science, the Cochrane Central Register of Controlled Trials, and Google Scholar from Jan 1, 2011, to Feb 12, 2017 with an updated search to Oct 13, 2022 (cohorts found in the second search were not included in the IPD). We included cohorts that collected TSH or FT4, and cardiovascular outcomes or mortality for adults (aged ≥18 years). We excluded cohorts that included solely pregnant women, individuals with overt thyroid diseases, and individuals with cardiovascular disease. We contacted the study investigators of eligible cohorts to provide IPD on demographics, TSH, FT4, thyroid peroxidase antibodies, history of cardiovascular disease and risk factors, medication use, cardiovascular disease events, cardiovascular disease mortality, and all-cause mortality. The primary outcome was a composite outcome including cardiovascular disease events (coronary heart disease, stroke, and heart failure) and all-cause mortality. Secondary outcomes were the separate assessment of cardiovascular disease events, all-cause mortality, and cardiovascular disease mortality. We performed one-step (cohort-stratified Cox models) and two-step (random-effects models) meta-analyses adjusting for age, sex, smoking, systolic blood pressure, diabetes, and total cholesterol. The study was registered with PROSPERO, CRD42017057576. FINDINGS: We identified 3935 studies, of which 53 cohorts fulfilled the inclusion criteria and 26 cohorts agreed to participate. We included IPD on 134 346 participants with a median age of 59 years (range 18-106) at baseline. There was a J-shaped association of FT4 with the composite outcome and secondary outcomes, with the 20th (median 13·5 pmol/L [IQR 11·2-13·9]) to 40th percentiles (median 14·8 pmol/L [12·3-15·0]) conveying the lowest risk. Compared with the 20-40th percentiles, the age-adjusted and sex-adjusted hazard ratio (HR) for FT4 in the 80-100th percentiles was 1·20 (95% CI 1·11-1·31) for the composite outcome, 1·34 (1·20-1·49) for all-cause mortality, 1·57 (1·31-1·89) for cardiovascular disease mortality, and 1·22 (1·11-1·33) for cardiovascular disease events. In individuals aged 70 years and older, the 10-year absolute risk of composite outcome increased over 5% for women with FT4 greater than the 85th percentile (median 17·6 pmol/L [IQR 15·0-18·3]), and men with FT4 greater than the 75th percentile (16·7 pmol/L [14·0-17·4]). Non-linear associations were identified for TSH, with the 60th (median 1·90 mIU/L [IQR 1·68-2·25]) to 80th percentiles (2·90 mIU/L [2·41-3·32]) associated with the lowest risk of cardiovascular disease and mortality. Compared with the 60-80th percentiles, the age-adjusted and sex-adjusted HR of TSH in the 0-20th percentiles was 1·07 (95% CI 1·02-1·12) for the composite outcome, 1·09 (1·05-1·14) for all-cause mortality, and 1·07 (0·99-1·16) for cardiovascular disease mortality. INTERPRETATION: There was a J-shaped association of FT4 with cardiovascular disease and mortality. Low concentrations of TSH were associated with a higher risk of all-cause mortality and cardiovascular disease mortality. The 20-40th percentiles of FT4 and the 60-80th percentiles of TSH could represent the optimal healthy ranges of thyroid function based on the risk of cardiovascular disease and mortality, with more than 5% increase of 10-year composite risk identified for FT4 greater than the 85th percentile in women and men older than 70 years. We propose a feasible approach to establish the optimal healthy ranges of thyroid function, allowing for better identification of individuals with a higher risk of thyroid-related outcomes. FUNDING: None.