Table 2.
Weighted baseline characteristics for cholesterol-related measures
| Cholesterol-lowering medication | LDL cholesterol (mmol/L) | HDL cholesterol (mmol/L) | Non-HDL cholesterol (mmol/L) |
Non-HDL cholesterol categories |
|||||
|---|---|---|---|---|---|---|---|---|---|
| <2·6 mmol/L | 2·6 to <3·7 mmol/L | 3·7 to <4·8 mmol/L | 4·8 to <5·7 mmol/L | ≥5·7 mmol/L | |||||
| All (n=398 846) | 12 311 (4·9%) | 3·6 (2·9–4·3) | 1·3 (1·1–1·6) | 4·3 (3·5–5·2) | 16 572 (5·1%) | 85 811 (26·2%) | 109 270 (33·3%) | 67 099 (20·4%) | 49 575 (15·1%) |
| Numbers available | 247 733 (59·7%) | 264 999 (60·1%) | 328 517 (74·4%) | 328 327 (74·3%) | 328 327 (74·3%) | 328 327 (74·3%) | 328 327 (74·3%) | 328 327 (74·3%) | 328 327 (74·3%) |
| Derivation (n=199 415) | 6081 (4·9%) | 3·6 (2·9–4·3) | 1·3 (1·1–1·6) | 4·3 (3·5–5·2) | 8194 (5·1%) | 42 941 (26·2%) | 54 771 (33·4%) | 33 463 (20·3%) | 24 804 (15·1%) |
| Numbers available | 123 904 (60·3%) | 132 424 (60·5%) | 164 276 (75·0%) | 164 173 (74·9%) | 164 173 (74·9%) | 164 173 (74·9%) | 164 173 (74·9%) | 164 173 (74·9%) | 164 173 (74·9%) |
| Validation (n=199 431) | 6230 (5·0%) | 3·6 (2·9–4·3) | 1·3 (1·1–1·6) | 4·3 (3·5–5·2) | 8378 (5·1%) | 42 870 (26·2%) | 54 499 (33·2%) | 33 636 (20·4%) | 24 771 (15·0%) |
| Numbers available | 123 829 (59·0%) | 132 575 (59·6%) | 164 241 (73·8%) | 164 154 (73·7%) | 164 154 (73·7%) | 164 154 (73·7%) | 164 154 (73·7%) | 164 154 (73·7%) | 164 154 (73·7%) |
| p value | 0·18 | 0·79 | 0·01 | 0·79 | 0·85 | 0·99 | 0·58 | 0·29 | 0·64 |
Data are n (%) for numbers available, median (IQR) for continuous variables, n (%) of participants or % (95% CI) for categorical variables. p values are given for the validation cohort versus the derivation cohort using weighted versions of the Mann-Whitney and χ2 test. Due to the presence of the Estonian case-cohort dataset, the summary estimates (% and median (IQR)) are weighted by the inverse of the inclusion probability for individuals in that cohort. The individuals from other cohorts are given weight 1 in the computation. Totals (n) are not weighted. From the Estonian data only, the sub-cohort is used in the table computations.