Table 4.
Adjusted hazard ratios for discontinuation and restarting statin treatment in secondary prevention group
| Adjusted hazard ratio (95% CI) | ||
|---|---|---|
| Discontinuation* | Restarting† | |
| Sex | ||
| Men | Reference | Reference |
| Women | 1.14 (1.12 to 1.17) | 0.93 (0.91 to 0.95) |
| Ethnicity | ||
| White or not recorded | Reference | Reference |
| Indian | 1.35 (1.23 to 1.49) | 1.40 (1.25 to 1.56) |
| Pakistani | 1.41 (1.19 to 1.66) | 1.13 (0.97 to 1.31) |
| Bangladeshi | 1.17 (0.84 to 1.61) | 1.58 (1.26 to 1.98) |
| Other Asian | 1.52 (1.30 to 1.79) | 1.41 (1.17 to 1.71) |
| Black African | 1.69 (1.33 to 2.16) | 1.15 (0.90 to 1.48) |
| Black Caribbean | 1.80 (1.54 to 2.10) | 1.35 (1.16 to 1.58) |
| Chinese | 1.17 (0.87 to 1.57) | 1.22 (0.94 to 1.58) |
| Other ethnic group | 1.23 (1.10 to 1.37) | 1.13 (1.00 to 1.29) |
| Smoking status | ||
| Non-smoker | Reference | Reference |
| Ex-smoker | 0.96 (0.94 to 0.98) | 1.04 (1.02 to 1.06) |
| Light smoker | 1.14 (1.09 to 1.19) | 1.11 (1.06 to 1.16) |
| Moderate smoker | 1.20 (1.16 to 1.24) | 1.04 (1.01 to 1.08) |
| Heavy smoker | 1.27 (1.21 to 1.33) | 1.09 (1.04 to 1.14) |
| Chronic conditions | ||
| Rheumatoid arthritis | 1.02 (0.96 to 1.08) | 1.03 (0.97 to 1.10) |
| Chronic renal disease | 1.08 (0.98 to 1.18) | 0.90 (0.81 to 0.99) |
| Liver disease | 1.22 (1.09 to 1.36) | 0.93 (0.82 to 1.07) |
| Chronic obstructive pulmonary disease | 1.22 (1.18 to 1.27) | 1.03 (1.00 to 1.08) |
| Cancer | 1.05 (1.03 to 1.08) | 0.96 (0.93 to 0.99) |
| Atrial fibrillation | 1.05 (1.01 to 1.09) | 0.94 (0.90 to 0.98) |
| Hypertension | 0.94 (0.92 to 0.96) | 1.01 (0.99 to 1.03) |
| Heart failure | 0.97 (0.94 to 1.01) | 0.97 (0.93 to 1.01) |
| Type 1 diabetes | 1.17 (1.08 to 1.26) | 1.02 (0.93 to 1.10) |
| Type 2 diabetes | 0.97 (0.95 to 1.00) | 1.06 (1.02 to 1.09) |
| Dementia, short term statin use‡ | 0.71 (0.61 to 0.82) | — |
| Dementia, long term statin use‡ | 1.22 (1.08 to 1.39) | — |
| Dementia in previous statin use | — | 0.75 (0.65 to 0.86) |
| Genetic characteristics | ||
| Family history of premature coronary heart disease | 1.00 (0.96 to 1.04) | 1.02 (0.97 to 1.06) |
| Familial hypercholesterolaemia | 1.74 (1.24 to 2.45) | 0.83 (0.54 to 1.29) |
| Use of other treatments | ||
| Aspirin | 0.83 (0.82 to 0.85) | 1.05 (1.02 to 1.07) |
| Anticoagulants | 0.89 (0.86 to 0.93) | 0.97 (0.93 to 1.02) |
| Each of other non-cardiovascular treatment | 1.03 (1.03 to 1.04) | 1.00 (0.99 to 1.00) |
| Townsend score groups (patient based)§ | ||
| 1 (most affluent) | Reference | Reference |
| 2 | 1.02 (0.98 to 1.06) | 0.99 (0.94 to 1.03) |
| 3 | 0.99 (0.95 to 1.03) | 1.00 (0.95 to 1.05) |
| 4 | 1.02 (0.97 to 1.06) | 1.05 (1.00 to 1.10) |
| 5 (most deprived) | 0.97 (0.92 to 1.01) | 1.06 (1.00 to 1.12) |
| Statin use at baseline | ||
| Simvastatin | Reference | Reference |
| Atorvastatin | 1.00 (0.97 to 1.03) | 0.91 (0.88 to 0.93) |
| Pravastatin | 0.94 (0.90 to 0.98) | 0.92 (0.87 to 0.97) |
| Rosuvastatin | 1.19 (1.11 to 1.28) | 0.77 (0.72 to 0.82) |
| Fluvastatin | 1.09 (0.99 to 1.22) | 0.92 (0.81 to 1.04) |
| Statin dose at baseline (potency units) | ||
| Less than 1 | 1.11 (1.08 to 1.14) | 0.88 (0.85 to 0.91) |
| One | Reference | Reference |
| Two | 0.86 (0.84 to 0.88) | 1.12 (1.10 to 1.15) |
| Three and more | 0.60 (0.57 to 0.64) | 1.37 (1.31 to 1.42) |
All adjusted models include sex, ethnicity, smoking status, chronic conditions, use of other drugs, practice based Townsend score groups, and year of entering the study.
*Model also includes type and dose of statin at baseline, fractional polynomials for age ((age/5) and (age/5)2), body mass index (BMI; ln(BMI), BMI0.5), total cholesterol:high density lipoprotein ratio (ratio0.5), and systolic blood pressure (SBP; (SBP/10)−0.5).
†Model also includes type and dose of statin at discontinuation, fractional polynomials for age ((age/5)−1, (age/5)3), body mass index (BMI), total cholesterol:high density lipoprotein ratio (ratio−0.5), and systolic blood pressure ((SBP/10)). Practice ID was included as clustering variable.
‡The analysis was stratified by dementia. Results for discontinuation were taken from additional analyses by terms of use (supplementary table 3).
For all models, Practice ID was included as clustering variable
§Estimates of Townsend score groups (patient based) are from the additional analysis (supplementary table 4).