Table 2.
Association of C-reactive protein levels with probable diagnosis of depression in the UK Biobank cohort.
| log CRP as continuous variable | CRP Q1 (n = 34,372) | CRP Q2 (n = 31,704) | CRP Q3 (n = 28,714) | CRP Q4 (n = 26,350) | CRP Q5 (n = 23,750) | Per-Q effect | P-value for trend | |
|---|---|---|---|---|---|---|---|---|
| All participants (cases = 8888; controls = 145,468) | ||||||||
| Model 1 (n = 144,890) | 1·27 (1·24–1·29) | 1 [reference] | 1·11 (1·03–1·19) | 1·19 (1·10–1·28) | 1·44 (1·34–1·54) | 2·05 (1·91–2·20) | 1·19 (1·17–1·21) | <0·001 |
| Model 2 (n = 144,600) | 1·12 (1·09–1·15) | 1 [reference] | 1·08 (1·00–1·16) | 1·10 (1·02–1·18) | 1·22 (1·13–1·31) | 1·41 (1·31–1·53) | 1·09 (1·07–1·10) | <0·001 |
| Model 3 (n = 138,766) | 1·09 (1·06–1·11) | 1 [reference] | 1·07 (0·99–1·15) | 1·08 (1·00–1·17) | 1·16 (1·07–1·26) | 1·28 (1·18–1·39) | 1·06 (1·04–1·08) | <0·001 |
| Model 4 (n = 138,765) | 1·09 (1·06–1·11) | 1 [reference] | 1·07 (0·99–1·16) | 1·08 (1·00–1·17) | 1·16 (1·07–1·26) | 1·29 (1·18–1·40) | 1·06 (1·04–1·08) | <0·001 |
| Women (cases = 5641; controls = 81,562) | ||||||||
| Model 1 (n = 81,610) | 1·28 (1·25–1·32) | 1 [reference] | 1·06 (0·96–1·16) | 1·22 (1·12–1·34) | 1·40 (1·28–1·53) | 2·11 (1·94–2·29) | 1·20 (1·18–1·23) | <0·001 |
| Model 2 (n = 81,454) | 1·12 (1·08–1·15) | 1 [reference] | 1·03 (0·94–1·13) | 1·13 (1·03–1·24) | 1·18 (1·07–1·30) | 1·41 (1·27–1·55) | 1·09 (1·06–1·11) | <0·001 |
| Model 3 (n = 77,818) | 1·10 (1·06–1·14) | 1 [reference] | 1·02 (0·93–1·13) | 1·13 (1·02–1·25) | 1·17 (1·06–1·30) | 1·33 (1·20–1·48) | 1·07 (1·05–1·10) | <0·001 |
| Model 4 (n = 77,818) | 1·10 (1·06–1·13) | 1 [reference] | 1·02 (0·93–1·13) | 1·13 (1·02–1·25) | 1·17 (1·06–1·30) | 1·33 (1·20–1·48) | 1·07 (1·05–1·10) | <0·001 |
| Men (cases = 3247; controls = 63,906) | ||||||||
| Model 1 (n = 63,280) | 1·22 (1·18–1·27) | 1 [reference] | 1·23 (1·09–1·38) | 1·17 (1·04–1·32) | 1·53 (1·36–1·72) | 1·87 (1·66–2·11) | 1·16 (1·13–1·19) | <0·001 |
| Model 2 (n = 63,146) | 1·13(1·08–1·17) | 1 [reference] | 1·16 (1·03–1·30) | 1·04 (0·92–1·18) | 1·27 (1·12–1·43) | 1·44 (1·27–1·64) | 1·08 (1·05–1·12) | <0·001 |
| Model 3 (n = 60,948) | 1·07 (1·02–1·11) | 1 [reference] | 1·12 (0·99–1·27) | 1·00 (0·88–1·14) | 1·14 (1·00–1·29) | 1·21 (1·06–1·39) | 1·04 (1·01–1·07) | 0·02 |
| Model 4 (n = 60,947) | 1·07 (1·03–1·12) | 1 [reference] | 1·13 (1·00–1·28) | 1·01 (0·89–1·15) | 1·15 (1·01–1·31) | 1·23 (1·07–1·41) | 1·04 (1·01–1·07) | 0·01 |
Data show OR and 95% CIs unless otherwise indicated. P for trend is from regression models with quintiles. Model 1, unadjusted; model 2, adjusted for age, sex, and BMI (body mass index); model 3, model 2 additionally adjusted for smoking, alcohol, physical activity, ethnicity, and TDI (Townsend deprivation index at recruitment); model 4, model 3 additionally adjusted for diabetes and cardiovascular disease; *: CRP concentration was log transformed; Median CRP level was 1·15 mg/L (range 0·08–78·22 mg/L).