Table 4:
Risk factors associated with chronic kidney disease, by study site
| Agincourt (n=126l) | Dikgale (n=844) | Nairobi (n=1356) | Nanoro (n=l850) | Navrongo (n=l630) | Soweto (n=825)* | |
|---|---|---|---|---|---|---|
| Male sex | 0·81 (0·60–1·08) | 0·61 (0·38–0·93)† | 0·71 (0·52–0·97)† | 1·04 (0·74–1·46) | 1·09 (0·79–1·52) | ‥ |
| Age | 1·04 (1·01–1·06)‡ | 1·04 (1·01–1·08)‡ | 1·04 (1·01–1·07)‡ | 1·04 (1·01–1·07)‡ | 1·04 (1·01–1·07)† | 1·04 (1·01–1·07)† |
| Body-mass index | 1·01 (0·99–1·04) | 1·00 (0·97–1·03) | 1·00 (0·97–1·03) | 0·98 (0·92–1·03) | 0·99 (0·95–1·04) | 1·03 (1·00–1·06)† |
| Diabetes | 1·86 (1·16–2·84)‡ | 2·28 (1·39–3·58)§ | 2·29 (1·46–3·45)§ | 2·99 (1·59–5·17)§ | 1·95 (0·59–4·68) | 2·62 (1·51–4·31)§ |
| Highest level of education | 1·01 (0·85–1·19) | 1·11 (0·85–1·46) | 0·96 (0·75–1·24) | 0·99 (0·69–1·35) | 1·10 (0·87–1·36) | 0·75 (0·54–1·06) |
| HIV positive | 1·41 (1·06–1·86)† | 1·35 (0·88–2·01) | 2·39 (1·68–3·33)¶ | ‥|| | ‥|| | 1·25 (0·79–1·90) |
| Hypertension | 1·44 (1·06–1·97)† | 1·77 (1·19–2·65)‡ | 2·31 (1·68–3·16)¶ | 2·10 (1·43–3·03)§ | 2·30 (1·64–3·20)¶ | 2·62 (1·67–4·22) |
| Socioeconomic status | 1·00 (0·90–1·10) | 1·05 (0·92–1·19) | 0·86 (0·76–0·96)‡ | 1·10 (0·97–1·25) | 1·04 (0·92–1·17) | 0·99 (0·83–1·18) |
| Current smoker | 0·79 (0·44–1·38) | 0·33 (0·15–0·71)‡ | 1·46 (0·89–2·32) | 2·13 (1·21–3·59)‡ | 1·37 (0·87–2·14) | 1·03 (0·70–1·52) |
| Current alcohol consumption | 1·37 (0·92–2·00) | 1·02 (0·61–1·66) | 1·62 (1·09–2·36)† | 1·12 (0·79–1·63) | 1·01 (0·71–1·44) | ‥ |
| History of cardiovascular disease | 0·76 (0·32–1·49) | 1·03 (0·46–1·98) | 1·29 (0·61–2·38) | 1·31 (0·32–3·48) | 1·35 (0·48–2·96) | ‥ |
Data are relative risk, with 95% CIs in parentheses. Relative risk for various effectors, with cofactors defined with directed acyclic graphs and six-step algorithms shown in the appendix (pp 13, 15). p values are derived from generalised linear model comparisons for each risk factor, for categorical variables this comparison was to the appropriate reference group—eg, diabetic vs non-diabetic—whereas for continous variables, such as age and body-mass index, an increase in risk is donoted by a 1 unit change in the variable.
Soweto participants did not have sufficient data on history of cardiovascular disease and alcohol consumption.
p≤0·05.
p≤0·01.
p≤0·001.
p≤0·0001.
Because HIV prevalence is less than 1% in Ghana and Burkina Faso, participants who had not been tested previously or had tested negative, were considered uninfected, and not offered further testing; therefore no data were available.