Self-reported alcohol intake is a better estimate of 5-year change in blood pressure than biochemical markers in low resource settings

Abstract

Despite criticism of self-reported alcohol intake, it is a valuable tool to screen for alcohol abuse as a risk factor for cardiovascular disease. We aimed to compare various self-reported estimates of alcohol use with γ-glutamyltransferase (GGT) and percentage carbohydrate deficient transferrin (%CDT) considering their relationship with blood pressure changes (%BP) over a 5-year period in black South Africans. We recruited 1994 participants and collected 5-year followed up data (N = 1246). Participants completed questionnaires on alcohol intake indicating their former and current alcohol use ('yes' response and 'no' if alcohol was never used). We assessed alcohol intake (in g) using a quantified food frequency questionnaire. We collected blood samples and measured GGT and %CDT. Brachial BP (bBP) was measured at baseline and follow-up and central BP (cBP) at follow-up only. Self-reported alcohol intake was significantly associated with the 5-year change in bBP before and after adjusting for confounders (%bSBP: R(2) = 0.263, β = 0.06, P = 0.023; %bDBP: R(2) = 0.326, β = 0.08 P = 0.005), as well as cSBP (R (2)= 0.286, β = 0.09, P = 0.010) and central pulse pressure (R(2) = 0.254, β = 0.06, P = 0.020). GGT and %CDT correlated well with self-reported alcohol intake (r = 0.44; P = 0.001; r = 0.34 P = 0.001), but did not associate significantly with %bBP or cBP at follow-up. Self-reported alcohol use was strongly associated with a 5-year increase in BP in Africans with a low socio-economic status. This was not found for biochemical measures, GGT and %CDT. Self-reported alcohol intake could be an important measure to implement in primary healthcare settings in middle to low income countries, where honest reporting is expected.