Liver enzymes are not directly involved in atrial fibrillation: a prospective cohort study.

Abstract

Epidemiological evidence proposes the direct involvement of the liver enzymes in atrial fibrillation. These relationships are controversial and mechanistically unclear. As part of the British Regional Heart Study, we investigated whether change in liver enzymes over time associates with atrial fibrillation in men initially free of this heart condition. We prospectively investigated change (delta) in liver enzymes and new-onset atrial fibrillation in a representative sample of 1428 men aged 60-79years. Mean follow-up was 12·3years, after which 108 new atrial fibrillation cases were identified. The liver enzymes did not differ at baseline or follow-up, except for gamma-glutamyl transferase which was higher at follow-up in men who developed atrial fibrillation compared to those who did not (P<0·0001). Change in GGT was greater in men who developed AF than those who did not (+6·12 vs. -2·60U/L, P=0·036). N-terminal pro-brain natriuretic peptide (baseline and follow-up, P<0·0001) and total bilirubin (follow-up only, P<0·0001) were also higher in men who developed atrial fibrillation while serum haemoglobin was similar at baseline and follow-up (P≥0·74). Atrial fibrillation was associated with change in gamma-glutamyl transferase (OR, 1·18; 95% CI, 1·01-1·37) after multiple adjustments and exclusions. However, after adjusting for baseline (P=0·088) or change (P=0·40) in N-terminal pro-brain natriuretic peptide, the association between atrial fibrillation and change in gamma-glutamyl transferase was lost. The direct relationship between atrial fibrillation and liver enzymes is absent and depends, at least in part, on the progression of heart failure as captured by N-terminal pro-brain natriuretic peptide.