1. The associations between the renal clearance of ingested lithium (used as a marker of renal proximal tubular sodium handling) and a number of biological variables have been investigated in an unselected sample of 592 healthy and untreated white men (aged 21–68 years) under their usual living conditions and drawn from a population at work.
2. Renal excretion of lithium was expressed both as clearance of lithium and as fractional excretion of lithium, i.e. normalized for unit of glomerular filtrate.
3. Clearance of lithium was positively associated with a number of anthropometric variables such as weight (r = 0.215; P <0.001), height (r = 0.212; P <0.001), body mass index (r = 0.122; P <0.01) and body surface area (r = 0.244; P <0.001). However, when expressed as fractional excretion of lithium many of these associations were lost or tended to be negative (weight, r = −0.107, P <0.01; body mass index, r = −0.119, P <0.01), greatly depending on the strong relationship between body size and/or mass and glomerular filtration rate. The associations between clearance of lithium and other renal variables reflected, in part, a spurious association mediated by the common relationship with glomerular filtration rate. However, when expressed as fractional excretion of lithium, only some associations were apparent (fractional excretion of sodium, r = 0.256, P <0.001; fractional excretion of uric acid, r = 0.336; P <0.001), probably indicating some plausible biological relationships. These results were confirmed by analysis of co-variance by quintiles of clearance of lithium and fractional excretion of lithium.
4. The results of the present study show that the renal excretion of ingested lithium (an approximate index of proximal tubular sodium handling in man) could be used in an epidemiological setting. However, the use of the fractional excretion of lithium rather than the clearance of lithium as such would be advisable to remove the confounding effects of age and anthropometry, at least in a white middle-aged male working population.