Plasma immunoreactive glucagon (IRG) concentrations were measured in 36 patients with chronic renal failure (CRF) and 32 normal subjects. In addition, the components of circulating IRG were analyzed by gel filtration in the fasting state and after physiological stimuli. Fasting IRG was elevated (P less than 0.001) in CRF patients (534 +/- 32 pg/ml) compared with the levels found in healthy subjects (113 +/- 9 pg/ml). Oral glucose suppressed plasma IRG in CRF patients from a basal level of 568 +/- 52 to a nadir of 354 +/- 57 pg/ml (120 min). This degree of suppression (38%) was comparable to that found in normal subjects (basal = 154 +/- 20 to 100 +/- 23 pg/ml) at 120 min (35%). Intravenous infusion of arginine (250 mg/kg) resulted in a 71% rise in IRG in CRF patients and a 166% increase in normal subjects. Gel filtration of fasting plasma from CRF patients showed three major peaks. The earliest (A) was found in the void volume (mol wt greater than 40,000) and constituted 16.5 +/- 4.7% of the elution profile. The middle peak (B) eluted just beyond the proinsulin marker (approximately 9,000 mol wt) and constituted the largest proportion of the elution profile (56.5 +/- 3.4%). The third peak (C) coincided with the standard glucagon and [125I]glucagon markers (3,485 mol wt) and comprised 27.0 +/- 4% of the IRG profile. In contrast, only peaks A and C were found in fasting plasma of normal subjects (53.6 +/- 10.4% in A and 46.4 +/- 10.4 in C). After oral glucose, glucagon immunoreactivity in the 3,500 mol wt peak (C) was markedly suppressed, while the B peak in patients with CRF declined to a lesser extent. The A peak in both groups was unchanged. After an arginine infusion only the C peak increased in both groups of subjects. Gel filtration of plasma in 3 M acetic acid gave similar profiles to those obtained in glycine albumin buffer. Exposure of serum to trypsin indicated that the B and C peaks were digestible, while the A peak was resistant to the action of the enzyme. In one sample, peak C increased after a 2-h exposure of serum to trypsin. We conclude that circulating IRG in normal subjects and patients with CRF is heterogenous. The hyperglucagonemia of renal failure is largely due to an increase in IRG material of approximately 9,000 mol wt, consistent with proglucagon, although the 3,500 mol wt component is also considerably elevated (threefold). The significance of circulating IRG levels should be interpreted with caution until the relative biological activity of the three components is established.