Impairment of glomerular filtration rate (GFR) after simultaneous pancreas and kidney (SPK) transplantation is an important marker of chronic renal rejection and recurrence of diabetic glomerulopathy. The use of unmodified serum creatinine to estimate GFR, however, is limited by variations in muscle mass. In this study, predictive factors for long-term GFR were evaluated in consecutive SPK recipients (n = 33) using a Tc99m DTPA GFR reference method between 90 days and 6 years after transplantation (n = 136 measurements). Substantial variability between serum creatinine and isotopic GFR after SPK (R2 = 0.30) high-lighted the inaccuracy of an unmodified serum creatinine in the evaluation of GFR. Factors which predicted GFR apart from serum creatinine included age, sex, height and body weight. A detailed formula was derived for accurate estimation of GFR (ml/min) = [71.4 (male) or 50.4 (female)] + 5520/creatinine (mumol/l) + 0.27 x body weight (kg) - 0.50 x age (yr) - 0.29 x height (cm). This formula was more accurate in estimation of GFR in SPK recipients than six published predictive methods which were derived from chronic renal failure patients using creatinine clearance. All of these methods overestimated GFR at lower levels of renal function. Most correlated poorly with Tc99m DTPA GFR and contained a generalized systematic overestimation of GFR which ranged from 4.7 to 8.4 ml/min (p < 0.05). A simplified version for rapid calculation was also derived as GFR (ml/min) = [25 (male) or 5 (female)] + 5000/creatinine (mumol/l).(ABSTRACT TRUNCATED AT 250 WORDS)