A percutaneous renal biopsy can be performed in several ways, including using a spring-loaded biopsy gun. As this form of renal biopsy has become more popular, a controversy has developed regarding tissue adequacy and the incidence of complications. To compare these two aspects in an automated biopsy and a manual biopsy, we studied 166 patients assigned to one of the two renal biopsy methods. In a randomized, prospective manner from June 1994 until February 1997, group 1 (67 patients) received a 14 G Tru-cut needle (Baxter, Deerfield, IL) manual biopsy while group 2 (99 patients) received an 18 G automated gun biopsy. There was no difference in sex, age, hemoglobin level, prothrombin time, partial thromboplastin time, or diastolic and systolic blood pressure prebiopsy in groups I and II. Indications for biopsy were proteinuria (38%), proteinuria accompanied by hematuria (31.3%), acute renal failure (9.6%), lupus nephropathy (9.6%), chronic renal failure (6%), and hematuria only (5.4%). In group I, the number of cores was 1.88 +/- 0.56, the glomeruli obtained were 27.3 +/- 13.8, and the number of glomeruli per core were 15.3 +/- 8.4. In group II, the values were 2.37 +/- 0.88, 20.7 +/- 11.1, and 9.95 +/- 6.9, respectively. These results showed a statistically significant difference (P < 0.05). In all cases, pathological diagnosis was possible. The histology showed IgA nephropathy in 25.9%, minimal change disease in 16.3%, lupus nephritis in 11.4%, membranous glomerulonephropathy in 9.3%, membranoproliferative glomerulonephritis in 5.4%, and others. The incidence of postbiopsy hematoma was marginally greater in group I (22.3% v 11.1%) and the area of perirenal hematoma shown on ultrasound 24 hours postbiopsy was larger in group I, as well (848 +/- 623 mm2 v 338 +/- 260 mm2). Hematocrit levels before and after biopsy showed a significant difference (34.9% +/- 7.9% and 34.0% +/- 7.6%, respectively; P < 0.05) in group I, but no significant difference was observed in group II (35.1% +/- 7.0% and 34.7% +/- 6.9%). Both techniques rendered adequate tissue sampling, but the extent of bleeding was more severe with the manual 14 G Tru-cut needle biopsy.