Percutaneous renal allograft biopsy. A comparison of two needle types and analysis of risk factors.

We retrospectively reviewed all (n = 369) percutaneous renal allograft biopsies performed at our institution between 1987 and 1992, comparing 14-gauge Franklin-Silverman (internal diameter = 2.0 mm, n = 169) and 18-gauge automated (internal diameter = 1.2 mm, n = 200) core biopsy needles. Visualization method, specimen adequacy, and complications were grouped by needle type. Five or more glomeruli were present in 88.9% of specimens obtained with Franklin-Silverman needles and in 82.7% with automated needles. A histologic diagnosis was obtained in 94.1% and 95.5% of Franklin-Silverman and automated biopsies, respectively. A complication was detected in 27 Franklin-Silverman biopsies (16.0%) and in 21 automated biopsies (10.5%) (not significant [NS], P > 0.05). Some procedures had more than one complication. Excluding asymptomatic gross hematuria, incidental hematomas, and incidental arteriovenous fistulas detected by routine ultrasonography, clinically significant complication rates were 6.5% for Franklin-Silverman biopsies and 2.5% for automated biopsies (NS). No allograft losses or patient deaths occurred as a result of allograft biopsy. Subgroup analysis of all biopsies performed with ultrasound marking alone (Franklin-Silverman, n = 119; automated, n = 148) revealed no significant (NS) difference in complication rates (15.1% vs. 10.8%). Additional subgroup analyses of palpation, ultrasound marking, and real-time ultrasonographic visualization techniques within each needle type also revealed no significant difference in the complication rate. Biopsy within 30 days of transplantation and no antihypertensive therapy were the only factors univariately associated (P < 0.05) with an increased complication rate. Multivariate analysis found biopsy within 30 days of transplantation (P = 0.007) was associated with the overall presence of one or more complications of any type. Type of needle (Franklin-Silverman vs. automated) achieved borderline significance (P = 0.047) when time to biopsy was statistically adjusted for; the Franklin-Silverman needle had a higher complication rate.