Matched Pair Analysis of Shock Wave Lithotripsy Effectiveness for Comparison of Lithotriptors

Abstract

In an effort to streamline a comparison of the effectiveness of a new lithotriptor with the standard HM3 lithotriptor (Dornier Medical Systems, Inc., Marietta, Georgia) we used a matched pair analysis design. A matched design often provides more efficient estimates (smaller variances) than an unmatched design given the same sample size. Patients with solitary renal or ureteral calculi treated on a LithoTron shock wave lithotriptor (HealthTronics, Marietta, Georgia) between October 1999 and February 2000 with a minimum followup of 3 months were identified. Evaluable patients treated with the LithoTron were matched using 5 parameters to a data base of patients treated with an unmodified HM3 shock wave lithotriptor between October 1997 and February 2000. Matching criteria consisted of calculus side, calculus location (1 of 7 categories), maximum stone diameter (+/-2 mm.), minimum stone diameter (+/-2 mm.) and patient body mass index (BMI +/-6). When more than 1 match was suitable, matching was directed by random numbers. Following matching, clinical charts and radiographic reports were evaluated for stone clearance and post-shock wave lithotripsy interventions. Stone treatment success was defined as residual fragments less than 2 mm. without need for further intervention. A total of 94 potentially evaluable patients treated with the LithoTron were identified and 38 matched pairs were created. Average maximum stone diameter, minimum stone diameter, and BMI were 9.6 and 9.9 mm., 6.7 and 6.8 mm. and 29.3 and 28.9 kg./m. for HM3 and LithoTron cases, respectively. All calculi were radiopaque and consisted of mixed calcium oxalate monohydrate (19 and 13), calcium oxalate dihydrate (1 and 1) or calcium phosphate (2 and 2) in the HM3 and LithoTron groups, respectively. Patients were not specifically matched on stone composition because of incomplete availability. Overall intervention-free, stone treatment success rate was 79% for the HM3 and 58% for the LithoTron. OR for failure of LithoTron versus HM3 treatment was 3.004 (McNemar test p = 0.08). There were 16 discordant pairs. In 4 cases LithoTron was successful and HM3 failed, and in 12 cases LithoTron failed and HM3 was successful. Subgroup analysis revealed a trend for LithoTron treatment failure for lower pole calculi, calculi 10 mm. or greater and BMI of 30 kg./m. or greater. In this initial evaluation the HM3, despite a relatively small study sample size, appeared to provide superior clinical results to the LithoTron (p = 0.08). The use of matched pair analysis using a large cohort of patients treated with the HM3 for retrospective matching may allow for accurate determination of the effectiveness of new lithotripsy technology with a relatively small clinical study group.