Introduction: The efficacy of different treatment in various fields of medical care has become more controversial ever from the introduction of new statistical techniques. The treatment of NS in idiopatic GN is an unsettled Issue where a standard search for evidence-based-medicine techniques Is necessary. Meta-analysis may be used to statistically speculate on observations from clinical trials to drive our clinical decision-making.

Objectives:

Methods: In order to evaluate the effect of immunosuppressive therapy on NS in adult patients with idiopatic GN we performed a meta-anatysis study based on the available randomized controlled clinical trials in the literature. We used MEDLINE and bibliographies from 1968 to 1998 as sources and identified 49 studies evaluating the effects of Prednisone (Pr), Cyclophosphamide (Cy), Chlorambucil (Ch), Azathioprine (Az) and Cydosporine A (CsA) on NS and renal function In patients with idiopatic GN. These include 23 randomized controlled dinical trials of which: 3 in Minimal Change Disease (MOD), 2 in Focal Segmental GlomenJtosderosis (FSGS), 14 In Membranous Glomerulonephritis (MGN) and 4 in Membranoproliferative glomerulonephrife (MPGN). Meta-analysis was performed according to the Mantel/Haenszel/Peto method

Results: The results are shown in the Table.

Therapy Disease No. of trials TOR (CIs)

Pr.Cy MCD 3 Controlled trials 0.10(0.03-0.27)
CsA FSGS coriico- 2 Controlled trials 0.14 (0.05-0.37)
Pr MGN (proteinuria) 3 Controlled trials 1.38 (0.84-2.26)
Pr MGN (renal function) 3 Controlled trials 0.74 (0.45-1.22)
Cy,Az MGN (proteinuriaL) 4 Controlled trials 0.25 (0.10-0.60}
Cy.Az MGN (renal function) 4 Controlled trials 1.20 (0.32-4.40)
Az, Ch,Cy MPGN 4 Controlled tirals 0.55 (0.25-1.20)

Discussion: The analysis shows a beneficial effect of corticosteroids and/or cytotoxic drugs on NS In MOD as compared to no treatment NS in FSGS completely or partially remitted in 50% of treated cases as shown by retrospective comparisons, but findings are not supported by controlled trials. Only 2 randomized controlled trials, for cortico-resistant FSGS patients treated with CsA, were selected for mete-analysis but date were not informative. Mete-analysis indicated a beneficial effect of corticosteroids alone or associated with cytotoxic drugs (Cy, Ch) in those trials where a greater percentage of MGN patients with stage 1 and 2 renal lesions were enrolled (as compared to patients with stages 3 and 4 renal lesions). In these trials the OD's were closer to 0 considering either remission of proteinuria or the outcome of renal function. No benefit of immunosuppressive therapy was shown for MPGN as the TOR was distant from 0 with wide 95% confidence intervals (CIs). Besides our specific interest in the field of gtomerulonephritides, this is an evidence that mete-analysis and the statistic tools may offer great insights to the decision making in the medical procedures. It is obvious that a number of studies need to be performed before standard procedures of treatment may be changed, and date of mete-analysis shall always be "handled with care". Yet, by performing this study we may conclude that patients with stage 1 and 2 membranous glomerulonephritis seem to benefit from therapy while patients with stages 3 and 4 do not respond to treatment.