Abstract Background and Aims Diabetic Nephropathy (DN) is usually a presumptive diagnosis based on clinical and biological evidences. However, Renal Biopsies (RB) may be performed when the suspicion of Non-Diabetic Renal Disease (NDRD) with potential specific treatment arises from atypical features mainly identified through studies focusing on the power of individual clinical and biological features to predict NDRD. We aimed to assess the actual value of RB indications (RBi) to discriminate NDRD from Type 2 Diabetes (T2D) Associated Renal Disease (T2D-ARD) such as DN and Hypertensive nephropathy which are frequently associated, seen in similar patients with high cardiovascular risk, lack for specific treatment and might be therefore be considered as “failure events” in the context of RB performed to diagnose NDRD with specific treatment. Method We conducted a retrospective multicenter study in four nephrology units and reviewed the charts of patients with T2D who underwent a first RB from 2006 to 2015. Clinical and biological characteristics, RBi and diagnoses were retrospectively collected from the patients’ medical charts. RBi were defined according to the sequence in Figure 1. As such, diabetic retinopathy may be absent in patients of all indication groups, conversely, patients in group 6 and 7 had stable proteinuria and renal function. Similarly, patients with monoclonal gammopathy were all included in group 1.We considered missing data regarding the absence/presence of Diabetic Retinopathy (DR) as meaningful, suggesting the data was unneeded to reach a decision of RBi. We performed univariate and multivariate logistic regression analyses to identify clinical or biological features associated with the diagnosis of NDRD. We also compared the number and percentage of actual NDRD diagnosed for each of the RBi. Results 464 first renal biopsies were performed in 464 patients during the study period. Two RB were excluded because the RBi could not be clearly identified through careful reviewing of the patients’ medical charts. Of the 462 remaining RB, 40% revealed NDRD. The most frequent were acute interstitial nephritis, IgA nephropathy and acute tubular necrosis (8.7%, 5.8% and 4.1% respectively). Patients with NDRD were more frequently >65yo (55 vs 44%), with serum creatinine (SCr) >250 µmol/l (52 vs 30%), hematuria (58 vs 36%), non-retrieved DR status (45 vs 17%), and less frequently with urinary-to-protein ratio (UPCR) >3g/g (33 vs 57%), HbA1c >7% (32 vs 50%) and duration of diabetes >5 years (52 vs 72%).Logistic regression analyses identified high SCr, UPCR, hematuria and low diabetes duration as predictive of NDRD. Surprisingly, missing RD status rather than absence of RD was predictive of NDRD. As expected NDRDs were frequent in patient with RBi 1 to 3 (53%, 40% and 21%), infrequent in patients with RBi 5 (7%) and totally absent from patients with RBi 6 and 7 despite their predictive value as regards with NDRD. Conclusion Our study confirms several well-known facts on the subject of renal biopsy in patients with T2D: 1) Renal biopsy is useful in selected patients with T2D and renal disease to diagnose NDRD with specific treatment. 2) Hematuria and short duration of diabetes are predictive of NDRD. On the other hand, our study suggests two novel and antidogmatic findings: 1) Assessing the presence of diabetic retinopathy is not required for RBi. 2) Hematuria is not indicative of NDRD in patients with otherwise typical diabetic kidney disease. Of course, our study is limited by its retrospective design precluding us of precisely defining what was considered rapid evolution of DFG or proteinuria emphasizing the need for prospective studies.
Read full abstract