Non-diabetic Renal Diseases in Patients with Diabetes Mellitus Clinicopathological Correlation

Introduction

Diabetic nephropathy (DN) is the leading cause of end stage renal disease worldwide. Despite this, most patients with diabetes mellitus (DM) are not formally evaluated with a renal biopsy. Kidney biopsy can differentiate DN from non-diabetic renal diseases (NDRD), but it is invasive, not suitable for every patient, associated with serious complications (though rare) and other issues like availability and reluctance, especially more in developing countries like India. The diagnosis is mostly made on clinical grounds. Proteinuria or renal failure in diabetic patients is usually interpreted as a clinical manifestation of DN. However, not all diabetic subjects with proteinuria or renal failure have DN. Non-diabetic renal diseases, though rarer than DN, have been seen to cause proteinuria or renal failure in diabetics.[1] The occurrence of NDRD in type 1 diabetes mellitus (T1DM) is rare compared with those with type 2 diabetes mellitus (T2DM). It has been well demonstrated that renal disease in patients suffering from T1DM for over 10 years is usually the result of DN, as proven histologically in >95% of these patients.[234]

It is usually believed that DN is hard to reverse and further management aims at prevention of progression of disease. But some NDRDs such as minimal change disease, immunoglobulin A (IgA) nephropathy and membranous nephropathy are often treatable and even curable. The therapy and prognosis of DN and NDRD are quite different and the differential diagnosis is of considerable importance. Hence, it is necessary to suspect, diagnose and treat the concurrent glomerular diseases or the unrelated renal disorders that are superimposed on DN because of the prognostic and therapeutic implications.[5]

Non-diabetic renal diseases are an important cause of renal involvement in diabetics but few research works focusing on this entity have been carried out. So, we carried out this study to assess the frequency and spectrum of NDRD in diabetic population and correlate differences in clinical and laboratory parameters between NDRD and DN groups.

Methods

This hospital-based prospective study was conducted between May 2011 and July 2012 at the Department of Nephrology, Institute of Post Graduate Medical Education and Research (IPGMER), Kolkata, India. Institutional ethics approval was obtained to conduct this study. 44 patients with T2DM were enrolled in this study.

Results

Isolated NDRD and NDRD superimposed on DN spectrum

Among patients having isolated NDRD, the most common NDRDs (≥5% of patients) were minimal change disease (19.2%) followed by lupus nephritis (11.5%), post-infectious GN (7.7%), antineutrophil cytoplasmic antibody (+) crescentic glomerulonephritis (7.7%), chronic interstitial nephritis (7.7%), membranoproliferative glomerulonephritis (7.7%), IgA nephropathy (7.7%), and HIV nephropathy (7.7%; Figure 1). Among patients with NDRD superimposed on DN, 33.3% of patients had DN + chronic pyelonephritis and 16.7% of patients each had DN + acute tubular necrosis, DN + kappa light-chain deposition disease, DN + thrombotic microangiopathy and DN + lupus nephritis Class 3 [Figure 2].

Close

Figure 1

Isolated Non-diabetic renal diseases spectrum. ANCA = Antineutrophil cytoplasmic antibodies; GN = Glomerulonephritis; HIV = Human immune deficiency virus; IgA = Immunoglobulin A

Export to PPT

Close

Figure 2

Non-diabetic renal diseases superimposed on diabetic retinopathy spectrum. DN = Diabetic retinopathy; LCDD = Light-chain deposition disease

Export to PPT

Discussion and Conclusions

This prospective hospital-based study was designed to highlight the prevalence and spectrum of NDRD in patients with DM with the intention of implementing more insight into the disease. The study also assessed the relationship of various clinical and laboratory parameters with development of NDRD. The effect of early specific intervention on outcome of the disease was finally evaluated.

The study aimed to include patients with T1DM and T2DM where there was clinical suspicion of NDRD and both types of patients were screened. But clinical suspicion of NDRD was found only in T2DM patients. This is in accordance with published literature where the occurrence of NDRD has been found to be rarer in patients with T1DM than in patients with T2DM. It has been well demonstrated that renal disease in patients suffering from T1DM for over 10 years is usually the result of DN, as proven histologically in >95% of these patients.[234] In fact, when biopsied for research purposes only and not for clinical indications, fewer than 1% of T1 DM patients with 10 or more years of diabetes and fewer than 4% of those with proteinuria and long diabetes duration will have conditions other than or in addition to DN.[5]

Of all the diabetics observed, clinical diagnosis of NDRD was made in 44 patients who were subsequently biopsied. The incidence of isolated NDRD in selected diabetic patients was 61.4% and when patients with NDRD superimposed on DN (13.6%) were clubbed with isolated NDRD, the overall incidence of NDRD rose to 75% (33 patients). The true incidence of NDRD in T2DM patients cannot be deduced from this study as this would require biopsies in all diabetic patients with nephropathy.

Although the exact incidence of NDRD is not known, the reported frequency of NDRD varies widely in published literature.[67] Our incidence of isolated NDRD (61.4%) is in accordance with figures reported in various studies from India and other parts of the world (42.5-64%).[8910] In a study, reported incidence of NDRD was very low (12.3%).[11] These discrepancies of variable frequencies is probably due to differences in the population being studied, different biopsy policies, geographic and ethnic factors, and lack of uniform criteria for biopsy interpretation.

For other baseline characteristics like age, sex, degree of anaemia, azotaemia, proteinuria, age at onset of DM, serum albumin and HbA1C levels, between-group difference was not statistically significant and were not found to be differential indicators of NDRD in this study.

This study found that duration of diabetes (mean years ± SD) was significantly (p = 0.02) shorter in the NDRD group (6 ± 4.6) than in the DN group (10.7 ± 5.85). Previous studies also concluded that shorter duration of diabetes was associated with NDRD.[1012] Thus, shorter duration of diabetes could be a predictor of underlying NDRD.

Prevalence of hypertension was significantly (p = 0.02) more in the DN group than in the NDRD group in this study. Similar observation was made in other studies.[913] However, prevalence of hypertension was similar in two groups in some studies.[1012] High prevalence of hypertension in the DN group can be explained by the fact that hypertension itself is responsible for progression of DN.

Absence of DR is said to be one of the important predictors of NDRD. In this study, DR was absent in equal percentage (54.5%) in patients in both DN and NDRD groups. Controversy exists with regard to absence of DR as an indicator for NDRD. According to a study, absence of DR is a predictor of NDRD,[14] However, others considered it as a poor indicator since the chance of DN and NDRD was 50%.[15] Therefore, the presence or absence of DR alone does not prove to be significant enough in distinguishing between DN and NDRD in T2DM.

In this study, before renal biopsy, all patients were categorized into eight syndromes based on their clinical presentations. Only the presence of CKD was statistically significantly more in the DN group than in the NDRD group. The reason for biopsing patients with CKD presentation was their atypical clinical features like active urinary sediments, duration of diabetes <5 years etc., The wide and variable spectrums of clinical presentation have been reported in literature.[91617] Thus, no clinical presentation is a strong predictor of NDRD other than CKD which is more suggestive of underlying DN.

On the basis of atypical clinical presentations, certain clinical and laboratory criteria were set which merit diabetic patients to undergo renal biopsy to look for NDRD or NDRD superimposed on DN. Between-group difference for any criteria was not statistically significant. Our results favour the results of many reported studies which deny any clinical and laboratory correlation that predict NDRD in diabetic population.[161819] But some other studies oppose these results.[1318] Thus, no clinical or laboratory marker strongly predict NDRD in diabetic patients and renal biopsy is the only investigation presently available to make a definitive diagnosis.

Histologically, NDRD comprised a heterogeneous group in the present study. We further classified the spectrum into those with isolated NDRD and those with NDRD superimposed on DN. In the isolated NDRD group, the most common NDRD was minimal change disease (19.2%) which is in accordance with the results of another Indian study.[9] In the NDRD superimposed on DN group, DN + chronic pyelonephritis (33.3%) was most common. From a review of the relevant literature, most of which consist of isolated case reports, it appears that a wide spectrum of NDRD could occur in patients with T2DM.[2021] Quite contradicting to the present series, some studies reported IgA nephropathy as the commonest NDRD.[1318] It is to be noted that IgA nephropathy is the most common primary glomerular disease in these countries. Incidence of IgA nephropathy of the total primary glomerular disease in an Indian study[22] was only 6%. Hence, its frequency in the NDRD group is also low. These results suggest that prevalence of different categories of biopsy proven renal disease in diabetic patients depends on the usual prevalence of renal disease in general population according to geographical area and ethnic characteristics and NDRD is merely coincidental in T2DM.

In the DN group, 72.7% of patients improved with conservative therapy, 18.2% patients were dependent on dialysis at the time of discharge from hospital whereas in the NDRD group, 78.8% of patients showed recovery in the renal function and clinical improvement with appropriate specific treatment, 15.1% patients were dialysis dependent at the time of discharge. Similar results were reported in a study in which 43% cases showed recovery with specific treatment.[11] In the NDRD group, 2 patients died due to therapy related complications (one crescentic glomerulonephritis patient died due to sepsis after administrating cyclophosphamide, while another patient suffering from thrombotic microangiopathy also died due to sepsis after initiating plasmapheresis). Rest of the NDRD patients (4 patients) were dialysis dependent at the time of discharge. In the DN group, one patient died due to myocardial infarction while undergoing haemodialysis.

The precise diagnosis of various NDRDs has obvious therapeutic and prognostic implications as many NDRDs (eg, minimal change disease, membranous nephropathy, IgA nephropathy) are often treatable and even curable with appropriate management. Even when a coexistent NDRD does not have a specific treatment, the information obtained by uncovering the histopathologic lesion by kidney biopsy helps in prognostication.

Thus, as seen in this study, accurate diagnosis by kidney biopsy and appropriate specific treatment depending on the coexistent NDRD leads to clinical improvement in significant number of patients.