Should all patients with diabetes have a kidney biopsy? Free

Diabetes remains the most frequent single cause of end-stage renal disease (ESRD) in the USA [1] and in other developed countries [2] and, along with cardiovascular disease [3], ESRD [4] is the main condition responsible for the excess mortality in patients with diabetes. Although data suggesting a reduction in ESRD incidence among patients with diabetes in the last decade [5] brought some optimism back into the field, sensitivity analyses of these same data revealed a similar risk of ESRD among patients with diabetes in 1990 as compared to 2010. Thus, when individuals with newly diagnosed diabetes (less than 2-year duration) were removed from the analysis, the ESRD rates in 1990 (27.9/10 000) were very similar to those observed in 2010 (24.1/10 000) [5]. Therefore, despite investments in research and improved clinical care, the rates of ESRD remain alarming. The increased incidence of type 2 diabetes in children and adolescents, particularly among minorities [6], and the lack of data on the natural history of diabetic nephropathy in this younger population with type 2 diabetes will ultimately add to this challenge. Data from the US Renal Data System already indicate that incident ESRD due to diabetes has increased among young minority patients [1]. Thus, there is no question that new and more efficacious preventative and therapeutic strategies are urgently needed. To develop new therapeutic strategies, it is paramount to have a more precise and complete understanding of the mechanisms of kidney injury and repair in diabetes. In part, this knowledge comes from proper analyses of clinical data, blood, urine and tissue samples, including kidney biopsy materials, derived from patients with diabetes. In addition, the more precise identification of individuals with diabetes in whom kidney dysfunction is truly due to diabetic nephropathy, so that they can be targeted for recruitment in studies evaluating the efficacy of new therapeutic strategies, can greatly influence the success of clinical trials.

In this issue of Nephrology Dialysis Transplantation, a pooled meta-analysis examines the prevalence of non-diabetic renal disease among patients with diabetes [2]. Although these results, at first glance, seem to suggest that almost half of the patients with diabetes have another form of kidney disease, either in addition to diabetic nephropathy or in isolation, these data need to be brought into perspective. Although there is no doubt that the correct diagnosis of the underlying kidney disease is important, as it can influence therapeutic strategies and prognostication, it is also important to consider the likelihood of an alternative explanation for kidney dysfunction in patients with diabetes before determining the need for a kidney biopsy. As described in the above-mentioned article [2], the likelihood that a patient with diabetes will have another kidney disease varies largely, and it clearly depends on the patient’s clinical presentation.

In the present report [2], data from 48 studies reporting kidney biopsy results in patients with diabetes were assessed. Although these studies involved a total of almost 5000 individuals, they were quite heterogeneous, and the majority were retrospective in design (75%), only included patients with type 2 diabetes (83%) and, more important, collected data from biopsies performed for clinical indications (95.5%) as opposed to research purposes. Another interesting aspect is that the majority of the studies included in these meta-analyses reported data from Asian populations, where the frequency of other kidney diseases, such as IgA nephropathy, in their non-diabetic population is greater than in other areas of the world, and a greater prevalence of other kidney diseases in the background population could also have influenced the results.

While other forms of kidney disease may be relatively common among patients with diabetes who follow an atypical course, they seem to be uncommon among those tracking the known natural history of the disease. This is illustrated in a multicenter study conducted in Italy comparing the rates of kidney diseases other than diabetes among centers with a restricted ‘versus’ those with an unrestricted biopsy policy [7]. In centers with a restricted policy, kidney biopsies were indicated if other renal disease was suspected (i.e. presence of hematuria, nephrotic syndrome, proteinuria in the absence of diabetic retinopathy, rapid progressive renal failure or unexplained renal insufficiency). Centers with an unrestricted policy performed kidney biopsies if a patient had proteinuria >500 mg/day and/or hematuria and/or renal dysfunction. However, even in centers with an unrestricted policy, if diabetic nephropathy seemed obvious (long duration of diabetes, presence of other chronic complications of diabetes—diabetic retinopathy, neuropathy, vasculopathy, etc.), a kidney biopsy was not performed. This report demonstrated that in centers with a restricted policy, where patients would necessarily have a greater positive predictive value of having another kidney disease, other kidney disease was found in association with diabetic nephropathy or in isolation in 38% and 33% of the cases, respectively. Among centers with an unrestricted biopsy policy, however, despite eliminating all patients with obvious diabetic nephropathy from their candidate biopsy pool, and then lowering the positive predictive value that these patients would have diabetic nephropathy, diabetic nephropathy was still found to be present in the majority of these patients with type 2 diabetes, and 73% of the biopsies confirmed the presence of diabetic nephropathy. Accordingly, in centers with an unrestricted policy, diabetic nephropathy was the sole pathology diagnosis in 51% of the kidney biopsies and it was associated with other glomerular disease in 22% of the cases, while nephrosclerosis or other glomerular diseases were observed in the remaining 27% of the biopsies. Hence, when the course is atypical, and there is clinical suspicion that non-diabetic renal disease may be present, the likelihood that this will be confirmed by a kidney biopsy is high, and a kidney biopsy may not only be justified but also clinically indicated.

Although the risks associated with kidney biopsies in patients with diabetes do not seem to be higher than in patients without diabetes, kidney biopsy remains an invasive procedure, and in the clinical setting it should be ordered after proper consideration. In type 1 diabetes, onset of proteinuria in patients with less than 5 years of disease duration, in the absence of other chronic complications of diabetes, is often considered an indication for kidney biopsy. In patients with type 2 diabetes, where duration of disease may be uncertain, other criteria are helpful in guiding the indication for kidney biopsy, and these include acute onset or worsening of renal disease, presence of active urine sediment—red cells (acanthocytes), cellular casts—and signs or symptoms of other systemic disease.

However, there is more to learn from kidney biopsies in patients with diabetes than whether or not other kidney diseases might be present. Based on research biopsy studies, we learned about the timing of renal structural changes in diabetes, the relationships between renal structural lesions and clinical signs of renal disease (increased urinary albumin levels and decreased glomerular filtration rate), as well as the relationships between these structural lesions and other clinical characteristics (glycemic control, blood pressure levels, etc.). We also learned that structural lesions of diabetic nephropathy can be present in patients who are normoalbuminuric and have a glomerular filtration rate within the normal range [8], and that when kidney biopsies are performed in research volunteers, the presence of diseases other than diabetic nephropathy is a very uncommon finding. Thus, diabetic nephropathy lesions can be detected in virtually all patients with type 1 diabetes with more than 5-year duration [9] and can clearly develop in clinical silence [8, 10]. Moreover, presence of more severe glomerular lesions among patients with type 1 diabetes who had normal urinary albumin levels at time of kidney biopsy was associated with lower glomerular filtration rate in cross-sectional studies [11] and with a greater likelihood of progressing to ESRD in longitudinal studies [10].

To have a more precise representation of the diagnostic challenges identified by this report [2], research biopsy studies would need to be conducted in patients with both type 1 and type 2 diabetes, and include patients recruited before and after development of clinical manifestations of renal dysfunction. Research biopsy studies in type 1 diabetes [12, 13] did not reveal the presence of other renal disease in these individuals. However, individuals recruited for these studies were selected to have minimal or no manifestation of kidney dysfunction, while the studies reviewed in the current meta-analyses reported data derived mostly from patients with clear manifestations of kidney disease. Also, it is very likely that most patients who underwent kidney biopsies on the studies evaluated on the current report followed an atypical course and thus had a kidney biopsy performed for clinical reasons, whose implications we discussed earlier. Additionally, patients with type 2 diabetes often have hypertension and dyslipidemia at time of diabetes diagnosis, while these comorbidities are usually not present in patients with type 1 diabetes until later on. Diabetic nephropathy is a serious disease, with a high mortality rate, and it deserves our full attention. The diabetes epidemic will add to the disease burden and present new diagnostic and therapeutic challenges. Kidney biopsies are a valuable research tool, and ought to be considered for diagnostic purposes in patients with atypical clinical presentation.

CONFLICT OF INTEREST STATEMENT

None to declare.

(See related article by Fiorentino et al. Renal biopsy in patients with diabetes: a pooled meta-analysis of 48 studies. Nephrol Dial Transplant 2017; 32: 97–110)

REFERENCES