FormalPara Key Summary Points

Renal biopsy for suspected or atypical diabetic kidney disease (DKD) classically reveals one of three possibilities: diabetic nephropathy (DN, pathological diagnosis of DKD), nondiabetic kidney disease (NDKD) and DN plus NDKD (mixed form). NDKD (including mixed form) manifests a wide spectrum of pathological lesions and variable prevalence across the world.

The emerging concept of DKD and the complexity of the routine application of renal biopsy has made the field of identifying DKD and NDKD from clinical indicators a hot topic.

The prevalence of these three types of DKD and pathological NDKD lesions in patients with type 2 diabetes mellitus (T2DM) are summarized and analyzed using data from 40 studies.

Clinical indicators for identifying DKD from NDKD are summarized, based on data from 42 studies. Regrettably, indicators with high sensitivity and specificity have not yet been identified.

Based these analyses, we conclude that, to date, renal biopsy is still the gold standard to diagnose diabetes complicated with renal disease, especially when T2DM patients present atypical DKD symptoms.

Future studies should place more focus on the identification indicators of DKD and NDKD.

Introduction

The number of persons with diabetes mellitus (DM) is increasing globally, especially type 2 DM (T2DM) [1]. In the USA, 23 million adults were diagnosed with diabetes in 2016, of whom 90.9% had T2DM and only 5.8% had type 1 diabetes (T1DM) [2]. Diabetes can lead to many microvascular complications. Diabetic kidney disease (DKD) [3, 4] is a long-term and major complication of DM that occurs in approximately 30% of patients with T1DM and approximately 40% of those with T2DM. It is the leading cause of end-stage renal disease (ESRD) worldwide and is associated with the occurrence of and death due to cardiovascular diseases, especially in T2DM patients [5,6,7,8]. In high-risk groups, such as middle-aged African Americans, native Americans and Hispanics, the incidence of ESRD caused by DKD continues to rise [9]. The various complications and comorbidities associated with DKD and ESRD places a heavy economic burden on healthcare systems [9, 10].

Over the past decades, renal biopsy in T2DM patients suspected of having DKD has shown that many patients actually suffer from nondiabetic kidney disease (NDKD) or diabetic nephropathy (DN) plus NDKD, with both conditions mistakenly identified as DKD. This translates into a difference in approach to treatment and overall prognosis: to date, controlling the incidence of DKD and improving prognosis of DKD are mainly attributed to the improvement of diabetes care. Innovative strategies to prevent, treat and reverse DKD are still unmet and need to be further explored [10, 11]. Unlike DKD, most of the glomerular and tubulointerstitial diseases caused by NDKD usually benefit from personalized treatments, such as immunosuppressive therapy, and can be alleviated or even cured if identified in a timely manner [12]. That is to say, early identification and treatment of NDKD is of great significance to reduce global ESRD prevalence and its various complications, such as cardiovascular complications. However, due to the limitation of renal biopsy applications, a confirmed diagnosis of DKD that at the same time excludes NDKD with certainty in clinical practice is challenging. The introduction of the DKD concept in 2007 [3] encouraged many researchers to search for indicators that can be used to distinguish NDKD from DKD on clinical grounds. However, to date, there has been lack of large published studies that can be considered to be representative, including those with a large amount of data. In addition, it is generally believed that the incidence of NDKD in patients with T1DM is comparatively rare; consequently, data in this area are particularly scarce [13]. Compared with T1DM, T2DM often occurs in middle-aged and elderly people on a background of kidney degenerative lesions. Hypertension, hyperlipidemia, high uric acid level and other metabolic syndromes together may also cause damage to the kidney. In terms of pathology, structural changes in patients with T2DM are more heterogeneous than those in T1DM patients and they correlate less with clinical manifestations, which are also highly heterogeneous [1, 14]. Given this context, studying T2DM patients may have more clinical significance.

Based on the above, the aim of this article is to review the pathological findings of renal biopsy in T2DM patients worldwide and review the clinical indicators used to distinguish DKD from NDKD, in order to help clinicians evaluate patients with T2DM and chronic kidney disease (CKD) and possibly deliver more effective medical management to patients.

Methods

We conducted a literature review using the PubMed literature database covering the time period between 1977 and 2019. The following keywords (and their combination) were used: “type 2 diabetes,” “diabetic kidney disease,” “renal biopsy,” “nondiabetic kidney disease,” “pathological lesions” and “clinical indicators.” Relevant published papers listed on the web page of the “similar articles” section of PubMed/automatically recommended were also included. Studies with a small number of cases and those failing to list the pathological NDKD lesions were excluded. All articles on renal biopsy results and pathological NDKD lesions in T2DM patients (40 articles in total) were included (summarized in Table 1, with studies describing pathological NDKD lesions listed as the top three). Additionally, articles on relevant clinical indicators to differentiate DKD from NDKD (42 articles in total) were included (summarized in Table 3).

This article is based on previously conducted studies and does not contain any studies with human participants or animals performed by any of the authors.

Results

Prevalence and Pathological Types

Analysis of the results of the 40 studies on renal biopsy results and pathological NDKD lesions in T2DM patients (Table 1) showed that the prevalence rate of DN alone ranges from 8.2 to 62.7%, with an average of 41.3%, which is similar to the T2DM rate of about 40% reported by Alicic et al. [1]. The overall prevalence of NDKD (including the mixed form) is 58.7%, which is consistent with the reported prevalence of NDKD in T2DM patients ranging from 8 to 93.5% [55]. The prevalence of isolated NDKD ranges from 0 to 68.6%, with an average of 40.6%, and the prevalence of DN plus NDKD ranges from 0 to 45.5%, with an average of 18.1%. As shown in Table 1, the prevalence and pathological types of isolated NDKD or DN plus NDKD are quite variable, which can be explained by the following:

  1. 1)

    Kidney biopsy criteria. Kidney biopsy is essential for identifying renal damage caused by DN, NDKD or DN plus NDKD [12]. Dong et al. [56] suggested that patients with type 2 diabetes complicated with renal damage should actively undergo renal biopsy for five reasons: high significance of the correct DKD diagnosis; limitations in diagnosing DKD diagnosis; necessity of biopsy-proven pathological diagnosis; indications of renal biopsy; safety of renal biopsy. However, there is no uniform standard for performing kidney biopsy worldwide, which may result in a selection bias. The criteria for renal biopsy in each region, in each country even in each nephrologist’s practice are not identical, which is one of the main reasons for the great variability in the reported prevalence of the pathological types [57]. Furthermore, even if renal biopsy is strongly recommended based on clinical factors, patient compliance is also a factor, leading again to great variability.

  2. 2)

    Geographical regions and heterogeneity of populations worldwide. The prevalence of DN, NDKD or DN plus NDKD differ according to geographical regions and populations, as indicated by two different studies carried out in India. Prakash et al. [58] reported the prevalence of NDKD in North India to be 12.3%, while John et al. [59] reported a higher prevalence of NDKD in South India, especially the proliferative glomerulonephritis type, which was as high as 21.5% among the patients studied. Data from studies conducted in the USA [40,41,42] show a high prevalence of focal segmental glomerulosclerosis (FSGS) in isolated NDKD (22%), which may be associated with the increased incidence of FSGS in Americans, especially in black individuals. Distribution of the pathological type DN plus NDKD appears to be relatively average [40,41,42], but glomerulonephritis is dominant overall [60]. However, in a multi-center study in China [61], membranous nephropathy (MN) was found to be the primary pathological type in NDKD (including mixed form) (40.8%), followed by immunoglobulin A nephropathy (IgAN) (19.8%)

Table 1 Renal biopsy results and pathological types of renal disease in patients with type 2 diabetes mellitus
Full size table

As illustrated in Table 2, the proportion of NDKD cases identified in renal biopsy performed on patients with T2DM is not insignificant. In all continents, with the exception of Europe and Oceania, the proportion of patients with NDKD exceeds that with DN. These studies may overestimate the true prevalence of NDKD in T2DM patients, as indications for biopsy favor atypical cases, such as nephrotic syndrome, abnormal urine sediment, rapid decline of renal function or other NDKD indicators. On the other hand, in clinical practice, patients with T2DM rarely undergo diagnostic renal biopsy. In the absence of renal biopsy, the registry usually identifies patients with DM plus CKD as having DKD, resulting in an increase in the reported prevalence of DKD and a decrease that of NDKD [62]. In Asia, Africa (specifically Morocco and Tunisia) and Europe, the most commonly isolated NDKD pathological type is MN (Table 2), possibly associated with the increasing incidence of MN worldwide in the past decades. According to the literature [63], the incidence of MN in China increased from 6.48% in 1997–1999 to 22.79% in 2009–2011, while in Korea, idiopathic MN (IMN) was identified in 12.0% of all renal biopsies. Similar to Asia, the prevalence of IMN increased from 11.2 to 29.4% in Europe. A large-scale data study by Xu et al. [64] explained the high incidence of MN in China in the past decades in the context of environmental pollution (especially the increase in PM 2.5 levels). In areas with PM 2.5 > 70 μg/m3, for every 10 μg/m3 increase in concentration, the risk of developing MN was noted to increase by 14% (odds ratio 1.14, 95% confidence interval 1.10–1.18), while the effect in areas with PM 2.5 < 60 μg/ m3 was much lower. In support of this explanation, the prevalence of other major glomerular diseases remained relatively stable with increasing air pollution in China over the past decades, leading the authors to conclude that patients with long-term exposure to high levels of PM 2.5 may be more prone to MN than to other kidney diseases in a nonlinear pattern. In North America (specifically the USA), FSGS is the primary pathological type in patients with isolated NDKD (22%) (Table 2), which may be also related to the high incidence of FSGS. Consistent with previously published data [65], FSGS has always represented the primary pathological lesion in the USA, reported to be an average of 25.3% from 1986 to 2015.

Table 2 Summary of renal biopsy results and pathological types in patients with type 2 diabetes mellitus in different continents
Full size table

Table 2 also illustrates the high rate of tubulointerstitial disease in the mixed group (21.7%), with the most common types being acute interstitial nephritis (AIN) (9.3%), followed by acute tubular necrosis (ATN) (9.0%). This high rate of tubulointerstitial disease may be associated with ischemic injury caused by diabetic vascular disease combined with other complex factors, such as susceptibility to infection, direct tubulointerstitial damage due to alterations in extracellular matrix synthesis and metabolism under the influence of glucose and/or the use of non-steroidal anti-inflammatory drugs, antibiotics, among others. These factors together lead to tubulointerstitial injuries. In conclusion, the high frequency of tubulointerstitial disease in patients with DKD may be related to the increased susceptibility of the kidney to ischemia and toxic injury [36].

Clinical Indicators

The natural history of progression of T2DM/DKD to ESRD is not as clear as that of T1DM/DKD to ESRD. In addition, the clinical manifestations of this progression are varied. As such, many researchers have focused on trying to identify clinical indicators to distinguish DKD, single NDKD or a combination of the two in T2DM patients. Relevant clinical indicators include proteinuria, hematuria, duration of diabetes, diabetic retinopathy (DR), hypertension, hemoglobin, glycated hemoglobin (HbA1c), serum creatinine, serum cystatin C (CysC), C3, immunoglobulin, among others (Table 3)

As illustrated in Table 3, DR is an important clinical indicator predicting DKD or NDKD [18, 19, 21, 22, 24, 29, 31, 32, 37,38,39,40, 44, 46, 48, 51, 53, 54, 69,70,71,72,73,74,75,76,77, 80]. The presence of DR strongly suggests DKD, and the absence of DR is a major indicator to predict NDKD. In a multi-variate analysis, the overall diagnostic efficiency of DR was found to be quite significant [75, 76, 80]. In the study by Castellano et al. [51], the presence of retinopathy had a predictive value of 100% for DN. However, in a meta-analysis by Liang et al. [72], 23.6% of patients with biopsy-proven DKD did not have DR and there was no evidence of co-existence of DN and DR in 17.6% of patients. In summary, the absence of DR can predict NDKD, but DKD cannot be ruled out. Similarly, even if DR is present, NDKD can also occur (DKD plus NDKD). As such, Zwi et al. [81] suggest that biopsy should be performed to exclude interstitial nephritis and FSGS even in the presence of retinopathy. Jiang et al. [77] also found that the severity of DR was not related to the presence of DKD.

The duration of diabetes is an important predictor of DKD and NDKD according to many published reports [16, 21, 22, 24, 26, 29, 37, 41, 44, 48, 54, 69, 72, 74,75,76, 80]. The major finding suggested by these reports is that a shorter duration of diabetes is suggestive of NDKD, while a longer duration is related to DKD. Some reports even provide a relatively specific time period: a duration of < 5 years from the time of diagnosing DM to developing a kidney disease is an independent predictor of NDKD [7, 16, 69, 75, 76], while a duration of > 10 years is an important indicator of DKD. More specifically, Yenigun et al. [ [37] and Sharma et al. 41] identified a duration of > 12 years as a predictor of DKD and NDKD with high sensitivity and specificity.

Hematuria is also a common predicting factor of DKD and NDKD [21, 26, 45, 48, 66, 74, 75]. Jiang et al. [82] further found dysmorphic erythrocytes were more valuable than microhematuria in the diagnosis of a nondiabetic lesion in T2DM with proteinuria. Lower HbA1c can also predict NDKD [16, 26, 32, 33, 53, 72]. Wang et al. [83] further pointed out that both the ratio of glycated albumin to HbA1c and glycated albumin were better biopsy-proven DKD indicators than HbA1c.

Other clinical predictors studied are proteinuria and serum CysC. Proteinuria as a predictor, most of the published literature suggest that sub-nephrotic range proteinuria is associated with the presence of NDKD, while nephrotic range proteinuria suggests DKD [20, 23, 37, 74, 80]. Ghani [34] also found that proteinuria level in patients with NDKD superimposed on diabetic glomerulosclerosis is typically higher; however, this finding needs to be supported by data from study with a large sample. Moreover, in a meta-analysis [67], researchers found that serum CysC was an early predictor of DN. Suzuki [78] also confirmed CysC can effectively detect grade 2 DN (urine albumin to creatinine ratio 30–300 mg/g) and will be especially useful for screening grade 2 CKD patients (K/DOQI–KDIGO [Kidney Disease Outcomes Quality Initiative–Kidney Disease Improving Global Outcomes] joint statement [3]).

Studies combining more than one predictor to distinguish DKD from NDKD have also been reported. Wong et al. [21] found that combination of absence of DR with hematuria or proteinuria (≥ 2 g/day) was the most sensitive indicator of a nondiabetic lesion, and Soni [29] considered that a short duration of diabetes and absence of retinopathy, especially when associated with nephrotic proteinuria, strongly predicts a nondiabetic lesion, consistent with results from a research paper from Australia [69]. Our analysis of the published literature revealed some contradictions. For example, in a study by Lin et al. [20], duration of DM of > 10 years and absence of DR failed to rule out NDKD, and in a study by Tone et al. [76] microscopic hematuria and granular casts had low specificity and sensitivity for NDKD. Furthermore, a study by Zhang et al. [68] showed that with different levels of proteinuria, low IgG and high C3 were independent indicators of NDKD in patients with T2DM, while Byun et al. [22] described an association between higher levels of IgG and NDKD, and Kanodia et al. [74] described an association of NDKD with low levels of C3 and/or C4. All of these contradictions require further study and verification by analysis of large-scale data.

Although single or combined indicators have resulted in higher sensitivity and specificity, most have been derived from single sample studies, which cannot exclude the effect of small sample size, selection bias and other such limitations. Christensen et al. [50] clearly pointed out that clinical or laboratory data could not be used to separate diabetes and nondiabetic glomerular lesions based on demographics. Clinically, if atypical DKD features are present, renal biopsy is still the only means to confirm the presence or absence of NDKD [2, 16,17,18,19,20, 22, 24,25,26,27, 31,32,33,34, 37,38,39,40, 46, 48, 50, 51, 69, 75, 80, 84]. To this end, we have divided these indicators from Table 3 into well-established indicators to NDKD and potential indicators to NDKD (Fig. 1). Whether potential indicators are able to predict NDKD requires more research data. However, when T2DM patients have the atypical DKD symptoms suggested by the above indicators, having kidney biopsy as early as possible to determine the presence and pathological type of NDKD is beneficial to the initiation of personalized treatment for patients.

Table3 Summary of clinical indicators for predicting diabetic kidney disease or nondiabetic kidney disease in patients with type 2 diabetes mellitus
Full size table
Fig. 1
figure 1

Clinical indicators of nondiabetic kidney disease. HbA1c Glycated hemoglobin IgG immunoglobulin G

Full size image

Discussion

Due to national policies and the paucity of relevant research, we were unable to collect comprehensive data for our review, particularly with regard to South America. Data from North America are limited to the USA, data from Oceania are also limited to New Zealand and data from Africa are only from Morocco and Tunisia; data from other countries on these three continents are lacking. However, our analysis of the available data allowed us to identify the high prevalence and wide spectrum of NDKD in patients with T2DM, as well as describe the sensitivity and specificity of the related clinical predictive indicators.

Overall, our findings suggest that, to date, renal biopsy is still an important strategy to differentiate between the three kinds of kidney disease complicating T2DM; this is particularly true when T2DM patients present atypical DKD symptoms (e.g. absence of DR, shorter duration of diabetes, microscopic hematuria, sub-nephrotic range proteinuria, lower HbA1c, lower fasting blood glucose). We look forward to collecting more data from Oceania, North America and Africa in the future to further verify or better explain our results.

Although the results from our review indicate the importance of kidney biopsy in T2DM complicated with renal disease, we admit that there are still some problems in applying this procedure. For example, kidney biopsy was considered to be safe in most studies [85,86,87], but patients with T2DM and kidney disease are usually older, and diabetes can cause damage to blood vessels and multiple systems, which might make renal biopsy more complicated. In addition, there is no consensus on the safety of kidney biopsy between kidney diseases complicating T2DM and kidney diseases without diabetes globally. Moreover, the decision-making process for kidney biopsy is mostly determined by the attending doctor, and research on the development of standardized criteria for performing renal biopsy in patients with diabetes is very limited.

In view of the above-mentioned issues, we suggest that future studies should focus on the identification of indicators of DKD and NDKD, not only those at the biochemical level but also at other levels, such as ultrasound and genomics. The availability of such indicators will increase the clinican’s confidence in making a diagnosis and optimize the decision-making process regarding the implementation of kidney biopsy. In addition, future research can make full use of modern technological advances to achieve data sharing, conduct analyses of large datasets and apply innovative experimental designs. By carrying out dynamic monitoring of and long-term follow-ups on T2DM patients, future studies can gain a deeper understanding of the medical methods needed to identify DKD and NDKD and allow a more comprehensive interpretation and evaluation of renal biopsy’s application value in T2DM complicated with renal diseases.

Our review has a number of limitations. First, we only used the PubMed database for our literature search, which may have resulted in omissions of some articles. In addition, we did not use additional statistical methods to quantify our conclusions, such as meta-analysis. In view of the deficiencies of our review, we will continue to pay attention to whether more research data appear, especially data from Africa, North America and Europe, and use a systematic approach to analyze the literature.

Conclusions

The high prevalence and various types of NDKD, as well as the complexity of DKD clinical manifestations, make it more difficult for clinicians to diagnose the three types of renal disease (DN, NDKD and DN plus NDKD [mixed form]). Although the application of renal biopsy in T2DM patients is controversial, our review suggests that, to date, renal biopsy is still an important strategy to identify T2DM complicated with renal disease, especially for patients with atypical DKD symptoms.