INTRODUCTION — Diabetes is the leading cause of chronic kidney disease (CKD) and end-stage kidney disease (ESKD) in the United States and worldwide. While the gold standard for diagnosis of diabetic nephropathy is defined by histology of the kidney, the majority of patients do not undergo kidney biopsy, as they are presumed to have diabetic kidney disease based upon clinical history and laboratory evaluation. This clinical practice is based in part upon the desire to avoid an invasive procedure that may not alter treatment, as well as the notion that there is a uniform clinical disease presentation, a traditional belief based upon observational studies performed several decades ago. However, diabetic kidney disease is now known to be clinically and pathologically heterogeneous.
The clinical manifestations, evaluation, and diagnosis of diabetic kidney disease are reviewed in this topic. Other topics discuss the following issues:
●Epidemiology and pathogenesis of diabetic kidney disease (see "Diabetic kidney disease: Pathogenesis and epidemiology")
●Treatment of diabetic kidney disease (see "Treatment of diabetic kidney disease")
●Management of hypertension in patients with diabetes (see "Treatment of hypertension in patients with diabetes mellitus")
TERMINOLOGY
Diabetic kidney disease versus diabetic nephropathy — The term "diabetic nephropathy" was historically defined by the clinical presence of albuminuria accompanied by retinopathy in patients with type 1 diabetes [1]. The presence of albuminuria was considered to be an early sign of classical diabetic glomerulopathy, which is characterized by glomerular basement membrane thickening, endothelial damage, mesangial expansion and nodules, and podocytes loss (picture 1 and picture 2). Diabetic nephropathy was further subdivided into "overt nephropathy" by "macroalbuminuria" and "incipient nephropathy" by "microalbuminuria." These albuminuria distinctions were proposed to reflect a disease spectrum from mild to severe.
Since the original description of diabetic nephropathy, it has become clear that there are various forms of kidney disease attributable to diabetes, including nonclassical glomerular lesions and tubulointerstitial disease [2]. "Diabetic kidney disease" is a clinical diagnosis based upon the presence of albuminuria, decreased estimated glomerular filtration rate (eGFR), or both, in patients with diabetes.
Notably, diabetic kidney disease (DKD) does not indicate the specific pathological phenotype of kidney damage due to diabetes. Rather, the designation DKD (or "CKD in diabetes") is used to clarify that the underlying pathologic phenotype is unknown in most cases [3]. The likelihood that diabetic glomerulopathy is the underlying pathology of DKD varies widely depending upon the clinical circumstances. It is highly likely that diabetic glomerulopathy is the pathologic phenotype of diabetic kidney disease in type 1 diabetes of five or more years duration with albuminuria, but the frequency can vary widely in type 2 diabetes (where superimposed or de novo kidney disease due to other disorders is common). Only a kidney biopsy can confirm with certainty that diabetic kidney disease is due to diabetic glomerulopathy, but biopsies are infrequently performed for clinical purposes in such patients.
Albuminuria and decreased glomerular filtration rate — The albumin excretion rate can be estimated (using spot, or random, urine measurements of albumin and creatinine and calculating the albumin-to-creatinine ratio), or it can be measured (using a 24-hour urine collection). (See "Assessment of urinary protein excretion and evaluation of isolated non-nephrotic proteinuria in adults", section on 'Detection and measurement of total urinary protein excretion'.)
In clinical practice, GFR is estimated using creatinine-based equations (calculator 1 and calculator 2). (See "Assessment of kidney function", section on 'Measurement of GFR with plasma clearance'.)
Moderately increased albuminuria (30 to 300 mg/g or mg/day) — Previously termed "microalbuminuria," moderately increased albuminuria is defined as either an estimated or a measured urine albumin excretion between 30 and 300 mg/g creatinine or mg/day (table 1). (See "Definition and staging of chronic kidney disease in adults", section on 'Kidney damage'.)
The change in terminology was intended to reflect more accurately the severity of moderately increased albuminuria (a range that is 10 to 100 times the upper limit of normal) and to avoid a common misperception that "microalbuminuria" signified the presence of "tiny albumin molecules."
Microalbuminuria was first described in patients with diabetes in the 1980s as the presence of small amounts of albuminuria that could be detected only with highly sensitive radioimmunoassays [4,5]. This was thought to be the earliest biomarker for underlying microvascular disease of the kidneys and was associated with later development of worsening albuminuria to the severely increased range (called "macroalbuminuria" at the time) and loss of GFR [6,7]. These early studies were performed in a small number of patients with mostly type 1 diabetes and at a time before hyperglycemia was established as a major modifiable risk factor for microvascular complications.
Any detectable albuminuria (even below the threshold for moderately increased albuminuria) forecasts a higher risk for future kidney disease and cardiovascular events in patients with diabetes; therefore, the presence of such low-grade albuminuria should not necessarily be considered "normal," particularly if persistent. (See "Moderately increased albuminuria (microalbuminuria) and cardiovascular disease".)
Research protocol kidney biopsies performed in a cohort of 34 patients with type 2 diabetes and moderately increased albuminuria revealed that while one-third had typical underlying diabetic glomerulosclerosis, one-third had "atypical" findings more consistent with hypertensive nephrosclerosis, and another one-third had normal kidney histology [8]. The findings that a number of patients with diabetes with moderately increased albuminuria have normal kidney histology and can spontaneously regress have led some experts to regard moderately increased albuminuria as being of uncertain clinical relevance [9]. Other experts maintain that it still holds important implications for disease prognosis, not only for future kidney disease progression but also for cardiovascular complications, major comorbidities in patients with diabetic kidney disease [10]. In our view, these perspectives are not mutually exclusive, as moderately increased albuminuria can occur as the result of multiple physiologic and structural aberrations, each with distinct consequences, thus reflecting and reinforcing the extensive underlying heterogeneity of diabetic kidney disease.
Severely increased albuminuria (>300 mg/g or mg/day) — Previously termed "macroalbuminuria," severely increased albuminuria is defined as either an estimated or a measured urine albumin excretion >300 mg/g or mg/day (table 1). (See "Definition and staging of chronic kidney disease in adults", section on 'Kidney damage'.)
Severely increased albuminuria forecasts an adverse prognosis. The risk for worsening GFR, end-stage kidney disease (ESKD), and cardiovascular events rises substantially with albuminuria levels that are severely increased [11-13], and levels above 1000 mg/g or mg/day are associated with even higher risk [14].
Decreased glomerular filtration rate — Normal GFR is generally considered to be above 90 mL/min/1.73 m2, although there is substantial interindividual variability, and even levels in the 90 to 100 mL/min/1.73 m2 range may be pathologic for some individuals, especially if the GFR is on a declining trajectory.
In clinical practice, GFR is estimated using equations that are not sufficiently accurate except when the eGFR is <60 mL/min/1.73 m2; therefore, the accepted definition of "decreased eGFR" is at this threshold, leaving individuals with eGFR 60 to 90 mL/min/1.73 m2 unclassified regarding the clinical implication of their reduction in kidney function.
Measurement (rather than estimation) of GFR, which is rarely performed in clinical practice, is discussed elsewhere. (See "Assessment of kidney function".)
NATURAL HISTORY
Common misconceptions — Clinical outcomes in diabetic kidney disease are quite variable, and there has been an evolution in what was once thought to be a single and unidirectional natural history. Historically, it was believed that patients began with either normal or elevated glomerular filtration rate (GFR; hyperfiltration), with the latter being most evident in type 1 diabetes in the setting of hyperglycemia, and that moderately increased albuminuria was the earliest clinically detectable biomarker of classical diabetic glomerulopathy (figure 1) [15]. A decline in estimated GFR (eGFR) to below 60 mL/min/1.73 m2 was thought to occur after the development of moderately and then severely increased albuminuria.
However, contemporary studies have shown that albuminuria may regress, even from the severely increased range. In addition, decreased eGFR can occur and progress to advanced stages of chronic kidney disease (CKD) before the onset or without ever developing increased albuminuria [16]. The limitations of albuminuria as the "earliest" marker of diabetic kidney disease have led many experts to propose that "rapid eGFR decline" is of greater prognostic importance [17]. (See 'Kidney function decline and end-stage kidney disease' below.)
Glomerular hyperfiltration and elevated GFR — Some patients with type 1 or type 2 diabetes have a paradoxically high GFR early in their disease course (ie, "glomerular hyperfiltration"). Glomerular hyperfiltration is usually defined as GFR approximately 20 percent or more above that in age-matched, healthy controls without diabetes. In younger individuals, the usual threshold for hyperfiltration is considered 120 to 140 mL/min/1.73 m2, whereas in older adults it may be closer to 100 to 120 mL/min/1.73 m2 [18,19]. The pathophysiology of glomerular hyperfiltration and its association with progression of diabetic kidney disease are discussed elsewhere. (See "Diabetic kidney disease: Pathogenesis and epidemiology".)
In studies of patients with diabetes that measured GFR, hyperfiltration was associated with greater risks of albuminuria progression and kidney function decline [20-22]. The kidney protective effects of renin angiotensin system (RAS) and sodium-glucose cotransporter 2 (SGLT2) inhibitors are thought to be mediated, at least in part, by reductions in glomerular hyperfiltration [23,24]. (See "Diabetic kidney disease: Pathogenesis and epidemiology" and "Treatment of diabetic kidney disease".)
Progression (or regression) of albuminuria — In patients with diabetes, regression from moderately increased albuminuria to normoalbuminuria is common and, in patients with type 1 diabetes, may occur more frequently than progression to severely increased albuminuria [25]. Regression of moderately increased albuminuria also occurs in type 2 diabetes but less commonly than in type 1 diabetes [26].
In addition, regression of severely increased albuminuria to moderately increased albuminuria can occur in both type 1 and type 2 diabetes [27-29]. Remission of nephrotic-range proteinuria has also been reported [30,31].
The following studies illustrate the range of findings:
●In a cohort of 386 patients with type 1 diabetes and moderately increased albuminuria, regression to a normal level of albumin excretion occurred in 59 percent at six years, and progression to severely increased albuminuria occurred in 19 percent [25]. Factors favoring regression of albuminuria include better glycemic control, lower blood pressure, and lower total cholesterol and triglyceride levels. In a study of 397 patients with type 2 diabetes and moderately increased albuminuria, 22 percent had regression to normal levels of albumin excretion and 19 percent experienced progression to severely increased albuminuria [26].
●In a large study of patients with type 1 diabetes, severely increased albuminuria regressed to moderately increased albuminuria in 56 percent and to normoalbuminuria in 19 percent [27,28]. Smaller studies in type 2 diabetes have reported that 36 to 48 percent regressed from severely increased to moderately increased albuminuria and that 2 to 13 percent regressed to normoalbuminuria [26,29].
●In two cohorts of patients with type 1 or type 2 diabetes and nephrotic range proteinuria, regression (defined as albuminuria persistent below 600 mg/day) was attained in 22 and 25 percent, respectively [30,31].
Regression of albuminuria is a favorable prognostic indicator. As an example, in a large meta-analysis of multiple cohorts and nearly 700,000 patients (including those with diabetes), a 30 percent regression in albuminuria was associated with a 22 percent reduction in the relative risk of ESKD [32]. In another study of patients with type 1 diabetes, the risk of rapid eGFR decline (defined as 3.3 percent or greater annual loss of eGFR over a period of up to 12 years) was similar in patients who always had normal albumin excretion and patients who had regressed to normal albumin excretion from moderately increased albuminuria [33]. Similar results have been reported in patients with type 2 diabetes who are able to achieve a remission, or even a 50 percent reduction, in albuminuria [29].
However, adverse prognostic consequences of any degree of albuminuria (even when it regresses to the normal range) are observed if patients are followed for a prolonged period. As an example, in the Diabetes Control and Complications Trial (DCCT)/Epidemiology of Diabetes Interventions and Complications (EDIC) study, the risk of developing low eGFR (<60 mL/min/1.73 m2) among those who had regression of albuminuria was fourfold higher at 30 years compared with patients who had persistent normoalbuminuria [11].
Kidney function decline and end-stage kidney disease — Defining the natural history of kidney function decline in patients with diabetic kidney disease is difficult for a variety of reasons [25,34-37]: interindividual variability, which is due in part to sociodemographic and clinical risk factors; temporal trends in the use of effective therapies to slow progression (improvements in glycemic and blood pressure control and use of RAS and SGLT2 inhibitors); nonlinear trajectories in eGFR decline; few studies with longitudinal direct measurements (rather than estimates) of GFR; and relatively low correlation between long-term changes in eGFR compared with measured GFR. In addition, studies of the natural history of GFR decline frequently involve clinical trial participants who typically have a better prognosis than those not enrolled in trials. Therefore, these studies may provide a lower-range estimate of GFR decline compared with the general population.
The accuracy and precision of longitudinal changes in eGFR are imperfect and not necessarily equivalent to longitudinal changes in measured (or "true") GFR. In addition, the accuracy of eGFR equations to estimate GFR slope over time may be worse in diabetic compared with nondiabetic kidney disease [35,37]. While changes in eGFR may not reflect changes in measured GFR, eGFR changes over time nevertheless have prognostic value for end-stage kidney disease (ESKD) and mortality [38].
The average annual rate of decline in eGFR due to age-related senescence of the kidney in otherwise healthy individuals is 0.5 to 1 mL/min/1.73 m2 per year [39]. In diabetes, the rate of eGFR decline varies, but it is typically more rapid (>3 mL/min/1.73 m2 per year), particularly in patients with a long duration of diabetes (>10 years), severely increased albuminuria, or low baseline eGFR (<60 mL/min/1.73 m2). The following studies illustrate the range of findings [35,36,40-42]:
●Among 997 patients with type 1 diabetes in the DCCT/EDIC study who had an average duration of diabetes of 5.5 years at the initial evaluation, the median decrease in eGFR at approximately three years was 3 mL/min/1.73 m2 [35]. However, variability was high, and one-quarter of the population had a three-year decline of 14 mL/min/1.73 m2 or greater, while some patients had an overall increase in GFR.
●In a long-term observational study of 410 people with newly diagnosed type 2 diabetes (mean age 59 years, eGFR 84 mL/min/1.73 m2, 13 percent with moderately increased albuminuria), 11 percent had an eGFR decline >3 mL/min/1.73 m2 per year and 27 percent had a decline between 1 and 2.9 mL/min/1.73 m2 per year. Independent predictors of a more rapid eGFR decline included baseline retinopathy and increased albuminuria [43].
●In another cohort of 1908 people with type 2 diabetes and chronic kidney disease followed over a median of 6.3 years, albuminuria status at baseline was associated with rapidity of kidney function decline. In those with albuminuria <30, 30 to 299, 300 to 1000, or >1,000 mg/24 hours at baseline, annual eGFR decline was -0.2, -1.4, -2.7, and -4.7 mL/min/1.73 m2, respectively [44].
The risk of progression to ESKD depends upon many factors, including the type of diabetes (type 1 versus type 2), access to care, the use of treatments to slow progression, and other risk factors:
●In a population-based study of Australian individuals from 2002 to 2013, the incidence of ESKD was 19 per 10,000 person-years in those with type 1 diabetes and 10 per 10,000 person-years in those with type 2 diabetes. In addition, the incidence rate was higher for Indigenous versus non-Indigenous people and for males versus females [45].
●In a cohort of 932 patients with type 1 diabetes in the United States diagnosed between 1965 and 1980, 20- and 40-year cumulative incidence of ESKD were 5.5 and 27 percent, respectively [46]. By contrast, in a cohort of 7871 patients with type 1 diabetes from Norway diagnosed between 1971 and 2012, the 20- and 40-year rates of ESKD were 0.7 and 5.3 percent, respectively [47]. This substantial difference is multifactorial but is likely due at least in part to temporal trends in glycemic and blood pressure control and use of RAS inhibitors.
●Long-term ESKD rates for type 2 diabetes in the United States also vary. Data from two large cohorts reported incidence rates of 2.5 to 6 cases per 1000 person-years; however, this follow-up time does not represent diabetes duration, which was different at the baseline visit for both studies [48,49]. In addition, many participants in these studies had CKD at baseline. Incidence rates were 0.9, 2.4, 10.5, and 97.8 cases of ESKD per 1000 person-years for patients whose baseline eGFR was >60, 45 to 60, 30 to 45, and <30 mL/min/1.73 m2, respectively.
Although the incidence of ESKD attributed to diabetes has declined negligibly over the past 10 years [50-52], the prevalence of ESKD attributed to diabetes has steadily risen to nearly 50 percent of all cases in the United States. As an attributable cause of ESKD, diabetes is most common and steadily increasing due to the rapidly expanding number of people living with diabetes worldwide [50,53]. Notably, in both diabetic and nondiabetic kidney diseases, African Americans have a higher risk for ESKD than other racial or ethnic groups [50] due to a combination of disparities in socioeconomic status, access to care, and possibly genetic predisposition [54-56]. (See "Diabetic kidney disease: Pathogenesis and epidemiology".)
There are also other populations with exceedingly high rates of diabetic ESKD, such as Aboriginal people in Canada and Northern Australia, wherein the lifetime risk for ESKD approaches 50 percent [57,58].
Nonalbuminuric diabetic kidney disease — Reduced eGFR in patients with diabetes can occur in the absence of albuminuria [59]:
●Among patients with type 1 diabetes and reduced eGFR (<60 mL/min/1.73 m2), 7 to 24 percent are nonalbuminuric (urine albumin <30 mg/day or <30 mg/g of creatinine) [27,60,61].
●Among patients with type 2 diabetes and reduced eGFR, 39 to 52 percent are nonalbuminuric [62-65].
The prevalence of nonalbuminuric CKD rises with age in parallel with albuminuric CKD and is less frequent among diabetic versus nondiabetic cohorts [66,67], indicating that it is probably not simply due to age-related senescence of the kidney [63]. The prevalence of nonalbuminuric CKD is consistently higher in women than in men [63,66-68]. Other factors associated with the presence of nonalbuminuric diabetic kidney disease include RAS inhibition [62,63], hypertension [64,68], dyslipidemia [64,68], and smoking [64,68].
The prevalence of diabetic retinopathy is lower in individuals with nonalbuminuric compared with albuminuric CKD (10 to 43 versus 22 to 62 percent, respectively) [60,63,67,69,70]. By contrast, the prevalence of macrovascular disease is similar between nonalbuminuric and albuminuric CKD groups in most [60,70], but not all [63], studies.
The lower prevalence of retinopathy in nonalbuminuric compared with albuminuric CKD could reflect mechanisms of nonalbuminuric disease apart from microangiopathy, such as macrovascular disease [62,68,71]. However, one study that investigated macrovascular disease as a pathogenic factor in nonalbuminuric CKD by performing renal duplex imaging of the interlobar renal arteries found that resistive index did not differ by albuminuria status [72]. Studies in which protocol biopsies were obtained in patients with normal levels of albumin excretion have revealed various findings, including classic diabetic glomerulopathy, predominantly vascular and tubulointerstitial disease, and nonspecific findings [73,74]. (See 'Pathology' below.)
Progression of diabetic kidney disease appears to be slower in patients with nonalbuminuric disease, regardless of diabetes type [27,60,75]. In the DCCT/EDIC study of patients with type 1 diabetes, the rate of eGFR decline was 1.2 percent per year for those with normoalbuminuria, 1.7 percent per year for those with moderately increased albuminuria, and 5.7 percent per year for those with severely increased albuminuria [27]. Of the 20 people who reached ESKD, all had severely increased albuminuria at the time their eGFR fell below 60 mL/min/1.73 m2. Similarly, in a Finnish cohort of patients with type 1 diabetes, ESKD developed in 63 percent of those with baseline moderately or severely increased albuminuria but only in 1.3 percent of those with nonalbuminuric CKD [60].
Despite having a slower rate of eGFR decline, patients with nonalbuminuric CKD are at high risk for eventual progression if followed over a sufficiently long period of time. In the United States, the crude incidence rate of kidney failure for nonalbuminuric CKD is approximately 7.4 cases per 1000 person-years [44]; in addition, cross-sectional studies have shown that, among patients with diabetes with stage 4 or 5 CKD (eGFR <30 mL/min/1.73 m2), 27 to 30 percent were normoalbuminuric [63,68].
Cardiovascular events and death — Although progression to advanced CKD and ESKD is a major concern, cardiovascular events and death occur more frequently than the need for kidney replacement therapy, particularly in patients with a urine albumin excretion <1000 mg/g of creatinine or an eGFR above 45 mL/min/1.73 m2 [13,14,76-79]. Higher levels of albuminuria (even those below 30 mg/g) and lower levels of eGFR independently and additively increase the risk for cardiovascular events and death [11,78,80-82]. In patients at high cardiovascular risk, the incidence of cardiovascular events is approximately 2.5-fold higher for every 10-fold increase in urine albumin excretion and approximately twofold higher for every halving of eGFR [78,82].
MANIFESTATIONS AND DIAGNOSIS
Manifestations and case detection — The most common clinical abnormalities of diabetic kidney disease are persistently elevated urine albumin excretion (ie, moderately increased or severely increased albuminuria) and/or persistently decreased estimated glomerular filtration rate (eGFR). In severe cases, albuminuria levels can be above the nephrotic threshold of 3.5 g per 24 hours, resulting in the nephrotic syndrome [28,83].
These manifestations are typically asymptomatic and are therefore usually detected through routine, periodic testing. We agree with guidelines from the American Diabetes Association (ADA) and the Kidney Disease: Improving Global Outcomes (KDIGO) organizations that patients with diabetes should undergo annual testing for kidney complications using serum creatinine-based eGFR and urine tests for abnormal levels of albumin excretion [84,85]. Abnormal results should be confirmed by repeat testing over a period of three to six months.
Type 1 diabetes generally presents more conspicuously than type 2 diabetes, which can be asymptomatic for years prior to diagnosis. As a result, testing is recommended to commence in patients with type 1 diabetes five years after diagnosis. In patients with type 2 diabetes, testing for kidney disease is recommended to be undertaken at the time of diagnosis.
Urine albumin excretion can be estimated using a spot urine sample and the albumin-to-creatinine ratio, which is the method most often used in clinical practice, or measured using the historical approach of a timed (24-hour or other) urine collection. At least two of three urine collections obtained over a three- to six-month period should reveal elevated levels of albumin before the presence of albuminuria can be established [84,86]. The basis for this recommendation is that there is significant intraindividual variability in daily urine albumin excretion [87]; in addition, various conditions (eg, uncontrolled hyperglycemia, exercise, urinary tract infection, heart failure exacerbations, high dietary protein intake) can induce transient albuminuria in the absence of abnormal kidney pathology [88].
Qualitative urine albumin dipsticks, which will turn "positive" in the presence of abnormally high albumin concentrations, are not sufficient to determine albuminuria progression or the response to treatment.
GFR can be estimated using several different creatinine-based equations; however, the Chronic Kidney Disease Epidemiology (CKD-EPI) [89] and Modification of Diet in Renal Disease (MDRD) [90] equations are most commonly used (calculator 1 and calculator 2). The advantages and disadvantages of these equations are presented separately. (See "Assessment of kidney function", section on 'Measurement of GFR with plasma clearance'.)
However, these equations were developed in primarily nondiabetic populations and their validity and precision may be reduced in patients with diabetic kidney disease [36,91], uncontrolled hyperglycemia [92], and overweight/obesity [93].
The urine sediment in diabetic kidney disease is usually bland, but patients with severely increased albuminuria may have microscopic hematuria [94,95], and those with nephrotic-range proteinuria often have oval fat bodies or lipid droplets. Dysmorphic red blood cells and red blood cell casts are uncommon in patients with diabetic kidney disease and are therefore suggestive of nondiabetic kidney disease and should prompt a search for glomerulonephritis [96].
The prevalence of hematuria was assessed in a large Taiwanese cohort of 1958 patients with diabetes, most of whom had severely increased albuminuria and eGFR <60 mL/min/1.73 m2 [95]. Microscopic hematuria was present in 48 percent (15 percent had moderate hematuria, defined as ≥5 red blood cells per high-powered field). Kidney biopsy was performed in 111 patients, 61 of whom were diagnosed with classic diabetic glomerulopathy alone. The prevalence of any microscopic hematuria was somewhat lower among those with classic diabetic glomerulopathy compared with those who had nondiabetic kidney disease (59 versus 84 percent), as was the prevalence of moderate hematuria (26 versus 50 percent).
In a meta-analysis of 35 studies and more than 4000 patients with diabetes who underwent urinalysis and kidney biopsy, the sensitivity and specificity of microscopic hematuria for detecting nondiabetic kidney disease were relatively poor (42 and 72 percent, respectively), suggesting that microscopic hematuria alone is insufficient to distinguish diabetic and nondiabetic kidney disease [94]. However, the presence of dysmorphic red blood cells in the urine, while insensitive, was highly specific (94 percent) for nondiabetic kidney disease. Thus, patients with dysmorphic hematuria are likely to have nondiabetic kidney disease, either alone or in combination with diabetic kidney disease.
Diagnosis — Diabetic kidney disease is typically a clinical (or presumptive) diagnosis. (See 'Clinical diagnosis of diabetic kidney disease' below.)
Kidney biopsy is rarely performed to confirm the diagnosis. Rather, a kidney biopsy is most commonly used when an alternative diagnosis (eg, glomerulonephritis or primary nephrotic syndrome) is suspected. (See 'Confirmatory biopsy if diagnosis is uncertain' below.)
Clinical diagnosis of diabetic kidney disease — A clinical diagnosis of diabetic kidney disease can be made if the following conditions are satisfied:
●Persistent albuminuria and/or persistent decreased GFR – As noted above albuminuria is defined as a urine albumin excretion ≥30 mg/day (if measured by a timed collection) or ≥30 mg/g (if using a spot urine albumin-to-creatinine ratio to estimate excretion). Decreased GFR is defined as an eGFR <60 mL/min/1.73 m2 using a creatinine-based formula. Persistence of these abnormalities for at least three months should be confirmed because transient abnormalities, which can be induced by a variety of unrelated disorders, are common. (See 'Albuminuria and decreased glomerular filtration rate' above.)
Albuminuria is not required to make a clinical diagnosis of diabetic kidney disease. A substantial minority of patients with diabetes and decreased eGFR have <30 mg/g of albuminuria, and such patients commonly have histopathologic findings consistent with diabetic kidney disease.
●Long duration of diabetes or established diabetic retinopathy – In patients with type 1 diabetes, five years is a sufficiently long duration.
By contrast, in patients with type 2 diabetes, diabetic kidney disease can be present at the time of diagnosis. Patients with type 2 diabetes are often asymptomatic and, if not receiving annual monitoring of fasting glucose and glycated hemoglobin, can have diabetes for many years before it is discovered. Even those patients who undergo routine monitoring, and therefore have their diabetes diagnosed shortly after onset, have been exposed to prediabetes and hyperglycemia for a prolonged period, and prediabetes is associated with albuminuria or decreased eGFR in a substantial proportion of patients [97]. Thus, at the time of diagnosis, patients with type 2 diabetes are considered to have had the disorder for a duration sufficiently long to produce kidney disease.
In patients with evidence of kidney disease, the presence of proliferative diabetic retinopathy correlates with pathologic features of diabetic kidney disease. Thus, a clinical diagnosis of diabetic kidney disease can be made in patients with established retinopathy, even if the duration of diabetes is short.
Persistently poor glycemic control and poor blood pressure control in these patients further increase the likelihood that kidney disease is due to diabetes.
●A judgment that alternate etiologies are unlikely – Making a presumptive diagnosis of diabetic kidney disease should be avoided if there are features present that suggest an alternative explanation for the kidney disease. Thus, a presumptive clinical diagnosis of diabetic kidney disease should not be made if any of the following apply:
•Severely elevated albuminuria (ie, ≥300 mg/day or mg/g) within five years of onset of type 1 diabetes, or severely elevated albuminuria for many years prior to the onset of type 2 diabetes (if the date of onset is known).
•Red blood cell casts, dysmorphic red blood cells, or white blood cell casts in the urine sediment.
•Presence of another systemic disease that is commonly associated with kidney disease (eg, systemic lupus erythematosus)
•A sudden increase albuminuria or a rapid decline in eGFR; in particular, a sustained rise in albuminuria of greater than 5- to 10-fold that occurs over a period of less than one to two years and an eGFR decline greater than 5 mL/min/1.73 m2 per year are atypical in diabetic kidney disease and should therefore raise suspicion for an alternative etiology.
The diagnosis of diabetic kidney disease is clinical and based upon the presence of albuminuria or reduced eGFR. In general, the cause of chronic kidney disease (CKD) is presumed to be diabetic kidney disease based upon the presence of longstanding diabetes (at least five years duration in the case of type 1 diabetes), particularly if retinopathy is present [86]. Kidney biopsy studies have suggested that the absence of retinopathy and presence of severe (particularly proliferative) retinopathy are strong predictors for nondiabetic and diabetic kidney pathology, respectively [98-101].
The history of glycemic and blood pressure control should also be taken into account as factors predisposing to diabetic kidney disease, although historic levels may be difficult to ascertain in patients with longstanding diabetes [102,103].
When diagnosing diabetic kidney disease, the magnitude of albuminuria and rapidity of eGFR decline should be considered in relation to other clinical factors, and alternative etiologies should be suspected if the pattern of albuminuria and eGFR decline are atypical.
Severely elevated albuminuria prior to five years of type 1 diabetes duration is rare, and this degree of albuminuria is not likely to occur until closer to 10 years' duration if glycemic control has been somewhat reasonable (eg, glycated hemoglobin [HbA1c] <9 percent) [28]. Nephrotic-range proteinuria is atypical in diabetes of less than 10 to 15 years' duration [28].
Sudden increases in the magnitude of albuminuria are uncommon in diabetic kidney disease. Although data are scarce, it is the opinion of the authors that a sustained rise in albuminuria of greater than 5- to 10-fold that occurs rapidly (over a period of less than one to two years) is unusual in diabetic kidney disease.
Sudden decreases in eGFR would also be reason to consider other causes of kidney disease. As described above, the rate of eGFR decline is quite variable in diabetes and can be as rapid as 5 mL/min/1.73 m2 per year. Deterioration more rapid than this rate should raise concern about other etiologies of kidney disease, particularly when accompanied by other atypical changes in albuminuria magnitude or the presence of dysmorphic hematuria.
Microscopic hematuria is common in patients with diabetic kidney disease and is most often identified in patients with severe albuminuria (albumin-to-creatinine ratio >1000 mg/g) and low eGFR [95]. However, red blood cell casts or dysmorphic red cells are uncommon and should raise the possibility of a glomerulonephritis [96]. Alternative diagnoses such as tubulointerstitial diseases should also be suspected if white blood cell casts are present.
The sensitivity of albuminuria and low eGFR for detection of early diabetic kidney disease is poor [104]. The limitations preclude early detection and prognostication of diabetic kidney disease and pose a challenge for the development of new therapeutics to prevent progression of early-stage disease. As a result, various circulating and urinary biomarkers are being developed for use as diagnostic and prognostic tools [105-118] but are not yet in clinical use.
Confirmatory biopsy if diagnosis is uncertain — In patients with diabetes with severe albuminuria or decreased eGFR, a kidney biopsy should usually be performed if nondiabetic kidney disease is suspected; reasons to suspect an alternate etiology are presented above. (See 'Clinical diagnosis of diabetic kidney disease' above.)
The histologic findings consistent with diabetic nephropathy are discussed below. (See 'Pathology' below.)
While the primary purpose of kidney biopsy is considered to be diagnostic, it also offers useful prognostic information. The glomerular disease class and degree of interstitial fibrosis yield information regarding time to end-stage kidney disease (ESKD) [119,120]. Acute tubular necrosis is common among patients with diabetes and may or may not be reversible [101,121].
In many regions, the health care culture is to avoid kidney biopsy in patients with diabetes, but there is concern that many patients with nondiabetic kidney disease go undiagnosed [101,122]. One study, for example, examined health centers that had a biopsy policy for patients with diabetes that was "restricted" (ie, a kidney biopsy was only performed if a nondiabetic pathology was suspected based upon clinical findings) and centers that had an "unrestricted" policy (kidney biopsy was performed if there was severe albuminuria, low eGFR, or hematuria) [123]. Even when biopsies were restricted to patients in whom an alternate diagnosis was suspected, 29 percent had diabetic kidney disease alone (with no other diagnosis). Conversely, when biopsies were unrestricted (and performed among patients with suspected diabetic kidney disease), 33 percent had an alternate diagnosis that might have otherwise been missed if a biopsy was not performed.
As another example, in a single-center cohort of 611 patients who underwent kidney biopsy, one-fourth were performed in patients with diabetes [121]. Of these, one-third revealed classic diabetic glomerulopathy alone, one-third showed diabetic glomerulopathy plus nondiabetic kidney disease, and one-third showed nondiabetic kidney disease alone. The most common nondiabetic kidney diseases identified were acute tubular necrosis (28 percent), immune-mediated glomerular diseases (25 percent), hypertensive nephrosclerosis (18 percent), and focal segmental glomerulosclerosis (18 percent; nearly all of which was due to secondary causes). In the United States, the majority of kidney biopsies performed in people with diabetes occur at late stages of disease, when median eGFR is 20 to 30 mL/min/1.73 m2 and median albuminuria is 3000 to 5000 mg/g [120,121]. This means that patients with diabetes with nondiabetic kidney disease are often not diagnosed until their disease is severely advanced.
Pathology — Patients with type 1 diabetes predominantly develop classical diabetic glomerulopathy, whereas patients with type 2 diabetes, particularly those without albuminuria, may have a myriad of pathologic findings in the kidney.
The initial abnormality of classical diabetic glomerulopathy is thickening of the glomerular basement membrane, which can occur as soon as two years following diagnosis of type 1 diabetes (picture 1 and picture 2) [124]. Other major diabetic glomerular changes include mesangial expansion, which can be diffuse or nodular (often termed "Kimmelstiel-Wilson nodules"), podocyte injury, and glomerular sclerosis (picture 1 and picture 2) [125,126]. The mesangial expansion and glomerulosclerosis do not always develop in parallel, suggesting that they may have different underlying pathogenesis [62]. Arteriolar hyalinosis and arteriosclerosis of larger vessels are common [127], likely representing the combined effect of hyperglycemia and hypertension (picture 3). Tubulointerstitial fibrosis usually occurs after the initial glomerular lesions and is a final pathway-mediating progression to advanced CKD and ESKD. As above, pathologic abnormalities in type 2 diabetes are more heterogeneous than in type 1 diabetes, with a large proportion of patients having dominant vascular and tubulointerstitial disease more than glomerular involvement [8,74,127].
Patients with diabetes who have kidney disease may also have other histopathologic findings, especially in those with type 2 diabetes and normal levels of urine albumin excretion (see 'Nonalbuminuric diabetic kidney disease' above). Research biopsies performed in patients with type 1 diabetes and albuminuria <30 mg/day found changes consistent with classic diabetic glomerulopathy (glomerular basement membrane thickening and mesangial expansion), particularly among those with GFR <90 mL/min/1.73 m2 [73]. Conversely, research biopsies from eight individuals with type 2 diabetes and nonalbuminuric CKD revealed classical diabetic glomerulopathy with predominant mesangial expansion and glomerular basement membrane thickening in three patients, predominant vascular and tubulointerstitial disease in three patients, and mild, nonspecific changes in two patients [74]. This is in contrast to five of six patients with type 2 diabetes, CKD, and moderately increased albuminuria who were observed to have typical diabetic glomerulopathy.
Renal Pathology Society classification — A classification of type 1 and type 2 diabetic kidney disease was developed by the research committee of the Renal Pathology Society [128], which provides a systematic and uniform approach to classify pathology of the diabetic kidney and also promotes the study of pathogenesis and prognosis of disease. Scores are assigned to all three kidney compartments: the glomeruli; vasculature; and interstitium:
●Class I – Isolated glomerular basement membrane thickening. Basement membranes are greater than 430 nm in males and 395 nm in females over the age of nine years. There is no evidence of mesangial expansion, increased mesangial matrix, or global glomerulosclerosis involving >50 percent of glomeruli.
●Class II – Mild (Class IIa) or severe (Class IIb) mesangial expansion. A lesion is considered severe if areas of expansion larger than the mean area of a capillary lumen are present in >25 percent of the total mesangium.
●Class III – At least one Kimmelstiel-Wilson lesion (nodular intercapillary glomerulosclerosis) is observed on biopsy, and there is <50 percent global glomerulosclerosis.
●Class IV – Advanced diabetic glomerulosclerosis. There is >50 percent global glomerulosclerosis that is attributable to diabetic nephropathy.
The clinical and prognostic significance of these glomerular classes has been evaluated in several retrospective cohort studies. Progression to ESKD is less common in classes I and IIa and more rapid in class IV. However, it is less clear whether or not class III (nodular glomerulosclerosis) carries a worse prognosis than class IIb (diffuse glomerulosclerosis) [119,120]. A limitation of this classification is that potentially important pathologic lesions are not included, such as the presence of focal and segmental sclerosis, mesangiolysis, capillary aneurysms, exudative lesions, and extracapillary hypercellularity [120,129].
Differential diagnosis — In patients with diabetes, albuminuria and reduced eGFR can result from nondiabetic kidney disease. Reasons to perform a kidney biopsy to detect nondiabetic kidney disease are presented above. (See 'Clinical diagnosis of diabetic kidney disease' above and 'Confirmatory biopsy if diagnosis is uncertain' above.)
The differential diagnosis of albuminuria and reduced eGFR is large. A detailed discussion is presented elsewhere. (See "Glomerular disease: Evaluation and differential diagnosis in adults".)
Conditions other than diabetes can cause nodular glomerulosclerosis, most of which are identified by characteristic findings on immunofluorescence or electron microscopy [126]. These include the following:
●Dysproteinemias such as amyloidosis and monoclonal immunoglobulin deposition diseases (MIDD), mostly kappa light chain deposition disease (see "Renal amyloidosis")
●Organized glomerular deposition diseases, fibrillary and immunotactoid glomerulonephritis, fibronectin glomerulopathy, and collagen III glomerulopathy (see "Glomerular diseases due to nonamyloid fibrillar deposits" and "Glomerulopathy with fibronectin deposits")
●Chronic hypoxic or ischemic conditions [130], such as cyanotic congenital heart disease, Takayasu arteritis with renal artery stenosis, or cystic fibrosis [131]
●Chronic membranoproliferative glomerulonephritis (type I) (see "Membranoproliferative glomerulonephritis: Classification, clinical features, and diagnosis")
●Idiopathic nodular glomerulosclerosis, which is often associated with smoking and hypertension [132]
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Diabetes mellitus in adults" and "Society guideline links: Chronic kidney disease in adults" and "Society guideline links: Diabetic kidney disease".)
SUMMARY
●"Diabetic nephropathy" is characterized pathologically by glomerular basement membrane thickening, endothelial damage, mesangial expansion and nodules, and podocyte loss. However, it has become clear that there are various forms of kidney disease due to diabetes including nonclassical glomerular lesions as well as tubulointerstitial disease. Thus, the term "diabetic kidney disease" is used to reflect the presence of albuminuria, decreased estimated glomerular filtration rate (eGFR), or both, but it is not intended to indicate a specific kidney disease phenotype in a patient with diabetes. (See 'Terminology' above.)
●Clinical outcomes in diabetic kidney disease are widely variable, and there has been an evolution in what was once thought to be a single and unidirectional natural history. Historically, it was believed that patients began with either normal or elevated GFR (ie, "hyperfiltration") and that moderately increased albuminuria was the earliest clinically detectable biomarker of classical diabetic glomerulopathy (figure 1). A decline in eGFR to below 60 mL/min/1.73 m2 was thought to occur after the development of moderately and then severely increased albuminuria. However, contemporary studies have shown that albuminuria may regress, even from the severely increased range. In addition, decreased eGFR can occur and progress to advanced stages of chronic kidney disease (CKD) before the onset or without the development of albuminuria. Although progression to advanced CKD and end-stage kidney disease (ESKD) is a major concern, cardiovascular events and death occur more frequently than the need for kidney replacement therapy. (See 'Natural history' above.)
●The most common clinical abnormalities of diabetic kidney disease are persistently elevated urine albumin excretion (ie, moderately increased or severely increased albuminuria) and/or persistently decreased eGFR. In severe cases, albuminuria levels can be above the nephrotic syndrome threshold of 3.5 g per 24 hours. Early manifestations are typically asymptomatic and are therefore usually detected through routine, periodic testing. Abnormal results should be confirmed by repeat testing over a period of at least three months. (See 'Manifestations and case detection' above.)
●The urine sediment in diabetic kidney disease is usually bland, but patients with severely increased albuminuria may have microscopic hematuria, and those with nephrotic-range proteinuria often have oval fat bodies or lipid droplets in the sediment. Dysmorphic red blood cells and red blood cell casts are uncommon in patients with diabetic kidney disease and therefore suggest nondiabetic kidney disease, particularly glomerulonephritis. (See 'Manifestations and case detection' above.)
●Diabetic kidney disease is typically a clinical (or presumptive) diagnosis. Kidney biopsy is rarely performed to confirm the diagnosis in clinical practice. Rather, a kidney biopsy is most commonly used when an alternative diagnosis (eg, glomerulonephritis or primary nephrotic syndrome) is suspected.
A clinical diagnosis of diabetic kidney disease can be made if the following conditions are satisfied (see 'Clinical diagnosis of diabetic kidney disease' above):
•Persistent albuminuria and/or persistent decreased GFR. Albuminuria is not required to make a clinical diagnosis of diabetic kidney disease. A substantial minority of patients with diabetes and a decreased eGFR have <30 mg/g of albuminuria, and such patients commonly have histopathologic findings consistent with diabetic kidney disease.
•Long duration of diabetes or established diabetic retinopathy. In patients with type 1 diabetes, five years is a sufficiently long duration. By contrast, in patients with type 2 diabetes, diabetic kidney disease can be present at the time that diabetes is diagnosed. In patients with evidence of kidney disease, the presence of proliferative diabetic retinopathy correlates with pathologic features of diabetic kidney disease. Thus, a clinical diagnosis of diabetic kidney disease can be made in patients with established retinopathy, even if the duration of diabetes is short.
•A judgment that alternate etiologies are unlikely. Making a presumptive diagnosis of diabetic kidney disease should be avoided if there are features present that suggest an alternative explanation for the kidney disease. Thus, a presumptive clinical diagnosis of diabetic kidney disease should not be made if any of the following conditions apply:
-Severely elevated albuminuria (ie, ≥300 mg/day or mg/g) within five years of onset of type 1 diabetes or severely elevated albuminuria for many years prior the onset of type 2 diabetes (if the date of onset is known).
-Red blood cell casts, dysmorphic red blood cells, or white blood cell casts in the urine sediment.
-Presence of another systemic disease that is commonly associated with kidney disease (eg, systemic lupus erythematosus).
-A sudden increase in albuminuria or a rapid decline in eGFR; in particular, a sustained rise in albuminuria of greater than 5- to 10-fold that occurs over a period of less than one to two years and an eGFR decline greater than 5 mL/min/1.73 m2 per year are atypical in diabetic kidney disease and should raise suspicion for an alternative etiology.
●In patients with diabetes with severe albuminuria or decreased eGFR, a kidney biopsy should usually be performed if nondiabetic kidney disease is suspected. (See 'Confirmatory biopsy if diagnosis is uncertain' above.)
●Conditions other than diabetes can cause nodular glomerulosclerosis, most of which are identified by characteristic findings on immunofluorescence or electron microscopy [126]. These include dysproteinemias, organized glomerular deposition disease, chronic hypoxic conditions, chronic membranoproliferative glomerulonephritis, and idiopathic nodular glomerulosclerosis. (See 'Differential diagnosis' above.)
آیا می خواهید مدیلیب را به صفحه اصلی خود اضافه کنید؟