ASIAN PACIFIC SOCIETY OF NEPHROLOGY CLINICAL PRACTICE GUIDELINE ON DIABETIC KIDNEY DISEASE

Abstract

The International Diabetes Federation (IDF) estimated that 425 million people were living with diabetes mellitus (DM) in 2017 (1), which had far exceeded what was originally predicted in 2003 (333 million people by 2025). Consequently, the IDF provided a new projection of a near doubling of 629 million people with DM in 2045, with 4 out of 5 people with DM living in low- and middle-income countries (LMICs), majority from South East Asia (82 million) and the Western Pacific (159 million) regions. The increasing prevalence of DM has contributed to the growing burden of ESKD, and it is estimated that DKD is responsible for about 50% of ESKD in the developed world (2). The global burden of DKD and ESKD has a major impact on healthcare costs and resources; making screening, early detection and preventive treatment important strategies to mitigate this worldwide pandemic. Although the prevalence of DM is well reported, the prevalence of DKD in the Asia-Pacific region is not well documented. Evidence suggests that among people living with type 2 DM in Asia-Pacific, the estimated prevalence of moderately increased albuminuria (or previously referred to as microalbuminuria) was 17.0-18.2%, while severely increased albuminuria (or traditionally known as overt proteinuria) and reduced estimated glomerular filtration rate (eGFR) were in the order of 2.1-14.4% and 15.3-61.6% respectively (3, 4). However, there is little or no report of the prevalence of DKD in people suffering from type 1 DM in the Asia-Pacific countries. Practice Point 1.1: Use the term "Diabetic Kidney Disease" instead of "Diabetic Nephropathy" to reflect the pathophysiology, histopathological heterogeneity and clinical course of the condition. The terms diabetic nephropathy (DN) and DKD are often used interchangeably. The concept of DKD was first proposed by the National Kidney Foundation's Kidney Disease Outcomes Quality Initiative (NKF/KDOQI) in 2007 (5), which was subsequently adopted by Kidney Disease: Improving Global Outcomes (KDIGO) (6) and the American Diabetes Association (ADA) (2). DN is the classical term used to describe a complication observed in the glomerulus as a consequent of hyperglycaemia due to DM (7), eventually resulting in overt proteinuria, hypertension and renal oedema. On the other hand, DKD is proposed to indicate a presumptive diagnosis of kidney disease caused by DM. While DN or diabetic glomerulopathy can only be diagnosed definitively by a kidney biopsy, DKD could exist for a long period of time before the traditional indications for a kidney biopsy develop (8). Moreover, DKD can become clinically apparent through careful screening in people with DM without the need for a kidney biopsy. In addition, lesions beyond diabetic glomerulosclerosis, such as tubulointerstitial lesions, are of equally important prognostic significance. It has also been shown that not all people with DKD will demonstrate pathological features of DN on kidney biopsy (9). Biopsy series performed in people with DKD from type 2 DM demonstrated heterogeneous biopsy findings including DN alone, non-diabetic kidney lesions alone and non-diabetic kidney lesions on a background of DN (10, 11). Therefore, ascribing the term DKD to this condition, rather than DN, would accrue greater clinically relevance and epidemiological accuracy. Clinical Recommendation 1.1: We recommend that urine albumin:creatinine ratio and estimated glomerular filtration rate be used for the screening of diabetic kidney disease in adults with diabetes mellitus. [1B] KDIGO had divided albuminuria categories into A1, A2 or A3, based on the quantity of urinary albumin loss (12) - normal-to-mildly increased albuminuria (A1) of < 30 mg/g creatinine; moderately increased albuminuria (A2; previously referred to as microalbuminuria) of 30-300 mg/g creatinine; and severely increased albuminuria (A3) of > 300 mg/g creatinine. Such categorization, when considered together with eGFR categories, has prognostic implications for the development and progression of DKD. Moderately increased levels of albuminuria is a well-recognized independent risk factor for the development of CKD (13, 14) and cardiovascular morbidities and mortality (15-17). Due to its molecular size, albumin is invariably filtered from the glomeruli, whereas total urinary protein consists of a variety of protein derived from the glomerular filtrate as well as secretory and excretory proteins from the renal tubules. Therefore, the use of uACR, with the ability to detect moderately increased levels of albuminuria, is often considered a more sensitive marker over urine protein:creatinine ratio (uPCR) for the detection of early DKD (18). Although a 24-hour urine collection was originally accepted as the gold standard for urinary albumin measurement, it is inconvenient and its accuracy is often flawed by inadequate or missed collections. On the other hand, the random measurement of uACR has an inherent day-to-day variability of about 30-40% (19), and may be elevated by factors independent of kidney disease, including strenuous exercise, urinary tract infections, menstruation, heart failure, marked hyperglycaemia and drugs e.g. non-steroidal anti-inflammatory drugs (NSAIDs). Consequently, while an early morning assessment of the uACR is commonly used as a screening tool for the detection of DKD, an elevated uACR should be confirmed in the absence of a urinary tract infection with 2-3 additional early-void urine specimens collected 3 to 6 months apart. In addition, 24-hour urine collections can be performed when uACR vary widely (20). Approximately 10-30% of people with both type 1 and type 2 DM develop kidney function decline before the onset of moderately increased or severely increased albuminuria (21, 22). Studies have shown that people with DM and normoalbuminuria may have progressive kidney insufficiency (eGFR < 60 ml/min/1.73m2), referred to as normoalbuminuric DKD (23). Moreover, this phenomenon is not a rare occurrence. In the UKPDS (UK Prospective Diabetes Study), only 51% of those who developed an estimated creatinine clearance of < 60 ml/min/1.73m2 ever tested positive for albuminuria (24). In addition, in the ADVANCE (Action in Diabetes and Vascular Disease: Preterax and Diamicron MR Controlled Evaluation) trial, which included 10,640 individuals with type 2 DM, a total of 62% of people with eGFR < 60 ml/min/1.73m2 had normoalbuminuria (16). Besides, the absence of albuminuria in this group of people with reduced eGFR did not confer a prognostic benefit and progression to advanced CKD had been described (25, 26). The RIACE (Renal Insufficiency and Cardiovascular Events) Italian multicentre study showed that normoalbuminuric CKD is also a strong predictor of mortality, thus supporting a major prognostic impact of kidney dysfunction irrespective of albuminuria (27). The eGFR can be calculated using either the Modification of Diet in Renal Disease (MDRD) equation (28, 29) or the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula (30). Both equations perform well when the GFR is < 60 ml/min/1.73 m2, but the CKD-EPI is more accurate at higher levels of kidney function. In a study that compared the performance of CKD-EPI and MDRD equations in estimating GFR in a large cohort of people with DM and various degrees of albuminuria, the CKD-EPI equation was noted to be a superior surrogate marker of GFR in people with normoalbuminuria and hyperfiltration (31). Since many people with early stages of DKD may have high-normal GFRs, the CKD-EPI may have a utility as a screening tool for early DKD. In addition, progressive DKD is best determined by the slope of sequential eGFR measurements, rather than a single estimate, while other markers of kidney damage (e.g. albuminuria) are additionally required to detect early stages of DKD with high to normal eGFR. Although eGFR is useful in assessing progressive changes in kidney function, it should not be used in situations where the kidney function is changing rapidly such as in acute kidney injury (AKI). When conditions that could cause AKI, such as infections, hypotension or volume depletion, are present, while assessment of the level of kidney function using eGFR may be necessary, it should not be used to determine or categorize the stages of CKD. Because kidney function can be transiently depressed, reduction in eGFR is required to be persistent before it is considered to be abnormal, and KDIGO recommends a persistent reduction in eGFR for a period of at least 90 days for staging CKD (12). Therefore, both eGFR and uACR are not only important determinants of DKD, but are independent predictors of cardiovascular and renal outcomes in individuals with DM. Screening people with DM for DKD using a combination of eGFR and uACR allows for early detection, risk stratification, directed intensified interventions and prognostic determination. Practice Point 1.2: In areas where urine albumin:creatinine ratio is not feasible or available, screening with spot urine albumin concentration, urine protein:creatinine ratio or urine dipstick may be considered. Despite the benefits of uACR in screening for early DKD, the availability and cost of the test may limit its use in LMICs. In the absence of uACR, it is the judgement of the WG that some form of albuminuria detection would still be preferable in screening for early DKD in people with DM. A study conducted in an Indo-Asian population in Pakistan compared the use of spot urine albumin concentration and uACR in detecting albuminuria (32). The authors found that both tests are comparable in screening for albuminuria, with similar areas under the receiver operator characteristic (ROC) curves across both methods and gender [spot urine albumin: men – 0.88 (0.84-0.92), women – 0.86 (0.82-0.9); uACR: men – 0.90 (0.86-0.93), women – 0.86 (0.82-0.89)]. A cross-sectional study performed on 216 people with DM at a general hospital in Sri Lanka demonstrated that total uPCR correlated well with spot uACR for the detection of moderately increased levels of albuminuria (30-300 mg/g creatinine) [R2 = 0.798; p<0.001] (33). Similarly in Japan, where the cost of uACR measurements is 16-fold higher than uPCR, a study among people with DM showed a significant positive correlation between uPCR and uACR, with uPCR measurements being able to detect 90% of cases of both moderately and severely increased albuminuria (34). A ROC analysis demonstrated that uPCR measurement had a sensitivity and specificity of 90.8% and 91.9% respectively, for the detection of albuminuria with a cut off value of 0.091 g/g creatinine. These results suggest that in areas where uACR is not available, uPCR can be considered as a screening tool for moderately increased albuminuria and could be a more cost-effective method in detecting early DKD. Moreover, larger randomized controlled trials (RCTs) have shown convincing prognostic value of uPCR for both long-term renal and cardiovascular outcomes. Likewise, improvement in analysis technique has enhanced the accuracy of uPCR in detecting a total urinary protein concentration of less than 0.5 g/g creatinine. The correlation of urine dipstick for albuminuria to uACR was reported in a study from India, where urine dipstick examinations were performed by trained technicians in 500 consecutive random urine samples of individuals with type 2 DM (35). The results were read visually as "negative" or as "positive" indicated as trace, 1+, 2+ or greater, representing a urine protein concentration of 0.15 g/L, 0.15-0.30 g/L, > 0.30 g/L or > 1.0 g/L respectively. 459 urine samples without proteinuria were evaluated for the presence of moderately increased albuminuria with a uACR and were used to determine the sensitivity and specificity of the dipstick to detect a negative or positive result. The authors demonstrated that the urine dipstick test with a reading of trace was highly specific at 98% and fairly sensitive at 68% in determining the presence of moderately increased albuminuria, with a positive and negative predictive value of 92% and 88% respectively. This suggests the possibility of using urine dipstick as an inexpensive tool in screening for DKD in countries where uACR is not available or economically plausible. A cross-sectional study from Tunisia suggested a slightly lack lustre benefit though the authors compared the urine dipstick to a 24-hour uACR as a screening tool for early DKD (36). 182 individuals with type 2 DM were recruited from 7 primary care centres, with 23% found to have moderately increased albuminuria by the 24-hour uACR. The Micral urine strip was used in the study and utilized a visual analogue scale – white (normoalbuminuria), light pink (albumin 20 mg/L), dark pink (albumin 50 mg/L) and very dark pink (albumin 100 mg/L). When compared to the 24-hour uACR, the urine dipstick had a sensitivity, negative and positive predictive value of 77%, 46% and 88% respectively. The overall ROC was 0.72 (95% CI 0.58-0.86, p=0.001) for the entire population, but this improved to 0.80 if cardiovascular parameters were taken into consideration. A small systematic review of 4 studies (37) using urine dipstick as a screening tool for DKD concluded that the Micral test has a high sensitivity but not specificity, and a lower positive predictive value in detecting moderately increased albuminuria, thereby, suggesting that it is not as effective as a laboratory comparator. Therefore, while uACR is the recommended screening tool for early detection and risk evaluation for DKD, some guidelines recommend the use of uPCR for monitoring the progression of established DKD and non-diabetic CKD. However, in a rural setting or in LMICs, rather than forgoing any form of albuminuria detection for DKD screening, spot urinary albumin concentration, uPCR or urine dipstick can be considered as an alternative screening method. Clinical Recommendation 1.2: We suggest that annual screening for diabetic kidney disease be performed beginning 5 years after diagnosis of type 1 diabetes mellitus and at diagnosis of type 2 diabetes mellitus. [2C] In people with newly diagnosed type 1 DM, a transient increase in albuminuria is thought to represent acute metabolic disturbances and usually resolves with glycemic control. The course of DKD in people with type 1 DM has been better documented and is more predictable. Many longitudinal cohorts reported significant increases in the prevalence of moderately increased albuminuria occurring only 5 years after diagnosis (38-40). Without specific interventions, up to 80% with sustained moderately increased albuminuria (≥ 30 mg/day) would develop severely increased albuminuria (> 300 mg/day) over 10-15 years, while in those with severely increased albuminuria, ESKD becomes established in 50% within 10 years and in greater than 75% of people by 20 years (41). On the other hand, elevated albuminuria was already detected at diagnosis in 6.5% of people with type 2 DM, suggesting that the presentation of DM had been delayed for an estimated period of 8 years from the onset of hyperglycemia. Thereafter, it is estimated that approximately 2% per year will progress from normoalbuminuria to moderately increased albuminuria and from moderately increased to severely increased albuminuria. At a median duration of 15 years after diagnosis, about 40% of affected individuals will develop albuminuria and 30% will have a reduced eGFR < 60 ml/min/1.73m2 or doubling of serum creatinine level (42). Table CR1.2 summarizes the recommendations on screening for DKD in published guidelines. Based on the epidemiology of DKD in type 1 and type 2 DM and taking into consideration cost-effectiveness, resource availability and healthcare infrastructure in the Asia-Pacific region, it is the judgement of the WG that annual screening for DKD in both types of DM should be performed, consistent with other published guidelines. However, in people where glucose control is extremely poor and where resources permit, more frequent screening may be required. Practice Point 1.3: Kidney biopsy is not routinely performed to diagnose diabetic kidney disease. However, kidney biopsy should be considered in individuals with diabetes mellitus if there is evidence suggesting the presence of non-diabetic kidney disease. (Table PP1.3) Kidney involvement in people with DM could be DKD, NDKD (non-diabetic kidney disease) or a combination of both conditions. Increasing evidence indicates that many individuals with DM had been erroneously labeled as having progressive forms of DKD when instead were developing NDKD or 'mixed' conditions, where typical features of DKD overlapped with other kinds of histological damage (10, 11). The prevalence of NDKD among people with DM varies widely depending on the regional and/or racial variations in different study populations, the selection criteria and the indications for kidney biopsy (46-48). A meta-analysis of kidney biopsies in individuals with DM showed that the prevalence of DKD was extremely variable across reports and ranged from 7 to 94% of all cases (49). Similar variability was seen with NDKD (3–83%) and mixed forms (4–46%), of which NDKD included IgA nephropathy [IgAN] (3–59%), membranous nephropathy [MN] (7–35%), focal segmental glomerulosclerosis [FSGS] (17–38%), and acute interstitial nephritis [AIN] (18–49%). The predominant types of NDKD among people with DM varies across different countries and ethnicities. MN and IgAN were the two most common histopathological findings in Asian (50-52) and Croatian populations (53). A recent Chinese study with the largest number of kidney biopsy findings across multiple centres reported similar findings of MN and IgAN as the two leading NDKD diagnoses among individuals with DM (46). However, FSGS was a much more common finding in developed countries, such as the USA and New Zealand (47, 54), reflecting the similar spectrum of glomerulonephritis in the same areas. In contrast, AIN was the leading cause of NDKD in Malaysia and India (55, 56), while another series reported a high incidence of acute tubular necrosis (47). Such disparity may, however, reflect different selection criteria and study design, reporting and and different biopsy criteria The between DKD and NDKD is important as the treatment and for both conditions In ESKD rate was found to be in the group with only DKD, in those with both DKD and NDKD and in individuals with only NDKD on the biopsies However, the predictive value of clinical parameters for NDKD had been with no clinical found to be superior in the diagnosis of NDKD before a kidney biopsy Although the lack of diabetic used to be considered an of a high for a cause of kidney disease, studies have shown that this is not the in people with type 2 DM and kidney disease. A study of and people with type 1 and type 2 DM respectively, showed that while there to be some correlation between the presence of diabetic and albuminuria with type 1 DM, this was not observed in type 2 DM, with of individuals having severely increased albuminuria and no evidence of This was also seen with increased albuminuria and in individuals with type 2 DM, suggesting that the absence of in people with type 2 DM not similar of kidney including DKD. Similar were also found in biopsy series between and histological findings of DKD among individuals with type 2 DM. In a analysis of people with type 2 DM, an evaluation of available kidney biopsies found that only 15 of individuals with histological features of DKD had diabetic while of the study with NDKD had to a negative predictive value of only 40% for the absence of diabetic in DKD The same group on to the kidney biopsies of among people with type 2 DM, persistent albuminuria and no and found histological features of DKD in only of the biopsy with no correlation to clinical or laboratory the of kidney biopsy in between DKD and These findings indicated that is not an of nephropathy while the presence of is a poor prognostic factor for kidney function. Therefore, the absence of diabetic in an with DKD should the of the kidney disease is due to a other than DM, in people with type 1 DM. However, if there are no other supporting or from the or laboratory there usually is little for biopsy. It is the judgement of the WG that while a kidney biopsy is not to diagnose DKD when classical features of the are present, a kidney biopsy should be considered when may suggest the presence of a Table summarizes the in which a kidney biopsy may be in people with DM. Clinical Recommendation 1.3: We suggest that in adults with diabetes mellitus and albuminuria, monitoring of albuminuria and estimated glomerular filtration rate be performed to the progression of diabetic kidney disease. is often the first of the onset of DKD and the progression of kidney damage in people with DM In the which on type 1 DM, compared to people who had albuminuria, those with moderately increased albuminuria that had with sustained moderately increased albuminuria and severely increased albuminuria were with an increased ratio of reduced eGFR of (95% CI (95% CI and (95% CI respectively Similarly in people with type 2 DM, moderately increased albuminuria was found to be with a increased risk of progression to overt independent of the duration of DM and and while a study in an Asian population noted that a random spot uACR predicted increased mortality, progression of albuminuria and of kidney function The UKPDS study found that progression from normoalbuminuria to moderately increased albuminuria at a rate of per while that of moderately increased to severely increased albuminuria and severely increased albuminuria to increased serum creatinine or at and per year respectively persistent albuminuria urinary albumin rate increases by approximately per year and eGFR to decline at a rate of about 10 ml/min/1.73m2 when urinary albumin rate 100 with an progression to ESKD in the population Consequently, monitoring both the urinary albumin and eGFR will people who are much more rapidly than where treatment could be intensified and the of to or the of kidney function be In the the risk for kidney increased for doubling of albuminuria level and albuminuria by at months with treatment the risk for kidney by more than 50% CI This was a similar finding in a which showed that the treatment on albuminuria correlated with the development of ESKD, where for 30% reduction in albuminuria, the risk for ESKD by (95% CI p=0.001) Moreover, this correlation was consistent of and while interventions have been shown to be effective in the progression to ESKD, this can only be through monitoring of albuminuria levels and eGFR in people with established DKD. Practice Point In people with established diabetic kidney disease, albuminuria and estimated glomerular filtration rate should be at least 6 should be performed more if resources or in people who are at high of progression to (Table The and of monitoring DKD progression have not been are no studies that have evaluated the of monitoring that would impact clinical outcomes or It the judgement of the that in those with established DKD, monitoring should be performed at least 6 that more frequent monitoring could be a in However, where resources permit, it is suggested that monitoring should be performed 2-3 In addition, Table the clinical situations where the more frequent monitoring is considered by the WG to be DKD is considered a major that approximately of individuals with DM with DKD are at increased risk of cardiovascular disease, progression to ESKD and Therefore, interventions to or the progression of DKD is very In the general population, as well as in individuals diagnosed with DKD, to a that to and from has been with better from cardiovascular and This will review the evidence from epidemiological and studies from Asia and other and recommendations the impact of factors on DKD incidence and Clinical Recommendation We recommend that people with diabetic kidney disease who are for Clinical Practice for Chronic Kidney Disease: and as a risk factor for disease progression in CKD. A large clinical trial, the men and found that was with an increased risk for studies have demonstrated that in people with DM the development of moderately increased albuminuria and DKD. In a the progression of DKD was increased among Moreover, the authors showed that as a risk factor for progressive DKD, was independent of duration of DM and This was a similar in other studies where there was a progression of DKD controlled and the use of in people with type 2 DM. In Asian was with higher mortality among people with DKD increased risk for DKD and greater disease progression of DKD a study had also reported that increased the risk of developing DM, and the and there are no to it is the judgement of the WG that all people with DKD be to Studies from the Asia-Pacific region are in Table Studies from populations with DM in and showed that is with a higher risk for and higher incidence and progression of DKD. Such correlated with the duration and to people with type 2 DM men with type 2 DM and DKD who for at least 1 with people who had more decline in eGFR = and tubular and interstitial on the kidney biopsy = In the no significant risk factor was shown about National and 629 men diagnosed with DM Diabetes included of the countries in the world with the from with at number and India at number two Consequently, it is that the burden of DKD would become a significant disease burden in countries. While is an many countries in the Asia-Pacific region, in resource not have a or on by the to as of its DKD and are Practice Point the of on diabetic kidney disease and general and the benefits of should be available to all people with DM. Clinical Recommendation We recommend that people with diabetic kidney disease a normal and have

Topics & Concepts