Worldwide, chronic kidney disease (CKD) attributable to diabetes is the single commonest cause of end-stage renal disease requiring renal replacement therapy. Two main forms are recognised. Classical diabetic nephropathy develops over many years. Urine albumin excretion increases gradually, accompanied by rising blood pressure and cardiovascular risk. Glomerular filtration begins to fall relatively late, but eventually end-stage renal disease occurs. Histologically, these individuals have diabetic glomerulosclerosis, with thickening of the glomerular basement membrane, mesangial expansion, podocyte changes and loss, and glomerulosclerotic nodules. Interstitial fibrosis occurs later, but is an important additional factor contributing to loss of glomerular function (see histology of chronic kidney disesae in diabetes). The non-classical presentation of diabetic nephropathy occurs in individuals in whom glomerular filtration declines progressively with no or very minimal increase in urine albumin excretion and without evidence of any other specific kidney disease. This form of CKD in diabetes is usually attributable to a combination of hypertension, renovascular disease and perhaps obesity, and is therefore more typical of type 2 diabetes.
Who develops chronic kidney disease in diabetes?
The glomerular filtration barrier in healthType 1 diabetic patients with any degree of classical nephropathy are more likely to be male, have poorer glycaemic control, higher blood pressure, more marked dyslipidaemia, and more marked insulin resistance, endothelial dysfunction and cardiovascular disease than their normoalbuminuric peers. The other microvascular complications of diabetes are also commoner in individuals with diabetic nephropathy. There is a genetic predisposition to classical diabetic nephropathy and its accompanying cardiovascular disease. The vast majority of people with Type 1 diabetes and chronic kidney disease have classical nephropathy, with only around 5 % having non-classical disease.
In contrast, in Type 2 diabetes, approximately two-thirds of individuals with chronic kidney disease have non-classical diabetic kidney disease. Non-classical chronic kidney disease is much less well studied, but several reports indicate that it is difficult clinically to distinguish individuals with Type 2 diabetes who will have classical diabetic glomerulosclerosis on biopsy from those with atypical disease. However, they are more likely to be female.
What causes chronic kidney disease in diabetes?
The glomerular filtration barrier in diabetic glomerulosclerosis, demonstrating greater basement membrane width, and widened and effaced podocyte foot processesThere is a genetic predisposition to classical diabetic nephropathy and its accompanying cardiovascular disease. First degree relatives of individuals with diabetes and nephropathy share the clinical phenotype of increased insulin resistance, higher blood pressure, more marked dyslipidaemia and more overt cardiovascular disease than first degree relatives of individuals without diabetic nephropathy. The actual genes involved remain unclear, and different genes may influence the initiation and progression of nephropathy, and also separately of albuminuria and glomerular filtration.
A host of metabolic abnormalities, coupled with haemodynamic factors, all secondary to hyperglycaemia drive the process. Independent accelerators such as cigarette smoking and hypertension also influence the rate of progression.
How does CKD in diabetes present clinically?
A proportion of individuals with short-duration diabetes have glomerular hyperfiltration which persists for some years after diagnosis. Whether or not this predisposes to later diabetic nephropathy remains controversial. In classical diabetic nephropathy, the first clinical sign of diabetic nephropathy is of increased urine albumin excretion into the microalbuminuric range.
Microalbuminuria is defined as urine albumin excretion above the normal range (urine albumin:creatinine ratio <3.5 mg/mmol) but below the level of proteinuria detected by conventional dip-stick (ACR >30 mg/mmol). There is a large day-to-day fluctuation in urine albumin excretion. In addition, in some individuals who have had low levels of microalbuminuria for several years, albumin excretion may revert back to within the normal range. Thus multiple measures of urine albumin excretion are needed, and one cannot be certain that an individual actually has diabetic nephropathy unless urine albumin excretion is observed over some time to be increasing gradually.
Individuals with low-level microalbuminuria should be regarded as being at high risk of developing nephropathy in the future, rather than as actually having nephropathy. Untreated, microalbuminuria will rise gradually, reaching the clinical proteinuric range over 5- 15 years. Glomerular filtration then begins to decline and end-stage renal failure is reached without treatment in 9 years.
In non-classical kidney disease in diabetes, urine albumin excretion remains within the normal range or in the low microalbuminuric range but does not progress. Instead, GFR falls steadily towards end-stage disease. The rate of decline in GFR appears to be at least as fast as in classical diabetic nephropathy.
How common is CKD in diabetes?
The life-time risk of an individual with diabetes developing microalbuminuria is approximately 50 %. Of those who do develop microalbuminuria, approximately one third will revert to normal urine albumin excretion, one third remain normoalbuminuric and one third progress to clinical proteinuria. Almost all of those who have clinical proteinuria will eventually reach end-stage renal disease, unless they succumb to cardiovascular disease beforehand.
Older data indicates that in both Type 1 and Type 2 diabetes, after 25-30 years duration, approximately 25 % of individuals will have clinical proteinuria. More recent figures in Type 1 diabetes suggest that in at least some cohorts, the incidence of proteinuria is now much less after 30 years duration, around 5-15 %. Only longer follow-up studies will confirm whether this is truly prevention of proteinuria or delay in its appearance.
Low GFR is common in diabetes, particularly in Type 2 diabetes. Large data-base analyses demonstrate that approximately 30 % of people with diabetes have eGFR <60 ml/min/1.73 m2. There has been an exponential rise in end-stage renal disease due to diabetes over the last twenty years, with diabetes being the attributable cause of ESRD in 50 % of new entrants to renal replacement therapy in some countries. This has been due mainly to an increase in older, frail individuals with Type 2 diabetes, whilst the numbers with Type 1 diabetes has remained steady or even fallen despite the increasing incidence of Type 1 diabetes.
Recently however, there are suggestions that CKD may develop more rapidly in obese people developing diabetes during adolescence and early adulthood, so that there may be an increase in ESRD in younger individuals in the future.
How should we screen for CKD in diabetes?
Current guidelines recommend that all individuals with diabetes should have screening for CKD annually. Urine albumin:creatinine ratio is measured, preferably in an early morning urine sample, using a sensitive assay. Serum creatinine is also measured, by standardised assay, and estimated GFR (eGFR) calculated. Abnormal values should be repeated within 3 months for confirmation. Individuals should then be categorised based the eGFR and ACR eg CKD stage 3b with microalbuminuria.
How can we prevent CKD in diabetes?
The risk of developing microalbuminuria is reduced by good glucose control. The lower the HbA1c, the lower the risk of developing microalbuminuria in both Type 1 and Type 2 diabetes. The open follow-up of the DCCT in Type 1 diabetes demonstrated that over 30 years of follow-up, the reduction in microalbuminuria translated into a reduction in clinical proteinuria and eventually end-stage disease.
In Type 2, but not Type 1, diabetes, there is also good evidence that good control of blood pressure reduces the likelihood of an individual developing microalbuminuria. The lower the blood pressure, the lower the risk. Most trials have used inhibitors of the renin angiotensin system, but it is likely that adequate control of blood pressure, to <140/80 mmHg, is more important than the actual class of agent used.
Whether these measures result in actual prevention of kidney disease or merely a delay in its appearance is hotly debated. Other factors, particularly not smoking, avoiding obesity and prescription of statins, probably also help in prevention, but evidence is lacking. There is no trial data about prevention of non-classical CKD, but similar measures are very likely to be appropriate.
How can we slow progression of CKD in diabetes?
There is only a little evidence that tight glucose control will prevent or delay progression of CKD in diabetes. Blood pressure control is vital. Maintaining blood pressure <130/80 mmHg reduces the fall in GFR from 10-12 ml/min/year to <5 ml/min/year and delays progression of albuminuria. Inhibitors of the renin angiotensin system are the backbone of treatment, in the maximum recommended or tolerated doses[^7][^8]. Many individuals require additional agents to achieve acceptable levels. Dietary protein restriction is generally no recommended because of concerns about malnutrition and difficulties with compliance. However, consideration should be given to advising a diet which does not contain more protein that actually needed eg >1.2 g/kg/day.
How do we tackle CVD risk?
Cardiovascular risk is increased 2-4 fold in microalbuminuria, ~9 fold in proteinuria and ~20-fold once serum creatinine is >180 μmol/l, compared to an individual with diabetes with no CKD. Aggressive management of CVD risk factors is thus appropriate. There is good evidence that target-driven, multifactorial management delivered in specialist clinics improves renal and cardiovascular outcomes substantially. Prescription of ACE inhibitors, statins and aspirin, coupled with attempts to improve glycaemic control and reduce smoking, reduce the risk of a cardiovascular event by approximately 50 %.
How can we manage end-stage disease?
Early referral to a specialist Nephrology service, preferably run jointly with specialist Diabetes services, improves outcome. Management of renal anaemia and bone disease can be optimised, cardiovascular screening investigations completed and there is time for physical and psychological preparation for commencement of renal replacement therapy. Individuals fit enough for transplantation should be referred when eGFR approaches 45 ml/min/1.73 m2, and others when eGFR reaches 30 ml/min/1.73 m2.
Earlier referral might be necessary where there is doubt about the diagnosis or difficulty in management. Simultaneous kidney and pancreas transplantation, done before dialysis is necessary, is the optimal approach. The decision to opt for haemo- or peritoneal dialysis depends on the individual patient. Survival on dialysis is poor, generally less than 5 years in the older, frailer patients.
How do we management diabetes in end-stage kidney disease?
The management of hyperglycaemia requires careful thought in ESRD. As GFR falls, the risk of hypoglycaemia rises, due to a combination of falling oral intake, decreased gluconeogenesis by the kidney and reduced clearance of glucose-lowering agents. Reductions in doses of insulin and sulphonylureas are often required.
Although the risk of lactic acidosis in renal failure is debated, current guidelines recommend that metformin should be stopped when eGFR approaches 30 ml/min/1.73m2. Fluid retention may be exacerbated by pioglitazone. Many of the gliptins are now licensed for use at low GFR and provide an attractive option, particularly because there is no risk of hypoglycaemia when used as monotherapy. None of the currently available GLP-1 analogues are currently licensed for use in CKD stage 4 or lower. Glucose control during dialysis is particularly problematic, because of wide swings in glucose levels.
How can we study kidney disease in diabetes?
Although there is much activity, research into kidney disease in diabetes is hampered by the lack of a suitable animal model which develops progressive renal failure secondary to diabetic glomerulosclerosis. In vitro monoculture work provides useful data, but is also limited in that often in the kidney cell-cell communication between different cell types is essential for normal function. Co-culture is thus often necessary.
In human studies, initial exploratory studies often use change in urine albumin excretion as the end-point. However, definitive studies with change in GFR as the primary end-point are then needed to ensure that the drug does indeed improve renal prognosis. Such studies are very often international, requiring large numbers of patients followed over several years and thus are very expensive. Although renal biopsy has been performed for research purposes, it does carry some risk. Magnetic resonance imaging and spectroscopy may provide much more information on renal structure, haemodynamics and biochemistry in a relatively non-invasive way, but techniques are in their infancy.
Despite these difficulties, our understanding of the pathophysiology of kidney disease in diabetes has improved considerably over recent years. Many new therapeutic targets have been identified and several new agents have been or are currently in clinical trial. Most centre around growth factor and hormone antagonism.
^ Thomas MC, Groop PH, Tryggvason K. Towards understanding the inherited susceptibility for nephropathy in diabetes. Curr Opin Nephrol Hypertens 2012;21:195-202.
^ Thorn LM, Forsblom C, Fagerudd J, Pettersson-Fernholm K, Kilpikari R, Groop PH; FinnDiane Study Group. Clustering of risk factors in parents of patients with type 1 diabetes and nephropathy. Diabetes Care 2007;30:1162-1167.
^ Gnudi L. Cellular and molecular mechanisms of diabetic glomerulopathy. Nephrol Dial Transplant 2012;27:2642-2649.
^ Marshall SM. Diabetic Nephropathy in Type 1 diabetes: Has the Outlook Improved Since the 1980’s? Diabetologia 2012; 55:2301-2306.
^ DCCT/EDIC Research Group. Intensive diabetes therapy and glomerular filtration rate in type 1 diabetes. N Engl J Med 2011;365:2366-2376.
^ UK Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes. UKPDS 38. Br Med J 1998; 317: 703–13.
^ Gæde P, Lund-Andersen H, Parving H-H, Pedersen O. Effect of a multifactorial intervention on mortality in type 2 diabetes. N Engl J Med 2008; 358: 580-591.