13.5.3 Diabetic nephropathy

Diabetic nephropathy is classically defined as a rise in urinary albumin excretion rate (UAER), often associated with an increase in blood pressure, with concomitant retinopathy but without evidence of other causes of renal disease (1). It is characterized by a progressive decline in glomerular filtration rate (GFR), eventually resulting in end-stage renal disease. Diabetic nephropathy occurs in approximately 30–35% of type 1 and type 2 patients and tends to cluster in families. These families also show a predisposition to cardiovascular disease and hypertension—and, hypertension, or a predisposition to it, appears a major determinant of diabetic renal disease. These data taken together clearly suggest an individual susceptibility to this complication.

The phases of diabetic nephropathy based on urine albumin excretion status and GFR are shown in Table 13.5.3.1 (2). Histological changes of diabetic glomerulopathy are present in over 95% of patients with type 1 diabetes and albuminuria (UAER >300 mg/day) and in approximately 85% of type 2 diabetic patients who develop albuminuria with concomitant diabetic retinopathy (1, 2). In the absence of diabetic retinopathy nearly 30% of patients with type 2 diabetes and proteinuria have nondiabetic renal lesions (1).

The all-cause mortality in patients with diabetic nephropathy is nearly 20–40 times higher than that in patients without nephropathy. In recent years it has become apparent that renal disease and cardiovascular disease are closely related and diabetic nephropathy is acknowledged as an independent and powerful risk factor for cardiovascular disease (3). Many patients with diabetes and renal impairment die from a cardiovascular disease event before they progress to end-stage renal disease. Diabetic nephropathy is the most common cause of end-stage renal disease worldwide and represent about 30–40% of all patients receiving renal replacement therapy in the Western World.

The earliest manifestation of diabetic renal injury is the detection of small amounts of the protein albumin in the urine (microalbuminuria) (3–5). Microalbuminuria is an early sign of renal microvascular disease and serves as a surrogate biomarker of vascular injury in diabetes, being a powerful predictor of cardiovascular disease and early mortality (3). Although the terms normoalbuminuria, microalbuminuria, and macroalbuminuria (clinical albuminuria) describe different categories of UAER, it is important to remember that they are part of a continuum in the relationship between albumin excretion and cardiorenal risk (3, 5). Recent studies have underscored that, irrespective of diabetes and/or hypertension, UAER in the general healthy population is related in an exponential fashion with cardiovascular risk with no evidence for a threshold (3, 6).

The evolution of diabetic nephropathy proceeds through several distinct but interconnected phases, an early phase of physiological abnormalities in renal function, a microalbuminuria phase, and a clinical phase of persistent clinical albuminuria progressing to a decline in GFR and ultimately to end-stage renal disease (Table 13.5.3.1).

Soon after the diagnosis of type 1 diabetes several renal abnormalities may be observed. Supra normal values of renal plasma flow and glomerular hyperfiltration (GFR above 135 ml/min/m²) are found in 20–40% of patients (7). This hyperfiltration is partially related to poor metabolic control and intensification of glycaemic control reduces GFR towards normal. An increase in kidney size (nephromegaly) is a prerequisite for hyperfiltration but its prognostic significance remains unclear. The increase in GFR is accounted for by an elevation in renal plasma flow, which contributes approximately 60%, and a rise in intraglomerular pressure. It is this glomerular hypertension which seems responsible for the progressive glomerular injury which eventually leads to loss of GFR. Glomerular hyperfiltration appears to increase the risk of progression to diabetic nephropathy in a meta-analysis, but uncertainty persists on its clinical significance (8).

About 97% of the small quantity of albumin filtered at glomerulus is reabsorbed nonselectively in the proximal tubules of the kidney. This reabsorptive process is at near maximal capacity so that moderate increases in filtered albumin results in elevated albumin excretion in the urine. The reabsorptive process is proportional to the filtered load of albumin and hence the excretion in the urine will change proportionately with the amount filtered. With the progression of microalbuminuria the proportion of albumin excreted increases. Thus, in patients with clinical albuminuria, albumin represents approximately 50% of total urinary protein. Most patients will exhibit microalbuminuria (i.e. UAER ranging between 20 and 200 μg/min) well before the onset of overt clinical albuminuria (Table 13.5.3.1). It is generally accepted that the rise in UAER seen in patients with microalbuminuria reflects an increased transglomerular flux of albumin as a consequence of an increased transglomerular pressure gradient and possibly loss in fixed negative charge on the glomerular basement membrane. As the disease progresses increases in glomerular membrane pore size also contribute to albuminuria. This paradigm has recently been challenged with the suggestion that the glomerulus is physiologically significantly less restrictive to the filtration of plasma albumin and that it is tubular damage which largely accounts for the increased albumin in the urine (9).

While the controversy remains unresolved, most would regard the glomerular wall as the main filtration barrier to albumin. The glomerular capillary wall is a complex structure which consists of fenestrated glomerular endothelial cells, the glomerular basement membrane and the glomerular epithelial cell or podocyte. All layers have been implicated in the restriction of albumin filtration but the podocyte has received particular attention in recent years as playing an important role in the development and progression of albuminuria (10). The mature podocyte via its interdigitating foot processes provides structural support for the glomerular capillaries and is the final barrier to the passage of proteins across the glomerulus into the urinary space. In human and experimental diabetes podocyte morphology is abnormal with broadening and effacement of foot processes. An intact slit diaphragm which bridges the space between foot processes is essential to preventing loss of albumin and other proteins into the Bowman’s space. The discovery of nephrin, a podocyte slit diaphragm protein, whose absence leads to gross proteinuria has further highlighted the importance of this cell (10). The pathophysiology of albuminuria is complex and likely to be heterogeneous even within a single disease entity such as diabetes with different cell types in the glomerulus, and possibly in the tubules, contributing in different degrees and at different stages of disease to the leakage of albumin.

Longitudinal studies in type 1 diabetes have demonstrated that microalbuminuria is associated with a 20-fold risk of progression to overt renal disease as compared to patients with normoalbuminuria (11). Without intervention microalbuminuria progresses towards clinical albuminuria over approximately 10–15 years. In healthy adults the normal UAER ranges between 1.5 and 20 μg/min with the median value around 6.5 μg/min (3). The average day-to-day variation of UAER is about 40%. In view of this high biological variability the diagnosis of microalbuminuria should ideally be made from the calculation of the median value of at least three timed nonconsecutive urine collections. In a clinic setting the calculation of the albumin to creatinine ratio (ACR) in an early morning urine sample has de facto replaced the measurement of albumin on timed urine collection and has proven of acceptable accuracy (3). ACR correlates closely with UAER and the relative constancy of urine creatinine excretion corrects to an extent for variability of urine albumin (3).

Figure 13.5.3.1 shows a screening and monitoring strategy for microalbuminuria in patients with diabetes. In patients with type 1 diabetes persistent microalbuminuria may be detected after 1 year of diabetes and in type 2 diabetes it can often be present at diagnosis. The exact significance of microalbuminuria in patients with short-term duration of diabetes is unclear; however, in patients with 5 or more years of diabetes the presence of microalbuminuria indicates definite, albeit early, renal injury. The prevalence of microalbuminuria varies between 6 and 19% within 1–5 years of duration of type 1 diabetes and may reach 40–50% after 30 years of diabetes, but recent reports suggest that this is falling (11, 12). Over 25–30 years of diabetes the cumulative incidence of microalbuminuria is around 30% (12).

Approximately 1.5–2.5% per annum of patients with type 1 diabetes and normoalbuminuria develop microalbuminuria with baseline UAER, poor glycaemic control, blood pressure, especially an increase of systolic blood pressure during sleep, presence of retinopathy, smoking, and dyslipidaemia being factors which influence this transition (11, 12). Morphological studies have clearly shown that structural abnormalities and lesions such as increased mesangial fractional volume and decreased filtration surface area are more pronounced and advanced in patients with microalbuminuria. Once microalbuminuria is established the UAER tends to rise over time. In early studies which followed patients from the late 1960s to early 1980s the rate of increase was about 14% per year and approximately 80% of patients with type 1 diabetes and microalbuminuria would develop overt clinical albuminuria. However, more recent studies suggest that in approximately 30% of patients UAER reverts back towards the normal range (<30 μg/min), in 50% it remains in the microalbuminuric range and in around 20% microalbuminuria progresses towards albuminuria over 5–9 years (11, 12). This change in the natural history of microalbuminuria most likely reflects changes and advances in medical care with ever more stringent glycaemic, lipid, and blood pressure control, as well as the widespread use in recent years of agents such as angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs). In those patients who never develop diabetic nephropathy UAER remains normal except perhaps during periods of poor glycaemic control and intercurrent illnesses where transient increases in UAER may be detected. Table 13.5.3.2 lists conditions that may affect interpretation of UAER results and should be considered when a patient initially presents with abnormal UAER.

The presence of microalbuminuria is consistently associated with higher blood pressure independent of age, duration of diabetes, gender, or body mass index, and this increase in pressure values of about 10–15% above that of normoalbuminuric patients occurs initially within the so called ‘normal’ blood pressure range. The transition from normo to microalbuminuria is accompanied by increases in blood pressure. Higher blood pressure values may indeed precede and predict the development of microalbuminuria suggesting that elevations in blood pressure and UAER initially through the normal range of values may represent concomitant manifestations of a common process leading to renal injury. Once developed, microalbuminuria not only denotes renal capillary damage, but also a heightened risk of vascular disease and alerts the physician to other modifiable risk factors. Box 13.5.3.1 shows the associations of microalbuminuria with other cardiovascular risk factors.

In those patients who develop clinical albuminuria (UAER >300 mg/day) GFR gradually declines in a linear fashion at variable rates (average 4.5 ml/min per year) depending on control of promoters of progression such as hypertension and degree of albuminuria and on individual response to treatment (11–13). Although variation in fall in GFR is seen from patient to patient the rate of fall remains relatively constant for each individual patient. With the advent of early and intensive treatment of hypertension the time from the onset of clinical albuminuria to death has virtually trebled from 7 to 21 years (14).

An elevated blood pressure is a feature of virtually all patients who develop albuminuria and blood pressure tends to rise by 7% per year, in parallel with progression of renal failure. Diabetic retinopathy and dyslipidaemia are also present in most patients with albuminuria. At this stage the course of renal failure does not appear to be reversible, however, available treatments can significantly slow the rate of decline of renal function and delay the need for renal replacement therapy.

The development of diabetic nephropathy in type 2 diabetes is in general very similar to that in type 1 diabetes, with some important differences. Microalbuminuria, and at times clinical albuminuria, can be seen at diagnosis of type 2 diabetes. The prevalence of microalbuminuria in type 2 diabetes ranges between 10% and 40% and depends on the population selected and ethnicity with the highest prevalences in South Asians, UK Asians, African Caribbeans, Maori and Pacific Islanders, and Pima Indians (5, 11).

In type 2 diabetes the rate of progression from normoalbuminuria to microalbuminuria and then to clinical albuminuria varies between 2% and 3% per annum and is affected by promoters of progression such as the degree of baseline UAER, glycaemic control, blood pressure, and dyslipidaemia (11, 15, 16). After 20 years of diabetes the cumulative incidence of proteinuria is between 20% and 50%, similar to type 1 diabetes, but with higher incidence in certain ethnic groups. At diagnosis of type 2 diabetes GFR can be normal or high and seldom, even reduced. This significant variability in GFR is partly explained by the range of intervals from onset of disease to time of diagnosis, and the concomitant presence of hypertension and nondiabetic renal disease. A significant proportion of patients with type 2 diabetes and raised UAER do not have the classical histological changes of diabetic glomerulosclerosis. In patients with type 2 diabetes and raised UAER but without diabetic retinopathy biopsy studies have reported prevalence of nondiabetic kidney diseases of approximately 30% (17). The concomitant presence of diabetic retinopathy reduces the prevalence of nondiabetic renal disease to about 16%. In certain populations such as the Pima Indians the GFR is found elevated already at the stage of impaired glucose tolerance. The GFR remains high with microalbuminuria and only begins to decline once clinical albuminuria develops. The rate of decline and the variation in rate of fall in GFR in type 2 diabetic patients with albuminuria is similar to that in type 1 diabetes. As in type 1 diabetes a rise in blood pressure is an early feature of fall in GFR in type 2 diabetes. However, the relationship between hypertension and diabetic nephropathy is often more difficult to dissect because of the high prevalence of arterial hypertension, especially systolic hypertension, in type 2 diabetes. There is significant ethnic variation in progression of diabetic nephropathy with one study from the USA showing that for equivalent blood pressure control, African Americans had a sevenfold greater decline in renal function compared with white subjects (18).

In type 2 diabetes microalbuminuria and clinical albuminuria are also strong predictors of cardiovascular disease perhaps more so that in type 1 diabetes, but this stronger association may be because of the older age of the type 2 population. Several mechanisms have been suggested for the relation of albuminuria with cardiovascular risk (3, 19).

A diabetic millieu is required for the diabetic glomerular lesion to develop. Clinical trials in both type 1 diabetes (Diabetes Control and Complication Trial (DCCT)) and in type 2 diabetes (United Kingdom Prospective Diabetes Study (UKPDS)) have established that the rate of development and progression of diabetic nephropathy is closely associated to glycaemic control (20, 21). Nevertheless in many patients with diabetes, diabetic nephropathy does not develop. Therefore in humans hyperglycaemia appears necessary but not sufficient to cause renal damage and other factors are needed for the clinical manifestation of this complication.

Hypertension plays a critical role in the initiation and progression of diabetic nephropathy. Indeed, the development of albuminuria in most cases is paralleled by a gradual rise in systemic blood pressure, and the levels of blood pressure closely relate to the rate of decline in GFR. Lowering of blood pressure by antihypertensive therapy has significant renoprotective and antialbuminuric effects. At the early stage of microalbuminuria the difference in arterial pressure levels between diabetic patients who will develop renal complications and those who will not can be numerically small but biologically highly relevant for the kidney in the presence of diabetes.

Under physiological nondiabetic conditions, intraglomerular capillary pressure is tightly regulated by precise adjustments in afferent and efferent arteriolar resistance. Hyperglycaemia induces vasodilatation with a marked reduction in afferent and a lesser reduction in efferent arteriolar resistance. This leads to an increase in glomerular capillary pressure levels and allows ready transmission of any increase in systemic blood pressure to the glomerular capillary circulation (22). Several observations support the notion that glomerular capillary hypertension is involved in glomerular injury and diabetic glomerulosclerosis. Conditions which increase glomerular capillary pressure, such as reduction of renal mass or superimposition of diabetes to genetic models of hypertension, significantly magnify glomerular damage. Manoeuvres which reduce glomerular hypertension, such as the use of inhibitors of the renin–angiotensin system (RAS) and low protein diet markedly slow the rate of progression of the glomerular injury. Autopsy studies in diabetic patients with unilateral renal artery stenosis show glomerulosclerotic lesions to be confined to the kidney with the patent renal artery. Elevated intraglomerular mechanical forces appear to damage and disrupt the normal structure of the multilayered glomerular barrier, eventually leading to the histological changes of diabetic glomerulopathy.

The molecular mechanisms of microvascular injury in diabetes are discussed in detail in Chapter 13.5.1. However the pathogenesis of microvascular disease is diverse and microcirculation-bed dependent. The glomerular capillary wall is quite distinct from other microvascular beds and many of the proposed common mechanisms of microvascular injury in diabetes do not fully explain the pathogenesis of diabetic glomerular disease. In the kidney both metabolic and haemodynamic perturbations trigger molecular pathways which contribute to glomerular injury (Fig. 13.5.3.2). In addition metabolic and capillary pressure stimuli interact through a mechanism of haemodynamic-metabolic coupling whereby any increase in glomerular capillary pressure exacerbates intracellular glucose metabolism thus leading, for any given level of prevailing glycaemia, to magnification of the deleterious effects of hyperglycaemia on glomerular cells (22). The pathogenesis of glomerular injury in diabetes has been recently reviewed in detail (22, 23).

Both metabolic and haemodynamic factors induce, via a series of molecular mediators, disruption of the function of glomerular capillary wall structures (endothelial cells, glomerular basement membrane and epithelial cells) which restrict filtration of plasma protein. This results in increased albumin leakage. Excessive glomerular filtration of albumin leads to an increase in tubular reabsorption of protein and accumulation of protein in tubular epithelial cells which respond by releasing vasoactive and inflammatory mediators that promote infiltration of mononuclear cells. This process causes tubulo-interstitial damage, renal scarring and progression of disease (15). A loss of functioning nephron units exacerbates haemodynamic pressure in surviving nephrons inducing further glomerular barrier damage and albuminuria, thus perpetuating a mechanism of both glomerular and tubular scarring. This results in progressive renal functional impairment. Indeed limitation of albuminuria has renal protective effects (15).

Histological changes occur early in the diabetic glomerulus. Within a few years of diagnosis, the thickness of the glomerular capillary basement membrane increases by about 30% above the nondiabetic width of between 250 and 450 nm, and this further worsens as the albuminuria becomes heavier. The mesangium undergoes changes of expansion with mesangial cell hypertrophy and hyperplasia and extracellular matrix accumulation. The relative volume of the mesangium, i.e. the fraction of the glomerulus occupied by mesangium, increases with the progression of albuminuria encroaching on the glomerular structure and reducing glomerular capillary filtration area, ultimately leading to glomerular occlusion. These processes are combined with hyalinosis of the glomerular afferent and efferent arterioles. Deposits of eosinophilic material accumulate in the arteriolar wall early within 4–5 years and contribute further to reduce blood flow to the glomerular capillary thus inducing ischaemia. Glomerular epithelial cells (podocytes) decrease in number (podocytopenia) and detach from the glomerular basement membrane leading to foot process effacement and fusion. With time, areas of basement membrane can become denuded of podocyte. If the disease progresses, the classical light microscopic features of established diabetic nephropathy would appear. They may be broadly divided into four categories.

The late histological change in the glomerulus are invariably associated with tubular and interstitium sclerosis. Indeed, it is the degree of these sclerotic changes that more closely relates to the loss of renal function and the decline in GFR. Tubular lesions, however, may also occur earlier in the form of tubular enlargement, tubular atrophy, glomerulotubular junction abnormalities, and atubular glomeruli (22, 24).

There is strong evidence of familial clustering for diabetic nephropathy. In cross-sectional studies, diabetic siblings of probands with type 1 diabetes and diabetic nephropathy have a prevalence of diabetic nephropathy of between 33% and 83%, as compared with a prevalence of 10–19% in siblings of probands without diabetic nephropathy (1, 11, 25). The cumulative incidence of diabetic nephropathy in siblings with type 1 diabetes of a proband with type 1 diabetes and diabetic nephropathy is 71.5%, as compared with 25.4% in diabetic siblings of a proband unaffected by diabetic nephropathy giving approximately a threefold increased risk of nephropathy for the sibling with positive family history of the complication. A family history of arterial hypertension is also important in the predisposition to diabetic nephropathy. An excess of hypertension and cardiovascular disease as well as reduced insulin sensitivity and hyperlipidaemia have been reported in relatives of diabetic patients with diabetic nephropathy compared to those without. In type 2 diabetes familial aggregation of type 2 diabetic nephropathy has been observed in Pima Indians and the heritability of UAER has been demonstrated in Caucasian patients, the resemblance being stronger between mothers and sons. Moreover familial aggregation of hypertension and cardiovascular disease has also been described in Caucasian patients with proteinuria. These observations suggest that diabetic nephropathy may form part of a wider propensity to cardiovascular disease, where genetic predisposition may interact with environmental determinants. However, the nature of the genetic susceptibility remains largely unknown. A recent review lists and details the genes that have been implicated (25). Many of these observations, however, remain unconfirmed or still controversial. The emerging picture seems to suggest a complex interaction between the effects of several genes and multiple environmental factors.

Essential hypertension has been consistently associated with the increased activity of a cell membrane cation transport system, the sodium lithium counter-transport, in red blood cells. This system is also overactive in both type 1 and type 2 diabetic patients with raised UAER (26). There is a significant correlation in the activity of this transport system between diabetic probands with nephropathy and their parents and in identical twins with diabetes. This suggests a heritability of the elevated activity of this system in diabetic nephropathy but its physiological relevance remains unclear. It could be related to tubular sodium reabsorption and in this respect studies which have demonstrated elevation in the activity of the Na⁺/H⁺ antiport in diabetic nephropathy are relevant. This ubiquitous membrane protein is involved in important cellular functions such as intracellular pH control, cell volume control, stimulus-response coupling and cell proliferation. Na⁺/H⁺ antiport activity is increased in leucocytes and red blood cells of type 1 diabetic patients with nephropathy. This abnormal phenotype is conserved in cultured skin fibroblasts and lymphoblasts suggesting that the over activity of the antiport is intrinsically determined. Furthermore cultured skin fibroblasts from patients with diabetic nephropathy show an impaired antioxidative response to high glucose, compared with diabetic patients without nephropathy and nondiabetic patients with nephropathy (22). This evidence suggests that the increased susceptibility to diabetic nephropathy resides in the host cell response to the metabolic disturbance of diabetes.

Most authorities now acknowledge the importance and need for screening for microalbuminuria in patients with diabetes. A calculation of the ACR in an early morning urine sample is generally used for this purpose. Screening is advised after the onset of puberty and on an annual basis. The detection of microalbuminuria indicates a heightened risk of cardiorenal disease and calls for intensification of treatment of modifiable risk factors (27–29).

To evaluate the progression of renal disease it is recommended that renal function is tested at least annually. The advent of equations that estimate the GFR (eGFR) from serum creatinine has made this easier. These equations take into account the effect of age, gender, weight and ethnicity and help in defining the stages of chronic kidney disease (Table 13.5.3.3). It must be noted however that the equations are accurate only in the range of GFR lower than 60 ml/min. There is significant underestimation of higher GFR values. In general all patients with stage 4 and 5 chronic kidney disease should be discussed with and referred to a renal physician if indicated. Patients with stage 3 disease should be referred to a renal physician if there is an unexpected and progressive fall in GFR of more than 5 ml/min per year, persistent proteinuria, micro or macroscopic haematuria, unexplained anaemia or metabolic disturbances in serum potassium, calcium or phosphate levels, and poorly controlled hypertension (29).

Studies in both type 1 and type 2 diabetes have demonstrated the benefits of intensive glycaemic control in reducing the risk of new onset microalbuminuria or clinical albuminuria. In the DCCT intensive glycaemic control with a mean HbA_1c of 7% (53 mmol/mol) reduced by 39% new onset microalbuminuria compared to conventional control with a mean HbA_1c of 9.1% (76 mmol/mol) in type 1 diabetes (20). Similarly in the UKPDS, a mean achieved HbA_1c of 7% (53 mmol/mol) in the intensive control group as compared with 7.9% (64 mmol/mol) in the conventional control group was associated with a relative risk reduction of developing microalbuminuria of nearly 30% after 9–12 years of follow-up in patients with type 2 diabetes (21). In the Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified Release Controlled Evaluation (ADVANCE) trial patients with type 2 diabetes treated by intensive glycaemic control, to a target HbA_1c of 6.5% (48 mmol/mol), had a 9% (75 mmol/mol) relative reduction in the risk of new onset microalbuminuria as compared to patients on standard control with a HbA_1c of 7.3% (56 mmol/mol) (30). These studies showed no HbA_1c threshold suggesting that the lower the HbA_1c the lower the risk for nephropathy. In the DCCT Epidemiology of Diabetes Interventions and Complications (DCCT-EDIC) study the incidence of microalbuminuria and clinical albuminuria remained lower in the group that was originally allocated to tight glycaemic control (31). In the UKPDS follow-up data, reduction of microvascular complications, which included renal disease also, persisted in the intensive control group treated with insulin and sulphonylurea (32). These follow-up results were obtained despite obliteration of the difference in glycaemic control between treatment groups. Importantly these beneficial ‘legacy’ effects extended to cardiovascular disease.

In the UKPDS better control of blood pressure (achieved mean blood pressure 144/82 mmHg) as compared with ordinary control (mean blood pressure 154/87 mmHg) translated into a 29% risk reduction of developing microalbuminuria over a 6-year period. Treatment with an ACE inhibitor (captopril) and a B-blocker (atenolol) was equally effective but the study was not powered to detect between drug differences (21). Continued good blood pressure control is critical because unlike glycaemic control, there was no legacy effect in the UKPDS follow-up on vascular complications as blood pressure control worsened (33). The Bergamo Nephrologic Diabetes Complications Trial (BENEDICT) found that the risk of developing microalbuminuria was reduced by about 50% by the use of the ACE inhibitor trandolapril but not by verapamil for equivalent blood pressure reduction in hypertensive patients with type 2 diabetes and normoalbuminuria (34). The ADVANCE blood pressure trial showed that in patients with type 2 diabetes treatment with perindopril and indapamide, a thiazide diuretic, reduced renal events (predominantly new onset microalbuminuria) by nearly 21% compared to conventional antihypertensive treatment. Achieved blood pressure was lower in the perindopril/indapamide group (35). At variance with these studies the Diabetic Retinopathy Candesartan Trial (DIRECT) failed to demonstrate, in a post hoc analysis, any effect of the ARB candesartan on development of microalbuminuria in type 1 and type 2 diabetic patients (36). This was despite lower blood pressure levels in the candesartan-treated group. Therefore the question of whether there are drug specific effects, over and above the effect of lowering blood pressure, in the prevention of microalbuminuria in diabetes remains at present unsettled.

In type 1 diabetes control of hyperglycaemia is likely to promote reversal to normoalbuminuria or delay progression to clinical albuminuria but not all studies are in accord. In an early study of intensive insulin therapy by infusion pump near normoglycaemia was associated with a significant and sustained reduction of UAER (37) and a meta-analysis of 16 randomized trials revealed that the risk of nephropathy progression was decreased significantly by intensive glycaemic control in microalbuminuric (as well as normo-albuminuric) patients (38). However the Microalbuminuria Collaborative Study, a small 5-year, randomized controlled trial failed to show an effect of intensive glycaemic control on progression of microalbuminuria. One important limitation of this study was that separation of glycaemic control between the intensive and conventional treatment groups did not persist beyond 3 years (39).

In type 2 diabetes the ADVANCE study demonstrated a significant reduction in the incidence of clinical albuminuria by intensification of glycaemic control in patients with microalbuminuria (30). The magnitude of this effect was greater than that achieved by strict control in lowering the rate of transition from normoalbuminuria to microalbuminuria.

In type 1 diabetes several earlier studies demonstrated the benefit of antihypertensive therapy on delaying progression and in some cases inducing reversal of microalbuminuria, by treatments which included calcium channel blockers, B-blockers and drugs that interfered with the RAS. A meta-analysis of all trials with ACE inhibitors confirmed the efficacy of this class of drugs, suggested that the effect may be independent of blood pressure lowering and showed that the higher the baseline UAER the greater the response to therapy (40). The extent of the reduction of microalbuminuria appeared to be attenuated after 4 years, raising the possibility that ACE inhibitors delay, rather than completely prevent the progression towards clinical albuminuria.

In type 2 diabetes progression of microalbuminuria is impaired by treatment with ARBs and regression to normoalbuminuria is enhanced (irbesartan in patients with type 2 diabetes and microalbuminuria (IRMA) study and MicroAlbuminuria Reduction with VALsartan (MARVAL) study) (41). These effects appear independent of blood pressure lowering in as much as similar blood pressure reductions by amlodipine, a dihydropyridine calcium channel blocker, have a very modest effect on microalbuminuria. In the ADVANCE trial the use of perindopril and indapamide resulted in a 18% reduction in the progression microalbuminuria to the composite nephropathy endpoint of clinical albuminuria, doubling of serum creatinine, need for renal replacement therapy, and renal death, which did not reach conventional statistical significance (35).

Rather than specific antihypertensive treatments the Steno-2 study applied a multifactorial intensive approach to care versus conventional care in type 2 diabetic patients with microalbuminuria. Intensive multifactorial treatment significantly lowered progression to clinical albuminuria and overt nephropathy after 4 years (42). Intensive multifactorial intervention included control of hyper-glycaemia, hypertension (mainly by ACE inhibitors or ARBs), dyslipidaemia, use of aspirin and behavioural modification and crucially this was associated with a lower rate of GFR decline (43).

A large percentage of patients in the intensive treatment arm achieved a blood pressure target of <130/80 mmHg and a total cholesterol target <4.5 mmol/l. However, less than 20% of patients achieved a glycaemic target of a HbA_1c <6.5% (48 mmol/mol). At the end of this 8-year study patients assigned to the intensive multifactorial approach also had a significant reduction in cardiovascular events. This beneficial cardiovascular disease effect as well as a reduction in total mortality persisted after a further 5 years of follow-up when all patients had been recommended intensification of treatment (44).

Observational studies have found an association between HbA1c and loss of renal function in diabetic patients with clinical albuminuria. These studies however do not prove cause-effect and no large controlled clinical trial has addressed this question. Small controlled studies in type 1 diabetes failed to show a beneficial effect of intensive glycaemic control on rate of GFR decline (45). Thus, whether intensification of glycaemic control in patients with clinical albuminuria impacts on progression of GFR loss remains an open question.

In type 1 diabetes treatment with an ACE inhibitor captopril compared with conventional antihypertensive treatment significantly slowed the rate of loss of renal function by approximately 50% measured as the risk of a doubling of serum creatinine, need for renal replacement therapy or death. These effects seemed to be ACE inhibitor specific because they persisted after adjustments for difference in mean arterial pressure (46).

In type 2 diabetes with overt nephropathy use of ARBs obtained similar effects although the magnitude of reduction of risk of progression of renal disease was smaller in the order of 25–30% (11, 41). In all these studies the degree of reduction in clinical albuminuria in the first 6 months of treatment was linearly related to the degree of preservation of renal function, strongly suggesting that lowering of albuminuria is causally directly related to renal protection.

New treatments have therefore targeted albuminuria as a modifiable risk factor for kidney disease progression. Recently in a proof of concept study the combination of Aliskiren, a new oral renin inhibitor, with losartan reduced albuminuria independently of blood pressure lowering by a further 20% as compared with losartan treatment alone in patients with type 2 diabetes and clinical albuminuria (47). Long-term controlled clinical trials are currently investigating the effects of aliskiren on clinical renal and cardiovascular outcomes in type 2 diabetes.

Several recent large clinical trials with primary cardiovascular outcomes have examined the effect of treatment on renal outcomes in populations of patients with type 2 diabetes. These studies were not designed for patients with diabetes on the basis of prespecified renal characteristics and outcomes and all analyses were post hoc. Nevertheless concordantly all these trials have shown that in type 2 diabetic patients at high cardiovascular risk therapy that inhibited the RAS resulted in a lower incidence of overt nephropathy compared to conventional treatment (41). Combination of ACE inhibitors and ARB has been advocated by some authors for greater cardiorenal protection in patients with type 2 diabetes. Results from a recent large randomized clinical trial, however, suggest that this approach does not translate into a reduction in cardiovascular or renal events and may potentially be detrimental to renal function (48).

The body of evidence from all these large trials has led to the formulation of guidelines which recommend ACE inhibitors or ARBs as first-line antihypertensive therapies for patients with diabetes (27, 28). Their renoprotective effects appear at least partly to be independent of their blood pressure lowering effects. Although the trial evidence supports the use of ACE inhibitors in type 1 diabetic nephropathy and ARBs in type 2 diabetic nephropathy, a lack of comparative studies leaves unresolved whether ACE inhibitors and ARBs can be used interchangeably in patients with type 1 or 2 diabetes (41). One study in subjects with type 2 diabetes (the majority with microalbuminuria) found no difference between the ACE inhibitor enalapril and the ARB telmisartan on change in measured GFR (41). Some current guidelines recommend targets of blood pressure control of ≤125/75 mmHg in patients with diabetic renal disease and proteinuria >1 g/day. However, whether such aggressive blood pressure control levels further lower the risk of end-stage renal disease in diabetic renal disease has not been put to the test.

A low protein diet reduces clinical albuminuria in patients with diabetic renal disease. However, the long-term significance of such diets on renal endpoints remains unclear with conflicting evidence. Most studies in this area are small and of short duration. In an early study of 19 patients with type 1 diabetes and a mean GFR of 60 ml/min, Walker et al. demonstrated a reduction in the rate of decline of GFR from 0.61 to 0.14 ml/min per month when protein intake was decreased from 1.13 g/kg body weight to 0.67 g/kg body weight per day (49). This effect which was accompanied by a significant reduction in albuminuria appeared independent of any systemic blood pressure effect. The individual responses to the low protein diet were, however, heterogeneous but the reason for this variability was unclear.

A systematic review and a meta-analysis of studies in this area have come to conflicting conclusions ranging from no evidence for an effect to a low protein diet induced reduction in the need for renal replacement therapy (50, 51). It is unclear whether a low protein diet affects disease progression but it may allow initiation of renal replacement therapy at a lower GFR. Acceptability and palatability of a low protein diet have also been important factors for long term compliance with such a therapeutic approach. There have also been concerns about malnutrition if significant reductions of protein intake are applied. Nevertheless in azotaemic (uraemic) patients restriction of protein lessens the signs and symptoms of uraemia and improves the adverse metabolic profile in the pre-dialysis stage (52). In conclusion low protein diet may benefit the diabetic patient with advanced renal failure (GFR <20 ml/min), but its value in the patient with better preserved renal function is uncertain.

Current treatments for diabetic kidney disease are effective in delaying disease progression, but do not obtain disease remission or regression. To address this unmet need new therapies that aim at new molecular targets in the pathogenesis of diabetic nephropathy have been tested in clinical trials in recent years. These have included inhibitors of the endothelin type A receptor (avosentan), inhibitors of PKC β (ruboxistaurin) and heparin-like glycosaminoglycans (sulodexide). Overall the results have been disappointing because of lack of effectiveness or safety concerns. A large clinical trial with renal and cardiovascular endpoints is ongoing in type 2 diabetes with an orally active renin inhibitor (aliskiren) recently introduced in the market for treatment of hypertension.

The general management of a patient with diabetes and end-stage renal disease requires a multidisciplinary approach. As concomitant cardiovascular disease risk is high optimization of medical treatments and regular review of medications is required along with unimpeded access to full cardiac (noninvasive and invasive) investigations. Recent UK public health guidelines advice hepatitis screening and hepatitis B vaccination if renal replacement therapy is anticipated. Once estimated GFR declines below 30 ml/min and/or serum creatinine rises above 150 µmol/l, metformin is generally stopped and short-acting sulphonylureas or insulin are used in preference as antidiabetic therapy. As GFR declines further often there is a need for reduction in insulin dose in particular and doses of sulphonylureas. Referral to a nephrology unit or a combined diabetes/renal unit is recommended at stage 4 CKD as early referral allows for optimization of medical therapy as well as the psychological and physical preparation for renal replacement therapy.

Renal transplantation should be the goal for all patients because both patient and graft survival are better post transplantation as compared to dialysis. Rehabilitation is also easier and with better outcomes. Transplantation should be preceded by a comprehensive cardiovascular work up and treatment if indicated. Transplantation should be considered in all patients with diabetes and end-stage renal disease but its widespread use is limited by reduced availability of donor organs. Ideally transplantation should be performed before dialysis is needed but this is often not possible. Simultaneous pancreas kidney transplant is the treatment of first choice in type 1 diabetes with end-stage renal disease. Ten-year graft and patient survival following combined transplantation can be nearly 20% greater than cadaveric donor kidney only transplantation in some observational studies, but no better than living donor kidney transplantation (53, 54). However, studies report better physical health and quality of life with simultaneous pancreas kidney transplantation and there is some evidence that pancreas transplantation may limit diabetic microvascular complications (54). Despite the clear benefits of transplantation most patients with diabetes and end-stage renal disease are treated with haemodialysis. In these patients vascular access may be more technically difficult and there is a greater premature failure rate of arteriovenous fistula as compared with nondiabetic patients. It is also often difficult to optimise metabolic control during dialysis with variable insulin doses being needed. Awareness of the risks of post dialysis blood pressure changes is required. Continuous ambulatory peritoneal dialysis (CAPD) can be used in patients with diabetes and in theory offers benefits because of easier blood pressure and volume control and the lack of need for vascular access. However, higher rates of peritonitis have limited its widespread use in patients with diabetes.

Currently diabetic nephropathy is the most common cause for need for renal replacement therapy worldwide. With the rising incidence of type 2 diabetes mellitus, the numbers of patients developing renal disease will increase. Patients with diabetic nephropathy are also at an increased risk of cardiovascular disease and increased albumin excretion rate remains the best bed side marker for predicting risk of both renal and cardiovascular disease. There is clear evidence that improvement of glycaemic control and reduction of elevated systemic blood pressure prevent/reduce the risk of diabetic nephropathy. Of the antihypertensive medications, those that interfere with the RAS appear particularly effective in delaying progression towards end-stage renal failure. It is paramount that a multifactorial treatment approach be initiated to prevent and delay progression of both cardiovascular and renal disease in this high-risk population.