3.1 Renal effects of angiotensin II
in diabetic nephropathy

Renal angiotensin II is delivered from the circulation or formed intrarenally from angiotensin I (22). Knowledge of the renal effects of angiotensin II have expanded considerably during the last three decades. Increasing evidence suggest that angiotensin II acts as a circulating vasoconstrictory hormone as well as a paracrine and autocrine peptide to modulate renal function (39) ( Figure 1 ).

3.1.1 Haemodynamic effects of angiotensin II

Although several vasoactive hormones exhibit effects upon renal haemodynamics, angiotensin II has been shown to be a major intrarenal hormone to regulate glomerular filtration rate (40). Early experimental studies of glomerular ultrafiltration with infusion of angiotensin II (41; 42) demonstrated by direct micropuncture measurements, that angiotensin II increases efferent and afferent arteriolar or at least preglomerular resistance, the latter primarily due to an autoregulatory response to an increased systemic blood pressure. As a consequence of a greater efferent than afferent arteriolar effect of angiotensin II, the glomerular capillary pressure increases.

In diabetic rats, glomerular capillary hypertension, proteinuria and renal structural abnormalities were found 14 months after induction of disease (43). However, ACE inhibitor treatment in a parallel group of diabetic rats, reduced glomerular capillary pressure to a level comparable to a non-diabetic control group which prevented development of proteinuria and structural abnormalities. These findings indicate that haemodynamic effects of angiotensin II are central in the pathogenesis of diabetic glomerulosclerosis and similar results have been found in non-diabetic rat models (44).

In type 2 diabetes, calculated glomerular pressure and efferent arteriolar resistance were found to be increased in patients with elevated urinary albumin excretion rate compared to normoalbuminuric patients (45). ACE inhibitor treatment lowered glomerular capillary pressure and albuminuria, in agreement with the experimental findings. Similar data in type 1 diabetic patients are presently not available, but our study of the time course of the antihypertensive and antiproteinuric effect of ARB treatment (46) discussed in section 6, demonstrated that systemic and renal haemodynamic mechanisms are of primary importance for the antiproteinuric effect. Hence, experimental and human data suggest that haemodynamic effects of angiotensin II are crucial in the pathogenesis of diabetic glomerulosclerosis.

3.1.2 Non-haemodynamic effects of angiotensin II

The recognized role of angiotensin II in the pathogenesis of diabetic nephropathy can not exclusively be attributed to the haemodynamic effects of angiotensin II (10). Several studies have investigated the importance of non-haemodynamic effects of angiotensin II in the pathogenesis of diabetic nephropathy. Accumulating data support the hypothesis that angiotensin II exerts diverse actions such as growth stimulation, induction of fibrogenesis and modulation of endothelial function.

Angiotensin II induce expression of TGF- β , collagen and fibronectin and stimulates mesangial cell proliferation resulting in increased synthesis of extracellular matrix (47; 48). Increased expression and synthesis of matrix proteins is not a direct effect of angiotensin II, but appear to be mediated through activation of PKC and p38 MAPK pathways which stimulate TGF- β gene expression (49). Numerous studies have indicated that hyperglycemia induce an increase in TGF- β protein and mRNA expression in experimental and human diabetes (50; 51) as well as in cell culture studies (51; 52). Other studies of mesangial cell cultures have disclosed striking similarities between the effects of high glucose-containing medium and incubation with angiotensin II on growth properties and induction of cytokines (53). Furthermore, increased gene expression of TGF- β in cultured mesangial cells stimulated with angiotensin II and high glucose concentration appear to be mediated by the same PKC and p38 MAPK pathways (49). Hence, the effect of high glucose on TGF- β gene expression may be mediated by increased angiotensin II production in mesangial cells. This hypothesis is supported by the observation that increased TGF- β secretion induced by high glucose may be inhibited by addition of an ARB (53).

Numerous other cytokines are involved in mesangial matrix accumulation. Recent studies have demonstrated that CTGF is an important factor in the pathogenesis of mesangial matrix accumulation acting downstream from TGF- β (54). VEGF is a family of potent cytokines which act to induce angiogenesis and markedly increase microvascular permeability, suggesting a key role in the pathogenesis of diabetic microvascular complications (55). VEGF is abundantly expressed in the renal glomerulus, especially in podocytes (56). Angiotensin II is a potent stimulator for VEGF mRNA expression and peptide production (55) and elevated plasma levels were recently demonstrated by Hovind et al (57) in type 1 diabetic patients with diabetic nephropathy. The exact role of VEGF on the pathogenesis of diabetic nephropathy is yet unclear, whereas evidence implicating VEGF in the pathogenesis of diabetic retinal neovascularization is much more substantial (58).

Plasmin is a key regulator of fibrinolysis and extracellular matrix turnover. Generation of plasmin from plasminogen by plasminogen activators are controlled by PAI-1. Angiotensin II has been shown to increase PAI-1 in cultured mesangial cells, which may lead to increased matrix accumulation (59).

Angiotensin II has been identified as a major modulator of endothelial function. Endothelial cells produce vasoconstrictive substances such as endothelin-1 and vasodilator substances such as nitric oxide (NO). NO serves as a functional antagonist for angiotensin II in the kidney, i.e. counterbalancing the constrictor actions of angiotensin II (60). Furthermore, NO has renal antiproliferative effects and decreases matrix protein synthesis (61). In vivo assessments of endothelial function by forearm pletysmography, has most often revealed an impaired state in diabetes (60). The NO system may be up-regulated in diabetic nephropathy, but according to Jaimes et al (62) angiotensin II also activates superoxide anion production. The superoxide anion interacts with NO, thus, the net effective NO availability in diabetic nephropathy appear to be reduced due to inactivation by superoxide anions. Angiotensin II has been identified as an important stimulator of the synthesis of the vasoconstrictor endothelin-1 in endothelial cells (63), vascular smooth muscle cells and mesanglial cells (64). The interaction between RAAS and endothelin-1 is emphasized by demonstration of reduction of endothelin-1 by ACE inhibition in human and experimental studies (65). Furthermore, this effect appears to be independent of bradykinin accumulation (65). Thus, angiotensin II may induce an imbalance between expression of the vasodilator NO and superoxide anions or increased endothelin-1 expression, which may enhance progression of diabetic nephropathy.

Further cytokines and growth factors such as insulin-like growth factors, interleukin-6 and platelet-derived growth factor are involved as primary or downstream effects of angiotensin II in the pathogenesis of diabetic nephropathy (66). Comprehensive research is ongoing to clarify the field of non-haemodynamic effects of angiotensin II in diabetic nephropathy.

In summary, angiotensin II stimulate fibrogenic growth factors and cytokines in mesangial cell culture studies, however, studies of the effects of high glucose concentration reveal similar results. Activation of cytokines by hyperglycemia may be mediated by angiotensin II formation in mesangial cells and inhibited by ARBs. Interactions between hyperglycemia, angiotensin II and intrarenal cytokines should be further investigated.

3.2 Angiotensin II receptors

Angiotensin II mediates its main effects through binding to AT1 and AT2 receptors. Further receptor subtypes have been identified, but the physiological significance of these are unknown (67). AT1 and AT2 receptors belong to a superfamily of 7 transmembrane-domaine G-protein coupled receptors, but differ significantly in distribution and expression in different tissues.

The AT1 receptors are responsible for the major actions of angiotensin II in adults. In the glomerulus, the presence of AT1 receptors has been demonstrated on mesangial cells (23), endothelial cells (68) and podocytes (69). AT1 receptor activation reduces renal blood flow and causes efferent arteriolar constriction (70). Activation of AT1 receptors in mesangial cells promote growth factors such as TGF- β (47; 48) and decrease matrix degradation due to PAI-1 activation (59), as discussed previously. AT1 receptor function on podocytes is unclear, but may influence the glomerular filtration barrier through contraction of foot processes (71), as discussed in section 4. In proximale tubules, AT1 receptor activation stimulates transport of sodium and water and induce expression of TGF- β (72) thereby stimulating the fibrogenic process.

The functions of AT2 receptors are incompletely understood. Experimental data suggest that activation of AT2 receptors oppose the actions mediated via the AT1 receptors, by endothelialcell-mediated renal vasodilation and exert antiproliferative effects (73; 74). Renal vasodilation induced by AT2 receptor activation may also be mediated through release of bradykinin and nitric oxide (75). The potential vasodilatory effect of AT2 receptor stimulation would be especially convenient during ARB treatment, since plasma concentrations of angiotensin II are increased. In contrast, recent data from experimental studies in non-diabetic kidney disease have suggested that that expression of VEGF may be mediated by AT2 receptors (76) and specific blockade of the AT2 receptor may confer a degree of renoprotection (77). Recently, a reduction of AT2 receptors was demonstrated in a model of early diabetes in streptozotocin diabetic rats (78) but the functional importance of possible downregulation of AT2 receptors in diabetes is unknown as is the state of the receptor in diabetic renal disease. Thus, evidence regarding the effect of AT2 receptor activation in diabetes is presently conflicting and limited.

3.3 Pathways for generation of angiotensin II
and escape of ACE inhibition

Angiotensin II is primarily generated by conversion of angiotensin I through the ACE pathway, but alternative enzymes such as chymase, also contribute to formation of angiotensin II (79). Thus, inhibition of ACE will probably not induce a complete blockade of the RAAS. Angiotensin II levels in plasma are lowered after initiation of ACE inhibitor treatment (80; 81), however, during prolonged treatment, the degree and duration of angiotensin II suppression may be reduced in agreement with partial escape of ACE inhibition (81). Incomplete RAAS blockade during chronic ACE inhibitor treatment due to ACE-escape and angiotensin II formation by alternative pathways may be overcome by inhibiting the action of angiotensin II at the site of the AT1 receptor by an ARB.

4.1 Glomerular permselectivity
in animal models

In normal rats, 90% of filtered albumin is reabsorped in proximal tubular reabsorption, thus, measured albuminuria, approximates only 10% of the corresponding filtered load (98). In streptozotocin diabetic rats, filtered albumin load become significantly elevated 50-70 days after induction of diabetes (99). Tubular reabsorption is markedly depressed, suggesting that albuminuira is not simply related to a defect in the glomerular capillary wall, but also ascribed to a tubular dysfunction in this early phase of diabetic glomerulopathy (99). The size-selective properties of the glomerular capillary wall and effect of ARB treatment has been investigated in a rat model analogue to early nephropathy in humans as judged by the presence of normal glomerular filtration rate, two-fold increase in proteinuria and low prevalence of glomerulosclerosis (100). Three groups of rats were followed for 12 months: two streptozotocin diabetic groups, one treated with insulin only, and one treated with insulin + losartan and a non-diabetic control group. Analysis of the data after 12 months for evaluation of size-selectivity revealed that sieving coefficients of Ficoll for all molecular radii were significantly higher in diabetic, insulin treated animals as compared with the control group. However, in insulin + losartan treated diabetic animals, Ficoll sieving coefficients were significantly lower than the insulin treated rats and comparable to that measured in the non-diabetic control group. Proteinuria increased significantly in all groups compared to baseline but was even lower in the losartan group compared to the control animals (100). Computations based on the hydrodynamic theory of hindered solute transport (92) was carried out according to the model assuming that the glomerular membrane is perforated by two pore-size "populations", a log-normal restrictive distribution and a non-restrictive shunt pathway. From this model, it was found that in insulin treated animals, as compared with controls, pore-size distribution was shifted towards larger sizes for restrictive and non-restrictive pore populations. However, treatment of the diabetic animals with losartan induced a uniform significant reduction in membrane pore sizes. Hence, by Ficoll clearance, this rat model of early diabetic nephropathy revealed major modifications of size-selective functions of the glomerular capillary wall, whereas treatment with an ARB prevented the development of the size-selective dysfunction.

The influence of angiotensin II on glomerular capillary wall function has been investigated using Ficoll as a probe molecule in a model of isolated perfused kidney in non diabetic rats (101). Infusion of angiotensin II significantly increased proteinuria and fractional clearance of Ficoll molecules of sizes greater than 34 Å. However, pre-treatment with and ARB, completely prevented these changes. An additional finding from this study was that angiotensin II infusion increased proteinuria by one order of magnitude more compared to clearance of Ficoll molecules corresponding to the size of albumin (36 Å). These observations indicated that changes in size-selectivity induced by angiotensin II, could not entirely explain the substantial proteinuria. Thus, angiotensin II may also affect other factors such as charge-selectivity or tubular function. However, caution should be taken before extrapolating data from an isolated kidney preparation to an intact human kidney.

4.2 Glomerular permselectivity
in diabetic nephropathy

Size-selectivity in type 1 diabetic patients with diabetic nephropathy and different levels of proteinuria have been extensively investigated by clearance of neutral dextran test molecules of graded size (30-60 Å) (16; 17; 82; 83; 102-104). Myers et al (83) originally demonstrated that advanced diabetic nephropathy with nephrotic range proteinuria is associated with elevated sieving coefficients above normal levels for large, nearly impermeant molecules > 50 Å. Analysis of dextran sieving data according to the concept of hindered solute transport (92), suggested that the elevated θ for the large molecules, could be explained by an expansion of a minor region of the glomerular capillary wall that behaves as a non-restrictive shunt pathway, whereas the radius of the restrictive pores (r 0 ) in the major membrane component was indifferent from normal control persons. Furthermore, data have revealed that advanced diabetic glomerular disease is characterized by a graded increase in the fraction of the filtrate penetrating this shunt pathway ( ω 0 ) (82; 102; 103). A relation between increasing proteinuria and the prominence of ω 0 is indicated by strong correlations between fractional clearance of albumin or IgG and ω 0 (82; 102; 103). Given the large size of immunoglobulins (r 0 = 55 Å) it is likely, that IgG is filtrated exclusively through the large pores in the shunt pathway. Although the smaller size of albumin (r 0 = 36 Å) it has been suggested that the major route of transcapillary escape of albumin could be via the non-restrictive pores in the shunt pathway rather than penetrating the smaller restrictive pores of the glomerular capillary wall because of the ellipsoid configuration and negatively charge of albumin (82; 102; 103). Similar results of size-selectivity in diabetic nephropathy have been found in type 2 diabetes (105) and by other research groups (104; 106; 107).

A comparable size-selective defect has been difficult to demonstrate in patients with lower grade proteinuria. In diabetic patients with sub-nephrotic range proteinuria or early nephropathy, significant elevations of θ compared to control persons could not be demonstrated, even for large dextran molecules (16; 103-105). A number of studies have demonstrated that the urinary clearance of the most anionic species of either albumin or IgG may be elevated relative to that of less anionic or cationic type of the same protein, which may suggest impairment of charge-selectivity (103; 104; 108-111). However, given the unknown contribution of tubular function, evaluation of endogenous protein clearance are difficult to interpret. An enhanced binding of cationic molecules in the tubules, which may favour clearance of anionic proteins into the urine, has been suggested as an alternative explanation for the changes in charge selectivity (112; 113).

Previous investigations of size-selectivity in early nephropathy were performed with dextran to probe the glomerular capillary wall. To circumvent the limitations with this method, we investigated whether early diabetic nephropathy, defined as a serum creatinine within normal range, is associated with impairment of barrier size selectivity, using Ficoll 70 characterized by the superior rigid structure compared to dextran to probe the glomerular membrane (114). In addition, we investigated the influence of the ARB losartan on size-selectivity and albuminuria. The double-blind cross-over study consisted of two treatment periods each lasting 2 months in which the patients received losartan 50 mg o.d. and placebo. Twelve normotensive, previously untreated type 1 diabetic patients with diabetic nephropathy and normal kidney function were included in the study. Ficoll clearances were performed in the end of each treatment period during water diuresis. Twelve healthy volunteers provided control values for glomerular function. Bolus injections of inulin and para-aminohippurate for determination of GFR and RPF, was followed by continuous infusion of each marker. After a 60 minute equilibration period, a bolus injection of 1 mCi of tritiated Ficoll 70 ( 3 H-Ficoll) was given. Thereafter, four timed urine collections were made at 20 minutes interval. Plasma was sampled to bracket each collection. The average renal clearance of inulin was equated with the GFR. The molecules of Ficoll 70 in infusate spanned a radius range of 10-90 Å with the peak radius at 30 Å. The mean Ficoll sieving curve in placebo treated patients compared with healthy control persons is illustrated in Figure 2 . As shown, diabetic nephropathy is associated with a selective elevation of θ for large nearly impermeant molecules in the 50-60 Å interval. Membrane parameters computed from the Ficoll sieving curves, revealed that neither the mean radius nor the breadth of the lower distribution of restrictive pores in placebo treated patients differed from control values. However, the upper distribution of shunt-like pores was one order of magnitude more prominent in placebo treated type 1 diabetic patients as compared to control persons (p < 0.05). Fractional IgG clearance was elevated by two orders of magnitude, whereas the corresponding elevation of θ for Ficoll of equivalent radius (56 Å) was approximately 50%. However, a sieving coefficient of IgG comparable to that of Ficoll of 56 Å radius combined with an expected reduction in reabsorption of the enhanced load of filtered IgG in diabetic nephropathy, could justify this disparity. Thus, we conclude that barrier size defects could account for the elevated level of immunoglobulinuria in early diabetic nephropathy. Comparison of fractional clearances of albumin and Ficoll molecules of equivalent radius (36 Å) failed to reveal similar relations between albuminuria and glomerular barrier function. As shown in Figure 2, θ of Ficoll of 36 Å radius is similar in diabetic nephropathy compared to the value in control persons. Furthermore, in healthy controls, θ of Ficoll of 36 Å radius exceeds the estimated sieving coefficient of albumin by three orders of magnitude. As discussed below, losartan treatment lowered albuminuria significantly, whereas no corresponding change in θ of Ficoll of 36 Å radius was observed ( Figure 3 ). These disparities imply that the reduction in barrier size-selectivity in diabetic nephropathy probably have a minor influence on relatively small macromolecules such as albumin, at least in early diabetic renal disease. Thus, the glomerular filtration barrier may restrict albumin in addition to size, probably by charge- or shape selectivity as suggested by previous studies (93; 94; 96; 104; 115). Administration of losartan 50 mg o.d. significantly lowered mean arterial blood pressure, fractional clearance of albumin, IgG and the anionic subclass IgG 4 , whereas GFR and renal plasma flow remained unchanged ( Table 1 ). As illustrated in Figure 3 , losartan lowered θ for the large Ficoll molecules towards normal values, though only approaching statistical significance for the molecule of 60 Å radius (p = 0.06). Computations of membrane parameters according to the heteroporous model of size-selectivity (92), revealed that the partly repair in the size-selective defect during losartan therapy may contribute to the observed reduction in proteinuria by a reduction in the fraction of the filtrate permeating the shunt-like pores ( ω 0 ). Furthermore, losartan treatment induced a minor reduction in mean radius of restrictive pores. However, a more pronounced effect may be expected from a higher dose of losartan, i.e 100 mg o.d. as suggested by our other studies (80; 116; 117). The effect of blockade of the RAAS by ACE inhibition on size-selectivity has been investigated by Morelli et al in type 1 diabetic patients with diabetic nephropathy (17). Data indicated that treatment with enalapril 10 mg, shifted both restrictive pores in the glomerular membrane and non-restrictive pores of the shunt pathway toward smaller sizes in keeping with our data. Similar results were found later by Remuzzi et al in patients with advanced diabetic renal disease (118). Effects of RAAS blockade on charge selectivity in microalbuminuric type 1 diabetic patients were investigated by Hansen et al (119) but a favourable influence of ACE inhibition could not be demonstrated. Two studies in non-diabetic kidney disease similarly suggested that the reduction in proteinuria induced by ARB treatment was associated with reductions of glomerular shunt volume (120; 121). Furthermore, comparison of ACE inhibition and ARB treatment in one of the studies (121) demonstrated similar effects on membrane parameters. Hence, the beneficial effect of blockade of the RAAS on glomerular size-selective function is likely to be a consequence of inhibition of the action of angiotensin II.

In summary, functional studies of size selectivity in type 1 diabetic patients have demonstrated the existence of a defect in the large shunt-like pores in early and advanced diabetic nephropathy which may be partly restored by ARB or ACE inhibitor treatment. However, loss of barrier size selectivity may be of minor importance in the genesis of modest albuminuria, thus, contributions from charge or shape selectivity, and the influence of RAAS blockade on these should be further investigated.

4.3 Structural and molecular relationships
of glomerular permselectivity

Recently, studies of glomerular permeability have been performed to relate the functional properties of the glomerular membrane to the structural and molecular features at the cellular level as reviewed by Deen et al (90). Although it is generally acknowledged that the glomerular basement membrane restricts large plasma proteins, there is increasing evidence indicating that the ultimate barriers for proteins larger than albumin are the cellular layers and slit diaphragms (90; 122; 123). Foot processes are joined laterally by slit diaphragms. However, knowledge of structure and molecular configuration of the slit diaphragm has been rather limited. It represents a tiny membrane bridging the 30-40 nanometer flitration slit. The most frequently cited structure is that of Rodewald and Karnowsky (124) who described a slit diaphragm as made up of rod like units connected in center to a linear bar forming a "zipper-like" appearance. Discovery of the transmembrane protein nephrin as a major component of the slit diaphragm complex, provided a decisive progress in podocyte biology and pathophysiology (125-127). Lack of proper expression of the nephrin gene NPHS1 has been shown by Tryggvason et al (123), to be linked to the congenital nephrotic syndrome of the Finnish type, a glomerular disorder that leads to severe proteinuria, loss of podocyte foot processes and slit diaphragms (128). Similarly, homozygous knock-out mice generated by inactivation of the nephrin gene, are nephrotic and fail to develop foot processes (129). Thus, the massive proteinuria associated with mutation or inactivation of the NPHS1 gene suggests a key role for nephrin in the function of the glomerular filtration barrier. However, the function of nephrin is unclear. It has been hypothesized that nephrin molecules, which extend from adjacent podocytes, may interact in a homophillic manner, to form the possible "zipper-like" structure discussed above (123), but this remain speculative.

Human and experimental studies of diabetic glomerulopathy have demonstrated podocyte injury and loss, along with broadening of podocyte foot processes (130; 131). Bonnet et al (132) demonstrated that expression of the nephrin gene and protein was significantly reduced in spontaneously hypertensive streptozotocin-diabetic rats compared to control animals. However, whether the decrease in nephrin expression is the cause of proteinuria or a consequence of advanced renal injury is unknown, but changes in nephrin expression were blunted by ARB treatment, which also prevented proteinuria.

Similar results have been found in animal studies of non-diabetic kidney disease, in which changes in nephrin expression were abolished by ARB and ACE inhibitor treatment (133). Furthermore, ACE inhibition has been shown to be associated with preservation of slit diaphragm function and protein zonula occludens-1, a component of the slit diaphragm (134). Changes observed during RAAS blockade in nephrin expression and slit diaphragm function could be related to the reduction in intraglomerular blood pressure rather than a direct action on podocytes. On the other hand, podocytes contribute significantly to the permeability properties of the glomerulus (71) and because of their elaborate cytoskeleton, suggestive of a contractile function, they are likely to participate in regulation of glomerular filtration. Furthermore, the demonstration of AT1 receptors on the podocyte surface (69), suggest that angiotensin II may influence the permselective properties of the glomerular filtration barrier through contraction of foot processes (71).

Thus, renoprotective effects by RAAS blockade may be partly related to podocyte and slit diaphragm function mediated directly via AT1 receptors located on podocytes on glomeruli.

In summary, podocytes and slit diaphragms may be key determinants of the glomerular filtration barrier. Discovery of nephrin as a component of the slit diaphragm and the association with proteinuria has provided seminal progress in research of podocyte biology. Furthermore, data suggest that ARB's and ACE inhibitors ameliorate the reduction of nephrin expression simultaneously with prevention of proteinuria.

5.1 Antihypertensive treatment
in incipient nephropathy

The concept of incipient nephropathy or microalbuminuria is based upon the observation that elevated urinary albumin excretion, defined as persistent urinary albumin excretion rate between 30-300 mg/24 hours or 20-200 μ g/min, predicts progression to overt diabetic nephropathy in type 1 and type 2 diabetes (138-141). Furthermore, microalbuminuria is associated with increasing blood pressure and hyperfiltration (140). Early intervention studies in normotensive type 1 diabetic patients demonstrated beneficial effects of antihypertensive therapy with beta blockers (142) on regression of microalbuminuria. Numerous subsequent studies in type 1 as well as type 2 diabetes have confirmed the renoprotective effects of antihypertensive treatment in microalbuminuria, even though the drug of choice has been debated (143).

5.1.1 Type 1 diabetes mellitus

How should we treat microalbuminuric type 1 diabetic patients in 2002? ACE inhibitors have been recommended as renoprotective therapy for all type 1 diabetic patients regardless of blood pressure based upon studies by Mathiesen et al (144), Marre et al (145), Viberti et al (146). A recent meta-analysis of individual patient data from 12 trials with 698 patients, addressed the question whether all type 1 diabetic patients with microalbuminuria should receive an ACE-inhibitor (143). Selected studies included at least 10 patients, had a placebo or conventional treatment group and at least 1 year of follow up. As illustrated in Figure 4 , data showed that ACE inhibitors reduced the risk of progression to macroalbuminuria to approximately one third of that in the placebo group. Furthermore, regression to normoalbuminuria was three times greater as compared to the placebo group. At two years, the urinary albumin excretion rate was approximately 70% lower on average in the patients receiving ACE inhibitors weighed against the placebo group. According to Mathiesen et al (147), the beneficial effect of ACE inhibition in prevention of progression from microalbuminuria to overt nephropathy is long lasting and associated with preservation of normal kidney function.

One small study investigated the short-term antiproteinuric effect of losartan in 9 microalbuminuric type 1 diabetic patients (148). Proteinuria and blood pressure were significantly reduced after few days of treatment, suggesting a renoprotective effect.

Thus, accumulating data shows that ACE inhibitors prevent progression of microalbuminuria to overt nephropathy and preserves kidney function during long-term treatment. Consequently, ACE inhibitors must be regarded as first-line therapy for all type 1 diabetic patients with microalbuminuria regardless of blood pressure. Dual blockade of the RAAS with ACE inhibition and ARB may apply an additional beneficial effect and should be investigated in a long-term study in microalbuminuric type 1 diabetic patients.

5.1.2 Type 2 diabetes mellitus

Several studies have demonstrated beneficial renoprotective effects of antihypertensive treatment in type 2 diabetic patients with microalbuminuria (149-157). Originally, Ravid et al showed that normotensive patients treated with enalapril for a 7 year period, experienced an absolute risk reduction in progression from microalbuminuria to macroalbuminuria by 42 percentage points (149). Evidence has been conflicting in hypertensive patients with microalbuminuria regarding the existence of a specific renoprotective effect beyond the hypotensive effect of agents such as ACE inhibitors (150-158). Superior renoprotective effects of ACE inhibitors compared to conventional treatment has been reported by Lebovitz et al (153), Trevisan et al (157) and Chan et al (158), whereas similar effects of were found in the FACET (150), the ABCD (152) and the UKPDS (151) studies. The reasons for these conflicting results may in part be explained by short duration of antihypertensive treatment, blood pressure differences between treatment groups and small sample sizes. An exception from this is the long lasting UKPDS study suggesting equal effect of ACE inhibition and beta blockade.

Due to the inconclusive nature of the evidence, a multicenter study, IRMA-2 (159) was conducted. The study aimed to investigate the effect of the ARB irbesartan in hypertensive patients with type 2 diabetes and microalbuminuria. We included 50 patients at Steno Diabetes Center of 590 patients who were randomized to treatment with either irbesartan 300 mg, 150 mg or placebo in combination with conventional treatment for two years. The primary endpoint was time to progression to diabetic nephropathy defined as a urinary albumin excretion rate above 200 μg/min and at least a 30% increase from baseline. During the 24 months of follow up, diabetic nephropathy developed in 30 patients in the placebogroup as compared to 19 patients in the irbesartan 150 mg group (p = 0.08 vs. placebo) and 10 patients in the 300 mg group (p < 0.001 vs. placebo) ( Figure 5 ). The blood pressure levels were similar in the three groups, 144/83 mmHg on average throughout the study in the placebo group, 143/83 mmHg in the 150 mg group and 141/83 mmHg in the 300 mg group. The hazard ratio for progression to diabetic nephropathy was 0.56 in the 150 mg group (p = 0.05) and 0.32 in the 300 mg group (p < 0.001 vs. placebo) after adjustment for a small difference in baseline albuminuria and systolic blood pressure during the study. Albuminuria remained unchanged in the placebo group, but was significantly decreased by 24% and 38% during treatment with 150 mg and 300 mg, respectively. Furthermore, regression to normoalbuminuria was more frequent in the 300 mg group as compared to placebo treatment, 34% vs. 21% (p = 0.006). Thus, the IRMA-2 study demonstrated that irbesartan 300 mg is renoprotective beyond that attributable to lowering blood pressure in hypertensive type 2 diabetic patients with microalbuminuria. Recent interventions trials in type 2 diabetic patients with microalbuminuria have further established this conclusion (160; 161). Whether blockade of the RAAS by ACE inhibitors induce similar beneficial effects, will require a head-to-head comparative study to investigate. Consequently, ARBs must be considered first-line therapy in the treatment of hypertensive type 2 diabetic patients with microalbuminuria.

5.2 Antihypertensive treatment
in diabetic nephropathy

5.2.1 Type 1 diabetes mellitus

Antihypertensive treatment in diabetic nephropathy has become most successful since the first studies by Mogensen et al (162) and Parving et al (18) demonstrated that aggressive antihypertensive treatment with beta blockers and diuretics lowered rate of decline in GFR from more than 10 ml/min/year to less than 5 ml/min/year. Original studies dealing with ACE inhibition were reported in 1986 by Björck et al (163) followed by Parving et al (164) suggesting beneficial effects of ACE inhibition on rate of decline in GFR and albuminuria. In 1992, Björck et al (165) reported an open prospective study with 40 type 1 diabetic patients treated with either enalapril or metoprolol for a mean of 2.2 years. GFR declined by 5.6 ml/min/ year in the metoprolol group, but only 2.0 ml/min/year in the enalapril group (p = 0.02). These data indicated a specific renoprotective effect of ACE inhibition on kidney function in addition to what might be expected from blood pressure reduction alone. A similar conclusion was made by the Captopril Collaborative Study Group (12) who randomized 409 type 1 diabetic patients with diabetic nephropathy to treatment with either captopril or placebo in combination with conventional antihypertensive treatment for three years. A significant risk reduction by captopril treatment of 68% (39-83; 95% CI) (p < 0.001) in time to doubling of serum creatinine was found in 102 patients with a baseline serum creatinine above 133 μ mol/l, whereas the risk reduction for the 307 patients with baseline serum creatinine below 133 μ mol/l of 33% (-44-69) did not reach statistical significance (p = 0.31). The risk reduction for advanced nephropathy remained significant after adjustment for a difference in mean arterial blood pressure of 4 mmHg in favour of the captopril group.

A similar conclusion was made in a recent review (166) of randomized placebo controlled trials investigating the effect of ACE inhibitors in diabetic and non-diabetic kidney disease. In the ACE inhibitor groups, 115 of 700 patients developed ESRD or doubling of serum creatinine, as compared to 193 of 689 individuals in placebo groups. The aggregate relative risk for these endpoints was 0.60 (0.49-0.70; 95% CI) for individuals treated with an ACE inhibitor compared with placebo. Thus, ACE inhibition may have renoprotective effects in type 1 diabetic patients with diabetic nephropathy in addition to the benefit from lowering systemic blood pressure, at least in advanced nephropathy.

Is the effect of ACE inhibitors mediated by interference in the renin-angiotensin system? ACE is not a specific enzyme for conversion of angiotensin I, but has other substrates such as bradykinin. Bradykinin is a potent vasodilator acting through the release of prostacyclin, nitric oxide and endothelial-derived factors (167). Whereas animal studies have reached conflicting results (168-170), a human study suggested that bradykinin may play a role in the effects of ACE inhibitors on blood pressure and kidney function (171). A recent study investigated the intraindividual, intrarenal haemodynamic responses to ACE inhibition and ARB treatment in type 1 diabetic patients on a high-salt diet (172). High correlation between the individual responses to the two antihypertensive agents was reported, which may suggest, that reduced formation of angiotensin II is the primary action of ACE inhibition, rather than bradykinin accumulation.

Multiple studies have suggested the existence of alternative ACE independent pathways for angiotensin II generation, as reviewed by Hollenberg et al (79). Thus, angiotensin II formation may not be completely suppressed by ACE inhibition. On the other hand, specific blockade of the action of angiotensin II at the receptor level omits the possibility of contribution from bradykinin accumulation to the antihypertensive effect. We compared the short-term renoprotective effects of the ACE inhibitor enalapril to the effect of specific intervention in the RAAS by the ARB losartan in 16 hypertensive type 1 diabetic patients with diabetic nephropathy (80). The double blinded cross-over study consisted of five treatment periods each lasting two months. The patients received losartan 50 mg, 100 mg, enalapril 10 mg, 20 mg and placebo o.d. in random order. Both doses of the drugs significantly reduced albuminuria and 24 hour arterial blood pressure without significant differences between the high doses ( Figure 6 ). Mean arterial blood pressure decreased from 104 ± 2 mmHg (mean ± SEM) in the placebo period to 95 ± 2 and 96 ± 2 mmHg during treatment with losartan 50 and 100 mg, and to 98 ± 3 and 93 ± 3 by treatment with enalapril 10 and 20 mg, respectively (all p values < 0.05 vs. placebo). Albuminuria in the placebo period was 1156 (643-2080) mg/24 hours (geometric mean; 95% CI) and was significantly reduced by 33% (12-51; 95% CI) by losartan 50 mg, 44% (26-57) by losartan 100 mg, 45% (23-61) by enalapril 10 mg and 59% (39-72) during treatment with enalapril 20 mg. GFR remained unchanged from baseline (90 ± 6 ml/min/1.73 ml/min/m 2 ) in all treatment periods. Thus, the results demonstrated that blockade of the RAAS at the angiotensin II receptor level by losartan, induce similar renoprotective effects as inhibition of ACE by enalapril as evaluated by reduction in albuminuria and blood pressure in hypertensive type 1 diabetic patients with diabetic nephropathy. Therefore, the study suggests that the primary action of ACE inhibition is mediated by interference in the RAAS, i.e. reduced formation of angiotensin II. A similar outcome was found in an analogous study in non-diabetic patients (173).

The long-term renoprotective effects of losartan 100 mg o.d. on a principal renal end-point, i. e. Δ GFR, was investigated in our prospective study of 54 hypertensive type 1 diabetic patients with diabetic nephropathy homozygous for the I or D allele of the ACE/ID polymorphism with an average follow-up time of 36 months (174). Considering the total cohort disregarding genotypes, the annual rate of decline in GFR was 3.2 (2.2-4.7) (95% CI) ml/min/year with a 24 hours mean arterial blood pressure of 95 ± 1 mmHg (mean ± SEM) during three years of treatment with losartan 100 mg o.d, in combination with conventional antihypertensive treatment other than ACE inhibition if required. The study was not designed to compare the renoprotective efficacy of different treatment modalities, but demonstrates the capacity of an ARB to reduce the rate of decline in GFR in hypertensive type 1 diabetic patients with diabetic nephropathy. A rate of decline in GFR of 3.2 ml/min/year is considerably lower compared to the outcome of the largest ACE inhibitor study, the Captopril Collaborative Study (12) in which the average rate of decline in creatinine clearance was 8.0 ml/min/year in the captopril group and 10.8 ml/min/year in the patients treated with antihypertensive drugs other than ACE-inhibitors or calcium-channel blockers. Mean arterial blood pressure level in the captopril group was similar to our study (174) but assesed by office measurements, which complicates the comparison.

What is the long term effect of calcium channel blockers on kidney function in type 1 diabetes? Our group compared the long-term renoprotective effect of the ACE inhibitor lisinopril with the long-acting dihydropyridine calcium channel blocker nisoldipine in 51 hypertensive type 1 diabetic patients with diabetic nephropathy (175). Patients were followed for 4 years in a randomised double-blinded (single-blinded after 12 months) design. During the study period, mean rate of decline in GFR was similar in treatment groups, 6.0 ml/min/1.73 m 2 , mean arterial blood pressure levels 100 ± 2 mmHg (mean ± SEM) and 103 ± 2 mmHg in the ACE inhibitor and calcium channel blocker group, respectively. Albuminuria was lowered by approximately 50% by ACE inhibitor treatment, but remained unchanged in the calcium channel blocker group. However, levels of albuminuria were not significantly different in the two groups during the study.

In summary, ACE inhibition may confer renoprotective effects beyond the effect of lowering arterial blood pressure, however, blockade of RAAS by ARB treatment induce similar short-term renoprotective effects as inhibition of ACE. Long-term ARB treatment imply beneficial effects on preservation of GFR, but large-scale studies of ARBs in type 1 diabetes on a principal renal and cardiovascular endpoint should be performed. Based on the literature, ACE inhibitors must be considered the first-line drug in the treatment of diabetic nephropathy in type 1 diabetes. However, increasing evidence suggest that the effect of ACE inhibitors is mediated by interference in the RAAS which indicate that ARBs represent an effective alternative to ACE inhibition. Calcium channel blockers may represent an effective class of drugs and long-term studies, e.g. in supplementary treatment to RAAS blockade in diabetic nephropathy, should be performed.

5.2.2 Type 2 diabetes mellitus

A number of studies have been performed comparing the effect of antihypertensive treatment with and without ACE inhibition on rate of decline in GFR in type 2 diabetic patients with diabetic nephropathy (152; 153; 176-179). Except from one study (176), no significant differences between treatment groups were found (152; 153; 177-179). However, recent evidence from the RENAAL and IDNT studies have outlined the treatment of this group of patients (180; 181). Both studies were large multinational, placebo controlled trials, investigating the long term renoprotective effect of ARBs in hypertensive type 2 diabetic patients with diabetic nephropathy and elevated serum creatinine. In RENAAL, 1513 patients were randomized to treatment with losartan or placebo alone or in combination with conventional antihypertensive treatment except ACE inhibitors. Mean follow up time was 3.4 years and primary efficacy measure was the time to first event of the composite end point of a doubling of serum creatinine concentration, end-stage renal disease or death. Losartan treatment caused a 16% risk reduction of the primary composite end point (p = 0.02). Analysis of individual components revealed a lowering of risk by losartan for doubling of serum creatinine of 25% (p = 0.006), end-stage renal disease by 28% (p = 0.002), whereas no differences in mortality were found between groups.The estimated rate of decline in GFR was 4.4 ml/min/1.73 m 2 /year in the Losartan group compared to 5.2 ml/min/1.73 m 2 /year in the conventional treated group (p = 0.01). In the IDNT study 1715 patients were randomized to treatment with irbesartan, amlodipine or placebo, alone or in combination with conventional therapy except ACE inhibitors, for a mean of 2.6 years. The primary composite end point was similar as described for the RENAAL study. Treatment with irbesartan was associated with a 20% risk reduction for the primary composite end point compared with placebo (p = 0.02) and 23% compared with amlodipine (p = 0.006). The risk of a doubling of serum creatinine was 33% lower in the irbesartan group than in the placebo group (p = 0.003) and 37% lower compared to the amlodipine group (p < 0.001). In both studies there was a small difference in blood pressure control between groups, but statistical adjustment for the disparity had no impact on study conclusions.

Consequently, RENAAL and IDNT demonstrated renoprotective effects of treatment with ARBs on progression of diabetic nephropathy beyond that attributable to lowering blood pressure in type 2 diabetes. Similar results may be obtained by ACE inhibitors, but would require investigation in a large-scale head-to-head comparative study with an ARB. Therefore, first-line therapy of newly diagnosed hypertensive type 2 diabetic patients with diabetic nephropathy should be ARBs within the recommended range. However, blood pressure control in hypertensive type 2 diabetic patients is rarely obtained by montherapy. In the IDNT study, patients received on average ≥ 3 non-study antihypertensive drugs to control the blood pressure.

8.1 Effect of angiotensin II receptor blockade
on TGF- β

Intrarenal angiotensin II stimulates mesangial cell proliferation and induces expression of TGF- β independent of blood pressure, as discussed previously (47). Increased renal production of TGF- β is a feature of diabetes expressed by elevated urinary TGF- β excretion noted in patients with type 1 and type 2 diabetes and diabetic nephropathy (193-195). Therefore, reduced stimulation of TGF- β expression may be part of the renoprotective effect of RAAS blockade in diabetic nephropathy.

Various in vivo experimental models of renal disease including diabetes have demonstrated that blockade of RAAS decrease expression of TGF- β and matrix proteins. ARB treatment of streptozotocin diabetic rats in an experimental model with overexpression of angiotensin II, revealed marked reduction of renal TGF- β and collagen IV mRNA compared to untreated animals 12 weeks after induction of disease (188). Furthermore, RAAS blockade by ACE inhibition in streptozotocin diabetic rats normalised expression of TGF- β type II receptor mRNA and proteins compared to untreated animals 30 days after induction of diabetes mellitus (196). In consistence, a recent study in a rat model of non-diabetic kidney disease revealed similar beneficial effect of ARB treatment and ACE inhibition on lowering expression of renal TGF- β mRNA (197).

Effects of ARB treatment on plasma levels of TGF- β in humans were originally investigated in patients with chronic allograft nephropathy by Campistol et al, who found a significant reduction of plasma TGF- β by losartan treatment for 8 weeks (198). Esmatjes et al (189) investigated fourteen type 2 diabetic patients with mild hypertension and microalbuminuria in an open study. Plasma TGF- β was higher in diabetic subjects compared to healthy controls, but significantly lowered by losartan treatment in seven patients with TGF- β levels at baseline above the mean. Levels remained unchanged in patients with TGF- β levels below the mean value. In conflict with these data, plasma TGF- β was unaffected by losartan in a recent study of a similar group of patients, whereas urinary TGF- β was significantly reduced (190). Furthermore, urinary TGF- β excretion at baseline correlated closely with indexes of glycemic control, possibly reflecting a relation between glucose concentration and TGF- β . Effect of RAAS blockade on TGF- β in diabetic nephropathy was investigated in serum samples of 58 type 1 diabetic patients from the Collaborative Study Group Captopril Trial (12; 199). Captopril significantly lowered serum TGF- β compared to an increase in the placebo group six months after initiation of treatment. Furthermore, inverse correlations of percentage changes in serum TGF- β during the six month period and the relative decline in GFR during the following two years of follow-up were found in both the captopril and placebo groups.

In summary, blockade of the RAAS is likely to reduce glomerular scarring, indicated by reductions in urinary TGF- β levels, thus, TGF- β may represent a supplementary therapeutic target in the treatment of diabetic nephropathy. Further studies evaluating effects of ARBs on urinary and blood levels of TGF- β should be performed.

8.2 Effect of angiotensin II receptor blockade
on circulating adhesion molecules

Type 1 diabetic patients with diabetic nephropathy are at extremely high risk of atherothrombotic disease (200). Adherence of circulating leukocytes to endothelial cells and their subsequent migration into the arterial intima is an early feature in the formation of atherosclerotic lesions (201). The endothelial attachment of leukocytes is mediated by an increased expression of adhesion molecules such as intercellular adhesion molecule-1 (ICAM-1), vascular adhesion molecule-1 (VCAM-1), and E-selectin. Soluble forms (e.g. sICAM-1) have been detected in plasma and are thought to reflect shedding of membrane bound forms (202). Angiotensin II induces expression of VCAM-1 on endothelial cell surfaces in animal studies, which may be inhibited by ARB treatment, thus mediated by the AT1 receptor (203).

A potential mechanism for increased expression of sVCAM-1 and accelerated vasculopathy in diabetes may be related to interaction between advanced glycation endproducts and their endothelial receptors (202). A recent study in type 2 diabetes (201) demonstrated that increased levels of sVCAM-1 are independently associated with the risk of cardiovascular mortality. Elevated levels of circulating adhesion molecules in type 1 diabetic patients with diabetic nephropathy have recently been demonstrated (204; 205). However, data dealing with the effect of blockade of the RAAS on circulating adhesion molecules were not available.

We performed a double-blind five-armed cross-over study investigating the effect of ARB treatment by losartan 50 and 100 mg and ACE inhibition by enalapril 10 and 20 mg compared to placebo on circulating adhesion molecules (206). Furthermore, values were determined in 29 healthy control persons. Plasma levels of sVCAM-1, sE-selectin and vWF were elevated during the placebo period as compared to the control persons (p < 0.05). General endothelial dysfunction was suggested by elevated vWF in all treatment periods. However, sVCAM-1 and sE-selectin were significantly decreased by RAAS blockade, even though lowering of sE-selectin did not reach statistical significance during losartan treatment (p = 0.08). Thus, inactivation of angiotensin II by ARB or ACE inhibition may reduce the proatherogenic leukocyte-endothelial adhesion in type 1 diabetic patients with diabetic nephropathy. Evidence from previous intervention studies is limited, but elevated plasma levels of sVCAM-1, sICAM-1 and sE-selectin have been reported in hypertensive type 2 diabetic patients with microalbuminuria (207) and non diabetic obese patients (208). Similarly, plasma levels of sVCAM-1 were reduced by RAAS blockade.

If plasma concentrations of sVCAM-1 are independently associated with the risk for cardiovascular disease, as suggested by the Hoorn study (201) and potentially modifiable by RAAS blockade, should sVCAM-1 be measured in all diabetic patients? According to present data, RAAS blockade may have an antiatherogenic effect in diabetic patients with elevated albumin excretion rate. sVCAM-1 may be elevated even in normoalbuminuric patients (205), even though most studies report values within normal range (204; 209). Before routine assessments of adhesion molecules are considered, further prospective studies are required to investigate the pathophysiological significance of increased expression of adhesion molecules and the potential beneficial effect of RAAS blockade on the cardiovascular risk profile.

10.1 ACE insertion/deletion genotypes
and angiotensin II receptor blockade
in diabetic nephropathy

Specific blockade of the AT1 receptor may overcome this suggested interaction between ACE genotype and ACE inhibition and potentially induce similar renoprotective effects independent of genotypes. We performed a study to evaluate the short-term renoprotective effect of ARB treatment by losartan assessed by reduction in albuminuria in hypertensive type 1 diabetic patients with diabetic nephropathy (116). Fifty-four patients homozygous for I (n = 26) D or (n = 28) allele of the ACE/ID polymorphism were investigated. Patients and the primary investigator were blinded to genotypes. The study consisted of two treatment periods each lasting two months. Patients received two doses of losartan starting with 50 mg followed by 100 mg o.d. in the second treatment period. No significant influence of the ACE/ID polymorphism on the antiproteinuric effect of losartan was found ( Table 2 ). Albuminuria, assessed by at least two 24 hour urine collections, was lowered by 55% (35-68; 95% CI) (p < 0.01 vs. baseline) in the II group and 46% (28-61) (p < 0.01 vs. baseline) in the DD group by treatment with losartan 100 mg (p < 0.01) (NS between groups). Similarly, 24 hour arterial blood pressure was significantly reduced by both doses of losartan without significant differences between genotype groups. Levels of ACE in serum were as expected approximately 50% higher in DD patients compared to the II group at baseline and remained stable during treatment. In contrast, renin and angiotensin II increased similarly in both groups in keeping with blockade of the RAAS (Table 2). Thus, blockade of the action of angiotensin II at the AT1 receptor may be independent of ACE genotype, as evaluated by changes in albuminuria and blood pressure. However, to investigate the long-term beneficial effect of blockade of the AT1 receptor on a principal renal endpoint such as rate of decline in GFR, the patients continued in a prospective trial with an average follow-up time of 36 months (174). Patients continued treatment with losartan 100 mg o.d. whereas additional antihypertensive therapy was given to reach a target blood pressure below 135/85 mmHg. Twenty-four hour arterial blood pressure, GFR and albuminuria were measured every sixth month, as illustrated in Figure 9 . Mean arterial blood pressure levels were significantly reduced to an average of 95 ± 1 mmHg in both genotype groups during follow up. We found no significant difference between rate of decline in GFR in the two genotype groups: GFR decreased by 2.9 (2.0-4.2) ml/min/year in the II group compared to 3.4 (2.3-5.1) ml/min/year in the DD group during follow-up (p = 0.4 II vs. DD). Thus, in contrast to previous studies dealing with renoprotection by ACE inhibiton, long-term treatment with an ARB conferred similar beneficial effects on progression of diabetic nephropathy in II and DD genotypes. Concentrations of angiotensin II in plasma were similar in both genotype groups throughout the study in accordance with previous data (235). Higher levels of ACE in the DD patients could be expected to result in elevated concentrations of angiotensin II. Enhanced pressor response to exogenous angiotensin I, probably caused by increased angiotensin II formation in individuals homozygous for the D allele as compared to II persons, has been reported in a number experimental and human studies (236-238). Systemic levels of angiotensin II do not specifically reflect RAAS activity at tissue level, as discussed previously. Possible differences in renal RAAS activity between II and DD patients can not be evaluated from systemic concentrations.

The studies discussed above are included in the recent defined concept of pharmacogenetics, i.e. how genetic differences influence the variability in individual patient responses to pharmacological intervention (239). Uncovering gene polymorphisms in patients with similar disease phenotypes may identify groups of patients who are most likely or unlikely to benefit from treatment of particular drugs. Treatment strategies and modalities may be individualized according to genotypes, which may also become a consequence of our studies if results are confirmed in large scale trials.

In summary, several studies have found an increased risk for renal function loss in DD patients compared to II patients, even during ACE inhibition. However, ARB treatment may confer similar long-term renoprotective effects in both genotypes. Large scale comparative trials with ACE inhibition should be performed before recommendations of selective therapy in different ACE/ID genotypes can be justified. Pharmacogenetic studies related to other RAAS polymorphisms may also be considered.

11.1 Effect of dual blockade of RAAS
in incipient diabetic nephropathy

Dual blockade of RAAS was investigated by the CALM study group (242) in 199 type 2 diabetic patients with hypertension and microalbuminuria. Combination therapy with candesartan 16 mg o.d. and lisinopril 20 mg o.d. was significantly more effective in reducing both systolic and diastolic blood pressure compared to monotherapy with either agent. Reduction of microalbuminuria tended to be greater with combination therapy, though only reaching statistical significance compared to Candesartan monotherapy. As discussed previously, ARBs must be considered first-line therapy in hypertensive type 2 diabetic patients with microalbuminruia according to the IRMA-2 study (159), but additional antihypertensive and antiproteinuric effect may be achieved by dual blockade of RAAS. Data of dual blockade in incipient nephropathy in type 1 diabetes are not available.

11.2 Effect of dual blockade of RAAS
in diabetic nephropathy

Rossing et al (243) investigated hypertensive type 2 diabetic patients with diabetic nephropathy with an insufficient response to previous antihypertensive medication, i.e. albuminuria > 1000 mg/24 h as well as arterial blood pressure > 135/85 mmHg despite treatment with at least two antihypertensive agents including maximal recommended dose of ACE inhibition. Eighteen patients were randomized in a double-blind crossover design to treatment with Candesartan 8 mg o.d. and placebo for two months in addition to previous antihypertensive medication. Dual blockade of RAAS lowered systolic blood pressure by 10 mmHg compared to placebo, whereas albuminuria was reduced by 25%. A similar study in type 1 diabetic patients with diabetic nephropathy has been performed by Jacobsen et al (244). Irbesartan 300 mg o.d. lowered albuminuria and diastolic blood pressure were significantly lowered by 37% and 5 mmHg, respectively, as compared to the placebo period, when Irbesartan was added to previous antihypertensive medication. Furthermore, plasma renin increased significantly during Irbesartan treatment, indicating an additional blockade of RAAS. A similar alteration in plasma renin during dual blockade in diabetic nephropathy has previously been reported (245). Hyperkalaemia was observed in patients with GFR < 35 ml/min/1.73 m 2 , but were not seen in patients with higher GFR levels.Thus, dual blockade of the RAAS may induce additional beneficial effects in therapy resistant patients with diabetic nephropathy as compared to conventional treatment.

The COOPERATE study (246) investigated the effect of monotherapy with either ACE inhibition or ARB compared to dual blockade with combined treatment in 263 patients with non-diabetic kidney disease followed for 2.9 years. In the combination group, 11% of the patients reached the combined primary endpoint, time to doubling of serum creatinine or end-stage renal disease compared to 23% in both monotherapy arms (p ≤ 0.018).

Thus, long-term studies of dual blockade in diabetic kidney should be initiated, however, dose-response relationships for optimal renoprotecive effects of monotherapy with both agents should be clarified before further action is taken.