OUP user menu

Reducing cardiovascular risk: protecting the kidney

Daniela Dobre, Hiddo J. Lambers Heerspink, Dick de Zeeuw
DOI: http://dx.doi.org/10.1093/eurheartj/sup027 F39-F46 First published online: 17 November 2009


Progressive decline of renal function in chronic kidney disease (CKD), measured by a reduced glomerular filtration rate or albuminuria, is linked to an increased risk of cardiovascular (CV) disease. Angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs), most likely because of their blockade of the pathophysiological effects of angiotensin II, provide renoprotection and are the treatment of choice in patients with CKD including those with diabetes. This renoprotective property appears to be partly independent of the blood pressure-lowering effect of drugs that intervene in the renin–angiotensin–aldosterone system. Recent data from the ONTARGET study show that the ARB telmisartan has similar CV and renal protection as the ACE-inhibitor ramipril in patients at risk but is better tolerated. Although extensively used by nephrologists, no additional benefit was observed with the combination of telmisartan plus ramipril compared with ramipril alone on the composite renal endpoint.

  • Angiotensin-converting enzyme inhibitor
  • Angiotensin II receptor blocker
  • Cardiovascular disease
  • Chronic kidney disease
  • Renal dysfunction
  • Ramipril
  • Telmisartan


The initial stages of renal dysfunction are usually asymptomatic. Early detection is essential to preserve renal function as much as possible and prevent progressive damage, not only to the kidneys, but also to the cardiovascular (CV) system. Different approaches are used to quantify renal dysfunction. Measurement of serum creatinine levels is used because of its simplicity; however, early renal disease cannot be detected this way because creatinine levels are insufficiently elevated. The serum creatinine levels can be converted to a so-called estimated glomerular filtration rate (eGFR) using equations such as used in the Modification of Diet in Renal Disease (MDRD) study; this takes into account age, gender, and race.1 Glomerular filtration rate can also be measured more accurately using substances that are not reabsorbed by the kidney (e.g. inulin or iohexol).2,3 A (e)GFR of >60 mL/min/1.73 m2 is usually considered normal, unless there is other clinical/laboratory evidence of other renal abnormalities.4

An alternative definition of renal dysfunction is the presence of increased levels of protein (usually albumin) in the urine. Urinary albumin excretion rate (UAER) is determined by the timed collection of urine (usually overnight or over 24 h). Microalbuminuria is defined as the urinary levels of albumin in the range 30–299 mg/24 h. Macroalbuminuria, or overt nephropathy, is present if urinary albumin levels are ≥300 mg/24 h.5 As an alternative to the cumbersome and time-consuming determination of UAER, the urinary albumin–creatinine ratio (UACR) in an early morning spot urine has become widely accepted.6 Microalbuminuria is defined as a UACR in the range 30–300 mg/g.5

Other biochemical parameters that may indicate decrease in renal function are haemoglobin levels and vitamin D levels. The kidneys are involved in the pathways of both these substances. With decreasing renal function, one also sees both haemoglobin and vitamin D levels going down.79

Using a GFR of ≤60 mL/min/1.73 m2 to define renal dysfunction, the Gubbio Population Study found that the prevalence in the general population varied according to age: <1% for ages 18–24 years to >30% in those aged 75 years.10 In the Prevention of Renal and Vascular End-stage Disease (PREVEND) population study, ∼7% of the adults between the ages of 28 and 75 years living in Groningen displayed microalbuminuria.11 By the detection of anaemia (using the definition of haemoglobin <11 g/dL), the Third National Health and Nutrition Examination Survey (NHANES III) estimated that the prevalence of chronic renal dysfunction is of the order of 10% in the USA.12

Correlation between glomerular filtration rate and cardiovascular disease

There is a well-established association between a reduced GFR and CV disease. Recent large studies in different populations all show the same: the lower the GFR, the higher CV risk a subject has. Go et al.13 demonstrated in over 1 million patients in the database of a healthcare provider that there was a direct relationship between the age-adjusted rate of CV events and the stage of renal dysfunction as defined by the National Kidney Foundation according to the eGFR.4 Similar correlations were also detected between the incidences of death and hospitalization and eGFR. In the Second National Health and Nutrition Examination Survey (NHANES II), participants who were 30–74 years old at baseline were followed up for 16 years. Subjects with an eGFR < 70 mL/min had a higher risk of CV mortality and all-cause mortality when compared with those with eGFR > 90 mL/min.14 The Cardiovascular Health Study has demonstrated that level of eGFR is an independent risk factor for CV disease, de novo CV disease, and all-cause mortality in the elderly (age >65 years).15 Finally, the Atherosclerotic Risk In Communities (ARIC) cohort study found that level of GFR is a risk factor for de novo and recurrent CV events, with an interaction (P = 0.007) between GFR and left ventricular hypertrophy (LVH; a decreased GFR being a stronger risk factor in subjects with LVH).16

Correlation between degree of albuminuria and cardiovascular disease

It is likely that leakage of albumin in the urine reflects not only renal damage but widespread vascular damage, and, as such, higher susceptibility to atherosclerosis and CV disease. The exact mechanism for the relation between albumin leakage from the kidney and vascular damage at other places is still not clear. However, recent data show that the vasculature contains a barrier function to albumin leakage, a so-called glycocalyx layer containing large amounts of negatively charged molecules which repel the negatively charged serum albumin.17 This layer is present in all capillary beds. The decrease in function of this layer will result in more albumin leakage in the kidneys as well as in other places in the body. This may well form a mechanism by which general vascular dysfunction (as reflected by albumin leakage) leads to general atherosclerosis and CV events.

Whatever the mechanism, there is overwhelming evidence that albuminuria is a risk predictor of CV disease in the general population, as well as in patients at increased risk, such as those with hypertension and diabetes. Arnlov et al.18 examined the association between urinary albumin excretion (UAE) and incidence of CV disease events in a sample of 1568 non-hypertensive and non-diabetic individuals from the Framingham cohort. After a median follow-up of 6 years, the incidence of CV events was incrementally higher with increasing UAE baseline levels. The data showed that even low levels of albumin excretion, below the microalbuminuria threshold, predicted the development of a composite of CV events. The group with UACR values at or above the sex-specific median was associated with three-fold risk for developing CV disease compared with the group below the median values. Intriguingly, the risk profile of the two groups (high and low albuminuria) was identical as far as classical CV risk factors (such as elevated blood pressure, cholesterol, or smoking, etc.) were concerned.

Data of 8206 patients with hypertension and LVH from the Losartan Intervention For Endpoint reduction in hypertension (LIFE) study showed similar results. The risk of the composite endpoint (CV death, fatal or non-fatal stroke, and fatal or non-fatal myocardial infarction) over the median follow-up of 4.8 years rose continuously as the baseline albuminuria increased.19 Similar to findings from a previous study, the data showed that there was no specific threshold for increased CV risk; even within the normoalbuminuria range, the incidence of CV events increased with urinary albumin levels. Again, other CV risk factors did not explain the predictive power of albuminuria for CV outcome.

In the Type 2 diabetes, Hypertension, Cardiovascular events and Ramipril (DIABHYCAR) study, Vaur et al.20 found a logarithmic relationship between level of albuminuria and heart failure incidence; each 10-fold increase in UAE was associated with double risk of heart failure.

Glomerular filtration rate or albuminuria?

Recently, Brantsma et al.21 studied which of the renal parameters (GFR or albumin) is the best for CV risk prediction. They showed that albuminuria is in fact the dominating factor in the prediction of CV risk. Most likely, eGFR predicts CV risk because patients with lower GFR have a higher chance of having albumin leakage.

Pharmacological renoprotection: the role of angiotensin-converting enzyme inhibitors and angiotensin II receptor blockers

Angiotensin II causes damage to the kidney, leading to renal dysfunction, by a number of mechanisms.22 These include: pressure-induced renal injury due to systemic and glomerular hypertension; ischaemia-induced renal injury secondary to intrarenal vasoconstriction and decreased renal blood flow; tubular injury secondary to angiotensin-induced proteinuria; and activation of renal fibroblasts leading to oxidative stress, local inflammation, and stimulation of vascular and mesangial cell proliferation and hypertrophy. Using angiotensin-converting enzyme (ACE)-inhibitors and angiotensin II receptor blockers (ARBs), which act upon the renin–angiotensin system (RAS) by limiting the production of angiotensin II and blocking binding of angiotensin II to the angiotensin II Type 1 receptor, respectively, should, therefore, confer renoprotection by both haemodynamic and non-haemodynamic mechanisms.

The risk of renal dysfunction is particularly high in patients with diabetes, and the evidence for the renoprotective effects of ACE-inhibitors comes largely from patients with diabetic nephropathy. The Collaborative Study Group showed captopril to be renoprotective in Type 1 diabetes.23 In Type 2 diabetes, the Diabetics Exposed to Telmisartan And enalaprIL (DETAIL) study showed renoprotection to be equal between ACE-inhibitors and ARBs (Figure 1).2426 Recently, the Renoprotection of Optimal Antiproteinuric Dose (ROAD) trial determined that both benazepril and losartan achieved similar renal outcomes in non-diabetic renal disease.27

Figure 1

Comparison of the effect of telmisartan as observed in the Diabetics Exposed to Telmisartan And enalaprIL (DETAIL) study25 on the progressive loss of renal function in patients with hypertension and Type 2 diabetes.

Non-haemodynamic effects of renin–angiotensin system intervention

Numerous studies have demonstrated the renoprotective effects of RAS intervention at different stages of renal dysfunction in diabetic patients and suggest that there is a non-blood pressure-lowering component.

In the placebo-controlled Bergamo Nephrologic Diabetes Complications Trial (BENEDICT), which was conducted in normoalbuminuric patients receiving monotherapy with trandolapril 2 mg, or sustained release verapamil 180 mg, the combination of both, or placebo, a lower incidence rate of microalbuminuria was observed (overnight albumin excretion ≥20 µg/min), whereas the effect of verapamil alone was similar to that of placebo.28 All treatment groups were allowed additional classes of antihypertensives to achieve the target SBP/DBP of 120/80 mmHg. Incidences of microalbuminuria were 5.7% with trandolapril plus verapamil, 6.0% with trandolapril, 11.9% with verapamil, and 10% with placebo over a median period of 3.6 years. The authors concluded that the effects of trandolapril plus verapamil and of trandolapril alone in preventing microalbuminuria exceeded expectations based on changes in blood pressure alone.

Irbesartan 150 or 300 mg given once daily over a period of 2 years prevented the onset of overt nephropathy, defined as persistent albuminuria and UAER > 200 µg/min and ≥30% higher than the baseline value in the placebo-controlled Irbesartan in Patients with Type 2 diabetes and MicroAlbuminuria (IRMA-2) trial.29 In the IRMA-2 trial, irbesartan reduced albuminuria in a dose-related manner compared with placebo independent of its blood pressure-lowering effect, as measured by 24 h ambulatory blood pressure measurement at baseline and 2 years post-randomisation.30

The Reduction of Endpoints in NIDDM with the Angiotensin II Antagonist Losartan (RENAAL) trial was conducted in patients with Type 2 diabetes and overt nephropathy [UACR ≥ 300 mg/g, or urinary protein excretion rate ≥500 mg/24 h, plus serum creatinine 1.3–3.0 mg/dL (115–265 µmol/L)]. The patients were treated for up to 4 years. Losartan 50–100 mg once daily (the 100 mg dose was used if SBP ≥ 140 mmHg), in comparison with placebo, significantly reduced the incidence of doubling of serum creatinine concentrations, the development of end-stage renal disease (ESRD) defined as the requirement for renal replacement therapy, or all-cause mortality.31 Statistical analyses that corrected for the small blood pressure differences between losartan and placebo showed that the renal protection conferred by losartan exceeded that attributable to any blood pressure differences.

The Irbesartan in Diabetic Nephropathy Trial (IDNT) showed that, in patients with more advanced overt nephropathy [urinary protein excretion rate ≥900 mg/day plus serum creatinine 1.0–3.0 mg/dL (88–265 µmol/L) in women and 1.2–3.0 mg/dL (106–265 µmol/L) in men], the primary composite endpoint of a doubling of the baseline serum creatinine concentration, the development of ESRD, or death from any cause was 23% lower over a mean follow-up of 2.6 years in patients treated with irbesartan 300 mg than in those treated with amlodipine 10 mg.32 Again, the conclusion was the renoprotective effect of irbesartan was independent of the reduction in blood pressure.

Are all angiotensin II receptor blockers equally renoprotective?

Studies directly comparing the renoprotective effects of different ARBs are not published to our knowledge, at least when looking at hard outcome measures. Comparisons of different ARBs with respect to their effect on surrogate markers such as blood pressure or albuminuria have been done. An example is the telMisartan vs. losArtan in hypertensive Type 2 DiabEtic patients with Overt nephropathy (AMADEO) trial. As would be expected from other studies evaluating ARBs,25,2932 both study drugs had a beneficial effect on renal function in patients with a morning spot urine protein–creatinine ratio (UPCR) ≥ 700 mg/g at baseline.33 However, at the end of the study, the lowering of UPCR was significantly greater (P = 0.027) in the telmisartan (29.8%) than in the losartan group (21.4%; Figure 2). Despite the statistically significant difference in favour of telmisartan on the reduction of proteinuria at the end of treatment, the reductions in blood pressure were similar with telmisartan and losartan. Therefore, a blood pressure lowering-independent effect suggests that the structural differences of the two ARBs impacted on the non-haemodynamic component of renoprotection. However, interpretation of these data has to be taken with care, since dose comparison over a wide range may be necessary to really determine a potential difference of ARBs on proteinuria. The recent Supra Maximal Atacand Renal Trial (SMART) showed that extremely high doses of candesartan still decrease proteinuria further, whereas blood pressure effect has already reached maximum at much lower doses.34

Figure 2

Comparison of the effects of telmisartan 80 mg vs. losartan 100 mg on urine protein–creatinine ratio (UPC) in patients with Type 2 diabetes and overt nephropathy. Reproduced with permission from Bakris et al.33

Controlling proteinuria rather than blood pressure

All the above studies on drugs that intervene in the renin–angiotensin–aldosterone system and renal protection have either used fixed dose of the drugs (usually at the recommended dose) or have titrated the drugs to the blood pressure. Hence, blood pressure differences between the treatment groups were small. The ROAD study27 is different, since it is the first study that tests renoprotection of an ACE-inhibitor and an ARB titrated to the lowest level of proteinuria. This study was conducted in patients with chronic kidney disease (CKD) [serum creatinine 1–5 mg/dL (88–442 µmol/L), persistent proteinuria >1 g/24 h]. At baseline, 57% of the patients were hypertensive. During the 8-week pre-titration period of the study, patients received open-label benazepril 10 mg or losartan 50 mg; additional antihypertensive(s) were given if SBP/DBP > 130/>80 mmHg after 4 weeks. After 8 weeks, the dose of benazepril or losartan was uptitrated to a maximum of 40 or 200 mg, respectively, in half the patients for optimal control of proteinuria. These doses were maintained throughout the follow-up period, a median of 3.7 years. The optimal doses of benazepril (20.8 mg) and losartan (117.7 mg) resulted in a 51% (P = 0.028) and 53% (P = 0.022) reduction, respectively, in the risk of doubling of serum creatinine, ESRD, or death compared with conventional doses of the same drugs. Despite increasing the dose, the blood pressure control was similar in patients receiving conventional or up-titrated doses. There were also no differences in major adverse events. Thus, there is further evidence that the renoprotective effect of the pharmacological targeting of the RAS has a non-blood pressure control-related component and that higher doses (based on proteinuria lowering) than those used for treating hypertension may be appropriate.

The control of excretion of protein in the urine is not only important in the management of CKD, but also in reducing the risk of CV morbidity and mortality. An analysis of the data from the LIFE study showed that losartan brought about a greater reduction in albuminuria during the study than did atenolol,35 thus once again emphasising the importance of targeting the RAS. A further post hoc analysis of the data demonstrated that a reduction in albuminuria improves the CV prognosis and that this reduction in CV events was not explained by the reduction of blood pressure during treatment.36 A recent meta-analysis has also shown that over and above blood pressure reduction, ACE-inhibitors and ARBs offer greater protective effects against congestive heart failure than calcium-channel blockers.36

Double blockade of the renin–angiotensin system

A systematic review and meta-analysis of randomized trials evaluating the combination of an ACE-inhibitor and an ARB in patients with microalbuminuric and proteinuric renal disease with or without diabetes was recently published.37 The combination of an ARB and an ACE-inhibitor brought about a further decrease in proteinuria of 25% compared with single-agent RAS blockade. These results are in accordance with a previous meta-analysis demonstrating that the addition of an ARB to ACE-inhibitor therapy significantly decreases proteinuria compared with ACE-inhibitor therapy alone (440 mg/day; 95% CI 289–591). In this meta-analysis combination therapy with an ACE-inhibitor and an ARB resulted in a non-statistically significant trend towards a decrease in GFR of 1.4 mL/min (95% CI −2.6 to 0.2).

The combination treatment of angiotensin-II receptor blocker and ACE-inhibitor in non-diabetic renal disease (COOPERATE) is the only hard endpoint trial that studied the effect of ACE-inhibitor and ARB combination in a (non-diabetic) renal population.38 Treatment with a combination of trandolapril 3 mg and losartan 100 mg for 3 years significantly reduced the incidence of the composite endpoint of doubling of serum creatinine or ESRD compared with either trandolapril (P = 0.018) or losartan (P = 0.011). Similar blood pressure reduction was achieved with the combination or either monotherapy, but median albuminuria was lower (P = 0.01) at the end of the study with the combination than with either monotherapy. However, the validity of the COOPERATE trial has recently been questioned due to discrepancies between reported statistical methods and published results.39

Renoprotection in ONTARGET

ONgoing Telmisartan Alone and in combination with Ramipril Global Endpoint Trial (ONTARGET) was designed to determine whether the CV protection provided by telmisartan was non-inferior to that of ramipril and whether use of a combination of telmisartan and ramipril, by more complete blockade of the RAS, would further reduce CV morbidity and mortality in high-risk patients.40 Data on the renoprotective effects on telmisartan, ramipril, and telmisartan plus ramipril from ONTARGET were reported in a recent publication by Mann et al.41 The renal outcomes of ONTARGET are surprising, since the combination therapy did lower blood pressure further and slowed the increase of albuminuria more than in the single treatment groups; however, the combination therapy did not offer more renal protection.42 There are probably a number of explanations.

In ONTARGET, CKD was not an entry criterion.40 At baseline, mean serum creatinine was 1.06 mg/dL (94 µmol/L) and mean eGFR was 73.6 mL/min/1.73 m2; 24% of 25 551 subjects evaluated had eGFR < 60 mL/min/1.72 m2 and eGFR was <30 mL/min/1.73 m2 in only 263 patients.41 Microalbuminuria was present in 13.5% and macroalbuminuria in 4.0%.

At the end of the study, there was an increase in albuminuria that was greater in patients receiving ramipril than in those treated with telmisartan (P = 0.033) or with telmisartan plus ramipril (P = 0.0028; Table 1).41 It should also be noted that telmisartan was better tolerated than ramipril, with fewer discontinuations.43

View this table:
Table 1

Effect of telmisartan 80 mg, ramipril 10 mg, and telmisartan on urinary albumin–creatinine ratio during ONTARGET40

TimepointMean UACR (95% CI)
TelmisartanRamiprilTelmisartan + ramipril
Baseline (mg/g)0.83 (0.80–0.86)0.81 (0.78–0.84)0.81 (0.78–0.84)
Two yearsa1.08 (1.05–1.12)1.17 (1.13–1.20)1.05 (1.02–1.08)
End of studya1.25 (1.20–1.29)*1.32 (1.27–1.37)1.22 (1.17–1.26)**
  • aRatio to baseline.

  • *P = 0.033 vs. ramipril.

  • **P = 0.0028 vs. ramipril.

When the composite renal outcome of a doubling of serum creatinine, dialysis or death was examined, its incidence for telmisartan was 13.4% and for ramipril was 13.5%. However, for telmisartan plus ramipril, the incidence was 14.5% (P = 0.037), suggesting that the combination may be less renoprotective.40 This observation needs to be put in perspective. Of the components of the composite endpoint, the incidence of death (telmisartan 11.6%, ramipril 11.8%, and telmisartan plus ramipril 12.5%) far exceeded the incidence of true renal endpoints of a doubling of serum creatinine and dialysis (telmisartan 2.38%, ramipril 2.18%, and telmisartan plus ramipril 2.62%). The majority of deaths were due to cardio- or cerebrovascular disease, as opposed to chronic renal disease.

The Kaplan–Meier curves for renal dialysis show that, even in the first year of the study, dialysis (whether acute or chronic) was required more frequently in patients receiving telmisartan plus ramipril (Figure 3).40 Because of the cumulative nature of a Kaplan–Meier plot, this difference is carried forward, but there were no significant differences between treatments in the cumulative rates at the end of the study. During the study, 98 patients required chronic dialysis and 74 underwent acute dialysis. At the start of treatment, the higher incidence of dialysis in the patients receiving telmisartan plus ramipril may be attributed to the need for acute dialysis due to hyperkalaemia, which was identified as a side effect in the meta-analysis of ACE-inhibitor/ARB combination therapy,42 rather than chronic dialysis for renal ESRD. This speculation needs to be confirmed. Notably, the majority of the adverse effects were observed for acute dialysis, which is potentially reversible, and is an outcome of less significance than long-term dialysis for ESRD. No separate effect on the risk of chronic renal failure (beneficial or harmful) was demonstrated.

Figure 3

Kaplan–Meier curves for dialysis in patients treated with telmisartan, ramipril, or telmisartan plus ramipril. Reproduced with permission from Mann et al.41

The changes in eGFR followed a similar pattern to the incidence of dialysis, with a greater reduction in patients receiving telmisartan plus ramipril than in those treated with either telmisartan or ramipril (Figure 4) when treatment was first started.41 After this initial difference in eGFR between the combination group and the single treatments, the curves for the different treatment are parallel, hence no difference in GFR slope between the treatments. Nephrologists often observed a decline in GFR when antihypertensive therapy with RAS intervention is initiated in patients with CKD.4446 This is probably due to a reduction in the high intraglomerular pressure due to dilation of the renal efferent arterioles.47 However, this initial decline in renal function due to haemodynamic changes does not have a negative impact, but rather a positive impact on long-term renoprotection; the more GFR drops initially, the better the patient is protected in the long run.48 One could thus speculate that the steeper initial fall of eGFR in the combination group of ONTARGET could be associated with better long-term renal protection. The reason why this was not observed in the trial may have to do with the unwanted effects of the combination therapy on serum potassium, which may have offset the beneficial effects of the combination therapy.

Figure 4

Changes in estimated glomerular filtration rate (eGFR) from baseline in patients at high cardiovascular risk treated with telmisartan, ramipril, or telmisartan plus ramipril. Reproduced with permission from Mann et al.41

During screening for ONTARGET, patients received ramipril and telmisartan combination to assess their tolerability.43 If an intolerability was found, at this stage, the patient was excluded. Similarly, during the trial, patients that experienced side effects on the combination therapy (like a rise in serum creatinine or hyperkalaemia) were taken off their combination medication. These two measures may have reduced the number of responders in the combination group. One could speculate that these patients would probably be the ones who would have benefited most from treatment, especially combination treatment.


There are several ways of evaluating renal function. Most useful are GFR (usually the estimated value derived from serum creatinine levels) and albuminuria. Using these parameters, one may obtain a good impression of the risk of a patient for renal and CV disease. Interestingly, albuminuria can also be useful as a target for treatment using drugs that target the RAS such as ACE-inhibitors or ARBs. These drugs have renoprotective effects on patients with chronic renal disease, including those with diabetes. Evidence is accruing that the renoprotection provided by ARBs (and ACE-inhibitors) cannot be attributed exclusively to a reduction in blood pressure; there are also blood pressure-independent effects that are due to blocking the varied pathophysiological effects of angiotensin II on kidney structure and function and which can be monitored by measuring the effect of the drug on albuminuria/proteinuria. Although RAS intervention has led to clear reduction in renal (and CV) risk, residual risk (and albuminuria) is quite high. Combination therapy of ACE-inhibitor and ARB further lowers albuminuria in renal patients and is thought to be the way forward to reduce the high renal risk in many patients. However, the recent results of the ONTARGET trial have clearly shed some dark clouds over the combination therapy optimism. New studies in renal patients are however needed, before we can abandon the potential positive effect of ACE-inhibitors plus ARB therapy in renoprotection. The ONTARGET trial was not powered or designed to look at those patients, and new trials are desperately needed and underway [Long-term Impact of RAS Inhibition on Cardiorenal Outcomes (LIRICO), Combination angiotensin receptor blocker and ACE-inhibitor for treatment of diabetic nephropathy (VA NEPHRON-D)].49,50

Conflict of interest: D.D. and H.J.L.H. have no conflict of interest to declare. D.Z. has received consultancy fees from Boehringer Ingelheim, Novarits, Abbot and AstraZeneca.


View Abstract