Cost–Effectiveness of Angiotensin-converting Enzyme Inhibitors in Nondiabetic Advanced Renal Disease

Charles Christian Adarkwah; Afschin Gandjour


Expert Rev Pharmacoeconomics Outcomes Res. 2011;11(2):215-223. 

In This Article


Overview & Model Design

The analysis was conducted from the perspective of the German statutory health insurance (SHI), which covers approximately 90% of the German population. We measured health outcomes in terms of quality-adjusted life years (QALYs). QALYs are the product of life years, a measure of health-related quality of life (preference weight or score) and enable the comparison of cost–effectiveness across diseases. Preference weights are anchored on a scale from 0 to 1, where 0 and 1 represent death and full health, respectively.[15] Economic outcomes were set in relation to clinical outcomes by dividing the incremental (i.e., additional) costs of providing ACE inhibitors compared with no therapy by the incremental QALYs gained (cost–effectiveness or cost–utility analysis).

In order to estimate the costs and cost–effectiveness over a lifetime, we adapted a Markov decision model previously developed for the prevention of diabetic nephropathy.[16] Based on the results of the RCTs by Ihle et al.[14] and Hou et al.,[1] we simulated the course of a cohort of 1000 patients at the age of 44 years with advanced renal insufficiency (serum creatinine: >3.0 mg/dl, glomerular filtration rate [GFR]: 15–26 ml/min/1.73 m2), proteinuria and hypertension (>150/85 mmHg), but without overt heart failure (New York Heart Association III or IV).[1]

A Markov model is an iterative process where patients stay in a defined health state for one cycle (in this case 1 year) and then make a transition to another cycle. Markov models are useful when a decision problem involves risk that is continuous over time, when the timing of events is important and when important events may happen more than once. Our model contains the following three health states (Figure 1), which represent the occurrence of events after model entry:

Figure 1.

Representation of the Markov decision models.
ESRD: End-stage renal disease.

  • Advanced renal disease (serum creatinine: >3.0 mg/dl, GFR: 15–26 ml/min/1.73 m2, CKD stage 4)

  • ESRD

  • Death

We did not include doubling of the serum creatinine level (an outcome in the RCT by Hou et al.[1]) as a separate Markov state, as doubling of the serum creatinine level is a surrogate for ESRD and shortly followed by ESRD. After each cycle, a specified proportion of patients move from advanced renal insufficiency to ESRD. This transition is delayed by ACE inhibitors. Simultaneously, a specified proportion of patients die. During each cycle, patients accumulate utility (measured by QALYs) and costs. A half-cycle correction was applied to both costs and outcomes to allow for transition events occurring midway through each 12-month cycle.

The simulation was carried out until the age of 101 years. Hence, the time horizon is 57 years. The age of 101 was chosen as a cut-off point as there are no German mortality data available beyond this age. Moreover, 99.8% of patients in the simulation are dead at this age.

Clinical Strategies

We considered two treatment strategies. In the ACE inhibitor treatment strategy, patients are treated with benazepril 10 mg twice a day. Patients in the control group receive no ACE inhibitor. In addition, both groups receive other antihypertensive agents (diuretics, α- or β-blockers, calcium-channel antagonists or some combination of these medications), but no other renin–angiotensin system agents.

Transition Probabilities

In order to identify placebo-controlled RCTs on the effect of ACE inhibitors on the progression of advanced renal insufficiency to ESRD, we used a review of the literature until July 2001 by Terajima et al. as a starting point.[17] This paper found one placebo-controlled RCT.[14] In order to identify additional placebo-controlled RCTs published after July 2001, we searched in the PubMed database (date: 1 February 2011) using the following search strategy: renal insufficiency, chronic AND creatinine AND antihypertensives AND end-stage renal disease AND angiotensin-converting-enzyme NOT(type-2-diabetes); limits: randomized controlled trial, all adult: ≥19 years of age. We obtained 39 hits, and of these, 37 papers were excluded because they did not compare ACE inhibitors with placebo, did not consider advanced renal insufficiency or did not include ESRD as an outcome. The only two studies left were by Ihle et al.[14] and Hou et al.[1] and were, therefore, used as the source of our effectiveness data. The two studies were double blinded (see Table 1 for baseline characteristics).[1,14] To combine effectiveness data from both studies and estimate the common odds ratio, we used the Mantel–Haenszel test.[18]

In order to determine the annual transition probability from advanced renal disease to ESRD without ACE inhibitor therapy, we first calculated a total probability for each of the two trials by dividing the number of events (ESRD) during the trial period in the control arm by the number of patients. Second, we determined the annual transition probability by assuming a constant annual hazard rate over the study time horizon.[19] And third, we determined a weighted average rate, with the number of individuals included in each study as weights. In the study by Hou et al., the number of events was obtained by directly contacting the authors and does not include death.[1]

Transition probabilities with and without an ACE inhibitor therapy are shown in Table 2. Mortality was regarded as a function of age in patients without ESRD. In particular, we used age-specific mortality rates of nondiabetic patients. To calculate the latter, age-specific mortality rates of the general population were multiplied with a standardized mortality ratio for patients with advanced renal disease compared with the general population, which was derived from a large community-based cohort study.[20] For patients with ESRD we assumed that mortality was age independent.

Preference Weights

For patients in the health state 'advanced renal disease' we considered a utility loss according to a survey using the time tradeoff (TTO) method in 65 patients.[21] This population had a mean age of 66 years and a GFR between 15 and 30 ml/min/1.73 m2. In addition, we considered an age-dependent loss of utility.[22] For patients treated with an ACE inhibitor we also considered a disutility from all side effects in a sensitivity analysis.[23] The preference weight for ESRD was taken from a systematic review of empirical studies in which TTO weights were provided by patients.[24] We tested the broad range of TTO weights in a sensitivity analysis. The TTO is the most commonly used method to elicit quality-of-life weights for QALYs.[25,26]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.