|A predictive model of the health benefits and cost effectiveness of celiprolol and atenolol in primary prevention of cardiovascular disease in hypertensive patients
|Milne R J, Hoorn S V, Jackson R T
This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.
Two beta-blocker strategies (Celiprolol andAtenolol) for reducing cardiovascular risk in patients with mild-to-moderate hypertension were studied and compared.
Type of intervention
Primary prevention; secondary prevention.
Economic study type
Cost-effectiveness analysis and cost utility analysis.
Non-smoking, non-diabetic men aged 60 with systolic blood pressure (SBP) of 160 mm Hg, total serum cholesterol levels (TC) of 6.0 mmol
Hospital. The economic study was carried out in Auckland, New Zealand.
Dates to which data relate
The data for the effectiveness analysis were obtained from studies published in 1986, and 1988-1993. The data for the resource use estimation were collected implicitly in unit costs based on 1992 data and assumptions, and from weighted means of dosages from the primary studies mentioned previously. The prices used in the final analysis were from 1997.
Source of effectiveness data
The effectiveness data were derived from a synthesis of previously completed studies and an accelerated failure time (AFT) model.
An accelerated failure time (AFT) model (regression equations) was used in order to estimate 5-year absolute risks of an initial coronary or cerebrovascular event or cardiovascular death. The model used data related to age, gender, smoking status, SBP, TC and HDL-C. The change in absolute risk was evaluated with the model by employing the estimated changes (from initiation to end of treatment) in SBP, and the ratio TC:HDC-L. Once calculated, this change was finally employed in estimating costs and benefits.
Outcomes assessed in the review
The outcomes assessed were mean changes (from initiation to termination of treatment in primary studies) in SBP and in the ratio TC:HDL-C, and the utilities attributed to health states.
Study designs and other criteria for inclusion in the review
Randomised controlled trials were included. The studies were selected if they included either atenolol or celiprolol.
Sources searched to identify primary studies
Not reported (the reader was referred to a previously published meta-analysis).
Criteria used to ensure the validity of primary studies
Methods used to judge relevance and validity, and for extracting data
Not reported. The data were extracted by means of summary statistics (outcomes) and original data (dosages).
Number of primary studies included
Fifteen (15) randomised controlled trials from 23 studies on antihypertensive drugs, used in an earlier meta-analysis, were included in the analysis. In addition, one "head-to-head comparative" study (reported as a randomised controlled trial) of both drugs in terms of blood pressure, but not serum lipids, was included.
Methods of combining primary studies
Meta-analysis. The authors weighted the results of each study by means of the inverse variance method, a random effects model being utilisedin the case of heterogeneity. Each treatment's dosage was pooled by means of sample-size weighting.
Investigation of differences between primary studies
The 'fixed effects model' assumption (implicit in the inverse variance method) was tested and those cases rejecting the null hypothesis of a same overall effect were analysed within the framework of the 'random effects model'. Homogeneity tests resulted in the effect differences of atenolol in TC across studies having a p value >0.05, as well as celiprolol's differences in effect upon HDL-C. All other cases had p values less than or equal to 0.05 and, thus, were analysed by means of the random effects method.
Results of the review
The estimated changes in percentage terms were reported as follows:
(a) in TC, atenolol +1.1% (+0.065 mmol/L, 95% CI: -0.04 to 0.17 ) and celiprolol -6.4% (-0.39 mmol/L, 95% CI: -0.75 to -0.04);
(b) in HDL-C terms, atenolol -5.5% (-0.07 mmol/L, 95% CI: -0.11 to -0.026) and celiprolol +4.1% (+0.05 mmol/L, 95% CI: 0.01 to 0.10);
(c) in TC:HDC-L, atenolol+7.7% (+0.36, 95% CI: 0.16 to 0.56) and celiprolol -10.2% (-0.51, 95% CI: -0.82 to -0.20);
(d)in terms of SBP, atenolol -10.2% (-16 mm Hg, 95% CI: 19.4 to -9.5) and celiprolol -9.1% (-14.4, 95% CI: 19.5 to -13.1).
The effect of atenolol on TC turned out to be the only effect with a p value higher than 0.05 (for the null hypothesis of zero treatment effect). The utilities attributed to the state of life after recovering from a non-fatal myocardial infarction, and to the state of life after a cerebrovascular event, were 0.9 and 0.8, respectively.
Measure of benefits used in the economic analysis
The benefits associated with each strategy were measured by means of life-years gained and quality adjusted life-years (QALYs) gained. These were in turn estimated using an AFT model which was evaluated by introducing the mean changes in SBP and the TC/HDL-C ratio as inputs in the model. Then, the final estimate was obtained by adding differences between the Kaplan-Meier survival curves resulting from the model and difference in 5-year cardiovascular death rates multiplied by life expectancy figures obtained from life tables.
Costs were discounted. Some quantities were reported separately from the prices. Cost items were not reported separately. The costs measured were operating costs and (partial) costs of complications. The boundary adopted was the hospital. The estimate of quantities was based upon authors' assumptions. The (broad) unit costs were estimated based on actual data from wholesale prices and on average 'direct medical' costs of events from the International Classification of Diseases (ICD) codes 410 to 414, for coronary events, and 434 and 436, for ischaemic cerebrovascular events. The costs were adjusted for inflation from 1992 to 1997 values by using the March quarter values of the health care component of the consumer price index. The costs excluded were those associated with recurrent events, the "ongoing" costs of care for patients in hospitalafter a major cerebrovascular event. No correction was made for different survival rates in each cohort, since the difference in absolute risk between treated and untreated cohorts was 2 percent (at 60 years of age) or less. In addition, it was assumed that the absolute risk reduction was effective only during the 5-year treatment period; once the therapy was finished, the absolute risk of cardiovascular event for the treated cohort became equal to that of the untreated one. Therefore, costs savings were restricted to the treatment period.
New Zealand dollars (NZ$).
The parameters analysed were life expectancy values, rate of nonfatal myocardial infarctions (as a proportion of total myocardial infarctions), rate of nonfatal strokes (as a proportion of total of cerebrovascular events), hospitalisation costs for coronary and cerebrovascular events, drug acquisition costs, drug dosage and discount rate (from 5% to 11.4%, which is used by the Treasury in New Zealand). These parameters were varied within threshold analysis, one-way simple sensitivity analysis and multi-way sensitivity analysis frameworks.
Estimated benefits used in the economic analysis
The estimated (undiscounted) average life-months gained (with respect to the untreated patients' group) for male nonsmokers of 60 years of age with SBP of 160 mm Hg,TC of6.0 mmol/L and HDL-C of 1.1 mmol/L (case 1), was 1.29 with atenolol, and 1.72 with celiprolol. The corresponding figures for male smokers60 years of age with SBP of 180 mm Hg, TC of 5.83 mmol/L and HDL-C of 1.0 mmol/L (case 2), were 2.53 and 3.53, respectively. In turn, for female smokers of 60 years of age with SBP of 180mm Hg, TC of 5.72 mmol/L and HDL-C of 1.0 mmol/L (case 3), the figures were 1.48 and 2.14, respectively.
Costs were discounted at 5%. The incremental total 5-year "direct" medical cost of treatment with celiprolol, relative to atenolol, was -NZ$33 per patient (case 1), and -NZ$ 64 per patient (case 2). Case 3 resulted in a corresponding figure of -NZ$44 per patient.
Synthesis of costs and benefits
The costs per life year gained (intervention relative to the comparator) was used to report the synthesis of costs and benefits, since "adjusting for quality of life by including morbidity avoided during the treatment period... had little effect on the analysis for individuals at 12% (base case) absolute cardiovascular risk because of the relatively low incidence of nonfatal myocardial infarction and stroke during the treatment period". The prices used were those prevailing in 1997, while the discount rate for costs and benefits was 5%. The cost per life-years gained of each intervention strategy against no-treatment was reported as follows:
Case 1,celiprolol, NZ$19,640; atenolol, NZ$26,746.
Case 2, celiprolol, NZ$8,680; atenolol, NZ$12,783.
Case 3, celiprolol, NZ$16,997; atenolol, NZ$ 25,394.
The results were sensitive to daily costs of treatment (doses and unit prices of the drug) and the discount rate. When the unit cost of celiprolol was increased by 44% or the unit cost of atenolol is decreased by 32%, both treatment strategies reach the same cost effectiveness ratio within the base case of 12% absolute risk of cardiovascular event (unit costs of celiprolol and atenolol at base case NZ$18.36 and NZ$15.45 per month, respectively). By using the 11.4% discount rate, the case 1 figures would change to NZ$39,167, for celiprolol, and NZ$53,250 for atenolol (100% increases).
The authors concluded that celiprolol and atenolol were shown to have opposite effects on serum lipids. Therefore, based on an epidemiological AFT model, it was predicted that the beneficial effects in terms of the reduction in the risk of coronary events will be increased, for celiprolol, and offset, for atenolol. Both strategies were predicted to have equivalent effects on cerebrovascular events. The authors stated that these strategies "will be cost effective by international standards when used to treat middle-aged individuals whose 5-year absolute cardiovascular risk is at least 10%", with these options being more cost-effective in patients at higher levels of absolute cardiovascular risk. Celiprolol was shown to be more cost-effective than atenolol for a wide range of patient characteristics and treatment courses and costs assumptions. The authors remarked, however, that the study results should be confirmed by "well controlled clinical trials".
CRD COMMENTARY - Selection of comparators
The reason for the choice of the comparator is clear.
Validity of estimate of measure of benefit
The estimate of measure of benefit used in the economic analysis is likely to be internally valid.
Validity of estimate of costs
Some resource quantities were reported separately from the prices. Adequate details of cost estimation were given although, from these details it appears that the cost analysis may have been over simplified, since it relied heavily on assumptions, and many items were left out of the analysis (given the "partial" societal perspective). The authors reported that home nursing care costs were omitted along with indirect costs (for middle-aged patients), cost savings associated with recurrences avoided or delayed, costs associated with extended life and adjustment to costs due to a small difference in mortality within the treatment period (although this was thought to be of an insignificant order of magnitude).
The conclusions were justified, given the uncertainties in the data. The issue of generalisability to other countries was not addressed (in particular the cost structure used was based on the particular conditions of New Zealand). Appropriate comparisons with other studies regarding the clinical study were made (although few were known to exist), with the differences in results being explained by the authors mostly in terms of the differences involved in the scope of the clinical endpoints selected.
Milne R J, Hoorn S V, Jackson R T. A predictive model of the health benefits and cost effectiveness of celiprolol and atenolol in primary prevention of cardiovascular disease in hypertensive patients. PharmacoEconomics 1997; 12(3): 384-408
Subject indexing assigned by NLM
Adrenergic beta-Antagonists /economics /therapeutic use; Adult; Aged; Atenolol /economics /therapeutic use; Cardiovascular Diseases /economics /prevention & Celiprolol /economics /therapeutic use; Female; Humans; Hypertension /complications /drug therapy /economics; Male; Middle Aged; Models, Economic; control
Date bibliographic record published
Date abstract record published