|Home | About | Journals | Submit | Contact Us | Français|
This analysis shows the economic benefit of antihypertensive treatment in patients 80 years of age or older from the perspective of the Swiss healthcare system. The cost-effectiveness analysis of antihypertensive treatment in the elderly was carried out applying the results of the Hypertension in the Very Elderly Trial study to the Swiss healthcare system. The analysis shows that hypertension treatment provides, compared with placebo, an additional life expectancy of 0.0457 years per patient, over a follow-up period of 2 years. The medication cost was covered by the reduction of costs related to the treatment of strokes, myocardial infarctions and heart failure: the total cost per patient in the active group resulted in a dominant strategy of savings compared with the placebo group. Sensitivity analysis yielded a stable estimate after varying the costs of medication, stroke, myocardial infarction, heart failure and life expectancy, confirming the robustness of these results. Moreover, considering that antihypertensive treatment also positively affects the incidence of dementia, those net benefits might even be underestimated.
Hypertension is a highly prevalent risk factor for ischaemic heart diseases and stroke worldwide. The increasing longevity and prevalence of contributing factors (for example obesity) strongly affect the morbidity and mortality of cardiovascular events in both developed and developing countries.1 Therefore, the prevention and management of high blood pressure is a major public health challenge.
In 2004 ischaemic heart diseases caused 40657 deaths in France, 152659 in Germany, 38840 in Spain and 106142 in the United Kingdom. Stroke-related deaths were 33487 in France, 68498 in Germany, 34250 in Spain and 60551 in the United Kingdom. In Switzerland, the deaths related to ischaemic heart diseases were about 9200. Half of the deceased men were older than 80 years and half of the women were older than 87 years. The deaths related to stroke were 4083; more than half of them were 80 years of age or older.2, 3
Several studies of hypertension have observed the relative risk reduction in total mortality in older and younger adults, indicating an overall significant benefit.4, 5 However, Turnbull et al.6 evidenced the relative paucity of data for patients over 80 years of age, even if this group report more than half of the hypertension-related deaths.
Becket et al.7 investigated in the ‘Hypertension in the Very Elderly Trial' (HYVET) the clinical benefit of treating patients with hypertension. In a double-blind, randomized, placebo-controlled, multicenter clinical trial involving 3845 patients 80 years old or older, they evaluated the effect of a stepped care therapy, beginning with 1.5mg indapamide SR (sustained release) and adding 2–4mg perindopril, if needed to reach the target blood pressure (BP150/80mmHg).
The results of the study provided evidence that antihypertensive treatment with indapamide SR, with or without perindopril, in very elderly people is beneficial: active treatment was associated with a 30% reduction in the rate of fatal or nonfatal stroke (95% confidence interval (CI), −1 to 51; P=0.06) and a 39% reduction in the rate of death from stroke (95% CI, 1 to 62; P=0.05). The rates of death from any cause or from cardiovascular causes were also reduced by 21% (95% CI, 4 to 35; P=0.02) and 23% (95% CI, −1 to 40; P=0.06), respectively. Finally Becket et al. determined a 64% reduction in the rate of heart failure in the active-treatment group. Fewer serious adverse events were reported in the active-treatment group (358 vs 448 in the placebo group; P=0.001). Therefore, it is not too late to start antihypertensive treatment in very elderly individuals from a clinical point of view.
In the following study, we estimated the cost-effectiveness of antihypertensive therapy in patients 80 years of age or older from the perspective of the Swiss healthcare system.
We carried out a retrospective cost-effectiveness analysis of the HYVET study published by Becket et al. in May 2008.7 This double-blind, randomized, placebo-controlled clinical trial was conducted in 13 countries in Western and Eastern Europe, China, Australasia and North Africa, and was used as the basis of efficiency assessment (Table 1).
It was assumed that the effect of antihypertensive treatment on clinical outcomes could be transferred to Switzerland. A summary of the effects on clinical outcomes for the particularly relevant clinical events is shown in Table 2. The detailed HYVET protocol has been published extensively.8
The end point of this cost-effectiveness analysis is the incremental cost per life-year gained in Switzerland, expressed in Swiss francs (CHF).
To determine the total costs of antihypertensive treatment vs placebo, we included four directly attributable cost groups: (i) the medication costs according to the dosages used in the HYVET study, (ii) the acute costs and 2-year follow-up costs of a myocardial infarction, (iii) the acute costs and 2-year follow-up costs of a stroke and (iv) the costs of heart failure over 2 years. Indirect costs (for example costs related to loss of work) and intangible costs (for example suffering, pain) were not included. We chose the perspective of third-party payers in Switzerland.
The costs of treatment with 1.5mg indapamide SR (Fludex SR tablet, Servier (Suisse) SA, CHF 0.67 per tablet) and 2–4mg perindopril (Coversum tablet 4mg, Servier (Suisse) SA, CHF 1.00 per tablet) were based on pharmacy retail prices in Switzerland, as mainly pharmacies are authorized to sell drugs to ambulatory patients.9 The mean daily treatment costs were calculated on the basis of the reported treatments in the HYVET study: 25.8% of the patient received 1.5mg indapamide SR alone, 23.9% received 1.5mg indapamide SR and 2mg perindopril, and 49.5% received 1.5mg indapamide SR and 4mg perindopril. The total drug costs were CHF 420544 per 1000 patients per year, based on a daily treatment cost of CHF 1.15. The total medication costs were calculated assuming a patient compliance rate of 100% throughout the treatment period. Medication costs were calculated for 100% of the patients, assuming that the acquisition costs also apply to those patients who no longer take the medication, and as <12% of patients died during the observation period. A patient co-payment of 10% was deducted from the pharmacy price.
The costs of an acute myocardial infarction, a stroke and a heart failure (medication costs, interventions, hospitalization, outpatient treatment, rehabilitation) have previously been calculated and published by international expert groups.10, 11 All costs were adjusted to 2007 values, taking into account inflation in Switzerland. A myocardial infarction with 2-year follow-up costs CHF 37934, a stroke with 2-year follow-up accounted for CHF 117729 and 2 years heart failure costs CHF 23173.
To determine the treatment costs that could be avoided by hypertension therapy, the avoidable events per 1000 patients over 2 years were taken from the HYVET study and multiplied by the treatment costs of each individual treatment (Table 3). The costs were discounted at 5% per year, which is the standard rate applied to health care investments in Switzerland.12
In this analysis, the effectiveness criterion is represented by life-years gained in the active-treatment group compared with the placebo. Estimation of the average life expectancy of very elderly patient with hypertension was based on the declining exponential approximation of life expectancy (DEALE) method.13, 14 The all-cause mortality of the Swiss population was adjusted in terms of age and gender distribution to the HYVET population. In 2006, the average life expectancy of a 65-year-old Swiss person was 83.3 years for men and 86.8 years for women.15 The remaining life expectancy of persons with a mean age of 83.6 years is therefore 1.93 years. In the HYVET study, the over mortality in the placebo group was 0.0233 years per patient. This figure was used as the disease-specific excess mortality. The combination of the excess mortality rate and the life expectancy according to the DEALE method, yielded an adjusted life expectancy results of 1.84 years.
Costs per life-year saved (YOLS) were deemed to be calculated by dividing the discounted cost difference between active-treatment and placebo group by the number of life-years gained in the active-treatment group.
To test accuracy and sensitivity of the results (that is, the costs per YOLS), a sensitivity analysis was carried out. The YOLS, the hypertension medication costs, and the treatment costs for stroke, myocardial infarction and heart failure were varied by ±20%. Threshold analyses for the most influencing variables were carried out.
Table 3 shows the medication and treatment costs per 1000 patients over 2 years in the active-treatment and placebo group. Each patient in the active-treatment group had an additional medication cost of CHF 841 over 2 years (1.15 CHF daily), but also showed a savings potential for the analysed end points: total costs of stroke, myocardial infarction and heart failure were CHF 1666 per patient over 2 years. Total costs of treatment in a placebo patient were CHF 2544. Thus, the total cost difference in favour of treatment group is nominal CHF 37 per patient per 2 year (discounted: CHF 35).
The effectiveness calculation yielded an additional life expectancy of 0.0457 years in the active-treatment group compared with the placebo group over a period of 2 years (Table 4).
This analysis resulted in a net cost savings, and hence a ratio of costs per YOLS is not required.
The sensitivity analysis (Figure 1) shows that the factors that mostly contribute to the costs variability are medication, stroke and heart failure costs. In case of a 20% increase of drug costs, the treatment would result in a cost-effectiveness of CHF 2734 per YOLS. In contrast, a 20% decrease results in a net benefit of CHF 4284. To reach the threshold for stroke and heart failure, it would be necessary to reduce their cost of 6.0% and 16.9%, respectively. That means the stroke and heart failure cost over a period of 2 years should be reduced by CHF 7012 and CHF 3912, respectively. A hypothetical increase of 20% of the stroke costs results in a cost saving of CHF 3378 (a dominant strategy). On the other side, a 20% reduction results in a cost-effectiveness of CHF 1828 per YOLS.
The HYVET study was stopped prematurely, after 140 strokes for ethical reason; the active-treatment group already showed evidence of a reduction in the rate of the primary and secondary end points. Despite this anticipated stop, the economic assessment of antihypertensive treatment in patients 80 years of age or older has shown that the administration of the diuretic indapamide SR and, if necessary, of the angiotensin-converting-enzyme inhibitor, perindopril, is cost-effective. The benefits are primarily attributed to the prevention of stroke and heart failure.
The incremental life expectancy of 0.0457 years (about 16–17 days) looks relatively small. However, as emphasized by Wright et al.16, the gain in life expectancy should be compared with gains from other interventions aimed at the same target population.
Unfortunately, strict comparison with other studies is difficult, as cost-effectiveness analyses of hypertension treatment usually include younger patients, or are based on country-specific situations. For example, the cost-effectiveness analyses of losartan (an angiotensin type 1 receptor blocker) in patient with hypertension and left ventricular hypertrophy based on the results of the LIFE study,17 reported an incremental life expectancy of 0.0495 years (about 18 days) and a potential cost-savings of CHF 24 in Switzerland, and a cost-effectiveness of 864 € (1380 CHF), 4188 € (6689 CHF), 3195 € (5103 CHF) and 1337 Can $ (1347 CHF) per life-year saved in the Netherlands, Sweden, UK and Canada, respectively.18, 19, 20, 21, 22 However, the average age in the LIFE study was 66.9 years, with all patients between 55 and 80 years of age.
In another study, Tsevat et al.23 have estimated that 35 years old men with hypertension would gain 1.1–5.3 years (13–64 months) by reducing their diastolic blood pressure to 88mmHg. For 35 years old women the incremental life expectancy would be 0.9–5.7 years (11–68 months).
In a more recent analysis, Gandjour et al.24 have estimated that antihypertensive treatment of 60–69 years old individuals would result in 0.2–0.6 life-years (2–7 months) gained for men and 0.1–0.6 life-years (1–7 months) for women.
It should be mentioned that the HYVET study also included an assessment of cognitive function, as earlier observational studies have shown a positive association between hypertension and incidence of dementia.25, 26, 27, 28, 29, 30, 31, 32 In this substudy,33 Peters et al. reported a trend for a positive influence of the active treatment on the incidence of cognitive function and dementia (with a hazard ratio of 0.86, 95% CI 0.67–1.09). As the study was stopped prematurely and because the follow-up period was short, this finding was not significant. Nevertheless, Peters et al. combined their results with other placebo-controlled trial of antihypertensive treatment in a meta-analysis, in which they could observe a significant risk reduction in treated patients (hazard ratio 0.87, 95% CI 0.76–1.00, P=0.045).
In this cost-effectiveness analysis, we decided not to include the effects of antihypertensive treatment on dementia incidence, as they were not significant in the HYVET per se. However, it is necessary to mention that the monetary and psychological costs of caring for people with dementia are very high.34, 35 Thus, the net savings of antihypertensive treatment in this population might be even more favourable and might be underestimated in this analysis, respectively.
In this analysis, we also excluded the costs of adverse events, which were significantly lower in the active-treatment group of the HYVET study (P=0.001). However, it should be considered that such costs would speak even more in favour for active treatment.
A limitation of the HYVET study, and consequently of this cost-effectiveness analysis, is the absence of data related to diabetes and renal failure. These diseases strongly influence the quality of life of the affected persons,36 are highly prevalent in the elderly and are quite expensive, as reported, for example, by Mullins et al. in their comparison of hypertension-related costs from multinational clinical studies.37, 38
Another limitation of this analysis is that the HYVET study implicitly assumed a homogeneous population across the countries participating in the study. But the study was conducted in 13 countries that partially strongly differ between each other in demographics characteristics (for example age distribution, ethnic composition, life style and so on), health care system, risk factors prevalence, and quality of diagnoses, treatment and monitoring. The results of the HYVET study were transferred into a Swiss context, using country-specific demographic rates and health care costs, but it was not possible to adjust the analysis for all differences between the Swiss and the original study population. In addition, it should not be forgotten that the HYVET study population is only partly representative for a total collective of hypertensive patients, as there were well-defined exclusion criteria. An additional limitation is that the HYVET study did not include data on the quality of life of the patients and on respective rehabilitation measures.
A last limitation of this analysis is the assumption of a 100% compliance of drug treatment.
The demography of our society is changing as fast as never before. In the last few decades the mean life expectancy has strongly grown up. The actual standards of study design, which seems to preferably include patient below 80 years of age, are becoming obsolete. New and more studies are needed to analyse the cost-effectiveness of medication in very elderly people.
In conclusion, the antihypertensive treatment of very elderly persons, on the basis of the HYVET study and performed from a Swiss perspective, reduced the stroke and heart failure incidence, increasing the life expectancy by 0.0457 years per patient and resulting to be net cost-saving, that is, reflecting a dominant strategy.
The authors declare no conflict of interest.