The HYVET sub-study on fractures is the largest placebo-controlled trial of thiazide type diuretic treatment in the prevention of fractures to have been started to date. The risks and benefits of anti-hypertensive treatment in the very elderly (over age 80) hypertensive has yet to be determined. The results of the pilot trial [42
], namely that for every stroke prevented there was one extra non-stroke death agree with the results of the INDANA meta-analysis [43
]. Thus, if the main trial should show both risks and benefits of treatment, other outcomes may have to be considered. These include any reduction in fracture rate, and any improvement (or deterioration) in cognitive function and quality of life. The very elderly may (or may not) have a greater interest in the prevention of fractures, strokes, and cognitive impairment than in the prolongation of life. The HYVET trial and its sub-studies therefore aim to give a rounded picture of what are the benefits and risks of anti hypertensive treatment in the very elderly, specifically those with mild to moderate hypertension who do not have heart failure or certain other complications of hypertension. The fracture sub-study is an integral part of the complete picture although not the primary end-point of the HYVET trial. The strengths of the sub-study are the inclusion of a high risk (very elderly) group and the attempts to gain full ascertainment and validation of fracture events. A possible, and as yet unconfirmed, weakness is that subjects may be recruited who are of above average fitness for their age group and thus lacking the expected fracture incidence. If this proves true then a trial needs to be performed in patients at even higher risk, for example the very elderly with previous fractures, who do not have a 'low' blood pressure. The HYVET study will also fail to detect many vertebral fractures. The Clinical Research Forms do not ask specifically about back pain and there is no routine radiology. We do, however, base our power calculations only on clinically detected peripheral fractures.
There are reasons to expect a benefit in preventing fractures as diuretics reduce the amount of calcium in the urine. In spontaneously hypertensive rats given large doses (1.5 mg/kg/day) of indapamide SR and a high salt intake, urinary calcium was decreased, sodium excretion increased and sodium-induced bone loss was prevented [44
]. The studies in man suggest that thiazide diuretics and indapamide SR reduce calcium excretion in the urine by about 40–50% in patients with hypercalcuria and, in a very small study of patients with essential hypertension, reduce calcium excretion by about 20% [16
It is relevant and interesting that renal stone production has been reduced by the use of diuretics in some trials but not in others. Four positive trials have been reported that employed hydrochlorothiazide [9
] bendroflumethiazide [10
] chlorthalidone [11
] and indapamide 2.5 mg/day [15
]. In the latter study stone formation was reduced by 0.2 stones/per patient/per year and the relative risk for remaining stone free with indapamide was 1.5. The HYVET trial, however, is not powered to detect a reduction in the rare event of passing renal stones.
Are there already data to prove that diuretics prevent fractures? To our knowledge there have not been any randomised controlled trials of the effects of diuretic treatment on the rate of fractures. Nevertheless there have been cross-sectional and longitudinal studies that have been summarised in a meta-analysis [26
]. These studies suggested a 20% reduction in hip fracture. On the other hand one case control study reported a 60% increased risk of hip fracture [45
]. A meta-analysis of longitudinal studies obviously carries more weight than one case control study.
Recently a pharmacoepidemiologic case-control study of the Danish population [22
] reported on 64,699 patients over age 40 with fractures. 194,111 controls used less corticosteroids, thyroxine, anxiolytics, antidepressants and diuretics than the cases. Thiazide diuretics were taken by 22% of controls and 23.7% of cases. However the thiazide users were older, more often women, took more of the other drug groups and had had more previous fractures than non-users. Thus for current users of thiazide diuretics, a crude odds ratio of 1.03 for getting any fracture was reduced to 0.90 (95%CI 0.88–0.93) by full adjustment [22
]. When only those subjects who had redeemed ≥2000 defined daily dosages of thiazides over a 5 year period were considered (in order to limit the study to those adherent to regular treatment), this consumption was associated with a 26% reduction in any fracture, a 19% reduction in hip fracture, a 5% reduction in fracture of the spine and a 31% reduction in forearm fractures. Moreover a second study has suggested that forearm fractures may be reduced by 37% in those using thiazide diuretics for more than 8 years [24
]. The results of a further large pharmacoepidemiological study suggested a 20% reduction in all fractures with the use of thiazide diuretic but also a 23% reduction with the use of beta blockers on their own [21
]. This has been confirmed recently but a benefit has also been suggested for angiotensin converting enzyme (ACE) inhibitors and calcium-Channel blockers [46
]. This raises the possibility that blood pressure reduction or confounding factors may be responsible for the benefits.