arznei-telegramm 2003; 34: 26


Barely two months after publication of the ALLHAT* (1) study, which demonstrated an advantage of the diuretic chlorthalidone (HYGROTON) compared to newer antihypertensive agents such as ACE inhibitors (a-t 2003; 34: 1-2), a further comparison of conventional and newer antihypertensive therapy was presented. The Australian ANBP2* trial was financed with funds from both the enalapril (e.g. XANEF) manufacturer MSD and public sources, and was conducted as an open study in medical practices (2). For each patient included in the study, the participating general practitioners received 100 dollars in addition to the costs of medical service. This money apparently came from MSD (3).

Including 6,083 patients ANBP2 is significantly smaller than double blind ALLHAT, which was financed by the state. Following randomisation 3,044 men and women between 65 and 84 years of age took an ACE inhibitor (recommended: enalapril) as initial antihypertensive, and 3,039 took a diuretic (recommended: hydrochlorothiazide [e.g. ESIDRIX]). Should the blood pressure goal - below 140/80 if tolerated - not be achieved, beta blockers, calcium antagonists and alpha blockers could be added (2). According to the study protocol published in advance, on step 3 a diuretic was also recommended for the ACE inhibitor group (4). This was not mentioned in the publication of the study (2).

There was essentially no difference between the two groups in the blood pressure achieved. Whether an equal treatment was guaranteed under the open design cannot be traced adequately: apart from information about some of the antihypertensives used, there are no data on comedications and other interventions. Information is lacking on how many patients assigned to an ACE inhibitor also took a diuretic. Possibly there was considerable crossover in both groups: 25% from the diuretic group are said to have taken an ACE inhibitor and vice versa. However, this does not emerge from the study itself but from a report about the study (5). This change in treatment (crossover) also occurred in the ALLHAT study. However, in ALLHAT it was not generally recommended but only if clinically indicated and was described in detail. In all three groups crossover increased from 9% in the first year to 22% - 24% in the fifth year (1).

ANBP2 failed to show a difference between ACE inhibitors and diuretics. With respect to the two primary end points, which include all or the first cardiovascular complications, plus deaths from any cause, after four years only a trend is found in favour of the ACE inhibitor (relative risk [RR] 0.89; 95% confidence interval [CI] 0.79 to 1.00 and 1.01 respectively). In two of the numerous secondary endpoints and in the post hoc defined male subgroup, the ACE inhibitor is said to have proved better than the diuretic. However, considering the multitude of endpoints, the level of significance of these results (p = 0.03 to 0.04) is too low for a positive conclusion (multiple testing). In the abstract the result in men is emphasised without mentioning that this is a post-hoc analysis. This admission is "hidden" in the discussion. Furthermore the blinded evaluation of the endpoints by an endpoint committee does not prevent bias due to the (open) pre-selection of clinical events as an endpoint.

The Australian hypertension study ANBP2 cofinanced by enalapril (e.g. XANEF) supplier MSD does not show a significant difference in the primary endpoints between ACE inhibitors and thiazide diuretics as initial antihypertensive treatment.
Alleged advantages of ACE inhibitors are derived from some of the numerous secondary or retrospectively defined analyses without sufficient evidence.
Possible bias due to the open design limits the validity of the study in comparison to the double blind ALLHAT study, according to which an ACE inhibitor comes off worse as first choice than the diuretic chlorthalidone (HYGROTON).
Withholding of data on the crossover and use of diuretics in the ACE inhibitor group, which are essential for evaluation, and further irregularities in the publication undermine confidence in the seriousness of the study.
Assertions that ANBP2 has contradicted ALLHAT lack any basis and appear to arise from a certain interest.


(R = randomised study)



ALLHAT Officers and Coordinators: JAMA 2002; 288: 2981-97



WING, L.M.H. et al.: N. Engl. J. Med. 2003; 348: 583-92



REID, C.M. et al.: Clin. Exp. Pharmacol. Physiol. 2001; 28: 663-7



Management Committee on behalf of the High Blood Pressure Research Council of Australia: Clin. Exp. Pharmacol. Physiol. 1997; 24: 188-92



GOTTLIEB, S.: BMJ 2003; 326: 415


ALLHAT = Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial
ANBP2 = Second Australian National Blood Pressure Study

© arznei-telegramm 03/03