¹Department of Cardiology, St Luke's-Roosevelt Hospital Center, Columbia University College of Physicians and Surgeons, New York, NY, USA

Correspondence: Dr FH Messerli, St Luke's-Roosevelt Hospital Center, Columbia University College of Physicians and Surgeons, 1000 Tenth Avenue, New York, NY 10019, USA. E-mail: FMesserli@aol.com

For almost half a century, the standard approach to the treatment of hypertensive emergencies has been based on the evidence from a case series by Gifford. In this study, the blood pressure was promptly reduced with either reserpine, the drug of first choice at the time, or sodium nitroprusside.¹ In 1991, the Lancet stated, 'Clearly the immediate goal of treatment should be a reduction in pressure that begins quickly and continues gradually.'² At that time, some of the preferred drugs to be used for achieving blood pressure control were sublingual captopril or nifedipine. Interestingly, the Lancet editorial also stated emphatically, 'In case of doubt, nifedipine will be the safer option.'² We, since then, have learned that this dictum was not necessarily true.

Although nifedipine is a powerful vasodilator and lowers blood pressure rapidly,^{3, 4, 5, 6, 7, 8} such a precipitous pressure reduction can, in susceptible patients, lead to significant morbidity and mortality. Indeed, numerous instances of acute myocardial infarction, devastating strokes and even death have been reported with the sublingual use of nifedipine. Although blood pressure studies with nifedipine are legion, its safety and efficacy has never been established. Not surprisingly, sublingual nifedipine was pulled from the antihypertensive arsenal for hypertensive emergencies after its dismal risk/benefit ratio was publicized and its widespread use came to a screeching halt.⁵ More recently, common practice has been to rapidly lower blood pressure by no more than 25% within the first 2 h and to achieve a blood pressure level of about 160/100 mm Hg after 6 h.⁹

To date, there are minimal evidence-based data to support the notion of a drug class being more effective or having a better impact on clinical outcomes. Of more concern is the publication in the current issue of the Journal of Human Hypertension, of a Cochrane review by Perez et al.,¹⁰ showing that there seems to be no evidence suggesting that antihypertensive medication in the setting of a hypertensive emergency can reduce a patient's morbidity and mortality. The authors embarked on a comprehensive review of the literature and extracted all the RCTs (randomized control trials) that met the inclusion criteria with the objective to analyse the effect of antihypertensive agents on morbidity and mortality. The largest RCT in this meta-analysis only involved 133 patients and the total patients' number in this review was 869. Because these 869 patients had to be divided into different treatment groups, only a small number of patients per treatment strategy could be evaluated. Furthermore, most of the trials did not follow the patients long enough. Not surprisingly, the authors were unable to answer the question whether there was a morbidity and mortality benefit from antihypertensive drugs in hypertensive emergencies. However, they were able to demonstrate some minor differences in the degree of blood pressure reduction among nitrates, angiotensin-converting enzyme inhibitors, diuretics, -adrenergic antagonists and calcium antagonists. The clinical significance of these small differences remains unknown. Surprisingly, some of the most commonly used drugs in hypertensive emergencies, such as -blockers and clonidine, are not mentioned in this review. This obviously indicates that there were no trials with these drugs that fulfilled the inclusion criteria and the evidence for the use of these drugs in hypertensive emergencies is even more meagre than for the drugs that the authors report.

Do we, therefore, have to conclude from this meta-analysis that antihypertensive therapy in hypertensive emergencies is merely blood pressure cosmetics and does not confer any benefit to the patient? We should also note that recent guidelines give less prominence to the management of malignant hypertension.^{11, 12} In this context, we perhaps should remember the old dictum that absence of evidence does not necessarily equal evidence of absence. Indeed, there are ample clinical data documenting that lowering blood pressure in hypertensive emergencies is beneficial: papillary oedema melt away, cotton wool spots regress, pulmonary oedema resolves, renal function improves, hypertensive encephalopathy vanishes and so on. However, there is equally ample evidence that indeed, in the context of hypertensive emergencies, abrupt lowering of blood pressure can be harmful. The fact that in the meta-analysis of Perez et al.¹⁰ no benefits of this treatment were documented was not surprising, as the individual studies were not designed to look for such benefits. This means very simply that in this day and age, the treatment of hypertensive emergencies remains as empiric as it was 50 years ago at the time of Gifford. Perhaps, we have become a bit more sophisticated in using antihypertensive drugs although such sophistication may reflect familiarity with pharmacokinetics and pharmacodynamics of these drugs rather than outcome evidence. Clearly, this is a deplorable state of the art in the year 2008. A randomized prospective study is urgently needed which will be looking at outcome in hypertensive emergencies by thoroughly comparing the risk/benefit ratio of lowering blood pressure among various drugs currently used for this reduction.