sign in sign up
<<
Home , Alacepril, Barnidipine, Ceronapril, Clentiazem, Fantofarone, Fenoldopam

Journal of Human (2008) 22, 596–607 & 2008 Macmillan Publishers Limited All rights reserved 0950-9240/08 $30.00 www.nature.com/jhh ORIGINAL ARTICLE Pharmacological interventions for hypertensive emergencies: a Cochrane

MI Perez and VM Musini Department of Anesthesiology, Pharmacology and Therapeutics, Faculty of Medicine, The University of British Columbia, Vancouver, British Columbia, Canada

Hypertensive emergencies occur when high blood tality event data were reported in seven trials. Due to pressure is associated with the presence of acute end- insufficient data, no meta-analysis was performed for organ damage, such as hypertensive encephalopathy. clinical outcomes. Differences in There is controversy as to when and which antihyper- changes between antihypertensives were minor. There tensive to use in these situations. Using a is no RCT evidence demonstrating that antihypertensive comprehensive search strategy in electronic sources, drugs reduce mortality or morbidity in patients with MEDLINE, EMBASE and Cochrane register, hypertensive emergencies. Furthermore, there is insuf- we conducted a systematic review to look all rando- ficient RCT evidence to determine which or drug mized control trials (RCTs) that compare an antihyper- class is most effective in reducing mortality and tensive drug versus placebo, no treatment or another morbidity. There were some minor differences in the . Fifteen RCTs (representing 869 degree of blood pressure lowering when one class of patients) met the inclusion criteria. Two trials included a antihypertensive drug is compared to another. However, placebo arm. All studies (except one) were open-label the clinical significance is unknown. RCTs are needed to trials. Seven drug classes were evaluated in those trials: assess different drug classes to determine initial and nitrates (nine trials), -converting enzyme longer-term mortality and morbidity outcomes. inhibitors (seven), (three), Journal of Human Hypertension (2008) 22, 596–607; blockers (six), a-1 adrenergic antagonists (four), direct doi:10.1038/jhh.2008.25; published online 17 April 2008 vasodilators (two) and agonists (one). Mor-

Keywords: hypertensive emergencies; randomized controlled trials; systematic review; meta-analysis

Introduction recommended by Gifford1 based on a series of eight cases with hypertensive encephalopathy that were A is the clinical setting treated with nitroprusside. On the basis of where a marked elevation of blood pressure is this case series evidence this approach has become associated with acute end-organ damage, for exam- and remained the standard of care and is currently ple, hypertensive encephalopathy or aortic dissec- recommended by most if not all guideline commit- tion. As such it is a life-threatening condition. The tees (such as JNC-72). At issue in this review is goal of treatment is to reverse the end-organ damage, whether randomized control trial (RCT) evidence prevent adverse outcomes and prolong life. This supports this approach and which drug classes are review focuses on blood pressure lowering drugs the most effective. that are used in this emergency setting. Two published systematic reviews have addressed The management of hypertension in these emer- these issues. One compares different antihyperten- gency situations represents a significant therapeutic sive drugs, but it pools hypertensive emergency and challenge. Many antihypertensive drug classes have urgency trials.3 Urgencies are defined as marked been used with the objective of rapidly reducing elevated blood pressure in an otherwise stable blood pressure, and the expectation of reducing patient (that is, without acute end-organ damage). adverse clinical outcomes. This approach was first In our opinion the urgency setting is very different from that of emergencies and needs to be reviewed Correspondence: Dr MI Perez, Department of Anesthesiology, separately. Pharmacology and Therapeutics, Faculty of Medicine, The The second systematic review, a Cochrane review University of British Columbia, 2176 Health Science Mall, of interventions that alter blood pressure after acute Vancouver, British Columbia, Canada V6T 1Z3. E-mail: [email protected] stroke is not limited to RCTs studying drugs to Received 7 December 2007; revised 26 January 2008; accepted 23 reduce blood pressure and includes RCTs whether February 2008; published online 17 April 2008 or not the patients had elevated blood pressure.4 Different interventions for hypertensive emergencies MI Perez and VM Musini 597 Therefore, it also does not answer the question threshold because of a higher incidence of rebleed- raised here. ing above this level.8 For patients with any other acute end-organ damage setting an SBPX180 and or DBPX110 mm Hg is the defined threshold. Objectives We included all RCTs that included patients with General. The general objective is to find and these minimum or higher thresholds. In the case that quantify the RCT evidence for antihypertensive drug an RCT does not define blood pressure inclusion treatment of patients with a hypertensive emer- criteria but had included only one category of gency, defined as marked hypertension associated patients (patients with pulmonary oedema, for with acute end-organ damage. example), then the mean baseline blood pressure had to be equal to or greater than these defined Specific. The specific objective is to answer the thresholds. In the event that an RCT had included following two questions: patients with different end-organ damage clinical Does antihypertensive drug therapy as compared settings, a mean baseline blood pressure of to placebo or no treatment affect mortality and SBPX180 and or DBPX110 mm Hg is acceptable morbidity in patients with a hypertensive emer- for inclusion. gency? Note that -related hypertensive emer- Does one first-line antihypertensive gencies are excluded from this review. offer a therapeutic advantage, in terms of mortality and morbidity, over another in patients with a hypertensive emergency? Types of interventions Intervention: A first-line antihypertensive drug class (first-line antihypertensive drug classes include: Criteria for considering studies for this review nitrates, b-blockers, angiotensin-converting enzyme Types of studies. All unconfounded, truly RCTs (ACE) inhibitors, diuretics, calcium channel block- that compare a first-line antihypertensive drug class ers (CCBs), dopamine agonists, a-adrenergic antago- versus placebo, no treatment or another first-line nists and direct vasodilators ( and antihypertensive drug class. Crossover trials were hydralazine). excluded. There was no language restriction. Control: placebo, no treatment or a different first- line antihypertensive drug class. Types of participants. Participants must meet the following hypertensive emergency definition: any clinical setting where patients present with marked Types of outcome measures elevation of blood pressure in the presence of acute Primary outcome measures include: end-organ damage. Examples of acute end-organ damage are the following: ,  total serious adverse events, unstable angina, acute left ventricular failure with  all-cause mortality and pulmonary oedema, acute aortic dissection, ence-  composite of non-fatal cardiovascular events, phalopathy, stroke and life-threatening bleeding including myocardial infarction, unstable angina, (intracerebral haemorrhage and subarachnoid hae- dissection of aortic aneurysm, acute renal failure, morrhage). stroke and respiratory failure (necessitating me- Thus, patients with marked elevation of blood chanical ventilation). pressure but without acute end-organ damage (defined as urgencies) were not included. Secondary outcome measures include: There is no evidence as to what constitutes ‘marked blood pressure elevation’. Therefore, we  weighted mean change in SBP, DBP and in heart have chosen blood pressure level(s) commonly used rate (HR), during the treatment period and in clinical practice to mandate the use of antihy-  withdrawals due to adverse effects. pertensive drugs (along with other acute therapy such as pain management) in relevant clinical Methods of the review settings. For example, for patients with acute myocardial infarction a systolic blood pressure Data abstraction (SBP) greater or equal to 180 and or diastolic blood The search strategy for identification of studies is pressure (DBP)X110 mm Hg is the threshold above summarized in Table 1. Two reviewers (MIP and which thrombolysis is contraindicated (ACC/ MVM) independently decided whether a trial was AHA-5). For patients with acute aortic dissection included. They also independently extracted and or with left ventricular failure and pulmonary entered the data from the included studies. Dis- oedema an SBP greater or equal to 120 mm Hg and crepancies were resolved by discussion. Absence of or DBPX70 mm Hg is the threshold for therapy.6,7 consensus was resolved by a third reviewer (JMW). For patients with intracranial haemorrhage or sub- A modified Cochrane quality scoring system was arachnoid haemorrhage an SBPX160 mm Hg is the used for concealment of allocation and blinding: A

Journal of Human Hypertension Different interventions for hypertensive emergencies MI Perez and VM Musini 598 Table 1 Cochrane search strategy: see also Collaborative review Table 1 Continued group search strategy 57 .mp. We searched randomized controlled trials of all antihypertensive 58 .mp. drugs used for hypertensive emergencies through the following 59 .mp. database of articles published from 1966 to August 2007, 60 .mp. MEDLINE, EMBASE and Cochrane clinical trial register. A 61 .mp. comprehensive search strategy was used to identify all relevant 62 Angiotensin converting enzyme inhibitor.mp. or angiotensin- articles. Review articles and trials reference lists were also converting enzyme inhibitors/ checked. Keywords: controlled clinical trial, randomized 63 Acebutolol.mp. controlled trials, meta-analysis, severe/accelerated/crisis(es), 64 .mp. hypertension, antihypertensive, emergencies: hypertensive 65 Bisoprolol.mp. encephalopathy, myocardial infarction, unstable angina, acute left 66 Esmolol.mp. ventricular failure, pulmonary oedema, stroke, subarachnoid/ 67 Labetalol.mp. intracranial haemorrhage, aortic dissection and list of drugs as 68 Metoprolol.mp. follows. 69 Nadolol.mp. 70 Practolol.mp. 1 Randomized controlled trial.pt. 71 .mp. 2 Randomized controlled trials.mp. 72 .mp. 3 Randomized controlled trial.mp. 73 .mp. 4 Controlled clinical trial.pt. 74 Carvedilol.mp. 5 Controlled clinical trials.mp. 75 Adrenergic beta-Antagonists.mp. 6 Controlled clinical trial.mp. 76 .mp. 7 Random allocation.mp. 77 .mp. 8 Exp double-blind method/ 78 .mp. 9 Double-blind.mp. 79 .mp. 10 Exp single-blind method/ 80 .mp. 11 Single-blind.mp. 81 .mp. 12 or/1–11 82 .mp. 13 Exp animal/ 83 .mp. 14 12 not 13 84 .mp. 15 Clinical trial.pt. 85 .mp. 16 Clinical trials.mp. 86 .mp. 17 Clinical trial.mp. 87 .mp. 18 Exp clinical trials/ 88 .mp. 19 (Clin$ adj25 trial$).mp. 89 Elgodipine.mp. 20 ((singl$ or doubl$ or trebl$ or tripl$)adj25 (blind$ or 90 Etafenone.mp. mask$)).mp. 91 .mp. 21 Random$.mp. 92 .mp. 22 Exp research design/ 93 .mp. 23 Research design.mp. 94 .mp. 24 or/15–23 95 .mp. 25 24 not 13 96 .mp. 26 25 not 14 97 .mp. 27 Comparative studies.mp. 98 .mp. 28 Comparative study.mp. 99 .mp. 29 Exp evaluation studies/ 100 .mp. 30 Evaluation studies.mp. 101 .mp. 31 Evaluation study.mp. 102 .mp. 32 Follow up studies.mp. 103 .mp. 33 Follow up study.mp. 104 .mp. 34 Prospective studies.mp. 105 .mp. 35 Prospective study.mp. 106 .mp. 36 (control$ or prospective$ or volunteer$).mp. 107 .mp. 37 or/27–36 108 .mp. 38 37 not 13 109 .mp. 39 38 not (14 or 26) 110 .mp. 40 14 or 26 or 39 111 .mp. 41 .mp. 112 Semotiadil.mp. 42 .mp. 113 .mp. 43 .mp. 114 Tiapamil.mp. 44 .mp. 115 .mp. 45 .mp. 116 Calcium .mp. or calcium channel blockers/ 46 Derapril.mp. 117 Nitroprusside.mp. 47 .mp. 118 Nitroglycerine.mp. 48 .mp. 119 Nitroglycerin/ or nitroglycerine.mp. or isosorbide dinitrate/ 49 .mp. 120 Nitrates.mp. or nitrates/ 50 Idapril.mp. 121 .mp. 51 .mp. 122 Trimethaphan/ or trimethaphan camsylate.mp. 52 .mp. 123 .mp. 53 .mp. 124 .mp. 54 Moveltopril.mp. 125 .mp. 55 .mp. 126 Labetalol.mp. 56 .mp. 127 Ketanserine.mp.

Journal of Human Hypertension Different interventions for hypertensive emergencies MI Perez and VM Musini 599 Table 1 Continued combined using a weighted mean difference 128 Hydralazine.mp. (WMD) method, whereby the trials are weighted 129 .mp. according to the number of subjects in the trial and 130 Fenoldopam.mp. the within-study variance. Some of the trials did not 131 Diazoxide.mp. report a within-study variance for blood pressure 132 .mp. reduction. In these studies standard deviation (s.d.) 133 $.mp. 134 .mp. was imputed using the following hierarchy: 135 Chlorthalidone.mp. or chlorthalidone/ 136 .mp. or furosemide/ (1) Pooled s.d. calculated either from the t-statistic 137 or/41–136 corresponding to an exact p-value reported or 138 40 and 137 from the 95% CI of the mean difference between 139 Myocardial infarction.mp. treatment group and comparative group. 140 Unstable angina.mp. (2) Standard deviation of blood pressure/HR at the 141 Acute left ventricular failure.mp. 142 Pulmonary / or pulmonary oedema.mp. end of treatment. 143 Stroke.mp. (3) Standard deviation of blood pressure/HR at 144 Life-threatening bleeding.mp. baseline (except if this measure is used for entry 145 Aneurysm, dissecting/ or aortic dissection.mp. criteria) 146 Intracranial hemorrhages/ or cerebral hemorrhage/ or intracranial haemorrhage.mp. (4) Weighted mean s.d. of change in blood pressure/ 147 Intracranial aneurysm/ or subarachnoid hemorrhage/ HR calculated from at least or subarachnoid haemorrhage.mp. (5) Other trials using the same drug and dose 148 or/139–147 regimen 149 Hypertension.ti,ab. (6) Weighted mean s.d. of change in blood pressure/ 150 High blood pressure.ti,ab. 151 Blood pressure.ti,ab. HR calculated from other trials using the same 152 or/149–151 drug 153 Pulmonary artery hypertension.mp. (7) Weighted mean s.d. of change in blood pressure/ 154 .mp. HR calculated from all other trials (any drug and 155 Portal hypertension.mp. 156 or/153–155 dose) 157 152 not 156 158 148 and 157 Several sensitivity analyses were pre-planned to 159 Hypertensive emergencies.ti,ab. test robustness, including the use of both fixed and 160 Hypertensive emergency.ti,ab. random effects models, 95 and 99% CIs, and quality 161 Hypertensive urgency.ab,ti. 162 Hypertensive urgencies.ti,ab. of trials. Also sensitivity analyses were pre-planned 163 Hypertensive crisis.ti,ab. according to the clinical setting and to the class of 164 Hypertensive crises.ti,ab. drug. 165 Acute end-organ damage.mp. 166 or/158–165 167 138 and 166 Results Description of studies Our search strategy yielded 86% of irrelevant (adequate and double blind), B (unclear and single studies in the first screening stage by reading titles blind or open label) and C (clearly inadequate and and abstracts (Figure 1). Fifteen RCTs (869 patients) open label). The two reviewers (MIP and MVM) were found that satisfied the inclusion criteria9–23 also independently assessed the quality of (Table 2). Two trials were placebo controlled.12,18 studies. Authors were contacted in case of missing Only one trial12 was confirmed to be double blind, information. whereas the rest were open label. No trial was designed for or had the power to detect differences in clinical outcomes .The largest trial consisted of Analyses 133 patients.20 The longest trial11 lasted 10 days. For the synthesis and analysis of the data Cochrane Most of the trials reported data for only 2–6 h. Seven review manager software Revman 4.2.9 was used. drug classes were evaluated: nitrates (9 trials), ACE Relative and absolute risk differences (with 95% inhibitors (7), CCBs (6), peripheral a-1 blockers (4), confidence interval (CI)) were calculated for dichot- diuretics (3), direct vasodilators (2) and dopamine omous outcomes for each trial on an intention to treat agonists (1). basis. Heterogeneity between trial results was tested All included trials had patients with elevated using w2-test, where po0.05 was taken to indicate blood pressure in the presence of acute significant heterogeneity. The fixed effect model end-organ damage. Blood pressure entry criteria was used when there was homogeneity and the differed among trials. Four trials were included random effect model was used to test for statistical on the basis of their mean blood pressure significance where there was heterogeneity. values at baseline.10,12,17,18 Seven trials included Trials were not subclassified according to dose or exclusively patients with acute pulmonary dosing regimen. Data for blood pressure was oedema.10,12,14,17,20,21,23 One trial included exclusively

Journal of Human Hypertension Different interventions for hypertensive emergencies MI Perez and VM Musini 600 Citations identified in literature search: N=5,413 Citations excluded by reading title and abstract. (clearly no relation to our work) n=4,660

Citations retrieved for more detailed evaluation: n=753 Excluded: n=711 Reasons : Non-randomized: 354 Non-acute end organ damage (AEOD) setting: n=270 (AEOD) setting but BP below our pre-established thresholds defining a hypertensive emergency: n=87

Potentially appropriate randomized controlled trials: n=42 Excluded: n=27 Reasons : (see text for details)

Included: n=15

Figure 1 Quorum flowchart of studies.

patients with hypertensive encephalopathy.16 There  One was a cross-over trial.47 was no trial that included exclusively patients  Two trials had wrong comparators (1 compared with acute aortic dissection or acute myocardial different doses of the same combination therapy;48 infarction. Thus, the rest of seven trials included a 1 compared two drugs of the same class.49 diverse population with different acute end-organ  One RCT only included responders to a previously damage. Only two trials9,15 reported the s.d. of the given antihypertensive therapy.50 change of blood pressure. In the rest of the trials this measure of variability was imputed from the s.d. at Two out of 27 excluded trials involved a b-blocker end point. arm and 18/27 excluded trials involved a calcium Additional information was required and re- channel blocker arm. One excluded trial studied quested from all included trials. One trialist9 exclusively patients with acute aortic dissection.49 provided missing information in the original pub- lication. The rest of the trialists did not reply to our request. Methodological quality of included studies 12 We excluded 27 clinical trials for several reasons: All studies, except one, were open-label trials. The method of randomization was not reported in eight  Several trials mixed patients with and without trials. The method to achieve concealment of acute end-organ damage in the same RCT (12 allocation was reported in only two trials.12,16 trials).24–35  Other trials included patients without explicitly stating whether patients had acute end-organ Comparisons according to outcomes damage or not (7 trials).36–42 Total serious adverse events. No trial reported total  Some trials included non-randomized partici- serious adverse events. pants in the trial’s results (1 trial).43  One trial did not report any of the outcomes of All-cause mortality. Mortality was reported in seven interest (1 trial).44 trials9,10,14,16,17,20,21 and totalled six deaths in three  Two trials did not fulfil blood pressure threshold RCTs. The group to which the dead patients were criteria.45,46 originally allocated was not reported for five of the

Journal of Human Hypertension Table 2 Summary of included studies

Author Comparators and dose n Blood pressure Mean SBP/DBP Clinical inclusion Mean SBP/DBP Mortality (mm Hg) inclusion (mm Hg) at criteria (mm Hg) at end criteria baseline point

Angeli et al.9 Nifedipine 10 mg s.l. Captopril 25 mg 10 DBP4140 247/158 Acute end-organ 204/115 0 (s.l.) 10 245/145 damagea 190/116 0 Beltrame et al.10 Nitroglycerine 2.5–10 mg i.v. infusion 37 Not definedb 161/NR Acute pulmonary 133/NR 3 Furosemide 40 mg i.v. boluses 32 164/NR oedema 139/NR Elliot et al.11 Nitroprusside 0.5-mgkgÀ 1 minÀ 1 i.v. 15 DBP4120 222/137 Acute end-organ 174/105 Not reported infusion 13 214/136 damagea 180/106 Fenoldopan 0.1 mgkgÀ 1 minÀ 1 i.v. infusion Halminton et al.12 Placebo 25 Not definedb 160/100 Acute pulmonary NR Not reported Captopril 25 mg (s.l.) 23 172/112 oedema NR Hirschl et al.13 Nitroprusside 0.5–3 mg k minÀ 1 i.v. 35 SBP4200 and/or 211/109 Acute end-organ 151/74 Not reported infusion 46 DBP4110 215/107 damagea 162/88 Urapidil 12.5 mg i.v. boluses Hirschl et al.14 Nitroglycerin 0.8 mg (s.l.) 23 SBP4200 or 206/116 Acute pulmonary 136/71 0 Enalaprill 2.5 mg i.v. boluses 23 DBP4100 211/115 oedema 139/70 Marigliano et al.15 Nifedipine 10 mg (s.l.) 22 SBP4210 208/139 Acute end-organ 154/70 Not reported Captoprill 25 mg (s.l.) 22 230/120 damagea 163/90 Mcnair et al.16 Diazoxide 75–150 mg i.v. boluses 28 DBP4135 228/118 Hypertensive 179/110 2 Dihydralazine 6.25–12.5 mg i.m. boluses 24 218/142 encephalopathy 169/101 Nelson et al.17 Isosorbide 50–200 mgkgÀ 1 hÀ 1 i.v. 14 Not definedb 130/75 Acute left ventricular 118/70 0 infusion 14 119/72 failure 116/70 0 À 1 Furosemide 1 mg kg i.v. boluses Musini VM and Perez emergencies MI hypertensive for interventions Different Pastorelli et al.18 Placebo 20 Not definedb 188/111 Acute end-organ 179/101 Not reported Nifedipine 10 mg (s.l.) 16 195/114 damagea 162/92 Captopril 50 mg (s.l.) and oral 41 196/114 168/97 Ketanserine 40 mg (s.l.) 15 185/117 165/100 Rubio et al.19 Isosorbide aeroso 1.25 mg oral 30 MAP4130 187/121 Acute end-organ 153/92 Not reported Nifedipine 10 mg (s.l.) 30 190/115 damagea 153/86 Schreiber et al.20 Nitroglycerine 0.8 mg (s.l.) 73 SBP4200 DBP4100 216/116 Acute pulmonary 134/72 0 Urapidiil 12.5 mg i.v. boluses 60 218/118 oedema 134/70 0 Verma et al.21 Furosemide 1 mg kgÀ 1 i.v. bolus 12 SBP4100 117/73 Acute left ventricular 112/70 0 Isosorbide 50–200 mg i.v. infusion 12 131/75 failure 118/70 0 Hydralazine 0.15 mg kgÀ 1 i.v. bolus 12 134/77 128/71 1 Wu et al.22 Nifedipine 10 mg (s.l.) 30 SBP4190 and/or 198/124 Cerebral signs or 140/78 Not reported Captopril 25 mg (s.l.) 35 DBP4120 198/122 symptoms 134/78 10 mg (s.l.) 27 198/128 160/90 Yang et al.23 Nitroprusside 1 mgkgÀ 1minÀ 1 i.v. 20 SBP4160 and/or 195/115 Acute pulmonary 144/85 Not reported infusion 20 DBP4100 196/114 oedema 135/79 Nicardipine 3 mgkgÀ 1minÀ 1 i.v. infusion ora fHmnHypertension Human of Journal Abbreviations: DBP, diastolic blood pressure; i.m., intramuscular; i.v., intravenous; MAP, mean arterial pressure; SBP, systolic blood pressure; s.l., sublingual. aAs stated in the article reflecting the inclusion of patients with different acute end-organ damage settings. bThis RCT was included on the basis of the mean blood pressure values at baseline according to our pre-defined thresholds for this category of patients. 601 Different interventions for hypertensive emergencies MI Perez and VM Musini 602 deaths. In one RCT, a patient treated with hydralazine significant greater reduction in both, SBP (WMD died of a rupture of the inter-ventricular septum.21 In À13.14, 95% CI, À19.48,À6.80) and DBP (WMD four trials mortality was reported as nil. In eight trials À8.03, 95% CI, À12.61,À3.45) with antihyperten- there was no mention of mortality. It is possible that sive therapy. There were no data on HR. there were no deaths during the short range of follow- It was not possible to extract blood pressure data up (6–24 h), but it is impossible to be certain. from the other placebo-controlled trial.12] In addi- tion to not reporting any measurement of variability, Non-fatal cardiovascular events this trial reported blood pressure data as change in Composite. Cardiovascular events were reported mean arterial pressure (MAP). in five trials.10,12,14,16,20 No trial reported cardiovascu- lar events as a composite. It was not possible to extract Nitrates vs diuretics. Three trials compared events from the original trials and analyse them as a nitrates to diuretics (Beltrame, Nelson and Ver- composite due to a risk of double-counting the events. ma)10,17,21. Furosemide was the common used in all of them with two nitrates, nitroglycerine Myocardial infarction. One placebo-controlled and isosorbide as comparators. Neither SBP- nor trial12 reported this outcome. There was no statisti- DBP lowering effect was statistically different cally significant difference between ACEi and between the two classes of drugs. However, in placebo (RR 0.72, 95% CI 0.31–1.72). Beltrame,10 the SBP lowering effect of both drugs Three head-to-head trials reported this out- was greater (À21 mm Hg for furosemide; come.10,16,20 There was no statistical difference in À23.75 mm Hg for nitroglycerin) than that reported myocardial infarction between nitrates (2.7%) and a- in the other two trials ( þ 1.0, þ 1.6 mm Hg for adrenergic antagonist (5%) (RR 0.55, 95% CI 0.09– furosemide groups and À6, À8 mm Hg for isosorbide 3.17); or nitrates (16%) vs diuretics (12.5%) (RR groups, respectively). The reasons for that difference 1.30,95% CI 0.40–4.19); or between diazoxide (3.5%) across trials are not clear. Despite these differences, vs dihydralazine (4%), (RR 0.86, 95% CI 0.06–12.98). heterogeneity was not present when pooling all these three trials. HR change was also not signifi- Pulmonary oedema requiring mechanical ventila- cantly different for both classes of drugs. tion. Three trials reported this outcome.12,14,20 There was no meta-analysis performed since there Nitrates vs a-1 antagonist. Two trials compared was only one trial for each comparison. There was the A1A, urapidil, with nitrates.13,20 The first trial no statistically significant difference between cap- used nitroprusside and the second used nitroglycer- topril and placebo (RR 0.40, 95% CI 0.09–1.86), ine as comparator. The SBP lowering effect of the nitrates and a-adrenergic antagonist (RR 4.12, 95% two nitrates was similar (À58.4 mm Hg for nitroprus- CI 0.20–84.24 or between nitrates and ACE Inhibitor side and À59.5 mm Hg for nitroglycerine). However, (RR 0.33, 95% CI 0.01–7.78). the effect of urapidil (administrated at the same dose Other than the above, the trials did not report any in both trials) was very different (À37.6 and of our list of cardiovascular events (unstable angina, À73.5 mm Hg). A similar discrepancy was seen for dissection of aortic aneurysm, acute renal failure or DBP. This heterogeneity precluded the pooling of stroke). An additional cardiovascular event was these trials in a meta-analysis for these outcomes. reported that was not on our list: asystole, which happened in one patient randomized to an ACE Nitrates vs . For this compari- inhibitor.14 son one trial was included.11 During 4 h of treat- ment, nitrates were associated with a statistically Withdrawals due to adverse events. Only one trial significant greater reduction in SBP as compared comparing an a-blocker with nitroglycerine reported with a dopamine agonist (WMD À14.00, 95% CI withdrawal due to adverse events.20 There were no (À27.72, À0.28). There were no differences between significant differences between these two drugs these classes in DBP or HR. classes (5 vs 2.7%; RR 3.38, 95% CI 0.17–68.84). Nitrates vs ACE inhibitors. One trial compared a Weighted mean change in blood pressure and HR nitrate with an ACE inhibitor.14 No statistically during treatment. For this secondary outcome all significant difference was found between the two trials provided some data and we were able to pool groups in SBP or DBP or HR. this data. Nitrates vs calcium channel blockers. By pool- Drug vs placebo or no treatment. Although we ing two trials19,23 CCBs were not associated with included two placebo-controlled trials, only one statistically significant differences in SBP or DBP as provided SBP or DBP data18 and this was limited to compared to nitrates. Using the fixed effect model, 1 h of follow-up. In this trial, three classes of CCBs were associated with statistically significant antihypertensives were included: calcium channel increase in HR as compared to the nitrates (WMD blocker, ACE inhibitors and a-1 adrenergic antago- 11.76 95% CI (4.45,19.07). However there was nists (A1A). The pooled effect showed a statistically significant heterogeneity across trials and this

Journal of Human Hypertension Different interventions for hypertensive emergencies MI Perez and VM Musini 603 increase was no longer statistically significant when drug treatment for hypertensive emergencies. a random effect model was used. A systematic review that combined hypertensive emergencies and urgencies3 did not include 11 trials Nitrates vs direct vasodilator. For this compar- included in our systematic review. Furthermore, ison one trial was included.21 There was no Cherney’s review mixed randomized with non- statistical difference in SBP or DBP reduction randomized trials. between the two drugs. There was also no signifi- The only other relevant systematic review in cant difference between these classes in HR change. relation to hypertensive emergencies is that con- ducted for acute stroke by BASC.4 We excluded one 46 ACE inhibitors vs calcium channel blockers.Four trial (n ¼ 16 patients) that the BASC 2001 systema- trials,9,15,18,22 compared an ACE Inhibitor with a CCB. tic review had included. The reason for excluding it The pooled data show that CCBs were associ- was because the blood pressure criteria in this trial ated with a significantly greater reduction (4170/95 mm Hg) did not meet our blood pressure in DBP as compared with ACE-I (WMD 7.86, 95% threshold criteria (SBPX180 and or DBPX CI (4.92, 10.81). No statistically significant 110 mm Hg). This exclusion does not affect our difference was found between the two groups in the conclusion for clinical outcomes as this trial did reduction of SBP. In three trials that reported not report clinical outcomes. The other BASC 2001 HR changes,9,15,22 CCBs were associated with trials were not included because blood pressure at a significant increase in HR as compared with ACE baseline was not elevated. Thus, these clinical trials inhibitors (WMD 22.91, 95% CI (19.8, 26.01). did not include hypertensive emergency patients, as However there was significant heterogeneity we have defined it. across trials and this increase was no longer signifi- One of the limitations in our review is that most cant when a random effect model was used. of the included trials were small (average 58 patients per trial). Furthermore, with the 12 ACE inhibitors vs a-1 adrenergic antagonist.Two exception of Hamilton et al. all trials were of poor trials18,22 compared an ACE Inhibitor with an A1A. quality. Three included trials deserve further discussion. Both trials used captopril as comparator but one trial 12 used prazosin and the other used . The Hamilton et al., the only double-blind trial, pooled data show that ACE-I were associated with a includes patients with acute pulmonary oedema significantly greater reduction in both SBP and DBP and high blood pressure, and it compared captopril as compared with A1A (SBP WMD À20, 95% CI vs placebo. It demonstrates that this high-quality and double-blind trial was ethical and feasible. The (À22.85,À17.39); DBP WMD À3.70, 95% CI 16 (À7.08,À0.31). For SBP outcome there was statisti- Danish II 1986 trial was the only trial that included cally significant heterogeneity across trials. How- patients exclusively with hypertensive encephalo- ever the difference was still significant when the pathy. This was a well-organized multicentre trial, random effects model was used. No statistically conducted in Denmark, comparing diazoxide vs significant difference was found between the two dihydralazine. Due to its study design, the ethical groups in the HR change in the only trial reporting committee accepted that the informed consent could that outcome.22 not be obtained from patients as all of them had symptoms of hypertensive encephalopathy. A downside of this study is the fact that the trialists Diazoxide vs hydralazine. For this compari- reported their results in duplicate publications that son one trial,16 which dealt with exclusively did not cite the other publications (the original hypertensive encephalopathy patients, was in- publication, Krogsgaard et al.,51 is not cited in the cluded. During 4 h of treatment, hydralazine was other duplicate publications,16,52,53 In addition, associated with a statistically significant greater blood pressure values were not the same in the reduction in both SBP (WMD 13.56, 95% CI different publications, and none of the publications (3.06,24.06)) and DBP (WMD 14.67, 95% CI reported measures of SBP or DBP variability. The (8.01,21.33)) as compared to diazoxide (WMD largest trial, Schreiber et al.,20 included 133 patients À14.00, 95% CI (À27.72, À0.28). It is important to with acute pulmonary oedema plus high blood mention, though, that there was no measure of pressure, in an out-of-hospital setting, who were variability reported in this trial. Therefore, we randomized to receive either nitroglycerin or urapi- imputed the s.d. of the change according to our dil. The ethical committee (Vienna, Austria) agreed hierarchy from other trials (last option: weighted that no informed consent had to be obtained at the mean s.d. of change from all trials; any drug any time of inclusion for randomization. However, the dose). There were no HR data reported. pitfall of this trial is that 16% of all randomized patients were excluded from the analyses Discussion that potentially biased the results. Consistent with this, there was significant heterogeneity when this This is the first systematic review investigating trial was combined with another trial studying the mortality and morbidity outcomes for all RCTs of same comparison groups.

Journal of Human Hypertension Different interventions for hypertensive emergencies MI Perez and VM Musini 604 In 19 of the excluded trials24–42 it was not possible pregnancy. However, Duley’s systematic review to determine how many patients had acute end- was not limited to patients with eclampsia and did organ damage or merely had elevation of blood not separately report outcomes in the eclampsia pressure. We believe that it would be misleading to patients. To the best of our knowledge there is no include these trials in this review as the impact of systematic review dealing exclusively with eclamp- antihypertensive drugs is potentially different. If sia and antihypertensive treatment. Thus, a sys- individual patient data could be obtained, the tematic review in this specific area is currently patients with acute end-organ damage could be needed. added to our review. The present review also does not provide any It was perhaps surprising and definitely disap- mortality and morbidity evidence from RCTs to pointing that we could find no RCT evidence to inform clinicians as to which first-line antihyper- answer the first question we have posed: Does tensive drug class provides more benefit than harm antihypertensive therapy as compared to placebo in hypertensive emergencies. This lack of evidence or no treatment change mortality and morbidity in was due to the fact that the trials were too small, did patients with hypertensive emergencies? The one not follow the patients for a long enough period of available placebo-controlled trial demonstrated that time and frequently failed to report all important blood pressure was reduced with drugs as compared outcomes. In addition all the RCTs except one were to the control treatment, however, it was too small open-label trials and therefore concealment of and of too short duration to assess morbidity and allocation was not possible in most cases. Although, mortality. We feel it is important for physicians to these shortcomings of the trials would not likely know that this is one of the clinical settings where affect mortality and morbidity outcomes, they could treatment is not supported by RCT evidence. bias blood pressure and HR data. Despite the lack of evidence it is not hard to Neither did we find RCTs that compared different accept the necessity of lowering blood pressure in strategies to reduce blood pressure. Thus, how fast those clinical settings where the excessive increases or how much blood pressure should be lowered in in blood pressure are the cause of the end-organ hypertensive emergencies remains unknown. damage. However, this is not necessarily the best Although it is unproven, it is highly likely that approach in settings where the excessive elevations antihypertensive therapy is an overall benefit in a of blood pressure are probably caused by end-organ hypertensive emergency and therefore a placebo- damage such as high blood pressure in the presence controlled trial to prove this would be unethical. of a cerebrovascular accident. The presently ac- What is clear is that this is a clinical area where cepted approach for the immediate treatment of properly conducted randomized trials are badly hypertensive emergencies in clinical practice is needed. At the present time RCTs could be con- primarily based on a series of cases published in ducted to compare different drug classes and 1959.1 In this study carried out over a period of 18 treatment strategies, for example, aggressive rapid months the author demonstrated the ability to lowering of blood pressure to a target versus reduced blood pressure with nitroprusside, within lowering the blood pressure slowly at a defined rate minutes, in eight patients with hypertensive emer- such as 5–10% every 2 h. What is also clear from this gencies (mostly patients with hypertensive ence- review is that any trial must follow patients long phalopathy), whose blood pressures had remained term and document mortality and morbidity. One of elevated after treatment with reserpine or hydrala- the best examples of an adequate RCT in an zine. However, he did not report clinical outcomes emergency setting is the corticosteroid randomiza- so we do not know whether these patients did better tion after significant head injury (CRASH) trial,55 as a result of the blood pressure lowering. Gifford where 10 000 patients with acute head injury were recommended prompt blood pressure reduction in randomized to intravenous steroids or placebo. Its clinical settings other than hypertensive encephalo- approach to handle ethical issues could serve as pathy such as intracerebral or subarachnoid hae- model when conducting a trial with hypertensive morrhage or acute left ventricular failure. The lack emergency patients. of RCT evidence leaves the distinct possibility that in some clinical settings defined as hypertensive emergencies immediate antihypertensive therapy Conclusions could be doing more harm than good. There is a hypertensive emergency not included Implications for practice in the present systematic review, eclampsia. Due to There is no evidence from RCTs that antihyperten- its pathophysiology and the involvement of the sive drugs reduce mortality or morbidity in patients infant as well as the mother, we felt this clinical with hypertensive emergencies, defined as marked entity must be studied separately from other hypertension associated with acute end-organ da- hypertensive emergencies and include outcomes in mage. Furthermore, there is insufficient RCT evi- the infant as well as the mother. There is a Cochrane dence to determine which drug or drug class is most systematic review54 that has studied the drugs for effective in reducing mortality and morbidity. There treatment of very high blood pressure during were some minor differences in degree of blood

Journal of Human Hypertension Different interventions for hypertensive emergencies MI Perez and VM Musini 605 pressure lowering between drug classes. However, infarction—executive summary: a report of the the clinical significance is unknown. American College of Cardiology/American Heart Asso- This review demonstrates a blood pressure low- ciation task force on practice guidelines (writing ering efficacy for nitrates, ACE inhibitors, diuretics, committee to revise the 1999 guidelines for the a-adrenergic antagonist, CCBs and dopamine ago- management of patients with acute myocardial infarc- tion). Circulation 2004; 110: 588–636. nists. Nitrates (including nitroprusside) have been 6 Dalen JE, Howe III JP. Dissection of the aorta. Current studied in most. Therefore, if a hypertensive diagnostic and therapeutic approaches. JAMA 1979; emergency patient cannot be treated as part of an 242: 1530–1532. RCT and a nitrate is available, it is a reasonable 7 Mattu A, Martinez JP, Kelly BS. Modern management choice of therapy. of cardiogenic pulmonary edema. Emerg Med Clin North Am 2005; 23: 1105–1125. 8 Wilson SR, Hirsch NP, Appleby I. Management of Implications for research subarachnoid haemorrhage in a non-neurosurgical RCTs are needed to assess different blood pressure centre. Anaesthesia 2005; 60: 470–485. lowering strategies and different first-line drug 9 Angeli P, Chiesa M, Caregaro L, Merkel C, Sacerdoti D, classes in patients with hypertensive emergencies. Rondana M et al. Comparison of sublingual captopril Outcomes in such trials must be mortality and total and nifedipine in immediate treatment of hypertensive emergencies. A randomized, single-blind clinical trial. serious adverse events at different times of follow- Arch Intern Med 1991; 151: 678–682. up such as 7 days, 1 month and including at least 10 Beltrame JF, Zeitz CJ, Unger SA, Brennan RJ, Hunt A, 6 months of follow-up of all patients. Moran JL et al. Nitrate therapy is an alternative to furosemide/morphine therapy in the management of acute cardiogenic pulmonary edema. J Card Fail 1998; Acknowledgements 4: 271–279. 11 Elliott WJ, Weber RR, Nelson KS, Oliner CM, Fumo MT, We acknowledge help and advice from Dr Ken Gretler DD et al. Renal and hemodynamic effects of Bassett, Stephen Adams for retrieving trials, and intravenous fenoldopam versus nitroprusside in se- Benji Heran, Michelle Wong and Jenny Chang for vere hypertension. Circulation 1990; 81: 970–977. 12 Hamilton RJ, Carter WA, Gallagher EJ. Rapid improve- comments on a draft. We also acknowledge the ment of acute pulmonary edema with sublingual trialists who provided us with additional informa- captopril. Acad Emerg Med 1996; 3: 205–212. tion from their studies. This paper has also been 13 Hirschl MM, Binder M, Bur A, Herkner H, Mu¨ llner M, published in the Cochrane Library as: MI Perez, VM Woisetschla¨ger C et al. Safety and efficacy of urapidil Musini. Pharmacological interventions for hyper- and sodium nitroprusside in the treatment of hyper- tensive emergencies. Cochrane Database of Systema- tensive emergencies. Intensive Care Med 1997; 23: tic Reviews 2008, Issue 1. art no.: CD003653. doi: 885–888. 10.1002/14651858.CD003653.pub3. 14 Hirschl MM, Schreiber W, Woisetschla¨ger C, Kaff A, Raab H. [Sublingual nitroglycerin or intravenous enalaprilat in preclinical treatment of hypertensive Conflict of interest patients with pulmonary edema]. [German] Z Kardiol 1999; 88: 208–214. There is no conflict of interest to declare. 15 Marigliano V, Santilli D, Fiorani M, Ariani A, Caccia- festa M, Ferri C et al. Hypertensive emergencies in old age: effects of angiotensin converting enzyme inhibi- References tion. J Hypertens Suppl 1988; 6: S91–S93. 16 McNair A, Krogsgaard AR, Hilden T, Nielsen PE. 1 Gifford RW. Treatment of hypertensive emergencies, Severe hypertension with cerebral symptoms treated including use of sodium nitroprusside. Mayo Clin Proc with furosemide, fractionated diazoxide or dihydrala- 1959; 34: 387–394. zine. Danish Multicenter Study. Acta Med Scand 1986; 2 Chobanian AV, Bakris GL, Black HR, Cushman WC, 220: 15–23. Green LA, Izzo Jr JL et al. The seventh report of the 17 Nelson GI, Silke B, Ahuja RC, Hussain M, Taylor SH. joint national committee on prevention, detection, Haemodynamic advantages of isosorbide dinitrate over evaluation, and treatment of high blood pressure: the frusemide in acute heart-failure following myocardial JNC 7 report. JAMA 2003; 289: 2560–2572. infarction. Lancet 1983; 1: 730–733. 3 Cherney D, Straus S. Management of patients with 18 Pastorelli R, Ferri C, Santucci A, Balsano F. New hypertensive urgencies and emergencies: a systematic therapeutic possibilities in hypertensive emergencies. review of the literature. J Gen Intern Med 2002; 17: Curr Ther Res, Clin Exp 1991; 50: 857–868. 937–945. 19 Rubio-Guerra AF, Vargas-Ayala G, Lozano-Nuevo JJ, 4 BASC. Interventions for deliberately altering blood Narvaez-Rivera JL, Rodriguez-Lopez L. Comparison pressure in acute stroke. Blood pressure in Acute between isosorbide dinitrate aerosol and nifedipine in Stroke Collaboration (BASC) (Cochrane Review). The the treatment of hypertensive emergencies. J Hum Cochrane Library, Issue 3, 2001, CD000039, Update Hypertens 1999; 13: 473–476. Software, John Wiley & Sons, Ltd. 20 Schreiber W, Woisetschla¨ger C, Binder M, Kaff A, Raab H, 5 Antman EM, Anbe DT, Armstrong PW, Bates ER, Hirschl MM. The nitura study—effect of nitroglycerin Green LA, Hand M. ACC/AHA guidelines for the or urapidil on hemodynamic, metabolic and respira- management of patients with ST-elevation myocardial tory parameters in hypertensive patients with

Journal of Human Hypertension Different interventions for hypertensive emergencies MI Perez and VM Musini 606 pulmonary edema. Intensive Care Med 1998; 24: pine, and clonidine: effect on blood pressure and heart 557–563. rate. J Cardiovasc Pharmacol 1988; 12: S154–S156. 21 Verma SP, Silke B, Hussain M. First-line treatment of 36 Ceyhan B, Karaaslan Y, Caymaz O, Oto A, Oram E, left ventricular failure complicating acute myocardial Oram A et al. Comparison of sublingual captopril and infarction: a randomised evaluation of immediate sublingual nifedipine in hypertensive emergencies. effects of diuretic, venodilator, arteriodilator, and Jpn J Pharmacol 1990; 52: 189–193. positive inotropic drugs on left ventricular function. 37 Guerrera G, Melina D, Capaldi L, Mauro R, Colivicchi J Cardiovasc Pharmacol 1987; 10: 38–46. F, Cardillo C et al. [Sublingually administered capto- 22 Wu SG, Lin SL, Shiao WY, Huang HW, Lin CF, Yang pril versus nifedipine in hypertension emergencies]. YH. Comparison of sublingual captopril, nifedipine [Italian] Minerva Cardioangiol 1990; 38: 37–44. and prazosin in hypertensive emergencies during 38 Pascale C, Zampaglione B, Marchisio M. Management hemodialysis. 1993; 65: 284–287. of hypertensive crisis: nifedipine in comparison with 23 Yang HJ, Kim JG, Lim YS, Ryoo E, Hyun SY, Lee G. captopril, clonidine and furosemide. Curr Ther Res, Nicardipine versus nitroprusside infusion as antihy- Clin Exp 1992; 51: 9–18. pertensive therapy in hypertensive emergencies. J Int 39 Pilmer BL, Green JA, Panacek EA, Elliot WJ, Murphy Med Res 2004; 32: 118–123. MB, Rutherford W et al. Fenoldopam mesylate versus 24 Bussmann WD, Kenedi P, von Mengden HJ, Nast HP, sodium nitroprusside in the acute management of Rachor N. Comparison of nitroglycerin with nifedipine severe systemic hypertension. J Clin Pharmacol 1993; in patients with hypertensive crisis or severe hyper- 33: 549–553. tension. Clin Investig 1992; 70: 1085–1088. 40 Pujadas R, Jane´ J, Forno´s C, Gago MJ, de la Concepcio´n 25 Conen D, Ru¨ ttimann S, Noll G, Schneider K, Mu¨ ller J. N. Comparison of sublingual captopril and nifedipine Short- and long-term cerebrovascular effects of nitren- in hypertensive crises. Arch Intern Med 1987; 147: dipine in hypertensive patients. J Cardiovasc Pharma- 175–176. col 1988; 12: S64–S68. 41 Reisin E, Huth MM, Nguyen BP, Weed SG, Gonzalez 26 Dadkar VN, Karnik ND, Izar M, Sharma SR, Gandhi YP, FM. Intravenous fenoldopam versus sodium nitroprus- Narawane NM et al. Sublingual nifedipine and side in patients with severe hypertension. Hyperten- captopril in hypertensive urgencies and emergencies. sion 1990; 15: I59–I62. Indian Heart J 1993; 45: 185–187. 42 Zampaglione B, Pascale C, Marchisio M, Santoro A. 27 Marghli S, Bouida W, Elatrous S, Boujdaria R, Nouira The use of lacidipine in the management of hyperten- S, Abroug F. A prospective, randomized and con- sive crises: a comparative study with nifedipine. trolled study of three intravenous antihypertensive J Cardiovasc Pharmacol 1994; 23: S116–S118. drugs in severe hypertension. Reanimation Urgences 43 Franklin C, Nightingale S, Mamdani B. A randomized 1997; 6: 289–296. comparison of nifedipine and sodium nitroprusside in 28 Moritz RD, de Queiroz LP, Pereira MR, Scotinni MA. severe hypertension. Chest 1986; 90: 500–503. [Comparative study of the use of nifedipine and 44 Bertel O, Conen D, Radu¨ EW, Mu¨ ller J, Lang C, Dubach captopril in hypertensive emergencies]. [Portuguese] UC. Nifedipine in hypertensive emergencies. Br Med Arq Bras Cardiol 1989; 52: 323–326. J (Clin Res Ed) 1983; 286: 19–21. 29 Neutel JM, Smith DH, Wallin D, Cook E, Ram CV, 45 Borghi C, Bacchelli S, Esposti DD, Bignamini A, Fletcher E et al. A comparison of intravenous nicardi- Magnani B, Ambrosioni E. Effects of the administra- pine and sodium nitroprusside in the immediate treatment of severe hypertension. Am J Hypertens tion of an angiotensin-converting enzyme inhibitor 1994; 7: 623–628. during the acute phase of myocardial infarction in 30 Nielsen PE, Krogsgaard A, McNair A, Hilden T. patients with arterial hypertension. SMILE Emergency treatment of severe hypertension evaluated Study Investigators. Survival of Myocardial Infarction in a randomized study. Acta Med Scand 1980; 208: Long-term Evaluation. Am J Hypertens 1999; 12: 473–480. 665–672. 31 Panacek EA, Bednarczyk EM, Dunbar LM, Foulke GE, 46 Lisk DR, Grotta JC, Lamki LM, Tran HD, Taylor JW, Holcslaw TL. Randomized, prospective trial of fenol- Molony DA et al. Should hypertension be treated after dopam vs sodium nitroprusside in the treatment of acute stroke? A randomized controlled trial using acute severe hypertension. Fenoldopam Study Group. single photon emission computed tomography. Arch Acad Emerg Med 1995; 2: 959–965. Neurol 1993; 50: 855–862. 32 Perez C, Dougnac A, Alvarez M, Andresen M, Dı´az O, 47 Nelson GIC, Nerma SP, Silke B, Abdulan S, Taylor Geni R et al. [Sublingual captopril versus nifedipine in SH. Management of hypertension complicating acute the treatment of hypertensive crisis]. [Spanish] Rev myocardial infarction (MI): comparison of sympathetic Med Chil 1991; 119: 402–405. (labetalol) and calcium channel (nifedipine) blockade. 33 Risler T, Bohm R, Wetzchewald D, Nast HP, Koch HH, [abstract] Aust NZ J Med (Suppl) 1984; 14: 577. Stein G et al. A comparison of the antihypertensive 48 Cotter G, Metzkor E, Kaluski E, Faigenberg Z, Miller R, efficacy and safety of felodipine IV and nifedipine IV Simovitz A et al. Randomised trial of high-dose in patients with hypertensive crisis or emergency not isosorbide dinitrate plus low-dose furosemide versus responding to oral nifedipine. Eur J Clin Pharmacol high-dose furosemide plus low-dose isosorbide dini- 1998; 54: 295–298. trate in severe pulmonary oedema. Lancet 1998; 351: 34 Rohr G, Reimnitz P, Blanke P. Treatment of hyperten- 389–393. sive emergency. Comparison of a new dosage form of 49 Yoshida N. Antihypertensive effect of the long-acting the calcium antagonist nitrendipine with nifedipine Ca antagonist amlodipine vs sustained-release capsules. Intensive Care Med 1994; 20: 268–271. nifedipine in patients with acute aortic dissection. 35 Spah F, Grosser KD. Treatment of hypertensive Rinsho to Kenkyu (Jpn J Clin Exp Med) 1998; 75: urgencies and emergencies with nitrendipine, nifedi- 1419–1422.

Journal of Human Hypertension Different interventions for hypertensive emergencies MI Perez and VM Musini 607 50 Annane D, Bellissant E, Pussard E, Asmar R, Lacombe 53 McNair A, Krogsgaard AR, Hilden T, Nielsen PE. F, Lanata E et al. Placebo-controlled, randomized, Reversibility of cerebral symptoms in severe hyperten- double-blind study of intravenous enalaprilat efficacy sion in relation to acute antihypertensive therapy. and safety in acute cardiogenic pulmonary edema. Danish Multicenter Study. Acta Medica Scandinavica Circulation 1996; 94: 1316–1324. Supplementum 1985; 693: 107–110. 51 Krogsgaard AR, Hilden T, McNair A, Nielsen PE. 54 Duley L, Henderson-Smart DJ, Meher S. Drugs for Cerebral symptoms and blood pressure during parent- treatment of very high blood pressure during preg- eral administration of , dihydralazine nancy (Cochrane Review). The Cochrane Library, Issue and diazoxide. Danish Mmulticentre Study on 3, 2006, CD001449, Update Software, John Wiley & acute severe hypertension. Acta Med Scand 1983; Sons, Ltd. 678(Suppl): 51–60. 55 Roberts I, Yates D, Sandercock P, Farrell B, Wasserberg 52 Krogsgaard AR, McNair A, Hilden T, Nielsen PE. J, Lomas G et al. Effect of intravenous corticosteroids on Reversibility of cerebral symptoms in severe hyperten- death within 14 days in 10 008 adults with sion in relation to acute antihypertensive therapy. clinically significant head injury (MRC CRASH trial): Danish Multicenter Study. Acta Med Scand 1986; 220: randomised placebo-controlled trial. Lancet 2004; 364: 25–31. 1321–1328.

Journal of Human Hypertension