Journal of Human (1999) 13, 413–417  1999 Stockton Press. All rights reserved 0950-9240/99 $12.00 http://www.stockton-press.co.uk/jhh ORIGINAL ARTICLE Double-blind comparison of and on cough and blood pressure in unselected hypertensive patients

WJ Elliott, for the Eprosartan Study Group

The effects of a new receptor antagonist, with a 3.45-fold higher risk of definite cough (14/261 vs -Overall cough incidence (from spon .(0.018 ؍ eprosartan (200 or 300 mg b.i.d.) and enalapril (5–20 mg 4/259, P u.i.d.) on cough and blood pressure were compared in taneous reports from patients, or investigator’s a 26-week, double-blind, randomised, parallel-group, observation) was also more frequent with enalapril, as multicentre, international study involving 528 patients compared to eprosartan. Both agents reduced blood with hypertension. Uptitration of doses was based on pressure significantly compared to baseline, although clinic blood pressure measurements during the first 12 the eprosartan-treated group had a slightly higher weeks, after which (12.5–25 response rate (defined as sitting diastolic blood press- mg/day) could be added. The frequency and intensity of ure Ͻ90 mm Hg, or at least a 10 mm Hg reduction from cough was assessed by a standardised questionnaire baseline), both at end of titration (70.3% vs 62.6%, administered at each clinic visit. The primary end-point P Ͻ 0.05) and after 26 weeks (81.7% vs 73.5%, These data suggest that, in unselected .(0.018 ؍ was the incidence of persistent, dry cough not due to P upper respiratory infection; change in sitting diastolic hypertensive patients, eprosartan is associated with blood pressure and overall incidence of cough were less cough and a somewhat higher responder rate secondary end-points. During the first 12 weeks of dou- than enalapril. ble-blind therapy, enalapril treatment was associated

Keywords: angiotensin II receptor antagonist; ACE inhibitor

Introduction Patients and methods Angiotensin-coverting enzyme (ACE)-inhibitors are Study design useful in treating hypertension, , and renal impairment, but are most often discontinued because of the development of a persistent, dry This 41-centre, international double-blind, double- cough, the prevalence of which has been reported dummy, parallel-group study consisted of three as high as 25–69%.1–6 Unlike ACE-inhibitors, angi- treatment periods: a 3- to 5-week single-blind pla- otensin II receptor antagonists do not inhibit brady- cebo run-in (‘baseline’) period, an 18-week double- kinin degradation in vitro, and have been associated blind titration period, and an 8-week maintenance with a lower incidence of cough in carefully selec- period. Initially, patients were randomised to ted patients who demonstrated the cough response receive either eprosartan 200 mg twice daily or enal- to an ACE-inhibitor.7–10 Eprosartan is a new, highly- april 5 mg once daily. At 3-week intervals, the office selective non-peptide, non-biphenyl, non-tetrazole blood pressure was measured and the dose titrated angiotensin II receptor blocker with few effects on as needed to a maximum dosage of 300 mg twice 11 the cyctochrome P450 system. This paper describes daily eprosartan or 20 mg once daily enalapril. At the results of a large clinical study designed to com- the end of week 12, the maximum doses of double- pare the effects of eprosartan and enalapril on cough blind medication were supplemented with HCTZ and blood pressure (BP) in unselected hyperten- 12.5 mg once daily, and increased if needed to a sive patients. maximum of 25 mg daily. This study was conducted in accordance with the Declaration of Helsinki as amended. The protocol and informed consent signed by each participant Correspondence: William J Elliott, Rush Presbyterian-St. Luke’s were approved by the Institutional Review Board or Medical Center, Department of Preventive Medicine, 1725 Harri- son Street, #117, Chicago, IL 60612, USA Ethics Committee at each study site. Received 21 July 1998; revised 8 January 1999; accepted 4 Febru- ary 1999 Cough: eprosartan vs enalapril WJ Elliott 414 Patients tarily discontinued medication because of coughing. Maximum cough was defined as the sum of definite, Patients were eligible if they were at least 18 years of probable, and possible cough, and tickle in throat. age and had essential hypertension. Premenopausal Antihypertensive efficacy was measured at trough as women of childbearing potential were required to be the mean change from baseline in sitting and stand- using hormonal, barrier, or intrauterine contracep- ing systolic blood pressure (SBP) and diastolic blood tion. pressure (DBP). Responders were patients whose Patients with secondary forms of hypertension, SitDBP was less than either 90 mm Hg, or 100 advanced hypertensive retinopathy, average sitting Ͼ mm Hg and had decreased from baseline by at least systolic blood pressure (SitSBP) 200 mm Hg, 10 mm Hg. advanced atrioventricular conduction defects, ven- Baseline values were obtained at the last visit of tricular tachyarrhythmias requiring therapy, brady- the placebo run-in period, before randomisation. For cardia, prior myocardial infarction or cerebrovascu- blood pressure only, baseline was the mean of the lar accident within the past 90 days, congestive last two visits prior to randomisation. Comparisons heart failure being treated with ACE inhibitors or at baseline with respect to categorical demographic , angina being treated with nitrates, beta- and clinical characteristics were performed using blockers, or calcium channel blockers, unstable dia- the chi-square technique with adjustment for sam- betes mellitus, or presence of clinically significant pling differences across centres. For continuous renal or hepatic disease or another concurrent sev- variables, baseline differences were assessed by ere disease, were excluded. Patients were also analysis of variance (PROC GLM in SAS), which excluded if they had conditions which could inter- included terms for centre, medication regimen, and fere with the assessment of cough: emphysema, interaction with centre. Centralised data collection asthma or chronic bronchitis, or upper respiratory and statistical analysis were performed by full-time infection within 2 weeks of screening. Use of antico- employees of the sponsor (SmithKline Beecham agulants or another investigational drug within 30 Pharmaceuticals, Inc, Philadelphia, PA, USA). days of enrolment, chronic sympathomimetic amine For the categorical variables of cough incidence, or nonsteroidal anti-inflammatory drugs (other than cough severity, and response rate, comparisons low-dose aspirin) within 7 days of enrolment, and between the eprosartan and enalapril regimens dur- concomitant use of antidepressants or medications ing the double-blind periods were performed using a known to affect blood pressure or cough were not Cochran-Mantel-Haenszel (CMH) statistic adjusting allowed. for centre interaction with regimen, which was assessed with the Breslow-Day test (PROC FREQ in Assessments of cough and blood pressure SAS). For vital signs, an analysis of variance was performed at each visit, which included terms for A pulmonary assessment (physician’s physical centre, medication regimen, and regimen-by-centre examination of the chest by auscultation and per- interaction. If the latter was not significant cussion, if abnormal) was performed at screening, at (P Ͼ 0.10), it was removed from the model. Sub- randomisation, at weeks 6 and 12 of the titration groups of age, gender, race, prior use of antihyper- phase, and at the end of the maintenance phase. No tensives, and baseline severity of hypertension were pulmonary function testing was performed. The compared using similar techniques for cough, blood presence and character of cough were assessed by pressure, and response rate. the investigator regarding type, duration, severity, frequency, and probable cause of cough. Cough was categorised according to a hierarchy of severity as Results definite, probable, possible, or a ‘tickle in throat’. At Patient demographics and treatment regimens each visit, the patient completed a quality-of-life questionnaire, which included a five-point toler- The demographic and clinical characteristics of the ability rating scale of both frequency (never, seldom, patients are shown in Table 1; there were no signifi- occasional, frequent, or constant) and severity for cant differences between groups in any of these each of 10 commonly-experienced adverse effects parameters. In the eprosartan group, 138 patients with antihypertensive medications (one of which were titrated to the maximum dose of 300 mg twice was cough). Cough that occurred at any time during daily, whereas 151 patients in the enalapril group the trial was recorded as an adverse experience. were titrated to the maximum dose of 20 mg once Blood pressure and heart rate were measured in trip- daily. In each group, 81 patients initially received licate at each visit during the four study periods, HCTZ 12.5 mg once daily, and then 51 patients from according to standard techniques. the eprosartan group and 50 patients from the enala- pril group received the maximum dose plus HCTZ 25 mg once daily. Statistical analyses

The primary end-point was definite cough of inter- Incidence of cough est, defined as the incidence of persistent, non-pro- ductive (dry) cough associated with treatment and A significantly higher incidence of definite cough not due to an upper respiratory infection as judged was found in the enalapril group prior to the intro- by the investigator. The cough must have been duction of HCTZ, the primary timepoint of interest, present for at least 2 weeks unless the patient volun- when compared with the eprosartan group Cough: eprosartan vs enalapril WJ Elliott 415 Table 1 Patient demographics and clinical characteristics at base- subjectively-reported cough, whether assessed as line ‘definite cough’ or ‘maximum cough’ (22 vs 13%, ␹2 = 8.15; P = 0.01). Seven patients randomised to Parameter Eprosartan Enalapril (n = 264) (n = 264) enalapril treatment discontinued participation because of cough; only two patients randomised to Age (years; mean ± s.e.m.) 55.6 ± 0.7 56.0 ± 0.7 eprosartan stopped due to this perceived adverse Sex (male:female; %) 57:43 56:44 effect; the cough resolved in all patients after dis- Race (Black: Caucasian: Asian: 21:225:2:16 19:231:4:10 continuing study medication. Other) There were no significant differences in demo- Weight (kg; mean ± s.e.m.) 82.8 ± 1.1 84.1 ± 1.1 Height (cm; mean ± s.e.m.) 168.1 ± 0.6 168.5 ± 0.6 graphic characteristics of patients who coughed, Mean SitSBP (mm Hg; 156.4 ± 0.9 156.2 ± 0.9 compared to those who did not report cough, with mean ± s.e.m.) the exception of gender and smoking history. Mean SitDBP (mm Hg; 100.7 ± 0.3 101.2 ± 0.3 Women were significantly more likely to report ± mean s.e.m.) cough (‘maximum cough’) during the monotherapy Prior antihypertensive therapy 223:41 215:49 = (Yes:No) treatment period than were men (odds ratio 1.86, Prior ACE-inhibitor therapy 157:107 140:124 95% CI: 1.15–3.0): a similar trend was seen for the (Yes:No or unknown) end-point of ‘definite cough’. Similarly, there was a Prior ACE-inhibitor-associated 1:263 3:261 trend for smokers to report more ‘maximum cough’ cough = Current cigarette smokers (n, 36 (14%) 31 (12%) than non-smokers (odds ratio 1.72, 95% CI: 0.95– %) 3.12), which was also seen for ‘definite cough’. Data were also collected regarding adverse experi- ences which might have contributed to cough, but were not attributed to ‘definite cough’, eg, viral or (Figure 1). Definite cough occurred in 14 (5.4%) upper respiratory tract infection, or pharyngitis. In patients randomised to enalapril and four (1.5%) each case, the incidence of symptoms which might patients receiving eprosartan; the incidence of defi- have been potential confounders for ‘definite cough’ nite cough during placebo-treatment was 1.9% or were found less frequently in the eprosartan-treated 0.8% in the groups subsequently treated with epros- group: any adverse experience (76% vs 81%), phar- artan or enalapril, respectively. Patients randomised yngitis (17% vs 24%), rhinitis (12% vs 16%), upper to enalapril were 3.45 (95% CI 1.26–10.0) times respiratory tract infection (12% vs 16%), viral infec- more likely to experience definite cough than tion (3% vs 6%). patients receiving eprosartan. During the entire 26- week double-blind treatment period, patients in the Analyses of blood pressure enalapril group were 3.85 times (95% CI 1.48–10.3) more likely to experience definite cough: the inci- Eprosartan and enalapril produced similar mean dences of definite cough were 6.1% and 1.5% in the changes from baseline for SitDBP, SitSBP, StaDBP, enalapril and eprosartan groups, respectively. Simi- and StaSBP when evaluated at the end of the lar results were obtained in analyses of ‘maximum titration phase, maintenance phase, or at the end of cough’ during monotherapy, with enalapril being study, when SitDBP and SitSBP were decreased by associated with a 1.41 (95% CI: 1.06–1.89) times 12.9 mm Hg and 15.5 mm Hg, respectively, in the higher risk of ‘maximum cough’ than eprosartan. eprosartan group; whereas in the enalapril group, Results from patient questionnaires also showed these measures decreased by 11.9 mm Hg and 14.7 that enalapril was associated with a higher rate of mm Hg, respectively. Analysis of variance revealed no significant difference between the two treatments on sitting or standing vital signs at the end of study.

Response rate A somewhat higher percentage of eprosartan-treated patients responded to treatment when measured at the end of the titration and maintenance phases (Table 2). The response rate at the end of study was

Table 2 Blood pressure response* rates after 12 and 26 weeks of eprosartan or enalapril (±HCTZ)

Eprosartan Enalapril P (n = 264) (n = 264)

Response rate 70.3% 62.6% Ͻ0.05 at 12 weeks Response rate 81.7% 73.4% Ͻ0.018 at 26 weeks

Figure 1 Incidence of ‘definite cough’ during 12 weeks of blinded *Patients with a SitDBP Ͻ90 mm Hg, OR Ͻ100 mm Hg and a therapy in patients randomised to eprosartan (left) or enalapril reduction of 10 mm Hg from baseline were defined as (right). ‘responders’. Cough: eprosartan vs enalapril WJ Elliott 416 significantly higher in the eprosartan group with treated with enalapril: at the end of titration, the 81.7% of eprosartan-treated patients vs 73.4% of response rate for enalapril was 57.9% of Black enalapril-treated patients (P Ͻ 0.05). In pre-specified patients and 85.3% of Caucasian patients; whereas subgroup analyses, the response rate at end of for eprosartan, the responder rates were 81.0% and titration in black patients was 81.0% for the eprosar- 87.5% for Black and Caucasian patients, respect- tan group and 57.9% for the enalapril group ively. (P Ͻ 0.05). The differences among Caucasian In this study, enalapril was found to be less effec- patients were not significant. tive for lowering blood pressure in Black patients There were no significant differences between than in other races, as has been seen previously with treatment groups in changes in quality of life this and other ACE-inhibitors.13,14 For example, measurements (using the General Well Being Index), when 1-year monotherapy was compared laboratory parameters, serious adverse events, or with drugs from five other pre-specified changes in vital signs of concern. classes, the ACE inhibitor was the least effective.13 Only 32% of Black patients under 60 years of age Discussion had their blood pressures controlled by captopril, whereas captopril was identical to placebo in Black The results from this study demonstrate that epros- patients over 60 years of age.15 In addition to the artan produces a lower incidence of persistent, non- lower efficacy of ACE inhibitors in Black patients, productive cough than enalapril in patients with there is a greater risk for both cough (2.6-fold) and essential hypertension. This type of cough has (4.5-fold) associated with ACE-inhibi- emerged as an important side effect of ACE inhibi- tors in Blacks, which may be independent of tors which limits their utility in many patients. Dur- dose.12,15 The mechanisms accounting for why Black ing this 26-week double-blind trial in patients not patients are less responsive to ACE inhibitors and selected for previous cough with ACE-inhibitors, the have a greater risk of ACE inhibitor-associated incidence of definite, persistent, non-productive adverse effects are unknown. cough was 1.5% for patients treated with eprosartan, In summary, patients receiving eprosartan had a vs 5.4% with enalapril; a similar pattern was seen statistically significantly lower risk of cough than in secondary analyses of ‘maximum cough’. did patients receiving enalapril, a result that was Although the 5.4% prevalence of ACE-inhibitor- consistent with the results of patient self-reported associated cough is far lower than that reported from cough and patient withdrawals due to cough. Both uncontrolled trials, it is not much lower than the agents lowered blood pressure to a similar degree; 7.7% reported in an unselected consecutive series however, patients treated with eprosartan had a of hypertensive patients which used higher response rate. Other than cough and pharyn- dechallenge/rechallenge methodology to be sure gitis (a symptom which could have been interpreted that the cough was associated with ACE-inhibitor by patients as ‘tickle in the throat’) occurring more therapy.12 Yeo and Ramsay3 have clearly shown that frequently in enalapril-treated patients, both agents the prevalence of cough with ACE-inhibitors is had similar adverse experience and safety profiles. inversely proportional to the amount of work Thus, eprosartan represents an attractive alternative expended to detect and ascribe a cause to it. In this to enalapril for treating patients with essential study, patient-reported ‘definite cough’ was hypertension. unchanged from baseline levels prior to receiving eprosartan, and the incidence of patient-reported Acknowledgement ‘maximum cough’ actually decreased from baseline while on treatment. This study was supported by grants and contracts Both eprosartan and enalapril produced similar from SmithKline Beecham Pharmaceuticals Inc. reductions in DBP and SBP; however, a significantly greater percentage of patients treated with eprosar- References tan responded to treatment. This observed differ- ence may be due to the relatively short elimination 1 Town GI, Hallwright GP, Maling TJ, O’Donnell TV. half-life of enalapril, the relatively low trough-peak Angiotensin converting enzyme inhibitors and cough. N Engl J Med 1987; 300: 161–163. ratio of once-daily enalapril, the twice-daily admin- 2 Ravid D, Lishner M, Lang R, Ravid M. Angiotensin- istration of eprosartan, or a combination of these and converting enzyme inhibitors and cough: a prospective other factors. The probability of responding to treat- evaluation in hypertension and in congestive heart ment was approximately 13% greater for the epros- failure. J Clin Pharmacol 1994; 34: 1116–1120. artan group than for patients treated with enalapril. 3 Yeo WW, Ramsay LE. Persistent dry cough with enala- Subgroup analyses of patients according to age, pril: incidence depends on method used. J Hum gender, race, prior use of hypertensive drugs, or Hypertens 1990; 4: 517–520. severity of hypertension confirmed the higher inci- 4 Israili ZH, Hall WD. Cough and angioneurotic edema dence of cough with enalapril, and suggested that associated with angiotensin-converting enzyme inhibi- eprosartan has a more uniform effect on blood press- tor therapy. Ann Intern Med 1992; 117: 234–242. 5 Semple PF. Putative mechanisms of cough after treat- ure across the patient population, as differences in ment with angiotensin converting enzyme inhibitors. responder rate between the two treatments were J Hypertens 1995; 13: (Suppl 3): S17–S21. consistent. When patients were subgrouped accord- 6 Simon SR, Moser M, Black HR, Berland WE. Cough ing to race, Black patients treated with eprosartan and ACE-inhibitors. Arch Intern Med 1992; 152: had a significantly greater response rate than those 1698–1700. Cough: eprosartan vs enalapril WJ Elliott 417 7 Messerli FH, Weber MA, Brunner HR. Angiotensin II ale San Carlo Borromeo; T Erwteman, MD, Medical receptor inhibition. A new therapeutic principle. Arch Center ‘Jan van Gooyer’; H Mulder, MD, Research Intern Med 1996; 156: 1957–1965. Organization ‘Good Clinical Practice’; Prof L 8 Lacourcie`re Y et al. Effects of modulators of the - Sorstein, Det Internasjonale Kardiologiske Institutt; angiotensin-aldosterone system on cough. C Calvo, MD, Hospital Xeral de Galicia; J Carcia- Cough Study Group. J Hypertens 1994; 12: 1387–1393. 9 Benz J et al. , a new angiotensin II receptor Puig, MD, Hospital La Paz; J Honorato, MD, Clinica antagonist: a double-blind study comparing the inci- Universitaria de Navarra; J Mora, MD, Fundacion dence of cough with and hydrochlorothiaz- Puigvert; J Redon, MD, Hospital de Sagunto; D Sanz, ide. J Clin Pharmacol 1997; 37: 101–107. MD, Clinical Puera de Hierro; Prof B Maharaj, Uni- 10 Chan P et al. Double blind comparison of losartan, lisi- versity of Natal Medicine; FJ Maritz, MB, ChB, FCP, nopril, and metolazone in elderly hypertensive MMed, University of Stellenbosch, Tygerberg Hospi- patients with previous angiotensin-converting enzyme tal; S Middlemost, MD, Johannesburg Hospital; Prof inhibitor–induced cough. J Clin Pharmacol 1997; 37: D Myburg, Institute of Aviation Medicine; M Chatt- 253–257. man, MD, Arizona Clinical Trials, Inc.; R Davidson, 11 Edwards RM et al. Pharmacological characterization of MD, University of Washington School of Medicine; the nonpeptide angiotensin II receptor antagonist, SK& F 108566. J Pharmacol Exp Ther 1992; 260: 175–181. SH Dyke, MD, FACC, New England Clinical 12 Elliott WJ. Higher incidence of discontinuation of Research Center; WJ Elliott, MD, PhD, Rush-Presby- ACE-inhibitors due to cough in Blacks. Clin Pharma- terian-St. Luke’s Medical Center; T Fagan, MD, Uni- col Ther 1996; 60: 582–588. versity of Arizona; V Gandhi, MD, Hines Veterans 13 Moser M. Black-white differences in response to anti- Affairs Hospital; H Gavras, MD, Boston University hypertensive medication. J Natl Med Assoc 1995; 87 Medical Center; D Ginsberg, DO, Harleysville Medi- (Suppl 8): 612–613. cal Associates, Inc.; S Kent, MD, Sansum Medical 14 Materson BJ et al. Single-drug therapy for hyperten- Research Foundation; B Levine, MD, Veterans sion in men. A comparison of six antihypertensive Affairs Medical Center West Los Angeles; T Little- agents with placebo. N Engl J Med 1993; 328: 914–921. john, MD, Piedmont Research Associates; NM 15 Brown NJ, Ray WA, Snowden M, Griffin MR. Black Americans have an increased rate of angiotensin con- Lunde, MD, Kidney Disease and Critical Care verting inhibitor-associated angioedema. Clin Pharma- Associates; DT O’Connor, Jr., MD, Veterans Affairs col Ther 1996; 60: 8–13. Medical Center; R Phillips, MD, PhD, FACC, The Mount Sinai Medical Center; D Ploth, MD, Medical University of South Carolina; W White, MD, Univer- Appendix sity of Connecticut Health Sciences Center; R Casey, MD, 407–1235 Trafalgar Road North; H Conter, MD, The following Investigators participated in the MSHJ Research Associates, Inc.; M Robitaille, MD, study: Hoˆpital Laval; JM Coupez, MD, Clinique de la Fais- anderie; G Creytens, MD, Salve Mater UPC; Prof R Bigi, MD, Ospedale Regionale di Bormio e Sond- Danchin, Hoˆpitaux de Brabois-Hoˆpital d’Adultes; alo; C Constantini, MD, Ospedale di Legnano; Prof J-Achille Lozic, MD, 59 Boulevard Gaston Serpette; G Maschio, Ospedale Civile Maggiore di Borgo- J Piro, MD, 215 Avenue de Prado; P Voiriot, MD, trento; E Ronchi, MD, Ospedale Generale Provinci- Centre ‘Les Nations’.