Journal of Human (2001) 15, 857–862  2001 Nature Publishing Group All rights reserved 0950-9240/01 $15.00 www.nature.com/jhh ORIGINAL ARTICLE Comparison of quality of life and cough on eprosartan and in people with moderate hypertension

E Breeze1, EC Rake2, MD Donoghue1 and AE Fletcher1 1Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK; 2Institute for Cancer Research, London School of Hygiene and Tropical Medicine, London, UK

The objective of this study was to compare quality of life were that 17.8% of enalapril patients and 13.2% of epros- and incidence of dry persistent cough among patients artan patients withdrew from randomised treatment. treated with eprosartan and enalapril for mild–moderate Those on enalapril were twice as likely to have gained hypertension. This was a randomised 26-week double- a definite or possible cough by study end point as those At monotherapy .0.099 ؍ blind controlled trial carried out in clinics in nine coun- on eprosartan (7.6% vs 3.2%) P tries of North America, Europe and South Africa. A total end point the differences were greater (9.9% vs 2.1%) Patients .0.001 ؍ of 529 patients aged 18 and over with diastolic blood and of statistical significance, P pressure between 95 mm Hg and 114 mm Hg were stud- treated with enalapril, however, had small but significant ied. Treatment comprised of eprosartan or enalapril improvements in measures of self-control and total monotherapy for 12 weeks with the option of hy- PGWB compared with those on eprosartan. The effect drochlorothiazide addition for the remaining 14 weeks. sizes of 0.2 or less indicated that there were small differ- The primary outcome measures were cough and the ences. In conclusion eprosartan was associated with Psychological General Wellbeing Index (PGWB) total fewer coughs than enalapril but it performed less well and subscales (anxiety, self-control, depression, gen- on some aspects of quality of life. eral health, positive wellbeing and vitality). The results Journal of Human Hypertension (2001) 15, 857–862

Keywords: quality-of-life; clinical trial; A-II receptor antagonist; ACE-I inhibitor

Introduction receptor antagonist.22 One hypothesised advantage of this generation of drugs is reduction in the ACE-inhibitor drugs for essential hypertension are adverse effect of cough. The objective of this trial effective at lowering blood pressure but are known was to compare the quality of life, including the for producing a dry unproductive tickly cough1–5 6 incidence of persistent non-productive cough, in which may reduce quality of life. As high blood mild-to-moderately hypertensive patients treated pressure is often symptomless and its treatment with eprosartan and enalapril. involves daily life-long medication, the quality of life of patients while on the drugs is an important factor in the choice of drug and may influence com- Methods pliance.7,8 Although studies have compared earlier Patients aged 18 or more with sustained mild– generations of drugs, including enalapril,9–13 and moderate hypertension were eligible for the trial cough side-effects of nonpeptide -II (AII) (mean sitting diastolic blood pressure (DBP) receptor antagonist,14–18 there is little information between 95 mm Hg and 114 mm Hg inclusive at about other aspects of quality of life on them19–21 three successive visits). The 41 centres in the study and none comparing these with ACE-I inhibitors. were situated in nine countries. Ethical approval Eprosartan is a highly selective non-peptide A-II was obtained from the relevant authorities at the London School of Hygiene and Tropical Medicine Ethics Committee and at each centre; patients pro- Correspondence: Ms Elizabeth Breeze, Epidemiology Unit, vided written consent to participate in the trial. Department of Epidemiology and Population Health, London Patients were screened before entering a 3–5 week School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, UK. E-mail elizabeth.breezeȰIshtm.ac.uk placebo run-in period. Eligible patients were ran- Received 20 December 2000; revised 3 June 2001; accepted 3 domised to treatment with either eprosartan 200 mg July 2001 twice daily or enalapril 5 mg once daily. Doses were Quality of life on eprosartan and enalapril E Breeze et al 858 titrated over an 18 week period up to a possible Table 1 Definitions for cough maximum of 300 mg eprosartan twice daily or 20 mg Category Persistent Dry/wet Duration of enalapril once daily. To ensure ‘blinding’, placebo a forms of the second drug were given alongside the cough active drug. After 12 weeks patients on maximum Definite cough Yes Dry At least 14 dose of the medication could be prescribed the days or caused diuretic (HCTZ) openly, the to withdraw initial dose being 12.5 mg once daily, later doubled Probable cough Yes Dry Not known if necessary. Finally, there was an 8-week mainte- Possible cough Yes Not known As definite cough nance period. or No/not known Dry As definite The main clinical outcomes were dry unproduc- cough tive persistent cough assessed by the investigator using a questionnaire, and control of blood pressure aThe difference between end date and start date or, if cough still (reduction of sitting DBP to less than 90 mm Hg or present at time of visit, the difference between visit date and a reduction to less than 100 mm Hg if this was at start date. least 10 mm Hg below baseline value). Based on evi- dence available at the time, the sample size was set first be prescribed. If there was no questionnaire for to detect a difference of 7% between treatments in this visit then the most recent questionnaire was the incidence of dry cough, with 90% power at a used. Patients included in the per protocol analysis significance level of 5%. The clinical outcomes are had completed all five questionnaires and the clin- reported elsewhere.23–25 ician assessed that they had completed the treat- The primary quality of life outcomes were self- ment period, but not necessarily complied strictly assessed dry unproductive persistent cough and the with pill-taking. Psychological General Wellbeing Index (PGWB)26 Incidence of cough by an end point is defined as with six subscales. This well-validated instru- the percentage of those without cough at baseline ment27,28 has been used in international studies who acquired a cough by the end-point. To cater for investigating the effects of antihypertensives on incomplete information, coughs were defined as quality of life.29 The study had 80% power to detect definite, probable or possible (Table 1). Cough out- a standardised difference30 of 0.3 or more on the comes were compared by chi-squared test. PGWB subscales at a significance level of 1%. The If patients had missed fewer than 20% of items on secondary outcomes, separate from the PGWB, were a PGWB scale their scores were increased by the sleep disturbance score,31,32 satisfaction with life, ratio of the total number of questions in the scale to job satisfaction score,33 postural hypotension34 and the number of questions answered by the individ- prevalence of 27 symptoms. ual. Patients with at least 20% of items on a scale A self-completion quality of life questionnaire missing were excluded from analysis of that scale. was given to patients at up to five clinic visits: at The wellbeing scores were highly skewed towards screening; at entry into the double-blind trial; at the the ‘best’ value but the changes in scores were closer second and fourth visits during the titration period to the normal distribution. The change in quality of (weeks 6 and 12); and at the final evaluation visit life score attributable to the treatment was estimated (week 26 for those who completed, otherwise the from a model using change in score as the outcome visit at which they withdrew or a follow-up visit). and treatment as the exposure with the logarithm of To avoid any influence from the clinician, the proto- the baseline score as covariate. Logarithms were col prescribed that the patient complete the ques- used to overcome the problem of skewed baseline tionnaire in privacy before seeing the clinician and values. The model also included the square of log that the clinic nurse seal each questionnaire in an (baseline score) as this improved the fit of the model envelope in the presence of the patient. The ques- (using an F test). Where there was statistically sig- tionnaire was translated into the primary language nificant interaction between treatment and baseline of each centre and checked by back-translation. values, separate models were run for baseline values Patients not literate in the primary language of their above and below the point at which the response centre were excluded. curves for the two treatments crossed over. SAS version 6.08 for Windows 3.1 and Stata Ver- sion 5 were used for the analyses which were done Results without knowledge of the treatment codes. Primacy was given to analysis by intention-to-treat with a A total of 265 patients were randomised to eprosar- per-protocol analysis as a subsidiary evaluation. tan and 264 to enalapril. Six patients were excluded Study end point was defined as the time that the from all analyses owing to lack of baseline and/or patient left the trial. The last questionnaire com- end point questionnaires, leaving 261 patients ran- pleted by the patient was used in the analyses, even domised to eprosartan and 262 patients to enalapril. if done at a follow-up visit or before the end of treat- Screening questionnaires were used as baseline ment. The monotherapy end point was week 12 of questionnaires for five patients (three eprosartan and the treatment period, the visit at which HCTZ could two enalapril). Fifteen patients did not complete a

Journal of Human Hypertension Quality of life on eprosartan and enalapril E Breeze et al 859 Table 2 Baseline characteristics and quality of life of patients study end point as those on eprosartan (7.6% vs 3.2%) P ϭ 0.099. At monotherapy end point the dif- Characteristic Eprosartan Enalapril ferences were greater (9.9% vs 2.1%) and of clear (n = 261) (n = 262) ϭ n (%) n (%) statistical significance, P 0.001. Among those who fulfilled the per protocol criteria the prevalence and Language group incidence of cough were slightly less than for the English 118 (45.2) 115 (43.9) intention-to-treat analysis for study and monother- French 58 (22.2) 60 (22.9) apy end points. As for the intention-to-treat analy- Spanish 36 (13.8) 37 (14.1) a ses, the treatment differences in incidence by study Other 59 (22.6) 60 (22.9) end point were not statistically significant Race ϭ White 225 (84.9) 231 (87.5) (P 0.144) but were at monotherapy end point Black 21 (7.9) 19 (7.2) (P ϭ 0.002). Other 19 (7.2) 14 (5.3) The main purpose of this paper was to consider Cough status patients’ own perceptions. Clinical investigators’ Definite 5 (1.9) 2 (0.8) Probable/possible 0 (0.0) 3 (1.2) definitions of cough differed from self-assessed cough in two ways: first, the investigator had to PGWB scores (no. Media IQRb n Media IQRb n judge that the cough was not of bacterial or viral ori- items) n n gin and second, the duration of the cough was calcu- lated not just up to the time of visit but up to the Anxiety (5) 24 21–28 248 24 21–27 252 time it disappeared. Of 15 patients identified by Depression (3) 17 15–18 259 16 15–18 259 Positive wellbeing 18 15–19 249 17 15–20 259 either the investigator or the patient as having a (4) cough at baseline, only two were identified by both. Self-control (3) 16 14–17 256 16 14–17 260 At study end point 20 out of 34 reported coughs General health (3) 15 13–17 258 14 13–16 259 were in common The incidences of cough by study Vitality (4) 19 16–20 254 18.5 15–20 256 Total 108 96–117 259 106 94–115 261 end point according to the investigators’ question- naires were 2.0% definite cough, 5.1% probable/ aAfrikaans, Dutch, Italian, and Norwegian. possible cough for enalapril, 0.4% and 2.4% bIQR, interquartile range. Minimum and maximum possible respectively for eprosartan (␹2(2df) ϭ 5.449, scores were: 3.18 for depression, self-control and general health; P ϭ 0.06). 4,24 for positive wellbeing and vitality; 5,30 for anxiety; 22, 132 for total score. PGWB questionnaire at the end of treatment so an earlier one was used. There were 181 patients eligible for Three of the six subscales (anxiety, depression, and the per protocol analysis from the eprosartan group self-control) showed small negative mean changes (68.3%) and 174 from the enalapril group (65.9%). between baseline and study end point for eprosartan Thirty-five (13.2%) patients withdrew from the but otherwise all the point estimates for within- eprosartan arm and 47 (17.8%) from the enalapril treatment change were positive for both treatments arm (P ϭ 0.15). Of these, five were excluded from (ranging from absolute values of 0.06 to 0.72 for the analyses because of non-completion of question- subscales and being 0.04 and 2.52 for total PGWB naires. Only two patients on eprosartan and seven for eprosartan and enalapril, respectively; detail on enalapril were withdrawn specifically because of not shown). an adverse experience of cough. Observed differences between treatments in At baseline, the two treatment groups were alike changes in PGWB score suggested that eprosartan in demographic characteristics, key clinical and might have some advantage (Table 4). After adjusting for baseline values a clear effect remained behavioural indicators, and scores on the quality of ϭ life scales (Table 2). Males accounted for 57% of the for self-control (P 0.02) at study end point (Table 5). Small between-group differences for anxi- eprosartan patients and 56% of the enalapril ones. ϭ ϭ The average ages were 55.5 years and 55.9 years and ety (P 0.08) and general health (P 0.06) were not the mean DBPs were 100.7 mm Hg and 101.1 mm Hg statistically significant. Interaction tests showed that for the two groups respectively. Over 80% of both the treatment effect for study end point differed groups had previously used antihypertensives but according to the baseline values for positive well- slightly more of the eprosartan than enalapril being, vitality and total score. Of these there was a patients had previously used ACE inhibitors (60.0% statistically significant advantage for enalapril for against 53.1%) and previously had ACE coughs the total score if the baseline value was no greater (10.4% against 6.5%). Similar percentages smoked than 119 out of 132 (83% of all patients at baseline). (13.8% on eprosartan and 12.2% on enalapril). At monotherapy end point there were no significant differences between treatments (not shown). The per-protocol results (not shown) were consist- Cough ent with those for the intention-to-treat analysis: the Table 3 shows that those on enalapril were twice as treatment difference for change for self-control was likely to have gained a definite or possible cough by −0.35 (P ϭ 0.07) and for total PGWB if baseline score

Journal of Human Hypertension Quality of life on eprosartan and enalapril E Breeze et al 860 Table 3 Self-assessed cough: incidence between baseline and end points

Cough category Study end point Monotherapy end point

Eprosartan Enalapril Eprosartan Enalapril (n = 247) (n = 249) (n = 245) (n = 247)

n (%) n (%) n (%) n (%) Definite 5 (2.0) 12 (4.8) 4 (1.6) 15 (6.1) Probable/possible 3 (1.2) 7 (2.8) 1 (0.5) 9 (3.6) None of these 239 (96.8) 230 (92.4) 240 (98.0) 222 (89.9) ␹2 (2df) = 4.647, P = 0.099a ␹2 (2df) = 13.468, P = 0.001a

aTest of distribution between categories of definite cough, probable/possible cough and no cough.

was up to 119 the treatment effect was −2.67 Table 4 Between treatment differences in changes in PGWB (P ϭ 0.03). a scores (mean, 95% CI, P value) ; observed differences before The end point values of scores for life satisfaction, adjustment for baseline values sleep disturbance and job satisfaction were similar Scale (no. items) Study end point Monotherapy end in the two treatment groups (not shown). The symp- (n = 259, 260) point (n = 259, 261) toms reported by over 10% of patients were having to get up at night to pass urine, head pains or head- Anxiety (5) −0.82 −0.58 aches, and pain in joints of hands. The mean num- (−1.55, −0.09) (−1.21, 0.05) ber of symptoms reported was the same for both Depression (3) −0.27 −0.07 treatments, being 2.0 at baseline (s.d. 3.0) and 1.8 at (−0.64, 0.11) (−0.40, 0.26) Positive wellbeing (4) −0.16 0.24 study end point (s.d. 2.9). (−0.68, 0.35) (−0.25, 0.72) Self-control (3) −0.50 −0.09 (−0.89, −0.10) (−0.44, 0.27) Discussion General health (3) −0.42 −0.00 (−0.82, −0.02) (−0.41, 0.41) Incidence of dry cough (definite or Vitality (4) −0.23 −0.21 probable/possible) among the enalapril group was (−0.75, 0.30) (−0.73, 0.31) Total (22) −2.48 −0.79 four times that among the eprosartan group at mono- (−4.63, −0.32) (−2.72, 1.15) therapy end point and twice as great at study end point, the latter estimate not being statistical sig- aMaximum change scores possible were 15 for depression, self- nificance. The evidence for less cough on ACE-II control and general health; 20 for positive wellbeing and vitality; antagonists agrees with studies comparing other 25 for anxiety; 110 for total score. ACE-II antagonists with enalapril during a 12-week trial35,36 and with among people with a history of ACE-inhibitor related cough.18,37 In the present study there were fewer people with Table 5 Treatment effects in PGWB and subscale scores between baseline and study end point; results from regression analyses coughs than anticipated. The expected incidence of adjusted for baseline valuesa 2% with definite non-productive persistent cough for eprosartan and 9% for enalapril was close to the Scale n Treatment 95% P value combined incidence of definite, probable and poss- effect confidence ible coughs at monotherapy end point and higher intervals than any other estimates. Part of the explanation for few definite coughs may be that approximately 12% − − Anxiety 490 0.60 1.28, 0.07 0.077 of patients gave incomplete answers. Also, 16 Depression 508 −0.19 −0.52, 0.15 0.275 Positive wellbeing patients who reported a cough at one visit during i) baseline score р19 376 −0.42 −0.97, 0.12 0.128 the double-blind period did not report it at a sub- ii) baseline score Ͼ19 125 0.65 −0.29, 1.60 0.175 sequent one. ACE coughs could disappear spon- Self-control 511 −0.45 −0.81, −0.08 0.016 taneously during treatment38 or patients may cease General health 510 −0.34 −0.70, 0.14 0.060 Vitality to report a cough because they become tolerant to i) baseline score р20 313 −0.27 −0.94, 0.39 0.423 it. The estimates of ACE-induced cough vary widely ii) baseline score Ͼ20 193 0.16 −0.53, 0.85 0.646 between studies, as do the means of assessing cough Total and the populations covered. In double-blind ran- р − − − i) baseline score 119 428 2.32 4.54, 0.10 0.041 domly controlled trials the net increase over base- ii) baseline score р19 91 −0.99 −6.13, 4.14 0.702 line ranged between 13% and 25%,6 all higher than aIn the regression model change in score was the outcome, treat- found in this study. The criteria used to define ment the exposure, and log (baseline score) log2 (baseline score) cough in this study were unusually stringent. the covariates. Mean baseline values of total PGWB were compa-

Journal of Human Hypertension Quality of life on eprosartan and enalapril E Breeze et al 861 rable with those found in other studies of hyperten- which have compared ACE inhibitors with newer sives in Europe and the US19,20,29,39 and the subscale beta-blockers (eg atenolol) or calcium channel block- means tallied with those found at baseline in a simi- ers the between-drug effect sizes for the total score lar study.13 The favourable effects of enalapril com- were 0.04 or less.41 pared with eprosartan on self-control and total score In summary, those on enalapril were more likely were seen at study end point but not at monotherapy than those on eprosartan to develop a persistent dry end point. Thus, the main improvements in quality cough by the monotherapy end point, a disadvan- of life might come after some weeks on treatment. As tage which appeared to moderate by study end the percentages of patients using HCTZ were almost point. On the other hand, enalapril appeared to have identical the addition of a diuretic is unlikely to a slight advantage over eprosartan in measures of account for this result. self-control and total PGWB score (if the patient was In a study of similar design that compared cilazap- not already at the high end of the scale), particularly ril with atenolol and nifedipine, changes in PGWB for those who stayed in the trial beyond monother- total scale tended to be more positive by week 24 apy end point. The effect sizes of around 0.2 were than they had been at week 12.29 However, in a of the same order of magnitude as those found for study comparing enalapril with amlodopine most of overall PGWB between and propranolol the benefit manifested in PGWB and subscale scores and larger than those observed in other trials com- came in the titration period of 12 weeks.13 In the paring ACE inhibitors with a range of other antihy- few studies which report scores for PGWB subscales pertensive agents. It is concluded there are small but results have been mixed with ACE I inhibitors show- real differences in the quality of life of patients on ing some quality of life improvements but not the two treatments. always to a greater extent than the comparator 11,13,29,39 drugs. Enalapril has not previously been Acknowledgements reported as having an advantage in self-control. One other study reporting PGWB results for ACE II Smithkline Beecham Pharmaceuticals (since merged drugs21 showed significant within-treatment into GlaxoSmithkline) funded the trial. Dr Diana improvements for for four of the six PGWB Elbourne of LSHTM was helpful in discussing some subscales in a 12-week study of losartan and amlo- of the statistical issues. dopine. However, self-control was one of the subs- cales without any improvement. References In two studies comparing quality of life for losar- tan with nifedipine GITS19 and amlodopine20 1 Berkin KE, Ball SG. Cough and angiotensin converting respectively there were no significant between-treat- enzyme inhibition. BMJ 1988; 296: 1279. ment differences in PGWB total, sleep disturbance, 2 Inman WH et al. Postmarketing surveillance of enalap- sexual function, social functioning, cognitive func- ril. I: results of prescription-event monitoring. BMJ tioning and overall health perception. The smaller 1988; 297: 826–829. 3 Coulter DM, Edwards IR. Cough associated with capto- scale and shorter duration of these studies would pril and enalapril. BMJ 1987; 294: 1521–1523. not account for the null finding, as the observed dif- 4 Israili ZH, Hall WD. Cough and angioneurotic edema ferences were very small. The only difference found associated with angiotensin-converting enzyme inhibi- (in losartan’s favour) was in reported oedema. tor therapy. Ann Intern Med 1992; 117: 234–242. In the absence of a placebo group it is difficult to 5 Simon SR, Black HR, Moser M, Berland WE. Cough gauge the importance of the changes which can be and ACE inhibitors. Arch Intern Med 1992; 152: ascribed to individual treatments. Our interpretation 1698–1700. is based on relative differences between treatments 6 Fletcher AE, Palmer AJ, Bulpitt CJ. Cough with angio- in their effects on quality of life. Effect size ‘stan- tensin converting enzyme inhibitors: how much of a dardises’ treatment differences by dividing the dif- problem? J Hypertens 1994; 12 (Suppl 2): S43–S47. 7 Gandhi SK, Kong SX. Quality-of-life measures in the ference by standard deviation. An effect size of 0.3 evaluation of therapy: is conventionally considered as small, one of 0.5 as reliability, validity, and quality-of-life domains. Clin 40 moderate and one of at least 0.8 as substantial. The Ther 1996; 18: 1277–1295. effect size of self control and total PGWB were of 8 Williams GH. Assessing patient wellness: new per- the order of 0.2. The 95% confidence bounds spectives on quality of life and compliance. Am J Hy- include effect sizes around 0.3 favouring enalapril pertens 1998; 11: 186S–191S. at one extreme (0.36 self-control and for the total 9 Applegate WB et al. A randomised controlled trial of PGWB score) and negligible between-treatment the effects of three antihypertensive agents on blood effect sizes at the other extreme (0.02, 0.09 pressure control and quality of life in older women. respectively). Arch Intern Med 1991; 151: 1817–1823. 10 Herrick AL et al. Comparison of enalapril and atenolol Although small, the between-drug effect sizes for in mild to moderate hypertension. Am J Med 1989; 86: self-control and total PGWB are of the same order of 421–426. magnitude favouring captopril observed in a study 11 Testa MA, Anderson RB, Nackley JF, Hollenberg NK comparing total PGWB for captopril and propranolol and the Quality-of-Life Hypertension Study Group. (the subscales were not reported). In other trials Quality of life and antihypertensive therapy in men. A

Journal of Human Hypertension Quality of life on eprosartan and enalapril E Breeze et al 862 comparison of captopril and enalapril. N Engl J Med 26 Dupuy HJ. The Psychological Wellbeing Index. In: 1993; 328: 907–913. Wenger NK, Mattson ME, Furberg CD, Eliason J (eds). 12 Croog SH et al. Effects of antihypertensive medications Assessment of Quality of Life in Clinical Trials of Car- on quality of life in elderly hypertensive women. Am diovascular Therapies. Le Jacq: New York, 1984, J Hypertens 1994; 7: 329–339. pp 170–183. 13 Omvik P et al. Double-blind parallel comparative 27 Wiklund I, Halling K, La˚ngstro¨m G. The Psychological study on quality of life during treatment with amlodip- General Wellbeing Index (PGWB), a reliable tool for ine or enalapril in mild or moderate hypertensive use in cross-cultural multicentre clinical trials. patients: a multicentre study. J Hypertens 1993; 11: (Abstract) Qual Life Res 1995; 4: 503. 103–113. 28 Badia X, Gutie´rrez F, Wiklund I, Alonso J. Validity and 14 Townsend R, Haggert B, Liss C, Edelman JM. Efficacy reliability of the Spanish version of the Psychological and tolerability of losartan versus enalapril alone or in General Wellbeing Index. Qual Life Res 1996; 5: combination with hydrochlorothiazide in patients 101–108. with essential hypertension. Clin Ther 1995; 17: 29 Fletcher AE et al. Quality of life with three antihyper- 911–923. tensive treatments. , atenolol, nifedipine. 15 Ramsay LE, Yeo WW. ACE inhibitors, angiotensin II Hypertension 1992; 19: 499–507. antagonists and cough. The Losartan Study Group. J 30 Kazis LE, Anderson JJ, Meenan RF. Effect sizes for Hum Hypertens 1995; 9 (Suppl 5): S51–S54. interpreting changes in health statistics. Med Care 16 Lacourciere Y. The incidence of cough: a comparison 1989; 27 (Suppl): S178–S189. of lisinopril, placebo and , a novel angio- 31 Bulpitt CJ, Fletcher AE. Measurements of quality of tensin II antagonist. Telmisartan Cough Study Group. life in hypertension: a practical approach. Br J Clin Int J Clin Pract 1999; 53:99–103. Pharmacol 1990; 30: 353–364. 17 Paster RZ et al. Use of losartan in the treatment of 32 Crown S, Crisp AH. A short diagnostic self-rating scale hypertensive patients with a history of cough induced for psychoneurotic patients. The Middlesex Hospital by angiotensin-converting enzyme inhibitors. Clin Questionnaire (MHQ). Br J Psychiatr 1966; 112: 917– Ther 1998; 20: 978–989. 923. 18 Benz J et al. , a new angiotensin II receptor 33 House JS. Work Stress and Social Support. Addison- antagonist: a double-blind study comparing the inci- Wesley: Reading Mass, 1981. dence of cough with lisinopril and hydrochlorothia- 34 Bulpitt CJ, Dollery CT, Carne S. A symptom question- zide. J Clin Pharmacol 1997; 37: 101–107. naire for hypertensive patients. J Chron Dis 1974; 27: 19 Weir MR et al. Efficacy, tolerability and effects on 309–323. 35 Tikkanen I, Omvik P, Jensen HA for the Scandinavian quality of life of losartan, alone or with hydrochloro- Study Group. Comparison of the angiotensin II antag- thiazide, versus nifedipine GITS in patients with onist losartan with the angiotensin converting enzyme essential hypertension. Clin Ther 1996; 18: 411–428. inhibitor enalapril in patients with essential hyperten- 20 Oparil S et al. Efficacy, tolerability, and effects on sion. J Hypertens 1995; 13: 1343–1351. quality of life of losartan, alone or with hydrochloro- 36 Mimran A et al. A randomised, double-blind compari- thiazide, versus amlodopine, alone or with hydro- son of the angiotensin II receptor anatagonist irbesar- chlorothiazide, in patients with essential hyperten- tan, with the full dose range of enalapril for the treat- sion. Clin Ther 1996; 18: 608–625. ment of mild-to-moderate hypertension. JHum 21 Dahlo¨fBet al for the LOA Study group. Main results Hypertens 1998; 12: 203–208. of the losartan versus amlodopine (LOA) study on drug 37 Ramsay LE, Yeo WW on behalf of the Losartan Cough tolerability and psychological general well-being. J Study Group. Double-blind comparison of losartan, Hypertens 1997; 15: 1327–1335. lisinopril and hydrochlorothiazide in hypertensive 22 McClellan KJ, Balfour JA. Eprosartan. Drugs 1998; 55: patients with a previous angiotensin converting 713–718. Discussion pp 719–720. enzyme inhibitor-associated cough. J Hypertens 1995; 23 Elliott W for the Eprosartan Study Group. Double- 13 (Suppl 1): S73–S76. blind comparison of eprosartan and enalapril on cough 38 Reisin L, Schneeweiss A. Spontaneous disappearance and blood pressure in unselected hypertensive of cough induced by angiotensin-converting enzyme patients. J Hum Hypertens 1999; 13: 413–417. inhibitors (captopril or enalapril). Am J Cardiol 1992; 24 Argenziano L, Trimarco B. Effect of eprosartan and 70: 398–399. enalapril in the treatment of elderly hypertensive 39 Croog SH et al. The effects of antihypertensive therapy patients: subgroup analysis of a 26-week, double- on the quality of life. N Engl J Med 1986; 314: 1657– blind, multicentre study. Eprosartan Multinational 1664. Study Group. Curr Med Res 1999; 15:9–14. 40 Machin D, Campbell MJ. Statistical Tables for the 25 Levine B. Effect of eprosartan and enalapril in the Design of Clinical Trials. Blackwell Scientific Publi- treatment of black hypertensive patients: subgroup cations: Oxford, 1987. analysis of a 26-week double-blind multicentre study. 41 Fletcher AE. Measuring the effects of hypertensive Eprosartan Multinational Study Group. Curr Med Res therapy upon the quality of life. Curr Opin Nephrol 1999; 15:25–32. Hypertens 1995; 4: 538–542.

Journal of Human Hypertension