Lisinopril Versus Hydrochlorothiazide in Obese Hypertensive Patients
A Multicenter Placebo-Controlled Trial
Abstract Because obesity-associated hypertension has unique hemodynamic and hormonal profiles, certain classes of antihypertensive agents may be more effective than others as monotherapy. Thus, we compared the efficacy and safety of the angiotensin-converting enzyme inhibitor lisinopril and the diuretic hydrochlorothiazide in a 12-week, multicenter, double-blind trial in 232 obese patients with hypertension. Patients with an office diastolic pressure between 90 and 109 mm Hg were randomized to treatment with daily doses of lisinopril (10, 20, or 40 mg), hydrochlorothiazide (12.5, 25, or 50 mg), or placebo. Mean body mass indexes were similar for all patients. At week 12, lisinopril and hydrochlorothiazide effectively lowered office diastolic (−8.3 and −7.7 versus −3.3 mm Hg, respectively; P<.005) and systolic (−9.2 and −10.0 versus −4.6 mm Hg, respectively; P<.05) pressures compared with placebo. Ambulatory blood pressure monitoring confirmed that lisinopril and hydrochlorothiazide effectively lowered 24-hour blood pressure compared with placebo (P<.001). Significant dose-response differences were observed between treatments. Sixty percent of patients treated with lisinopril had an office diastolic pressure <90 mm Hg compared with 43% of patients treated with hydrochlorothiazide (P<.05). Responses to therapies differed with both race and age. Neither treatment significantly affected insulin or lipid profiles; however, plasma glucose increased significantly after 12 weeks of hydrochlorothiazide therapy compared with lisinopril (+0.31 versus −0.21 mmol/L; P<.001). Hydrochlorothiazide also decreased serum potassium levels by 0.4 mmol/L from baseline. In conclusion, lisinopril was as effective as hydrochlorothiazide in treating obese patients with hypertension. Treatment with angiotensin-converting enzyme inhibitors may show greater efficacy as monotherapy at lower doses compared with thiazide diuretics, may have a more rapid rate of response, and may offer advantages in patients at high risk of metabolic disorders.
Longitudinal studies have shown an increased prevalence of hypertension as patient weight and age increase.1 In fact, up to 30% of hypertension cases can be attributed to obesity,2 and numerous studies have demonstrated a dramatic drop in BP with weight loss.3 4 5 6 7 8 Unfortunately, long-term control of BP by weight reduction has been limited because compliance with weight-reduction programs is poor, with dropout rates ranging from 50% to 70% at 1 to 2 years.9 10 Consequently, pharmacological treatment of hypertension in obese patients is indicated when nonpharmacological approaches fail or when patients are diagnosed initially with moderate or severe hypertension.11 The class of antihypertensive medication best suited for first-line treatment of hypertension in obesity has not been determined because of the lack of large, placebo-controlled studies comparing different classes of antihypertensive agents in this population.
We report findings from the first multicenter, placebo-controlled study comparing monotherapeutic treatment effects of the ACE inhibitor lisinopril with the diuretic HCTZ in obese patients with hypertension (TROPHY). The TROPHY study is also the first trial to use ABP monitoring to evaluate 24-hour BP control in obese patients.
Obese men and women (n=334) recruited for this study were between 21 and 75 years of age and had stage I or II hypertension, as defined by the DBP criteria of the fifth report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure (JNC-V).11 Patients were enrolled in the study if their BMI was within 28 to 40 kg/m2 for men and 27 to 40 kg/m2 for women. Patients were eligible for randomized treatment with lisinopril, HCTZ, or placebo if the average of three seated office DBP measurements was ≥90 mm Hg and ≤109 mm Hg at two consecutive visits during the placebo lead-in period and at the final visit before randomization. A total of 126 patients completed successful baseline ABP monitoring before randomization.
Written informed consent was obtained from all patients before they entered into the study. The protocol was approved by the Institutional Review Board at each center.
Patients were excluded from the study for the following reasons: secondary hypertension; target-organ damage (renal failure, congestive heart failure, myocardial infarction, or cerebrovascular accident 6 months preceding the study); presence of second- or third-degree heart block, valvular heart disease, hepatic disease, hypothyroidism, or diabetes mellitus treated with insulin or oral hypoglycemics; arm circumference greater than 40 cm; weight loss of more than 10 kg or participation in more than two weight-loss programs within 6 months of the present study; and hypersensitivity or contraindication to ACE inhibitors, HCTZ, or other sulfonamide-derived drugs.
Patients were required to maintain their normal eating habits for the duration of the study and not to smoke or consume alcohol in the morning before office BP measurements.
Lisinopril, HCTZ, and placebo were compared in a multicenter, randomized, double-blind, double-dummy, parallel-group trial. The trial consisted of a 2- to 4-week single-blind placebo lead-in period and a 12-week double-blind treatment phase with once-daily doses of lisinopril (10, 20, or 40 mg), HCTZ (12.5, 25, or 50 mg), or placebo.
After randomization to treatment, patients returned for office visits at 4-week intervals. All patients started at the lowest dose of study medication, and response to treatment was determined from office BP measurements and defined as a DBP <90 mm Hg with a minimal decrease of ≥10 mm Hg from baseline (prerandomization measurement). Patients who did not meet response criteria during an office visit were titrated to the next dose. When the patient met response criteria, he or she remained at that dose for the duration of the study.
Office BP measurements were taken 24 hours (±1 hour) after the previous morning dose of study medication with the use of a standard mercury sphygmomanometer and cuff size adapted to the patient’s arm circumference. The mean of three seated, resting measurements was recorded. One standing BP measurement was also recorded to monitor patients for orthostatic hypertension. Pulse rates were measured coincident with BP.
Twenty-four-hour ABPs and heart rates were monitored twice during the study with a SpaceLabs monitor (model 90207). ABP was monitored 24 hours before randomization to double-blind treatment and 24 hours before the last day of treatment. BPs and heart rates were recorded every 20 minutes from 5 am to 11 pm and every 30 minutes thereafter until 5 am. Successful monitoring meant that at least 80% of readings were valid; ie, there were no more than 3 daytime hours with only one valid reading and no more than 1 nighttime hour with zero valid readings.
Clinical and Laboratory Assessments
Baseline assessments included a physical examination, height and weight measurements, chest radiograph, and electrocardiogram. Patients were weighed at each visit after randomization to treatment. Baseline laboratory assessments included serum chemistry, hemogram, urinalysis, plasma glucose, insulin, thyroid-stimulating hormone, triglycerides, total cholesterol, and high-density and low-density lipoprotein cholesterols. All laboratory assessments, except thyroid-stimulating hormone, were repeated on the last day of double-blind treatment.
The principal statistical analysis was a between-treatment comparison to detect statistically significant differences in changes from baseline to week 12 among the three treatment groups. Comparisons of interest were active treatments versus placebo and lisinopril versus HCTZ.
Sample size (n=80 per treatment group) calculations were based on an 80% power to detect a difference of 5 mm Hg between treatments (standard deviation=11.0) in office DBP change from baseline after 12 weeks of therapy (significance level of .05 for a two-sided T test).
Efficacy analyses included all patients who had baseline data and completed all scheduled visits. Initially, treatments were compared with respect to demographic and office DBP measurements and heart rate. ANCOVA, with the baseline as the covariate, was used to compare office SBP and DBP changes from baseline between treatment groups and to test for center and treatment-by-center interactions. The proportion of responders (based on office DBP criteria) was analyzed with Fisher’s exact test.
For 24-hour ABP, ANCOVA was used to determine treatment differences in average change from baseline (hour 0) in systolic and diastolic ABP measurements across 24 hours as well as during daytime (hours 1 to 12), nighttime (hours 13 to 24), and trough (hours 22 to 24), including tests for treatment, center, and treatment-by-center interaction. Area under the curve was computed for each patient using a trapezoid rule after reducing the number of data points per patient into hourly averages. Mean reduction in the area under the curve from baseline was then compared between treatments.
Mean values for standard laboratory test data were compared for lisinopril, HCTZ, and placebo using descriptive statistics.
Of 334 patients who entered the study, 232 were randomized to treatment with lisinopril (n=77), HCTZ (n=76), and placebo (n=79). As seen in Table 1⇓, patients were similarly divided among treatment groups with respect to sex, race, age, height, weight, and BMI. Mean patient weights were similar for all treatment groups.
A total of 21 patients were withdrawn from the study for the following reasons: adverse events (n=10), refused to continue (n=10), and informed consent withdrawn (n=1). More patients were withdrawn in the placebo group (n=11) than in the lisinopril (n=7) or HCTZ (n=3) groups. In the lisinopril group, 1 patient was withdrawn because of chest pain and increased BP over the limits allowed by the protocol. Patients were withdrawn from HCTZ treatment because of headache (n=3) and dry cough (n=1). Patients in the placebo group were withdrawn because of headache (n=6), dry cough (n=1), a hypertensive crisis (n=1), and non–insulin-dependent diabetes mellitus (n=1) that required pharmacological treatment.
Office BP: Efficacy Analysis
Distribution of Patients and Doses
A total of 211 patients completed the final study visit and were included in efficacy analyses. Baseline office SBP and DBP were similar among treatment groups (lisinopril, 147±15/98±6 mm Hg; HCTZ, 148±14/98±5; placebo, 146±13/96±4). At week 12, both lisinopril and HCTZ significantly reduced DBP from baseline compared with placebo (−8.3 and −7.7 versus −3.3 mm Hg, respectively; P<.005); a significant reduction was also observed in SBP after treatment with lisinopril and HCTZ compared with placebo (−9.2 and −10.0 versus −4.6 mm Hg, respectively; P<.05). No significant difference was observed between the antihypertensive efficacy of lisinopril and HCTZ at week 12; however, lisinopril significantly (P≤.05) decreased DBP at weeks 4 and 8 compared with HCTZ.
Heart rate was not significantly affected by active treatments or placebo. Baseline and final heart rates for the three treatment groups were as follows: lisinopril, 71±7 and 71±9 beats per minute; HCTZ, 74±9 and 73±9; and placebo, 72±9 and 75±8, respectively.
In the HCTZ group, the majority of patients required the highest dose of study medication (50 mg), whereas a large proportion of patients in the lisinopril group were on the initial 10-mg dose. Overall, 40% of patients responded to lisinopril and 33% responded to HCTZ (P>.05). Of patients who responded to lisinopril treatment (n=28), 57% responded to dose 1 (10 mg/d), 18% responded to dose 2 (20 mg/d), and 25% responded to dose 3 (40 mg/d). Of patients who responded to HCTZ treatment (n=24), only 29% responded to dose 1 (12.5 mg/d), 25% responded to dose 2 (25 mg/d), and 46% responded to dose 3 (50 mg/d). At the end of treatment, 60% of patients treated with lisinopril had a DBP <90 mm Hg compared with 43% of patients treated with HCTZ (P<.05).
Response According to Race
Response to treatment varied with race. Black patients responded better to HCTZ (52%) than to lisinopril (33%). The mean reduction in office DBP was −10.8 mm Hg from baseline to week 12 for black patients in the HCTZ group; this decrease was significant across treatments (P<.01) and compared with placebo (−1.3 mm Hg, P<.05) but was not statistically different from lisinopril (−7.0 mm Hg). SBP reduction in black patients after 12 weeks of HCTZ therapy was significant across treatments (P<.02) and compared with lisinopril and placebo (−13.7 versus −4.7 and −4.7 mm Hg, respectively; P<.05).
Conversely, white patients responded better to lisinopril (43%) than to HCTZ (20%). The mean reduction in office DBP was −9.4 mm Hg from baseline to week 12 for white patients in the lisinopril group; this reduction was significant across treatments (P<.001) and compared with HCTZ and placebo (−5.2 and −1.9 mm Hg, respectively; P<.05). SBP reduction in white patients after 12 weeks of lisinopril therapy (−11.4 mm Hg) was significant across treatments (P<.001) and compared with placebo (−1.4 mm Hg; P<.05) but was not statistically different from HCTZ (−6.4 mm Hg).
Response According to Age, Sex, and Weight
Response to lisinopril was better in young patients (ages 20 to 39 years) of both races than to HCTZ (46% versus 15%, respectively). Patients 40 years of age or older responded similarly to treatment with lisinopril and HCTZ. Women and men responded similarly to lisinopril treatment (38% and 42%, respectively), although women responded better to HCTZ than did men (41% versus 26%, respectively). Patient response to therapy did not appear to be related to baseline BMI. No statistically significant change in weight was observed for any treatment group (Table 3⇓). Furthermore, no correlation existed between changes in weight or BMI and changes in DBP with treatment.
Ambulatory BP: Efficacy Analysis
Table 2⇓ summarizes the results of 24-hour ABP monitoring. Mean baseline 24-hour SBP and DBP were similar among the three treatment groups. At week 12, both lisinopril and HCTZ effectively lowered BP throughout the 24-hour monitoring period compared with placebo (P<.001); however, no significant difference was observed between active treatments (Figs 1⇓ and 2⇓). Lisinopril and HCTZ showed similar efficacy regardless of whether ABP was elevated during daytime hours (hours 1 to 12) or at night (hours 13 to 24). Although lisinopril effectively lowered BP through hours 22 to 24 compared with placebo, HCTZ was slightly more effective during these hours.
Clinical Laboratory Assessments
Table 3⇓ details results of several laboratory assessments. Significant (P<.001) treatment differences were found between plasma glucose levels of patients in the lisinopril and HCTZ groups. At week 12, plasma glucose levels decreased by 0.21 mmol/L from baseline in the lisinopril group, while increasing by 0.31 mmol/L in the HCTZ group. The overall change from baseline in insulin levels and the glucose-insulin ratio was not statistically significant among treatments. An analysis of several lipid subfractions revealed no statistically significant changes between treatments. At week 12, a significant decrease from baseline was observed in potassium levels of patients receiving HCTZ (4.3 to 3.9 mmol/L; P<.0001).
Significant associations (P<.05) were found, at both baseline and week 12, between increased patient weights and higher glucose and insulin levels and glucose-insulin ratio. However, no significant association was observed during the study between changes in weight or BMI and changes in glucose, insulin, or the glucose-insulin ratio.
Headache and hypokalemia (potassium <3.5 mmol/L) were the only treatment-related adverse experiences that occurred in 3% or more of patients in any treatment group. Headache was reported in 5.1% of patients in the placebo group; hypokalemia was reported in 3.9% of patients in the HCTZ group and 1.3% of patients in the placebo group. The incidence of cough was low and reported with similar frequency among treatment groups (lisinopril, 2.6%; HCTZ, 1.3%; placebo, 2.5%).
Findings from this multicenter, placebo-controlled study in obese patients with hypertension showed similar reductions in office SBP and DBP after 12 weeks of monotherapy with lisinopril or HCTZ. Most patients who responded to lisinopril therapy (57%) were controlled with the initial 10-mg dose, whereas a large percentage of patients (46%) who responded to HCTZ required the highest dose (50 mg) of study medication to maintain BP control. Additionally, this study showed that obese white and young patients responded better to ACE inhibitor therapy, whereas obese black patients responded better to diuretic therapy.
We did not find a correlation between changes in weight or BMI and changes in DBP. Thus, we concluded that the antihypertensive effects were correlated solely with the therapy received.
Another unique feature of this study was that it examined ABP monitoring in obese hypertensive patients. ABP tracings showed that at week 12, both antihypertensive agents effectively lowered systolic and diastolic ABPs throughout a 24-hour period.
Analysis of clinical laboratory data revealed treatment differences between lisinopril and HCTZ. Twelve weeks of lisinopril therapy reduced plasma glucose levels, whereas HCTZ therapy increased plasma glucose. HCTZ also significantly reduced serum potassium levels after 12 weeks of therapy.
Earlier studies described some pathophysiological abnormalities associated with hypertension in obese patients, such as renin dependency5 and salt sensitivity.12 13 Sodium retention, a characteristic of salt sensitivity, increases extracellular fluid volume, elevates cardiopulmonary volume, and increases cardiac output in obesity hypertension.14 Salt sensitivity is associated with an increase in glomerular filtration fraction and proteinuria in patients with obesity hypertension.12 13 These renal hemodynamic changes may not only fuel the hypertensive process but may also lead to progressive renal injury. Insulin resistance may also be associated with similar renal hemodynamic changes.15
We used an ACE inhibitor (lisinopril) in this study because its primary mechanism of action is selective control of BP through blockade of the renin-angiotensin-aldosterone system.16 17 This action can occur at both systemic and tissue autocrine-paracrine levels18 and may, through dilating the efferent glomerular arterioles, restore the ability of the kidney to excrete salt and water as well as control glomerular hyperfiltration.19
Lisinopril effectively controlled BP, with 40% of patients meeting strict response criteria (ie, DBP <90 mm Hg and a decrease in DBP of ≥10 mm Hg from baseline). Moreover, the rate of response to lisinopril was more rapid than to HCTZ. These findings support the importance of targeting the renin-angiotensin system and salt sensitivity when treating obese patients for hypertension.
The significant reduction in plasma glucose levels after 12 weeks of lisinopril therapy may relate to the described effects of ACE inhibition on insulin sensitivity17 ; however, no relationship was observed in our study between reductions in DBP and glucose or insulin levels or the glucose-insulin ratio. Hence, the effect of lisinopril on glucose levels in this patient population may or may not be attributed to improved insulin sensitivity.
We used HCTZ in this study because it controls hypertension by causing diuresis and natriuresis,20 which first reduces extracellular fluid volume, thereby decreasing cardiac output.21 Extracellular fluid volume is reduced 3 to 4 days after an initial dose of HCTZ, and subsequently a new equilibrium is established at a lower volume and BP is reduced.20 HCTZ may also decrease peripheral resistance.21 Consequently, diuretics, like ACE inhibitors, may be a good pharmacological approach when treating obese patients for hypertension, and may improve the hemodynamic pattern—increased intravascular volume and cardiac output—that is a hallmark of obese hypertensive patients.19
Despite the high dosages of HCTZ (50 mg/d) used in our study, the lipid profile of our patients was not affected as in previous studies.22 We did, however, show a significant increase in plasma glucose and a decrease in serum potassium levels in the HCTZ group when compared with the lisinopril or placebo groups.
We conclude that a rational monotherapy approach for initial treatment of obese patients with hypertension may include an ACE inhibitor or diuretic. The choice of agent may be determined by the patient’s race and age. Lisinopril was more effective than HCTZ in white and young patients, whereas HCTZ was more effective in black patients. However, further research is needed with larger numbers of patients to confirm our findings. Patients generally responded to low doses of lisinopril and high doses of HCTZ. As previously demonstrated,23 ACE inhibitors may also offer a treatment advantage when comparing certain metabolic profiles.
Selected Abbreviations and Acronyms
|ABP||=||ambulatory blood pressure|
|BMI||=||body mass index|
|DBP||=||diastolic blood pressure|
|SBP||=||systolic blood pressure|
This study was supported by a grant from Zeneca Pharmaceuticals. The authors thank the following people for participating in the TROPHY Study Group: Dr Kenneth V. Adams, Dr M. Eileen Cook, Dr Gary Enzmann, Dr Barry Gould, Dr Ronald J. Graf, Dr Edward Meyer, Dr David T. Nash, Dr Dennis A. Ruff, Dr Ziad Zawaideh, Dr Dale Murphy, Dr H. Morgan Ashurst, Dr Robert Finklehor, Dr F. Gilbert McMahon, and Stacy Wilson. The authors also thank Mary Jo Psomas for editing the manuscript.
- Received October 7, 1996.
- Revision received October 31, 1996.
- Accepted December 26, 1996.
Reisin E, Abel R, Modan M, Silverberg DS, Eliahou HE, Modan B. Effect of weight loss without salt restriction on the reduction of blood pressure in overweight hypertensive patients. N Engl J Med. 1978;298:1-6.
Langford HG, Davis BR, Oberman A, Wassertheil-Smoller S, Hawkins CM, Zimbaldi N for the TAIM Study Group. Effect of drug and diet treatment of mild hypertension on diastolic blood pressure. Hypertension. 1991;17:210-217.
Stevens VJ, Corrigan SA, Obarzaneck E, Bernaver E, Cook NR, Hebert P, Mattfeldt-Beman M, Oberman A, Sugars C, Dalcin AT, Whelton PK, for the TOHP Collaborative Research Group. Weight loss intervention in phase I of the trials of hypertension prevention. Arch Intern Med. 1993;153:849-858.
NIH Technology Assessment Conference Panel. Methods of voluntary weight loss and control. Ann Intern Med. 1992;116:942-949.
Tuck ML. Role of salt in the control of blood pressure in obesity and diabetes mellitus. Hypertension. 1991;17(suppl I):I-135-I-142.
Weir MR, Dengel DR, Behrens MT, Goldberg AP. Salt induces increases in systolic blood pressure effect renal hemodynamics and proteinuria. Hypertension. 1995;25:1339-1344.
Dengel DR, Goldberg AP, Mayuga RS, Kairis GM, Weir MR. Insulin resistance, elevated glomerular filtration fraction, and renal injury. Hypertension. 1996;28:127-132.
Frohlich ED. Obesity hypertension: converting enzyme inhibitors and calcium antagonists. Hypertension. 1992;19(suppl I):I-119-I-123.
Sanchez RA, Marco E, Gilbert HB, Raffaele GP, Brito M, Gimenez M, Moledo LI. Natriuretic effects and changes in renal hemodynamics induced by enalapril in essential hypertension. Drugs. 1985;30:149-158.
Wilson IM, Freis ED. Relationship between plasma and extracellular fluid volume depletion and the antihypertensive effect of chlorothiazide. Circulation. 1959;20:1028-1036.
Frohlich ED, Schnaper HW, Wilson IM, Freis ED. Hemodynamic alterations in hypertensive patients due to chlorothiazide. N Engl J Med. 1960;262:1261-1263.