beta-blocker: meta-analysis of 7 clinical trials

AJH–May 2003–VOL. 16, NO. 5, PART 2

OR-20 DETERMINANTS OF BLOOD PRESSURE RESPONSE TO ACE INHIBITOR MONOTHERAPY IN HYPERTENSIVE AFRICAN AMERICANS AND CAUCASIANS Evan Mokwe, Suzanne E Ohmit, Amanda I Dudley, John M Flack. Internal Medicine, Wayne State University, Detroit, MI. Objectives: Data collected during a clinical trial evaluating rapidity of medication up-titration with blood pressure (BP) response were used to investigate multiple factors, including race, as predictors of blood pressure (BP) response to monotherapy with the ACE inhibitor, quinapril. Methods: Participants with JNC-VI stages 1 or 2 hypertension (SBP 140-169 or DBP 90-104) were randomly assigned to fast (every 2 weeks; duration 6 weeks) or slow (every 6 weeks; duration 18 weeks) drug titration. Participants received 20 mg of quinapril once daily at randomization; at intervention-specific (fast or slow) follow-up visits, medication dose was doubled (up to 80 mg/day) if BP remained ⱖ 140/90 mm Hg. Means and interquartile ranges (IQR, middle 50% of the SBP response distribution) were determined and compared by race. Factors associated with SBP responses in the lower 25% or the upper 25% of the response distributions were determined in multivariate logistic regression models. Results: Data were available for 533 African Americans and 2,046 Caucasians; 53% of participants were women. Participant mean age was 51 years, and 48% were obese (BMI ⱖ30). Mean SBP responses were -10.5 ⫾ 13.4 and -15.3 ⫾12.2 mm Hg, respectively, in African Americans and Caucasians. IQR boundaries were -2.2 to -20 and -7.3 to -23.5 mm Hg in African Americans and Caucasians, respectively; these IQR were 3.7 and 3.4 times, respectively, larger than the racial difference in SBP response (4.8 mm Hg). Factors associated with BP response in the lower 25% of the SBP response distribution included fast treatment group (p⫽.06), early study visits (time), older age, African American race, obesity, higher medication dose, and lower baseline SBP (all p⬍.01). Factors associated with BP response in the upper 25% of the SBP response distribution included slow treatment group, later study visits (time), younger age, Caucasian race, lower medication dose and higher baseline SBP (all p⬍.01). Conclusions: Racial difference in SBP response to treatment was documented even though the middle 50% of the SBP response distribution in both groups largely overlapped. Within-race IQRs of SBP response were several-fold greater than the mean difference in SBP response between the races. Race, along with other factors that vary at the level of the individual, predict extreme SBP responses to monotherapy with the ACE inhibitor quinapril. Key Words: Blood Pressure Response, Ace Inhibitor, African American Race

OR-21 RELATION OF SERUM URIC ACID TO CARDIOVASCULAR ENDPOINTS IN HYPERTENSION: THE LIFE STUDY

ORALS: JNC VII

9A

contrast to atenolol and to other AII antagonists. The LIFE study was a double-masked, randomized, parallel-group trial in 9193 patients (54% female) with essential hypertension and left ventricular hypertrophy. The participants received once-daily losartan- or atenolol-based treatment. We used Cox regression analysis to compare regimens. Baseline SUA was significantly associated with increased cardiovascular risk (hazard ratio [HR] ⫽ 1.024 [95% C.I. 1.017-1.032] per 10 ␮mol/L, p⬍0.0001) with evidence of a significant gender interaction. Baseline SUA remained significantly associated with increased risk in women (HR⫽1.025 [1.013-1.037], p⬍0.0001) but not in men (HR⫽1.009 [0.998-1.019], p⫽0.108). After adjustment for Framingham risk score (FRS), SUA was no longer significant in the entire study population (HR⫽1.006 [0.998-1.014], p⫽0.122) or in men (HR⫽1.006 [0.995-1.017], p⫽0.291) but independently predicted composite endpoints in women (HR⫽1.013 [1,1.025], p⫽0.0457). The baseline-to-end-of-study increase in SUA (standard deviation, SD) differed between the treatment groups, 44.4 (72.5) ␮mol/L atenolol and 17.0 (69.8) ␮mol/L losartan, p⬍0.0001. The contribution of SUA to the treatment effect of losartan in terms of the primary composite endpoint was 29% [14, 107], p⫽0.004. Although the SUA-by-gender interaction was not significant (p⫽0.079), the association between time-varying SUA and increased cardiovascular risk tended to be stronger in women (p⬍0.0001) than in men (p⫽0.0695). In conclusion, treatment with losartan significantly attenuated the increase in SUA compared with treatment with atenolol in the LIFE study. This difference seemed to explain 29% of the treatment effect on the primary composite endpoint. Correlation between SUA and FRS, especially in men, complicates the interpretation. Further investigations are warranted. Key Words: serum uric acid, losartan, risk factors

OR-22 CARDIOVASCULAR EVENTS DURING INITIAL ANTIHYPERTENSIVE THERAPY WITH EITHER A CALCIUM ANTAGONIST OR A DIURETIC/BETA-BLOCKER: META-ANALYSIS OF 7 CLINICAL TRIALS William J. Elliott. Preventive Medicine, RUSH Medical College, Chicago, IL. Despite the proven benefits of a low-dose diuretic or beta-blocker as initial antihypertensive therapy, recent prescription monitoring data show that a calcium antagonist remains very popular in this role. Several large clinical trials directly comparing representatives of these drug classes have released their results only recently, and these have not as yet been aggregated. We therefore gathered data from 7 trials (MIDAS, INSIGHT, NORDIL, NICS-EH, STOP-Hypertension 2, VHAS, CONVINCE) that directly compared a calcium antagonist against either a diuretic or betablocker as initial therapy across several cardiovascular (CV) endpoints, including myocardial infarction (MI) and congestive heart failure (CHF). Data from each trial report were combined using the Mantel-Haentzel method to give a relative risk and a 95% confidence interval (CI). Final data from ALLHAT were embargoed at the time of this submission, but will be included in May. The number of patients with events/number of patients in each group, and the results of the meta-analysis were:

Aud Høieggen, Sverre E Kjeldsen, Stevo Julius, Richard B Devereux, Michael Alderman, Cong Chen, Bjorn Dahlof. Div. Nephrology, Ullevaal University Hospital, Oslo, Norway; Department of Cardiology, Ullevaal University Hospital, Oslo, Norway; Department of Internal Medicine, Division of Hypertension, University of Michigan, Ann Arbor, MI; Division of Cardiology, Cornell Medical Center, New York City, NY; Department of Epidemiology and Social Medicine, Albert Einstein College of Medicine, New York City, NY; Merck Research Laboratories, Merck & Co., Blue Bell, PA; Department of Medicine, Sahlgrenska Hospital/Ostra, Gothenburg, Sweden.

Death CV Death Stroke MI Major CV Events CHF

Serum uric acid (SUA) has been proposed to be an independent risk factor for cardiovascular morbidity and death. Losartan is uricosuric, in

These data indicate that, for all cardiovascular events except CHF, the relative risk for an initial calcium antagonist was not significantly dif-

© 2003 by the American Journal of Hypertension, Ltd. Published by Elsevier Inc.

Diuretic/ Beta-Blocker

Calcium Antagonist

Relative Risk (95% CI)

1 103/20 306 538/20 306 640/20 306 554/20 306 1 429/20 306 346/20 306

1 121/20 507 565/20 507 583/20 507 588/20 507 1427/20 507 405/20 507

1.03 (0.94–1.12) 1.06 (0.94–1.20) 0.92 (0.81–1.03) 1.07 (0.95–1.21) 1.01 (0.93–1.09) 1.19 (1.02–1.38)

0895-7061/03/$30.00

10A

ORALS: JNC VII

AJH–May 2003–VOL. 16, NO. 5, PART 2

ferent than for an initial diuretic or beta-blocker. In view of the similar (and not better) efficacy for a typically more expensive calcium antagonist, compared to a generally less expensive diuretic or beta-blocker in preventing most cardiovascular events, it is likely that the current large disparity in acquisition costs would support a public policy recommendation for an initial diuretic (or possibly a beta-blocker) as the more cost-effective initial choice.

Since the absolute number of new diuretic prescriptions in the US continues to rise annually, as does the population of incident ESRD patients, we propose that a prospective RCT trial would be justified to conclusively show the nephrotoxic causality of prolonged diuretic exposure within the population.

Key Words: Meta-analysis, Calcium channel blocker controversy, Guidelines for initial therapy

OR-24 ANTIHYPERTENSIVE CALCIUM CHANNEL BLOCKERS MAY PROMOTE HUMAN CANCER CELL GROWTH

OR-23 POPULATION-WIDE DIURETIC USE IS DIRECTLY ASSOCIATED WITH THE INCIDENCE OF ENDSTAGE RENAL FAILURE IN THE UNITED STATES Ralph G. Hawkins, Mark C. Houston. Hypertension Institute, St. Thomas Hospital, Nashville, TN. Treatment of hypertension over the past 3 decades using predominantly diuretic and beta-blocker drug regimens has been promoted as a factor in the reduction of cardiovascular disease related (CVD) and stroke-related mortality. Concurrent long-term measurement of end-stage renal disease (ESRD) morbidity in the US shows an increasing incident population for both hypertensive nephrosclerosis (HT-NS) and all-cause ESRD. Utilizing the technique of data fusion, where concurrent observations of the same total population using different measurement tools are merged for analysis, we have evaluated available national databases for CVD mortality, stroke mortality, HT-NS incidence and ESRD incidence and have compared those data with a commercial database (IMS Health©) reflecting antihypertensive medication supply to US distributors for the period 1990 to 2001. The annual percentage change in manufacturer supply for diuretics was accepted to reflect the annual percent change in drug consumption by the US population at large. This premise permits analysis of annual variations in population-wide disease behavior concurrent with annual changes in population-wide drug category use. Annual changes in prescription diuretic supply were observed to have the same pattern of fluctuations as the annual variability in ESRD incidence time-lagged two years later. Analysis of the relationship shows a significant (p⫽0.0001) direct curvilinear relationship (r⫽0.99795) existing between these variables over time. This relationship allows for the development of a simple linear model that relates annual percentage changes in diuretic supply and annual percentage changes in ESRD incidence two years later (Y⫽0.7X ⫹ 1.16); (r⫽0.754, r2⫽0.568, p⫽0.03). The data from 1990 to 2001 show that changes in US diuretic supply and consumption were directly associated with changes in ESRD incidence in the US two years later. A hypothesis can be proposed that ESRD incidence rates might attenuate if diuretic use is deliberately minimized in the United States as much as possible.

Key Words: ESRD, Diuretics, Data Fusion

Arne Melander, Antoni M. Zawadzki, Qing Liu, Yusheng Wang, Bengt Jeppsson, Henrik Thorlacius. Department of Community Medicine, Medical Research Centre, Malmo University Hospital, Malmo, Sweden; The NEPI Foundation, Malmo and Stockholm, Sweden; Department of Surgery, Malmo University Hospital, Malmo, Sweden. It has been suggested that calcium ion (Ca2⫹) channel blocking drugs (CCB), commonly employed as antihypertensive agents, may promote growth of pre-existing cancer cells by inhibition of apoptosis. So far, however, there are no experimental data to support this notion, and results from clinical trials and epidemiological studies have been negative or ambiguous, respectively. Therefore, we examined if vascular type (L-type) Ca2⫹ channels are expressed in human colon cancer cells and if L-type Ca2⫹ channel activators (CCA) and blockers (CCB) influence Ca2⫹ influx and subsequent apoptosis in such cells. Both primary (collected at operation) and commercially available human colon cancer cell lines were used. The cells were incubated with three different CCB (verapamil, diltiazem or nifedipine) and one CCA (BayK 8644) at clinically relevant concentrations. RNA was determined by reversetranscription polymerase chain reaction (RT-PCR SuperScript one-Step). Intracellular Ca2⫹ levels were measured by fluorometry (Fluo-4-AM). Apoptosis was quantified by flow cytometry (Annexin V binding). Cells from both lines expressed vascular (L-type) Ca2⫹ channel mRNA but neither N- nor P-type channel mRNA. The L-type Ca2⫹ channels were composed of a1D and b3- subunits. The selective L-type CCA BayK 8644 markedly increased Ca2⫹ influx and apoptosis in the cells, and CCB pretreatment abolished BayK 8644-induced Ca2⫹ influx and apoptosis. Accordingly, while there is no evidence that CCB induce transformation of normal cells to cancer cells, these drugs may promote growth of pre-existing cancer cells in humans. As this would become clinically relevant only in subjects who already harbor a critical mass of cancer cells, the negative results in clinical trials and the conflicting results of epidemiological studies are not surprising. Until studies have been conducted in appropriate patient groups to substantiate or refute the notion that CCB treatment poses a real cancer threat, it seems prudent to minimize the use of CCB in hypertensive patients. Key Words: Calcium channel blockers, Cancer, Apoptosis