I CUNICAL STuDlES

Effect of Gemfibrozil in Men With Primary Isolated Low High-Density Lipoprotein Cholesterol: A Randomized, Double-Blind, Placebo-Controlled, Crossover Study MICHAELMILLER. M.D., PAULS. BACHORIK, Ph.D., BRIAN W. MCCRINDLE, M.D., PETERO. KWITEROVICH, JR., M.D., Baltimore, Mary/and

PURPOBE: To evaluate the efficacy of gemfibro- xii in men with primary isolated low high-densi- ty lipoprotein cholesterol (HDL-C) levels.

PATlENlB AND lKEl’HOD& Fourteen men wlth low levels of HDL-C but desirable total choles- terol levels received gemfibroxil in a random- ixed, double-blind, placebo-controlled, crossover trial. The men were placed on a National Cho- lesterol Education Program Step-Two Diet. They were randomly assigned to receive placebo and gemfibroxil each for 3 months, with a l- month washout period between phases.

RESULT& Overall, gemfibroxil increased the to- tal HDL-C concentration by 9.2% (p = O.OOl), re- duced triglyceride (TG) levels by 38% (p <O-01), and significantly lowered the total cholester- ohHDI,-C ratio (p = 0.01). Those with fasting TG levels of 1.07 mmol/L (95 mg/dL) or greater had a significant elevation in the HDEC level (14.6%, p = 0.005) and a reduction in TG levels (50%, p = 0.002) with gemfibroxih those with fasting TG levels less than 1.07 nuuol/L had a smalIer increase in the HDL-C level (4.1%, p >0.05) and a smaller reduction in TG levels (15%, p = 0.02). There were no significant dif- ferences in the plasma levels of low density lipo- protein-cholesterol, HI&&, apolipoproteins (ape) A-I and B, or Lp(a). IIDk-C and apo A-II levels rose slightly. The adverse effects attribut- able to gemfibroxil were ’ ’ 1.

CONCLUSIONz In men with desirable total cho- lesterol levels, gemfibroxil raises HDL-C and

From the Lipid Research-Atherosclerosis Unit. Deoartments of Medi- cine, Laboratory Medicine and Pediatrics, Johns Hopkins Medical Institu- tions (MM, PSB. BWM. PDK). and the Department of Medicine, University of Maryland Medical Systems (MM), Baltimore, Maryland.

The study was supported by Grants HL 31497 and National Research Service Fellowship Award 5F32HL07596 from the National Institutes of Health, and a grant from Parke-Davis. division of the Warner-Lambert Company.

Requests for reprints should be addressed to Michael Miller, MD. Uni- versitv of Marvland Medical Systems. Division of CardioloavN3W84, 22 SouthGreeneStreet. Baltimore. Maryland 21201. -

Dr. McCrindle’s current address: Hospital for Sick Children, Cardiology Division, 555 Universlty Avenue, Toronto, Ontario, Canada.

Manuscript submitted November 19. 1991, and accepted in revised form June 8.1992.

lowers TG levels to a similar extent as reported for hyperlipidemic men in the Helsii Heart Study. These lipid-altering effects were most pronounced in those with the highest fasting TG levels.

A lthough epidemiologic studies have long dem- onstrated the strong, and often independent,

inverse correlation between high-density lipopro- tein cholesterol (HDL-C) and coronary heart dis- ease (CHD) [l-3], the separate effect of increasing the HDL-C level on reducing the incidence of CHD has only recently been confirmed [4]. Not only is there a high prevalence of low HDL-C levels among CHD patients with “desirable” total cholesterol (TC) levels (less than 5.18 mmoliL or 200 mg/dL) [5,6], but a low HDL-C level has also been found to predict subsequent CHD events [7].

Despite the apparent high risk of CHD in sub- jects with a low HDL-C level, specific therapeutic recommendations exist solely in those manifesting concomitant hypercholesterolemia. This, in part, reflects the lack of controlled studies specifically addressing whether low HDL-C levels can effective- ly be increased in the absence of other lipid and lipoprotein abnormalities. In one study attempting to address this issue, nearly one third of the subjects were concurrently receiving p blockers, a known secondary cause of low HDL-C levels [ES]. In the current study, we eliminated potential confounders [9] in order to study subjects whose sole abnormali- ty was an isolated low level of HDL-C.

PATIENTS AND METHODS Men between the ages of 21 and 60 years were

recruited using the following criteria: average plas- ma levels of TC less than 5.18 mmol/L (200 mg/dL), HDL-C less than 1.04 mmol/L (40 mg/dL), triglyc- eride (TG) less than 1.69 mmol/L (150 mg/dL), and low-density lipoprotein cholesterol (LDL-C) less than 3.37 mmol/L (130 mg/dL) from two baseline samples obtained after a la-hour fast and analyzed in the Johns Hopkins Core Lipoprotein Analytical Laboratory. Those men who were cigarette smok-

lanuary 1993 The American Journal of Medlclne Volume 94 7

GEMFlGROZfL FOR ISOLATED LOW HDLC / MILLER El AL

Phase Baseline I Washout II

I I I I I I I I

Week 1 9 18 21 25 34 37

Blood Sample t 1 t I

Flgure 1. Outline of experimental protocol. After the baseline

period, patients were randomly assigned to receive drug or

placebo for 3 months (Phase I). After a l-month washout pe-

riod, they were crossed over (Phase II) to the alternate treat-

ment plan for 3 months. Blood samples were collected during

the 9th and 12 weeks of each phase.

em, obeae (greater than 30% ideal body weight), or receiving any medications known to alter HDL-C levels (e.g., /3 blockers, steroids) were excluded. In addition, men with unstable angina, diabetes melli- tus, a thyroid disorder, nephrotic syndrome, or pre- existing gall bladder disease or hepatic disease were also excluded. After screening approximately 100 individuals referred to the Johns Hopkins Lipid Clinic because of a low HDL-C leve1, we identified 14 subjects who met the above criteria and vohm- teered to participate. Informed consent was ob- tained, and the study was approved by the Johns Hopkins Joint Committee on Clinical Investigation.

study Design The experimental design of the study is illus-

trated in Figure 1. To control for dietary variation, each subject was instructed by a registered dietitian to follow the Step-Two Diet of the National Choles- terol Education Program, which consists of deriving 55% of dietary calories from carbohydrates, 15% from protein, and 7% and 30% from saturated fat and total fat, respectively, and ingesting less than 200 mg of cholesterol daily [lo]. Dietary compliance was monitored using 24-hour recalls during routine visits and 72-hour written dietary recalls at the completion of each phase of the study. Each subject was asked to maintain his usual level of exercise throughout the course of the study and not to con- sume more than one drink (2 oz or 60 mL) of alcohol per day.

If all of these conditions were met, each subject was then randomly selected to receive either place- bo or gemfibrozil(600 mg twice a day) for 3 months. This was followed by a l-month washout period during which the subjects received no medication. Seven subjects received placebo in Phase I and were then assigned to gemfibrozil for Phase II; the other seven received gemfibrozil first, followed by place-

8 January 1993 The American Journal of Mdlcinr Volume 94

bo. Neither the participants nor the clinical or labo- ratory staff knew whether a placebo or active drug was being taken.

Laboratory Measurements Following an overnight (12-hour) fast, venous

blood was obtained after the subject was in the su- pine position for 10 minutes (to minimize postural variation [ 111). Plasma was separated within 30 minutes of collection, and stored at 4°C until it was analyzed in the Johns Hopkins Lipoprotein Analyt- ical Laboratory. This laboratory is standardized for cholesterol, TG, and HDL-C measurements accord- ing to the Collaborative Centers for Disease Con- trol, National Heart, Lung, and Blood Institute Cri- teria for Standardization [12]. TC and TG concentrations were measured enzymatically on a Hitachi 704 clinical chemistry analyzer (Boeh- ringer-Mannheim, Indianapolis, IN). HDL-C was measured after precipitation of apolipoprotein (ape) B-containing lipoproteins with heparin sul- fate and manganese chloride (1.3 g/L and 0.092 M, respectively) [13]. LDL-C was calculated from the Friedewald formula: [LDL-C] = [TC] - [HDL-C] - [TG/5] 1141.

Plasma was adjusted to density 1.12 g/L using liquid sodium chloride:potassium bromide [15] and subjected to ultracentrifugation for 30 hours at 10°C. HDLs-C was measured in the ultracentrifugal infranatant fraction. HDLz-C was estimated as the difference between total HDL-C and HDLe-C (161. Apo A-I and B concentrations were determined us- ing commercially available immunodiffusion plates (apo A-I, Tago, Inc., Burlingame, CA; apo B, Beh- ring Diagnostics, La Jolla, CA). Apo A-II was mea- sured with a modification of an established proce- dure [ 171. Lp(a) was measured by an enzyme-linked immunosorbent assay using a commercially avail- able method (Terumo, Elkton, MD). During the course of the study, the coefficients of variation for these measurements were as follows: TC = 1.5%; TG = 4.4%; HDL-C = 2.9%; apo A-I = 4.2%; apo A-II = 7.4%; apo B = 3.9%; and Lpfa) = 6.8%.

Safety Assessments At each visit, subjects were asked whether they

experienced any adverse reactions; alI positive re- sponses were recorded by the study coordinator. Weight, blood pressure, and heart rate were mea- sured at each visit. A complete physical examina- tion and electrocardiography were performed dur- ing the baseline period and at the final visit. Complete blood cell counts, blood chemistry tests, and coagulation profiles were performed at baseline and at the completion of both study phases.

GEMFIGROZIL FOR BCR.ATED LOW HDLC / MILLER ET AL

TABLE I

Baseline Characteristics of Subjects With Isolated Low HDL-C Levels Prior to Randomization*

Subject ?f; BMI ETOH

TC TG HOL-C LDL-C (kg/m*) (mL/d) Ex FHCAD

: ;: 155 148 :; 36 102 24.3 23.0 t30 0 t

i 3": 134

;! ii z: 25.5 <30 t tFMI-53

162 zi

114 24.6 t30 t tF Ml-59 z 61 153 139 128 146 ;: 25.8 22.6 <30 0 tFMI-45 t

ii :; 147 134 105 100 :; 91

2': 3306 39 102 :z

21.8 19.6 <30 <30 t t tF MI-31

1: 131 155 :: 22.6 20.5 t30 <30 t t FCHD-53

i: i8 145 151 125 <60 t tFMI-46

13 3": 133 1:;

;: z: 25.0 21.8

i t tFMI-58

26.4 t tMCABG-50 14 57 171 118 zz 1;; 26.0 <60 t sip CABG-47

I

BMI = body mass index; F = father; M = mother; Ex = aerobic exercise; s/p = status post; CABG = coronary artery bypass graft; MI = myocardial infarctron; FHCAD = famrly history of coronary artery disease; ETOH = alcohol consumed (ml/d). Other abbreviations as in text. ‘Lipid and lipoprotein values (mg/dL) represent the average of two samples rounded off to the nearest integer.

Compliance was monitored by calculating the difference between the prescribed number of tab- lets to be taken during the given interval and the number returned during the subsequent visit.

Statistical Analysis For each phase of the study, plasma levels of lip-

ids, lipoproteins, and apo A-I and apo-B were calcu- lated as the average of two measurements obtained at Weeks 18 and 21 in Phase I and Weeks 34 and 37 in Phase II (Figure 1). HDL-C subfractions and apo A-II levels were determined at the completion of each phase; apo A-II levels were not measured in two subjects because the samples were insufficient. A paired t-test (two-tailed) was used for the statisti- cal analysis, with p <0.05 as the designated level of significance.

RESULTS Table I lists the clinical characteristics of each

participant in the study. The average age of the participants was 35 f 11 years. The majority of the subjects gave strong, positive family histories for premature CHD. Patient 14 had a personal history of CHD. None of the patients was obese (mean body mass index = 23.5), and most exercised aerobically [18] several times weekly.

The average caloric intake, dietary cholesterol consumed, and the percent distribution of calories derived from protein, carbohydrate, and fat while receiving placebo or drug are shown in Table II. The average fat content was well within the range defined by the National Cholesterol Education Pro- gram Step-Two Diet, and no differences were ob- served between the groups. Similarly, no significant differences were noted between the various nutri-

TABLE II

Total Caloric Intake and Percentage of Calories Derived From Energy Sources During Placebo or Gemfibrozil Treatment

I % of Calories as Fat

‘~~~~o~ilc~~~~~c~p~~~~~~~~~ CHO = carbohydrates: Sat = saturated; Mono = monounsaturated; Poly = polyunsaturated.

ent groups or within various types of fats consumed. The amount of alcohol consumed by the group (av- erage less than 60 mL/d) did not differ between the placebo and drug phases.

Mean lipid, lipoprotein, and apo and Lp(a) levels are summarized in Table III. During treatment with gemfibrozil, the mean TG levels and the TC:HDL-C ratio were significantly decreased, whereas the HDL-C levels significantly increased, compared with placebo. In addition, the concentra- tions of HDLs-C and apo A-II rose slightly, al- though not significantly. Neither the HDLs-C nor apo A-I levels increased with gemfibrozil treatment, and there were no significant differences in the lev- els of LDL-C, apo B, or Lp(a) during drug therapy.

Gemfibrozil was well tolerated, and no significant side effects were reported, three subjects reported mild gastrointestinal discomfort during treatment with gemfibrozil. No alterations were detected in complete blood cell counts, results of blood chemis- try tests, or coagulation indices during drug treat- ment. The compliance rates for drug treatment and placebo were 91.4% and 91.3%, respectively.

1 lanuaty 1993 The American Journal of Medicine Volume 94 9

GEMFIBROZIL FOR l3Ol.ATED LOW HDLC / MILLER ET AL

TABLE III

Effect of Gemfibrozil in Men With Primary Low HDL-C Levels*

Measurement Placebo Gemfibrozil

TC 3.80 2 0.10 3.76 ?r 0.11 (147.2 + 3.8) (145.5 ? 4.3)

TG 1.15 * 0.11 0.71 ? 0.04’ (102.0 + 9.9) (63.6 + 3.3)

HDL-C 0.82 + 0.02 0.89 + 0.03t (31.6 2 0.9) (34.5 2 1.1)

HDLz-C 0.22 + 0.04 0.21 + 0.04 (8.4 2 1.6) (8.1 + 1.4)

HDLj-C 0.57 f 0.03 0.64 2 0.03 (22.1 * 1.1) (24.9 _+ 1.2)

LDL-C 2.45 f 0.10 (94.8 + 3.7)

2.53 + 0.11 (98.1 + 4.3)

TC/HDL-C 4.7 2 0.16 4.3 + 0.201

ApoA-1 1.23 2 0.03 1.23 ? 0.03

Apo A-II 0.23 f 0.09 0.25 f 0.13

ApoB 1.10 lr. 0.04 1.10 T 0.06

Lp(a)

breviations as in text.

0.09 f 0.03 0.11 + 0.02

_-. alues are expressed in mmol/L and lmg/dL) r standard error (St1 ot the mean tar plasma hplds

;nd;\ytTteins and in g/L ? SE for apolipoproteins and Lp(a).

$= 0.001.

COMMENTS The significance ascribed to aggressive manage-

ment of hyperlipidemia has been borne out by evi- dence demonstrating reduction of primary CHD events and retardation or regression of established atherosclerotic lesions [19-211. While reduced HDL-C levels may accompany elevations in TG and LDL-C concentrations in the setting of CHD, there is increasing evidence that a low HDL-C level is often associated with both the presence and sever- ity of angiographically proven coronary disease when other lipids are not elevated [5,6,22]. How- ever, it is unclear whether pharmacologic agents that provide beneficial alterations in HDL-C and triglyceride levels in hyperlipidemic patients have similar efficacy when the TC concentration is in the desirable range. As such, the purpose of the current study was to examine whether a primary, isolated low HDL level can be increased by treatment with gemfibrozil, an agent with established efficacy in the treatment of hyperlipidemia, especially hyper- triglyceridemia associated with a low HDL-C level

WI. Nonpharmacologic approaches to manage pa-

tienta with a low HDL-C level often include recom- mendations for dietary changes, reduction of satu- rated fat intake [23], cigarette smoking cessation [24], weight loss in obese persons [25], and aerobic

exercise [26]. Many persons with a low HDL-C lev- el, however, are not obese [5], and exercise is often less effective in those displaying the lowest levels of HDL-C [27]. The men in the current study were nonsmokers, nonobese, and had been following a low-fat diet prior to randomization; furthermore, approximately 80% performed aerobic exercise on a regular basis. Yet, despite adherence to accepted hygienic measures, their levels of HDL-C were low.

Although a number of pharmacologic studies in hyperlipidemic patients have demonstrated reduc- tions of elevated LDL-C and TG levels with recipro- cal increases in HDL-C [19-211, only two studies have specifically addressed the management of low HDL-C [8,28]. However, both studies employed higher entry criteria for TC and TG concentrations than ours, and use of 0 blockers was not excluded. The current study utilized more stringent entry cri- teria in an effort to focus exclusively on the primary low HDL-C level as the sole abnormality.

Medications that are most effective in raising the HDL-C level include nicotinic acid and the fibric acid derivative, gemfibrozil. Nicotinic acid may raise HDL-C levels by inhibiting hepatic produc- tion of very-low-density lipoprotein cholesterol (VLDL); this reduces the transfer of core choles- teryl esters from HDL-C, thereby inhibiting its ca- tabolism [29]. Although nicotinic acid is very effec- tive in raising HDL-C levels in hyperlipidemic patients, side effects may limit ita utility in some individuals [30]. In a recent study, nicotinic acid raised HDL-C level 27% in men with a desirable TC level. However, a 10% increase in TG levels was also observed [28].

Gemfibrozil is believed to increase HDL-C by two mechanisms. First, gemfibrozil reduces production of VLDL. In addition, the drug enhances the activi- ty of lipoprotein lipase, resulting in greater hydro- lytic efficiency of triglyceride-rich lipoproteins and an increased transfer of surface constituents to na- scent HDL-C [31]. In the Helsinki Heart Study, use of gemfibrozil led to a mean increase in HDL-C of 10% and a mean reduction in TG of approximately 40% [19]. Importantly, the Helsinki Heart Study was the first study to demonstrate the independent effect of an elevation in the HDL-C concentration on the primary CHD event rate, thus underscoring the notion that small elevations in HDL-C may have substantial biologic effects [4]. Although simi- lar increases in the HDL-C concentration were ob- served in our study, no changes were found in the apo A-I levels and small but statistically insignifi- cant increases were noted for apo A-II. The absolute change in the plasma concentrations of HDL-C was 2.9 mg/dL (p = O.OOl), and the final concentration of HDL-C was 34.5 mg/dL.

10 January 1993 The Amerlcan Journal of Medicine Volume 94

The elevation in HDL-C was, in large part, ac- counted for by a concomitant reduction in TG levels (R = -0.49, p = 0.07). This is not surprising, given the well-known inverse relationship between HDL- C and TG-rich lipoproteins [32]. In fact, patients with the highest TG levels had the best response to gemfibrozil. Thus, patients with fasting TG levels of 95 mg/dL (1.07 mmol/L) or more during the pla- cebo phase had a mean increase of 4.6 mg/dL (14.6%) in HDL-C (p = 0.005) and a mean reduction of 66 mg/dL (50%) in TG (p = 0.002) during drug therapy. In subjects with TG levels below 95 mg/dL (1.07 mmol/L), the average HDL-C level rose only 1.2 mg/dL (4.1%) (NS) and TG levels were reduced 11 mg/dL (15%) (p = 0.02). This suggests that with fasting TG levels between 95 and 150 mg/dL, or within the range classified as desirable (i.e., less than 250 mg/dL) [33], gemfibrozil appears to be effective in both increasing HDL-C and reducing TG concentrations. In contrast, with fasting TG levels below 95 mg/dL, gemfibrozil appears to pro- duce a substantially smaller effect.

Overall, the results of our study mirror those of the Helsinki Heart Study, except that our subjects were neither hypercholesterolemic nor hypertrigly- ceridemic. The overall 9.2% increase in total HDL- C was primarily accounted for by elevations in HDL&. That this finding may have particular rel- evance is reflected in the Physicians’ Health Study, which found that among nearly 15,000 men, the HDLs-C level was the most potent inverse predictor of subsequent myocardial infarction [34].

In the current study, most patients gave family histories of premature CHD. A genetic predisposi- tion for a low HDL-C level is characteristic of famil- ial hypoalphalipoproteinemia, an autosomal domi- nant disorder strongly associated with premature CHD [35]. In view of the attendant high risk of CHD in this subgroup, the search for effective treatment deserves further consideration. None- theless, until definitive data become available, we believe that all individuals with CHD or those eval- uated for CHD risk (i.e., familial predisposition) should be screened for a low HDL-C level as recent- ly outlined by the National Institutes of Health Consensus Development Conference on triglycer- ides and HDL-C. Hygienic measures (i.e., cigarette cessation, reduced saturated fat intake, and aerobic activity) remain the cornerstone of treatment. Cur- rently, however, there are no recommendations for pharmaceutical treatment in subjects with an iso- lated low HDL-C level.

Because retardation of coronary atherosclerotic lesion progression has been demonstrated in CHD patients even without overt hyperlipidemia, we be- lieve that pharmacologic therapy should be consid-

ered in this subgroup [20]. Our study demonstrates that gemfibrozil increases isolated low HDL-C lev- els, particularly when fasting TG levels exceed 95 mg/dL. Whether gemfibrozil is effective in reducing the subsequent cardiac event rate in patients with an isolated low HDL-C level will be answered in the ongoing multicenter Veterans Administration HDL-C Intervention Trial.

ACKNOWLEDGMENT We thank Mr. Thomas Weber for serving as the study coordinator, Ms. Hazel Smith for serving as the primary phlebotomist, and Ms. Theresa Ckxy for coordinating the laboratory measurements and data entry. We would also like to thank registered diiticians, Ms. Virginia Hattmuller, Ms. Martha Cecil, and Ms. Judith Chiostri, for dietary consultations and analyses, and the staff of the Johns Hopkins Lipoprotein Analytical Laboratory.

REFERENCES 1. Gordon T, Castelli WP, Hjortland MC, Kannel WB. Dawber TR. High density lipoprotein as a protective factor against coronary heart disease. The Framing- ham Study. Am J Med 1977; 62: 707-14. 2. Miller NE, Forde OH, Thelle DS, Mjos OD. The Tromso Heart Study. Hi density lipoprotein and coronary heart disease: a prospective case-control study. Lancet 1977; 1: 965-7. 3. Goldbourt U, Medalie JH. High density lipoprotein cholesterol and incidence of coronary heart disease-the Israeli lschemic Heart Disease Study. Am J Epidamiol 1979; 109: 296-308. 4. Manninen V. Elo MO, Frick MH, et a/. Lipid alterations and decline in the incidence of coronary heart disease in the Helsinki Heart Study. JAMA 1988, 260: 641-51. 5. Miller M. Mead L. Kwiterovich PO, Pearson TA. Dyslipidemias with desirable plasma total cholesterol levels and angiographically demonstrated coronary artery disease. Am J Cardiol 1990; 65: l-5. 6. Ginsburg GS. Safran C, Pasternak RC. Frequency of low serum high-density lipoprotein cholesterol levels in hospitaliied patients with “desirable” total cho- lesterol levels. Am J Cardiol 1991; 68: 187-192. 7. Miller M. Seidler A, Kwiterovich PO, Pearson TA. Long-term predictors of subsequent cardiovascular events with coronary artery disease and “desirable” levels of plasma total cholesterol. Circulation 1992; 86: 1166-70. 8. Vega GL. Grundy SM. Comparison of lovastatin and gemfibrozil in normolipi- demic patients with hypoalphalipoproteinemia. JAMA 1989; 262: 3148-53. 9. Miller M, Kwiterovich PO. Isolated low HDL-cholesterol as an important risk factor for coronary heart disease. Eur Heart J 1990; 11 Suppl H: 9-14. 10. Report of the National Cholesterol Education Program Expert Panel on detection, evaluation and treatment of hi blood cholesterol in adults. Arch Intern Med 1988; 148: 36-69. 11. Miller M, Bachorik PS. Cloey TA. Normal variation of plasma lipoproteins: postural effects on plasma concentrations of lipids, lipoproteins, and apolipopro teins. Clin Chem 1992; 38: 569-74. 12 Myers GL. Cooper GR. Winn CL, Smith SJ. The Centers for Disease Control- National Heart, Lung and Blood Institute Lipid Standardization Program. An approach to accurate and precise lipid measurements. Clin Lab Med 1989; 9: 105-35. 13. Bachorik PS, Kwiterovich PO. Measurement of plasma cholesterol, Iow- density lipoprotein cholesterol and high-density lipoprotein cholesterol. In: Hommes FA. editor. Techniques in diagnostic human biochemical genetics. A laboratory manual. New York: Wiley-L&, 1991: 425-59. 14. Friedewald Wl, Levy RI, Fredrickson DS. Estimation of the concentration of low density lipoprotein cholesterol in plasma without use of the preparative ultracentrifuge. Clin Chem 1972; 18: 499-502. 11 Have1 RJ. Eder HA, Bragdon JH. The distribution and chemical composition of ultracentrifugally separated lipoproteins in human serum. J Clin Invest 1955; 34: 1345-53. 16. Lipid Research Clinics Program. Manual of laboratory operations, vol 1. Lipid and lipoprotein analysis. DHEW publication no. (NIH) 75-628. Washington, DC:

GEMFIGRGZIL FDR l9DlATED LDW HDLC / WIUER ET AL

January 1993 The Amrrlcan Journal ol Medlclno Volume 94 11

GEMFIGROZIL FOR l9Ol.ATED LOW HDL-C / MILLER ET AL

U.S. Government Printing Office, 1974. 17. Cheung MC. Albars JJ. The measurement of apolipoprotein A-l and A-II levels in man and women by immunoassay. J Clin Invest 1977; 60: 43-50. 18. American College of Sports Medicine. Guidelines for exercise testing and prescription. Philadelphii: Lea and Feblger, 1986. 19. Frick MH. Elo 0. Haapa K, et a/. Helsinki Heart Study: primary prevention trial wfth pmffbrozil in middle-a&d men with dyslipidemia. N Engl J Med 1987: 317: 1237-45. 29. Cashin-Hemphill L, Mack WJ, Pogada JM. et al. Beneficial effects of corn blned colestipol-niacin therapy on coronary atherosclerosis and coronary ve- nous bypass grafts. JAMA 1990; 264: 3013-7. 21. Brown G, Albers JJ, Fisher LD, et al. Regression of coronary artery disease as a result of intensive lipid-lowering therapy in men with high levels of apolipo- protein B. N En@ J Med 1990; 323: 1289-98. 22. Romm PA, Green CE. Reagan K. Rackley CE. Relation of serum lipoprotein cholesterol levels to presence and severity of angiographic coronary artery disease. Am J Cardii 1991; 67: 479-83. 23. Mensink RP. Katan MB. Effect of a diet enriched with monounsaturated or polyunsaturated fatty adds on levels of low density and hii density lipoprotein cholesterol in healthy men and women. N Engl J Med 1989; 321: 436-41. 24. Crlqui M, Wallace RB, Heiss G. Cigarette smoking and plasma high density lipoprotein cholesterol. Circulation 1980; 62 (Suppl IV): 70-6. 25. Caegiula AW. Christakis G. Farrand M, et a/. The Multiple Risk Factor Inter- vention Trial: intervention on blood lipids. Prev Med 1981; 10: 443-75.

26. Wood P, Williams PT, Haskell WL. Physical activity and high density lipopro- teins. In: Miller NE. Miller GJ. editors. Clinical and metabolic aspects of high density lipoproteins. New York: Elsevier Science Publishers, 1984: 13365. 27. Krauss RM. Exercise, lipoproteins, and coronary artery disease. Circulation 1989; 79: 1143-5. 28. Lavie CJ, Mailander L, Milani RY. Marked benefit with sustained-release niacin therapy in patients with isolated very low levels of high-density lipoprotein cholesterol and coronary artery disease. Am J Cardiol 1992; 69: 1083-5. 29. Miller NE. PharmacolMc intervention for altering lipid metabolism. Drugs 1990; 40 (Suppl 1): 26-32. 39. Blum CB, Levy RI. Current therapy for hypercholesterolemia. JAMA 1989; 261: 3582-6. 31. Nikkila EA. Ylikahri R. Huttunen JK. Gemfibrozil: effect on serum lipids, lipoproteins, post-heparin plasma lipase activities and glucose tolerance in pri- mary hypertriglyceridemia. Proc R Sot Med 1976; 69 SuppI 58-63. 32. Gotto AM, Pownall HJ, Have1 RJ. Introduction to the plasma lipoproteins. Methods Enzymol 1986; 128: 3-41. 33. National Heart, Lung and Blood Institute Consensus Development Panel: Treatment of hypertr@yceridemia. JAMA 1984; 251: 1196-2000. 34. Stampfer MJ, Sacks FM. Salvini S, Willett WC, Hennekens CH. A prospective study of cholesterol, apolipoproteins, and the risk of myocardial infarction. N Engl J Med 1991; 325: 373-82. 35. Glue& CJ, Melser MA, Borecki IB, Third JHLC, Rao DC, Laskarzewski PM. Familial hypoalphalipoproteinemia. Adv Exp Med Biol 1986; 201: 83-92.

12 January 1993 The American Journal of Medicine Volume 94