Efficacy in Reducing Low\x=req-\ Density Lipoprotein Cholesterol Levels on High- vs Low-Fat Diets Lovastatin

Margaret M. Cobb, MD, PhD; Howard S. Teitelbaum, DO, PhD; Jan L. Breslow, MD The effectiveness of lovastatin was compared with both a high-fat vs low-fat diet. Hypercholesterolemic subjects were studied under metabolic ward conditions for diet periods of 3 weeks while receiving lovastatin (40 mg/d) or placebo. Multiple lipoprotein levels were measured during the final week of each diet period. Nineteen subjects completed the study on the high-fat (43% of kilojoules) diet and 16 on the low-fat (25% of kilojoules) diet. Lovastatin reduced total cholesterol by 23% and low-density lipoprotein cholesterol by 30%, compared with placebo on both diets, with no significant diet-drug interaction. High-density lipoprotein cholesterol was raised by 7% to 8% on the diet regimens. Addition of lovastatin to the low-fat diet permitted 80% of subjects on this diet, but less than 50% of those on the high-fat diet, to achieve current guidelines. Although lovastatin produces a comparable percentage reduction in lipoprotein profiles on either diet, the accompanying low-fat diet remains advisable for additional reduction of low-density lipoprotein cholesterol levels to specified goals. (JAMA. 1991;265:997-1001)

THE ADULT treatment panel of the National Cholesterol Education Pro¬ gram (NCEP) has specified a patientbased approach for the treatment of hy-

percholesterolemia.[ Individuals with

elevated total cholesterol (TC) levels are directed to have their low-density lipoprotein (LDL) cholesterol concen¬ trations measured. The LDL cholester¬ ol measurement then guides therapeu¬ tic strategies. The goal is to reduce LDL cholesterol levels below 3.36 mmol/L in high-risk patients. Initially the recom¬ mended mode of therapy is a low-fat diet, with drug therapy started only if this dietary regimen fails. The sequence of diet first and drug second is appropri¬ ate and avoids the overuse of medicines with their expense and side effects. Unfortunately, in clinical situations, LDL cholesterol-lowering drugs are of¬ ten introduced without an adequate di¬ etary trial, because physicians lack con¬ fidence in diet efficacy while their patients lack the discipline necessary to

complete protracted dietary regimens. The most commonly used of these drugs is lovastatin (Mevacor), a 3-hydroxy-3methylglutaryl CoA reductase inhibi¬ tor.2 This drug blocks the rate-limiting From the Laboratory of Biochemical Genetics and Metabolism, The Rockfeller University, New York, NY (Drs Cobb and Breslow), and Department of Community Health Science, Michigan State University, East Lan-

sing (Dr Teitelbaum). Reprint requests to 1735 York Ave, New York, 10128 (Dr Cobb).

NY

step in endogenous cholesterol biosyn¬ thesis, thereby diminishing the hepatic

pool of cholesterol. Low-cholesterol di¬ ets, on the other hand, limit the exoge¬ nous cholesterol supply and, thus, de¬ plete hepatic cellular cholesterol.3,4 Reduction of hepatic cholesterol levels by either means triggers a compensa¬ tory increase in LDL-receptor activity, enhancing LDL cholesterol clearance from plasma.5 Since hepatic LDL-re¬ ceptor regulation is a site of action com¬ mon to both drug and diet, lovastatin combined with a low-fat diet may syner-

gistically enhance receptor activity, producing even greater reductions in plasma cholesterol than diet plus drug alone. Diet-drug synergism (interac¬

tion) has been demonstrated in the treatment of other chronic diseases.6"9 We sought to determine first whether lovastatin plus a low-fat diet reduced lipid levels further than the individual effects of diet alone or drug alone (ie,

diet-drug interaction) and, second, if the NCEP LDL cholesterol goals of less

than 3.36 mmol/L

are

achieved

by

a

produces

a

greater proportion of subjects on a lowfat diet in conjunction with lovastatin. We found that lovastatin

comparable (percentage) improvement in plasma lipoprotein profiles on both diets. The accompanying low-fat diet

remains advisable for further reduction of plasma LDL cholesterol levels to

specified goals.

Downloaded From: http://jama.jamanetwork.com/ by a Michigan State University User on 03/11/2014

SUBJECTS AND METHODS

Subjects Twenty-nine hypercholesterolemic subjects with plasma LDL cholesterol

levels greater than the upper quartile,10 stratified for age and gender, were re¬ cruited for this study; 26 of these sub¬ jects completed the study. This group, which was composed of 12 women and 17 men, had a mean age ( ± SD) of 53 ± 13 years (Table 1), with a body mass index of 26 ±3. No subject had evidence of hepatic, renal, or secondary lipid disor¬ ders. Five subjects presented with clin¬ ical criteria suggestive of monogenic fa¬ milial hypercholesterolemia.11 Eight subjects exhibited coronary and/or pe¬ ripheral vascular disease. Subjects who were taking ß-blockers and calcium channel blockers continued these medi¬ cations during the study. Over the year prior to the study, two subjects had been treated with cholestyramine (Questran), one with gemfibrozil (Lopid) and one with lovastatin. Subjects

discontinued

hypolipidemic medica¬

tions for at least 1 month prior to initia¬ tion of this study. All subjects gave in¬ formed consent for participation as approved by the Rockefeller University Institutional Review Board for Human Investigation, and were admitted to the Rockefeller University Metabolic Unit. Baseline parameters (medical histo¬ ry, physical examination, and biochemi¬ cal analysis) were determined prior to the study. The subjects' weights were monitored weekly, and the Harris-Ben¬ edict12 equation was used to estimate the subjects' initial energy requirements. Subjects continued prior exercise regi¬ mens during this investigation, with body weights maintained within ± 2 kg.

Study Design All subjects were randomly assigned to one of two metabolically controlled diets, the 43% fat (high-fat) diet or the 25% fat (low-fat) diet. After completion of the first diet, subjects were offered the second alternative. Only nine of the initial 29 subjects consumed both meta¬ bolic diets with intervening 10-day washout periods. In addition to these

Table 1.—Patient Characteristics, Diet

Groups, and 'Counseled' Baseline Lipld Profiles* Total Lipids,

No. in

Study Group

Age,

Male

y

Cholesterol

mmol/Lt

%

TC

29

53±13

59

8.00

19

52±13

58

8.20 ±1.62

19

54±11

63

7.95 ±1.55

+

Subfraction,

TG All Subjects 1.62 ±0.83

1.59

HDLC

LDLC/ HDLC Ratio

5.72 ±1.40

1.28 ±0.29

4.75 ±1.81

5.86 ±1.44

1.29 ±0.29

4.91 ±2.11

5.62±1.37

1.25±0.31

4.90 + 2.15

VLDLC

1.01 ±0.55 Low-Fat Diet Group 1.71 ±0.93 1.05 ±0.59 High-Fat Diet Group 1.76±0.95 1.09±0.61

*TC indicates total cholesterol; TG, triglycérides; VLDL C, very low-density lipoprotein cholesterol; lipoprotein cholesterol. Values represent mean ± SD. tData shown are values following the 37% fat counseled diet.

LDL C,

mmol/L

LDLC

low-density llpoprotein cholesterol;

and HDL C,

high-density

LOVA(5) Totals

M/W1 Washout Metabolic

\

Weeks

I

I

I

I

I

0

1

2

3

4

5

6

Dropouts

I

I

1

I

I

I

I

I

I

I

I

I

I

7

8

9

10

11

12

13

14

15

16

17

18

19

20

Fig 1.—Study design. Refer to text for description. Numbers in parentheses represent numbers of subjects who were randomized (LOVA) or placebo (PLA).

to receive lovastatln

nine diet-crossover subjects, 10 of the remaining 20 subjects consumed only foods from the 43% fat diet and 10 con¬ sumed only foods from the 25% fat diet 1 (Fig 1). Within each diet group, subjects were further randomized in a doubleblind fashion by our pharmacy to re¬ ceive lovastatin (20 mg orally, to be tak¬ en twice a day) or placebo in a crossover design, for 3-week periods. The resul¬ tant drug sequence for subjects is dis¬ played in Fig 1. Diets Counseled Diet.—An initial 3-week washout period was employed during which subjects were counseled as out¬ patients to follow a "typical" US diet.13 Referred to as the "counseled diet," it approximated a 37% fat, 16% protein,

and 47% carbohydrate diet, with a polyunsaturated to saturated fat ratio ap¬ proximately equal to 0.5, and an esti¬ mated dietary cholesterol intake of 400 mg/d. This diet was reemployed for all subsequent 10-day washout periods. A review of 24-hour recall logs showed similar adherence across both the diet groups. Blood samples were drawn at the end ofthis initial counseled diet peri¬ od and lipid profiles were averaged to represent the baseline (Table 1). To qualify for the study, all LDL cholester¬ ol levels had to remain above the 75th percentile for those subjects on the counseled diet. As presented, the base¬ line parameters in the low- and high-fat diet groups were statistically equiva¬ lent. A comparison ofbaseline values for subjects who consumed foods from both

Downloaded From: http://jama.jamanetwork.com/ by a Michigan State University User on 03/11/2014

diets with those subjects who consumed foods from only one of the diets showed no statistically significant differences. For those subjects who consumed foods from only one diet, the percentage re¬ ductions in lipid profiles did not differ from the nine crossover subjects; hence, the combined groups were used for the basis of this analysis. Three subjects who participated in the 25% fat diet group stopped the study before comple¬ tion and were excluded from the analy¬ sis; a comparison of their baseline pa¬ rameters and profiles with those subjects who did complete the study showed no significant differences. Metabolic Diets.—Both metabolic diets were prepared in the clinical re¬ search center kitchen and were based on two standards: the 25% fat (low-fat)

diet, following the American Heart As¬ sociation's step 2 diet,14 and the 43% fat (high-fat) diet, representing the 75th

outpatients (Fig 1). Subjects were mon¬ itored for adverse reporting and toxicity studies weekly, with no adverse re¬ actions noted.

rank for dietary fat intake based on the most recent National Health and Nutrition Examination Sur¬ 13 vey. The two metabolic diets consisted of natural foods, weighed to 0.1 g, with a 2-day matched rotating menu. A natu¬ ral food supplement matched in compo¬ sition to the diet treatments was given as needed to maintain body weight. The formulations were initially calcu¬ lated from the Human Nutrition Infor¬ mation Service15 food tables (Table 2). A homogenized aliquot of the two con¬ trasting diets was further analyzed for actual dietary cholesterol content16 and nutrient composition (Analytical Chem¬ ical Labs, New York, NY).17 The actual fatty acid composition of the oils that were fed to the subjects in the two diets was determined by gas-liquid chromatography. " The calculated and actual di¬ etary cholesterol content and nutrient compositions for the two diets were deemed similar.

percentile

Laboratory Analysis Blood samples were collected after a 12-hour fast, and plasma was separated at 4°C by centrifugation. Total choles¬ terol and triglycéride levels were as¬ sayed enzymatically by reagents in kits supplied by Boehringer Mannheim Biochemicals (Indianapolis, Ind). The highdensity lipoprotein (HDL) plus LDL cholesterol concentrations were deter¬ mined after air centrifugation to sepa¬ rate the very low-density lipoprotein (VLDL) cholesterol subfraction; LDL and VLDL cholesterol levels were de¬ termined by difference as reported pre¬ viously.1819 Frozen aliquots of plasma were assayed for apolipoprotein B (apo B) levels by sandwich enzyme-linked immunosorbent assay.19'20 The interassay coefficients of variation for TC, tri¬ glycérides, VLDL, LDL, HDL choles¬ terol, and apo B were 1.0%, 2.7%, 3.8%, 3.8%, 2.4%, and 5.2%, respectively, as previously reported.19 Total cholesterol, HDL cholesterol, and apo B levels were standardized by a reference serum pool

Drug.—Each drug treatment was separated by an intervening 10-day washout period, during which subjects were

returned to the counseled diet

Table 2—Diet

as

Composition* 43% Fat Diet

25% Fat Diet

Calculated Diet

S, %

M, %

Calculated

P, %

Actual

M, % Component_Actual Fat, %kj_45J_42_51_7_208_30_26

P, %

S, %

44

ratio_ _017_L_J_"L47_ 207_216_67_78_ Carbohydrate, %kJ_37J_41_55J_59_ P:S

Cholesterol, mg/4184 kj Protein, %kJ

17.8

17.5

16

16

"Actual refers to composition measured by food analyses. Calculated refers to composition computed from food tables S indicates total saturated fatty acids; M, total monounsaturated fatty acids; and P, total polyunsaturated fatty acids. Table 3— Influence of Diet and

Drug Treatments on Lipid Profiles and Apollpoprotein 43% Fat Diet

Placebo, mmol/L

B

(apo B)

Levels

from the Centers for Disease Control, Atlanta, Ga. Statistical Analysis A retrospective analysis was per¬ formed on lipid profiles that were ob¬ tained from our clinic population and treated with doses of lovastatin to esti¬ mate sample sizes. On average, the SD of TC was 1.29 mmol/L. A significant therapeutic response to lovastatin was predicted to be a 1.29-mmol/L TC re¬ duction. Thus, an approximate sample size of 17 per group was required for a Type I error equaling .05 and power of

0.80.21

Three research questions were inves¬ tigated: (1) Were the subjects' lipid pro¬

files different when treated with lova¬ statin relative to the placebo (drug effect)? (2) Were the subjects' lipid pro¬ files different when the subjects were fed high- or low-fat diets with placebo (diet effect)? (3) Was there a diet by drug interaction?22 All statistical tests employed the 5% significance level. The z test for proportions tested for differ¬ ences between the proportion of sub¬ jects who achieved NCEP goals while receiving lovastatin with the 43% fat and 25% fat diets.23 Statistical treat¬ ments were performed by blinded in¬ vestigators at Michigan State Universi¬ ty using the Statistical Analysis Software system (SAS Institute Inc, Cary, NC) for calculation of repeated multivariate analysis of measures covariance. RESULTS Dietary Effects on Lipid and Apolipoprotein Levels The effects of the 43% and 25% fat diets alone (placebo vs placebo) on lipid, lipoprotein, and apo B levels are shown for the pooled study groups (Table 3). Compared with the 43% fat diet, the

(Mean ± SD)* 25% Fat Diet Group

Group (n 19)

Lovastatin, mmol/L

=

=

Difference

Difference^

-"-*

A

(n 16)t

Placebo,

%A

mmol/L

Lovastatin, mmol/L

•-'->

A

%A

Lipid

23 ±7 6.63 1.86H_5.09 1.47_1.55±1.09§ 23±12 JC_7.99 ±1.76_6.13 ±1.45_1.86±0.70§ 11 19 1.62±0.79_1.41+0.68_0.21 ±0.62||_7±33 TG_1.38 ±0.48_1.20 ±0.43_0.18 0.36|| 13±29 1.09 ±0.54_0.83 0.52_0.31 ±0.54§ 21 ±5 VLDLC_0.98 ±0.39_0.80 ±0.26_0.18 0.31§ 30 ±9 30 11 LDLC_5.72 ±1.66_3.98 ± 1.42_1.73 0.541 4.42±1.73||_3.10 1.16_1.32 ±0.8511 -8 9 -7 9 HDLC_1.27 0.26_1.37 ±0.26_-0.10±0.101 1.11±0.14||_1.19 0.34_-0.08±0.08H 4.7 ±1.9_3.0 ±1.3_1.7 LDLC/HDLC ±0.71_36±8_4.3 ±2.2_2.9 1.7_1.4 ±1.011_33±12 +

+

+

+

+

+

+

+

+

+

+

+

+

+

apoB

1.69±0.42g/L

1.17±0.31

g/L

0.52±0.321

30±14

1.39±0.35g/L

1.04 + 0.35g/L

0.35±0.171l

26±13

*TC indicates total cholesterol; TG, triglycérides; VLDL' C, very low-density lipoprotein cholesterol; LDL C, low-density lipoprotein cholesterol; HDL C, high-density lipoprotein cholesterol; LDL C/HDL C, low-density lipoprotein cholesterol/hlgh-density lipoprotein cholesterol ratio tAt entry, 19 subjects were randomized to the 25% fat diet group (Table 1 ) ; three subjects did not complete both phases of the diet-drug regimen and were excluded from the

analysis (see text and Fig 1 ).

tDiet effect: 43% fat diet and placebo vs 25% fat diet and placebo. Drug effect: lovastatin

text).

vs

placebo,

same

§P