53

Acta Med Scand (Suppl) 701: 53-57, 1985

Randomised Controlled Trial of the Treatment of Hyperlipidaemia on Progression of Atherosclerosis

Barry Lewis

Department of Chemical Pathology and Metabolic Disorders

St Thomas’ Hospital London, UK.

ABSTRACT. Two types of end-point measure- ment are presently available in clinical trials of the effect of treatment of hyperlipidaemia on cardiovascular disease: these are the incidence of clinical events and the arteriographic assess- ment of progression or regression of athero- sclerosis. These approaches are briefly re- viewed. In the present trial, 25 hyperlipidaemic men with symptomatic femoral atherosclerosis underwent biplanar femoral arteriography at baseline. They were then randomised into treat- ment and usual-care groups: treatment was in- dividualized, comprising a lipid-lowering diet with cholestyramine, nicotinic acid or clofibrate as appropriate for the lipoprotein disorder. Mean cholesterol and triglyceride levels were 19 Vo and 37 Vo lower in the treatment group. Arteriography was repeated after a mean period of 19 months. With attention to blinding of ob- servers, changes in arteriograms were quanti- tated using computerised image analysis and visual methods, and expressed both by patient and by arterial segment. All end-points were in conformity and showed a lower rate of progres- sion of arterial disease in the treatment group, and a higher frequency of segmental regression in treated patients. In this small trial of patients with functionally-significant atherosclerosis, ef- fective treatment of hyperlipidaemia favourably affected the course of the arterial disease.

Key words: atherosclerosis, hyperlipidaemia, image analysis, progression, regression

Abbreviations: HDL = high density lipoprotein, LDL = low density lipoprotein, VLDL = very low density lipoprotein

Two complementary approaches have been employed to evaluate the effects of risk factor reduction on cardiovascular disease. In the past 15 years numerous controlled trials have been reported in which coronary heart disease mor- tality and morbidity have been the end points (1). With progress in controlled trial design it has become clear that many of these trials have been of marginal statistical power, and studies of adequate size such as the Lipid Research Clinics cholestyramine trial (2) have been re- markably costly. The second approach, emerg- ing in the past 7 years, depends on the direct measurement of atherosclerotic narrowing of the arterial lumen and has been made possible by recent advances in angiographic technique and, especially, in computerized image analysis (3). Such is the sensitivity of this methodology that far smaller trials appear to be feasible.

Each experimental design has its advantages. Assessment of clinical end-points provides in- formation of the greatest interest to the prac- ticing doctor, and the size of such trials permits

54 B. Lewis

evaluation of untoward effects of treatment. However the strength of such trials, a function of the number of participants, is in a sense also a weakness; there is a clear tendency for the ef- fectiveness of the intervention under investiga- tion to be inversely related to the size of the study. While dietary intervention in some smaller trials has led to a fall of 12-16 Yo in plasma cholesterol ( l ) , the reduction in recent large- scale trials using diet or drug treatment has been in the range 1-9 Yo (2, 4). There is clear evi- dence of a dose-response relationship between the fall in cholesterol levels and the reduction in coronary heart disease events ( 1 , 2, 5). Trials based upon arteriographic end-points, because of the smaller sample size required, may permit more effectively-supervised intervention (6). They also potentially offer insights into the effect of treatment on the pathological process underlying the clinical events, though arterio- graphic methods have the limitation that they provide information about luminal encroach- ment rather than data on the arterial wall. A comparison of various approaches to the assess- ment of the course of atherosclerosis has been made in a recent authoritative review (7).

In 1977 we commenced a randomised con- trolled trial to assess the results of effective treatment of hyperlipidaemia on atherosclerosis in man. For ethical and other reasons the study was confined to patients with symptomatic dis- ease; and because of the impressive demonstra- tion by Blankenhorn et a1 (3) that femoral atherosclerotic lesions could be quantified ar- teriographically we elected to study the femoral artery, a site of predilection for atherosclerosis. The trial is best regarded as a feasibility study, for we had at that time no basis for estimating sample size; because of rigorous exclusion char- acteristics the number of participants was rela- tively small. In view of prior evidence (reviewed in ref 6) that atherosclerosis, though generally progressive, can regress even in the absence of treatment of risk factors (at least over periods of 1-2 years), a controlled trial design was clearly mandatory; our own findings showed

that localised regression occurred in 15 Yo of femoral artery segments in patients receiving no lipid-lowering treatment.

DESIGN

Patients aged < 65 years with intermittent clau- dication due to femoral atherosclerosis were eli- gible if mean plasma cholesterol level was > 6.5 mmol/l and/or mean plasma triglyceride level was > 1.8 mmol/l. Fully informed written con- sent was obtained, and patients were random- ised into an intervention group and a usual care group. Patients requiring surgery, i.e. those with rest pain, skin ischaemia or severely limited claudication distance, and those with diabetes, hypertension, arteritis and major organ disease, (16 of 41 potential participants) were excluded. Matching was assessed for age, duration of symptoms, arteriographic severity, levels of cholesterol, triglyceride, low density lipoprotein (LDL) cholesterol, high density lipoprotein (HDL) cholesterol, very low density lipoprotein (VLDL) cholesterol, cigarette consumption, blood pressure and haematocrit. The two groups were satisfactorily matched, and there were no significant differences between the 13 patients in the treatment group and the 12 in the usual care group.

Usual care, provided to participants in both groups, included antismoking counselling, cor- rection of overweight and supervision directed to the ischaemic limb. The patients in the inter- vention group received instruction in a fat- modified lipid-lowering diet (30 Yo energy de- rived from fat, polyunsaturated: saturated fatty acid ratio 1.0) together with medication as ap- propriate for the lipoprotein disorder. Patients with elevation of LDL cholesterol level (n = 8) received cholestyramine 12-24 g/day, and five of them also received nicotinic acid 3-6 g/day. Those with increased levels of VLDL triglyceride received nicotinic acid in the same dose range, and those with Type I11 hyperlipoproteinaemia were treated with clofibrate 1.5-2 g/day.

Controlled Trial of Treatment of Hyperlipidaemia 55

One patient, in the intervention group, with- drew from the study.

Femoral arteriography was performed at baseline and repeated an average of 19 months later. The technique which had certain unique features was developed by Dr AB Ayers and Dr RGM Duffield. Biplanar arteriography was employed, to provide greater information in analysing eccentric stenoses and in estimating edge irregularity than conventional uniplanar techniques, and to permit identification of the direction of the artery so avoiding errors in the estimation of cross-sectional areas from non- perpendicular Xray beams. By timing the expo- sure to late diastole changes in luminal volume due to the pulse wave were avoided, and homog- eneous dilution of contrast medium in blood was ensured. An optimised combination of ex- posure time (20 msec) and focal spot size was chosen to maximise definition.

Rigorous precautions were adopted to provide reproducible positioning of the !imb during both arteriograms. Magnification was mini- mised by the long tube-to-limb distance, a fur- ther factor enhancing definition; comparison of serial arteriograms was also facilitated, for the position of the artery in the limb had a minimal effect on magnification of the image. This ge- ometry, together with the use of a long cassette, potentially provided images of the entire length of the femoral artery, by contrast with other procedures that permit visualization of a com- paratively short segment of the vessel.

Image analysis was performed by procedures developed by Drs JNH Brunt and ACF Col- Chester together with Dr Duffield, at the Uni- versity of Manchester Department of Biophys- ics. The technique, which empIoys the Joyce- Loebl Magiscan I1 image analyser, has been described (6). A video camera generates a signal which is digitized and logarithmically trans- formed before storage in the memory. Trans- verse density profiles are generated, and an index of edge irregularity derived. In addition, arteriographic appearances were visually quan- tified in the magnified image, using Vernier cali-

pers to measure cross-sectional diameters and cross-sectional areas of atheromatous plaques seen in profile. All end-point measurements were made by observers who were blinded to the treatment status of the patients.

RESULTS

Major findings have recently been reported by us (6 ) . The intervention was successful in pro- ducing a substantial reduction in elevated levels of plasma cholesterol, LDL cholesterol and tri- glyceride, respectively 19, 24 and 37 To lower in the treatment group (Table I) . HDL cholesterol was 41 To higher in the treated patients. The magnitude of these responses is likely to be at- tributable to individualization of the treatment regime, close supervision in a clinical setting and use of multiple therapy as described.

The magnitude and direction of change in edge irregularity index showed significant rela-

Table I. Mean risk factor levels during the trial.

Treatment Usual care

Cigarettes per day 16 14 Blood pressure 145/87 1501’86 LDL cholesterol, mmol/l 3.9** 5.1 HDL cholesterol, mmol/l 1.5** 1.1 VLDL cholesterol, mmol/l 0.60** 1.3

** P<0.01

Table 11. Correlations between changes in edge irregularity index during study and mean plas- ma levels of lipids and lipoproteins (mmol/l), n = 24.

Cholesterol 0.48 LDL cholesterol 0.59* Triglyceride 0.02 HDL cholesterol -0.32 LDL/HDL cholesterol 0.54

* P < 0.05

56 B. Lewis

Table 111. Assessment of arteriograms by segment

Treatment Usual care P

Change in edge irregularity index 070 segments with decrease in e.i. index Segments showing progression Increase in plaque area, mm/y/segment

0.019 0.047 <0.05

33 15 0.05 10 of 144 21 of 156 < 0.01

0.58 1.72 0.06

tionships with mean plasma lipid and lipopro- tein levels during the period of observation. These relationships were similar in the interven- tion and usual care groups, hence these data were pooled (Table 11). A significant positive re- lationship was evident between LDL cholesterol level and progression of arterial disease as esti- mated by the change in edge irregularity index; though numerically-similar correlation coeffi- cients were obtained for total cholesterol level and for the ratio of LDL-to-HDL cholesterol, these escaped formal statistical significance. An inverse but non-significant correlation was present between progression of disease and HDL cholesterol level.

Worsening of disease, as assessed by an in- crease in edge irregularity index, occurred in 4 of 12 patients in the intervention group and 8 of 12 in the usual care group (P = 0.11). Though the two-fold difference did not achieve statisti- cal significance it was in the same direction as the differences between the groups in evolution of disease analysed per segment of artery. The two forms of analysis are complementary, for analysis by patient masks the fact that progres- sion and regression commonly coexisted in the same patient (in 11 of 24). Furthermore the evidence of regression of disease in 4 of 12 usual-care patients (i.e. with no anti-hyper- lipidaemic therapy) emphasizes that open un- controlled trials of atheroma regression may be inadmissible in that ))regression)) changes can occur in the absence of treatment of lipid risk factors.

Worsening of plaque area was also less pro- nounced in the treatment group; the increase in

mean plaque area per patient was 2.9-fold greater in the usual-care group (P < 0.05).

Arteriograms were also analysed by 1 cm seg- ments visually and by 2.5 cm segments using image analysis (Table 111). The mean increase in edge irregularity index in the intervention group was 40 Vo as great as that in the usual care group, and 33 Vo of segments showed regres- sion in the former group, 15 Vo in the latter, by the same criterion. Visual end-points showed approximately one-third as much progression in the treated groups as in the usual care group.

DISCUSSION

This is the first reported study in which a ran- domised controlled trial design was employed to test the effect of reduction of hyperlipidaemia on atherosclerosis, using quantitative arterio- graphy as an end point. The feasibility of this approach has been established; and despite the small sample size and relatively short duration evidence was obtained that the intervention favourably influenced all measured criteria of atherosclerosis progression.

The trial design was such as to reflect clinical practice, and optimal treatment of the several forms of hyperlipidaemia was achieved by indi- vidualising the drug management, and by fre- quent clinic visits; thus the choice of drug was that indicated by' the lipoprotein disorder, and the dose, as well as the introduction of a second drug were determined by the initial response to treatment. Hence the study was a test of reduc- tion of hyperlipidaemia and not a trial of any

Controlled Trial of Treatment of Hyperlipidaemia 57

single mode of therapy. While observer-blind- ing was rigorously maintained the design did not permit blinding of the clinicians in charge of patients.

The various visual and automated criteria for assessing the arteriograms yielded concordant results. In this series of patients with extensive, functionally significant atherosclerosis and a mean age of 55 years, the overall results suggest that effective treatment of hyperlipidaemia may lead to a major reduction in the rate of progres- sion of the disease, within a relatively short treatment period.

The extent to which the present findings on the superficial femoral artery are applicable to coronary and other vessels is uncertain and is the subject of current studies employing similar methodology.

The trial was supported by a grant from Bris- to1 Myers Company Limited.

REFERENCES

1. Mann JI, Marr JW. Coronary heart disease preven- tion-trials of diets to control hyperlipidaemia. In: Miller NE, Lewis B (eds). Lipoproteins, Athero-

2.

3.

4.

5 .

6.

7.

sclerosis and Coronary Heart Disease, ch. 12. Am- sterdam: Elsevier, 1981. Lipid Research Clinics Program. The Lipid Re- search Clinics Coronary Primary Prevention Trial results. 1. JAMA 1984; 251: 351-64. Crawford DW, Brooks SH, Selzer SH, Barndt R, Beckenbach S . Computer densitometry for angio- graphic assessment of arterial cholesterol content and gross pathology in human atherosclerosis. J Lab Clin Med 1977; 88: 378-92. WHO European Collaborative Group. Multifac- torial trial in the prevention of coronary heart dis- ease: 3. Europ Heart J 1983; 4: 141-47. Lewis B. The LDL theory and HDL hypothesis. In: Diet and Drugs in Atherosclerosis, eds G Noseda, B Lewis, R Paoletti, p 1-8. New York: Raven, 1980. Duffield RGM, Lewis B, Miller NE, Jamieson CW, Brunt JNH, Colchester ACF. Treatment of hyper- lipidaemia retards progression of symptomatic fem- oral atherosclerosis. Lancet 1983; ii: 639-42. Bond MG, Insull W Jr, Glagov S, Chandler AB, Cornhill JF (eds). Clinical Diagnosis of Athero- sclerosis. New York: Springer, 1983.

Correspondence address: Professor Barry Lewis Department of Chemical Pathology and Metabolic Disorders St Thomas’ Hospital London SEl 7EH, UK.