Methods

Top Methods Conclusion Author & Article Info References

We did a MEDLINE search of the English-language literature (1966 to January 1994) using the keywords atherosclerosis progression, atherosclerosis regression, lipid reduction therapy, and coronary angiography. Articles were selected if they included data on the effects of cholesterol reduction (by diet, lifestyle modification, medications, or surgery) on angiographic coronary artery disease or if they described animal models of atherosclerosis progression and regression. The following information was extracted: study population, treatment strategy, duration of therapy and follow-up, and reported changes in luminal dimensions on serial coronary angiography.

Studies of Coronary Atherosclerosis Regression

The major lipid reduction studies and the angiographic findings are summarized in Table 1. Despite the attempts of these studies to show a beneficial effect of different lipid-lowering treatment strategies on atherosclerosis regression, investigators observed only modest effects on luminal calibers (<0.3-mm change in absolute luminal dimension or 10% in diameter stenosis) using both qualitative and quantitative angiographic methods. Furthermore, no consensus was reached in these trials on how to accurately assess angiographic progression or regression of atherosclerosis. Using various definitions of progression and regression, four of these studies [9-12] have shown that atherosclerosis can be stabilized, five [8, 13-16] have shown that atherosclerosis is reversible, and one [13] has shown that new lesion formation can be reduced. These studies also have shown that changes in lesion morphology appear to be a continuum, with lesion progression, stability, and regression often occurring in different lesions in the same patient [17].

In the Leiden Intervention Trial [10], plasma lipid levels were modified by diet alone. A modest reduction (10.1%) in the total cholesterol value was achieved in 39 patients over the 2-year study period. Although overall progression was noted in 307 lesions (percent diameter stenosis by computer-assisted angiographic quantitation, 44.1% ±31.6% at baseline and 48.6% ±30.9% at follow-up), 18 patients had no worsening of lesion percent diameter stenosis. Notably, the total cholesterol level and the total cholesterol/high-density lipoprotein (HDL) cholesterol ratio were lower in patients without lesion progression than in patients with lesion progression (absolute change in mean vessel diameter, 0.002 ±0.23 mm in patients with a total cholesterol/HDL cholesterol ratio <6.9 compared with 0.237 ±0.32 mm in patients with a total cholesterol/HDL cholesterol ratio >6.9; P = 0.001).

Lipid Reduction by Combined Diet and Lifestyle Changes

In the Lifestyle Heart Trial [8], 41 patients were randomly assigned to either a change in both diet and lifestyle or no intervention and followed for 1 year. In the treatment group, total cholesterol was reduced by 24%, and low-density lipoprotein (LDL) cholesterol was reduced by 37%. Using computer-assisted quantitative angiographic methods, the investigators found 105 lesions at baseline angiography in the 22 patients receiving intervention and 95 lesions in the 19 controls. During the study period, the intervention group showed an overall decrease in percent diameter stenosis (40.4% ±16.9% to 37.8% ±16.5%), whereas controls showed an overall increase in percent diameter stenosis (42.7% ±15.5% to 46.1% ±18.5%; P = 0.001). In the intervention group, progression was noted in 4 (18%) patients, and regression was noted in 18 (82%); conversely, in the control group, no change or progression was noted in 11 patients (58%), and regression was found in 8 (42%).

In a similar study of 113 patients randomly assigned to usual care or diet and exercise (but no drug therapy), body weight decreased by 5%, total cholesterol decreased by 10%, triglycerides decreased by 24%, and HDL cholesterol increased by 3% in the treatment group [12]. Paired coronary arteriograms obtained 12 months apart were analyzed using computer-assisted methods, and a change in lesions of more than 10% in diameter stenosis or of 0.18 mm in minimal luminal diameter was scored as lesion progression (+1), lesion regression ( –1),or no change^[0]. Segment scores were added to show an overall net change. In the treatment group, 30% of patients showed regression, 50% of patients had no change, and 20% of patients had progression. In the control group, 42% of patients showed progression, 54% had no change, and 4% showed regression (P = 0.03). The average change in minimal luminal diameter for the 122 lesions was 0.0 ±0.38 mm in the treatment group and –0.13±0.45 mm in the control group (P = 0.01).

Lipid Reduction by Drug Therapy

The National Heart, Lung, and Blood Institute study tested the effect of cholestyramine on coronary atherosclerosis in 116 hyperlipidemic patients; cholesterol levels in treated patients were reduced by 26% [11]. Paired baseline and 5-year cineangiograms were analyzed subjectively by a majority consensus of three observers. Patients were classified as showing definite lesion progression (at least one lesion with definite progression, no lesion with regression), probable lesion progression (at least one lesion with probable progression, no lesion with either definite regression or definite progression), probable lesion regression (at least one lesion with probable regression, no lesion with either definite regression or definite progression), definite lesion regression (at least one lesion with definite regression, no lesion with progression), mixed lesion response (both progression and regression in the same patient), and no lesion change. Using these criteria, the study found probable lesion progression in 32% of treated patients and 49% of control patients and definite progression in 25% and 35% of treated and control patients, respectively. Because many patients showed a mixed lesion response, statistically significant differences in the primary analysis were not observed. In the Cholesterol Lowering Atherosclerosis Study [13], 162 men receiving coronary bypass grafting were randomly assigned to treatment with colestipol and nicotinic acid or to placebo. Reductions of 26% in total cholesterol and 43% in LDL cholesterol and an increase of 37% in HDL cholesterol were observed in treated patients after 2 years of follow-up. In addition, new lesions developed in 10% of treated patients and in 22% of placebo recipients during the 2-year period. By 4 years [18], the frequency of new lesion development was 14% in treated patients and 38% in placebo recipients (P = 0.001). Furthermore, 18% of lesions in the treated group had regressed compared with 6% of lesions in the placebo group (P = 0.04); similarly, 52% of lesions in the treated group had not progressed compared with 15% of lesions in the placebo group (P < 0.001).

In the Familial Atherosclerosis Treatment Study [7], 146 patients with angiographic evidence of coronary artery atherosclerosis and at least one segment of significant stenosis, a familial history of coronary artery disease, and elevated lipid concentrations were randomly assigned to conservative therapy, treatment with a combination of lovastatin and colestipol, or treatment with niacin and colestipol. In the two treatment groups, LDL concentrations decreased (46% in the lovastatin-colestipol group and 32% in the niacin-colestipol group) and HDL concentrations increased (15% and 43%, respectively) significantly. Paired angiograms were analyzed using the worst stenosis in each of nine standard proximal segments of the three major epicardial coronary arteries. The average percent diameter stenosis of the worst lesion in all segments was calculated, and regression or progression was expressed as the change in this value between the baseline and follow-up angiograms. In the placebo group, the mean percent diameter stenosis progressed by +2.1% ±3.9%; in the niacin-colestipol group, the mean percent diameter stenosis regressed –0.9%±3.0%; in the lovastatin-colestipol group, the mean percent diameter stenosis regressed –0.7%±5.3% (P = 0.003). Similarly, the mean change in minimal lesion diameter was –0.05±0.14 mm in the placebo group, +0.035 ±0.13 mm in the niacin-colestipol group, and +0.012 ±0.16 mm in the lovastatin-colestipol group (P = 0.01). In the placebo group, lesion progression was noted in 46% of lesions, lesion regression in 11% of lesions, and no change in 43% of lesions. In the niacin-colestipol group, lesion progression was noted in 25% of lesions, regression in 39% of lesions, and no change in 36% of lesions. In the lovastatin-colestipol group, lesion progression was noted in 21% of lesions, lesion regression in 32% of lesions, and no change in 47% of lesions. Only 5 of the 120 patients had a mixture of segment progression and regression.

Lipid Reduction by Diet plus Medication

In a trial of 72 patients with heterozygous familial hypercholesterolemia [15], patients were randomly assigned either to diet and low-dose colestipol or to diet, high-dose colestipol, and niacin (switching to lovastatin when it became available). Paired angiograms obtained 26 months apart were analyzed using criteria similar to those used in the Familial Atherosclerosis Treatment Study [7]. Notably, only three patients had objective evidence of coronary artery disease before the initial angiogram. In the low-dose colestipol group, LDL cholesterol levels decreased 38.1%, serum triglycerides decreased 18.9%, and HDL cholesterol levels increased 28%. In the 457 lesions analyzed, mean change in percent area stenosis was +0.80% ±5.07% in the low-dose colestipol group and –1.53%±4.34% in the high-dose colestipol group (P = 0.039). Moreover, in the low-dose colestipol group, 13 patients had definite lesion progression and 4 had definite lesion regression; in the high-dose colestipol group, 8 had lesion progression and 13 had lesion regression (P = 0.06).

In the St. Thomas' Atherosclerosis Regression Study [16], 90 men with coronary artery disease and mild hypercholesterolemia were randomly assigned to usual care, diet alone, or diet and cholestyramine. Patients assigned to diet alone showed a 14.2% decrease in total cholesterol and a 16.2% decrease in LDL. Those assigned to diet and cholestyramine showed a 25.3% decrease in total cholesterol and a 35.7% decrease in LDL. Paired angiograms obtained 39 months apart were analyzed using computer-assisted methods, and the mean lesion diameter of 13 arterial segments was obtained. Notably, the mean arterial diameter progressed in the usual care group by 0.201 ±0.062 mm, regressed by 0.003 ±0.087 mm in the diet alone group, and regressed by 0.103 ±0.051 mm in the diet plus cholestyramine group (P = 0.012). Overall progression was less frequent in the diet group (15%) and the diet plus cholestyramine group (12%) than in the control group (46%, P = 0.01). Similarly, the proportion of patients who showed overall improvement in mean luminal diameter was greater in the diet group (38%) and in the diet plus cholestyramine group (33%) than in the control group (4%, P = 0.01).

Lipid Reduction by Surgical Treatment

The Program on the Surgical Control of the Hyperlipidemias study [9] compared 421 patients treated by partial ileal bypass with 417 patients in a randomly selected control group. Over the 9.7-year follow-up period, an average reduction in total cholesterol of 23.2%, a reduction in LDL cholesterol of 37.7%, and an increase in HDL of 4.3% was seen in the treated group. Paired angiograms were graded by two observers according to the semi-quantitative criteria of the Cholesterol Lowering Atherosclerosis Study [18]. During the follow-up period, 38.9% of lesions remained static in the treated group compared with 11.3% of lesions in the control group (P < 0.001). Similarly, marked progression was noted in only 4.2% of lesions in the treated group, whereas it was present in 26.2% of lesions in the control group (P < 0.00001). Lesion regression occurred less commonly; it was seen in 6.4% of lesions in the treated group and in 3.8% of lesions in the control group (P = 0.51).

Limitations of Coronary Angiography in Assessing Atherosclerosis Regression

Several reports have addressed additional inherent limitations of current angiographic methods used in regression studies [19, 20], including technical difficulties of reproducible data acquisition on serial studies, intra- and interobserver variability of interpreting the acquired data, and inability of angiograms to provide insight into the mechanism of atherosclerosis progression or regression (Table 2).

Data Acquisition

Careful attention to the details of cineangiographic acquisition on serial studies is most critical to a successful regression study [19, 20]. These technical details include control of factors such as same magnification and camera angle for image acquisition, attainment of maximal coronary vasomotor tone using intracoronary nitroglycerin, minimal vessel motion artifact with consistent cardiac (end-diastolic) or respiratory (during held inspiration) phase, and geometric considerations such as minimal vessel overlap or foreshortening.

Data Interpretation

Even if the serial angiograms were consistently and accurately obtained, the data interpretation may be hampered by intra- and interobserver variability [21-24], underestimation of the extent of atherosclerosis from diffuse nature of coronary atherosclerosis [25-41], and misinterpretation of the initial compensatory enlargement as atherosclerosis regression [42].

An important explanation for the discrepancy between the extent of atherosclerosis and lack of angiographic luminal compromise is that dilatation of the wall of the diseased arterial segment occurs to compensate for the accumulation of atherosclerotic plaque. This compensatory mechanism has been noted in primates with diet-induced atherosclerosis [43] and in persons with coronary atherosclerosis, in whom it has been shown pathologically [44, 45], by high-frequency epicardial echocardiography [46], and by intravascular ultrasound [37, 47, 48]. Cross-sectional arterial enlargement occurs in direct relation to the cross-sectional area of plaque accumulated both concentricly and eccentricly; initial eccentric plaque accumulation [47] is followed by eccentric dilatation, whereas concentric plaque accumulation is accompanied by concentric dilatation. Luminal compromise detectable by arteriography is delayed until the atherosclerotic lesion occupies more than 40% [44] to 50% [45] of the potential area within the internal elastic lamina. This phenomenon has occurred in all major epicardial coronary arteries [45]. Thus, both the plaque size and the extent of disease are underestimated using standard angiographic methods that visualize the lumen alone. Moreover, small plaques can be associated with an ectatic, larger-than-normal lumen area, suggesting overcompensation at the initial phase of the disease process [42], and may be misinterpreted as atherosclerosis regression.

Also, because relative measurements of luminal narrowing are based on the determination of absolute diameters of both the apparently normal reference segment and the most severely narrowed segment, and based on the assumption that the normal segment is free of significant atherosclerotic disease, the use of relative measurements of plaque regression (for example, percent diameter stenosis) may underestimate the degree of plaque change that occurs as a result of lipid-lowering therapy. Pathologic [25-34], epicardial echocardiographic [35], and intravascular [36-41] ultrasound studies have shown that significant diffuse coronary atherosclerosis often is present when angiograms reveal only discrete lesions. Absolute measurements in millimeters (such as mean and minimal lesion diameter) may be better markers of progression or regression of atherosclerosis [25-27, 49-52] than relative measurements of percent luminal narrowing, although these quantitative methods require extensive efforts to obtain accurate image calibration.

Alteration in baseline coronary tone also may result in substantial interstudy variability; some of these differences can be minimized, although not entirely eliminated, by use of periprocedural nitroglycerin. Atherosclerotic vessels react abnormally to endothelial mediated vasodilatation [47]. Concomitant therapy that improved endothelial function or acts as a vasodilator can increase luminal size and give the spurious impression of a reduction in the mass of atherosclerotic tissue. For example, hyperlipidemia can have a direct effect on endothelial function more widespread than just within diseased segments of the artery [48]. Studies in animals have shown that changes in coronary vasomotor tone (and luminal diameters) accompany induced changes in plasma lipid levels [49]. Similarly, changes in coronary tone can result from antianginal drug therapy.

In addition to the difficulties of correctly interpreting data, the reported variabilities in the serial measurements of percent diameter stenosis and minimal luminal diameter and modest changes in their values over time may mandate substantial sample sizes to show significant plaque regression using angiography. For instance, the variations in arterial dimension may range from 5% to 10% simply from phasic pulsations, equivalent to reported annual atherosclerosis progression or regression [20]. Likewise, the standard deviation for mean luminal diameter for paired angiograms can be as great as the reported changes in this measurement during a progression study [50]. Finally, the presence of multiple lesions in the same patient and his or her independent course of progression or regression [53] may further complicate the interpretation of the overall effect of the therapy.

Mechanistic Aspects

Even if the serial angiograms were accurately obtained and interpreted regarding luminal compromise, angiography still cannot assess lesion composition and does not measure atherosclerosis. Lesion composition may be an important factor influencing the rate of atherosclerosis progression and regression. The clinically and hemodynamically significant lesion of human coronary atherosclerosis is the fibrocalcific atheromatous plaque. The connective tissue changes leading to development of fibrous plaques begin in the second decade of life and are present in virtually all persons older than 35 years in Western societies; however, several decades of slow growth are usually required before these changes produce clinical symptoms [54]. Animal models of atherosclerosis progression and regression show that cholesterol is deposited into and removed from plaques much faster than collagen [55-58]. Collagenous accumulations of intimal plaques induced by diet in monkeys do not regress even after 4 years of drastic plasma cholesterol reduction [57]. If the lessons of the animal models are correct and applicable to humans, significant regression of lesions causing high-grade obstruction is unlikely other than over a long time [59].

Conversely, atherosclerosis in proximal reference vessels or in diffuse disease of modest severity typically contains less fibrous tissue and calcium and more lipid. The propensity of plaques to become unstable and undergo fissuring and thrombosis (and therefore rapid progression) or to become unstable and cause acute coronary events [60, 61] is related directly to the amount of extracellular lipid contained [62-64]. Thus, the optimal approach to halting progression of atherosclerosis may be lesion suppression, stabilization, and regression early in the disease process [17]. Angiography cannot distinguish among any of these atherosclerotic components and cannot provide mechanistic insight into atherosclerotic regression or progression.

Conclusion

Top Methods Conclusion Author & Article Info References

Although serial coronary angiography can detect changes in luminal dimensions over time, this method has many limitations, including difficulties with consistent data acquisition on serial studies, accurate interpretation of the data, and limited value in understanding the pathogenesis of atherosclerosis. Even if these shortcomings are overcome, the changes noted over long-term follow-up with lipid reduction therapies have been modest, usually within the standard deviation of the repeated measurements. Thus, new approaches, such as intravascular ultrasound [65-67], may be needed to better comprehend the dimensional changes as well as the pathogenesis of these changes.

Future Directions

Intravascular ultrasound [65-67] is a recent development that permits detailed, high-quality, cross-sectional pictures of the coronary arteries in vivo. Unlike coronary arteriography, intravascular ultrasound may enable quantification of the amount of atherosclerotic plaque [36], determination of composition of atherosclerotic plaque [36, 68], detection of significant atherosclerosis in angiographically normal reference coronary arterial segments [36-40], and accurate detection of atherosclerosis progression or regression caused by the motorized pullback and its ability to precisely determine the original site of interest [69]. Thus, intravascular ultrasound has the potential for answering many of the questions regarding atherosclerosis progression or regression not currently possible with coronary arteriography. Another imaging method with limited utility for detecting atherosclerosis progression or regression is coronary angioscopy [70], which can assess the morphologic features of the intimal surface but cannot evaluate the pathologic characteristics and plaque composition of the subintimal layer. Finally, noninvasive imaging such as ultrafast computed tomography [71] or magnetic resonance imaging [72], which can detect atherosclerosis, would be preferable for longitudinal studies, but at present cannot quantify the plaque burden.