Colour-Doppler ultrasonography of carotid vessels in patients treated with antiretroviral therapy: a comparative study

Objectives: To evaluate the correlation between antiretroviral therapy (ART) and lesions of the carotid vessels using an ultrasound colour-Doppler technique.

Design: A total of 293 HIV-1 infected patients underwent epiaortic vessel ultrasonography: 105 on treatment with protease inhibitors (PI) (group I), 125 PI-naive patients treated with a non-nucleoside reverse transcriptase inhibitor-including regimen (group II), and 63 patients treated with two nucleoside reverse transcriptase inhibitors or naive to ART (group III).

Methods: Intima characteristics, pulsation and resistance indexes, and minimal, peak and mean speed were evaluated using a colour power doppler. Atherosclerotic plaques were described. Independent risk factors and values for glycaemia, cholesterolaemia and triglyceridaemia were considered. Statistical analysis included the Wilcoxon tests, the χ² test, the Cochran Armitage trend test and the Mantel-Haenszel test and, when necessary, logistic regression analysis.

Results: Of the 150 group I patients, 55 (52.4%) presented acquired lesions of the vascular wall at ultrasonography, whereas similar lesions were found in 19 out of 125 (15.2%) patients in group II and in nine of 63 (14.3%) in group III. ART, age, smoking and CD4 T-cell count were the main predictive risk factors for vascular lesions. However, the highest significance was with the use of PI.

Conclusions: These data confirm the higher prevalence of premature carotid lesions in the PI-treated patients. A periodic ultrasonographic study of the vascular wall should be included in the follow-up of HIV infected patients.

Introduction

Treatment with protease inhibitors (PI) have been associated with the onset of metabolic disturbances, such as insulin resistance and lipodystrophy [^1-4]. Moreover, various vascular complications (stroke, angina pectoris, myocardial infarctions) have been described in patients subjected to prolonged treatment with this class of antiretroviral drugs. These observations prompted the hypothesis that PI can contribute to atherosclerotic damage of the vascular wall [^5-8], although the actual role of these molecules in increasing the risk of cardio- and cerebro-vascular events in HIV-positive patients still remains a highly controversial issue.

Ultrasound colour-Doppler is a well-established method for measuring the degree of atherosclerosis. Furthermore, the examination can be conducted early in the development of atherosclerosis and can be used to monitor the progression of the lesions.

To obtain a more specific correlation between antiretroviral therapy (ART) and lesions of the carotid vessels, we evaluated the presence of premature atherosclerotic lesions and alterations in the flow and the wall of the common and internal carotid arteries using a ultrasound colour-doppler technique in PI-treated patients, in PI-naive patients treated with a non- nucleoside reverse transcriptase inhibitor (NNRTI)-including regimen, and in patients treated with two nucleoside reverse transcriptase inhibitors (NRTI) or naive to ART.

Patients and methods

A total of 293 HIV-1 patients including 105 in treatment with PI (group I), 125 PI-naive patients treated with a NNRTI-including regimen (group II) and 63 patients treated with two NRTI or naive to ART (group III) were evaluated. All of the treated patients had been in therapy for at least 12 months. Subjects affected with hypertension were excluded from the study. The main patient characteristics are reported in Table 1.

Table 1

HIV-1 seropositivity had been documented for a median period of 7 years (range, 1-18 years). In group I, all patients were treated for a median period of 26 months (range, 12-77 months) with combination therapies including indinavir (50 patients), nelfinavir (27 patients), saquinavir (six patients), ritonavir (four patients), amprenavir (one patient). Ten patients were treated with ritonavir plus saquinavir, four with nelfinavir plus saquinavir, two with indinavir plus ritonavir and one with nelfinavir plus ritonavir. In group II, 71 subjects were treated with combination regimens including nevirapine, 52 with efavirenz, and two with delavirdine. The median treatment period was 24 months (range,12-68 months). In group III, 49 patients were naive to ART and 14 were in treatment with two NRTI. All treated patients were in therapy for a median period of 23 months (range, 12-84 months).

All patients were subjected to ultrasonography of the epi-aortic vessels using an AU 5 ESAOTE power colour-Doppler with 7.5 mgHz probes. Characteristics of the intima, pulsation index, resistance index, minimal speed, peak speed, and mean speed, were evaluated; an intima media thickness (IMT) of > 1 mm was considered to be pathological. Atherosclerotic plaques, if present, were described.

Ultrasonography was performed by physicians specifically trained on carotid vessels (L.A., P.F., F.S., R.G., M.A.) with 10 years of experience in the ultrasound colour-Doppler technique and at least 1000 documented epi-aortic examinations; they were blinded to the patient's treatment history and status. Moreover, during the study, periodical meetings were held using filmed reports aimed at the comparison and standardization of the technique.

Patients were submitted to the investigation in a supine position after at least 10 min of acclimatization in a comfortable room. Patients were informed that the investigation was non-invasive. The common carotid, the bifurcation and at least the first 2 cm of the internal and external carotid vessels were examined in the short and long axis during the tele-diastolic phase (T wave of the electrocardiogram). During the investigation the head of the patient was extra-rotated from the opposite side. The percentage of stenosis was always quantified by calculating the stenosis areas in the short axis using a strong magnification. This was intended to distinguish correctly the real lumen from plaques markedly hypoechoic with the colour or the power Doppler. The speed measurements were performed at a 45-60° inclination with respect to the lumen. The morphological investigation of the plaque was performed using both ultrasonography and the ultrasound power colour-Doppler in order to better characterize the profile of the plaque and the IMT [^9-12]. None of the patients was subjected to angiography because of poor patient compliance to an invasive procedure. Moreover, previously published literature confirms that the technique used is the 'gold standard' for investigation of carotid plaques [¹³].

Risk factors for cardiovascular diseases were evaluated, such as familial history (angina, myocardial infarction, cerebral stroke, transitory ischemic attack), sedentary lifestyle (< 1 h/week of sport activity), cigarette smoking, alcohol consumption (> 80 g/day), active drug addiction, hyperglycaemia, hypercholesterolaemia, hypertriglyceridaemia; in addition, risk factors for HIV-1 infection including Centers for Disease Control and Prevention (CDC) stage, CD4 cell count, viral load and ART were assessed.

Differences among groups in terms of epidemiological characteristics and risk factors were evaluated by Wilcoxon tests and χ² test. The Cochran Armitage trend test and the Mantel-Haenszel test were performed when necessary. The probability of developing vascular lesions in general, or atherosclerotic plaques in particular, as a function of the factors analysed was evaluated by logistic regression models. The odds ratio (OR) and the relative 95% confidence intervals (CI) were evaluated for the variables which were significant by the logistic regression models.

Results

The distribution of the risk factors independent of HIV-1 infection is shown in Table 1.

In group I no abnormal findings at ultrasonography were observed in 50 of 105 patients (47.6%), whereas 55 patients (52.4%) demonstrated acquired lesions of the vascular wall including 25 patients with pathological IMT, 24 with pathological IMT plus atheromatous plaques and six with plaques. The median IMT was 1.2 mm for the right carotids (range, 1.01-2.47 mm) and 1.3 mm for the left carotids (range, 1.01-3.00 mm). In the patients with atheromatous plaques, the median percentage of stenosis of the vascular lumen was 42.9% for the right carotids (range 15-54) and 41.9 for the left carotids (range, 15-70%).

In group II, ultrasonography excluded abnormal pathologies in 106 of 125 patients (84.8%) while 19 (15.2%) showed acquired lesions of the vascular wall 10 of whom demonstrated increased IMT, three plaques and six increased IMT plus plaques. The median IMT value was 1.31 mm for the right carotids (range, 1.01-2.33 mm) and 1.36 mm for the left carotids (range, 1.01-2.08 mm). In the patients with plaques, the median percentage of stenosis of the vascular lumen was 35% for the right carotids (range, 25-66 %) and 38% for the left carotids (range, 25-52%).

Of the 63 patients in group III, 54 had no abnormal findings at ultrasonography (85.7%), whereas lesions were observed in nine (14.3%) including three patients with increased IMT, four with increased IMT plus plaques and two with plaques. The median IMT value was 1.24 mm for the right carotids (range, 1.02-1.4 mm) and 1.4 mm for the left carotids (range, 1.1-3.5 mm). In the patients with plaques, the median percentage of stenosis of the vascular lumen was 30% for the right carotids (range, 20-40%). Only one patient had a left stenosis (46.7%).

Regarding congenital abnormalities of the internal carotid arteries, in group I four kinkings were observed (two associated with increased IMT, one with increased IMT plus plaque); in group II seven kinkings were noted (one associated with increased IMT) and two in group III together with one coiling.

No alterations in the pulsation index, resistance index, minimal speed, peak speed, and mean speed were observed in the three groups and no patients had neurological symptoms.

From the statistical point of view, the logistic regression model evidenced that the risk of vascular lesions is significantly linked with type of therapy (Wald χ², 44.24; P < 0.0001), age (Wald χ², 14.78; P = 0.0001), cigarette smoking (Wald χ², 4.99; P = 0.025) and CD4 cell count (Wald χ², 6.88; P = 0.032), whereas the risk of lesions is significantly higher for patients with a CD4 cell count < 200 × 10⁶/l or between 200 × 10⁶/l and 500 × 10⁶/l compared to those with a CD4 cell count > 500 × 10⁶/l. The model expression and the OR are reported in Table 2.

Table 2

The other risk factors evaluated were not statistically significant, including the disease stage both in terms of interaction with therapy given the association between stage of disease and therapy (χ², 16.88; P = 0.002), and as a main effect. This result was reinforced by the Mantel-Haenszel test which demonstrated a strong relationship between vascular damage and use of PI during every stage of the disease (χ², 33.89; P = 0.0002); conversely, our findings indicated the absence of a relationship between vascular lesions and the CDC stage for the three therapeutic regimens (χ², 0.71; P = 0.4).

Evaluating the association between the single risk factors and the presence of vascular damage, the only two significant factors were the PI-based therapy (χ², 46.65; P = 0.0001) and hypertriglyceridaemia (χ², 6.68; P = 0.01). Nevertheless, as reported previously [¹⁴], the association between hypertriglyceridaemia and vascular lesions was no longer apparent when stratification according to therapy was performed, thus confirming the close relationship between hypertriglyceridaemia and PI-based therapy. Moreover, the hypercholesterolaemia findings were highly correlated to the use of PI (respectively χ², 5.3, P < 0.001 for hypertriglyceridaemia and χ², 17.15, P < 0.001 for hypercholesterolaemia).

When considering the presence of atheromasic plaques as a depending variable in the logistic regression model, the predictive risk factors were, once again, PI-based therapy and age. Interestingly, in this analysis the role of the CDC stage emerged; in fact, the regression model evidenced that patients in the CDC group B have a 2.5-fold higher odds of developing atheromasic plaques with respect to groups A and C. When the logistic regression model was applied to the subgroup of 83 patients with vascular lesions (pathological IMT and/or plaques), the risk of developing plaques was not increased by the presence of the other risk factors evaluated (data not shown).

Discussion

Various studies have hypothesized an increased risk of cerebro- and cardiovascular diseases in HIV-positive patients submitted to PI-based therapies. Many of the studies dealing with this issue are controversial and the role of these molecules in determining vascular damage still remains uncertain, although recent data from a large international trial has evidenced an increased relative risk for myocardial infarction during the first 7 years of highly active ART [¹⁵].

Ultrasound colour-Doppler, which incorporates the ultrasonographic and Doppler methods, is a safe, inexpensive and rapid technique which allows an accurate evaluation of the vascular vessel. Moreover, the images obtained can be highly magnified with the 'power' technique. Using this procedure, in a preliminary report of the PREVALEAT study conducted on 102 patients [¹⁴], we observed a higher than expected prevalence of premature carotid vessel lesions in the group treated with PI (52.7%) with respect to a cohort previously untreated with these drugs (14.9%), although PI and NNRTI-based regimens were not compared directly. In addition, we observed an overwhelming difference between the percentage of acquired lesions between healthy individuals (6.7%) and both seropositive groups, thus suggesting that HIV-1 positive patients have a much higher risk of endothelial damage which becomes quite remarkable in the group of patients treated with PI-containing regimens for prolonged periods of time.

Other studies have been performed using the same technique, but the results have been controversial and the number of patients scarce [^16-20]. Therefore, our PREVALET study was extended to include a larger patient cohort. To our knowledge, this is the largest study using the ultrasound colour-Doppler technique to evaluate HIV-positive patients. Moreover, in order to obtain a more specific correlation between ART and carotid vessel lesions, the PI-treated patients were compared with PI-naive patients treated with a NNRTI-including regimen and to patients treated with two NRTI or naive to ART.

The results of the present study largely confirm the previous data. In fact, a significantly higher percentage (52.4%) of acquired lesions of the epi-aortic vessels was observed in patients treated with the PI-based regimens (group I) as compared with those treated with NNRTI-based regimens (group II, 15.2%) and to naive patients or patients treated with two NRTI (group III, 14.3%). Moreover, no statistical differences emerged when comparing group II to group III.

When applying the logistic regression model, of the risk factors independent of HIV infection, age was found to be the most important with a 1.09 increase of the OR yearly (Table 2). This signifies that a 50-year-old patient could have a 2.5-fold odds for developing vascular diseases with respect to a 40-year-old patient. Interestingly, the CD4 cell count also seems to have a significant role in vascular damage, determining a higher risk for patients with CD4 between 200 × 10⁶/l and 500 × 10⁶/l (OR, 2.45; CI, 1.23-4.89).

When considering the presence of atheromasic plaques as a depending variable in the logistic regression model, once again the PI-based therapy was confirmed as the main predictive risk factor, followed by age. Another interesting point is the role of the CDC stage; in fact, the patients in CDC group B seem to have a greater liklihood of developing atheromasic plaques with respect to groups A and C (OR, 2.45; CI, 0.93-6.46).

As shown in the previous PREVALEAT report, stratification according to therapy demonstrates that hypertriglyceridaemia is related to PI-based therapy and does not seem to exert a direct role in the genesis of the vascular lesions observed.

It is noteworthy that drug abuse was not related to an increased risk of vascular lesions, nor was the viral load, thus suggesting that HIV per se does not exert direct damage on the endothelium of the arterial vessels.

In 14 of the 203 patients in the present study, ultrasonography revealed evidence of congenital changes (kinking or coiling). These alterations in the course of the carotid vessels should be monitored as they could be a predisposing factor for premature flow variations if associated with acquired lesions.

In conclusion, these data suggest that the risk of premature damage to the carotid wall in HIV-1 infected patients is complicated by factors such as cigarette smoking and age, but that PI-containing regimens appear to play a major role. However, it would seem that the vascular damage exerted by PI therapy, at least in its early onset, might not be mediated by hypercholesterolaemia or hyperglycaemia, but additional evidence is necessary to clarify the mechanism involved. Also the role of the CD4 cell count and CDC stage should be better investigated. In the light of these results, a periodic ultrasonographic study of the vascular wall should be included in the general follow-up of these patients. In the case of rapid and progressive worsening of acquired lesions, the shift toward a PI-sparing combination therapy would be an appropriate choice.

Acknowledgements

The authors are grateful to Ms Paulene Butts and Mr Felipe Torio for their assistance in the preparation of the manuscript.