| |
Sustained improvement of dyslipidaemia in HAART-treated patients replacing stavudine with tenofovir- 48 weeks results
|
| |
| |
AIDS: Volume 20(10) 26 June 2006 p 1407-1414
Llibre, Josep Ma; Domingo, Pereb; Palacios, Rosarioc; Santos, Jesusc; Perez-Elias, Maria Jd; Sanchez-de la Rosa, Rainele; Miralles, Celiaf; Antela, Antoniod; Moreno, Santiagod; the Lipo-Rec Study Group
From the aHospital de Calella, Barcelona, Spain
bHospital Sant Pau, Barcelona, Spain
cHospital Virgen de la Victoria, Malaga, Spain
dHospital Ramon y Cajal, Madrid, Spain
eGilead Sciences, Madrid, Spain
fHospital Xeral, Vigo, Spain.
Abstract
Objective: To describe the 12-month evolution of lipid profile in HIV-infected virologically suppressed patients substituting tenofovir for stavudine.
Design and methods: 'Recover' was a prospective, multicenter, switch study conducted at 120 HIV units across Spain designed to identify single nucleoside analogue substitution due to adverse events in real practice. Tenofovir substituted stavudine in 873 adult patients. No other substitutions were allowed. This lipid sub-study included 352 randomly recruited patients with complete follow-up and lipid parameters.
Main outcome measures: Changes in fasting levels of total cholesterol (TC), high and low-density lipoprotein cholesterol (HDL-C and LDL-C), and triglycerides (TG) at 48 weeks, and their cardiovascular risk (CVR) translation.
Results:
At 48 weeks, there was a sustained reduction in median TC (-17.5 mg/dl; P < 0.001), LDL-C (-8.1 mg/dl; P < 0.001), and TG (-35 mg/dl; P < 0.001). HDL-C remained roughly unchanged (-0.8 mg/dl).
Patients with baseline hyperlipidaemia showed greater reductions in LDL-C (-29 mg/dl; P < 0.001) and TG (-76 mg/dl; P < 0.001). The greatest TG reduction was observed in patients with severe hyper-TG (-266 mg/dl; P < 0.001).
The estimated 10-year CVR decreased in all patients (P < 0.001), and to a higher extent in patients with baseline hyperlipidaemia. There was a trend towards reduction according to the use of lipid-lowering agents (11.6% to 9,9%; P = non-significant).
Conclusions: The substitution of tenofovir for stavudine causes a sustained improvement of dyslipidaemia. The reduction, although modest, is robust and sustained over time, and significantly reduces the CVR. This switch strategy is safe and contributes to an improvement in the lipid profile, especially TG, in HAART-treated patients.
Introduction
Dyslipidemia has emerged as an important problem in HIV infected individuals receiving antiretroviral treatment [1-4]. It worsens the risk of coronary diseases up to 50% over 10 years and increases the treatment burden, the overall resource use, and the cost of care of HIV infection [1,5-7]. Recent studies have also confirmed an increased risk of cardiovascular events in patients receiving HAART [8-10].
The primary goals in dyslipidaemia treatment in HIV-infected patients are reductions of low-density lipoprotein cholesterol (LDL-C) and raised triglycerides (TG). Changes in lifestyle including diet, exercise programs and smoking cessation, are the first line of treatment. However, although universally recommended, their results in HIV infection have been inconsistent and rarely reach the LDL-C and TG goals defined in the National Cholesterol Education Program (NCEP) guidelines [4,11,12].
There are two non-exclusive options for drug therapy: using lipid-lowering drugs and switching antiretroviral agents. Drugs within each class of antiretroviral agents have different impacts on lipid profiles. Protease inhibitors (PI) have been extensively linked to dyslipidemia and PI-replacing strategies have demonstrated significant improvements in LDL-C and TG in randomized trials [13-15].
The impact on lipid profile of nucleoside analogue reverse transcriptase inhibitor (NRTI) replacement has rarely been evaluated in spite of evidence showing a linkage between dyslipidaemia and some NRTI, especially stavudine [16,17]. The results of most studies have been inconclusive because studies were actually designed to ascertain the evolution of lipoatrophy [18,19]. Stavudine has been associated wit elevations in non-fasting LDL-C and TG, especially when combined with didanosine [17,20-22]. Unlike tenofovir, a contribution of stavudine to high-density lipoprotein cholesterol (HDL-C), LDL-C and TG abnormalities has been recently seen in the GS 903 trial [23,24].
The Recover Study is the first large-scale trial designed to evaluate the most frequently occurring toxicities in clinical practice causing a NRTI replacement by tenofovir [25].
The present analysis describes the 48-week evolution of complete lipid parameters in patients substituting tenofovir for stavudine.
Results
Patient characteristics
Baseline characteristics of the 352 patients are shown in Table 1. They had a high mean stavudine exposure and the main reasons for stavudine replacement were lipodystrophy (mainly lipoatrophy, 77%) and peripheral neuropathy (13%). Hyper-TG was the main reason for stavudine replacement in only 26 patients. Body fat redistribution and dyslipidaemia were frequent at baseline. Patients had been heavily exposed to antiretroviral drugs. PIs were included in the current regimen in 112 (34%) patients, boosted with ritonavir in 60 (54%) of them. Most of the patients were receiving a non-NRTI based HAART.
74% of patients had HIV=RNA <200 c/ml. Median CD4 count was 529. Most patients had been on multiple regimens previously: 51% 4th line or higher; 23% third line; 19% second line; 7$ first line. Mean d4T exposure was 4 years (range: 2-5). 34% of patients were on a PI regimen: 16% non-boosted, 18% boosted (mostly Kaletra). 70% were on a NNRTI regimen & 53% on Sustiva, 19% on NVP. 12% of patients were on a lipid lowering agent. 49% of patients had dyslipidemia: 35% >220 mg/dl cholesterol; 43% >200 mg/dl tryglicerides; 8% >500 mg/dl TG. 67% had lipoatrophy, 3% fat accumulation, 15% both.
Lipid outcomes
The 48-week outcomes of lipid parameters are shown in Table 2. All parameters, with the exception of HDL-C significantly improved (Wilcoxon rank sum test, P < 0.01) after tenofovir for stavudine replacement. All parameters significantly improved independently of the presence of baseline hyperlipemia. However the magnitude of the reduction of median plasma LDL-C and TG values was greater in patients with baseline hyperlipemia than in patients with normal baseline lipid values. LDL-C decreased by a median of 29 versus 8 mg/dl, and TG 76 versus 35 mg/dl, respectively (P < 0.001 for all comparisons with respect to baseline).
Both in TG and total-C, all the improvement attained at 48 weeks was already seen as early as at 12 weeks. In fact, in some patients the TG levels at month 3 were even lower than thereafter (data not shown). Quite the opposite, total and LDL-C levels showed a steady improvement from baseline to 48 weeks.
The 24- and 48-week evolution of total-C and its sub-fractions in the whole group and in selected patients with baseline dyslipidaemia are shown in Fig. 1.
Regarding the whole population (n = 352), 136 (39%) had LDL-C > 130 mg/dl at baseline. Also 67 of these 136 subjects had LDL-C > 160 mg/dl. The evolution of this group is shown in Table 2. Patients with baseline dyslipidemia -- baseline, week 24, week 48 numbers: median triglycerides were 283, 213, & 207 mg/ml; median total cholesterol were 262, 231, and 225 mg/dl; median LDL-C were 163, 141, and 134 mg/dl. For patients with >500 tryglicerides median numbers were 679 at baseline, 353 at week 24, and 413 mg/dl at week 48. For patients with LDL-C >160 mg/dl numbers were 182 at baseline, 154 at week 24, and 151 at week 48.
HDL-C did not improve, but instead showed a 0.8 mg/dl decrease (P < 0.05). There was a significant improvement in the ratio total-C/HDL-C: 4.89 to 4.56 (P < 0.01), both in the whole group and in patients with baseline hyperlipidemia: 5.53 to 4.88 (P < 0.01).
The evolution of TG levels at 24 and 48 weeks is shown in Fig. 2. The greatest lipid reduction was observed in patients with severe baseline hyper-TG with a 48-week reduction in TG of 265.5 mg/dl (P < 0.001) (Fig. 2).
Among patients whose stavudine replacement was primarily due to the presence of hyper-TG, a 48-weeks improvement was seen in 16 (61%), a resolution in eight (31%) and a worsening in only two (8%) (P < 0.01 for improvement/resolution versus worsening).
The proportion of patients receiving lipid-lowering agents decreased during the study follow-up (12% at baseline, 10% at 48 weeks), but the difference was not statistically significant. However, it must be reiterated that the design of the study did not allow the substitution of other antiretroviral agents present at baseline (i.e., boosted PI) with known deleterious influence on lipid parameters. All patients on lipid lowering therapy were dyslipidemic at baseline and we specifically analysed the change in lipid values in this group (n = 41). There was also a non-significant increase in the rate of patients achieving normal lipid values at 48 weeks in all subgroups.
Concerning all patients with baseline dyslipidemia, in individuals with raised baseline TG the rate of hyper-TG decreased from 81% to 60% (P = 0.21), and the rate of severe hyper-TG (> 500 mg/dl) decreased from 37% at baseline to 20% at 48 weeks (P = 0.68). In patients with hypercholesterolemia, the rate of subjects with total-C > 220 mg/dl decreased from 66% at baseline to 31% at 48 weeks (P = 0.09), and the rate of patients with LDL-C > 160 mg/dl decreased from 42% at baseline to 23% at 48 weeks (P = 0.50).
Cardiovascular risk impact
There was a small but significant decrease of 1% raw point in the mean estimated 10-year CVR in all patients, independent of smoking status and sex (Table 2). Mean systolic blood pressure, body mass index, percentage diabetes, and percentage smokers did not differ between baseline and 48 weeks (122 mmHg; 22.8 kg/m2; 7.4% and 7.7%; (54% and 53%, respectively).
The greatest improvement in the mean estimated CVR should be obtained in a male subject with baseline dyslipidaemia (smoker, 15.5-10%; non-smoker, 4-2% 10-year CVR).
During the 48-week follow-up three patients (all with baseline dyslipidaemia) suffered cardiovascular events: one acute myocardial infarction, one angina pectoris, and one cerebrovascular accident, that yield an incidence of cardio or cerebrovascular events of 8.53 per 1000 person-years (95% confidence interval, 8.48-8.57).
Safety and tolerability
The safety profile of the study has been reported elsewhere [25]. Briefly, the switch did not increase the rate of virological failure, and the median CD4 cell count increased at the end of the study by 42 cells/_l, despite patients being virologically controlled at baseline. There were no new adverse events during the 48-week follow-up.
Discussion
Dyslipidaemia is very common in heavily pretreated patients receiving HAART including stavudine, and with a high rate of lipodystrophy [3,7,29-31]. So, this is a relevant scenario to evaluate the impact of a drug switch in the lipid profile.
The substitution of tenofovir for stavudine caused a sustained significant improvement in total-C, LDL-C, and TG. Although HDL-C decreased, it was clinically insignificant, and the TC/HDL-C ratio decreased in patients with and without baseline dyslipidemia.
The drop in lipid values was even more pronounced in patients with baseline dyslipidaemia, with TG showing a striking median decrease in patients with severe hyperTG.
In a small subset of our patients hyper-TG was the main reason for stavudine withdrawal. At the time the study was designed there was little evidence to support this strategy, but the treating physicians of this subset of patients considered that this switch could be beneficial. In fact, 24 out of 26 patients improved or resolved their hyper-TG.
We observed an early lipid decay effect, greater at 12 weeks, probably due to the effect of switching off the offending drug causing dyslipidaemia. The improvement was maintained thereafter at every time point because tenofovir itself does not induce hyperlipidaemia [23,32]. This early decrease in lipid levels after stavudine replacement by tenofovir was also described in a small retrospective pilot study, although blood analysis were non-fasting [33]. It must be emphasized that both of these limitations have been avoided in our study. Contributing factors to hyperlipidaemia in HIV infection are increased hepatic synthesis of VLDL, impaired lipid clearance, and circulating cytokines [30]. However, all of these mechanisms have been related mainly to PI [11].
The NRTI substitution approach has rarely been considered in the management of HAART-associated dyslipidaemias. Most emphasis has been directed to PI replacement by nevirapine, efavirenz, abacavir or, more recently, atazanavir [1,4,34].
Exposure to stavudine has been linked to dyslipidaemia and also to insulin resistance through the inhibition of mitochondrial DNA polymerase-gamma, which plays a key role in adipose tissue metabolism and dyslipidaemia [21,31].
Although there are currently no investigational data supporting an intrinsic lipid-lowering effect of tenofovir, in fact adding solely tenofovir to a stable antiretroviral regimen reduced the rate of hypertriglyceridemia in a previous crossover controlled study [35]. Moreover, tenofovir has recently been compared with abacavir in the RAVE study, designed to evaluate the improvement of lipoatrophy in patients receiving thymidine analogues (stavudine and zidovudine) [36]. In this first head-to-head switching comparison there were no between-group differences in limb fat increase. However, although TG, total-C and LDL-C declined in both arms, the decline was significantly greater with tenofovir. So, the whole improvement in lipid profile in our study should not be attributed solely to the replacement of stavudine, but could also be due to the absence of deleterious effect on lipids of tenofovir.
The metabolic effects of any given antiretroviral agent are one of several factors to consider when selecting a regimen. Therefore, tenofovir appears to be an optimal candidate to substitute stavudine as a safe switch with the aim of improving the lipid profile.
Our results are in absolute agreement with the single-arm GS903e substudy, that assessed reversibility of lipid changes induced by stavudine at the conclusion of the randomized study period [24]. Patients who switched to tenofovir showed significant improvements in their TG, total-C and LDL-C levels, while HDL-C remained unchanged. These results completely parallel our findings. The absence of improvement in HDL-C in our trial could be related to the presence of PI in the backbone regimen in some patients, the only agents clearly linked to reduced HDL-C [7,37,38]. In fact, most studies of PI substitution with nevirapine, and to a lesser extent efavirenz, have shown improvements in HDL-C [14,39]. However, patients in the GS 903 trial replacing stavudine with tenofovir did not receive PI and showed unchanged HDL-C values at 48 weeks [24]. So, probably it must be assumed that the overall lipid improvement shown after only stavudine replacement with tenofovir will have little impact on HDL-C. Levels of HDL-C decrease soon after HIV-1 seroconversion and remain low thereafter in spite of receiving successful antiretroviral therapy [2]. It has also been assumed that low HDL-C could be unrelated to treatment, but rather is result of HIV infection itself as a marker of chronic inflammatory activity. The fact that it does not return to normal despite HAART-induced viral suppression and immune reconstitution could suggest that inflammatory activity is not completely restored.
The present substitution study was not designed to obtain a maximal improvement in lipid profile. The antiretroviral combination during the study was composed in some patients of drugs that are known to have deleterious effects on lipid parameters. Of note, one-third of the patients were receiving PI boosted by low-dose ritonavir in most of them. Furthermore, half of them were receiving efavirenz, a non-nucleoside that impairs lipogenesis to a greater extent than nevirapine through suppression of the transcription factor sterol regulatory element binding protein-1c (SREBP-1c) pathways [40-42].
Although switching stavudine to tenofovir may not resolve dyslipidaemia adequately in some patients, its role as an adjuvant to direct lipid-lowering agents seems clear. Both strategies must not be contemplated as exclusive but rather as complementary.
The changes attained in lipid parameters by this NRTI switch translated into a small but significant median decrease in the estimated overall CVR of the patients. Moreover, the individualized 10-year CVR benefit obtained is even greater in individuals with a worse cardiovascular profile: male subjects with baseline dyslipidaemia. The mean systolic blood pressure and BMI did not differ between baseline and 48-week values, nor did the percentages of diabetes and smokers. So, the CVR reduction was entirely related to dyslipidaemia improvement.
Our study further defines lipid improvement dynamics and will foster appropriate studies to gain a more complete understanding of this issue.
There main limitation of our study is the lack of a randomized stavudine-treated control arm. However the GS903 study showed that in the stavudine-continuing arm lipid abnormalities did not subside; ratherr a worsening was shown in TG, total-C and LDL-C levels with respect to baseline [23]. So, given that we have compared the observed lipid parameters at 48 weeks to their baseline values, the existence of a stavudine control arm would probably even have increased the differences and have raised the significance of the observed improvement in lipid profile and cardiovascular risk estimates.
In conclusion, the first large-scale prospective substitution study from stavudine to tenofovir unveils that this replacement may reverse, at least partially, lipid abnormalities and reduces the CVR and the atherogenic profile in HIV-1-treated patients. This safe switch strategy could reduce the use of hypolipidaemic drugs and the rate of severe hypertriglyceridemia and must be considered as another possible approach to the treatment of dyslipidaemia associated with HIV-infection and HAART.
|
|
| |
| |
|
|
|
