Impact of Antiretroviral Treatments on Coronary Heart Disease

Although it is now clearly established that cART is linked to metabolic disorders, the long-term impact of these disorders is still being discussed. The first questions con­cerning an increase in the risk of CHD and the possible responsibility of PIs were raised in May 1998.

Some studies [26-32], but not others [33-36, 37], have shown a link between the risk of CHD and exposure to PIs (Table 1). Some suffered from methodological prob­lems such as small sample size, median PI treatment periods of less than 12 months (two studies), and likely underreporting rel­ative to the general population.

Far from increasing the incidence of car­diovascular events, a study from a database on administrative data [36] showed that cART tended to reduce the short-term risk in a population of more than 36,000 HIV- infected U.S. army veterans. Between 1995 and 2001, the rate of hospital admission for cardiovascular or cerebrovascular disease fell from 1.7 to 0.9 per 100 person-years (PY), while the overall mortality rate fell from 21.3 to 5.0 per 100 PY during the same period.

A study based on the French Hospital Database on HIV (FHDH) using DMI2 soft­ware (property of the French Ministry of Health) included 34,976 men with a total fol­low-up of 88,029 PY between 1996 and 1999 [31]. There were 49 cases of MI during 39,023 PI-exposed PY. The incidence rate of MI among subjects exposed to PIs between 1996 and 1999 was estimated according to the duration of PI therapy, based on three peri­ods: infarction Between December 1999 and January 2007 33 347 subjects

157 912 PY 517 RR=1.90 (1.47-2.45) recent use abacavir vs no recent use

RR=1.49 (1.14-1.95)

recent use didanosine vs no recent use

PY, person-years; med, median

The expected incidence of MI in the general male population with the same age distribu­tion was 10.8 cases per 10,000 PY. The esti­mated incidence of MI was 8.2 per 10,000 PY (95% CI=4.7-11.7) in group 1, 15.9 (95% CI=7.9-23.9) in group 2, and 33.8 (95% CI=15.4-52.1) in group 3 (Fig. 1). No signifi­cant difference was observed between the general male population of the same age and the patients treated with PIs for less than 18 months.

How can this discordance between the results of Bozzette et al. [36] and Mary- Krause et al. [31] be explained? A direct comparison of the results of the two studies is impossible because they did not examine the same types of event. Mary-Krause et al. [31] examined admissions for and deaths from MI, whereas Bozzette et al. [36] only looked at admissions for cardiovascular dis­ease, without taking deaths into account (especially deaths occurring outside hospi­tal), probably leading to an underestimation of the number of cases. Also, this latter study grouped together deaths and admissions for cardiovascular and cerebrovascular disease. Holmberg et al. [28], who found an effect of PIs regarding the risk of MI, did not observe an increased risk of cerebrovascular disease. Results from the published DAD study (Data Collection on Adverse Events of Anti-HIV Drugs) indicated that there is an increasing risk of first cardio- and cerebrovascular events with longer exposure to combination antiretroviral treatment [38]. The ninth ver­sion of the international classification of dis­eases [46], used by Bozzette et al. [36], includes codes 410 (acute MI), 411 (other acute and subacute forms of ischemic heart disease), 413 (angina pectoris), and 414 (other forms of chronic ischemic heart dis­ease), whereas Mary-Krause et al. [31] used only code 410.

Fig. 1 Incidence of myocardial infarction (MI) per 10,000 person-years (PY) among HIV-infected men ac­cording to the duration of PI exposure (in months) compared to the incidence of MI among the general male population of the same age. The standardized morbidity ratios (SMR) with their 95% confidence in­tervals in brackets were used to test the differences between incidence rates. (From [31], with permission)

A recent update of the study of Bozette et al [37] continues to find hazards for seri­ous cardiovascular events near 1.0 for expo­sure to cART. Nevertheless, the HR of cART that includes an NNRTI or a PI, even non significant, is equal to 1.22 (95% CI=0.77- 1.92) and the HR of cART that includes a PI only is equal to 1.28 (95% CI=0.71-2.30).

The DAD study [29], a prospective obser­vational study of 23,468 patients enrolled in 11 cohorts on 3 continents from December 1999 to July 2002, showed a similar relation­ship between exposure to combination anti­retroviral therapy (including PIs or NNRTIs) and the risk of MI to that found by Mary- Krause et al. [31], the risk increasing with the duration of cART. In the same way, Moore et al.

An update of the DAD study [32] on 23,437 subjects corresponding to 94,469 person-years (PY) of follow-up among them 345 MI showed an incidence of 3.7/1,000 PY with a decrease between 1999 and 2005. The incidence increased with increasing length of exposure to PI with an incidence increased from 1.5 per 1,000 PY among those not exposed to PI to 6.0 per 1,000 PY among those exposed to PI. In addition, the relative rate per year of expo­sure to PI was 1.16 (95% CI=1.10-1.23), result found as well by Klein et al. [47], which corre­sponds to a doubling of the risk over a 5-year period of exposure, whereas for non-nucleo- side reverse transcriptase inhibitors it was 1.05 (95% CI=0.98-1.13).

Recent results reported by the D:A:D group at the 2008 CROI meeting [39], and published in The Lancet [40], showing that recent use of abacavir and didanosine (cur­rent use or interruption less than 6 months previously) was associated with an increased risk of myocardial infarction (MI) need to be confirmed or not by other studies.

The Strategies for Management of Anti­retroviral Therapy (SMART) study [48] recently showed an increased risk of cardio­vascular disease in patients whose antiretro­viral therapy was interrupted when their CD4+ cell count reached a certain level as opposed to patients who received continuous treatment. This international randomized trial, conducted in 318 sites in 33 countries, had enrolled 5,472 patients when it was stopped.

The increased risk of MI observed just after the advent of PI could be in part explained by a population who lived longer, and so the aging of the population which is at higher risk of cardiovascular diseases; by the chronic inflammatory state of HIV-infection, by the impact of known risk factors or CHD as smoking, and by the metabolic effects of antiretroviral treatment. Nevertheless, although today it is admitted that incidence of CHD was higher in HIV-infected patients than in the general population, links between viral infection, antiretroviral treat­ment, especially the impact of new antiretro­viral drugs which have potentially less meta­bolic disturbances than PI, and the increase of MI incidence need to be precise even if we are not able to study each drug separately.