HIV Aids - Gender Effects, HIV, ARV

Author

Gender Effects, HIV, ARV

GMCarter

2005-03-29, 6:22 pm

forwarded from NATAP - www.natap.org

Gender Affect in HIV & HAART

Sex (Gender) and the course of HIV infection in the pre- and highly
active antiretroviral therapy eras

AIDS, 4 March 2005

Prins, Mariaa; Meyer, Laurenceb; Hessol, Nancy Ac
>From the aCluster Infectious Diseases, HIV and STI Research, Municipal Health Service, Amsterdam, The Netherlands
bService d'Epidémiologie et de Santé Publique, INSERM
U569/INED/Paris XI, Hopital de Bicatre, Le Kremlin-Bicatre cedex,
France
cDepartment of Medicine, university of California, San Francisco,
California, USA.

Article Outline
* Abstract
* Introduction
* Sex and clinical disease progression following HI...
* Sex, HIV-1 RNA levels and CD4 lymphocyte counts
* Specific manifestations in HIV-infected women
o Pregnancy and the course of HIV infection in the mother
o Impact of antiretroviral therapy on reproductive life
o Amenorrhea, (an)ovulation and menopause
o Cervical disease
* Sex differences in pharmacology and response to t...
o Drug levels, viral suppression, immunologic response and
clinical outcomes
o Adverse drug reactions
o Metabolic effects
+ Body habitus changes
+ Insulin resistance
+ Bone metabolism
+ Cardiovascular events
* Conclusions

Abstract
We reviewed the available literature on the potential effects of sex
on the course of HIV infection and found that there is little evidence
for sex differences in the rate of disease progression in the
pre-highly active antiretroviral therapy (HAART) and HAART era.

Compared to men, women appeared to have lower HIV RNA levels and
higher CD4 cell counts shortly after infection with HIV, but studies
were inconclusive regarding whether these differences diminish over
time. Differences in viral load or CD4+ cell count might cause women
to delay initiation of HAART. Nonetheless, we found no substantial sex
difference in the benefit of antiretroviral therapy.

The studies we reviewed failed to find any harmful effect of pregnancy
on HIV disease progression. With the availability of effective
antiretroviral agents, HIV-infected women have increasingly decided to
have children.

Conflicting results exist on the effect of HAART on regression of
cervical intra-epithelial neoplasia (CIN). Unlike CIN, invasive
cervical cancer has not been found to be much higher in HIV-infected
women than in HIV-uninfected women. Although publication bias cannot
be ruled out, published studies suggest higher rates of adverse events
among HIV-infected women on therapy as compared to men. As more
pharmacological agents are developed, it is especially important that
potential sex differences in pharmacodynamics are assessed. The
relationship between metabolic abnormalities, changes in body habitus,
and endocrine perturbations has not been extensively studied. Whether
sex differences are due to unalterable genetic factors or social and
environmental conditions, it is imperative that all HIV-infected
individuals have equal access to interventions that can slow disease
progression.

Introduction
An estimated 50% of the adults living with HIV are women. This
proportion differs across regions; in sub-Saharan Africa, where most
HIV-infected individuals live, women represent 58% of the 26.5 million
adults living with HIV, whereas in Australia and New Zealand 7% of the
15 000 HIV-infected adults are women [1].

In 1981 the first cases of AIDS were described in homosexual men [2]
and subsequently most research on the course of HIV infection has been
derived from studies in gay men. This accounts, at least in part, for
the lack of early data describing HIV infection in women, to whom the
epidemic spread more recently. However, nowadays the number of women
included in clinical trials evaluating the effect of new treatment
regiments and potential HIV vaccines is still disproportionately
small.

The HIV epidemic is greatly impacted by issues related to sex and
gender [3], ranging from susceptibility and vulnerability to
infection, prevention behaviors, disease progression and
manifestations, access and response to treatment and vaccines, and
pharmacology. Conducting sex comparisons within research studies is
important to delineate sex differences or similarities. Sex
comparisons of the rates of disease progression in the era before
highly active antiretroviral therapy (HAART) gave us information about
the natural history of HIV infection. Comparisons in the era after the
introduction and widespread use of HAART tell us whether women and men
equally benefit from HAART. Moreover, attention to the possibility of
sex differences in pharmacokinetics and -dynamics is necessary, along
with examination of differences in side-effects and toxicity. Specific
aspects related to HIV infection in women, including pregnancy and
gynecological diseases, can be studied only in women. Next to cultural
and social attitudes, reproductive decision-making has probably been
influenced by the availability of measures to reduce mother-to-child
transmission and the improved life expectancy of the HIV-infected
mothers due to effective antiretroviral therapy.

Herein, we review the effect of sex on clinical manifestations and
laboratory markers of HIV disease progression, and the influence of
sex on the response to HAART, including pharmacokinetics and
-dynamics. Using the available English language scientific literature,
identified by MEDLINE using appropriate keywords and supplemented with
perusal of reference lists of relevant publications, we aimed to
synthesize the findings and to provide recommendations.

Sex and clinical disease progression following HIV seroconversion
A few studies published in the early 1990s among HIV-infected
individuals, mainly people with AIDS, suggested that women had a
shorter survival time than men [4-6]. Rather than differences due to
biology, these sex differences probably reflected more limited access
and use of health care for women compared with men. In addition,
factors related to the unknown duration of HIV infection might have
biased these results [7]. To avoid bias related to duration of
infection, sex differences in HIV disease progression are best
evaluated in cohorts of HIV-infected individuals for whom HIV
seroconversion is documented. These studies are often called
seroincident or seroconverter cohorts. In some seroconverter studies
health care is provided for participants who return for follow-up,
thereby minimizing any differential access and use of health care.

We looked for studies evaluating the effect of sex on disease
progression in seroconverter cohorts of men and women from the same
risk group in the era before the widespread use of HAART. This search
resulted in six studies reporting relative risks of AIDS or death, for
adult women compared to adult men, that were adjusted for age [8-13].
Adjustment for age is needed as older age has consistently been shown
to be associated with faster disease progression in the pre-HAART era
[12,14] and therefore might confound the association by sex. Two of
the six studies [11,12] are a pooled analysis of individual data from
different cohorts and therefore have overlapping data. Although none
of the studies demonstrated a statistically significant effect, most
studies found somewhat slower rates of progression to AIDS and death
in women than men. With the exception of one study [9], the sex effect
was slightly stronger for death as an outcome compared to AIDS.

In the HAART time period, seroincident cohorts continue to play a
vital role in evaluating temporal effects, in addition to those of
sex. We found three recent seroconverter studies that investigated
whether the effect of sex had changed in the HAART era versus the
pre-HAART era by including interaction terms between sex and the
calendar period at risk in their multivariate models. These studies
showed that the effect of sex on progression to AIDS and death did not
substantially change after the introduction of HAART [15-17], except
that in the HAART period HIV-infected female injecting drug users
(IDU) tended to live longer than their male counterparts [17].
All-cause mortality was the endpoint in this collaborative study,
therefore competing non-HIV-related causes of death including
non-natural deaths might be largely responsible for the latter
finding. Interestingly, a statistically significant sex effect on
progression from seroconversion to AIDS and death was identified in
the Spanish cohort of mainly IDU (adjusted hazard ratio for women
compared with men for the overall period (pre-HAART and HAART era)
0.68 [95% confidence interval (CI), 0.46-0.99] and 0.53 (95%CI,
0.33-0.87) respectively) [15].

The finding that, in general, the sex effect on HIV disease
progression did not change in the HAART era suggests that at the
population level women and men equally benefit from HAART. This is in
agreement with the finding of the CASCADE collaboration that time from
HIV seroconversion to initiating antiretroviral therapy did not differ
by sex, although women initiated therapy at higher CD4+ cell counts
than men [18].

Sex, HIV-1 RNA levels and CD4 lymphocyte counts
HIV RNA levels and CD4+ cell count are the two markers most closely
correlated with and predictive of HIV disease progression.
Recommendations for when to initiate therapy in people with chronic
asymptomatic HIV infection focus primarily on the CD4+ cell count,
whereas in the past more weight was placed on HIV RNA levels [19].
Aside from recommendations for treatment during pregnancy to reduce
mother-to-child transmission, treatment guidelines have been similar
for both sexes and have largely been based on studies in men.

Shortly after the test for HIV RNA became available, viral load was
established as a key marker, along with CD4+ cell count, for HIV
disease progression in men [20]. The first studies evaluating sex
differences in HIV RNA levels were published and whereas one study
found a statistically significant sex effect [21] the other study did
not [22]. In 1998 investigators reported an increased rate of disease
progression in women compared with men with the same HIV RNA levels
[23]. They suggested that the HIV RNA threshold for initiation of
therapy in women should be revised downwards. However, others reported
similar progression rates in men and women with equivalent HIV RNA
levels [13,24]. In any case, at that point in time the debate about a
potential sex effect on the outcome of HIV infection started again.

Recently two review articles on sex differences in HIV-1 RNA levels
have been published [25,26]. In each publication, 13 studies were
reviewed and 12 were included in both reviews. Napravnik et al. [25]
used data from the published studies and in a meta-analysis estimated
that in CD4+ cell count-unadjusted analysis women had, on average,
half the HIV RNA levels [-0.30 log10 difference (interquartile range
(IQR), -0.20 to -0.42)] as men. After adjusting for CD4+ cell counts,
on average, women had 41% lower plasma HIV RNA than men [-0.23
log10(IQR), -0.16 to -0.31]. Gandhi et al. [26] distinguished
cross-sectional from longitudinal studies: seven of the nine
cross-sectional studies demonstrated a 0.13-0.35 log10 lower level of
HIV RNA in women after controlling for CD4+ cell count, whereas four
longitudinal studies revealed that women had 0.33-0.78 log10 (2- to
6-fold) lower levels of HIV RNA than men, even when controlling for
time since seroconversion. Both reviews clearly show that, in general,
the higher the CD4+ cell count, the larger the HIV RNA
female-to-male-difference reported in each individual study (adapted
from Gandhi et al. [26]). This indicates that sex differences may
diminish over time.

As with clinical disease progression, sex differences in laboratory
marker patterns are best evaluated in individuals for whom the
duration of infection is known. Five research groups evaluated the sex
effect on HIV RNA levels longitudinally in adult HIV seroconverters.
Sterling et al. [27] compared fast with slow progressors and found
that shortly after infection HIV RNA levels are lower in women, but
that HIV RNA levels tended to increase more rapidly in women than in
men and subsequently converged. In another study in the same cohort of
drug users, Sterling et al. [28] stated that by applying treatment
recommendations twice as many men as women are eligible for treatment
during the first year after seroconversion, but that the sex
difference becomes less pronounced with time since infection. The
narrowing of the sex gap with increasing duration of HIV infection is
in accordance with two other studies [13,29]. However, Lyles et al.
[30] found, in Italian HIV seroconverters, a significantly lower viral
load in women and no sex difference in the slope, suggesting that a
sex effect is present throughout infection. A recently published study
by the CASCADE collaboration also suggested that women had lower HIV
RNA levels and that this difference persisted throughout infection
[31].

Finally, lower HIV RNA levels in women than in men have also been
observed in cross-sectional studies outside North America and Europe,
where other HIV-1 subtypes are prevalent [32-34]. Interestingly, in a
longitudinal study among children with vertically acquired HIV
infection, HIV RNA peaked at 3 months of age higher for girls than
boys, but after 4 years HIV RNA levels were consistently lower for
girls, at 5 years being about half a log10 below that of boys [35].
This study also found that sex differences in HIV RNA are still
present, but less pronounced under therapy.

Although the CD4 lymphocyte counts have been established as an
important prognostic marker for HIV disease progression well before
HIV RNA was identified as prognostic marker, less attention has been
paid to sex differences in CD4+ cell count. Nonetheless, CD4+ cell
counts have been reported to be higher among HIV-uninfected women than
among men [36-40]. Several studies, the majority of which are
cross-sectional in design, have evaluated sex differences in CD4+ cell
count in HIV-infected individuals.

Simple pooling of the aggregate data (i.e. reported CD4+ cell count
difference between women and men) from each published study showed
that the median female-to-male difference in HIV-uninfected
individuals is 98 × 106 cells/l (IQR, 73 to 131) higher in women than
in men [36-40]. Among adult HIV-infected individuals in the pre-HAART
era, at seroconversion, AIDS, and death the median female-to-male
differences were 50 (IQR, 27 to 90), 1 (IQR, -7 to 8) and 22 × 106
cells/l, respectively [8,11,13,27-30,41-46]. As progression rates to
AIDS and death seem to be similar for men and women, this suggests
that the higher CD4+ cell count in women in comparison with men
becomes less pronounced over time since infection. However, as for HIV
RNA, results of longitudinal studies among HIV seroconverters are
inconclusive, with some studies suggesting similar CD4 slopes over
time since infection [11,41,43], whereas others found steeper slopes
for women compared with men [13]. In HIV-uninfected children, CD4
lymphocyte levels have been reported to be systematically higher for
girls than boys, with sex differences increasing until about 4 years
of age, whereas the CD4+ cell counts of HIV-infected girls were
initially lower than those for infected boys [47].

Based on studies in women, it has been suggested that both markers are
affected by sex hormones [39,48-53], which differ by sex and change
over the course of time, as we discuss below. However, the underlying
mechanism is unclear.

Specific manifestations in HIV-infected women
A large prospective, multicenter cohort study in the pre-HAART era
found a striking similarity in the occurrence of most initial and
subsequent specific AIDS-defining conditions between men and women
[45]. After excluding homosexual men to compare men and women from the
same risk group and adjusting for potential confounders, this study
found that HIV-infected women with AIDS were at increased risk of
developing toxoplasmosis and herpes simplex virus ulcerations. Another
multicenter study among men and women from all risk groups found a
decreased risk of developing Kaposi's sarcoma in women, mirroring the
higher prevalence of human herpesvirus 8 (HHV8) infections in
homosexual men, and an increased risk of bacterial pneumonia, which
was primarily evident among injecting drug users [6]. In conclusion,
AIDS diagnoses vary by risk group rather than by sex, with risk group
being a surrogate marker for some unmeasured related factor, such as
HHV8.

Pregnancy and the course of HIV infection in the mother
Pregnancy has been reported to modify the course of several autoimmune
diseases. Therefore when the HIV epidemic began, it was reasonable to
question whether pregnancy altered the course of HIV infection. A
number of studies addressed this question [54-59]. All these studies
led to the same reassuring conclusions that there is no effect of
pregnancy on disease progression, at least in developed countries and
before HAART became widespread. Not surprisingly, pregnant
HIV-infected women tended to be healthier than non-pregnant
HIV-infected women [58-60]. Better health status in pregnant versus
non-pregnant women could explain why, in some studies, pregnancy even
appeared to delay disease progression [54,59,61]; this is probably due
to residual confounding despite the fact that most analyses were
adjusted for time since seroconversion, CD4+ cells at baseline or
before conception [58].

In comparison with developed countries, fewer studies of HIV infection
and pregnancy have been performed in developing countries and they are
hampered by smaller sample sizes. Deschamps et al. [62] in Haiti
observed no deleterious effect of pregnancy on progression to AIDS,
and a non-significant increase in the risk of death. An adverse effect
of pregnancy on disease progression was described in an Indian study;
however women were enrolled at a very advanced stage of HIV disease
[63].

The absence of a harmful effect of pregnancy on HIV infection is
consistent with results from immunological studies. In a large study
of untreated women [50], a decline in the absolute CD4+ cell counts
was observed during the first trimester of pregnancy, followed by a
gradual increase thereafter until delivery; CD4 cell counts reached
baseline levels by 6 months postpartum. In contrast, CD4 percentages
appeared to be more stable during pregnancy. Additionally, pregnancy
had no effect on the overall CD4 dynamics over a 7-year period in
women with a documented date of seroconversion [52].

Another study assessed viral load during pregnancy in pregnant women,
most of whom received zidovudine during pregnancy [64]. A slight
decrease in HIV RNA levels during pregnancy followed by a modest but
significant rebound during the 12 weeks immediately following delivery
was observed, even among women continuing antiretroviral therapy.

These overall CD4 lymphocyte and HIV RNA patterns suggest a mechanism
related to physiological changes during pregnancy that reverses
postpartum. A dilutional mechanism secondary to volume expansion might
partly explain the observed changes. The elevated estradiol and
progesterone levels during pregnancy could also have an effect on
viral replication, such an effect being removed after delivery when
hormone levels return to pre-pregnancy levels.

Although breastfeeding can confer substantial health benefits to the
infant, the high rate of mother-to-child HIV transmission through this
route has led to the recommendation that HIV-infected mothers refrain
from breastfeeding when safe alternatives are available. Three studies
have evaluated whether breastfeeding affects the health of the
HIV-infected women. A three-fold higher mortality rate was found in
mothers in Kenya who breastfed their infants compared with those who
fed their infants with formula [65]. An excess weight loss was
observed in the breastfeeding group, which did not however fully
explain the excess in mortality. Conversely, in a study in South
Africa, a non-significant decrease in mortality was observed in
HIV-infected women who breastfed their infants in comparison with
those giving formula [66]. A recently published observational study in
Tanzania also indicated that breastfeeding was not detrimental to the
health of the HIV-infected mother [67].

Impact of antiretroviral therapy on reproductive life
The dramatic decrease in mother-to-child transmission rates due to
effective antiretroviral agents may have led to an increase in the
number of HIV-infected women deciding to have children. Indeed, the
birth rate tended to increase in the period 1994-1996 in HIV-infected
women who were followed in two French cohorts when compared with the
years before 1994 [68]. Data from pregnant women living in southern
France showed that the percentage of voluntary abortions fell from 59%
before 1994 to 38% after 1994 [69].

Until recently, few studies have evaluated the impact of HAART on
reproductive behavior. In one study of European HIV-infected women,
the proportion of births among pregnancies increased from around 30%
before 1995 to 45-50% after [70]. Abortion among women with HIV also
became less common after the introduction of HAART among US women in
the WIHS study [71].

Since 1995, significant increases in HIV seroprevalence among pregnant
women have been found in some West-European countries [72]. The number
of children born from HIV-infected mothers in a French study also
increased since 1996 [73]. These trends may reflect migration of
HIV-infected women from high prevalence countries to Western Europe.

Amenorrhea, (an)ovulation and menopause
It is uncertain whether HIV-positive women are more likely to have
abnormal menstrual function when compared with well-matched
HIV-negative women. Few studies have been performed and most rely on
self-reported symptoms and cycle rhythm. Comparisons with HIV-negative
women are necessary since anovulatory cycles occur in 5-30% of
HIV-uninfected women with regular menses [74,75]. These comparisons
should be made after adjustment for factors that are important
determinants of menstrual function, such as age, body mass index,
health status, substance use, and socio-economic status [76]. Most
studies could not enroll a sufficiently large sample of HIV-uninfected
women that were demographically and behaviorally comparable to
HIV-infected women.

One study evaluated menstrual cycles in 248 HIV-infected and 82
demographically similar HIV-uninfected women in the pre-HAART era
[77]. They found that HIV-infected women were more likely to
experience intervals of >6 weeks and >3 months between menstrual
cycles, and had lower rates of premenstrual breast symptoms and
dysmenorrhea than HIV-uninfected women; suggesting that anovulatory
cycles are more common in HIV-infected women. However, the frequency
of amenorrhea >3 months over a 6-month period was only 5% among
HIV-infected women, which is not very different from what is generally
observed in women of reproductive age [76]. In a larger study,
investigators concluded that although HIV infection may slightly
increase the probability of having very short cycles (<18 days), HIV
serostatus appeared to have little overall effect on menstrual cycle
length [78]. Contrary results were found regarding the association
between low CD4+ cell counts and menstrual cycle abnormalities, some
studies found an association [78,79], but others did not [77]. There
is little to no published data on the menstrual cycle among
HIV-infected women receiving HAART, an important area of investigation
in those experiencing therapy-related lipodystrophy.

The effect of sex hormones on the CD4 lymphocytes and HIV RNA levels
during menstrual cycles in HIV-infected women was evaluated in two
small studies. A decrease of 0.16 log10 in HIV RNA levels from early
follicular to the midluteal phase was observed in HIV-infected
untreated women [51]. Conversely, no variation in plasma HIV RNA over
two menstrual cycles was observed in another study of ovulating women,
who mostly received combination therapy [80]. Among women who
ovulated, absolute CD4+ cell counts decreased over the cycle, although
the effect did not reach statistical significance [51].

There are two inter-related issues regarding menopause: (1) the effect
of HIV on onset and duration of menopause; and (2) the effect of
menopause on HIV infection. One cross-sectional study that evaluated
the first issue found a mean age of menopause of 47 years among
HIV-infected post-menopausal women [81]. However, this cross-sectional
study might have underestimated the age at menopause since
HIV-infected women who had not yet become menopausal, and probably do
so at older age, were not included. Regarding the second issue,
post-menopausal HIV-infected women had marginally lower CD4+ cell
counts than pre-menopausal women, 66 × 106 cells/l at 3 years after
seroconversion, but the monthly CD4 decline was not associated with
menopause [52].

Cervical disease
The relationship between human papillomavirus (HPV) infection and the
development of HPV-associated cervical intraepithelial lesions and
cancer is now well established. A much higher rate of cervical lesions
were observed in HIV-infected women than in uninfected women,
including studies that controlled for sexual behavior [82]. The
incidence of squamous intraepithelial lesions (SIL) in a prospective
study in the pre-HAART era showed an incidence of 8.3 per 100
person-years in HIV-infected women compared with 1.8 per 100
person-years in socio-demographically similar uninfected women [83].
In that study, SIL incidence also increased with decreasing CD4 cell
counts; this relation has been described in several other studies
[84,85]. HPV infection, apart from being more common in HIV-infected
women, is also less transient [83,86,87]. Moreover, the likelihood of
spontaneously clearing HPV was found to decrease with decreasing CD4+
cell counts [82]. Most interestingly in this study, the association of
HIV infection with increased risk of cervical intraepithelial
neoplasia (CIN) was fully explained by repeated HPV positivity.

Conflicting results exist about whether immune restoration associated
with receiving HAART can enhance the regression of CIN. Among
HIV-infected women presenting with CIN, the rate of regression was
twice as high in women receiving HAART in comparison with women not on
HAART [88]. A trend for a higher increase in CD4+ cell count at 6
months after HAART initiation was observed in women who regressed
compared with women who did not; finally, the calendar period of CIN
diagnosis (after 1996 compared with before) was also associated with
an increased probability of CIN regression. Similarly, Minkoff et al.
[89] reported that women, all bearing an oncogenic HPV subtype, on
HAART were more likely to demonstrate regression of cervical
cytological abnormalities and less likely to demonstrate progression
than women off-HAART with similar CD4+ cell counts. Conversely, in two
other studies no beneficial effect of HAART was demonstrated on
incidence of SIL or progression/regression rates of cervical
abnormalities [85,90], but duration of HAART was not taken into
account in these analyses.

Unlike CIN, the incidence of invasive cervical cancer (ICC) has not
been found to be much higher in HIV-infected women than in
well-matched HIV-uninfected women, as shown in a large, recently
published prospective study [91]. Based on the results, the authors
concluded that ICC is uncommon in HIV-infected US women who have
regular access to ICC prevention and treatment.

Sex differences in pharmacology and response to treatment
In the past, women were excluded from clinical phase I and early phase
II trials due to possible risks associated with childbearing [92].
Therefore information about pharmacokinetics in women and sex
differences in the relationship between dose and adverse drug
reactions, or dose and efficacy has historically been less. That began
to change in 1993 when the US Food and Drug Administration recognized
the need to include women in early phases of drug studies [93,94].
Since then, increasing attention has been paid to the role of
biological sex in the pharmacokinetics and pharmacodynamics of drugs
[95].

Pharmacokinetics, defined as the process by which a drug is absorbed,
distributed, metabolized, and eliminated by the body, can be
influenced by hormones (endogenous or exogenous), gastric acid
secretion, gastrointestinal (GI) blood flow, renal blood flow, body
weight and fat, and muscle mass. Pharmacodynamics, defined as the
study of the action or effects of drugs on living organisms, is the
result of drug pharmacokinetics.

Drug levels, viral suppression, immunologic response and clinical
outcomes
Nearly all anti-HIV regimens include a nucleoside reverse
transcriptase inhibitor (NRTI). Both pharmacokinetic and
pharmacodynamic sex differences have been found among women taking
zidovudine and lamivudine, in which women had higher NRTI drug
concentrations and achieved a more rapid reduction in viral load than
men [96], and abacavir, which found that one female patient had
consistently higher drug concentrations [97]. However, both these
reports are hampered by a small number of women studied. Another study
evaluated the role of sex in nevirapine pharmacokinetics and observed
statistically significant inter-group differences in the plasma
concentration [98]. These sex differences in plasma levels may be
accounted for by variations in lean body mass and composition. In a
large study by Fletcher et al. [99], concentrations of saquinavir were
significantly higher in women than in men.

Two observational studies found that, among virological responders,
women had a greater immunological response than men [100,101]. Several
studies by Moore and colleagues have shown differing results
[102-104]. The first study found that after starting HAART, women
achieved virologic suppression at a faster rate than men and have a
more durable response [102]. A follow-up publication using the same
cohort had less conclusive evidence for a sex advantage with respect
to clinical progression [103]. A more recent study, using a larger
multicenter study population found that after controlling for
potential confounders there were no significant sex differences in
virologic, immunologic, or clinical outcomes after starting HAART
[104]. These findings are in agreement with those from Mocroft et al.
[105] who found that response to treatment was similar between the
sexes. After adjustment for potential confounders, Le Moing et al.
[106] found a higher short-term increase in CD4+ cell counts in women
initiating a protease inhibitor (PI)-containing regimen compared with
men, but 4 months after initiation of therapy the increase in CD4+
cell counts was similar for men and women. Another observational
cohort study found no statistically significant sex difference in
disease progression within any of the HIV RNA or CD4 lymphocyte count
strata [107]. This is consistent with the findings from two recent
studies from the Antiretroviral Cohort Collaboration [108,109].
Adverse drug reactions TOP

NRTIs are the backbone of HAART and continue to be widely used
worldwide. A few studies found higher rates of adverse events, such as
neuropathy, pancreatitis, and toxicity-driven regimen changes, among
women taking NRTIs (didanosine) in comparison with men [110,111]. In
one of these studies, women taking nucleoside analogue combination
therapy were more likely to reduce dose and develop severe symptoms
than men, whereas men had higher rates of severe laboratory
abnormalities than women [110].

PI-therapy heralded in the HAART era and with them began reports of
side effects. Along with higher drug concentrations in women than men
with the same dose, a few studies have found concomitant side effects
to be more frequent in women than in men, including GI symptoms,
allergic reactions, nephrolithiasis, paresthesias, and taste
perversion with PIs [112-114].

Non-nucleoside reverse transcriptase inhibitors (NNRTIs) are
increasingly important components of HAART; especially the use of
nevirapine in women since it is generally less expensive, easier to
administer, and is effective at reducing vertical transmission [115].
Nonetheless, sex-specific side effects have also been found with
NNRTIs, such as an increase risk of rash in women taking nevirapine or
efavirenz compared with men given the same dose of drug [116-119].

One multicenter study found that women who started HAART were 1.4
times more likely than men to interrupt at least one drug because of
toxicity [120]. However, the risk of developing a clinical HIV-related
event was found to be no different between women and men. Whether
women are more likely to switch medications than men deserves further
study and may be influenced by many factors, including the severity or
tolerance of adverse events and the decision of the patient and health
care provider to make this change.

Metabolic effects

Body habitus changes
Prior to the advent of HAART involuntary weight loss and muscle
wasting were unique identifying characteristics of people with HIV
infection. In the HAART era, changes in body habitus have become
significant clinical problems. Several investigators have identified a
syndrome characterized by the redistribution of body fat, from the
extremities and face to the trunk, often called lipodystrophy. While
the true nature and etiology of body habitus changes in HIV infection
are ill-defined, despite efforts to improve the case definition [121],
many patients have been greatly troubled by them, and may discontinue
otherwise effective antiretroviral therapies in the belief that
stopping therapy may cause these changes to reverse.

As women and men naturally differ in the amount and distribution of
body fat and muscle mass, there has been speculation that sex
differences in metabolic changes are possible. One large study
measured body composition by bioelectrical impedance analysis and
found that control men and women weighed more and had more body cell
mass, fat-free mass, and fat than did HIV-infected men but not
HIV-uninfected women [122]. However, differences in body composition
between HIV-infected and control groups were strongly influenced by
sex. Of the differences in weight between HIV-infected and uninfected
subjects, fat-free mass accounted for 51% in men but only 18% in
women, in whom the remainder was fat. Race and environment had smaller
effects than sex and HIV infection. This study did not consider the
effect of HIV therapy.

Several studies have evaluated the sex-specific metabolic effects of
HIV therapy [123-128]. One study found adverse metabolic effects of a
PI-containing HAART regimen were more pronounced in women than in men
[123], but numbers were small: female HIV-infected patients seem to
lose part of their natural protection from arteriosclerosis during
therapy. In a large multicenter study, the authors found that women
were at higher risk of developing antiretroviral treatment-related
adipose tissue abnormalities than men and that women show a particular
and complex pattern of these abnormalities [128].

Case reports and studies found that the majority of HIV-infected women
with fat changes reported breast enlargement, increases in abdominal
fat and decreases in lower limb fat [125-127]. These reports have been
in women taking only NRTIs or PI-containing HAART regimens. Two larger
cohort studies [129,130], have evaluated lipodystrophy in women. The
first study found that women after the initiation of PI-containing
regiment had a higher risk of isolated peripheral atrophy when
compared with men. The second study found a significantly higher
incidence of peripheral and central lipoatrophy in HIV-infected women
compared with HIV-uninfected women after adjusting for age and race.

Insulin resistance
Outside of the realm of HIV infection, there have been numerous
reports of sex-specific differences regarding insulin resistance. For
example, women have higher insulin levels and a greater degree of
insulin resistance than men [131-133] and among individuals with
insulin resistance syndrome hypertension is more common in women than
men [134].

HIV-infected women with normal weight or weight <90% ideal body weight
had higher mean insulin levels and higher mean insulin: glucose ratio
when compared with healthy controls, unrelated to PI usage [124]. This
study also found that HIV-positive women with elevated trunk to
extremity ratio as measured on dual X-ray absorptiometry (DXA) had
increased insulin: glucose ratio when compared with HIV-positive women
with normal trunk to extremity ratio.

Studies of HIV-uninfected individuals with acquired and congenital
lipodystrophy also demonstrated an association with insulin
resistance, hypertriglyceridemia, and steatosis [135,136]. As HIV is
also associated with lipoatrophy, there has been postulation that
there will be more insulin resistance in people with HIV infection.
Although the insulin resistance seen in HIV has been attributed to the
changes in fat distribution, these metabolic changes occurred early in
the course of antiretroviral treatment and before any change in fat
distribution [137]. PI therapy leads to small increases in glucose and
significant increases in plasma insulin levels indicative of insulin
resistance [137,138]. In addition, strong epidemiologic and
pathophysiologic evidence in non-HIV-infected patients exists, linking
visceral obesity with insulin resistance and diabetes mellitus
[139,140]. Thus changes in fat distribution may worsen insulin
resistance and, given that women may be more predisposed to develop
insulin resistance, further investigation of sex differences in
insulin resistance among people with HIV is warranted.

Bone metabolism
An increased prevalence of osteoporosis and osteopenia has been
reported in cross-sectional investigations of HIV-infected men. An
association with PI therapy has been suggested [141], although others
[142,143] suggested that factors other than antiretroviral therapy
might increase osteopenia in HIV infection including HIV itself [144].
A study comparing bone density collected from HIV-infected male
patients with wasting in the pre-HAART era and those on HAART found
lower bone mineralization measures on average when compared with
matched controls [144]. However, another study in healthy HIV-infected
men in the pre-HAART era found only marginally lower bone mineral
density (BMD) at the lumbar spine and no difference in total body or
hip BMD by DXA when compared to controls [145]. The determinants of
osteopenia in HIV infection remain inconclusive and clearly needs to
be studied in women, in whom the fracture rate is higher.

A small study of HIV-infected women with AIDS wasting found decreased
bone mineral density when compared to HIV-uninfected matched controls
and no association between PI use and BMD among the HIV-infected women
[146]. A case report of osteoporosis with vertebral fracture in two
young HIV-infected African women also suggests factors other than
antiretroviral therapy [143]. A recent study by Dolan et al. [147]
suggested that altered nutritional status, hormonal function and body
composition may contribute the observed lower bone density in
HV-infected women compared to uninfected women.

Cardiovascular events
The incidence of cardiovascular events may be increased due to
atherosclerosis resulting from prolonged exposure to HAART [148-150].
One study assessed the distribution of risk factors for coronary heart
disease in men and women treated with PI-containing regimens with the
distribution of these factors in the general HIV-uninfected population
[151]. They found a higher prevalence of tobacco smoking, waist-to-hip
ratio, higher triglyceride levels, and lower high-density lipoprotein
cholesterol levels in PI-treated individuals than in the general
population. Treated HIV-infected women had higher total cholesterol
than uninfected women, which was not observed when comparing treated
HIV-infected men with uninfected men. In conclusion, there was an
excess distribution of risk factors for cardiovascular events in both
HIV-infected men and women.

Conclusions
In developed countries and before the widespread use of HAART there
was little evidence for sex differences in disease progression
following HIV seroconversion. If anything, women live somewhat longer,
which is also the case for the uninfected population [152]. Much less
is known about these sex differences in developing countries. Often,
differences in progression rates between men and women can be
explained by underlying conditions, such as inequities in social
support, access to health care and socio-economic status.

Studies consistently showed that women have higher CD4+ cell count and
lower HIV RNA levels the first years of their HIV infection, but they
were inconclusive regarding whether these differences diminish over
time. Many studies adjusted the female-male HIV RNA comparisons for
CD4+ cell counts, assuming that CD4+ cell counts do not differ by sex.
If this is not true, at least for the first years following HIV
seroconversion, this adjustment will bias towards higher HIV RNA
levels in women and smaller HIV RNA sex differences. Additional
studies in HIV seroconverters over a prolonged period of time are
needed to evaluate whether and when sex differences in both markers
change during the natural course of HIV infection, although this might
be difficult to do in the presence of HAART. Along with studies on the
underlying biological mechanisms, studies on the sex differences in
HIV RNA and CD4 cell count trajectories in HIV-infected adults on
HAART might help to better understand the marker dynamics.

Studies performed in the pre-HAART era have led to the conclusion that
there is no harmful effect of pregnancy on HIV disease progression, at
least in developed countries. However, the increasing complexity of
clinical management of pregnant women in the HAART era makes it
necessary to continue monitoring both mother-to-child transmission and
clinical impact of pregnancy.

There has been limited research regarding the effect of HIV infection
on amenorrhea, ovulation, menopause, and osteoporosis. The prolonged
life expectancy due to HAART makes it necessary to better describe
these women-specific conditions in well-designed prospective studies
of women. Unlike CIN, invasive cervical cancer has not been found to
be higher in HIV-infected women with regular access to health care
than in well-matched HIV-uninfected women. The high incidence of CIN
in HIV-infected women and the conflicting results regarding potential
regression in women on HAART make it necessary to offer HIV-infected
women regular gynecological follow-up, including Pap smear testing,
colposcopy and biopsy if indicated.

Initiation of therapy may be delayed in women due to the sex
differences in laboratory markers of HIV disease progression.
Consequently women might benefit less from HAART than men under
current treatment guidelines. However, seroconverter studies failed to
demonstrate a major difference between men and women in the rate of
disease progression at the population level in the HAART era. In
addition, women responded equally well or even better to HAART than
men. Therefore, there is little support for revision of the CD4 cell
count and HIV RNA thresholds for HAART initiation for women.
Synthesizing findings, sex differences in CD4+ cell count and HIV RNA
levels are likely to be affected by sex hormones and to have no
functional significance. However, since there is still debate about
the optimal time for initiating HAART, at this moment it might be hard
to detect whether any delay in starting HAART in women resulted in
disadvantage in terms of progression. Alternatively, the large benefit
of HAART might overrule disadvantages due to the later start in women.
Nonetheless, as follow-up time in the HAART era is still limited, a
sex difference in the long term cannot be excluded.

Our review of sex differences and HIV infection may be subject to
publication bias, meaning that some studies may have evaluated sex
difference but failed to report these results if they were
insignificant or inconclusive. Although publication bias cannot be
ruled out, published studies suggest higher rates of drug-related
adverse events among HIV-infected women on therapy in comparison with
men. Administering the same drug dosage to women and men at the same
frequency, without considering physical differences, might predispose
women to develop dose-related adverse reactions to certain drugs.
Adjusting the adult dosage for weight might correct for some of these
adverse reactions but sex-specific effects may still exist. With the
growing number of pharmacological agents for the treatment of HIV
infection and their narrow therapeutic index, it becomes especially
important to assess potential sex differences in drug
pharmacodynamics.

The relationship between metabolic abnormalities changes in body
habitus, and endocrine perturbations has been very little studied
until now. Sorting out the effects of HIV itself, specific HIV
therapeutics, and the aging process on the metabolism can only be done
in well-designed studies of men and women.

After reviewing the literature, it is clear that some of the reported
effects of sex on HIV disease progression were a function of social
rather than biological constructs. Nonetheless, whether sex
differences are due to unalterable genetic factors or social and
environmental conditions it is imperative that all HIV-infected
individuals have equal access to interventions that can slow disease
progression.