Objectives: To estimate the current extent of cutane- disease were enrolled; 65.3% had at least 1 skin mani-
ous disorders in a large population of ambulatory and festation during the course of HIV infection. A history
hospitalized human immunodeficiency virus (HIV)– of skin disorder was reported for 269 patients (59.8%),
infected patients and to describe characteristics associ- and 199 patients (44.2%) had clinical dermatologic mani-
ated with the presence of current skin manifestations. festations at the time of the study. The most frequent
causes of cutaneous disorders were infections (fungal, vi-
Design: Cross-sectional survey. ral, and bacterial), neoplasia, and miscellaneous disor-
ders. Skin diseases were more likely to be reported in ho-
Setting: Hospital units participating in the hospital- mosexual and bisexual men and in patients with clinically
based information system of the Groupe d’Epidémiologie and biologically advanced HIV infection.
Clinique du SIDA en Aquitaine, southwestern France.
Conclusions: Cutaneous disorders occur more fre-
Patients: All the patients seen by physicians between No- quently as HIV infection advances and immune func-
vember 18 and December 20, 1996, in the participating tion deteriorates; however, they are common and of vari-
units (hospital ward, outpatient clinic, or day hospital). ous types throughout the course of HIV disease. Taking
cutaneous disorders into consideration for case manage-
Main Outcome Measures: Prevalence and cumula- ment is essential to improve quality of life for HIV-
tive incidence of cutaneous disorders. infected patients.
Results: Four hundred fifty patients at all stages of HIV Arch Dermatol. 1998;134:1208-1212
S
KIN DISORDERS are common text of HIV infection, cutaneous disorders
manifestations of human im- can present with particular clinical mani-
munodeficiency virus (HIV) festations: unusual anatomical sites, in-
disease: they affect between creased severity, treatment failure, and un-
80% and 95% of HIV-infected usual clinical appearance.11 Moreover, it
patients according to the literature,1-5 oc- is argued that some cutaneous disorders
curring at anytime in the course of infec- reflect the progression of HIV disease,4-6
tion. Skin is often the first and only organ but this relation is still controversial.2,9
affected during most of the course of HIV This survey estimated the preva-
disease.1,6,7 lence of cutaneous disorders in a large
population of HIV-infected patients seen
From INSERM Unit 330, For editorial comment in public hospitals in the Aquitaine re-
Université Victor Segalen see page 1290 gion, southwestern France, and we stud-
Bordeaux 2 (Drs Spira and ied the link of these skin manifestations
Dabis), and the Departments of Cutaneous disorders during HIV in- with the progression of HIV disease.
Dermatology (Drs Mignard and fection are numerous.4-9 Some have drawn
Doutre) and Internal Medicine attention because their onset defines some RESULTS
(Dr Morlat), Bordeaux of the Centers for Disease Control and Pre-
University Hospital, Bordeaux,
vention (CDC) acquired immunodefi- SAMPLE CHARACTERISTICS
France. Participants in the
Groupe d’Epidémiologie ciency syndrome (AIDS) clinical catego-
Clinique du SIDA en Aquitaine ries, eg, oral candidiasis, zoster, herpes Four hundred eighty-eight patients were
are listed in the simplex, oral hairy leukoplakia, and Ka- seen in the participating hospital units dur-
Acknowledgment section at the posi sarcoma,10 but most have been docu- ing the study, among whom 38 were ex-
end of the article. mented solely in case reports. In the con- cluded from the analysis because of a lack
*Patients may have more than 1 diagnosis. Values are given as numbers (percentages). HIV indicates human immunodeficiency virus; CDC, Centers for Disease
Control and Prevention.
†Value has not been computed.
to 0.200 3 109/L and 35.2% of those with a count higher AIDS or had at least non–AIDS-defining symptoms. Simi-
than 0.200 3 109/L (P,.001). Prevalence of cutaneous dis- larly, the prevalences of cutaneous disorders reported by
orders was higher among patients receiving a tritherapy Uthayakumar et al5 and Goldstein et al6 were higher than
of antiretroviral drugs than among other patients (56.8% those in our study; however, in the former the study popu-
vs 36.5%; P,.001). This association was not observed any- lation included almost exclusively homosexual and
more after adjustment for CD4 cell count was made, sug- bisexual men, and in the latter, the study setting was a
gesting that this trend was completely explained by the department of dermatology, with an obvious referral bias.
difference in disease advancement between patients treated The GECSA hospital-based information system deals
by protease inhibitors and the others. with at least 80% of the HIV-infected patients known in
Prevalence of the 6 most frequent cutaneous disor- the Aquitaine region.16,17 Thus, our results provide an es-
ders according to clinical and biologic markers of HIV timation of the frequency of skin disorders among a sample
infection is described in Table 4. Because Kaposi sar- that is representative of the current population of HIV-
coma is an AIDS-defining disease, its occurrence con- infected patients followed up by the hospital wards in
cerned exclusively patients in CDC category C. Xerosis, Aquitaine, the third region in France for AIDS reporting.
seborrheic dermatitis, and pruritus prevalences did not Using data from a specific questionnaire filled in by
statistically differ according to the clinical stage of HIV participating physicians in addition to results of the regu-
disease. Oral candidiasis and molluscum contagiosum oc- lar medical examination may still have led to an under-
curred more frequently as HIV infection clinically ad- estimate of the frequency of cutaneous disorders: skin
vanced. All 6 skin disorders except seborrheic dermati- diseases present at the time of the study but for which
tis were significantly more frequent in patients with a CD4 diagnosis requires a careful examination, eg, dermato-
cell count below 0.200 3 109/L than in other patients. phyte infections, may have been missed by a nonspecial-
None of the other observed cutaneous disorders was sta- ist. Moreover, history of minor cutaneous disorders may
tistically associated with the progression of HIV disease have been underestimated in relation to a recall bias, and
(data not shown). previous cutaneous disorders whose occurrence de-
fined a CDC clinical category were more likely to be well
COMMENT remembered by physicians. In addition, our estimation
of prevalence of previous skin disorders used 2 sources
Results of this study show a high prevalence of skin dis- of information for the classifying conditions but only 1
orders among a population of HIV-infected patients seen source for the others because they were not reported in
in public hospitals. To our knowledge, this is the first pub- the GECSA standardized questionnaire: the capture-
lished prevalence survey on cutaneous disorders in pa- recapture method used when information was provided
tients at all stages of HIV disease and followed up in hos- by the 2 sources took into account the cases notified by
pital units that do not specialize in dermatology. each of the 2 sources but corrected the possible overlap.
Furthermore, the survey was carried out after the intro- The overall predominance of skin disorders in ho-
duction of large-scale, combination, antiretroviral drug mosexual and bisexual men compared with the other HIV
therapies. transmission groups, even after exclusion of Kaposi sar-
Our prevalence estimate of skin disorders, 65.3%, coma, may be explained by the fact that these patients
is much lower than that in previous reports: Pitche et al3 usually report more easily their discomfort than the oth-
reported a prevalence of 82.5% in HIV-infected patients ers. Moreover, some skin disorders, such as condyloma
hospitalized in Togo, West Africa; Coldiron and Berg- or molluscum contagiosum, are known to be caused by
stresser2 observed a prevalence of 92.0% among 100 se- sexually transmitted agents.
rial outpatients in Texas; and Goodman et al9 reported Our results show that cutaneous disorders are more
higher prevalence figures of all skin diseases among 117 and more frequent as HIV infection clinically advances
HIV-infected outpatients and inpatients. However, in these and immune function deteriorates, ie, in patients with
3 studies, patients were much more advanced in HIV dis- full-blown AIDS, with a low CD4 cell count, and whose
ease than in our sample: most patients had full-blown status justifies the prescription of a tritherapy of antiret-