This article was downloaded by: [Gazi University] On: 04 February 2015, At: 01:29 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
AIDS Care: Psychological and Socio-medical Aspects of AIDS/HIV Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/caic20
Frequency and correlates of late presentation for HIV infection in France: older adults are a risk group – results from the ANRS-VESPA2 Study, France ab
ab
ab
cd
Kayigan d'Almeida Wilson , Rosemary Dray-Spira , Cindy Aubrière , Christine Hamelin , efg
cd
Bruno Spire , France Lert
& the ANRS-Vespa2 Study Group
a
Department of Social Epidemiology, INSERM, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Paris, France b
Click for updates
Department of Social Epidemiology, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Sorbonne Universités, UPMC Univ Paris 06, Paris, France c
Department of Epidemiology of Occupational and Social Determinants of Health, INSERM, U1018, Center for Research in Epidemiology and Population Health, Villejuif, France d
UFR Sciences de la Santé, University Versailles Saint-Quentin-en-Yvelines, Villejuif, France
e
INSERM UMR912 (SESSTIM), Marseille, France
f
Aix Marseille Université, UMR_S912, IRD, Marseille, France
g
ORS PACA, Observatoire Régional de la Santé Provence-Alpes-Côte-d'Azur, Marseille, France Published online: 14 Apr 2014.
To cite this article: Kayigan d'Almeida Wilson, Rosemary Dray-Spira, Cindy Aubrière, Christine Hamelin, Bruno Spire, France Lert & the ANRS-Vespa2 Study Group (2014) Frequency and correlates of late presentation for HIV infection in France: older adults are a risk group – results from the ANRS-VESPA2 Study, France, AIDS Care: Psychological and Socio-medical Aspects of AIDS/HIV, 26:sup1, S83-S93, DOI: 10.1080/09540121.2014.906554 To link to this article: http://dx.doi.org/10.1080/09540121.2014.906554
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
AIDS Care, 2014 Vol. 26, No. Supplement 1, 83–93, http://dx.doi.org/10.1080/09540121.2014.906554
Frequency and correlates of late presentation for HIV infection in France: older adults are a risk group – results from the ANRS-VESPA2 Study, France Kayigan d’Almeida Wilsona,b*, Rosemary Dray-Spiraa,b, Cindy Aubrièrea,b, Christine Hamelinc,d, Bruno Spiree,f,g, France Lertc,d and the ANRS-Vespa2 Study Group a
Department of Social Epidemiology, INSERM, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Paris, France; bDepartment of Social Epidemiology, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Sorbonne Universités, UPMC Univ Paris 06, Paris, France; cDepartment of Epidemiology of Occupational and Social Determinants of Health, INSERM, U1018, Center for Research in Epidemiology and Population Health, Villejuif, France; dUFR Sciences de la Santé, University Versailles Saint-Quentin-en-Yvelines, Villejuif, France; eINSERM UMR912 (SESSTIM), Marseille, France; fAix Marseille Université, UMR_S912, IRD, Marseille, France; gORS PACA, Observatoire Régional de la Santé Provence-Alpes-Côte-d’Azur, Marseille, France
Downloaded by [Gazi University] at 01:29 04 February 2015
(Received 5 November 2013; accepted 13 March 2014) Correlates of late presentation (LP) for HIV infection in Metropolitan France and French overseas departments (FODs) were assessed among HIV-infected patients recently diagnosed, using data from a large cross-sectional survey, representative of the French HIV-infected population, conducted in 2011. LP was defined as presentation with either clinical AIDS events within the calendar year of diagnosis or CD4 < 350/mm3 and presentation with advanced disease (PAD) was defined as presentation with either clinical AIDS events or CD4 < 200/mm3. Correlates of LP/PAD were assessed through logistic modelling, separately in Metropolitan France and FODs. In Metropolitan France, 47.7% of participants were late presenters and 29.3% presented with advanced disease. LP was more frequent among male and female migrants from sub-Saharan Africa (SSA; 58.5% and 56.4%) and non-African heterosexual males (61.8%) than among men who have sex with men (34.8%). In FODs, 53.2% of participants were late presenters and 36.8% presented with an advanced disease. LP was more frequent among men than women (60.6% vs. 45.3%) and among those with a lower level of education (56.6% vs. 47.5%). A consistent positive association was found in adjusted analyses between LP/PAD and increasing age at diagnosis among all subpopulations, in both settings. In Metropolitan France, among men who have sex with men, those self-declaring as bisexual were at higher risk of LP/PAD; among non-African heterosexual males and females, religiosity was associated with increased risk of LP/PAD; and among SSA migrants, those diagnosed within the year following their arrival in France were at higher risk of LP/PAD. Older age at diagnosis is a major risk factor for LP/PAD independently of any other socio-demographic characteristics. Promotion of HIV testing should be renewed to target each subgroup at risk while paying a particular attention to middle-aged or older adults whose attitudes and beliefs towards HIV/AIDS might prevent them from seeking testing.
Keywords: HIV; late presentation; screening; older age; France
Introduction In the recent years, consistent findings have contributed to clearly establish that expanded testing, early diagnosis and early treatment of HIV infection reduce morbidity, mortality and further transmission of the virus (Braitstein et al., 2006; Cohen et al., 2011; HIV-CAUSAL Collaboration et al., 2010). Thus, in the era of highly effective antiretroviral therapies (ART), delayed presentation to care represents a missed opportunity for both HIVpositive individuals and the community. Since the mid-2000s, national health authorities throughout high-income countries have released guidelines to encourage early testing (Branson et al., 2006; Health Protection Agency, 2010; Ministère de la Santé et des Sports, 2010). However, despite promotion campaigns and easy access to testing facilities, nearly a *Corresponding author. Email: [email protected] © 2014 Taylor & Francis
third of HIV-infected persons are still diagnosed at an advanced stage disease in Europe (Mocroft et al., 2013). Therefore, late presentation (LP) remains a major barrier to reducing the burden of HIV infection and controlling the epidemic (May et al., 2011; Montlahuc et al., 2013a). Many studies have assessed the correlates of late diagnosis of HIV infection in Europe (Camoni, Raimondo, Regine, Salfa, Suligoi, & Regional representatives of the HIV Surveillance System, 2013; Celesia et al., 2013; Couturier et al., 1998; Delpierre et al., 2007; Ndiaye et al., 2011; Vives et al., 2012; Zoufaly et al., 2012) and found that older age, low-education level, male sex and foreign birth increased the risk of delayed diagnosis. Nevertheless, most studies were restricted to specific subpopulations and were not representative of the whole newly diagnosed HIV-infected population.
S84
K. d’Almeida Wilson et al.
The purpose of the present paper was to assess the correlates of LP for HIV infection in France, using data from a nationally representative study carried out in 2011–2012.
Material and methods
Downloaded by [Gazi University] at 01:29 04 February 2015
Study design We used data from the ANRS-VESPA2 study, a large national representative survey primarily aimed at assessing the socio-economic and clinical situation of HIVinfected people in Metropolitan France and French overseas departments (FODs) of Guadeloupe, Martinique, La Reunion, French Guiana (Cayenne) and St Martin (French part). The study design has been detailed elsewhere (Dray-Spira, Spire, Lert, & le groupe VESPA2, 2013). Data were collected from January 2011 to January 2012 in 81 hospital outpatient departments randomly selected among all hospitals delivering HIV care in France. All adult outpatients (aged 18 or older), HIV-infected for at least 6 months, and either French citizens or living in France for at least 6 months, were eligible. Physicians invited a sample of eligible patients randomly selected according to the order of appointments to participate in the study. After signing an informed consent, participants answered a face-to-face standardised questionnaire in French or Creole language in FODs. The questionnaire covered a wide range of information on social status, occupation, income, sexual behaviour, use of medical care, mental health and social network at the time of data collection. Main bio-clinical HIV-related markers and comorbidities were documented directly from the patient’s medical record. The ANRS-VESPA2 study was approved by the French National Commission for Data Protection and Liberties (CNIL).
Participants In Metropolitan France, a total of 5617 individuals were randomised. Of them, 378 were not solicited (212 because of major cognitive or health impairment and 166 due to language barriers) and 1767 declined participation. Thus, 3022 participants were included. In FODs, a total of 1065 individuals were randomised, 125 were not solicited (83 because of major cognitive or health impairment and 42 because of language barriers) and 317 declined to participate. Thus, 598 participants were included. Compared to non-participants, participants were more likely to be French citizens, unemployed, infected through homo/bisexual contacts and to have good biological parameters (CD4 cell count >500/ mm3 or undetectable viral load). A weighting procedure taking into account the participation bias was
implemented in order to allow extrapolating the results to the entire eligible population. Variables of interest LP was defined according to the CD4 cell count at presentation and the clinical stage within the calendar year of diagnosis. We considered that CD4 at presentation was missing when the delay between diagnosis and the first CD4 cell count exceeded one calendar year. Participants who presented with 350 (N = 457)
Late presenters CD4 < 350 (N = 519)a
Presentation with advanced stage disease CD4 < 200 (N = 308)a
Unknown presentation status (N = 120)
Total (N = 1096)
320 (43.7) 137 (36.0)
329 (45.0) 190 (52.3)
204 (28.9) 104 (29.8)
74 (11.3) 46 (11.8)
723 373
259 2 6 23 83 34 50
174 6 3 66 83 13 14
47 3 4 10 29 13 14
(9.5) (19.9) (27.0) (13.8) (9.7) (11.6) (17.1)
480 11 13 99 253 130 110
19 (10.4) 101 (11.8)
207 886
(55.7) (20.3) (36.6) (27.6) (33.9) (26.6) (42.7)
(34.8) (59.8) (36.4) (58.5) (56.4) (61.8) (40.2)
69 (34.4) 388 (42.7)
119 (55.2) 397 (45.6)
129 156 119 41 12
(47.1) (42.4) (37.8) (36.7) (19.1)
93 168 142 81 35
(36.7) (44.7) (53.3) (59.0) (73.6)
302 16 5 114 20
(48.6) (70.1) (19.7) (31.4) (40.2)
251 7 16 220 25
(41.0) (27.9) (49.7) (56.7) (45.1)
85 111 97 164
(36.9) (32.7) (43.3) (53.4)
135 171 98 113
(52.6) (57.4) (43.3) (34.1)
99 3 2 43 80 59 22
(19.6) (30.2) (20.4) (41.2) (34.2) (45.3) (17.4)
80 (37.8) 225 (26.8) 43 99 86 54 26
(18.6) (27.4) (34.5) (38.6) (52.5)
39 49 23 7 2
(16.2) (12.9) (8.8) (4.2) (6.5)
261 373 284 129 49
146 (23.9) 2 (11.4) 11 (35.9) 133 (36.5) 16 (26.2)
62 1 4 45 8
(10.3) (2.0) (30.6) (11.9) (14.7)
615 24 25 379 53
31 27 27 34
(10.5) (10.0) (13.4) (12.6)
251 309 222 311
84 97 63 63
(34.9) (34.3) (27.2) (18.9)
Note: ANRS-VESPA 2 study, weighted frequencies. a Non-exclusive categories. b Per cent of non-late presenters, late presenters, presenters with advanced stage disease, unknown presentation status among the whole study population including those with a missing CD4 cell count at presentation.
S86
K. d’Almeida Wilson et al.
Table 2. Factors associated with presentation with advanced stage disease or LP, among French MSM, in logistic regression models. LP
Downloaded by [Gazi University] at 01:29 04 February 2015
Characteristics
Number of participants
Age at diagnosis (years) 60 28 Educational attainment Basic or low 34 Secondary or 446 higher Self-reported sexual orientation Homosexual 385 Other 91 Religiosity Not important at all 219 Not so important 132 Important 97 Very important 30
%a
OR (95% CI)
Presentation with advanced stage disease aOR (95% CI)
%b
1 2.0 2.5 6.8 7.3
5.1 1 14.7 3.2 (1.1–9.1)* 24.7 6.1 (2.1–17.6)* 40.2 12.3 (3.9–39.1)* 60.2 27.7 (7.9–97.2)*
OR (95% CI)
aOR (95% CI)
19.6 33.7 38.2 64.6 70.9
1 2.0 2.5 7.3 9.7
52.4 36.9
1 0.5 (0.2–1.2)
1 1.0 (0.4–2.5)
47.4 19.2
1 0.3 (0.1–0.6)*
1 0.5 (0.2–1.4)
52.1 34.9
1 2.0 (1.1–3.6)*
1 1.4 (0.7–2.7)
18.9 31.4
1 1.9 (1.03–3.7)*
1 0.9 (0.5–1.8)
1 1.0 (0.6–1.8) 1.3 (0.6–2.9) 1.2 (0.4–3.7)
18.8 21.4 23.7 34.5
1 1.2 (0.6–2.3) 1.3 (0.7–2.7) 2.3 (0.9–5.9)
1 1.1 (0.6–2.2) 1.1 (0.5–2.3) 1.6 (0.5–4.8)
(0.9–4.4) (1.1–5.5)* (2.6–20.1)* (3.2–29.1)*
34.6 1 35.2 1.0 (0.5–1.9) 46.4 1.6 (0.8–3.1) 47.3 1.7 (0.7–4.2)
(0.9–4.2) (1.1–5.3)* (2.4–19.7)* (2.2–24.0)*
1 3.1 5.6 11.9 21.4
(1.1–9.0)* (1.9–16.4)* (3.7–38.2)* (5.9–76.9)*
Note: ANRS-VESPA2 study, Metropolitan France (2003–2011). N = 480. OR, Odds ratio; aOR, adjusted OR. a Per cent of presentation with advanced stage disease among those with a CD4 cell count at presentation. b Per cent of LP among those with a CD4 cell count at presentation. *p < 0.05.
Age at diagnosis ranged from 16 to 85 years with a median of 36 years (Inter Quartile Range [IQR] = 30–45). In the study population, 47.7% were late presenters, 29.3% presented with advanced disease and 11.5% had an unknown status at presentation (Table 1). The proportion of missing CD4 cell count at presentation decreased with age at diagnosis (11% among those aged 40 years, p = 0.01) and increased with time since diagnosis (5.4% when diagnosis occurred after 2008 vs. 16% for diagnosis in 2003–2004). LP increased with age at diagnosis (from 36.7% under 30 to 73.6% among those aged over 59, p = 0.0006) and was higher among male/female SSA migrants and non-African heterosexual males compared to MSMs (respectively, 58.5%, 56.4%, 61.8% vs. 34.8%, p < 0.0001). Results were similar for presentation with advanced disease (PAD). Factors associated with LP In multivariate analyses, LP significantly increased with age at diagnosis among MSM [adjusted odds ratio (aOR; 95% CI) for >40 years vs. 350 (N = 158)
Late presenters: CD4 < 350 (N = 243)a
Presentation with advanced stage disease CD4 < 200 (N = 174)a
Unknown presentation status (N = 40)
Total (N = 441)
75 (31.1) 83 (46.5)
149 (60.6) 94 (45.3)
112 (41.9) 62 (31.2)
22 (8.3) 18 (8.2)
246 195
63 (38.9) 95 (38.4)
93 (48.5) 150 (56.1)
70 (36.5) 104 (36.9)
24 (12.6) 16 (5.4)
180 261
5 10 15 5 5
(7.3) (8.2) (7.7) (9.0) (10.0)
88 122 117 66 48
22 (6.8) 18 (10.6)
262 179
13 11 9 7
(12.2) (7.0) (5.9) (9.0)
89 144 119 89
40 40 36 24 18
(50.1) (40.7) (29.6) (36.9) (40.8)
93 (36.6) 65 (41.9) 44 51 39 24
(54.0) (41.2) (32.1) (25.2)
43 72 66 37 25
(42.6) (51.1) (62.7) (54.0) (49.2)
147 (56.6) 96 (47.5) 32 82 71 58
(33.9) (51.7) (61.9) (65.9)
31 54 44 27 18
(28.6) (37.4) (41.1) (37.4) (35.9)
114 (41.9) 60 (28.2) 18 58 52 46
(20.7) (33.6) (42.6) (52.6)
83 (40.6) 10 (32.0) 65 (37.6)
125 (49.4) 20 (66.4) 98 (54.8)
86 (32.0) 17 (53.4) 71 (38.7)
21 (10.0) 1 (1.5) 18 (7.6)
229 31 181
59 64 21 13
121 79 24 15
94 49 19 10
19 12 3 5
(8.3) (6.6) (7.3) (17.4)
199 155 48 33
(34.5) (42.6) (49.5) (36.0)
(57.2) (50.8) (43.2) (46.6)
(42.2) (31.8) (30.4) (32.2)
4 (19.8) 15 (38.0) 139 (39.6)
7 (69.1) 22 (59.4) 210 (51.8)
3 (21.2) 13 (33.6) 154 (36.7)
1 (11.1) 2 (2.6) 36 (8.6)
12 39 385
79 (44.1) 77 (33.0)
98 (47.2) 144 (59.1)
67 (30.8) 106 (42.6)
25 (8.7) 15 (7.8)
202 236
15 (43.0) 142 (38.1)
19 (43.9) 222 (54.0)
11 (27.3) 161 (37.4)
4 (13.1) 36 (7.9)
38 400
Note: ANRS-VESPA 2 study, weighted frequencies. a Non-exclusive categories. b Per cent of non-late presenters, late presenters, presenters with advanced stage disease, unknown presentation status among the whole study population including those with a missing CD4 cell count at presentation.
(Fernández-Dávila et al., 2013; Ross et al., 2013), which might be due to a greater fear of social consequences among those men with high levels of internalised homophobia (Herek, Gillis, & Cogan, 2009). Those associations were no longer significant after adjustment for age at diagnosis, suggesting that these factors might
be combined to explain LP, since older generations were less educated and less inclined to self-declare as gay. Among non-African heterosexual males and females, those who declared religion as important in their lives had a higher risk of LP. Such an association was not found among SSA migrants; indeed, religiosity was
S90
K. d’Almeida Wilson et al.
Table 6. Factors associated with presentation with advanced stage disease or LP, among HIV-infected individuals, in logistic regression models.
Downloaded by [Gazi University] at 01:29 04 February 2015
Characteristics
All Observed number
Gender Male Female Diagnosis period 1996–2002 2003–2011 Department Martinique Guadeloupe French Guiana La Reunion Saint Martin Educational attainment Basic or lower Secondary or higher Age at diagnosis (years) 49 Nationality Native French Naturalized French Foreigner Religiosity Very important Important Not so important Not important at all Injecting drug use Yes Not at time of diagnosis Never Declared sexual orientation Homosexuals Others Cohabiting couple Yes No
LP %
a
Presentation with advanced stage disease
OR (95% CI)
aOR (95% CI)
%b
1 0.6 (0.4–1.0)
41.9 31.3
1 0.6 (0.4–1.0)
36.5 36.9
1 0.9 (0.6–1.5)
28.6 37.4 41.1 37.4 35.9
1 1.5 1.7 1.5 1.4
OR (95% CI)
aOR (95% CI)
246 195
60.6 45.3
1 0.5 (0.3–0.8)*
180 261
48.5 56.1
1 0.7 (0.5–1.1)
88 122 117 66 48
42.6 51.1 62.7 54.0 49.2
1 1.4 2.3 1.6 1.3
262 179
56.5 47.5
1 0.7 (0.4–1.1)
1 0.6 (0.3–0.9)*
41.9 28.2
1 0.5 (0.4–0.9)*
1 0.5 (0.3–0.8)*
89 144 119 89
33.9 51.7 61.9 65.9
1 2.1 (1.1–4.1)* 3.2 (1.6–6.3)* 3.8 (1.8–8.0)*
1 2.0 (0.9–4.4) 2.4 (1.1–5.4)* 3.6 (1.5–8.7)*
20.7 33.6 42.6 52.6
1 1.9 (0.9–4.3) 2.8 (1.3–6.4)* 4.2 (1.8–9.9)*
1 1.9 (0.8–4.1) 2.3 (1.03–5.1)* 3.4 (1.4–8.2)*
229 31 181
49.4 66.4 54.8
1 2.0 (0.8–5.3) 1.2 (0.8–2.0)
32.0 53.4 38.7
1 2.4 (0.9–6.3) 1.3 (0.8–2.2)
199 155 48 33
46.6 43.2 50.8 57.2
1 0.8 (0.4–1.3) 0.6 (0.3–1.2) 0.6 (0.3–1.4)
32.2 30.4 31.8 42.2
1 0.6 (0.4–1.1) 0.6 (0.3–1.2) 0.6 (0.3–1.5)
12 39 385
69.1 59.4 51.8
1 0.6 (0.1–2.7) 0.5 (0.1–1.6)
21.2 33.6 36.7
1 1.9 (0.4–8.5) 2.1 (0.6–7.8)
38 400
43.9 54.0
1 1.5 (0.7–3.0)
27.3 37.4
1 1.5 (0.7–3.3)
202 236
47.2 59.2
1 1.6 (1.02–2.6)*
30.8 42.6
1 1.6 (1.04–2.7)*
1 1.4 1.7 1.7 1.3
(0.7–2.7) (1.2–4.3)* (0.8–3.1) (0.6–2.8)
(0.7–2.7) (0.9–3.5) (0.8–3.3) (0.6–2.8)
1 1.6 (1.0–2.6)
1 0.6 (0.4–1.0)
(0.7–2.9) (0.9–3.4) (0.7–2.9) (0.6–3.1)
1 1.6 (0.9–2.5)
Note: ANRS-VESPA 2 study, FODs (1996–2011). N = 441. a Per cent of presentation with advanced stage disease among those with a CD4 cell count at presentation. b Per cent of LP among those with a CD4 cell count at presentation. *p < 0.05.
unlikely to discriminate between late presenters and others, as in this subgroup most individual declared religion as important in their lives. Divergent results have been published on the association between religion and HIV testing. Most studies were conducted among African populations and while some showed that religious affiliation was associated with an increased risk of HIV and a lower uptake of HIV testing (Addis et al.,
2013; Hawkes et al., 2013; Mohamed & Mahfouz, 2013), others concluded that strong religious beliefs were unlikely to act as a barrier to accessing HIV testing (Fakoya et al., 2012). More broadly, religiosity has been reported to be negatively associated with use of sexual and reproductive health services (Hall, Moreau, & Trussell, 2012) and with lack of disclosure (Préau, Bouhnik, Roussiau, Lert, & Spire, 2008).
Downloaded by [Gazi University] at 01:29 04 February 2015
AIDS Care LP or PAD was particularly high among SSA migrants. It is of importance to note that the figures we provide include those who declared they had been diagnosed in their country of origin (7%), but we do not know if they had been linked to care prior to their arrival in France. For SSA migrants, PAD was higher among those diagnosed during the year following their migration, probably due to the delay between contamination in the country of origin and date of migration (88% of SSA migrants diagnosed within the year following their migration declared they were contaminated in their country of origin vs. 30% of those diagnosed more than one year after migration). Similar results were reported in a previous analysis of the correlates of LP among HIV-infected SSA migrants living in France (Delpierre et al., 2007).The high level of LP in that subgroup contrasts with the high uptake of HIV tests declared by SSA migrants in France (Lapostolle, Massari, Beltzer, Halfen, & Chauvin, 2013). Further investigation will be necessary to disentangle the respective roles of migration after HIV infection and delayed access to testing when living in France. In FODs, age at diagnosis and education level was associated with LP and PAD, and the association with education level was more marked than in Metropolitan France. LP was associated with gender and cohabiting partnership. Previous studies have reported earlier HIV diagnosis in women (Delpierre et al., 2007) probably because of routine prenatal testing, gynaecological follow-up and higher attendance to primary care. Our study has some limitations. Its cross-sectional design might have led to underestimate late diagnosis due to higher mortality rates among persons diagnosed with advanced HIV infection (Lanoy et al., 2007). Nevertheless, this underestimation is likely to be limited given the decrease in mortality since 1996 and the low mortality rate among HIV-infected people diagnosed since 2003 in Metropolitan France (Aouba et al., 2008). Similarly, late diagnosis might be underestimated because of the non-inclusion of people diagnosed for
AIDS Care: Psychological and Socio-medical Aspects of AIDS/HIV Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/caic20
Frequency and correlates of late presentation for HIV infection in France: older adults are a risk group – results from the ANRS-VESPA2 Study, France ab
ab
ab
cd
Kayigan d'Almeida Wilson , Rosemary Dray-Spira , Cindy Aubrière , Christine Hamelin , efg
cd
Bruno Spire , France Lert
& the ANRS-Vespa2 Study Group
a
Department of Social Epidemiology, INSERM, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Paris, France b
Click for updates
Department of Social Epidemiology, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Sorbonne Universités, UPMC Univ Paris 06, Paris, France c
Department of Epidemiology of Occupational and Social Determinants of Health, INSERM, U1018, Center for Research in Epidemiology and Population Health, Villejuif, France d
UFR Sciences de la Santé, University Versailles Saint-Quentin-en-Yvelines, Villejuif, France
e
INSERM UMR912 (SESSTIM), Marseille, France
f
Aix Marseille Université, UMR_S912, IRD, Marseille, France
g
ORS PACA, Observatoire Régional de la Santé Provence-Alpes-Côte-d'Azur, Marseille, France Published online: 14 Apr 2014.
To cite this article: Kayigan d'Almeida Wilson, Rosemary Dray-Spira, Cindy Aubrière, Christine Hamelin, Bruno Spire, France Lert & the ANRS-Vespa2 Study Group (2014) Frequency and correlates of late presentation for HIV infection in France: older adults are a risk group – results from the ANRS-VESPA2 Study, France, AIDS Care: Psychological and Socio-medical Aspects of AIDS/HIV, 26:sup1, S83-S93, DOI: 10.1080/09540121.2014.906554 To link to this article: http://dx.doi.org/10.1080/09540121.2014.906554
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
AIDS Care, 2014 Vol. 26, No. Supplement 1, 83–93, http://dx.doi.org/10.1080/09540121.2014.906554
Frequency and correlates of late presentation for HIV infection in France: older adults are a risk group – results from the ANRS-VESPA2 Study, France Kayigan d’Almeida Wilsona,b*, Rosemary Dray-Spiraa,b, Cindy Aubrièrea,b, Christine Hamelinc,d, Bruno Spiree,f,g, France Lertc,d and the ANRS-Vespa2 Study Group a
Department of Social Epidemiology, INSERM, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Paris, France; bDepartment of Social Epidemiology, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Sorbonne Universités, UPMC Univ Paris 06, Paris, France; cDepartment of Epidemiology of Occupational and Social Determinants of Health, INSERM, U1018, Center for Research in Epidemiology and Population Health, Villejuif, France; dUFR Sciences de la Santé, University Versailles Saint-Quentin-en-Yvelines, Villejuif, France; eINSERM UMR912 (SESSTIM), Marseille, France; fAix Marseille Université, UMR_S912, IRD, Marseille, France; gORS PACA, Observatoire Régional de la Santé Provence-Alpes-Côte-d’Azur, Marseille, France
Downloaded by [Gazi University] at 01:29 04 February 2015
(Received 5 November 2013; accepted 13 March 2014) Correlates of late presentation (LP) for HIV infection in Metropolitan France and French overseas departments (FODs) were assessed among HIV-infected patients recently diagnosed, using data from a large cross-sectional survey, representative of the French HIV-infected population, conducted in 2011. LP was defined as presentation with either clinical AIDS events within the calendar year of diagnosis or CD4 < 350/mm3 and presentation with advanced disease (PAD) was defined as presentation with either clinical AIDS events or CD4 < 200/mm3. Correlates of LP/PAD were assessed through logistic modelling, separately in Metropolitan France and FODs. In Metropolitan France, 47.7% of participants were late presenters and 29.3% presented with advanced disease. LP was more frequent among male and female migrants from sub-Saharan Africa (SSA; 58.5% and 56.4%) and non-African heterosexual males (61.8%) than among men who have sex with men (34.8%). In FODs, 53.2% of participants were late presenters and 36.8% presented with an advanced disease. LP was more frequent among men than women (60.6% vs. 45.3%) and among those with a lower level of education (56.6% vs. 47.5%). A consistent positive association was found in adjusted analyses between LP/PAD and increasing age at diagnosis among all subpopulations, in both settings. In Metropolitan France, among men who have sex with men, those self-declaring as bisexual were at higher risk of LP/PAD; among non-African heterosexual males and females, religiosity was associated with increased risk of LP/PAD; and among SSA migrants, those diagnosed within the year following their arrival in France were at higher risk of LP/PAD. Older age at diagnosis is a major risk factor for LP/PAD independently of any other socio-demographic characteristics. Promotion of HIV testing should be renewed to target each subgroup at risk while paying a particular attention to middle-aged or older adults whose attitudes and beliefs towards HIV/AIDS might prevent them from seeking testing.
Keywords: HIV; late presentation; screening; older age; France
Introduction In the recent years, consistent findings have contributed to clearly establish that expanded testing, early diagnosis and early treatment of HIV infection reduce morbidity, mortality and further transmission of the virus (Braitstein et al., 2006; Cohen et al., 2011; HIV-CAUSAL Collaboration et al., 2010). Thus, in the era of highly effective antiretroviral therapies (ART), delayed presentation to care represents a missed opportunity for both HIVpositive individuals and the community. Since the mid-2000s, national health authorities throughout high-income countries have released guidelines to encourage early testing (Branson et al., 2006; Health Protection Agency, 2010; Ministère de la Santé et des Sports, 2010). However, despite promotion campaigns and easy access to testing facilities, nearly a *Corresponding author. Email: [email protected] © 2014 Taylor & Francis
third of HIV-infected persons are still diagnosed at an advanced stage disease in Europe (Mocroft et al., 2013). Therefore, late presentation (LP) remains a major barrier to reducing the burden of HIV infection and controlling the epidemic (May et al., 2011; Montlahuc et al., 2013a). Many studies have assessed the correlates of late diagnosis of HIV infection in Europe (Camoni, Raimondo, Regine, Salfa, Suligoi, & Regional representatives of the HIV Surveillance System, 2013; Celesia et al., 2013; Couturier et al., 1998; Delpierre et al., 2007; Ndiaye et al., 2011; Vives et al., 2012; Zoufaly et al., 2012) and found that older age, low-education level, male sex and foreign birth increased the risk of delayed diagnosis. Nevertheless, most studies were restricted to specific subpopulations and were not representative of the whole newly diagnosed HIV-infected population.
S84
K. d’Almeida Wilson et al.
The purpose of the present paper was to assess the correlates of LP for HIV infection in France, using data from a nationally representative study carried out in 2011–2012.
Material and methods
Downloaded by [Gazi University] at 01:29 04 February 2015
Study design We used data from the ANRS-VESPA2 study, a large national representative survey primarily aimed at assessing the socio-economic and clinical situation of HIVinfected people in Metropolitan France and French overseas departments (FODs) of Guadeloupe, Martinique, La Reunion, French Guiana (Cayenne) and St Martin (French part). The study design has been detailed elsewhere (Dray-Spira, Spire, Lert, & le groupe VESPA2, 2013). Data were collected from January 2011 to January 2012 in 81 hospital outpatient departments randomly selected among all hospitals delivering HIV care in France. All adult outpatients (aged 18 or older), HIV-infected for at least 6 months, and either French citizens or living in France for at least 6 months, were eligible. Physicians invited a sample of eligible patients randomly selected according to the order of appointments to participate in the study. After signing an informed consent, participants answered a face-to-face standardised questionnaire in French or Creole language in FODs. The questionnaire covered a wide range of information on social status, occupation, income, sexual behaviour, use of medical care, mental health and social network at the time of data collection. Main bio-clinical HIV-related markers and comorbidities were documented directly from the patient’s medical record. The ANRS-VESPA2 study was approved by the French National Commission for Data Protection and Liberties (CNIL).
Participants In Metropolitan France, a total of 5617 individuals were randomised. Of them, 378 were not solicited (212 because of major cognitive or health impairment and 166 due to language barriers) and 1767 declined participation. Thus, 3022 participants were included. In FODs, a total of 1065 individuals were randomised, 125 were not solicited (83 because of major cognitive or health impairment and 42 because of language barriers) and 317 declined to participate. Thus, 598 participants were included. Compared to non-participants, participants were more likely to be French citizens, unemployed, infected through homo/bisexual contacts and to have good biological parameters (CD4 cell count >500/ mm3 or undetectable viral load). A weighting procedure taking into account the participation bias was
implemented in order to allow extrapolating the results to the entire eligible population. Variables of interest LP was defined according to the CD4 cell count at presentation and the clinical stage within the calendar year of diagnosis. We considered that CD4 at presentation was missing when the delay between diagnosis and the first CD4 cell count exceeded one calendar year. Participants who presented with 350 (N = 457)
Late presenters CD4 < 350 (N = 519)a
Presentation with advanced stage disease CD4 < 200 (N = 308)a
Unknown presentation status (N = 120)
Total (N = 1096)
320 (43.7) 137 (36.0)
329 (45.0) 190 (52.3)
204 (28.9) 104 (29.8)
74 (11.3) 46 (11.8)
723 373
259 2 6 23 83 34 50
174 6 3 66 83 13 14
47 3 4 10 29 13 14
(9.5) (19.9) (27.0) (13.8) (9.7) (11.6) (17.1)
480 11 13 99 253 130 110
19 (10.4) 101 (11.8)
207 886
(55.7) (20.3) (36.6) (27.6) (33.9) (26.6) (42.7)
(34.8) (59.8) (36.4) (58.5) (56.4) (61.8) (40.2)
69 (34.4) 388 (42.7)
119 (55.2) 397 (45.6)
129 156 119 41 12
(47.1) (42.4) (37.8) (36.7) (19.1)
93 168 142 81 35
(36.7) (44.7) (53.3) (59.0) (73.6)
302 16 5 114 20
(48.6) (70.1) (19.7) (31.4) (40.2)
251 7 16 220 25
(41.0) (27.9) (49.7) (56.7) (45.1)
85 111 97 164
(36.9) (32.7) (43.3) (53.4)
135 171 98 113
(52.6) (57.4) (43.3) (34.1)
99 3 2 43 80 59 22
(19.6) (30.2) (20.4) (41.2) (34.2) (45.3) (17.4)
80 (37.8) 225 (26.8) 43 99 86 54 26
(18.6) (27.4) (34.5) (38.6) (52.5)
39 49 23 7 2
(16.2) (12.9) (8.8) (4.2) (6.5)
261 373 284 129 49
146 (23.9) 2 (11.4) 11 (35.9) 133 (36.5) 16 (26.2)
62 1 4 45 8
(10.3) (2.0) (30.6) (11.9) (14.7)
615 24 25 379 53
31 27 27 34
(10.5) (10.0) (13.4) (12.6)
251 309 222 311
84 97 63 63
(34.9) (34.3) (27.2) (18.9)
Note: ANRS-VESPA 2 study, weighted frequencies. a Non-exclusive categories. b Per cent of non-late presenters, late presenters, presenters with advanced stage disease, unknown presentation status among the whole study population including those with a missing CD4 cell count at presentation.
S86
K. d’Almeida Wilson et al.
Table 2. Factors associated with presentation with advanced stage disease or LP, among French MSM, in logistic regression models. LP
Downloaded by [Gazi University] at 01:29 04 February 2015
Characteristics
Number of participants
Age at diagnosis (years) 60 28 Educational attainment Basic or low 34 Secondary or 446 higher Self-reported sexual orientation Homosexual 385 Other 91 Religiosity Not important at all 219 Not so important 132 Important 97 Very important 30
%a
OR (95% CI)
Presentation with advanced stage disease aOR (95% CI)
%b
1 2.0 2.5 6.8 7.3
5.1 1 14.7 3.2 (1.1–9.1)* 24.7 6.1 (2.1–17.6)* 40.2 12.3 (3.9–39.1)* 60.2 27.7 (7.9–97.2)*
OR (95% CI)
aOR (95% CI)
19.6 33.7 38.2 64.6 70.9
1 2.0 2.5 7.3 9.7
52.4 36.9
1 0.5 (0.2–1.2)
1 1.0 (0.4–2.5)
47.4 19.2
1 0.3 (0.1–0.6)*
1 0.5 (0.2–1.4)
52.1 34.9
1 2.0 (1.1–3.6)*
1 1.4 (0.7–2.7)
18.9 31.4
1 1.9 (1.03–3.7)*
1 0.9 (0.5–1.8)
1 1.0 (0.6–1.8) 1.3 (0.6–2.9) 1.2 (0.4–3.7)
18.8 21.4 23.7 34.5
1 1.2 (0.6–2.3) 1.3 (0.7–2.7) 2.3 (0.9–5.9)
1 1.1 (0.6–2.2) 1.1 (0.5–2.3) 1.6 (0.5–4.8)
(0.9–4.4) (1.1–5.5)* (2.6–20.1)* (3.2–29.1)*
34.6 1 35.2 1.0 (0.5–1.9) 46.4 1.6 (0.8–3.1) 47.3 1.7 (0.7–4.2)
(0.9–4.2) (1.1–5.3)* (2.4–19.7)* (2.2–24.0)*
1 3.1 5.6 11.9 21.4
(1.1–9.0)* (1.9–16.4)* (3.7–38.2)* (5.9–76.9)*
Note: ANRS-VESPA2 study, Metropolitan France (2003–2011). N = 480. OR, Odds ratio; aOR, adjusted OR. a Per cent of presentation with advanced stage disease among those with a CD4 cell count at presentation. b Per cent of LP among those with a CD4 cell count at presentation. *p < 0.05.
Age at diagnosis ranged from 16 to 85 years with a median of 36 years (Inter Quartile Range [IQR] = 30–45). In the study population, 47.7% were late presenters, 29.3% presented with advanced disease and 11.5% had an unknown status at presentation (Table 1). The proportion of missing CD4 cell count at presentation decreased with age at diagnosis (11% among those aged 40 years, p = 0.01) and increased with time since diagnosis (5.4% when diagnosis occurred after 2008 vs. 16% for diagnosis in 2003–2004). LP increased with age at diagnosis (from 36.7% under 30 to 73.6% among those aged over 59, p = 0.0006) and was higher among male/female SSA migrants and non-African heterosexual males compared to MSMs (respectively, 58.5%, 56.4%, 61.8% vs. 34.8%, p < 0.0001). Results were similar for presentation with advanced disease (PAD). Factors associated with LP In multivariate analyses, LP significantly increased with age at diagnosis among MSM [adjusted odds ratio (aOR; 95% CI) for >40 years vs. 350 (N = 158)
Late presenters: CD4 < 350 (N = 243)a
Presentation with advanced stage disease CD4 < 200 (N = 174)a
Unknown presentation status (N = 40)
Total (N = 441)
75 (31.1) 83 (46.5)
149 (60.6) 94 (45.3)
112 (41.9) 62 (31.2)
22 (8.3) 18 (8.2)
246 195
63 (38.9) 95 (38.4)
93 (48.5) 150 (56.1)
70 (36.5) 104 (36.9)
24 (12.6) 16 (5.4)
180 261
5 10 15 5 5
(7.3) (8.2) (7.7) (9.0) (10.0)
88 122 117 66 48
22 (6.8) 18 (10.6)
262 179
13 11 9 7
(12.2) (7.0) (5.9) (9.0)
89 144 119 89
40 40 36 24 18
(50.1) (40.7) (29.6) (36.9) (40.8)
93 (36.6) 65 (41.9) 44 51 39 24
(54.0) (41.2) (32.1) (25.2)
43 72 66 37 25
(42.6) (51.1) (62.7) (54.0) (49.2)
147 (56.6) 96 (47.5) 32 82 71 58
(33.9) (51.7) (61.9) (65.9)
31 54 44 27 18
(28.6) (37.4) (41.1) (37.4) (35.9)
114 (41.9) 60 (28.2) 18 58 52 46
(20.7) (33.6) (42.6) (52.6)
83 (40.6) 10 (32.0) 65 (37.6)
125 (49.4) 20 (66.4) 98 (54.8)
86 (32.0) 17 (53.4) 71 (38.7)
21 (10.0) 1 (1.5) 18 (7.6)
229 31 181
59 64 21 13
121 79 24 15
94 49 19 10
19 12 3 5
(8.3) (6.6) (7.3) (17.4)
199 155 48 33
(34.5) (42.6) (49.5) (36.0)
(57.2) (50.8) (43.2) (46.6)
(42.2) (31.8) (30.4) (32.2)
4 (19.8) 15 (38.0) 139 (39.6)
7 (69.1) 22 (59.4) 210 (51.8)
3 (21.2) 13 (33.6) 154 (36.7)
1 (11.1) 2 (2.6) 36 (8.6)
12 39 385
79 (44.1) 77 (33.0)
98 (47.2) 144 (59.1)
67 (30.8) 106 (42.6)
25 (8.7) 15 (7.8)
202 236
15 (43.0) 142 (38.1)
19 (43.9) 222 (54.0)
11 (27.3) 161 (37.4)
4 (13.1) 36 (7.9)
38 400
Note: ANRS-VESPA 2 study, weighted frequencies. a Non-exclusive categories. b Per cent of non-late presenters, late presenters, presenters with advanced stage disease, unknown presentation status among the whole study population including those with a missing CD4 cell count at presentation.
(Fernández-Dávila et al., 2013; Ross et al., 2013), which might be due to a greater fear of social consequences among those men with high levels of internalised homophobia (Herek, Gillis, & Cogan, 2009). Those associations were no longer significant after adjustment for age at diagnosis, suggesting that these factors might
be combined to explain LP, since older generations were less educated and less inclined to self-declare as gay. Among non-African heterosexual males and females, those who declared religion as important in their lives had a higher risk of LP. Such an association was not found among SSA migrants; indeed, religiosity was
S90
K. d’Almeida Wilson et al.
Table 6. Factors associated with presentation with advanced stage disease or LP, among HIV-infected individuals, in logistic regression models.
Downloaded by [Gazi University] at 01:29 04 February 2015
Characteristics
All Observed number
Gender Male Female Diagnosis period 1996–2002 2003–2011 Department Martinique Guadeloupe French Guiana La Reunion Saint Martin Educational attainment Basic or lower Secondary or higher Age at diagnosis (years) 49 Nationality Native French Naturalized French Foreigner Religiosity Very important Important Not so important Not important at all Injecting drug use Yes Not at time of diagnosis Never Declared sexual orientation Homosexuals Others Cohabiting couple Yes No
LP %
a
Presentation with advanced stage disease
OR (95% CI)
aOR (95% CI)
%b
1 0.6 (0.4–1.0)
41.9 31.3
1 0.6 (0.4–1.0)
36.5 36.9
1 0.9 (0.6–1.5)
28.6 37.4 41.1 37.4 35.9
1 1.5 1.7 1.5 1.4
OR (95% CI)
aOR (95% CI)
246 195
60.6 45.3
1 0.5 (0.3–0.8)*
180 261
48.5 56.1
1 0.7 (0.5–1.1)
88 122 117 66 48
42.6 51.1 62.7 54.0 49.2
1 1.4 2.3 1.6 1.3
262 179
56.5 47.5
1 0.7 (0.4–1.1)
1 0.6 (0.3–0.9)*
41.9 28.2
1 0.5 (0.4–0.9)*
1 0.5 (0.3–0.8)*
89 144 119 89
33.9 51.7 61.9 65.9
1 2.1 (1.1–4.1)* 3.2 (1.6–6.3)* 3.8 (1.8–8.0)*
1 2.0 (0.9–4.4) 2.4 (1.1–5.4)* 3.6 (1.5–8.7)*
20.7 33.6 42.6 52.6
1 1.9 (0.9–4.3) 2.8 (1.3–6.4)* 4.2 (1.8–9.9)*
1 1.9 (0.8–4.1) 2.3 (1.03–5.1)* 3.4 (1.4–8.2)*
229 31 181
49.4 66.4 54.8
1 2.0 (0.8–5.3) 1.2 (0.8–2.0)
32.0 53.4 38.7
1 2.4 (0.9–6.3) 1.3 (0.8–2.2)
199 155 48 33
46.6 43.2 50.8 57.2
1 0.8 (0.4–1.3) 0.6 (0.3–1.2) 0.6 (0.3–1.4)
32.2 30.4 31.8 42.2
1 0.6 (0.4–1.1) 0.6 (0.3–1.2) 0.6 (0.3–1.5)
12 39 385
69.1 59.4 51.8
1 0.6 (0.1–2.7) 0.5 (0.1–1.6)
21.2 33.6 36.7
1 1.9 (0.4–8.5) 2.1 (0.6–7.8)
38 400
43.9 54.0
1 1.5 (0.7–3.0)
27.3 37.4
1 1.5 (0.7–3.3)
202 236
47.2 59.2
1 1.6 (1.02–2.6)*
30.8 42.6
1 1.6 (1.04–2.7)*
1 1.4 1.7 1.7 1.3
(0.7–2.7) (1.2–4.3)* (0.8–3.1) (0.6–2.8)
(0.7–2.7) (0.9–3.5) (0.8–3.3) (0.6–2.8)
1 1.6 (1.0–2.6)
1 0.6 (0.4–1.0)
(0.7–2.9) (0.9–3.4) (0.7–2.9) (0.6–3.1)
1 1.6 (0.9–2.5)
Note: ANRS-VESPA 2 study, FODs (1996–2011). N = 441. a Per cent of presentation with advanced stage disease among those with a CD4 cell count at presentation. b Per cent of LP among those with a CD4 cell count at presentation. *p < 0.05.
unlikely to discriminate between late presenters and others, as in this subgroup most individual declared religion as important in their lives. Divergent results have been published on the association between religion and HIV testing. Most studies were conducted among African populations and while some showed that religious affiliation was associated with an increased risk of HIV and a lower uptake of HIV testing (Addis et al.,
2013; Hawkes et al., 2013; Mohamed & Mahfouz, 2013), others concluded that strong religious beliefs were unlikely to act as a barrier to accessing HIV testing (Fakoya et al., 2012). More broadly, religiosity has been reported to be negatively associated with use of sexual and reproductive health services (Hall, Moreau, & Trussell, 2012) and with lack of disclosure (Préau, Bouhnik, Roussiau, Lert, & Spire, 2008).
Downloaded by [Gazi University] at 01:29 04 February 2015
AIDS Care LP or PAD was particularly high among SSA migrants. It is of importance to note that the figures we provide include those who declared they had been diagnosed in their country of origin (7%), but we do not know if they had been linked to care prior to their arrival in France. For SSA migrants, PAD was higher among those diagnosed during the year following their migration, probably due to the delay between contamination in the country of origin and date of migration (88% of SSA migrants diagnosed within the year following their migration declared they were contaminated in their country of origin vs. 30% of those diagnosed more than one year after migration). Similar results were reported in a previous analysis of the correlates of LP among HIV-infected SSA migrants living in France (Delpierre et al., 2007).The high level of LP in that subgroup contrasts with the high uptake of HIV tests declared by SSA migrants in France (Lapostolle, Massari, Beltzer, Halfen, & Chauvin, 2013). Further investigation will be necessary to disentangle the respective roles of migration after HIV infection and delayed access to testing when living in France. In FODs, age at diagnosis and education level was associated with LP and PAD, and the association with education level was more marked than in Metropolitan France. LP was associated with gender and cohabiting partnership. Previous studies have reported earlier HIV diagnosis in women (Delpierre et al., 2007) probably because of routine prenatal testing, gynaecological follow-up and higher attendance to primary care. Our study has some limitations. Its cross-sectional design might have led to underestimate late diagnosis due to higher mortality rates among persons diagnosed with advanced HIV infection (Lanoy et al., 2007). Nevertheless, this underestimation is likely to be limited given the decrease in mortality since 1996 and the low mortality rate among HIV-infected people diagnosed since 2003 in Metropolitan France (Aouba et al., 2008). Similarly, late diagnosis might be underestimated because of the non-inclusion of people diagnosed for
