Inequalities in HIV disease management and progression in migrants from Latin America and sub-Saharan Africa living in Spain

Authors

Errata

This article is corrected by:

  1. Errata: Corrigendum Volume 15, Issue 1, 64, Article first published online: 3 December 2013

Correspondence: Dr Susana Monge, Centro Nacional de Epidemiología, Instituto de Salud Carlos III, Av. Monforte de Lemos, 5. 28029 Madrid, Spain. Tel: 0034 91 822 26 60; fax: 0034 91 387 78 15; e-mail: smonge@isciii.es

Abstract

Objectives

The objective of the study was to analyse key HIV-related outcomes in migrants originating from Latin America and the Spanish-speaking Caribbean (LAC) or sub-Saharan Africa (SSA) living in Spain compared with native Spaniards (NSP).

Methods

The Cohort of the Spanish AIDS Research Network (CoRIS) is an open, prospective, multicentre cohort of antiretroviral-naïve patients representing 13 of the 17 Spanish regions. The study period was 2004–2010. Multivariate logistic or Fine and Gray regression models were fitted as appropriate to estimate the adjusted effect of region of origin on the different outcomes.

Results

Of the 6811 subjects in CoRIS, 6278 were NSP (74.2%), LAC (19.4%) or SSA (6.4%). For these patients, the follow-up time was 15870 person-years. Compared with NSP, SSA and LAC under 35 years of age had a higher risk of delayed diagnosis [odds ratio (OR) 2.0 (95% confidence interval (CI) 1.5–2.8) and OR 1.7 (95% CI 1.4–2.1), respectively], as did LAC aged 35–50 years [OR 1.3 (95% CI 1.0–1.6)]. There were no major differences in time to antiretroviral therapy (ART) requirement or initiation. SSA exhibited a poorer immunological and virological response [OR 0.8 (95% CI 0.7–1.0) and OR 0.7 (95% CI 0.6–0.9), respectively], while no difference was found for LAC. SSA and LAC showed an increased risk of AIDS for ages between 35 and 50 years [OR 2.0 (95% CI 1.1–3.7) and OR 1.6 (95% CI 1.1–2.4), respectively], which was attributable to a higher incidence of tuberculosis. However, no statistically significant differences were observed in mortality.

Conclusions

Migrants experience a disproportionate diagnostic delay, but no meaningful inequalities were identified regarding initiation of treatment after diagnosis. A poorer virological and immunological response was observed in SSA. Migrants had an increased risk of AIDS, which was mainly attributable to tuberculosis.

Introduction

Unlike other European countries, Spain only started to experience significant economic immigration in the 21st Century. Excluding people from the European Union (EU) and North America, migrants represented 8.8% of the Spanish population in 2010, the largest proportion of whom originated from Latin America and the Spanish-speaking Caribbean (LAC); for these LAC immigrants there is virtually no language barrier and a smaller cultural divide than for migrants from other regions, such as sub-Saharan Africans (SSA) [1].

In the 1980s and 1990s, Spain suffered one of the largest HIV epidemics in Europe, mainly among Spanish injecting drug users (IDUs) and their sexual partners. A shift has been observed in the last decade towards a predominantly sexually transmitted HIV epidemic, and an increasing proportion of migrants among those infected with HIV. In 2010, 38.4% of new HIV diagnoses were in migrants, of whom 21.5% were LAC and 8.0% SSA [2]. This pattern is unique within the EU; in other EU countries, SSA are by far the largest group of HIV-infected migrants [3, 4].

Migrants often experience poor social support and discrimination in host countries which, together with language, cultural or legal barriers to health care, can have a negative impact on their health, even in universal health care contexts, such as in Spain. The health of migrant populations is a fundamental issue for social integration, public health policies and health service planning and delivery. A number of studies in the HIV/AIDS field have addressed health inequalities for migrants living in Europe [5, 6]4 [7-17]. However, most of these studies have focused on the SSA population, while HIV/AIDS in LAC has not been studied other than in the USA.

The aim of the present study was to investigate inequalities in the management and progression of HIV disease in LAC and SSA migrants living in Spain compared with the native population. Specifically, this study analysed the risks of a delay in the diagnosis of HIV infection, access and response to antiretroviral therapy (ART), risk of AIDS and overall survival.

Methods

The Cohort of the Spanish AIDS Research Network (CoRIS) is an open, multicentre, prospective cohort of patients over 13 years of age with confirmed HIV infection who are naïve to ART at entry; recruitment started in 2004. Thirty-one HIV units from 13 of the 17 Autonomous Communities of Spain participate in CoRIS. Ethics approval is obtained and every patient provides written informed consent to participate in the cohort. English and French versions of consent forms are available to encourage participation of migrants with a poor knowledge of Spanish. Detailed descriptions of the cohort have been published previously [18, 19].

Patients are followed up according to routine clinical practice and information is collected on sex, age, mode of HIV transmission, educational level, Centers for Disease Control and Prevention (CDC) stage, date of HIV diagnosis, CD4 count and plasma HIV viral load (VL), along with clinical and treatment information. For each individual, self-referred region of origin is recorded; only patients originating from Spain, LAC or SSA were included in the present study. Administrative censoring was October 2010.

χ2 and Kruskall–Wallis tests were used to evaluate differences across the three regions of origin. To characterize the management and progression of HIV infection, seven different outcomes were analysed: diagnostic delay (DD), time to ART requirement, time to ART initiation, immunological and virological response to ART, time to AIDS diagnosis and time to death.

DD was defined as a CD4 count < 350 cells/μL or an AIDS-defining illness in the first year following HIV diagnosis [20, 21]. A multivariate logistic regression was fitted to analyse the association between region of origin and DD after adjustment for any confounding variables.

ART requirement was defined according to minimum criteria for ART initiation in Spain throughout the study period; that is, CD4 count ≤ 350 cells/μL or an AIDS diagnosis. To assess differences across regions in time to ART requirement, time to ART initiation, time to AIDS diagnosis and overall mortality, a survival analysis was performed, censoring each patient at the date of the last follow-up visit (and, in the case of the first outcome, at the date of ART initiation if it preceded the last follow-up visit). For each survival outcome, patients who fulfilled the outcome definition at recruitment were excluded from that particular analysis. Taking into account the fact that subjects were ART naïve at entry, patients who started ART on the day of recruitment were artificially given 1 day of follow-up. A sensitivity analysis was performed excluding these patients. All four models were adjusted by CD4 count and VL at recruitment, to account for duration of infection at recruitment. Sensitivity analyses excluding patients with missing information for either CD4 count or VL were performed.

A multivariate Fine and Gray regression was fitted to estimate the effect of region of origin, accounting for competing risks and adjusting for any confounders. Deaths occurring prior to the event of interest were considered as competing events in every model [22]. For time to ART requirement, initiating ART before fulfilling the requirement criteria was also considered a competing event. Alternatively, inverse probability weighting (IPW) was performed to adjust a possible informative censoring. Patients who had not had any follow-up visit in the 1 year prior to administrative censoring were considered to be lost to follow-up (LTFU). The probability of not being LTFU in each month of a patient's follow-up was calculated, adjusting by baseline and present CD4 count, VL and clinical and treatment status. A pooled logistic regression was weighted by the inverse of the probability of being not LTFU in each month, using stabilized weights and considering each patient a cluster [23, 24].

For analysis of immunological and virological responses to treatment, the patients included were those who had at least one CD4 count of any value, or a VL measurement > 50 HIV-1 RNA copies/ml, respectively, in the 6 months prior to ART initiation and at least two post-ART determinations, one within the first year. The virological response date was considered the date of the first of two consecutive VLs < 50 copies/ml, and the immunological response date was the first of two consecutive CD4 counts at least 100 cells/μL higher than the pre-ART determination. Patients not experiencing the event were censored at their last CD4 count or VL assessment, respectively.

For every outcome, estimative modelling was performed, retaining variables that produced a change of > 10% in the coefficient of interest, and its interaction with sex was systematically explored. Statistical significance was evaluated using the Wald test and stata software was used (V.11.1; Stata Corporation, College Station, TX).

Results

Socioeconomic and clinical characteristics of the sample

Of the 6811 subjects recruited to CoRIS up to October 2010, 6278 (92.2%) were native Spaniards (NSP; n = 4657; 74.2%), Latin Americans or Caribbeans (LAC; n = 1221; 19.4%) or sub-Saharan Africans (SSA; n = 400; 6.4%). NSP were older, were more frequently infected through IDU and had higher VLs, both at recruitment and at treatment initiation (Table 1). The majority of LAC were men who have sex with men (MSM); they had a higher educational level than the other groups and a higher percentage of them were at CDC stage A at recruitment. SSA had the highest proportions of female patients, those infected via heterosexual transmission, those with DD and those at CDC stage C, and had a lower educational level than the other groups and the lowest CD4 counts at recruitment. Both LAC and SSA were enrolled in CoRIS more frequently in more recent years. Importantly, no differences were found in CD4 counts at ART initiation. Figure 1 shows a flow chart specifying which population contributed to the assessment of each of the outcomes and which exclusion criteria were applied.

Figure 1.

Flow chart for the sample. Populations used to assess each outcome were as follows: (a) patients assessed for delayed diagnosis; (b) patients assessed for risk of death; (c) patients assessed for time until antiretroviral therapy (ART) initiation; (d) patients assessed for risk of AIDS; (e) patients assessed for time until ART requirement; (f) patients assessed for virological response to ART; (g) patients assessed for immunological response to ART.

Table 1. Characteristics of the sample by geographical origin of the patients
 NSPLACSSATotal 
(n = 4657)(n = 1221)(n = 400)(n = 6278)P*
  1. NSP, native Spaniards; LAC, Latin Americans and Caribbeans; SSA, sub-Saharan Africans; CDC, Centers for Disease Control and Prevention; HCV, hepatitis C virus; IQR, interquartile range.
  2. *P-values are for χ2 or Kruskal–Wallis (as appropriate) for differences across regions.
Sex [n (%)]         
Male3884(83.4)986(80.7)183(45.8)5053(80.5)< 0.001
Female773(16.6)235(19.2)217(54.2)1225(19.5)
Age [median (IQR)]37 (30–43)32 (27–38)31 (27–38.5)35 (29–42)<0.001
Mode of transmission [n (%)]         
MSM2364(50.8)770(63.1)8(2.0)3142(50.1)< 0.001
Heterosexual1324(28.4)406(33.2)358(89.5)2088(33.3)
IDU798(17.1)17(1.4)9(2.3)824(13.1)
Other/NA171(3.7)28(2.3)25(6.3)224(3.6)
Educational level [n (%)]         
Secondary or higher2281(49.0)664(54.4)87(21.7)3032(48.3)< 0.001
Primary or none1624(34.9)386(31.6)218(54.5)2228(35.5)
Unknown752(16.1)171(14.0)95(23.8)1018(16.2)
CDC stage [n (%)]         
A3510(75.4)971(79.5)280(70.0)4761(75.8)< 0.001
B425(9.1)91(7.5)32(8.0)548(8.7)
C599(12.9)133(10.9)81(20.3)813(13.0)
Unknown123(2.6)26(2.1)7(1.8)156(2.5)
History of HCV infection [n (%)]         
No3391(72.8)1104(90.4)336(84.0)4831(77.0)< 0.001
Yes967(20.8)50(4.1)33(8.2)1050(16.7)
Unknown299(6.4)67(5.5)31(7.8)397(6.3)
Year of recruitment [n (%)]         
2004–20061872(40.2)408(33.4)151(37.8)2431(38.7)< 0.001
2007–20102785(59.8)813(66.6)249(62.3)3847(61.3)
Delayed diagnosis [n (%)]         
No1865(40.1)472(38.7)123(30.8)2460(39.2)< 0.001
Yes1723(37.0)503(41.2)208(52.0)2434(38.8)
Unknown1069(23.0)246(20.2)69(17.3)1384(22.1)
Losses to follow-up [n (IR)]996(8.1)309(11.1)168(19.7)1473(9.3)<0.001
CD4 count at enrolment (cells/μL)n444011603795979<0.001
Median (IQR)364 (180–572)350.5 (171.5–527)284 (130–474)357 (175–555)
Viral load at enrolment (log copies/ml)n442011523735945<0.001
Median (IQR)4.66 (4.05–5.16)4.56 (3.97–5.07)4.58 (3.65–5.11)4.64 (4.01–5.14)
CD4 count at ART initiation (cells/μL)n216751019628730.945
Median (IQR)210 (85–306)210 (97–302)208.5 (92.5–287.5)210 (89–302)
Viral load at ART initiation (log copies/ml)n21295011932823<0.001
Median (IQR)4.99 (4.44–5.40)4.80 (4.27–5.28)4.83 (4.06–5.18)4.94 (4.37–5.37)

Delay in the diagnosis of HIV infection

Patients who could not provide a CD4 count measurement in the 6 months following their HIV diagnosis were excluded from this analysis; a total of 1384 patients (22%). They were mainly patients who had been diagnosed before the initiation of CoRIS or who started their clinical follow-up in centres not participating in CoRIS. Exclusion from the DD analysis was more frequent among women (25.6%), IDUs (52.4%), NSP (23.0%) and patients with a low level of education (26.4%), reflecting the predominant HIV-infected population profile in the late 1990s and early 2000s.

Of the 4894 patients assessed, 2434 (49.7%) experienced DD. The prevalence of DD was lower in NSP (48%) compared with both LAC (51.6%) and SSA (62.8%). Transmission route was a confounder, and an interaction was found between region of origin and age, so adjusted odds ratios (ORs) are shown stratified by age (Table 2). SSA less than 35 years old had a risk of DD double that of young NSP, while older SSA showed no difference relative to older NSP. For LAC, an excess risk compared with NSP was seen for subjects less than 50 years old, and was more pronounced for those under 35 years old.

Table 2. Results of the univariate and multivariate analyses for all outcomes in the study
 SSA (OR or HR [95% CI])LAC (OR or HR [95% CI])
OutcomeNo. of events/nb(%; IRc)CrudeAdjusteddCrudeAdjustedd
  1. ART, antiretroviral therapy; SSA, sub-Saharan Africans; LAC, Latin Americas and Caribbeans; OR, odds ratio; HR, hazard ratio; IR: incidence rate per 100 person-years; TB, tuberculosis.
  2. a*P < 0.05; **P < 0.01.
  3. bThe denominator differs between outcomes. For diagnostic delay (DD), it is the number of subjects assessed for DD in each age group. For the other outcomes, it is person-years at risk.
  4. cThe percentage is shown for the DD outcome; IR is shown for the other seven outcomes.
  5. dFor the DD outcome, the OR is shown adjusted by transmission route and age at recruitment. For time to ART requirement, the HR was adjusted by CD4 count and viral load (VL) at recruitment. For time to ART initiation, the HR was adjusted by CD4 count and VL at recruitment and transmission route. For time to immunological response and time to virological response, the HR was adjusted by CD4 count and VL at treatment initiation. For time to AIDS and time to AIDS excluding TB, the HR was adjusted for CD4 count and VL at recruitment, transmission route and age at recruitment. For time to death, the HR was adjusted for CD4 count and VL at recruitment, transmission route, age and history of hepatitis C virus infection.
Diagnostic delayAge  1.83 [1.45–2.31]aAge1.15 [1.00–1.33]aAge 
<35 yearss948/2382(39.8)<35 years2.02 [1.47–2.78]a<35 years1.69 [1.39–2.06]a
35–50 years1116/1987(56.2)35–50 years1.15 [0.78–1.72]35–50 years1.28 [1.00–1.64]a
>50 years370/525(70.5)>50 years0.42 [0.14–1.24]>50 years0.84 [0.44–1.62]
ART requirement 1897/4306(44.1)1.22 [1.00–1.49]a 0.90 [0.67–1.21]1.13 [1.01–1.26]a 1.02 [0.88–1.18]
ART initiation 3521/5765(61.1)1.43 [1.16–1.62]a 1.06 [0.91–1.24]1.01 [0.93–1.09] 0.91 [0.84–1.00]a
Immunological response 2226/2578(86.3)0.81 [0.69–0.96]a 0.80 [0.67–0.96]a1.02 [0.92–1.13] 1.04 [0.94–1.16]
Virological response 2064/2726(75.7)0.79 [0.65–0.94]a 0.74 [0.61–0.90]a1.01 [0.90–1.13] 1.00 [0.89–1.13]
AIDSAge  1.17 [0.77–1.78]Age0.80 [0.60–1.06]Age 
<35 years109/6422(1.7)<35 years0.94 [0.46–1.91]<35 years0.71 [0.44–1.14]
35–50 years201/5334(3.8)35–50 years2.05 [1.12–3.74]a35–50 years1.58 [1.05–2.40]a
>50 years64/1140(5.6)>50 years1.47 [0.27–7.87]>50 years0.25 [0.04–1.79]
AIDS (excluding TB) 250/12897(1.9)1.04 [0.60–1.80]1.26 [0.70–2.26]0.85 [0.61–1.19]1.19 [0.81–1.74]
Death 231/15868(1.5)0.62 [0.32–1.21]0.74 [0.36–1.53]0.54 [0.35–0.81]a0.92 [0.60–1.42]

Access to ART

Of the 3889 subjects assessed [providing 4306 person years (py) of follow-up], 1897 subjects required treatment (48.8%; incidence rate 44.1/100 py). The incidence was higher for SSA (68.1/100 py) and LAC (49.5/100 py) compared with NSP (41.6/100 py). Crude hazard ratios (HRs) showed an excess of risk for both groups, but the effect disappeared after adjusting for CD4 counts and VLs at recruitment (Table 2).

Overall, 3521 of 5991 patients (58.8%) initiated ART (providing 5765 py of follow-up; incidence rate 61.1/100 py). The incidence was higher for SSA (102.7/100 py) than for LAC (62.5/100 py) and NSP (58.6/100 py). Crude HRs showed earlier ART initiation for SSA, which disappeared after adjusting for CD4 count and VL at recruitment and transmission route. LAC did not show any difference compared to NSP in the crude analysis, but a slightly higher risk of delayed initiation of ART was observed after adjustment (Table 2). Initial antiretroviral regimes for NSP, LAC and SSA are shown in Table 3; no significant differences were observed.

Table 3. Initial combined antiretroviral therapy (ART) regimen by region of origin
Initial ART regimeNSP [n (%)]SSA [n (%)]LAC [n (%)]Total [n (%)]
  1. Fisher's exact P-value for the table = 0.36.
  2. NSP, native Spaniards; SSA, sub-Saharan Africans; LAC, Latin Americans and Caribbeans; NRTI, nucleoside reverse transcriptase inhibitor; NNRTI, nonnucleoside reverse transcriptase inhibitor; PI, protease inhibitor; PI/r, ritonavir-boosted protease inhibitor.
2 NRTIs + 1 NNRTI1448(56.6)134(55.1)376(58.0)1998(56.8)
2 NRTIs + 1 PI/r877(33.4)85(35.0)202(31.2)1164(33.1)
2 NRTIs + 1 PI55(2.1)7(2.9)16(2.5)78(2.2)
2 NRTIs + 1 integrase Inhibitor37(1.4)1(0.4)16(2.5)54(1.5)
3 NRTIs42(1.6)3(1.2)7(1.1)52(1.5)
Other131(4.9)13(5.3)31(4.8)175(5)
Total2630(100)243(100)648(100)3521(100)

Immunological and virological responses to ART

Of the 2724 subjects assessed for immunological response to treatment, 2226 (81.7%) had a CD4 count increase of 100 cells/μL or higher at some point in time, 1896 (69.6%) within the first year of ART. The global response rate was 86.3/100 py, and was lower for SSA (71.8/100 py) compared with NSP (86.9/100 py) and LAC (89.4/100 py). A total of 2641 patients were assessed for virological response to treatment, of whom 2064 responded at some time (78.2%) and 1852 (70.1%) within the first year. The virological response rate was 75.7/100 py globally and again was significantly lower for SSA (61.5/100 py) compared with NSP (76.5/100 py) and LAC (77.6/100 py). Higher risks of delayed immunological and virological responses in SSA were observed in the crude analyses and after adjustment for CD4 count and VL at ART initiation, while no differences were found for LAC (Table 2).

Risks of AIDS and death

Of the 5242 patients evaluated (12898 py of follow-up), 374 were diagnosed with AIDS during the follow-up period (7.1%; incidence rate 2.9/100 py), 280 of them within the first year. Initial AIDS-defining illnesses are shown in Figure 2. AIDS incidence was highest for SSA (3.7/100 py), lowest for LAC (2.4/100 py) and intermediate for NSP (3.0/100 py). Crude HRs showed no significant differences across regions. Age was found to modify the effect of the region of origin, so HRs are shown stratified by age and adjusted for CD4 count and VL at recruitment and transmission route (Table 2). Adjustment revealed a higher risk of AIDS for both SSA and LAC in those aged 35–50 years. A subanalysis was performed excluding tuberculosis (TB) as an AIDS defining illness, and no differences in the risk of AIDS were then observed.

Figure 2.

Initial AIDS-defining illnesses at incident AIDS diagnosis (n = 374) by region of origin. Pn jirovecii, Pneumocystis jiroveci pneumonia; TB, Mycobacterium tuberculosis; pulm., pulmonary; diss., disseminated; extrapulm., extrapulmonary; PML, progressive multifocal leucoencephalopathy; pneumonia (recurrent): bacterial pneumonia (recurrent); ADIs: AIDS-defining illnesses; NSP, native Spanish; SSA, sub-Saharan Africans; LAC, Latin Americans and Caribbeans. *One patient can simultaneously experience more than one ADI at AIDS diagnosis, so total ‘n’ is larger than the total number of patients.

Among the 5991 subjects (15868 py) included in the overall survival analysis, 231 patients died (3.9%; incidence rate 1.5/100 py). The death rate was highest for NSP (1.6/100 py), intermediate for LAC (1.1/100 py) and lowest for SSA (0.9/100 py). Crude analysis showed longer survival times for LAC and SSA, which reached statistical significance for LAC. After adjusting for CD4 count, VL, transmission route, age at recruitment and history of hepatitis C virus infection (either active or cleared), the difference was no longer significant.

No differences were found by sex in any of the analyses. Results were largely unchanged in the sensitivity analyses and when using inverse probability weighting for censoring.

Discussion

Differences in the management and progression of HIV infection were found in migrants from Latin America or the Caribbean (LAC) and sub-Saharan Africa (SSA) living in Spain compared with Spaniards (NSP). Both migrant groups experienced longer delays in the diagnosis of HIV infection at younger ages and, although no meaningful inequalities were observed in initiation of treatment, SSA exhibited poorer virological and immunological responses. Both groups show an increased risk of AIDS for those aged between 35 and 50 years, while no significant difference was seen in mortality rates.

To our knowledge, this is the largest study to be carried out in HIV-positive migrants from Latin America in a European setting. Cultural differences, legal status and language barriers affect LAC and SSA communities living in Spain differently and produce different vulnerabilities, even in the context of a free universal health care system. Together with their different epidemiological and behavioural backgrounds, this results in very different HIV-infected population profiles, as shown in the present study and previously [2, 25], which suggests that these groups should be studied independently.

A longer DD in migrants compared with NSP was previously described in Europe [7, 8]. What our study adds is the identification of a higher risk in LAC migrants under 50 years of age, probably reflecting legal, administrative or cultural barriers to accessing testing. The lack of differences in those over 50 years of age could reflect, rather than an improved situation for LAC, a much more frequent DD in NSP [21]. SSA also experienced a higher risk of DD below 35 years of age, but not after that age. It is possible that their ethnic visibility leads physicians to comply with HIV testing recommendations when subjects contact the health care system, thus protecting them from DD.

Time to requiring ART was shorter and more prompt ART initiation was found for both migrant groups in the crude analysis, but not after adjusting for CD4 count and VL, suggesting an effect of recruitment at later stages of disease. Delayed ART initiation was found for LAC after adjustment, although the magnitude of the delay was small and probably without clinical relevance. These results are consistent with existing literature for European settings [12, 13]. On 1 September 2012, policy changes became effective in Spain regarding coverage of the public health care system. Under this new legislation, migrants of uncertain legal status will no longer be entitled to access public and free health care, which will probably worsen inequalities with regard to DD, and generate new barriers to accessing treatment for this population [26].

Compared with NSP, immunological and virological responses were poorer for SSA, but not for LAC. Differences could not be attributed to initial ART regimes, which were independent of geographical origin, or to different frequencies of CD4 count and VL determinations. Some biological factors have been associated in the literature with a worse response to treatment and thus could partially explain our results, such as specific viral subtypes and/or resistance mutation profiles [27-31], as there could be an effect of differences in laboratory reference parameters [32, 33] or drug metabolism [34]. However, socioeconomic factors may also have a significant effect[16, 35, 36], probably related to ART adherence.

Adherence has been a challenge since HIV treatments became available, requiring commitment from the patient, who must cope with frequent adverse events and regular clinical follow-up visits. An increased rate of adverse events has been described in SSA and could affect adherence [37]. Also, high geographical mobility, communication, language and administrative barriers, stigma and discrimination, different health beliefs and habits, low literacy, lack of social support and depression affect SSA communities living in Spain and are negatively associated with adherence [38].

Previous studies have found no difference in virological response when using a threshold of 400–500 copies/ml, but a poorer response when using a target VL of 50 copies/ml [9, 11, 14, 16, 39]. Achieving a VL as low as 20 copies/ml may have an impact on sustained virological response, so studies should consider a target VL as low as possible to capture meaningful inequalities.

A higher risk of AIDS in those aged between 35 and 50 years was found for SSA and LAC as a result of a higher rate of tuberculosis, which is prevalent in groups with poor living conditions or who come from endemic areas [40-42].

In contrast to our findings, previous studies have found a similar or slightly decreased progression to AIDS in SSA compared to native populations in Europe [10, 15, 17]. A recent study [13] did not find any difference in a large seroconverter cohort. Although such a cohort is ideal for studying the natural history of the disease, it preferentially includes migrants who were infected in Europe, have good access to the system and do not experience DD. Seroprevalent cohorts are more representative of the HIV epidemic as a whole and can better reflect the situation for migrants regarding HIV/AIDS. In fact, a previous work [12] in a seroprevalent cohort comparing SSA and northwestern Europeans showed a statistically nonsignificant 55% increased risk of AIDS in SSA, which was attenuated when tuberculosis was excluded as the first AIDS defining illness.

Studies in the literature [10, 12, 13, 15] have reported lower mortality in migrants, which was attributed to a ‘healthy migrant’ pattern, in which there is self-selection of healthier individuals for migration [43, 44]. Our results are consistent with those of previous studies regarding the direction of the association, although in this case the difference was not statistically significant.

The main limitation of this study is the lack of statistical power for some analyses, especially those in SSA, and this could explain the absence of observed differences by gender. Gender and social class are inextricably linked to migration in their effects on health [43, 45], but we were not able to fully explore these factors, which would require collaboration among cohort studies.

Regarding potential information biases, educational level was used in this study as a proxy for social class [46-48], but it is possible that it may not completely capture socioeconomic level in migrants, who are often faced with loss of social status in the host country and whose educational curriculum may not be fully comparable. Biological factors such as viral subtype or the prevalence of resistance mutations were also not explored, and nor were adherence and other behaviours that have an impact on prognosis. Also, informative censoring could have occurred as rates of loss to follow-up were higher in migrants [19, 49], so inverse probability weighting was performed to minimize this source of bias. Loss to follow-up was also higher in patients with better clinical status, so the residual bias, if any, would tend to overestimate incidence rates in migrants. Finally, selection bias could have been present in the DD analysis, as this was restricted to newly diagnosed patients recruited to the cohort, the implications of which were discussed in a previous publication for this cohort [50]. The results of this study indicate different sources of inequality in HIV management and disease progression which are important for public health policies and service planning and delivery. DD was one of the main sources of inequality between migrants and native Spaniards, especially in younger subjects, reflecting the existence of barriers to HIV testing for migrant communities. Expanding HIV testing in migrants should be a priority for public health programmes. Once the patient has accessed the system, no major barriers were identified for treatment initiation, although recent policy developments in Spain may change this situation for migrants of uncertain status, and effective access to treatment will need to be monitored over the next few years. The poorer response in SSA should be further studied to identify factors that may be limiting ART effectiveness in this group. Finally, more attention should be paid to preventing tuberculosis among migrants living with HIV.

Acknowledgements

This study would not have been possible without the collaboration of all the patients, physicians, nurses, and data managers who have taken part in the project. The RIS cohort (CoRIS) is funded by the Instituto de Salud Carlos III through the Red Temática de Investigación Cooperativa en Sida (ISCIII-RETIC RD06/006).

Conflicts of interest: The authors declare that they have no conflict of interest.

Author contributions: JdA, FG and SM initiated this project. SM, BA, FD, JdR, JAI, FP, RR, JMM, FG and JdA were responsible for data collection. SM and BA performed the analyses. SM drafted the manuscript which was reviewed by JdA, FG and BA. All authors were involved in the study design, revised the manuscript for important intellectual content and contributed to the final version of the manuscript.

Appendix: Appendix: Members of Cohorte de la Red de Investigación en Sida (CoRIS)

Steering committee: Juan Berenguer, Julia del Amo, Federico García, Félix Gutiérrez, Pablo Labarga, Santiago Moreno and María Ángeles Muñoz. Fieldwork, data management and analysis: Paz Sobrino, Belen Alejos, Victoria Hernando, Debora Alvarez, Susana Monge and Inma Jarrin. BioBank: María Ángeles Muñoz, Isabel García-Merino, Coral Gómez. Jorge Gallego and Almudena García. Hospitals: Hospital General Universitario de Alicante (Alicante): Joaquín Portilla Sogorb, Esperanza Merino de Lucas, Sergio Reus Bañuls, Vicente Boix Martínez, Livia Giner Oncina, Carmen Gadea Pastor, Irene Portilla Tamarit and Patricia Arcaina Toledo. Hospital Universitario de Canarias (Santa Cruz de Tenerife): Juan Luis Gómez Sirvent, Patricia Rodríguez Fortúnez, María Remedios Alemán Valls, María del Mar Alonso Socas, Ana María López Lirola, María Inmaculada Hernández Hernández and Felicitas Díaz-Flores. Hospital Carlos III (Madrid): Vicente Soriano, Pablo Labarga, Pablo Barreiro, Pablo Rivas, Francisco Blanco, Luz Martín Carbonero, Eugenia Vispo and Carmen Solera. Hospital Universitario Central de Asturias (Oviedo): Victor Asensi, Eulalia Valle and José Antonio Cartón. Hospital Clinic (Barcelona): José M. Miró, María López-Dieguez, Christian Manzardo, Laura Zamora, Iñaki Pérez, Mª Teresa García, Carmen Ligero, José Luis Blanco, Felipe García-Alcaide, Esteban Martínez, Josep Mallolas and José M. Gatell. Hospital Doce de Octubre (Madrid): Rafael Rubio, Federico Pulido, Silvana Fiorante, Jara Llenas, Violeta Rodríguez and Mariano Matarranz. Hospital Donostia (San Sebastián): José Antonio Iribarren, Julio Arrizabalaga, María José Aramburu, Xabier Camino, Francisco Rodríguez-Arrondo, Miguel Ángel von Wichmann, Lidia Pascual Tomé, Miguel Ángel Goenaga, Mª Jesús Bustinduy and Harkaitz Azkune Galparsoro. Hospital General Universitario de Elche (Elche): Félix Gutiérrez, Mar Masiá, José Manuel Ramos, Sergio Padilla, Andrés Navarro, Fernando Montolio, Yolanda Peral and Catalina Robledano García. Hospital Germans Trías i Pujol (Badalona): Bonaventura Clotet, Cristina Tural, Lidia Ruiz, Cristina Miranda, Roberto Muga, Jordi Tor and Arantza Sanvisens. Hospital General Universitario Gregorio Marañón (Madrid): Juan Berenguer, Juan Carlos López Bernaldo de Quirós, Pilar Miralles, Jaime Cosín Ochaíta, Matilde Sánchez Conde, Isabel Gutiérrez Cuellar, Margarita Ramírez Schacke, Belén Padilla Ortega and Paloma Gijón Vidaurreta. Hospital Universitari de Tarragona Joan XXIII, IISPV, Universitat Rovira i Virgili (Tarragona): Francesc Vidal, Joaquín Peraire, Consuelo Viladés, Sergio Veloso, Montserrat Vargas, Miguel López-Dupla, Montserrat Olona, Alba Aguilar, Joan Joseph Sirvent, Antoni Soriano and Rami A. A. Qaneta. Hospital Universitario La Fe (Valencia): José López Aldeguer, Marino Blanes Juliá, José Lacruz Rodrigo, Miguel Salavert, Marta Montero, Eva Calabuig and Sandra Cuéllar. Hospital Universitário La Paz (Madrid): Juan González García, Ignacio Bernardino de la Serna, José María Peña Sánchez de Rivera, José Ramón Arribas López, María Luisa Montes Ramírez, José Francisco Pascual Pareja, Blanca Arribas, Juan Miguel Castro, Fco Javier Zamora Vargas, Ignacio Pérez Valero, Miriam Estebanez and Raphael Mohr y Francisco Arnalich Fernández. Hospital de la Princesa (Madrid): Ignacio de los Santos, Jesús Sanz Sanz, Johana Rodríguez, Ana Salas Aparicio and Cristina Sarriá Cepeda. Hospital San Pedro-CIBIR (Logroño): José Antonio Oteo, José Ramón Blanco, Valvanera Ibarra, Luis Metola, Mercedes Sanz and Laura Pérez-Martínez. Hospital San Pedro II (Logroño): Javier Pinilla Moraza. Hospital Universitario Mutua de Terrassa (Terrassa): David Dalmau, Angels Jaén Manzanera, Mireia Cairó Llobell, Daniel Irigoyen Puig, Laura Ibáñez, Queralt Jordano Montañez, Mariona Xercavins Valls, Javier Martinez-Lacasa, Pablo Velli and Roser Font. Hospital de Navarra (Pamplona): Julio Sola Boneta, Javier Uriz, Jesús Castiello, Jesús Reparaz, María Jesús Arraiza, Carmen Irigoyen and David Mozas. Hospital Parc Taulí (Sabadell): Ferrán Segura, María José Amengual, Eva Penelo, Gemma Navarro, Montserrat Sala, Manuel Cervantes and Valentín Pineda. Hospital Ramón y Cajal (Madrid): Santiago Moreno, José Luis Casado, Fernando Dronda, Ana Moreno, María Jesús Pérez Elías, Dolores López, Carolina Gutiérrez, Beatriz Hernández, María Pumares and Paloma Martí. Hospital Reina Sofía (Murcia): Alfredo Cano Sánchez, Enrique Bernal Morell and Ángeles Muñoz Pérez. Hospital San Cecilio (Granada): Federico García García, José Hernández Quero, Alejandro Peña Monje, Leopoldo Muñoz Medina and Jorge Parra Ruiz. Centro Sanitario Sandoval (Madrid): Jorge Del Romero Guerrero, Carmen Rodríguez Martín, Teresa Puerta López, Juan Carlos Carrió Montiel, Cristina González and Mar Vera. Hospital Universitario Santiago de Compostela (Santiago de Compostela): Antonio Antela, Arturo Prieto and Elena Losada. Hospital Son Dureta (Palma de Mallorca): Melchor Riera, Javier Murillas, Maria Peñaranda, Maria Leyes, Mª Angels Ribas, Antoni Campins, Concepcion Villalonga and Carmen Vidal. Hospital Universitario de Valme (Sevilla): Juan Antonio Pineda, Eva Recio Sánchez, Fernando Lozano de León, Juan Macías, José del Valle, Jesús Gómez-Mateos and Rosario Mata. Hospital Virgen de la Victoria (Málaga): Jesús Santos González, Manuel Márquez Solero, Isabel Viciana Ramos and Rosario Palacios Muñoz. Hospital Universitario Virgen del Rocío (Sevilla): Pompeyo Viciana, Manuel Leal, Luis Fernando López-Cortés and Mónica Trastoy.

Ancillary