Clinicoepidemiological Characteristics of HIV-Infected Immigrants Attended at a Tropical Medicine Referral Unit


  • José A Pérez-Molina MD, PhD,

    Corresponding author
    1. Tropical Medicine and Clinical Parasitology Unit, Infectious Diseases Department, Ramón y Cajal Hospital, Madrid, Spain
      José A. Pérez-Molina, MD, PhD, Tropical Medicine and Clinical Parasitology Unit, Infectious Diseases Department, Ramón y Cajal Hospital, Carretera de Colmenar Km 9,100, Madrid 28034, Spain. E-mail:
    Search for more papers by this author
  • Rogelio López-Vélez MD, PhD,

    1. Tropical Medicine and Clinical Parasitology Unit, Infectious Diseases Department, Ramón y Cajal Hospital, Madrid, Spain
    Search for more papers by this author
  • Miriam Navarro MD,

    1. Tropical Medicine and Clinical Parasitology Unit, Infectious Diseases Department, Ramón y Cajal Hospital, Madrid, Spain
    Search for more papers by this author
  • María J. Pérez-Elías MD,

    1. Infectious Diseases Department, Ramón y Cajal Hospital, Madrid, Spain
    Search for more papers by this author
  • Santiago Moreno MD, PhD

    1. Infectious Diseases Department, Ramón y Cajal Hospital, Madrid, Spain
    Search for more papers by this author

  • This article has been partially presented at the 11th European AIDS Conference, Madrid, Spain, October 24 to 27, 2007.

José A. Pérez-Molina, MD, PhD, Tropical Medicine and Clinical Parasitology Unit, Infectious Diseases Department, Ramón y Cajal Hospital, Carretera de Colmenar Km 9,100, Madrid 28034, Spain. E-mail:


Background Migration is a growing phenomenon with a well-known impact in infectious diseases epidemiology. Currently, immigrants represent almost 10% of the Spanish population. The majority come from countries where the prevalence of chronic viral illnesses is higher than in Spain.

Methods To describe clinicoepidemiological features of human immunodeficiency virus (HIV)–infected immigrants attending our Unit and to compare differential characteristics depending on geographical origin, information from all new immigrants from January 1997 to December 2006 was collected. Study design: noninterventional retrospective chart review.

Results We screened 1,609 patients of whom 77 (4.8%) were HIV antibody (Ab) positive. Of these, 80% were sub-Saharan Africans (SSAFR) and 20% were South–Central Americans (SCA). HIV prevalence was higher in SSAFR (5.6% vs 3.2%; p= 0.04). Overall, of those who were HIV Ab positive, 70% were male (median age 30 years), 59% heterosexuals, 9% hepatitis C virus coinfected, 8.6% hepatitis B virus coinfected, and 34% showed a positive tuberculin skin test. Median CD4 cell count was 263 cells/μL, median HIV-ribonucleic acid viral load 4.6 Log/mL, and 48% had a late diagnosis [acquired immunodeficiency syndrome (AIDS)–defining illness or <200 CD4 μL at the time of diagnosis]. Only 68% of patients for whom antiretroviral therapy was indicated actually started therapy and 22% were lost to follow-up just after diagnosis. SCA had lower CD4 cell counts (26 vs 168 cells/μL; p= 0.016), higher viral loads (5.3 vs 4.8 Log; p= 0.001), and were more likely to have an AIDS-defining illness (53% vs 21%; p= 0.04) compared to SSAFR. Tuberculin skin test reactivity was more common among SSAFR versus SCA [adjusted by CD4 count, odds ratio (OR) 6.3 and 95% confidence interval (CI): 0.65–60.5]. The main risk factor for late diagnosis was geographical origin: OR 4.6 (95% CI: 1.11–19.3) (SCA vs SSAFR; adjusted by the interval between the date of arrival in Spain and the date of HIV diagnosis).

Conclusions Almost half the HIV-infected immigrants were diagnosed in late stages. Patients were frequently lost to follow-up, and a significant minority did not start highly active antiretroviral therapy when indicated. SCA seem to have more severe immunosuppression at the time of diagnosis than SSAFR. Early voluntary routine HIV screening should be promoted.

Migration is a growing social phenomenon with a well-recognized impact on the epidemiology of infectious diseases. The number of global migrants increased from 84 million in 1975 to 175 million in 2000, and it is estimated that this number will reach 230 million in 2050.1 The health characteristics of migrants soon after arrival are influenced by the health risks in the country of origin and the disease risk area visited during migration to the intended target country.2,3

Immigration has been an increasing phenomenon in Spain especially in the past 10 years. In 2007, immigrants represented almost 10% of the Spanish population.4 The majority come from regions, such as sub-Saharan Africa, South-Central America, and Eastern Europe,5–7 where the prevalence of chronic viral illnesses, namely human immunodeficiency virus (HIV), hepatitis C virus (HCV), and hepatitis B virus (HBV), is higher than in Spain. Of note, 30% of patients newly diagnosed with HIV in Spain in the past 3 years were of foreign origin.8 Lack of knowledge of the local language, social exclusion, and cultural factors may act as barriers for prompt medical attention and early diagnosis in vulnerable subjects.9

Currently, most frequent risk behaviors for HIV infection in Spain are unsafe heterosexual sex (46.1% of newly infected patients), unsafe homosexual sex (31.2%), and intravenous drug use (IVDU) (15.3%). Among immigrants, unsafe heterosexual sex is the leading cause for HIV infection (61.4%), followed by unsafe homosexual sex (26.1%) and IVDU (6.2%). Nevertheless, these figures change depending on the patients’ geographical origin: IVDU is most common in immigrants coming from Eastern Europe (28.4%), unsafe homosexual sex in SCA (47%), and unsafe heterosexual sex among SSAFR (92%).

Due to their expertise in imported infections, travel clinics and tropical medicine units are increasingly expected to provide health care for sick immigrants among whom chronic infections such as HIV are increasing. Prompt diagnosis, education on preventive measures, and treatment when necessary have dual benefits for the patient and for the country of destination. Our objectives in this study were to describe the clinical and epidemiological features of HIV-infected immigrants attending our Unit, factors related to delay in diagnosis, and to compare geographically determined characteristics, to enable factors to be identified that could improve the management of these patients.

Patients and Methods

The study was carried out at the Department of Infectious Diseases at the Ramón y Cajal Hospital in Madrid. The Tropical Medicine and Clinical Parasitology Unit of this department has been a referral center for immigrants and particularly for SSAFR for many years. For all patients, a standardized study protocol is applied and active screening for transmissible infectious diseases is performed. Information from all new immigrants who attended our Unit during the period January 1997 to December 2006 was collected. Country of origin, duration of residence in Spain prior to visiting the clinic, social status, risk behaviors for imported diseases, and current and prior medical problems were recorded and a thorough medical and analytical workup was performed. Tests included a complete blood count, serum biochemistry, basic urine analysis, serology for HIV, HBV, HCV, syphilis, and Mantoux skin test for all immigrants, as well as other examinations needed for specific clinical conditions. Patients are informed in their own language about all the tests intended to be performed and the confidentiality of the results prior to obtaining verbal permission. For a patient to be defined HIV-infected, two serological HIV screening tests have to be positive (enzyme-linked immunosorbent assay test identifying both HIV-1 and HIV-2). Positive results are confirmed by Western blot.

This study was designed as a noninterventional retrospective chart review. Late diagnosis of HIV infection was defined by a prior or current diagnosis of an acquired immunodeficiency syndrome (AIDS)–defining illness or presence of <200 CD4 cells/μL at the time of the first visit. The interval between arrival in Spain to the date of HIV diagnosis was calculated for all patients.

Categorical variables were expressed as absolute frequencies and percentages. Quantitative variables were expressed as medians and interquartile ranges (IQR). Numeric parameters were compared between groups using the t test or the Mann–Whitney U test when appropriate. Categorical variables were compared between groups using the Pearson chi-squared test or Fisher’s exact test if the group was small. Simple linear regression was used to assess the linear relationship between numeric variables. To calculate crude and adjusted odds ratios (ORs), logistic regression was used. The inclusion of variables in the model was decided according to their significance in the univariate analysis, together with the importance given to each variable regardless of whether or not it was clearly significant in this analysis. An explicative strategy was followed, including the variable. “Geographical origin” in all models. The remaining variables for analysis were entered into the models at the same time, and a backward strategy was used for modeling.


During the study period, 1,609 new patients were seen: 29% SCA, 66% SSAFR, and 5% from other regions. All of them were tested for HIV infection in our Unit or at the primary care level before attending our clinic. Seventy-seven subjects (4.8%) were HIV positive, of whom 74 had sufficient information recorded to allow the description of clinicoepidemiological characteristics.

Regarding HIV-infected subjects, 80% were from sub-Saharan Africa (n= 59) and 20% from South-Central America (n= 15), with HIV prevalence being higher among the former: 5.6% versus 3.2% (p= 0.045) (Table 1). An increase in the number of HIV diagnoses was observed during the study period with the slope of the regression line significantly greater than zero: [1.4 (95% confidence interval (CI): 0.18–2.63; p= 0.03)]. Overall, 70% of the patients were male, with a median age of 30 years (IQR 27–37 y); 59% heterosexuals, with a median CD4 count of 263 cells/μL (IQR 83–458) and median VL 4.6 Log (IQR 4.1–5.2); 9% HCV coinfected; and 8.6% HBV coinfected. No patients were simultaneously coinfected with HCV and HBV. The tuberculin skin test was positive in 34% of the patients (Table 2).

Table 1.  Number of new immigrant patients attended at the Tropical Medicine Unit during the study period
YearNew immigrants patientsSCA% SCA% HIV+ within SCASSAFR% SSAFR% HIV+ within SSAFR
  1. SSAFR = sub-Saharan Africans; SCA = South and Central Americans; HIV = human immunodeficiency virus.

Table 2.  Comparison between South and Central Americans and sub-Saharan patients
CharacteristicSCA (%)SSAFR (%)p Value
  1. SSAFR = sub-Saharan Africans; IQR = interquartile range; SCA = South and Central Americans; TST = tuberculin skin test; HAART = highly active antiretroviral therapy; HIV = human immunodeficiency virus; HCV = hepatitis C virus; HBV = hepatitis B virus.

Age (years) (IQR)29 (24–33)32 (26–38)0.24
Female gender35.729.30.75
Risk behavior for HIV infection
 Unsafe heterosexual sex46.262.7p < 0.0001
 Unsafe homosexual sex30.80.0
 Exposure to tainted blood products/syringes0.03.4
HCV coinfection0.09.80.57
HBV coinfection0.011.10.33
CDC category
Late diagnosis80.041.10.009
Median number of days from arrival in Spain to date of HIV diagnosis (IQR)1,327 (20–2,802)63 (16–1,219)0.27
Positive TST7.141.80.024
Median CD4 cell count/μL (IQR)26 (21–68)168 (74–231)0.016
Viral load (Log) (IQR)5.35 (4.9–5.65)4.85 (4.3–5.2)0.001
HAART indicated at diagnosis86.753.70.034
HAART started when indicated75.064.30.73

HIV was diagnosed a median of 118 days (38–901) after arrival in Spain. Forty-eight percent of patients met criteria for a late diagnosis. Of these, 57% had an AIDS-defining illness at the time of first visit, the most frequent being tuberculosis (40%), Pneumocystis jiroveci pneumonia (20%), histoplasmosis (15%), and esophageal candidiasis (10%). Sixteen patients (22%) were lost to follow-up just after diagnosis. Highly active antiretroviral therapy was recommended to 45 (60.6%) patients according to national and international guidelines10,11 but only 31 of them (68%) started therapy.

We compared the history of sexual behavior of SCA with SSAFR. SCA patients reported homosexual risk behavior more frequently and had lower CD4 cell counts, higher viral load, and more frequently presented with late diagnosis and/or AIDS-defining illness compared to SSAFR. There were no differences with respect to the etiology of AIDS-associated diseases. A positive tuberculin skin test was more common among SSAFR [crude OR 11.5 (95% CI: 1.36–96.7)]. After adjusting for CD4 cell count, the difference was not statistically significant [OR 6.3 (95% CI: 0.65–60.5)].

The main risk factor for a late diagnosis was geographical area of origin. Patients from South and Central America showed a fivefold increased risk of being diagnosed with HIV concomitantly with an opportunistic infection and/or of having a CD4 cell count <200 [OR 5.74 (95% CI: 1.45–22.64)]. To control for confounding factors in the association between geographical origin and a late diagnosis, we performed a multiple logistic regression analysis. Maximum model included the following variables: age, risk behavior, viral load, diagnostic delay (days), and geographical origin. Adjusted OR for a late diagnosis of HIV infection was 4.60 (95% CI: 1.11 – 19.31), very close to the value for the crude OR.


HIV infection among immigrants is increasingly recognized as a health problem in Spain, mainly in those coming from countries with high prevalence rates. In our Unit, 4.8% of attendees were infected with HIV, a prevalence approximately 10 to 12 times higher than that of the Spanish population.8,12 Almost half of them were diagnosed in late stages and up to 22% of them were lost to follow-up just after diagnosis. In addition, one-third failed to start treatment when this was indicated. This may be explained by several factors, which may act synergistically in this population, including cultural differences, prior beliefs about AIDS, language barriers, social exclusion, ignorance regarding rights, and increased geographical mobility while searching for work.13–16 When the two main groups of immigrants were compared, those coming from South and Central America had a significantly higher risk of having an opportunistic infection at the time of HIV diagnosis or a low CD4 count compared to those coming from sub-Saharan countries.

There are several significant differences in the characteristics of HIV-infected immigrant and the Spanish HIV-infected population.17–19 In our study, immigrants had a lower median age at diagnosis and we did not see any patients who acquired the infection by IVDU. This has resulted in lower prevalences of HCV coinfection (0 among Americans and 9% among Africans). In contrast, HBV coinfection prevalence and tuberculin reactivity were higher among immigrants from sub-Saharan Africa, reflecting hepatitis prevalence in their countries of origin. Immunovirologic status among immigrants at diagnosis appeared to be worse, with median CD4 cell counts <200/μL both in SSAFR and in SCA compared to higher figures found in Spaniards according to several reports (228–317 CD4/μL). Despite these lower median CD4 counts, the diagnostic delay and the occurrence of opportunistic infections at the moment of HIV-infection diagnosis were not different compared to that reported in indigenous patients.17–19

Compared with SSAFR, SCA showed a higher degree of immunosuppression at diagnosis (26 vs 168 CD4 cells μL) and a higher risk of being in late stages of the disease (adjusted OR 4.6). Rates for starting therapy when indicated, although very low, were equivalent in both groups, and early loss of follow-up was frequent (22%) but was not different among groups. Although this loss rate seems to be lower than in their countries of origin,20 it is still higher than figures for indigenous patients as immigrants are lost to follow-up more often.21,22 Patients who do not seek routine medical care cannot fully benefit from the increasingly effective treatments available to them. Consistently, attending medical appointments plays a central role in both prolonging life and enhancing quality of life for persons living with HIV/AIDS. In the HIV outpatient clinic of our department, missing visits has been an independent factor for predictive of death (OR 6.6) and admission to the hospital (OR 2.4), emphasizing the importance of an adequate medical follow-up, encouraging patient survival and health resource use.22

A potential source for bias in this study must be considered. There is also an active HIV outpatient clinic in our department, while traditionally the Tropical Medicine Unit has been a referral center for immigrants (both legally and illegally resident in Spain), mainly from sub-Saharan Africa. One factor leading to a later diagnosis among SCA could be that the majority with known HIV infection are referred to the HIV clinic, whereas those detected at our Unit did not know their status and came for other reasons. Instead, SSAFR are routinely referred to the Unit both with and without specific health complaints. Although it could be argued that this could lead to a referral bias of people in more advanced clinical stages and with a worse social status, it is more likely that immigrants being referred to the Tropical Medicine Unit for any reason (mostly sub-Saharan patients) are routinely tested for HIV and have an earlier diagnosis than those who attend the HIV Unit first once they have HIV-related complaints (mostly SCA).

In summary, prevalence of HIV infection is considerably higher in the immigrant population from sub-Saharan Africa and South and Central America compared to Spaniards. Almost half the HIV-infected immigrants attending our Unit were diagnosed in late stages and were frequently lost to follow-up or did not start antiretroviral therapy when this was indicated. Extensive educational campaigns and early voluntary routine HIV screening should be promoted to avoid late diagnosis in this population.


The clinical research team acknowledges the support provided by the Red de Investigación Cooperativa de Enfermedades Tropicales (RICET; RD06/0021/0020).

Declaration of Interests

The authors state that they have no conflicts of interest.