Effective treatment for HIV has been available in Europe since the mid 1990s and has led to a dramatic reduction in the incidence of AIDS events and HIV-related deaths. Many HIV-positive individuals are now living with HIV as a chronic condition rather than an inevitably fatal illness.

However, of the approximately 2.3 million HIV-infected individuals living in the European region, it is estimated that one in three is unaware of his/her HIV serostatus (i.e. 700,000–900,000 individuals) [1], resulting in significant levels of late diagnosis and transmission across the region. In Western Europe, 45–50% of newly diagnosed HIV-positive individuals are diagnosed and enter care late (i.e. with a CD4 count ≤350 cells/μL) [2, 3].

Late presentation is associated with: increased HIV-related morbidity and mortality [4–7], shorter survival [8], poor response to treatment [9, 10], increased healthcare costs [11] and increased rates of HIV transmission [12]. If a person is diagnosed early and HIV treatment is introduced early in the course of infection before severe impairment of the immune system has occurred, life-expectancy may approach that of the general population [13].

The aim of the present study is to analyze the characteristics and the factors associated with being diagnosed late or at an advanced stage of disease among persons newly diagnosed with HIV infection in Italy in 2010 and 2011.

We used individual data on new HIV diagnoses reported in Italy in 2010–2011 to the HIV surveillance system. The HIV surveillance system is coordinated by the National AIDS Unit of the National Institute of Health (Rome) and was established in 2008 by a decree of the Ministry of Health [14] which included HIV infection to the list of mandatory notifiable infectious diseases.

Individual data on HIV diagnoses reported by clinicians are available since 2010. In 2010–2011, 19 out of 21 Italian regions provided individual information on socio-demographic characteristics, transmission mode, and CD4 count at HIV diagnosis. These regions account for 87.5% of the total Italian population. CD4 count was measured within 15 days after HIV diagnosis.

Persons with CD4 ≤350 cells/μL or diagnosed with AIDS (regardless of the CD4 cell count) were defined as late presenters (LP); persons with CD4 ≤ 200 cells/μL or AIDS (regardless of the CD4 cell count) were defined as presenting with advanced HIV disease (AHD).

The proportion of LP and AHD was calculated for every variable; the characteristics of LP and AHD were compared with those of non-LP or non-AHD and differences were analyzed with the Chi-square test. To compare median age we performed statistical analysis using Mann-Whitney U-test.

Variables included in the univariate analysis were: year of reporting, gender, age, nationality, area of notification, transmission mode.

In order to identify factors associated with being LP or AHD, we built two multivariate logistic regression models in which being LP or AHD was considered as dependent variable. Variables with a P value of <0.05 were entered in the model. The fitness of the final model was assessed with the likelihood ratio test. Gender was forced into the model regardless of its significance.

The statistical analysis was conducted using IBM SPSS 20.0.

Our results show that more than half of individuals with a new HIV diagnosis were LP; among these, more than two thirds were in an advanced stage of HIV disease and one third was diagnosed with AIDS.

The late HIV diagnosis and the consequent late linkage to care is a worrying scenario. Indeed, the threshold below which initiation of ART is strongly recommended by recent international and Italian guidelines is 350 CD4 cells/uL [15].

The proportion of LP at HIV diagnosis observed in our study was similar to that reported in other European countries (Denmark: 55.9%; Germany: 53%; France: 51.9%; United Kingdom: 49.2%; Spain: 45.5%) [1].

In our sample, non-nationals were more likely to be diagnosed later than Italians, as already reported elsewhere [16, 17] suggesting a low access to HIV testing sites, or a poor perception of exposures at risk for HIV, or the presence of language, cultural or socioeconomic barriers. However, in Italy HIV testing, care and treatment are provided free-of-charge to both legal and illegal immigrants.

The median age was significantly higher among LP and persons with AHD compared to non-LP and non-AHD, and the probability of being LP or AHD increased with increasing age, as shown by other Italian studies [16, 17]. This finding may be attributed to a lower awareness among older persons of having been exposed to HIV infection [18]. Moreover, older persons presenting with HIV-associated diseases are often misdiagnosed by general practitioners who may not consider the possibility of HIV infection in these patients.

Persons reported in Southern and Central Italy were more likely to be diagnosed late than those reported in Northern Italy, suggesting a poorer access to HIV testing centers or a lower risk perception of HIV infection. In the period 2010–2011 the incidence of new HIV diagnoses reported in Southern Italy has always been lower compared to that in Central and Northern Italy (South 3.4 per 100,000 Central 5.7 per 100,000, North 6.3 per 100,000). This finding might have led both health care personnel and general population to think that the risk of acquiring HIV infection was actually lower in Southern Italy, thus minimizing education and information campaigns aimed at HIV prevention in this area [19].

Despite the strong reduction of the HIV epidemic among IDU in Italy since 1990 [19], our results evidence a high proportion of LP and the highest proportion of persons with AHD among IDU; similar results were reported in a study conducted in Spain [20]. These findings can be explained by a lower HIV testing uptake in drug-treatment centers in the last decade [21] mainly attributable to the decline in HIV prevalence among IDU [22] and in the number of persons who inject drugs [21]. These changes may have led healthcare services to consider the HIV epidemic under control in this population thus reducing the active offer of HIV testing. This has also caused a minimization of the risk of HIV infection through unprotected sexual contact among both injecting and non-injecting drug users [23].

In our study, complaining of clinical conditions associated with HIV infection was the main reason for undergoing HIV testing among LP, confirming the relevant role of healthcare professionals (especially general practitioners) in recommending HIV testing not only in presence of AIDS-defining diseases but also for specific HIV indicator conditions [24].

Among LP, having engaged in sexual risk behavior was not a relevant reason for undergoing HIV testing, suggesting that unprotected sexual contact is not perceived as being at risk [25].

IDU and HET had a similar probability of late diagnosis, higher than that found for MSM. Also other studies conducted in different countries [18, 26] report that MSM are less likely to be diagnosed late compared to IDU and HET; this finding may be associated with a higher probability among MSM of undergoing HIV testing which reflects a higher perception of sexual risk behavior [19].

Some weaknesses and strengths of this study need to be mentioned. In the first place, national data on HIV surveillance system are available only from 2010. Moreover this study shares the limitation of all surveillance systems: underreporting, missing data, timelessness and regional differences in health system.

However, our study has several advantages: it provides a picture of both LP and AHD for two recent years, is based on a robust sample of individuals, has an almost-national geographical coverage, offers information on the testing attitude of individuals exposed to HIV infection based on the reasons for HIV testing.

Despite the recommendation for initiating ART early, in Italy the interventions to discover hidden HIV cases are few and the proportion of LP is quite high. In this study, more than half of new HIV diagnoses are LP, mainly persons older than 50 years of age, non-nationals, persons living in Central-Southern Italy, and persons who acquired the infection through injecting drug use or heterosexual contact. These results reinforce the importance of identifying HIV infection as early as possible and suggest the need for a more active offer of HIV testing in these groups. The presence of HIV-related symptoms represents an effective trigger for early diagnosis; general practitioners play a crucial role in the early identification of HIV-infected individuals and should consider HIV infection even among older individuals, encouraging HIV testing when appropriate. At-risk sexual behavior instead is not recognized as a sufficient reason for being tested. An effort to improve the awareness of exposures at risk is required, to optimize the early linkage to care and treatment, and to reduce HIV morbidity, mortality and transmission.