Mozambique is a country with stable P. falciparum transmission, where malaria-related morbidity and mortality is still quite frequent, with more than four million cases and 4,000 attributed deaths reported each year . The study was performed during two high-risk months for malaria; in fact, transmission is seasonal, mainly between November and July with peaks from December to April in the Central region of the country.
At the same time, the proportion of individuals coexisting with HIV in the whole population is about 16% according to the last epidemiological survey in 2007 , with the Northern and Central regions presenting a higher prevalence compared to the Southern part of the country. In particular, the Sofala province shows a prevalence rate between 17% and 33%, with the highest peaks involving urban areas. Among these, Beira, which is the 2nd largest city in Mozambique (population of more than 400,000 inhabitants), has one of the highest adult HIV seroprevalence rates in the country. The national survey based on sentinel posts throughout the country shows that the rate of HIV prevalence measured in pregnant women reached a peak of 36% in Beira in 2002, stabilizing around 29% in the last report in 2007  while the 2009 WHO report gives a range between 25% and 39.9% for Beira . This rate is extra ordinarly high and is probably due to geographical and historical reasons linked to its role as a key commercial harbor on the Indian Ocean, thus probably favouring a high sexual promiscuity. This is reflected in the impressive HIV prevalence (66.7%) in patients hospitalized in the Internal medicine Clinic of the Central Hospital of Beira serving a wide area of Central Mozambique.
Only a few studies have evaluated the epidemiological and clinical aspects concerning concurrent HIV and malaria in this country  and, none in the above-mentioned area. In particular, data are lacking on the reciprocal influence of the two infections on the morbidity of the Mozambican population following the contemporary introduction in 2004 of both ART for HIV infection and a new artemisinin-based treatment for malaria.
Malaria, according to our analysis, is still the most frequent cause of adult admissions to Internal Medicine wards in Mozambique (>25% of all motives for hospitalization). However, when evaluating patients based on their HIV serological status, results show that, although HIV-infected patients are considered a high risk group for malaria, malaria is not the most frequent diagnosis in this patient group, and the prevalence of both malaria infection and clinical malaria is lower among HIV-infected patients (26% and 21%, respectively) than among HIV-negatives (33% and 32%, respectively). This might imply at least two different explanations: The first is that, naturally, the prevalence of malaria infection in hospitalized patients does not directly reflect the disease distribution throughout the entire population, due to the greater number of pathologies, especially AIDS-related opportunistic infections, which could require hospitalization in HIV-infected individuals compared to HIV-negatives. In fact, a previous study performed in Maputo, Mozambique, has already evidenced that the fraction of febrile illness attributable to malaria is lower in HIV-positive adults than in HIV-negatives . However, a second explanation is possible as a statistically significant and independent association between CTX use and decreased risk of malaria infection was observed, while, on the contrary, the malaria distribution did not differ between untreated-HIV-positive and HIV-negative patients. These results are not surprising as the protective effect of cotrimoxazole on parasite infection has been well-documented in clinical trials in children , pregnant women  and adults [6, 13]. The present study confirms this protective effect also in the “real word” scenario of an African urban hospital, outside the selected conditions of a randomized clinical trial.
The effect of anti-virals on malaria parasitaemia is still questionable [4, 5], and only protease inhibitors are currently considered to provide a proven anti-malarial efficacy [14–19], even if in a recent study, a low anti genaemia was demonstrated for suppressed HIV-infected patients treated with ART and CTX for long periods . In this study it was not possible to individuate aviremic patients, because neither viral load nor CD4 count were available; however, the median duration of ART was only four months and 77% of patients were classified as WHO stage III-IV. The majority (67/69) of treated patients were on NNRTIs; an association between the use of antiretrovirals and lower risk of malaria was observed at univariate analysis but it most likely depended on the protective effect of CTX rather than on the direct impact of ART, as statistical significance was not confirmed with multivariate analysis. Hence, further studies directly assessing this topic are required to validate this association.
A wider implementation of prophylaxis for HIV-infected subjects in sub-Saharan Africa in future years might, therefore, substantially reduce the burden of malaria infection and consequently its morbidity in these patients. Moreover, it also appears that the problem of a greater access to HIV testing is the fundamental issue. In fact, the extensive use of HIV testing in the hospital demonstrates that more than one-half of HIV-positive subjects were unaware of their seropositivity before hospitalization. Thus, among HIV-infected patients, only 31% and 35% were already receiving antiretroviral therapy and/or chemoprophylaxis for opportunistic infections, respectively. Moreover, low rates of regimen switches were described (only two patients were on a second line regimen): this is not surprising as the average ART-treatment time in the present study was only four months; moreover the lack of testing for viral load renders it difficult to early individuate therapy failure. These data are consistent with the study of Auld et al.  on ART implementation in Mozambique, even if in disagreement regarding the rate of prescription of CTX, which was lower in their survey compared to ours.
Despite the low frequency of infection and clinical malaria diagnosed in HIV-positive patients, no statically significant differences were noted either in parasitaemia or in occurrence of severe malaria, but this study presents two limitations: the use of a semi quantitative method for measuring parasitaemia whose low accuracy does not permit us to deduct any definite conclusions, and the lack of many laboratory parameters which render incomplete our criteria for classification of complicated malaria. However, a trend toward a more severe presentation characterized the HIV-malaria co-infection, with more frequent symptoms, such as dyspnea and diarrhea, occurring in the HIV population. Remarkably, anaemia was significantly associated with HIV-infection in malaria patients, and a trend towards a lower platelet number was also observed in co-infected individuals. Cytopenia of all major blood cell lines have been recognized in patients with chronic HIV since the beginning of the epidemic  and its incidence directly correlates with the degree of immuno suppression, although isolated abnormalities, particularly thrombocytopenia, may be encountered also as the initial presentation of HIV infection. In this analysis, anaemia was also associated with use of ART. ART has been demonstrated to improve preexisting anemia after use for at least 6 months , but the median length of therapy in our series was only four months. Therefore, the association between anaemia and ART probably only indicates a more advanced stage of HIV infection in treated patients. These results confirm and extend previous observations  emphasizing the need to avoid the dual infection, thus limiting hematological complications in these patients.
Lastly, the study was carried out in a 3rd level hospital where blood slides for confirming malaria have been largely used for some years; for the purposes of the analysis, MBSs were utilized for all hospitalized patients in Internal medicine, thus permitting us to reject about 10% of the presumptive malaria diagnoses; contrary to previous observations , an overestimation of malaria in HIV-positives when compared to HIV-negatives was not noted. The results of RDT were comparable to those of MBS in both HIV-infected and non-infected patients, independently of simultaneous treatment with ART; RDT use did not seem to provide an added value for correct diagnosis in patients already submitted to MBS, but the high sensitivity (94%) and specificity (96%) obtained in our survey confirmed that they are suitable in all fields in which MBS is not feasible and accuracy cannot be assured due to lack of expertise.