Impact of public health strategies on reducing AIDS mortality in southern Brazil

Abstract

Summary

In Brazil, all patients who fulfill the criteria for AIDS have had free access to antiretroviral therapy since 1996. We performed this cross-sectional study to evaluate the causes of death among 643 HIV-infected patients over three non-consecutive years (2000, 2006, and 2010), using their epidemiological, clinical, and laboratory data. The causes of death were classified as AIDS-defining or non-AIDS-defining conditions. We observed a progressive increase in the prevalence of HIV infection over the study period, although there was also a decrease in the mortality rate for various groups, and especially among pediatric patients. An AIDS-defining condition was recorded as the cause of death for approximately 30% of the patients. There was also a high frequency (>70%) of infectious and parasitic diseases, including opportunistic infections, and the most common diagnoses were septicemia, pneumonia, tuberculosis, and pneumocystosis. Acute respiratory failure was the underlying cause of death in 30% of these cases. Despite advances in HIV therapy, the mortality rate remains high in Brazil. As few Brazilian studies have investigated HIV/AIDS-related mortality, it is important to evaluate and improve the mortality notification databases, in order to provide information regarding the effects of HIV and to guide the implementation of appropriate healthcare measures.

Keywords

Introduction

The global use of highly active antiretroviral therapy (HAART) to treat HIV-infected individuals has helped to reduce the related morbidity and the mortality rates. Before HAART was introduced, approximately ten in 100,000 patients with AIDS died, although the mortality rate has subsequently fallen to approximately six in 100,000.¹ The increasing number of available therapies has also increased the likelihood of achieving positive outcomes, especially with the inclusion of more potent drugs that are safe and have more convenient dosing regimens.^2–6

In Brazil, patients with AIDS have had free access to antiretroviral drugs since 1996, and there has been a progressive increase in the access to these treatments. Administration of these treatments should follow the guidelines that were established by Brazilian experts and adopted by the Ministry of Health. In principle, the indication for antiretroviral therapy (ART) was only patients with a CD4 + T lymphocyte (TL) count of <350 cells/mm³ or patients who were diagnosed with an AIDS-defining condition. However, patients with a CD4 + TL count of <500 cells/mm³ became eligible for treatment in 2010, and the Ministry of Health has recently approved ART for all people who are living with HIV infection (PLWH) throughout Brazil. In 1992, a small number of reported patients with AIDS were treated from the date of their diagnosis (zidovudina monotherapy), although this proportion has progressively increased. In 2006, 180,000 patients were receiving antiretroviral drugs, and currently approximately 400,000 of the reported patients are receiving treatment, representing 44% of PLWH. In 2009, 17 antiretroviral drugs were available through the Brazilian Public Health Service.^5,7–9

The introduction of ART has led to a dramatic reduction in the mortality rate among patients with AIDS-defining conditions. However, there have been no corresponding changes for non-AIDS-defining conditions, which have subsequently become important causes of death.^10,11 Furthermore, despite universal access to treatment, the mortality rate has not decreased equally worldwide¹² and across Brazil, which may be related to social and regional heterogeneities that could affect the care of PLWH.¹³

The HIV epidemic in Brazil is characterised by increases in men who have sex with men (MSM) transmission, the number of cases among women, the number of cases among >65-year-old people, the number of cases among users of injectable drugs, disease spread to small and mid-sized cities, and impoverishment of the infected population.⁷ In Curitiba, Southern Brazil, notification has been performed for all patients with HIV infection/AIDS (even those who are not receiving treatment) since 2000, and in 2013, there were 10,244 PLWH (6798 men and 3446 women). This epidemic has been characteristically concentrated among young (30–39 years old) white men with a medium/high level of education, and with incidences of 127.7 cases per 100,000 male inhabitants and 53.9 cases per 100,000 female inhabitants.¹⁴

This study aimed to analyse the epidemiological profile, mortality rate, and the causes of death among PLWH from Curitiba, southern Brazil. The study was conducted over three non-consecutive years (2000, 2006, and 2010), which corresponded to a period of widespread access to medication for patients who were eligible for treatment based on the Brazilian criteria.¹⁵ Furthermore, this period included an increasing number of available drugs, which affected the management of patients with HIV infection.

Material and methods

This cross-sectional study was conducted at the Hospital de Clínicas/Universidade Federal do Paraná (HC-UFPR) and the Municipal Health Department of Curitiba during 2013. The HC-UFPR is an academic tertiary care hospital to which PLWH are referred. The HC-UFPR institutional review board approved this study (IRB#17442813.3.0000.0096).

Definitions

Some causes of death are directly related to AIDS, while other causes are not directly related to AIDS. Therefore, we classified the causes of death as AIDS-defining conditions or non-AIDS-defining conditions. The former conditions are related to immune system suppression, which may permit opportunistic infections and AIDS-related malignancies. The latter are related to long-term exposure to viremia and the side effects of antiretroviral drugs, which include cardiovascular disease, liver-related events, chronic kidney disease, and non-AIDS-related malignancies.

In the mortality notification database, each patient's cause of death is classified according to the International Statistical Classification of Diseases and Related Health Problems, 10^th Revision (ICD-10). In this study, we used the definitions from the World Health Organization's Underlying Causes of Death and Associated Causes of Death,¹⁶ which describe the disease or injury that initiated the sequence of events that led directly to death. The associated causes of death include the terminal and intervening causes, which may or may not be related to the events that led directly to death. The underlying and associated causes of death were assigned to groups of multiple causes of death.^17,18

The causes of death were recorded after considering all causes (single or multiple) and were classified into three groups based on the information from the databases:

Group A included cases with AIDS-defining conditions, which were a group of infections or malignancies that were based on the Definition of AIDS Cases in Adults and Children criteria.¹⁵ These conditions included salmonella septicemia, cryptosporidiosis, isosporidiosis, extrapulmonary tuberculosis, progressive multifocal leukoencephalopathy, herpes simplex, cytomegalovirus, pulmonary candidiasis, histoplasmosis, cryptococcosis, acute Chagas disease, toxoplasmosis, pneumocystosis (PCP), Kaposi sarcoma, cervical cancer, Burkitt tumor, other lymphomas, and chorioretinal inflammation.

Group B included cases with AIDS-suggesting conditions, which were a group of signs, symptoms, or diseases that generate scores for the epidemiological diagnosis of AIDS, based on the Rio de Janeiro/Caracas criteria.¹⁹ These conditions included herpes zoster, oral candidiasis, prolonged diarrhea (for ≥1 month), pulmonary tuberculosis, chronic anemia, malnutrition, dementia, inflammatory disease of the central nervous system, bacterial pneumonia, lymphadenopathy, asthenia, convulsions, wasting syndrome, and mental confusion.

Group C included all other cases of infections or conditions that were not included in Groups A and B, but could be related to long-term exposure to viremia or the side effects from ART. These conditions included acute gastroenteritis, pancreatitis, cholangitis, common meningitis, hepatitis and liver disease, non-AIDS-defining malignancies, cardiac and cerebrovascular diseases (e.g., flutter, atrial fibrillation, acute myocardial infarction, and cardiac malformation), circulatory diseases (e.g., hemorrhage and deep vein thrombosis), endocrine and metabolic disorders, renal insufficiency, mental disorders, and other conditions.

Methods

Data were collected from two national databases: the Mortality Information System (SIM) and the Notifiable Diseases Information System (SINAN). SIM is a national information system whose data are from the death certificate that is a pre-numbered instrument, which is available throughout Brazil. Health professionals must complete a standard death certificate, which facilitates the collection of information regarding all deaths throughout the country. The SINAN database contains epidemiological information regarding diseases with mandatory reporting requirements, which includes AIDS, and aims to provide information for morbidity profile analyses among the general population.

In this study, the SIM and SINAN databases were screened for all cases with a cause of death that was linked to HIV infection or AIDS during 2000, 2006, and 2010, among individuals from Curitiba, Southern Brazil. Table 1 lists the various ICD-10 codes that were evaluated to confirm or exclude cases of HIV infection or AIDS. The Epidemiology Division staff of the Municipal Health Secretary of Curitiba conducted these active searches to find these patients. To identify patients in both databases, a manual search was performed using the patient's name, birth date, and mother's name, which allowed us to collect their epidemiological, clinical, and laboratory data. The data that we recorded for the included patients were sex, age, cause of death, years of HIV infection, years since AIDS diagnosis, source of HIV infection, CD4 + TL count, and viral load. Epidemiological parameters, such as HIV prevalence, mortality rate, age- and sex-specific mortality rates, case fatality rate, and one-year survival rate for each studied year were estimated using surveillance data that were provided by the Epidemiology Division of the Municipal Health Secretary, Curitiba, Brazil.

Table 1.

Diseases and their ICD codes that led investigators to confirm or exclude HIV/AIDS involvement.

Disease	ICD-10 code
AIDS	B20–B24
Anemia	B63
Candidiasis	B37
Cytomegalovirus	B25
Cryptococcosis	B45
Malnutrition	E46
Chronic diarrhea	K52
Herpes zoster	B02
Histoplasmosis	B39
CNS lymphoma	C71.1
Isosporiasis	A07.3
Opportunistic mycoses	B48.7
Pneumocystosis	B59
Bilateral pneumonia	J15–J16
Kaposi's sarcoma	C46
Septicemia	A41
Toxoplasmosis	B58
Tuberculosis	A15–A19
Viral CNS infection	A86

Statistical analysis

Demographic and clinical data were compiled using JMP software (version 5.2.1; SAS Institute Inc., North Caroline, USA) and analysed using GraphPad Prism® software (version 5.03; Graph Pad Software Inc., La Jolla, CA, USA). Kaplan-Meier curves were fitted to describe the probability of survival according to time since HIV diagnosis for each study year, and the findings were compared using the log-rank test. These curves were constructed using information regarding the date of HIV diagnosis (from the SINAN database) and the date of death (from the SIM database). The frequencies for the main associated causes of death were reported as percentages, which were calculated relative to the number of male, female, and total deaths; these calculations assumed that deaths could have one or more associated causes. Fisher exact test or the χ² test was used to assess the inter-group differences, and the Mann-Whitney test was used for continuous variables. Results for continuous data are expressed as median ± interquartile range. All p-values were two-tailed, and a p-value of <0.05 was considered statistically significant.

Results

Among HIV-infected patients from Curitiba, we identified 214 deaths in 2000, 236 deaths in 2006, and 193 deaths in 2010. The demographic and laboratory findings for these patients are shown in Table 2. In 2000, deaths occurred in significantly younger individuals, and the median age of mortality among women was significantly lower than that among men (difference: 5.1 years, p = 0.0045). Between 2006 and 2010, patients exhibited a significant increase in the time since HIV diagnosis (p < 0.0001), as well as significantly higher CD4 + TL counts (p = 0.024) and significantly lower viral loads (p < 0.0001).

Table 2.

Demographic, clinical, and laboratory data recorded on the death certificates of HIV-infected patients in Curitiba.

Characteristic	2000 N = 214	2006 N = 236	2010 N = 193	P-value^a
Gender, n (%)				0.6212
Female	59 (27.6)	75 (31.8)	57 (29.7)
Male	155 (72.4)	161 (68.2)	135 (70.3)
Age, years, median (IQR)	36 (30–42)	41.2 (33.5–48.8)	42.6 (35.8–49.7)	<0.0001 0.0045
Female	32.2 (26–40.7)	40.5 (33.9–47.8)	41 (32.7–46.6)
Male	37.3 (31–43.1)	41.4 (33.3–49.3)	42.3 (37.2–50.8)
Time since HIV diagnosis, days, median (IQR)	100.5 (10.3–633.3)	482 (34–2,427)	588 (57.3–3,371)	<0.0001
Risk category, n (%)
Sexual behavior
Heterosexual	58 (27.5)	99 (42)	32 (16.5)
MSM	25 (11.7)	23 (9.5)	13 (7)
Drug use	28 (12.5)	32 (13.5)	9 (5)	0.13
Blood transfusion	0	0	1 (0.5)	NA
Not recorded	103 (48)	82 (35)	136 (71)	NA
Nadir TL CD4+, cells/mm³, median (IQR)	–	57 (26–161)	110 (31.5–223)	0.10
Not recorded, n (%)	214 (100)	179 (76)	108 (56)
Current TL CD4+, cells/mm³, median (IQR)	–	63 (29.2–189.7)	146 (52–285)	0.02
Not recorded, n (%)	236 (100)	179 (76)	108 (56)
Current viral load, copies/mL, median (IQR)	–	115,000 (11,150–233,500)	48,769 (8,549–118,210)	<0.0001
Not recorded, n (%)	236 (100)	175 (74)	121 (62.6)

Comparisons between the same sub-population across different years.

Note: IQR = Interquartile range, TL = T lymphocyte, MSM = men who have sex with men, NA = not applicable. Bold = statistically significant.

The prevalence of HIV infection was calculated to evaluate the burden of HIV in Curitiba during the study years, and we observed a progressive increase in the prevalence of HIV infection during that time (Table 3). However, the mortality rates in 2000 and 2006 were very similar, and a reduction in the mortality rate was observed in 2010. Further analysis according to age revealed significant mortality risk reductions in some subgroups, including children, adolescents, and young adults. Men had a higher mortality rate than women in all three study years, although the difference became smaller in the later years, which reflects a relative increase in AIDS-related mortality among women. Figure 1 shows Kaplan-Meier survival curves for each study year, and we observed significant differences in the median time from HIV diagnosis to death between the three study years (log-rank p < 0.001).

Table 3.

Epidemiological data recorded on the death certificates of HIV-infected patients in Curitiba (HIV Prevalence, Mortality rate, Age- and Gender-specific mortality rate were calculated using surveillance data provided by Epidemiology Division, Municipal Health Secretary, Curitiba, Brazil).

Characteristic	2000	2006	2010
	N = 214	N = 236	N = 193
Proportion of deaths by gender
Male to Female	2.6:1	2.1:1	2.3:1
HIV prevalence, % (95% CI)	0.22 (0.21–0.23)	0.37 (0.36–0.38)	0.50 (0.49–0.51)
HIV mortality rate (per 100,000)	13.8	13.1	11
HIV age-specific mortality rate (per 100,000)
≤13 y	1.3	0.9	0.3
14–19 y	0.5	0.4	0
20–29 y	14.2	6.7	6.1
30–39 y	32.3	25.3	17
40–49 y	28.8	32.8	28.4
≥50 y	7.2	18.6	12.5
HIV gender-specific mortality rate (per 100,000)
Male	20.3	18.9	16.2
Female	7.1	8.0	6.2
Case fatality rate, % (95% CI)	6.1 (5.3–6.9)	3.5 (3.1–3.9)	2.1 (1.9–2.5)
One-year survival rate, % (95% CI)	96 (95.3–96.6)	97.7 (97.3–98)	98.4 (98.1–98.6)

Note: CI = confidence interval.

Figure 1.

Kaplan-Meier survival curves according to time since HIV diagnosis.

One AIDS-defining condition was recorded as the cause of death for approximately 30% of the patients (Table 4). A significant difference in the cause of death was observed between 2000 and 2006 (p = 0.039), due to an increase in the incidences of Group C infections and conditions. However, this difference was not observed when we compared 2006 and 2010 (Table 4). The average number of diagnoses that are listed on the death certificate is an important criterion for recovery, although there were no differences in the mean number of diagnoses for 2000, 2006, and 2010.

Table 4.

Total deaths of HIV-infected patients in Curitiba and the cause(s) of death recorded on the death certificates.

	2000		2006		2010
	n	%	n	%	n	%	P-value
Cause(s) of death
Group A	76	36	70	30	67	35	0.039
Group B	37	17	24	10	21	11
Group C	101	47	142	60	104	54
Total deaths	214		236		193		NA
Mean number of diagnoses at time of death	1.48 (±0.9)		2.54 (±0.6)		2.42 (±0.7)		NA

Note: NA = not applicable. Bold = statistically significant.

Table 5 shows the frequencies of the main associated causes of death. There were high frequencies of infectious and parasitic diseases, including opportunistic infections, and at least one of these causes were reported on >70% of the death certificates. Important conditions among these causes included septicemia, pneumonia, tuberculosis, PCP, toxoplasmosis, and cryptococcosis. Furthermore, acute respiratory failure was the underlying cause of death in approximately 30% of the patients.

Table 5.

Causes of death of HIV-infected patients in Curitiba, Southern Brazil, according to gender.

Cause of death	2000 N = 214		2006 N = 236		2010 N = 193
Cause of death	Male n (%)	Female n (%)	Male n (%)	Female n (%)	Male n (%)	Female n (%)
Acute respiratory failure	50 (23)	15 (7)	56 (24)	19 (8)	24 (12)	15 (8)
Septicemia	31 (14)	10 (5)	43 (18)	19 (8)	46 (24)	16 (8)
Pneumonia	29 (13)	8 (4)	49 (21)	18 (8)	36 (18)	17 (9)
Pneumocystosis	20 (9)	2 (1)	16 (7)	8 (3)	12 (6)	4 (2)
Tuberculosis	18 (8)	5 (2)	8 (3)	1 (0.5)	7 (4)	3 (1.5)
Wasting syndrome	13 (6)	6 (3)	8 (3)	4 (2)	7 (4)	1 (0.5)
Hepatic disease	10 (5)	2 (1)	9 (4)	4 (2)	12 (6)	3 (1.5)
Neurocryptococcosis	10 (5)	6 (3)	7 (3)	4 (2)	6 (3)	4 (2)
Neurotoxoplasmosis	9 (4)	7 (3)	6 (2.5)	5 (2)	7 (4)	7 (4)
Extrapulmonary tuberculosis	7 (3)	1 (0.5)	8 (3)	3 (1)	7 (4)	4 (2)
Neoplasm AIDS-related	6 (3)	4 (2)	5 (2)	4 (2)	10 (5)	6 (3)
Candidiasis	6 (3)	1 (0.5)	1 (0.5)	0	1 (0.5)	0
Other neoplasm	5 (2)	3 (1)	4 (2)	2	3 (1.5)	0
Anemia	4 (2)	2 (1)	1 (0.5)	3 (1)	2 (1)	1 (0.5)
Other interstitial lung disease	4 (2)	2 (1)	2 (1)	3 (1)	3 (1.5)	1 (0.5)
Diarrhea	2 (1)	1 (0.5)	2 (1)	3 (1)	0	1 (0.5)
Cardiovascular disease	1 (0.5)	1 (0.5)	2 (1)	1(0.5)	2 (1)	0
HCMV disease	–	1 (0.5)	2 (1)	2 (1)	1 (0.5)	0

Note: HCMV = human cytomegalovirus.

Discussion

In Brazil, the diagnosis and treatment of PLWH follows the National STD/AIDS Program's clinical guidelines. Over the years, HIV has been diagnosed earlier, and new therapeutics have been introduced. In 2002, the dispensing program for antiretroviral drug program for patients was widely established. In subsequent years the strategies have continued to improve by extending the network of laboratories able to perform viral load monitoring and genotyping-resistance testing (for treatment failure management) in 2006 and by adding new classes of therapeutic drugs in 2009.^5,20 Seeking to evaluate changes over the decade, we selected the year 2006 and the farthest years from this date (2000 and 2010) to carry out this study. It was not included patients before 2000 because most of them had restricted access to ARV.

Several tools have been used to evaluate the effectiveness of these strategies, among these the system of disease notification, causes of death, and mortality and fatality rates could be valuable sources of information. Therefore, the SIM and SINAN databases have been important tools for the Brazilian Health Ministry to analyse the epidemiological information and provide the official data regarding the effects of these diseases on public health. To find individuals in distinct systems, linkage methods are typically used to evaluate large databases, though this approach is associated with identification failures that are mainly due to typing errors. Thus, the manual method has achieved better results when attempting to identify a relatively small number of cases.

Despite the decreasing overall AIDS mortality rate in Brazil during the last ten years, this trend has not been observed for all regions.⁵ In the present study, we did not observe an overall reduction in the mortality rate in Curitiba, Southern Brazil. However, when we evaluated the data according to age, we observed a reduced mortality rate among patients who were <13 years old, which is likely related to decreased vertical transmission and improved care for infected patients. There was also a reduced mortality rate among individuals who were 14–39 years old during the study years, which may be related to their better access to therapy. Though, we did not observe any difference among patients who were 40–49 years old and observed an increased mortality rate among patients who were >50 years old. These findings may be related to an increased detection rate for PLWH in these respective age groups.¹⁴

The slight reduction in the mortality rates between 2006 and 2010 was not significant, and we did not observe any significant changes in the causes of death. Despite these patients having significantly higher CD4 + TL counts, the majority of patients still exhibited severe immunodeficiency (a median CD4 + TL count of ≤ 200 cells/mm³) and had AIDS-defining conditions at the time of their death (Group A). Furthermore, > 30% of patients were diagnosed with AIDS within 90 days of their date of death, which indicates that there is an urgent need for earlier HIV detection.²⁰ Analysis of Kaplan-Meier survival curves according to time since HIV diagnosis revealed that patients had significantly longer survival times in 2006 and 2010, which indicates that these patients likely had access to ART before their death. However, as these patients' medical records were not available, it was not possible to evaluate whether the earlier deaths were associated with limited access to antiretroviral drugs or therapies for opportunistic infections. Nevertheless, the patients were significantly older in 2010, compared to those in 2000, which suggests an increased survival, even with a higher frequency of infection among the young patients. Therefore, the accurate diagnosis and early treatment of AIDS-associated diseases remain significant challenges to the management of PLWH in Brazil and worldwide.²¹

In contrast to the findings of Lopes et al.,⁹ we observed a significantly higher number of deaths among men, compared to the number among women. The proportion of death by sex in PLWH in Curitiba (2.6, 2.1, and 2.3 for 2000, 2006, and 2010, respectively) was similar to, although slightly higher than, the ratio for the general Brazilian population (1.9).⁵ During this time, we did not observe any significant changes in the proportions of men and women who were infected with HIV. In this context, women are more likely to seek medical care and to have better treatment adherence, which may have affected our findings.⁹ Furthermore, men typically exhibit a greater time of infection, compared to women. In the present study, we did not observe any significant sex-related differences in mortality, and the mortality rate remained approximately 2–3-fold higher for men, which is similar to the previously reported findings.^5,22

Acute respiratory failure after pulmonary infection was the leading cause of death during all three study years. However, the pathogen that was associated with the infection was not reported in most cases. Furthermore, pulmonary involvement is a common concern in most Brazilian hospitals, as the investigation and identification of the related microorganism(s) is uncommon. Nevertheless, Soeiro et al.²³ reported their pulmonary findings from 250 autopsies of HIV-infected patients in Brazil and found bacterial bronchopneumonia and Pneumocystis jiroveci pneumonia in > 50% of their cases. These findings are similar to our results. Pulmonary involvement has also been reported in 80–94% of Indian patients with HIV infection or AIDS.²⁴ Moreover, the high burden of PCP and bacterial infections in the present study indicates the importance of prophylactic PCP treatment, implementation of prevention programmes, cessation of tobacco use, and immunisations against influenza and pneumococcus among PWLH.^21,25 Although tuberculosis has been frequently diagnosed in PLWH, we did not observe a high frequency of this disease in the present study. TB was the fourth cause of death found among the studied patients. Nonetheless, some cases with diagnosis of ‘pneumonia’, ‘wasting syndrome’ and ‘other interstitial lung disease’ could have been an under diagnosis of TB, but there was no report of this investigation in SINAM and SIM database. Moreover, it must be noted that extrapulmonary tuberculosis is more common among very immunosuppressed patients, and this condition is difficult to diagnose because it is typically paucibacillary tuberculosis.

In the present study, the frequency of AIDS-related malignancies did not increase significantly over time, despite reports of increased frequencies for non-AIDS-related malignancies in this population.^26–28 Furthermore, increases in the frequencies of tumors and other non-AIDS-related events have been reported in patients who are receiving ART, older, and who have a longer survival.²⁴ This discrepancy may be related to our including mainly patients with a recent diagnosis, and the increased frequency of infections among patients who are >50 years old, which may have contributed to the low frequency of these events.

Due to the substantial reductions in morbidity and mortality among PLWH, HIV infection is currently considered a potentially manageable chronic illness. However, the success of therapy depends on several factors, such as access to treatment and health services, viral, immune response characteristics, and individual behavioral factors (e.g., treatment adherence).^29,30 Therefore, an improved understanding of the changes in these factors may help to explain the regional heterogeneity in AIDS-related outcomes and allow for a more focused approach to controlling the AIDS epidemic in Brazil.

This study has some limitations. First, the information was based on a review of death certificates, and medical records were not available to confirm the epidemiological, clinical, and laboratory data. Second, the mean of number of diagnoses at the time of death in 2000 was very low (1.48), and unexpectedly low numbers were also observed in the subsequent years (despite small increases in the numbers of diagnoses). This finding suggests a lack of laboratory investigation or even limited commitment from the healthcare professionals to correctly complete the required forms. Nonetheless, the staff of the Epidemiologic Division reviewed all forms and completed the data before including them in the systems. Third, the Medication Logistics Control System (SICLOM) was implemented in 2002 to monitor the use of ART, and though after 2006 it has already included all the medication dispensing units, access to this information is currently controlled; therefore, we could not evaluate this data in the present study. Fourth, although some information regarding antiretroviral drug use, viral load, and CD4 + TL counts are available in the national database, these data were scarce in 2000, and even the present records do not contain all relevant information, which makes the search process laborious and subject to error. Fifth, we excluded cases with various causes of death, as suicide, homicide, and accidental death, because information regarding HIV infection was not included on these cases' death certificates. Finally, it is important to note that we evaluated a relatively short period, and it was not possible to assess the duration of HIV infection in these patients, which is a critical mortality risk factor that can confound other correlations.³ Nevertheless, despite these limitations, database analyses have been important methods for assessing the effects of HIV infection and AIDS on population health, and our findings may help guide the introduction of additional therapeutic and preventive approaches by the Brazilian Ministry of Health.

The findings of the present study highlight the fact that, despite advances in HIV therapy, the mortality rate remains unacceptably high. Furthermore, the late diagnosis of HIV infection is a serious public health issue in Brazil, which may be due to limited access to healthcare services. Thus, measures are needed to address this challenge, such as expanding the HIV testing network, opening testing centers in medium and small cities, supporting non-governmental organisations that serve at-risk populations, providing commercial tests for self-diagnosis, and informing the population regarding the importance of early diagnosis. In addition, studies assessing the mortality of PLWH in the Brazilian population should be encouraged through critical analysis of the databases available. Therefore, it is important to ensure that Brazilian mortality notification databases are up-to-date, organised, and improved, as their information is crucial to evaluating the effect of HIV infection, and can be used to promote improvements in the control of this disease.

Footnotes

Acknowledgments

SMR has a fellowship from Conselho Nacional de Pesquisa-CNPq.

Author contributions

SMR and CER designed the study and analysed the results. JPMMT, MA, and GAS collected information from the database. SMR and SMA wrote the manuscript.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

UNAIDS. Global AIDS response progress reporting 2014. http://www.unaids.org/sites/default/files/media_asset/GARPR_2014_guidelines_en_0.pdf (accessed 26 February 2015).

Palella

Delaney

Moorman

. Declining morbidity and mortality among patients with advanced immunodeficiency virus infection. NEJM 1998; 338: 853–860.

Bhaskaran

Hamouda

Sannes

. Changes in the risk of death after HIV seroconversion compared with mortality in the general population. JAMA 2008; 300: 51–59.

Zolopa

Andersen

Komarow

. Early antiretroviral therapy reduces AIDS progression/death in individuals with acute opportunistic infections: a multicenter randomized strategy trial. PLoS One 2009; 4: e5575.

Brazilian Ministry of Health, Division of Health Surveillance, National Program on STD/AIDS. Epidemiologic Bulletin AIDS. 2013. http://www.aids.gov.br/sites/default/files/anexos/publicacao/2013/55559/_p_boletim_2013_internet_pdf_p__51315.pdf (accessed 15 March 2015).

Murray

CJL

Ortblad

Guinovart

. Global, regional, and national incidence and mortality for HIV, tuberculosis, and malaria during 1990-2013: a systematic analysis for the global burden disease study 2013. Lancet 2014; 384: 1005–1070.

Greco

Simão

. Brazilian policy of universal access to AIDS treatment: sustainability challenges and perspectives. AIDS 2007; 21: 37–45.

Brazilian Ministry of Health, Division of Health Surveillance, National Program on STD/AIDS. Adesão ao tratamento antiretroviral no Brasil: Coletânea de estudos do projeto ATAR. 2010. http://www.aids.gov.br/sites/default/files/atar-web.pdf (accessed 16 February 2015).

Lopes

LAB

Silva

EMK

. Biological, behavioral, and socioeconomic factors associated with death from AIDS in Brasilia, Brazil, in 2007. Rev Soc Bras Med Trop 2012; 45: 448–452.

10.

Lewden

May

Rosenthal

. Changes in causes of death among adults infected by HIV between 2000 and 2005: the “Mortalité 2000 and 2005” surveys (ANRS EN19 and Mortavic). J Acquir Immunodefic Synd 2008; 48: 590–598.

11.

Antiretroviral Therapy Cohort Collaboration. Causes of death in HIV-1 infected patients treated with antiretroviral therapy, 1996-2006: collaborative analysis of 13 cohort studies. Clin Infect Dis 2010; 50: 1387–1396.

12.

Carriquiry

Fink

Koethe

. Mortality and loss to follow-up among HIV-infected persnos on long-term antirretroviral therapy in Latin America and the Caribbean. J Int AIDS Soc 2015; 18: 20016.

13.

Malta

Silva

Magnanini

. Improvement of HAART in Brazil, 1998-2008: a nationwide assessment of survival times after AIDS diagnosis among men who have sex with men. BMC Public Health 2015; 15: 226.

14.

Municipal Health Department, Curitiba-PR. HIV/AIDS Epidemiological Bulletin. 2014. http://multimidia.curitiba.pr.gov.br/2014/00156481.pdf (accessed 26 March 2015).

15.

Brazilian Health Ministry. Recomendações para terapia antirretroviral em adultos infectados pelo HIV. 2008. http://www.aids.gov.br/sites/default/files/consensoAdulto005c_2008montado.pdf (accessed 15 July 2015).

16.

WHO. International Statistical Classification of Diseases and Related Health Problems 10th Revision Geneva: World Health Organization; 1992. http://apps.who.int/classifications/icd10/browse/2015/en (accessed 23 March 2015).

17.

Santo

Pinheiro

Jordani

. Causas básicas e associadas de morte por Aids, Estado de São Paulo, Brasil, 1998. Rev Saude Publica 2000; 34: 581–588.

18.

State Department of Health of São Paulo. STD/AIDS Reference and Training Center. Dados para repensar a AIDS no estado de São Paulo. 2011. http://produtos.seade.gov.br/produtos/publicacoes/pub_Repensar_Aids_2011.pdf (accessed 23 March 2015).

19.

Brazilian Health Ministry. Critérios para definição de casos de AIDS em crianças e adultos. 2004. http://www.aids.gov.br/sites/default/files/criterios_aids_2004.pdf (accessed 15 July 2015).

20.

Grangeiro

Escuder

Cassanote

. The HIV cohort study: design, methods and participant characteristics. PLoS One 2014; 9: e95673.

21.

Ford

Shubber

Meintjes

. Causes of hospital admission among people living with HIV worldwide: a systematic review and meta-analysis. Lancet HIV 2015; 2: e438–e444.

22.

Pinto

Dos Santos Fernandes

de Oliveira

. Transitioning through AIDS epidemics – gender and temporality. Braz J Infect Dis 2015; 19: 657–659.

23.

Soeiro

Hovnanian

Parra

. Post-mortem histological pulmonary analysis in patients with HIV/AIDS. Clinics 2008; 64: 497–502.

24.

Lanjewar

Duggal

. Pulmonary pathology in patients with AIDS: an autopsy study from Mumbai. HIV Med 2001; 2: 266–271.

25.

Luz

Bruyand

Ribeiro

. AIDS and non-AIDS severe morbidity associated with hospitalizations among HIV-infected patients in two regions with universal access to care and antiretroviral therapy, France and Brazil, 2000-2008: hospital-based cohort studies. BMC Infect Diseases 2014; 14: 278.

26.

Herida

Mary-Kruase

Kaphan

. Incidence of non-AIDS-defining cancer before and during the highly active antiretroviral therapy era in a cohort of human immunodeficiency virus-infected patients. J Clin Oncol 2003; 21: 3447–3453.

27.

Clifford

Polesel

Rickenbach

. Cancer risk the Swiss HIV Cohort study: associations with immunodeficiency, smoking, and highly active antiretroviral therapy. J Natl Cancer Inst 2005; 97: 425–432.

28.

Castilho

Luz

Shepherd

. HIV and cancer: a comparative retrospective study of Brazilian and U.S. clinical cohorts. Infect Agent Cancer 2015; 10: 4.

29.

Paterson

Swindells

Mohr

. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med 2000; 133: 21–30.

30.

Berg

Arnsten

. Practical and conceptual challenges in measuring antiretroviral adherence. J Acquir Immune Defic Syndr 2006; 43: S79–S87.