Incidence and determinants of severe morbidity among HIV-infected patients from Rio de Janeiro, Brazil, 2000-2010

Supervised machine learning algorithms to predict the duration and risk of long-term hospitalization in HIV-infected individuals: a retrospective study

Objective

The study aimed to use supervised machine learning models to predict the length and risk of prolonged hospitalization in PLWHs to help physicians timely clinical intervention and avoid waste of health resources.

Methods

Regression models were established based on RF, KNN, SVM, and XGB to predict the length of hospital stay using RMSE, MAE, MAPE, and R2, while classification models were established based on RF, KNN, SVM, NN, and XGB to predict risk of prolonged hospital stay using accuracy, PPV, NPV, specificity, sensitivity, and kappa, and visualization evaluation based on AUROC, AUPRC, calibration curves and decision curves of all models were used for internally validation.

Results

In regression models, XGB model performed best in the internal validation (RMSE = 16.81, MAE = 10.39, MAPE = 0.98, R2 = 0.47) to predict the length of hospital stay, while in classification models, NN model presented good fitting and stable features and performed best in testing sets, with excellent accuracy (0.7623), PPV (0.7853), NPV (0.7092), sensitivity (0.8754), specificity (0.5882), and kappa (0.4672), and further visualization evaluation indicated that the largest AUROC (0.9779), AUPRC (0.773) and well-performed calibration curve and decision curve in the internal validation.

Conclusion

This study showed that XGB model was effective in predicting the length of hospital stay, while NN model was effective in predicting the risk of prolonged hospitalization in PLWH. Based on predictive models, an intelligent medical prediction system may be developed to effectively predict the length of stay and risk of HIV patients according to their medical records, which helped reduce the waste of healthcare resources.

Early mortality in a cohort of people living with HIV in Rio de Janeiro, Brazil, 2004-2015: a persisting problem

BACKGROUND

Global mortality from AIDS-related diseases has been declining since 2005, resulting primarily from the widespread use and early initiation of combination antiretroviral therapy. Despite the significant improvements, high rates of early mortality, usually defined as that occurring within the 1st year of entry to care, have been observed, especially in resource-limited settings. This analysis draws upon data from an observational cohort of people with HIV (PWH) followed at a reference center for HIV/AIDS care and research in the city of Rio de Janeiro, Brazil, to identify the pattern and factors associated with early mortality.

METHODS

The study population includes PWH aged 18 or older followed at the National Institute of Infectious Diseases Evandro Chagas who were enrolled between 2004 and 2015. The primary outcome was early mortality, defined as deaths occurring within 1 year of inclusion in the cohort, considering two follow-up periods: 0 to 90 days (very early mortality) and 91 to 365 days (early mortality). Cox proportional hazards models were used to identify the variables associated with the hazard of very early and early mortality.

RESULTS

Overall, 3879 participants contributed with 3616.4 person-years of follow-up. Of 220 deaths, 132 happened in the first 90 days and 88 between 91 and 365 days. Very early mortality rate ratios (MRR) show no statistically significant temporal differences between the periods 2004-2006 to 2013-2015. In contrast, for early mortality, a statistically significant decreasing trend was observed: mortality rates in the periods 2004-2006 (MR = 5.5; 95% CI 3.9-7.8) and 2007-2009 (MR = 3.9; 95% CI 2.7-5.7) were approximately four and three-fold higher when compared to 2013-2015 (MR = 1.4; 95% CI 0.7-2.7). Low CD4 count and prior AIDS-defining illness were strongly associated with higher hazard ratios of death, especially when considering very early mortality.

CONCLUSIONS

The present study shows an excess of mortality in the 1st year of follow-up with no changes in the mortality rates within 90 days among PWH from Rio de Janeiro. We note the significant impact of initiating treatment with immunosuppression, as evidenced by the increased risk of death among those with low CD4 cell count and with AIDS-defining illnesses.

Iron intake is positively associated with viral load in antiretroviral naïve Brazilian men living with HIV

BACKGROUND

Iron homeostasis contribute for the human immunodeficiency virus (HIV) pathogenesis.

OBJECTIVES

We assessed the iron intake pattern in antiretroviral naïve Brazilian men living with HIV correlating with clinical and nutritional parameters.

METHODS

The iron consumption mean was estimated according to a food frequency questionnaire (FFQ), and a 3-day food record (3dFR) submitted to the patients. HIV viral load, CD4+ T cell counts, serum iron, haematological and anthropometrics parameters were recorded.

FINDINGS

Fifty-one HIV-infected adult men naïve for antiretroviral therapy (ART) were enrolled. The mean age of participants was 35 (SEM ± 1.28) years old, with mean time of HIV-1 infection of 1.78 (0-16.36, min-max) years. Majority (41.18%) had complete secondary, and 21.57% had tertiary educational level. The income was around 1x (54.90%) to 2x (41.18%) minimum wage. Fifty-four percent showed normal weight, while 40% were overweight. The patients showed normal mean values of haematological parameters, and mean serum iron was 14.40 µM (SEM ± 0.83). The FFQ showed moderate correlation with the 3dFR (ρ = 0.5436, p = 0.0009), and the mean values of iron intake were 10.55(± 0.92) mg/day, recorded by FFQ, and 15.75(± 1.51) mg/day, recorded by 3dFR. The iron intake, recorded by FFQ, negatively correlated with serum iron (ρ = -0.3448, p = 0.0132), and did not have influence in the CD4+ T cell counts [e.B 0.99 (0.97-1.01, 95% confidence interval (CI), p = 0.2]. However, the iron intake showed a positive effect in HIV viral load [e.B 1.12 (1.02-1.25, 95%CI), p < 0.01].

MAIN CONCLUSIONS

This study draws attention for the importance of iron intake nutritional counseling in people living with HIV. However, more studies are required to clarify the association between high iron intake and HIV infection and outcome.

The cost of hospital care for HIV patients in Colombia: an insurer's perspective

The aim of this study was to evaluate the cost derived from the hospitalization of people living with HIV (PLHIV) in Colombia between 2011 and 2015. This is an analysis of the direct cost of PLHIV hospitalization from the perspective of an insurer of the Colombian General Social Security System. The costs were calculated in Colombian pesos and corrected for inflation on the basis of the 2017 Consumer Price Index of the Bank of the Republic of Colombia. It was converted to US dollars at the Market Representative Exchange Rate of the same year. We analyzed 1129 hospitalizations in 612 PLHIV, of which 12% started with a diagnosis of HIV during the same hospitalization, with the majority in the AIDS stage (63%). The median overall cost of hospitalizations was US$1509 (25th and 75th percentiles: US$711-US$3254), being even higher in patients with AIDS and as the CD4 T lymphocyte count decreased. The cost derived from the medical care of PLHIV increases as the clinical control of the disease worsens, and it is a key indicator of the impact of the strategies implemented for the timely identification of the infection and subsequent management of the disease.

Severe infection increases cardiovascular risk among HIV-infected individuals

BACKGROUND

The identification and management of cardiovascular risk factors became a major clinical issue among HIV-infected individuals in the post-cART era. As in the past decades the link between acute infections and cardiovascular diseases became clear in the general population, we sorted to investigate the role of severe infections on incident cardiovascular diseases (CVDs) among HIV-infected individuals.

METHODS

HIV-infected individuals aged ≥18 years, with no history of CVD were followed from January 2000 to December 2013 until the occurrence of the first CVD event, death or end of study, whichever occurred first. To explore the effect of severe infections on the incidence of CVD we used extended Cox regression models and stratified post-hospitalization follow-up time into three periods: < 3 months, 3-12 months and > 12 months post discharge.

RESULTS

One hundred-eighty four persons from 3384 HIV-infected individuals developed incident CVD events during the follow-up (incidence rate = 11.10/1000 PY (95%CI: 9.60-12.82)). Risk of an incident CVD was 4-fold higher at < 3 months post-hospitalization for severe infections (adjusted hazard ratio [aHR], 4.52; 95% confidence interval [CI] 2.46-8.30), after adjusting for sociodemographic and clinical factors as well as comorbidities. This risk remained significant up to one year (3-12 months post hospital discharge aHR 2.39, 95% CI 1.30-4.38). Additionally, non-white race/ethnicity (aHR 1.49, 95% CI 1.10-2.02), age ≥ 60 years (aHR 2.01, 95% CI 1.01-3.97) and hypertension (aHR 1.90, 95% CI 1.38-2.60) were associated with an increased risk of CVD events. High CD4 (≥ 500 cells/mm3: aHR 0.41, 95% CI 0.27-0.62) and cART use (aHR 0.21, 95% CI 0.14-0.31) reduced the risk of CVD events.

CONCLUSIONS

We provide evidence for a time-dependent association between severe infection and incident cardiovascular disease in HIV-infected individuals. cART use, and high CD4 count were significantly associated with reduced hazards of CVD.

Community-acquired lower respiratory tract infections in HIV-infected patients on antiretroviral therapy: predictors in a contemporary cohort study

Community-acquired pneumonia represents the most frequent bacterial infection in patients with HIV/AIDS.

PURPOSE

We aimed to assess variables associated with lower respiratory tract infection (LRTI) among HIV-infected adults using ART.

METHODS

A cohort study of HIV-infected patients aged ≥18 years, enrolled from 2000 to 2015, on ART for at least 60 days, with primary outcome as the 1st episode of LRTI during follow-up. The independent variables included were sex at birth, age, race/skin color, educational level, tobacco smoking, alcohol use, cocaine use, diabetes mellitus, CD4 count, HIV viral load, influenza and pneumococcal vaccination. Extended Cox proportional hazards models accounting for time-updated variables were fitted to assess LRTI predictors.

RESULTS

2669 patients were included; median follow-up was 3.9 years per patient. LRTI was diagnosed in 384 patients; incidence rate was 30.7/1000 PY. In the unadjusted Cox extended models, non-white race [crude hazard ratio (cHR) 1.28, p = 0.020], cocaine use (cHR 2.01, p < 0.001), tobacco smoking (cHR 1.34, p value 0.007), and HIV viral load ≥400 copies/mL (cHR 3.40, p < 0.001) increased the risk of LRTI. Lower risk of LRTI was seen with higher educational level (cHR 0.61, p < 0.001), rise in CD4 counts (cHR 0.81, p < 0.001, per 100 cells/mm3 increase), influenza (cHR 0.60, p = 0.002) and pneumococcal vaccination (cHR 0.57, p < 0.001). In the adjusted model, aHR for CD4 count was 0.86, for cocaine use 1.47 and for viral load ≥400 copies 2.20.

CONCLUSIONS

LRTI has a high incidence in HIV-infected adults using ART. Higher CD4 counts and undetectable viral loads were protective, as were pneumococcal and influenza vaccines.

Thirty-day Readmission Rates in an HIV-infected Cohort From Rio de Janeiro, Brazil

BACKGROUND

The 30-day readmission rate is an indicator of the quality of hospital care and transition to the outpatient setting. Recent studies suggest HIV infection might increase the risk of readmission although estimates of 30-day readmission rates are unavailable among HIV-infected individuals living in middle/low-income settings. Additionally, factors that may increase readmission risk in HIV-infected populations are poorly understood.

METHODS

Thirty-day readmission rates were estimated for HIV-infected adults from the Instituto Nacional de Infectologia Evandro Chagas/Fiocruz cohort in Rio de Janeiro, Brazil, from January 2007 to December 2013. Cox regression models were used to evaluate factors associated with the risk of 30-day readmission.

RESULTS

Between January 2007 and December 2013, 3991 patients were followed and 1861 hospitalizations were observed. The estimated 30-day readmission rate was 14% (95% confidence interval: 12.3 to 15.9). Attending a medical visit within 30 days after discharge (adjusted hazard ratio [aHR] = 0.73, P = 0.048) and being hospitalized in more recent calendar years (aHR = 0.89, P = 0.002) reduced the risk of 30-day readmission. In contrast, low CD4 counts (51-200 cells/mm³: aHR = 1.70, P = 0.024 and ≤ 50 cells/mm³: aHR = 2.05, P = 0.003), time since HIV infection diagnosis ≥10 years (aHR = 1.58, P = 0.058), and leaving hospital against medical advice (aHR = 2.67, P = 0.004) increased the risk of 30-day readmission.

CONCLUSIONS

Patients with advanced HIV/AIDS are most at risk of readmission and should be targeted with prevention strategies to reduce this risk. Efforts to reduce discharge against medical advice and to promote early postdischarge medical visit would likely reduce 30-day readmission rates in our population.

Hospitalization rates, length of stay and in-hospital mortality in a cohort of HIV infected patients from Rio de Janeiro, Brazil

In this study, we evaluated trends in hospitalization rates, length of stay and in-hospital mortality in a cohort of HIV-infected patients in Rio de Janeiro, Brazil, from 2007 through 2013. Among the 3991 included patients, 1861 hospitalizations occurred (hospitalization rate of 10.44/100 person-years, 95% confidence interval 9.98–10.93/100 person-years). Hospitalization rates decreased annually (per year incidence rate ratio 0.92, 95% confidence interval 0.89–0.95) as well as length of stay (median of 15 days in 2007 vs. 11 days in 2013, p-value for trend < 0.001), and in-hospital mortality (13.4% in 2007 to 8.1% in 2013, p-value for trend = 0.053). Our results show that, in a middle-income setting, hospitalization rates are decreasing over time and non-AIDS hospitalizations are currently more frequent than those related to AIDS. Notwithstanding, compared with high-income settings, our patients had longer length of stay and higher in-hospital mortality. Further studies addressing these outcomes are needed to provide information that may guide protocols and interventions to further reduce health-care costs and in-hospital mortality.

Traditional and HIV-specific risk factors for cardiovascular morbidity and mortality among HIV-infected adults in Brazil: a retrospective cohort study

BACKGROUND

Antiretroviral therapy (ART) agents potentially associated with adverse metabolic profiles are commonly used in low- and middle-income countries. We assessed risk factors for cardiovascular disease (CVD)-related morbidity and mortality in a cohort of HIV-infected, ART-treated adults in Rio de Janeiro, Brazil.

METHODS

Hospital records and mortality data between 2000-2010 were examined for incident CVD-related ICD-10 and Coding of Death in HIV diagnoses among adults ≥18 years old on ART, enrolled in an observational cohort. Poisson regression models assessed associations between demographic and clinical characteristics and ART agent or class on CVD event risk.

RESULTS

Of 2960 eligible persons, 109 had a CVD event (89 hospitalizations, 20 deaths). Participants were 65 % male, 54 % white, and had median age of 37 and 4.6 years on ART. The median nadir CD4(+) T lymphocyte count was 149 cells/mm(3). The virologic suppression rate at the end of study follow-up was 60 %. In multivariable models, detectable HIV-1 RNA prior to the event, prior CVD, less time on ART, age ≥40 at study baseline, nadir CD4(+) T lymphocyte count ≤50 cells/mm(3), non-white race, male gender, and a history of hypertension were significantly associated with CVD event incidence (p < 0.05), in order of decreasing strength. In multivariate models, cumulative use of tenofovir, zidovudine, efavirenz and ritonavir-boosted atazanavir, darunavir and/or lopinavir were associated with decreased CVD event risk. Recent tenofovir and boosted atazanavir use were associated with decreased risk, while recent stavudine, nevirapine and unboosted nelfinavir and/or indinavir use were associated with increased CVD event risk.

CONCLUSIONS

Virologic suppression and preservation of CD4(+) T-lymphocyte counts were as important as traditional CVD risk factor burden in determining incident CVD event risk, emphasizing the overall benefit of ART on CVD risk and the need for metabolically-neutral first- and second-line ART in resource-limited settings.

Survival benefits of antiretroviral therapy in Brazil: a model-based analysis

Objective

In Brazil, universal provision of antiretroviral therapy (ART) has been guaranteed free of charge to eligible HIV-positive patients since December 1996. We sought to quantify the survival benefits of ART attributable to this programme.

Methods

We used a previously published microsimulation model of HIV disease and treatment (CEPAC-International) and data from Brazil to estimate life expectancy increase for HIV-positive patients initiating ART in Brazil. We divided the period of 1997 to 2014 into six eras reflecting increased drug regimen efficacy, regimen availability and era-specific mean CD4 count at ART initiation. Patients were simulated first without ART and then with ART. The 2014-censored and lifetime survival benefits attributable to ART in each era were calculated as the product of the number of patients initiating ART in a given era and the increase in life expectancy attributable to ART in that era.

Results

In total, we estimated that 598,741 individuals initiated ART. Projected life expectancy increased from 2.7, 3.3, 4.1, 4.9, 5.5 and 7.1 years without ART to 11.0, 17.5, 20.7, 23.0, 25.3, and 27.0 years with ART in Eras 1 through 6, respectively. Of the total projected lifetime survival benefit of 9.3 million life-years, 16% (or 1.5 million life-years) has been realized as of December 2014.

Conclusions

Provision of ART through a national programme has led to dramatic survival benefits in Brazil, the majority of which are still to be realized. Improvements in initial and subsequent ART regimens and higher CD4 counts at ART initiation have contributed to these increasing benefits.

Predictors of opportunistic illnesses incidence in post combination antiretroviral therapy era in an urban cohort from Rio de Janeiro, Brazil

Background

Opportunistic illnesses still account for a huge proportion of hospitalizations and deaths among HIV-infected patients in the post combination antiretroviral therapy (cART) era, particularly in middle- and low-income countries. The aim of this study was to assess predictors of the top four most incident opportunistic illnesses (tuberculosis, esophageal candidiasis, cerebral toxoplasmosis and Pneumocystis jiroveci pneumonia) in an HIV clinical cohort from a middle-income country in the post cART era.

Methods

A total of 2835 HIV infected participants aged ≥ 18 years at enrollment were followed from January 2000 to December 2012 until the occurrence of their first opportunistic illness, death or end of study, whichever occurred first. Extended Cox proportional hazards regression models, stratified by use of cART, were fitted to assess predictors of opportunistic illness incidence during follow-up.

Results

The incidence rates of tuberculosis, esophageal candidiasis, cerebral toxoplasmosis and Pneumocystis jiroveci pneumonia were 15.3, 8.6, 6.0, 4.8 per 1000 persons-year, respectively. Disease specific adjusted Cox models showed that presence of an opportunistic illness at enrollment significantly increased disease incidence while higher nadir CD4+ T lymphocyte count had a significant protective effect in patients not in use of cART. Duration of cART use also significantly reduced disease incidence.

Conclusions

Our findings show that, still in the post-cART era, prevention of opportunistic infections can be achieved by preventing immune deterioration by instituting early use of cART. Interventions focusing on early diagnosis and linkage to care in addition to the prompt initiation of cART are essential to reduce the incidence of opportunistic illnesses among HIV infected patients in post-cART era.

HIV and cancer: a comparative retrospective study of Brazilian and U.S. clinical cohorts

BACKGROUND

With successful antiretroviral therapy, non-communicable diseases, including malignancies, are increasingly contributing to morbidity and mortality among HIV-infected persons. The epidemiology of AIDS-defining cancers (ADCs) and non-AIDS-defining cancers (NADCs) in HIV-infected populations in Brazil has not been well described. It is not known if cancer trends in HIV-infected populations in Brazil are similar to those of other countries where antiretroviral therapy is also widely available.

METHODS

We performed a retrospective analysis of clinical cohorts at Instituto Nacional de Infectologia Evandro Chagas (INI) in Rio de Janeiro and Vanderbilt Comprehensive Care Clinic (VCCC) in Nashville from 1998 to 2010. We used Poisson regression and standardized incidence ratios (SIRs) to examine incidence trends. Clinical and demographic predictors of ADCs and NADCs were examined using Cox proportional hazards models.

RESULTS

This study included 2,925 patients at INI and 3,927 patients at VCCC. There were 57 ADCs at INI (65% Kaposi sarcoma), 47 at VCCC (40% Kaposi sarcoma), 45 NADCs at INI, and 82 at VCCC. From 1998 to 2004, incidence of ADCs remained statistically unchanged at both sites. From 2005 to 2010, ADC incidence decreased in both cohorts (INI incidence rate ratio per year = 0.74, p < 0.01; VCCC = 0.75, p < 0.01). Overall Kaposi sarcoma incidence was greater at INI than VCCC (3.0 vs. 1.2 cases per 1,000 person-years, p < 0.01). Incidence of NADCs remained constant throughout the study period (overall INI incidence 3.6 per 1,000 person-years and VCCC incidence 5.3 per 1,000 person-years). Compared to general populations, overall risk of NADCs was increased at both sites (INI SIR = 1.4 [95% CI 1.1-1.9] and VCCC SIR = 1.3 [1.0-1.7]). After non-melanoma skin cancers, the most frequent NADCs were anal cancer at INI (n = 7) and lung cancer at VCCC (n = 11). In multivariate models, risk of ADC was associated with male sex and immunosuppression. Risk of NADC was associated with increased age.

CONCLUSIONS

In both cohorts, ADCs have decreased over time, though incidence of KS was higher at INI than VCCC. Rates of NADCs remained constant over time at both sites.

Factors associated with tobacco smoking and cessation among HIV-infected individuals under care in Rio de Janeiro, Brazil

Worldwide the prevalence of smoking among people living with HIV/AIDS is elevated compared to the general population. This probably reflects the cluster of individual characteristics that have shared risk factors for HIV infection and smoking. A cross-sectional study, enrolling a convenience sample from a Brazilian HIV clinical cohort was conducted to evaluate the prevalence of tobacco smoking and the factors associated with current smoking and abstinence. A total of 2,775 HIV-infected individuals were interviewed: 46.2% have never smoked, 29.9% were current smokers and 23.9% were former smokers. Current smokers had a higher prevalence of alcohol and illicit drug use when compared to the other two groups. A higher proportion of heterosexual individuals were former smokers or never smokers while among men who have sex with men (MSM) a higher proportion were current smokers. Former smokers had been more frequently diagnosed with high blood pressure, diabetes mellitus, cardiovascular diseases and depression, while for current smokers lung diseases were more frequent. Former smokers and current smokers were more likely to have had any hospital admission (42.0% and 41.2%, respectively) than participants who never smoked (33.5%) (p<0.001). Multivariate model results showed that current smokers (versus never smokers) were more likely to be less educated, to report the use of alcohol, crack and cocaine and to present clinical comorbidities. Former smokers (versus current smokers) were more likely to be older, to have smoked for a shorter amount of time and to have smoked >31 cigarettes/day. MSM (compared to heterosexuals) and cocaine users (versus non-users) had lower odds of being former smokers. Considering our results, smoking cessation interventions should be tailored to younger individuals, MSM and substance users.