Advertisement

Mortality predictive factors of people living with human immunodeficiency virus and bloodstream infection

Open AccessPublished:June 19, 2021DOI:https://doi.org/10.1016/j.ijid.2021.06.032

      Highlights

      • Between 2007 and 2016, bloodstream infections (BSI) occurred in 6.8/100 people living with HIV (PLWHIV) admitted to the infectious diseases ward.
      • PLWHIV were newly diagnosed in 25% of BSI admissions, most commonly late presenters.
      • Corticoid use, intravenous drug use, and APACHE II score were predictors of 30-day mortality.
      • The influence of corticotherapy in nosocomial BSI and the short-term prognosis is suggested.
      • Chronic diseases, lymphoma, and a CD4 cell count <200/µl had an impact on 3-year mortality.

      Abstract

      Introduction

      Portugal has one of the highest mortality rates for people living with HIV (PLWHIV) in Europe. After antiretroviral therapy introduction, HIV-associated mortality declined, included the one associated with bloodstream infection (BSI). However it is still high, and European data are scarce . Therefore, characterizing BSI and defining prognostic factors may improve our approach.

      Methods

      This was a 10-year retrospective study of predictive factors for 30-day and 3-year mortality in PLWHIV with BSI in a tertiary infectious diseases ward.

      Results

      Of 2134 PLWHIV admissions, 145 (6.8%) had a BSI, mostly respiratory and catheter-related bacteremia and globally community-acquired. Nosocomial infections occurred in 42 (36%) cases, mostly caused by Enterococcus spp, Staphylococcus aureus, and Candida spp. PLWHIV with a BSI had higher 30-day mortality (27%) compared to those without a BSI (14%). APACHE II score, corticotherapy, and current intravenous drug use (IDU) had a prognostic impact on 30-day mortality. Three-year survival was 54% in PLWHIV with a BSI; a CD4 <200 cells, vascular or chronic pulmonary disease, and lymphoma were prognostic factors.

      Conclusions

      Patients with a BSI were more likely to present advanced HIV disease, have more comorbidities, a longer length of stay, and higher 30-day mortality. IDU and severity of infection determined the short-term prognosis. Three-year mortality was primarily influenced by lower CD4 cell counts, hematological tumor, and cardiopulmonary comorbidities. Systemic corticotherapy may influence nosocomial BSI and short-term prognosis.

      KEYWORDS

      1. Introduction

      According to the World Health Organization (WHO), early HIV diagnosis and the accessibility of combination antiretroviral therapy (cART) has led to an increased overall survival in people living with HIV (PLWHIV) and AIDS. In 2000, according to the National Health Institute Doutor Ricardo Jorge (INSA), 60% of deaths occurred in the first year of diagnosis. Mortality due to HIV has continuously decreased and, nowadays, in Portugal, most deaths occur more than 10 years after the HIV diagnosis (

      DGS – Programa Nacional para a Infeção VIH e SIDA INSA – Departamento de Doenças Infecciosas Portugal. Ministério da Saúde. Direção-Geral da Saúde/Instituto Nacional de Saúde Doutor Ricardo Jorge. Infeção VIH e SIDA em Portugal - 2019. Lisboa: DGS/INSA; 2019. Direção-Geral da Saúde (DGS) e Instituto Nacional de Saúde Doutor Ricardo Jorge (INSA, IP): 2019 Retrieved July 11 from https://www.sns.gov.pt/wp-content/uploads/2019/11/Relat-VIH-SIDA-2019.pdf.

      ). In-hospital mortality dropped by 14%, between 2000 and 2010 (
      • Catumbela E.
      • Freitas A.
      • Lopes F.
      • Mendoza Mdel C.
      • Costa C.
      • Sarmento A.
      • et al.
      HIV disease burden, cost, and length of stay in Portuguese hospitals from 2000 to 2010: a cross-sectional study.
      ).
      The mortality rate of bloodstream infection (BSI) has also reduced with antiretroviral therapy (ART), yet it still remains high both in the short-term (
      • Afessa B.
      • Morales I.
      • Weaver B.
      Bacteremia in hospitalized patients with human immunodeficiency virus: A prospective, cohort study.
      ;
      • Ortega M.
      • Almela M.
      • Soriano A.
      • Marco F.
      • Martínez J.A.
      • Muñoz A.
      • et al.
      Bloodstream infections among human immunodeficiency virus-infected adult patients: epidemiology and risk factors for mortality.
      ;
      • Declercq S.
      • De Munter P.
      • Derdelinckx I.
      • Verhaegen J.
      • Peetermans Willy E.
      • Vanderschueren S.
      • et al.
      Characteristics, causes, and outcome of 54 episodes of bloodstream infections in a cohort of HIV patients.
      ) and long-term, with a 3-year mortality rate of 59% (
      • Declercq S.
      • De Munter P.
      • Derdelinckx I.
      • Verhaegen J.
      • Peetermans Willy E.
      • Vanderschueren S.
      • et al.
      Characteristics, causes, and outcome of 54 episodes of bloodstream infections in a cohort of HIV patients.
      ). BSI in PLWHIV has a wide spectrum of possible etiologies, heavily influenced by the availability of cART and the geographic area. Most epidemiological reports on this matter have been from developing countries of Africa and Asia, which have disclosed mycobacterial and fungal infections (
      • Taramasso L.
      • Tatarelli P.
      • Di Biagio A.
      Bloodstream infections in HIV-infected patients.
      ) rather than bacterial infections, in contrast to European reports (
      • Ortega M.
      • Almela M.
      • Soriano A.
      • Marco F.
      • Martínez J.A.
      • Muñoz A.
      • et al.
      Bloodstream infections among human immunodeficiency virus-infected adult patients: epidemiology and risk factors for mortality.
      ;
      • Rodríguez-Créixems M.
      • Alcalá L.
      • Muñoz P.
      • Cercenado E.
      • Vicente T.
      • Bouza E.
      Bloodstream infections: evolution and trends in the microbiology workload, incidence, and etiology, 1985-2006.
      ;
      • Declercq S.
      • De Munter P.
      • Derdelinckx I.
      • Verhaegen J.
      • Peetermans Willy E.
      • Vanderschueren S.
      • et al.
      Characteristics, causes, and outcome of 54 episodes of bloodstream infections in a cohort of HIV patients.
      ;
      • Taramasso L.
      • Liggieri F.
      • Cenderello G.
      • Bovis F.
      • Giannini B.
      • Mesini A.
      • et al.
      Bloodstream infections in patients living with HIV in the modern cART era.
      ). Commonly, BSI is community-acquired, but up to 38% are hospital-associated (
      • Taramasso L.
      • Liggieri F.
      • Cenderello G.
      • Bovis F.
      • Giannini B.
      • Mesini A.
      • et al.
      Bloodstream infections in patients living with HIV in the modern cART era.
      ), mainly caused by Staphylococcus spp and Candida spp (
      • Ortega M.
      • Almela M.
      • Soriano A.
      • Marco F.
      • Martínez J.A.
      • Muñoz A.
      • et al.
      Bloodstream infections among human immunodeficiency virus-infected adult patients: epidemiology and risk factors for mortality.
      ). In-hospital and long-term mortality factors have been published, with these being heterogeneous: older age, AIDS, malignancy, current intravenous drug use (IDU), end-stage liver disease, corticotherapy, neutropenia, candidemia, APACHE II score, and septic shock (
      • Afessa B.
      • Morales I.
      • Weaver B.
      Bacteremia in hospitalized patients with human immunodeficiency virus: A prospective, cohort study.
      ;
      • Petrosillo N.
      • Viale P.
      • Nicastri E.
      • Arici C.
      • Bombana E.
      • Casella A.
      • et al.
      Nosocomial bloodstream infections among human immunodeficiency virus-infected patients: incidence and risk factors.
      ;
      • Ortega M.
      • Almela M.
      • Soriano A.
      • Marco F.
      • Martínez J.A.
      • Muñoz A.
      • et al.
      Bloodstream infections among human immunodeficiency virus-infected adult patients: epidemiology and risk factors for mortality.
      ;
      • Taramasso L.
      • Liggieri F.
      • Cenderello G.
      • Bovis F.
      • Giannini B.
      • Mesini A.
      • et al.
      Bloodstream infections in patients living with HIV in the modern cART era.
      ).
      Given the lack of Portuguese data on an issue that contributes to additional mortality, the authors assessed the incidence, demographic, immunological, and microbiological characteristics of PLWHIV with BSI and studied risk factors for 30-day and long-term mortality in a Portuguese tertiary center. This appears to be the first such report on this issue in Portugal.

      2. Methods

      2.1 Study design and population

      This retrospective observational cohort study included International Classification of Diseases ninth revision (ICD-9) codified HIV-infected adults with positive blood cultures (bacteria, fungi, and mycobacteria) admitted to the infectious diseases ward of a university hospital between January 2007 and December 2016. The main aim was to analyze factors predictive of 30-day and 3-year mortality for this BSI cohort. The secondary aim was to compare 30-day mortality with that of 376 PLWHIV admitted during the same time interval for bacterial or fungal infections but with negative blood cultures and no previous BSI. Sex, age, Centers for Disease Control and Prevention (CDC) stage, CD4 cell count, length of stay (LOS), malignant disease, palliative care, septic shock, neutropenia (<1500 cells/μl), and intensive care unit (ICU) admission were recorded and compared between the two groups. Additionally, the incidence of BSI was assessed and the community-acquired and nosocomial infections were characterized microbiologically. Organisms considered as blood contaminants were excluded. Three-year follow-up data were recorded by assessing clinical appointments, CD4 cell counts, and death records. Patients without a CD4 cell count or who lost follow-upwere excluded. Details of the population setting and study design are given in Figure 1.
      Figure 1
      Figure 1Schematic diagram of the study methodology: population setting and study design.
      1Patients with a single diagnosis of cutaneous herpes, cutaneous or mucous fungal infection, hepatitis, chronic sinusitis, syphilis, or molluscum contagiosum, and those without antibiotic therapy were excluded. BSI, bloodstream infection; ICD-9, International Classification of Diseases ninth revision.

      2.2 Data collection and definitions

      Demographic data, comorbidities, and drug consumption were obtained: sex, age, race, diabetes, chronic kidney disease (estimated glomerular filtration rate ≤60 ml/kg/min2 and dialysis), hepatitis B and C virus co-infections, history of malignancy (early or advanced stage according to Ann Arbor or ACTG classification system), chemotherapy, end-stage liver disease (image, biopsy, or transient elastography-compatible), and previous or current tobacco, alcohol, or intravenous drug use (IDU). Functional impairment was defined as a bedridden patient or a patient who needed care for daily living activities. Viral and immunological data included late presenters according to the
      • Antinori A.
      • Coenen T.
      • Costagiola D.
      • et al.
      Late presentation of HIV infection: a consensus definition.
      consensus: CD4 count <350 cells/μl or presenting with an AIDS-defining event, regardless of the CD4 cell count. HIV viral load, lymphocyte count, CD4 cell count 120 days before until 30 days after hospital admission, ART on admission and azithromycin and/or trimethoprim–sulfamethoxazole prophylaxis (taking into account non-compliance according to the medical records), neutrophil count <1500 cells/μl, and acute or chronic systemic corticotherapy (≥10 mg prednisolone daily dose equivalent for at least 5 days during hospitalization) were also recorded. The microorganism, source of infection, and European Committee on Antimicrobial Susceptibility Testing (EUCAST) profile were assessed. A nosocomial infection was defined as one acquired at least 48 hours after admission, and a catheter-related BSI (CRBSI) was one in which the same organism grew in one blood culture and in a culture of the catheter tip, or when two samples (from the catheter and from a peripheral vein) were positive. Organ dysfunction was assessed according to the APACHE II scoring system and Sepsis-3 consensus (
      • Singer M.
      • Deutschman C.S.
      • Seymour C.W.
      • et al.
      The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3).
      ). The CD4 cell count reached at the 3-year follow-up was obtained and registered according to three levels: complete recovery of the CD4 count or maintaining a count ≥500 cells/µl, CD4 count 200–499 cells/µl, and poor or no CD4 recovery (<200 cells/µl).

      2.3 Statistical analysis

      Data were analyzed using IBM SPSS Statistics version 22.0 (IBM Corp., Armonk, NY, USA). Continuous variables were expressed as the mean and standard deviation (±SD) or median and interquartile range (IQR) according to the Kolmogorov–Smirnov test with Lilliefors significance correction. Categorical variables were compared using the Chi-square test or Fisher's exact test. Quantitative variables were compared using the Student–Fisher t-test or analysis of variance (ANOVA). Non-parametric tests were used when the application conditions were not applicable. Linear and logistic regression was exhibited after verified independence, linearity, absence of multicolinearity and residual standard deviation inferior to 3.. Kaplan–Meier and Cox regression analyses were used for the survival analysis. Statistical significance was defined as a two-tailed P-value <0.05. Missing data were imputed using regression methods.

      3. Results

      3.1 Incidence and sample characterization

      From a total of 2134 PLWHIV admissions to the infectious diseases ward over the 10-year period, 145 (6.8%) admissions had at least one BSI, corresponding to 123 distinct patients and a total of 171 microorganisms. The incidence was similar over the study period. The characteristics of the study patients are presented in Table 1. About two thirds (78.9%) were male, with the median age of 44.5 years (IQR 13) (range 19–84 years) and white European. Overall, 22 (17.9%) were current IDU and 61 (52.1%) reported tobacco and/or alcohol abuse. Regarding comorbidities, 15 (12.2%) individuals had diabetes, 22 (17.6%) had a vascular or chronic pulmonary disease, 48 (39.0%) had previously diagnosed with hepatitis B or C, 12 (9.8%) had liver cirrhosis, and 29 (23.6%) had neoplasia. Of the patients with neoplasia, 22 (75.9%) had an AIDS-defining tumor (lymphoma or Kaposi's sarcoma) and theremaining had lung, pharynx, bone, pancreas, and gynecological tumors. In 107 (87.0%) admissions, the patient was in CDC stage 3, and 87 (70.7%) patients had an AIDS-defining disease: tuberculosis (28.5%), Cryptococcus meningitis (17.9%), and Pneumocystis jirovecii pneumonia (14.6%) were the most common. Only three (2.4%) patients were in CDC stage 1.
      Table 1Characterization of 123 PLWHIV with a first BSI episode
      Patient characterizationNumber (%)
      Demographic characteristics
      Sex Male FemaleRace White/white European Age (years) Median (IQR); range age group ≥60 years97 (78.9%)26 (21.1%)107 (87.0%)44.5 (13); 19–8417 (13.8%)
      Comorbidities
      Total missing rate: 1.6% to 6.5%.
      Current intravenous drug useTobacco or alcohol use
      Previous or current use.
      Hepatitis C and/or B contactChronic kidney disease (≤60 ml/min/m2)DiabetesEnd-stage liver diseaseVascular or pulmonary chronic disease
      Previous documented stroke, myocardial infarction, chronic pulmonary obstructive disease, or peripheral vascular disease.
      Functional impairmentHistory of malignancy AIDS-defining tumor Non-AIDS-defining tumor Hematological cancer Advanced stageAll AIDS-defining diseases (≥1)
      22/123 (18%)61/117 (52%)48/119 (39%)12/121 (9.8%)15/123 (12%)12/121 (9.8%)21/119 (17.6%)18/121 (14.6%)29/123 (23.6%)22/29 (75.9%)7/29 (24.1%)13/29 (10.6%)23/29 (79.3%)87/123 (70.7%)
      CDC stage and CDC stage
      All AIDS patientsCD4 count, median (IQR); rangeCD4 %, median (IQR); rangeCD4 nadir, median (IQR)107 (87.0%)96 (223); 2–79611 (16); 1–4260 (155)
      Viral load, new versus known HIV
      Undetectable (<200 copies/ml)<5 log≥5 log0 (0%) vs 31 (27.0%)7 (6.1%) vs 28 (24.3%)21 (18.3%) vs 28 (24.3%)
      ART taking at admission
      Patients on ART: excluding 31 new HIV patients. NRTI, nucleoside reverse transcriptase inhibitors; NNRTI, non-nucleoside reverse transcriptase inhibitors; PI, protease inhibitors; II, integrase inhibitors. ART missing data: 3.3%.
      , 40/89 (44.9%)
      NRTI ± NNRTINRTI ± PINRTI ± IIOther cART regimen7 (17.5%)25 (62.5%)5 (12.5%)3 (7.5%)
      Other medication
      On antibiotic prophylaxisLong-term corticoid therapySystemic corticotherapy during hospital stay19/89 (21.3%)11/123 (8.9%)49/123 (39.8%)
      PLWHIV, people living with HIV; BSI, bloodstream infection; IQR, interquartile range; CDC, US Centers for Disease Control and Prevention; ART, antiretroviral therapy; cART, combination antiretroviral therapy.
      a Total missing rate: 1.6% to 6.5%.
      b Previous or current use.
      c Previous documented stroke, myocardial infarction, chronic pulmonary obstructive disease, or peripheral vascular disease.
      d Patients on ART: excluding 31 new HIV patients. NRTI, nucleoside reverse transcriptase inhibitors; NNRTI, non-nucleoside reverse transcriptase inhibitors; PI, protease inhibitors; II, integrase inhibitors. ART missing data: 3.3%.
      The HIV diagnosis was made from the year 1993 to 2014: 92 patients (74.8%) had a previously known diagnosis (one third known for over 10 years) and 31 (25.2%) had a new diagnosis at the time of BSI. The latter were more frequently late presenters (93.5% vs 64.8%, P = 0.002) and had a higher viral load (P < 0.001).

      3.2 BSI microorganisms and sources of infection

      Blood cultures were requested in 1240 (58%) of all admissions. Of these episodes, 214 (17.3%) had ≥1 positive blood culture for bacteria, mycobacteria, or fungus. Positive cultures for coagulase-negative staphylococci (CoNS), Bacillus spp, or fungus in a single blood culture that were not treated, were excluded (19%). All 171 microorganisms from 2007 to 2016 were considered in this step. Admission was directly related to the BSI in 70 (57%) episodes. For the remaining episodes, the most frequent causes of admission were malignancy-related (n = 15; 12.2%) and P. jirovecii pneumonia (n = 7; 5.7%). Bacterial (non-mycobacterial) BSI were the most common, occurring in 130 (76%) episodes (Figure 2). Of these, 72 (55.4%) were gram-positive, more specifically 23 (17.6%) Staphylococcus aureus, 18 (13.8%) CoNS, 15 (11.5%) Streptococcus spp, 12 (9.2%) Enterococcus spp, and four others (≤3%). The most frequent gram-negative bacteria were Escherichia coli (n = 17; 13.1%), Klebsiella spp (n = 13; 10%), Pseudomonas spp (n = 7; 5.4%), and non-typhi Salmonella (n = 6; 4.6%). A respiratory tract infection was observed in 34 (26.2%) individuals, catheter-related infection (including one peripheral line) in 23 (17.7%), skin and subcutaneous infection in 14 (10.8%), urinary infection in 13 (10%), gastrointestinal or biliary infection in 13 (10%), and endocarditis with uncommon central nervous system or joint infections in eight (6.2%). Twenty-two (16.9%) cases were actually codified as primary BSI.
      Figure 2
      Figure 2Sources of the infections and MDR rate for bacterial (non-mycobacterial) BSI out of the total episodes. Uncommon sources (≤3% of cases) are not represented: Acinetobacter spp, Stenotrophomonas spp, Enterobacter spp, Povidencia stuartii, C. striatum, Listeria spp, Proteus mirabilis, R. mucilaginosa, and Serratia spp. Uncommon infections (one in a joint and two in the central nervous system) are not represented. Fungi and mycobacteria are not represented. GE, gastrointestinal; CA, community-acquired; MDR, multidrug resistance.
      PLWHIV who were current IDU had more cutaneous infections and endocarditis (P = 0.016).
      Fungemia occurred in 29 (17%) cases, 19 (67%) due to Cryptococcus neoformans, in patients with lower CD4 cell counts (median 23, IQR 49), lower CD4 cell count nadir (median 22, IQR 45), and higher viral loads when compared to those with bacteremia (P < 0.005). Mycobacterial BSI occurred in 12 (7%) admissions, with no difference in prevalence according to race, and with similar immunological and viral load profiles as in Cryptococcus BSI. The remaining 10 fungi were Candida albicans, Candida krusei, and Candida parapsilosis, of which eight (80%) were nosocomially acquired.
      Central lines, including dialysis catheters, were inserted in 52 (35.8%) episodes, with associated infection in 15 (28.8%) of them and a propensity for candidemia or primary BSI (P = 0.01).
      Polymicrobial cultures occurred in five (4%) admissions and multiple microorganisms were isolated during the same admission in 16 (11%) cases.

      3.3 Community versus nosocomial infection

      Regarding 116 positive bacterial or Candida cultures, 74 (63.8%) were community-acquired (CA), 36 (48.6%) of them with healthcare-associated factors: nursing care residence/healthcare facilities, hemodialysis, or previous hospitalization within 90 days, but with the same community multidrug resistance (MDR) rate (33% vs 30%). Overall CA-MDR infections occurred in 39 (31.7%) patients.
      Nosocomial BSI (NBSI) accounted for 42 (36.2%) of all bacteremia/candidemia episodes, with a rate of 2.34 NBSI/100 admissions, which remained stable during the study period; these presented at a median of 29 days (IQR 35 days) of hospitalization. Respiratory tract infections, catheter infections, and primary BSI were the most frequent infections, and Enterococcus spp and Candida spp were the main microorganisms. A step-wise logistic regression analysis was performed selecting the probability of having bacteria/Candida NBSI: central line, advanced tumor, and systemic corticotherapy appeared to be predictors of NBSI (Table 2). Additionally, NBSI patients had more MDR (69%) (P = 0.009), longer LOS (42 days (IQR 54 days) vs 16 days (IQR 30 days)), and greater 30-day mortality (odds ratio (OR) 2.63, 95% confidence interval (CI) 1.15–6.00, P = 0.02).
      Table 2Bivariate and multivariate analysis of predictors of nosocomial BSI in 116 distinct bacteremia and candidemia episodes
      VariableBivariate OR (95% CI)P-valueMultivariate OR (95% CI)P-value
      Male sex2.70 (1.00–7.32)0.045*0.48 (0.18–1.26)0.136
      AIDS stage12.23 (1.56–95.59)0.003*Not applicable
      Not applicable: assumptions criteria were not present. The residual standard deviation was ≥3.0 when AIDS or functional impairment was considered in the model, although it seemed to have a role as a nosocomial predictor in the bivariate analysis.
      -
      Functional impairment3.16 (1.19–8.37)0.02*Not applicable
      Not applicable: assumptions criteria were not present. The residual standard deviation was ≥3.0 when AIDS or functional impairment was considered in the model, although it seemed to have a role as a nosocomial predictor in the bivariate analysis.
      -
      Lymphoma4.53 (1.56–13.22)0.003*1 (0.30–3.39)0.995
      Advanced tumor2.03 (1.29–3.18)0.004*3.54 (1.29–9.71)0.01*
      Central line
      Central line: Chi-square (1) = 29.847; P < 0.001; Nagelkerke R2 = 0.311. Area under the curve (AUC) = 0.772.
      5.79 (2.23–15.04)<0.001*4.85 (2.00–11.76)<0.001*
      Candidemia7.02 (1.44–34.18)0.016*Not applicable
      Not applicable: assumptions criteria were not present. The residual standard deviation was ≥3.0 when AIDS or functional impairment was considered in the model, although it seemed to have a role as a nosocomial predictor in the bivariate analysis.
      -
      Corticotherapy3.70 (1.72–9.01)0.001*3.48 (1.40–8.68)0.007*
      BSI, blood stream infection; OR, odds ratio; CI, confidence interval. *P < 0.05.
      a Central line: Chi-square (1) = 29.847; P < 0.001; Nagelkerke R2 = 0.311. Area under the curve (AUC) = 0.772.
      b Not applicable: assumptions criteria were not present. The residual standard deviation was ≥3.0 when AIDS or functional impairment was considered in the model, although it seemed to have a role as a nosocomial predictor in the bivariate analysis.

      3.4 Prognostic factors for 30-day mortality

      Only the first hospitalization during the study period was included in the analysis of 30-day mortality. Thirty-four patients (27.6%) died, 16 (45.7%) due to an AIDS-defining event.
      The Kaplan–Meier curve comparing 30-day mortality in PLWHIV with a BSI (group 1, G1) and PLWHIV without a BSI (group 2, G2) is shown in Figure 3. These patients had similar sex and age distributions and similar new diagnosis and neutropenia rates (<1500 cells/μl), but PLWHIV with a BSI had a worse immunological state, more comorbidities, longer LOS, and more septic shock than those without a BSI, resulting in up to 22% higher mortality at day 30.
      Figure 3
      Figure 3Kaplan–Meier curve of cumulative survival for the two groups of PLWHIV: with and without BSI. (BSI, bloodstream infection; PLWHIV, people living with HIV; IQR, interquartile range; CDC, Centers for Disease Control and Prevention; LOS, length of stay; ICU, intensive care unit.)
      Regarding the total population of PLWHIV (G1 and G2), age ≥50 years (mean age at death 52 ± 14 years), AIDS stage (P = 0.001), corticotherapy (P < 0.001), and <1500 neutrophils/μl (P = 0.009) had prognostic value regarding 30-day mortality, the latter being associated with early mortality.
      Thirty-day mortality was assessed (Table 3) for G1 patients. Patients who died had higher APACHE II scores (mean 11 ± 4, P < 0.001) and SOFA scores (median 2 (IQR 3), P = 0.001), with no difference in severity scores according to the microorganisms involved. Septic patients had higher APACHE II scores (mean 17 ± 5 points) and SOFA scores (mean 5 ± 3 points). Time-dependent Cox regression was applied after non-proportional risk suspicion for corticotherapy at day 2 (not confirmed), showing that current IDU (hazard ratio (HR) 5.60), corticotherapy use (HR 3.74), and higher APACHE II score (HR 1.12) were predictors of mortality, with a lower likelihood ratio for the model containing IDU, in contrast to the one that included cutaneous infections.
      Table 3Cox regression analysis for predictors of 30-day mortality after first BSI (123 patients), from 2007 to 2016
      Variable
      Variables without predictive value (P ≥ 0.05): age, sex, race, year of BSI, new diagnosis, late presenter, CD4 cell count nadir, AIDS CDC stage, number of AIDS-defining diseases, chemotherapy, malignant disease, CD4 cells, CD8 cells, lymphocyte count, CD4%, CD4/8 ratio, viral load, alcohol or tobacco use, diabetes, advanced liver disease, hepatitis B or C contact, chronic kidney disease, on cART, antibiotic prophylaxis, home care facility residence, previous 90-day hospital admission, ICU admission.
      Unadjusted HR (95% CI)P-valueβAdjusted HR (95% CI)P-value
      age group ≥50 years
      Dichotomized as per the ROC curve. Proportional risk was not verified for another infection during hospitalization or neutropenia during the time considered
      2.60 (1.35–5.00)0.004*-1.91 (0.82–4.48)0.134
      APACHE II value1.21 (1.10–1.33)<0.001*0.1031.12 (1.03–1.21)0.008*
      SOFA score value1.20 (1.10–1.31)<0.001*-1.05 (0.94–1.16)0.401
      Septic shock3.91 (2.00–7.62)<0.001*-1.33 (0.56–3.15)0.518
      Corticotherapy (acute/chronic)4.18 (2.05–8.52)<0.001*1.3193.74 (1.65–8.48)0.002*
      Current IDU1.58 (0.71–3.49)0.261.7225.60 (2.03–15.40)0.001*
      Cutaneous infection2.28 (0.997–5.20)0.073---
      BSI, bloodstream infection; HR, hazard ratio; CI, confidence interval; APACHE II, Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment; IDU, intravenous drug use; CDC, Centers for Disease Control and Prevention; cART, combination antiretroviral therapy; ICU, intensive care unit; ROC, receiver operating characteristics. *P < 0.05.
      a Variables without predictive value (P ≥ 0.05): age, sex, race, year of BSI, new diagnosis, late presenter, CD4 cell count nadir, AIDS CDC stage, number of AIDS-defining diseases, chemotherapy, malignant disease, CD4 cells, CD8 cells, lymphocyte count, CD4%, CD4/8 ratio, viral load, alcohol or tobacco use, diabetes, advanced liver disease, hepatitis B or C contact, chronic kidney disease, on cART, antibiotic prophylaxis, home care facility residence, previous 90-day hospital admission, ICU admission.
      b Dichotomized as per the ROC curve.Proportional risk was not verified for another infection during hospitalization or neutropenia during the time considered

      3.5 Predictive factors for 3-year mortality

      Additional data from 85 survivors at ≥30 days after discharge were recorded over a median of 2 years and 3 months (IQR 5 years 8 months): (1) subsequent BSI admissions, (2) CD4 cell counts, and (3) the 3-year mortality rate, excluding four patients with limited data.
      Overall, 16 (13%) patients had at least another admission (up to four admissions) because of BSI, with a median interval of 6.5 months (IQR 11 months). The CD4 cell count after 3 years was ≥500 cells/µl in 26 patients (30.9%), 200–499 cells/µl in 31 (37.2%), and <200 cells/µl in 27 (31.9%).
      During follow-up, 57 (46.3%) patients died. Mortality predictors were analyzed (Table 4): a CD4 cell count ≥200 cells/µl at the 3-year follow-up had a prognostic impact, with HR 13.9 at 3 years and median survival 3.6 times higher (Figure 4); there was no difference (P = 0.282) between patients with CD4 200–499 cells/µl and CD4 ≥500 cells/µl. Vascular or chronic pulmonary and lymphoproliferative diseases had an impact on long-term mortality.
      Table 4Cox regression analysis for predictors of 3-year follow-up mortality after first BSI for 85 PLWHIV
      VariableUnadjusted HR (95% CI)P-valueβAdjusted
      AUC = 0.80.
      HR (95% CI)
      P-value
      age group ≥65 years3.59 (1.06–12.19)0.04*-0.55 (0.83–3.62)0.53
      Long-term corticoid8.78 (2.37–32.47)0.001*-3.33 (0.44–25.41)0.25
      Chemotherapy4.16 (1.69–10.20)0.002*-7.50 (0.99–54.49)0.051*
      Hematological tumorAdvanced tumor
      Advanced solid or hematological tumor.
      9.21 (3.80–22.30)5.64 (2.41–13.21)<0.001*<0.001*3.083-21.82 (2.08–228)1.59 (0.12–20.86)0.016*0.73
      Functional impairment2.79 (1.10–7.11)0.03*-1.38 (0.38–4.97)0.62
      Chronic vascular or pulmonary disease
      Previous documented stroke, myocardial infarction, chronic pulmonary obstructive disease, or peripheral vascular disease.
      2.86 (1.16–7.02)0.02*1.8586.41 (1.74–26.65)0.015*
      CD4 cells ≥200, at 3 years2.97 (1.25–7.06)0.01*1.9456.99 (1.97–24.82)0.013*
      BSI, bloodstream infection; PLWHIV, people living with HIV; HR, hazard ratio; CI, confidence interval; AUC, area under the curve. *P < 0.05.
      a AUC = 0.80.
      b Advanced solid or hematological tumor.
      c Previous documented stroke, myocardial infarction, chronic pulmonary obstructive disease, or peripheral vascular disease.
      Figure 4
      Figure 4CD4 count on follow-up was an independent predictive factor for mortality.

      4. Discussion

      This study focused on factors predictive of 30-day and 3-year mortality in PLWHIV and BSI over a 10-year period among a tertiary hospital cohort. Additionally, BSI incidence and the microorganisms in CA-BSI and NBSI were evaluated. From a total of 2134 admissions, 68/1000 had a BSI, similar to other European reports (
      • Ortega M.
      • Almela M.
      • Soriano A.
      • Marco F.
      • Martínez J.A.
      • Muñoz A.
      • et al.
      Bloodstream infections among human immunodeficiency virus-infected adult patients: epidemiology and risk factors for mortality.
      ;
      • Taramasso L.
      • Liggieri F.
      • Cenderello G.
      • Bovis F.
      • Giannini B.
      • Mesini A.
      • et al.
      Bloodstream infections in patients living with HIV in the modern cART era.
      ). However, the incidence was higher than that in a previous Italian report (
      • Bonadio M.
      • Gigli C.
      • Maccanti O.
      • Longo B.
      • Smorfa A.
      Bloodstream infections in HIV-positive patients: a review of sixty-eight episodes.
      ), which did not include Cryptococcus and mycobacteria (58/1000 per year), and in a report from Spain, with 37/1000 per year (
      • Pedro-Botet M.L.
      • Mòdol J.M.
      • Vallés X.
      • Romeu J.
      • Sopena N.
      • Giménez M.
      • et al.
      Changes in bloodstream infections in HIV-positive patients in a university hospital in Spain (1995-1997).
      ), where the incidence of BSI in PLWHIV was continuously decreasing from 1995 to 2006 (
      • Rodríguez-Créixems M.
      • Alcalá L.
      • Muñoz P.
      • Cercenado E.
      • Vicente T.
      • Bouza E.
      Bloodstream infections: evolution and trends in the microbiology workload, incidence, and etiology, 1985-2006.
      ). Nevertheless, we cannot draw conclusions about previous changes in Portugal, as many publications were restricted to short periods of time and particular geographical areas. From 2007 to 2016, the incidence of BSI was the same in the study cohort. The predominant heterosexual transmission and the sex and age distribution were in accordance with a 2019 INSA report for the HIV population.
      Predictably, the spectrum of BSI was heavily influenced by the geographical area. European data are in agreement regarding the predominance of bacteremia (63–85%). However, in developing countries, bacterial BSI has been reported to occur in only 16% (
      • Taramasso L.
      • Tatarelli P.
      • Di Biagio A.
      Bloodstream infections in HIV-infected patients.
      ), and lower CD4 cell counts and high HIV viral loads have been associated with fungal and mycobacterial infections (
      • Kiertiburanakul S.
      • Watcharatipagorn S.
      • Chongtrakool P.
      • Santanirand P.
      Epidemiology of bloodstream infections and predictive factors of mortality among HIV-infected adult patients in Thailand in the era of highly active antiretroviral therapy.
      ).
      Nosocomial infections occur worldwide, affecting both resource-rich and resource-poor countries, with a significant burden for the patient and for public health. Such infections have been estimated to occur in 6.5/100 hospital admissions in the European Union and European Economic Area, with 0.7/100 of these being BSI (
      • Suetens C.
      • Latour K.
      • Kärki T.
      • Ricchizzi E.
      • Kinross P.
      • Moro M.L.
      • et al.
      The Healthcare-Associated Infections Prevalence Study Group. Prevalence of healthcare-associated infections, estimated incidence and composite antimicrobial resistance index in acute care hospitals and long-term care facilities: results from two European point prevalence surveys, 2016 to 2017.
      ). However, the incidence of NBSI appears to be higher in PLWHIV. A multicenter study performed in the pre-ART era (1989–1995) revealed an NBSI incidence of 8/100 (
      • Stroud L.
      • Srivastava P.
      • Culver D.
      • Bisno A.
      • Rimland D.
      • Simberkoff M.
      • et al.
      Nosocomial Infections in HIV-Infected Patients Preliminary Results from a Multicenter Surveillance System (1989-1995).
      ). Data after ART introduction showed a decreased incidence over the years: 4.7/100 in Italy in 1998 (
      • Petrosillo N.
      • Viale P.
      • Nicastri E.
      • Arici C.
      • Bombana E.
      • Casella A.
      • et al.
      Nosocomial bloodstream infections among human immunodeficiency virus-infected patients: incidence and risk factors.
      ) and 2.5/100 in Brazil in 2003-2004 (
      • Panis C.
      • Matsuo T.
      • Reiche E.M.V.
      Nosocomial infections in human immunodeficiency virus type 1 (HIV-1) infected and AIDS patients: major microorganisms and immunological profile.
      ). More recent rates (2008–2015) are 1.6/100 in Italy (
      • Taramasso L.
      • Liggieri F.
      • Cenderello G.
      • Bovis F.
      • Giannini B.
      • Mesini A.
      • et al.
      Bloodstream infections in patients living with HIV in the modern cART era.
      ) and 2.3/100 in the present study cohort. Nevertheless, the distinct geographic areas and lack of large pre-ART and after cART series could create a bias. Not only was NBSI associated with higher mortality in several pre-ART and after ART publications (
      • Tumbarello M.
      • Tacconelli E.
      • Donati K.G.
      • Leone F.
      • Morace G.
      • Cauda R.
      • et al.
      Nosocomial bloodstream infections in HIV-infected patients: attributable mortality and extension of hospital stay.
      ;
      • Petrosillo N.
      • Pagani L.
      • Ippolito G.
      Gruppo HIV e Infezioni Ospedaliere. Nosocomial infections in HIV-positive patients: an overview.
      ), but also with LOS (
      • Tumbarello M.
      • Tacconelli E.
      • Donati K.G.
      • Leone F.
      • Morace G.
      • Cauda R.
      • et al.
      Nosocomial bloodstream infections in HIV-infected patients: attributable mortality and extension of hospital stay.
      ;
      • Petrosillo N.
      • Viale P.
      • Nicastri E.
      • Arici C.
      • Bombana E.
      • Casella A.
      • et al.
      Nosocomial bloodstream infections among human immunodeficiency virus-infected patients: incidence and risk factors.
      ) and the multidrug resistance rate (
      • Franceschini E.
      • Santoro A.
      • Menozzi M.
      • Bacca E.
      • Venturelli C.
      • Zona S.
      • Bedini A.
      • Digaetano M.
      • Puzzolante C.
      • Meschiari M.
      • Cuomo G.
      • Orlando G.
      • Sarti M.
      • Guaraldi G.
      • Cozzi-Lepri A.
      • Mussini C.
      Epidemiology and Outcomes of Bloodstream Infections in HIV-Patients during a 13-Year Period.
      ). In the present study, the MDR rate in NBSI was 69% (compared to 31.7% for CA-BSI), and this was mostly due to methicillin-resistant S. aureus, extended-spectrum β-lactamase-producing bacteria, and Enterococcus spp. Healthcare assistance leads to invasive procedures and even in the pre-ART era, CRBSI was one of the most implicated factors (
      • Stroud L.
      • Srivastava P.
      • Culver D.
      • Bisno A.
      • Rimland D.
      • Simberkoff M.
      • et al.
      Nosocomial Infections in HIV-Infected Patients Preliminary Results from a Multicenter Surveillance System (1989-1995).
      ;
      • Franceschini E.
      • Santoro A.
      • Menozzi M.
      • Bacca E.
      • Venturelli C.
      • Zona S.
      • Bedini A.
      • Digaetano M.
      • Puzzolante C.
      • Meschiari M.
      • Cuomo G.
      • Orlando G.
      • Sarti M.
      • Guaraldi G.
      • Cozzi-Lepri A.
      • Mussini C.
      Epidemiology and Outcomes of Bloodstream Infections in HIV-Patients during a 13-Year Period.
      ). Gram-positive bacteria (Enterococcus spp, Staphylococcus spp) caused most of those infections (
      • Declercq S.
      • De Munter P.
      • Derdelinckx I.
      • Verhaegen J.
      • Peetermans Willy E.
      • Vanderschueren S.
      • et al.
      Characteristics, causes, and outcome of 54 episodes of bloodstream infections in a cohort of HIV patients.
      ), although Candida species were also implicated (
      • Declercq S.
      • De Munter P.
      • Derdelinckx I.
      • Verhaegen J.
      • Peetermans Willy E.
      • Vanderschueren S.
      • et al.
      Characteristics, causes, and outcome of 54 episodes of bloodstream infections in a cohort of HIV patients.
      ;
      • Taramasso L.
      • Liggieri F.
      • Cenderello G.
      • Bovis F.
      • Giannini B.
      • Mesini A.
      • et al.
      Bloodstream infections in patients living with HIV in the modern cART era.
      ).
      CA-BSI remained the most prevalent, at 62–82% (
      • Afessa B.
      • Morales I.
      • Weaver B.
      Bacteremia in hospitalized patients with human immunodeficiency virus: A prospective, cohort study.
      ;
      • Ortega M.
      • Almela M.
      • Soriano A.
      • Marco F.
      • Martínez J.A.
      • Muñoz A.
      • et al.
      Bloodstream infections among human immunodeficiency virus-infected adult patients: epidemiology and risk factors for mortality.
      ;
      • Taramasso L.
      • Liggieri F.
      • Cenderello G.
      • Bovis F.
      • Giannini B.
      • Mesini A.
      • et al.
      Bloodstream infections in patients living with HIV in the modern cART era.
      ), with Streptococcus and Escherichia coli as the main agents and respiratory and urinary tract infections as the main sources (
      • Ortega M.
      • Almela M.
      • Soriano A.
      • Marco F.
      • Martínez J.A.
      • Muñoz A.
      • et al.
      Bloodstream infections among human immunodeficiency virus-infected adult patients: epidemiology and risk factors for mortality.
      ;
      • Declercq S.
      • De Munter P.
      • Derdelinckx I.
      • Verhaegen J.
      • Peetermans Willy E.
      • Vanderschueren S.
      • et al.
      Characteristics, causes, and outcome of 54 episodes of bloodstream infections in a cohort of HIV patients.
      ;
      • Taramasso L.
      • Liggieri F.
      • Cenderello G.
      • Bovis F.
      • Giannini B.
      • Mesini A.
      • et al.
      Bloodstream infections in patients living with HIV in the modern cART era.
      ).
      Blood culture contaminations occurred in 41 (3.3%) admissions with a blood culture request.
      • Rodríguez-Créixems M.
      • Alcalá L.
      • Muñoz P.
      • Cercenado E.
      • Vicente T.
      • Bouza E.
      Bloodstream infections: evolution and trends in the microbiology workload, incidence, and etiology, 1985-2006.
      reported 5.3% contaminations in a teaching hospital. They also reported 10.7% polymicrobial cultures (compared to 4% in the present study).
      Although the BSI mortality rate has decreased since the introduction of ART, it still remains high, both in the short-term (
      • Declercq S.
      • De Munter P.
      • Derdelinckx I.
      • Verhaegen J.
      • Peetermans Willy E.
      • Vanderschueren S.
      • et al.
      Characteristics, causes, and outcome of 54 episodes of bloodstream infections in a cohort of HIV patients.
      ) and in the long-term, with a 3-year mortality of 41% (
      • Taramasso L.
      • Liggieri F.
      • Cenderello G.
      • Bovis F.
      • Giannini B.
      • Mesini A.
      • et al.
      Bloodstream infections in patients living with HIV in the modern cART era.
      ). The present study focused on factors associated with 30-day and 3-year mortality in PLWHIV who were admitted for BSI or had NBSI. The 30-day mortality rate was 27.6% (34 patients) in this cohort, probably reflecting the late presenters and the AIDS rate (87%). It is important to underline that 93.5% of the patients with concomitant BSI and a new HIV diagnosis were late presenters. In fact, despite new HIV diagnosis decreasing, Portugal has the highest rate (49%) of new cases and AIDS in Western Europe (

      DGS – Programa Nacional para a Infeção VIH e SIDA INSA – Departamento de Doenças Infecciosas Portugal. Ministério da Saúde. Direção-Geral da Saúde/Instituto Nacional de Saúde Doutor Ricardo Jorge. Infeção VIH e SIDA em Portugal - 2019. Lisboa: DGS/INSA; 2019. Direção-Geral da Saúde (DGS) e Instituto Nacional de Saúde Doutor Ricardo Jorge (INSA, IP): 2019 Retrieved July 11 from https://www.sns.gov.pt/wp-content/uploads/2019/11/Relat-VIH-SIDA-2019.pdf.

      ).
      Unexpectedly, older age had no direct impact on mortality, and the relationship between age and mortality is inconsistent (
      • Ortega M.
      • Almela M.
      • Soriano A.
      • Marco F.
      • Martínez J.A.
      • Muñoz A.
      • et al.
      Bloodstream infections among human immunodeficiency virus-infected adult patients: epidemiology and risk factors for mortality.
      ;
      • Taramasso L.
      • Liggieri F.
      • Cenderello G.
      • Bovis F.
      • Giannini B.
      • Mesini A.
      • et al.
      Bloodstream infections in patients living with HIV in the modern cART era.
      ). A higher APACHE II score was associated with an unfavorable BSI outcome (OR 1.12). Models based on APACHE II at admission have shown better performance than those based on the SOFA score to predict mortality in severe illness, with a higher area under the receiver operating characteristic curve (AUC) (0.73 for APACHE II and 0.63 for SOFA score), but with comparable performance after sequential evaluation (
      • Minne L.
      • Abu-Hanna A.
      • de Jonge E.
      Evaluation of SOFA-based models for predicting mortality in the ICU: A systematic review.
      ;
      • Hosseini M.
      • Ramazani J.
      Evaluation of Acute Physiology and Chronic Health Evaluation II and sequential organ failure assessment scoring systems for prognostication of outcomes among Intensive Care Unit's patients.
      ). It is possible that significance was lost in the present study because the SOFA score was only determined at admission and a combination of factors must be taken into consideration to estimate the prognosis. The definition of septic shock was based on the Sepsis-3 consensus according to the need for inotropic support, the fluid response, and lactate values. However, the evaluation of the fluid response and the decision regarding who should receive inotropic support and when this should be started could limit the significance of the SOFA score. Nevertheless, the APACHE II score has been shown to have strong discriminatory power for mortality (AUC 0.83) in septic patients and to be a reliable score to stratify disease severity with significance as a prognostic tool (
      • Sadaka F.
      • Ethmane Abou El Maali C.
      • Cytron M.A.
      • Fowler K.
      • Javaux V.M.
      • O'Brien J.
      Predicting Mortality of Patients With Sepsis: A Comparison of APACHE II and APACHE III Scoring Systems.
      ). Systemic corticotherapy (OR 3.74) and current IDU (OR 5.60) were also found to be unfavorable prognostic indicators.
      Successful treatment depends on the virulence of the microorganism, host barriers, and the immune response. Corticotherapy appears to have a role in bacteremia (
      • Fardet L.
      • Petersen I.
      • Nazareth I.
      Common Infections in Patients Prescribed Systemic Glucocorticoids in Primary Care: A Population-Based Cohort Study.
      ), and the incidence of early S. aureus bacteremia (
      • Bassetti M.
      • Trecarichi E.M.
      • Mesini A.
      • et al.
      Risk factors and mortality of healthcare-associated and community-acquired Staphylococcus aureus bacteraemia.
      ) was found to be influenced by the cumulative dose exposure (Smit et al., 2016 in a population-based case–control study. Also, in the study by
      • Ortega M.
      • Almela M.
      • Soriano A.
      • Marco F.
      • Martínez J.A.
      • Muñoz A.
      • et al.
      Bloodstream infections among human immunodeficiency virus-infected adult patients: epidemiology and risk factors for mortality.
      , corticotherapy had a prognostic impact. Interestingly, corticotherapy had an impact in both BSI (G1) and non-BSI HIV (G2) patients, leading to the hypothesis that the effects of corticotherapy go beyond bacteremia events. Of course, these results should be critically evaluated given the absence of corticosteroid specification, reduced number of patients on chronic therapy (8.9%), and cumulative dose information, which could be better evaluated in prospective studies. Evidence suggests that corticotherapy may influence the host immune response (
      • Heming N.
      • Sivanandamoorthy S.
      • Meng P.
      • Bounab R.
      • Annane D.
      Immune Effects of Corticosteroids in Sepsis.
      ), the persistence of respiratory tract colonization, antibiotic tolerance, and bacterial gene regulation (
      • Earl C.S.
      • Keong T.W.
      • An S.-Q.
      • Murdoch S.
      • McCarthy Y.
      • Garmendia J.
      • Ward J.
      • et al.
      Haemophilus influenzae responds to glucocorticoids used in asthma therapy by modulation of biofilm formation and antibiotic resistance.
      ). Among cancer patients, NBSI rate was higher in corticosteroid users (
      • Smit J.
      • Kaasch A.J.
      • Søgaard M.
      • Thomsen R.W.
      • Nielsen H.
      • Frøslev T.
      • et al.
      Use of Glucocorticoids and Risk of Community-Acquired Staphylococcus aureus Bacteremia: A Population-Based Case-Control Study.
      ).
      In Portugal, cases attributed to IDU reduced in 93%. However, in Portugal (

      DGS – Programa Nacional para a Infeção VIH e SIDA INSA – Departamento de Doenças Infecciosas Portugal. Ministério da Saúde. Direção-Geral da Saúde/Instituto Nacional de Saúde Doutor Ricardo Jorge. Infeção VIH e SIDA em Portugal - 2019. Lisboa: DGS/INSA; 2019. Direção-Geral da Saúde (DGS) e Instituto Nacional de Saúde Doutor Ricardo Jorge (INSA, IP): 2019 Retrieved July 11 from https://www.sns.gov.pt/wp-content/uploads/2019/11/Relat-VIH-SIDA-2019.pdf.

      ), in Europe and North America, drug addiction continues to predict a higher standard mortality rate (
      • Fontela C.
      • Aguinaga A.
      • Moreno-Iribas C.
      • Repáraz J.
      • Rivero M.
      • Gracia M.
      • et al.
      Trends and causes of mortality in a population-based cohort of HIV-infected adults in Spain: comparison with the general population.
      ). In Portugal, 49% of deaths occurred in IDU, at younger age. And globally, IDU has been estimated to cause 4.0% (2.82 million) of disability-adjusted life years due to HIV (
      • Degenhardt L.
      • Charlson F.
      • Stanaway J.
      • Larney S.
      • Alexander L.T.
      • Hickman M.
      • et al.
      Estimating the burden of disease attributable to injecting drug use as a risk factor for HIV, hepatitis C, and hepatitis B: findings from the Global Burden of Disease Study 2013.
      ). On the other hand, socioeconomic disadvantage, adverse lifestyle factors, and poorer mental health are strong predictors of hospitalization in IDU (
      • Rein S.M.
      • Smith C.J.
      • Chaloner C.
      • Stafford A.
      • Rodger A.J.
      • Johnson M.A.
      • et al.
      Prospective association of social circumstance, socioeconomic, lifestyle and mental health factors with subsequent hospitalisation over 6–7 year follow up in people living with HIV.
      ). The authors hypothesize that reduced previous ART taking, a lack of constant healthcare attendance, and possible socioeconomic circumstances may have contributed to the mortality in this cohort.
      Early diagnosis and cART accessibility have led to an increased overall survival in AIDS, with death more than 10 years after HIV diagnosis. In this study, we considered the 3-year follow-up, although we found a wide range of published follow-up durations. The overall 3-year mortality rate in this study was 46.3%, in agreement with
      • Taramasso L.
      • Liggieri F.
      • Cenderello G.
      • Bovis F.
      • Giannini B.
      • Mesini A.
      • et al.
      Bloodstream infections in patients living with HIV in the modern cART era.
      . As mentioned, age was found to be an inconsistent prognostic factor and the focus on age seemed not to be on age itself, but rather on age-associated comorbidities, like chronic vascular or pulmonary disease, which led to a 6.41 times higher chance of dying during follow-up. In a Spanish cohort, cardiovascular disease, respiratory disease, and malignancies were indicated to be non-HIV-related causes of excess mortality (
      • Fontela C.
      • Aguinaga A.
      • Moreno-Iribas C.
      • Repáraz J.
      • Rivero M.
      • Gracia M.
      • et al.
      Trends and causes of mortality in a population-based cohort of HIV-infected adults in Spain: comparison with the general population.
      ). AIDS-defining and non-AIDS-defining malignancies were the most common causes of death in the HIV population, contributing to 47% of all causes in a study performed in Tokyo (
      • Nishijima T.
      • Inaba Y.
      • Kawasaki Y.
      • Tsukada K.
      • Teruya K.
      • Kikuchi Y.
      • et al.
      Mortality and causes of death in people living with HIV in the era of combination antiretroviral therapy compared with the general population in Japan.
      ). Moreover, cardiovascular diseases and malignancy were the main causes of death in non-HIV Portuguese people (
      • Simões J.
      • Augusto G.
      • Fronteira I.
      • Hernández-Quevedo C.
      Portugal: Health System Review.
      ).
      According to the WHO, the CD4 cell count is the best predictor of disease status and the immediate risk of death. Indeed, the lowest standardized mortality ratio was seen in patients who did not have AIDS at cART initiation (
      The Antiretroviral Therapy Cohort Collaboration
      Mortality of HIV-infected patients starting potent antiretroviral therapy: comparison with the general population in nine industrialized countries.
      ;
      • May M.
      • Gompels M.
      • Delpech V.
      • Porter K.
      • Post F.
      • Johnson M.
      • et al.
      Impact of late diagnosis and treatment n life expectancy in people with HIV-1: UK Collaborative HIV Cohort (UK CHIC) Study.
      ,
      • May M.T.
      • Gompels M.
      • Delpech V.
      • Porter K.
      • Orkin C.
      • Kegg S.
      • et al.
      UK Collaborative HIV Cohort (UK CHIC) Study. Impact on life expectancy of HIV-1 positive individuals of CD4+ cell count and viral load response to antiretroviral therapy.
      ), who had attained a viral load of ≤500 copies/ml, and who had a CD4 cell count of ≥350 cells/μl by 6–12 months after starting cART (
      The Antiretroviral Therapy Cohort Collaboration
      Survival of HIV-positive patients starting antiretroviral therapy between 1996 and 2013: a collaborative analysis of cohort studies.
      ;
      • Katz I.T.
      • Maughan-Brown B.
      Improved life expectancy of people living with HIV: who is left behind?.
      ), although mortality remained 4.8 times higher than in non-HIV patients, in absolute terms due to non-AIDS cancer and cardiovascular disease (
      • Fontela C.
      • Aguinaga A.
      • Moreno-Iribas C.
      • Repáraz J.
      • Rivero M.
      • Gracia M.
      • et al.
      Trends and causes of mortality in a population-based cohort of HIV-infected adults in Spain: comparison with the general population.
      ). Sociodemographic factors, lifestyle factors, and treatment adherence are possible modifiers (
      • Sabin C.A.
      Do people with HIV infection have a normal life expectancy in the era of combination antiretroviral therapy?.
      ;
      • Poorolajal J.
      • Hooshmand E.
      • Mahjub H.
      • Esmailnasab N.
      • Jenabi E.
      Survival rate of AIDS disease and mortality in HIV-infected patients: a meta-analysis.
      ).
      Although this study was performed in a university hospital, covering a 10-year period, and in the absence of previously published data from Portuguese cohorts, it has several limitations. This was a retrospective single-center study, restricted to the post-ART era. Despite the exclusion of supposed contaminants, the review by
      • Petrosillo N.
      • Pagani L.
      • Ippolito G.
      Gruppo HIV e Infezioni Ospedaliere. Nosocomial infections in HIV-positive patients: an overview.
      reported that colonization appears to contribute to the development of nosocomial infections. Also, the duration of central line use could clarify the relationship between CRBSI, microorganisms, and nosocomial infection. The impact of early and suitable empiric antimicrobial therapy was not assessed, although it could be a predictor of mortality, as shown by
      • Ortega M.
      • Almela M.
      • Soriano A.
      • Marco F.
      • Martínez J.A.
      • Muñoz A.
      • et al.
      Bloodstream infections among human immunodeficiency virus-infected adult patients: epidemiology and risk factors for mortality.
      The 3-year follow-up was recorded. However, treatment adherence and events without hospitalization were not assessed. Nevertheless, studies have rarely reported long-term follow-up after discharge.
      In conclusion, BSI occurred in 6.8/100 admissions of PLWHIV. These patients were in the advanced stage of HIV, with 25% being newly diagnosed and significantly more late presenters, comorbidities, longer LOS, and higher 30-day mortality. Conditions related to IDU and the severity of infection determined the short-term prognosis. Factors predicting 3-year mortality were non-AIDS-defining events in the presence of lower CD4 cell counts and hematological malignancies. Systemic corticotherapy may influence nosocomial BSI and short-term prognosis.

      Uncited References:

      • Tacconelli E.
      • Tumbarello M.
      • Donati KdG
      • Bertagnolio S.
      • Pittiruti M.
      • Leone F.
      • et al.
      Morbidity associated with central venous catheter-use in a cohort of 212 hospitalized subjects with HIV infection.

      Funding source

      This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

      Ethical approval

      This study was performed according to the institutional ethics committee and the principles of the Declaration of Helsinki. Privacy rights of human subjects were ensured.

      Conflict of interest

      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.

      Sample CRediT author statement

      De Matos, Andreia: conceptualization, methodology, investigation, data collection, formal analysis, writing – original draft preparation; Brandão Lopes, Sara: investigation, data collection; Serra, José Eduardo: writing review and editing; Ferreira, Eugénia: writing review and editing; Saraiva da Cunha, José: project administration, writing review and editing.

      REFERENCES

        • Afessa B.
        • Morales I.
        • Weaver B.
        Bacteremia in hospitalized patients with human immunodeficiency virus: A prospective, cohort study.
        BMC Infect Dis. 2001; 1: 13
        • Antinori A.
        • Coenen T.
        • Costagiola D.
        • et al.
        Late presentation of HIV infection: a consensus definition.
        HIV Med. 2011; 12: 61-64
        • Bassetti M.
        • Trecarichi E.M.
        • Mesini A.
        • et al.
        Risk factors and mortality of healthcare-associated and community-acquired Staphylococcus aureus bacteraemia.
        Clin Microbiol Infect. 2012; 18: 862-869
        • Bonadio M.
        • Gigli C.
        • Maccanti O.
        • Longo B.
        • Smorfa A.
        Bloodstream infections in HIV-positive patients: a review of sixty-eight episodes.
        J Chemother. 1998; 10: 243-247
        • Catumbela E.
        • Freitas A.
        • Lopes F.
        • Mendoza Mdel C.
        • Costa C.
        • Sarmento A.
        • et al.
        HIV disease burden, cost, and length of stay in Portuguese hospitals from 2000 to 2010: a cross-sectional study.
        BMC Health Serv Res. 2015; 15: 144
        • Declercq S.
        • De Munter P.
        • Derdelinckx I.
        • Verhaegen J.
        • Peetermans Willy E.
        • Vanderschueren S.
        • et al.
        Characteristics, causes, and outcome of 54 episodes of bloodstream infections in a cohort of HIV patients.
        Infect Dis (Lond). 2015; 47: 611-617
        • Degenhardt L.
        • Charlson F.
        • Stanaway J.
        • Larney S.
        • Alexander L.T.
        • Hickman M.
        • et al.
        Estimating the burden of disease attributable to injecting drug use as a risk factor for HIV, hepatitis C, and hepatitis B: findings from the Global Burden of Disease Study 2013.
        Lancet Infect Dis. 2016; 16: 1385-1398
      1. DGS – Programa Nacional para a Infeção VIH e SIDA INSA – Departamento de Doenças Infecciosas Portugal. Ministério da Saúde. Direção-Geral da Saúde/Instituto Nacional de Saúde Doutor Ricardo Jorge. Infeção VIH e SIDA em Portugal - 2019. Lisboa: DGS/INSA; 2019. Direção-Geral da Saúde (DGS) e Instituto Nacional de Saúde Doutor Ricardo Jorge (INSA, IP): 2019 Retrieved July 11 from https://www.sns.gov.pt/wp-content/uploads/2019/11/Relat-VIH-SIDA-2019.pdf.

        • Earl C.S.
        • Keong T.W.
        • An S.-Q.
        • Murdoch S.
        • McCarthy Y.
        • Garmendia J.
        • Ward J.
        • et al.
        Haemophilus influenzae responds to glucocorticoids used in asthma therapy by modulation of biofilm formation and antibiotic resistance.
        EMBO Mol Med. 2015; 7: 1018-1033
        • Fardet L.
        • Petersen I.
        • Nazareth I.
        Common Infections in Patients Prescribed Systemic Glucocorticoids in Primary Care: A Population-Based Cohort Study.
        PLoS Med. 2016; 13e1002024
        • Fontela C.
        • Aguinaga A.
        • Moreno-Iribas C.
        • Repáraz J.
        • Rivero M.
        • Gracia M.
        • et al.
        Trends and causes of mortality in a population-based cohort of HIV-infected adults in Spain: comparison with the general population.
        Sci Rep. 2020; 10: 8922
        • Franceschini E.
        • Santoro A.
        • Menozzi M.
        • Bacca E.
        • Venturelli C.
        • Zona S.
        • Bedini A.
        • Digaetano M.
        • Puzzolante C.
        • Meschiari M.
        • Cuomo G.
        • Orlando G.
        • Sarti M.
        • Guaraldi G.
        • Cozzi-Lepri A.
        • Mussini C.
        Epidemiology and Outcomes of Bloodstream Infections in HIV-Patients during a 13-Year Period.
        Microorganisms. 2020; 8: 1210
        • Heming N.
        • Sivanandamoorthy S.
        • Meng P.
        • Bounab R.
        • Annane D.
        Immune Effects of Corticosteroids in Sepsis.
        Front. Immunology. 2018; 9: 1736
        • Hosseini M.
        • Ramazani J.
        Evaluation of Acute Physiology and Chronic Health Evaluation II and sequential organ failure assessment scoring systems for prognostication of outcomes among Intensive Care Unit's patients.
        Saudi J Anaesth. 2016; 10: 168-173
        • Katz I.T.
        • Maughan-Brown B.
        Improved life expectancy of people living with HIV: who is left behind?.
        The Lancet HIV. 2017; 4: e324-e326
        • Kiertiburanakul S.
        • Watcharatipagorn S.
        • Chongtrakool P.
        • Santanirand P.
        Epidemiology of bloodstream infections and predictive factors of mortality among HIV-infected adult patients in Thailand in the era of highly active antiretroviral therapy.
        Jpn J Infect Dis. 2012; 65: 28-32
        • Taramasso L.
        • Liggieri F.
        • Cenderello G.
        • Bovis F.
        • Giannini B.
        • Mesini A.
        • et al.
        Bloodstream infections in patients living with HIV in the modern cART era.
        Nature Scientific Reports. 2019; 9: 5418
        • May M.
        • Gompels M.
        • Delpech V.
        • Porter K.
        • Post F.
        • Johnson M.
        • et al.
        Impact of late diagnosis and treatment n life expectancy in people with HIV-1: UK Collaborative HIV Cohort (UK CHIC) Study.
        BMJ. 2011; 343: d6016
        • May M.T.
        • Gompels M.
        • Delpech V.
        • Porter K.
        • Orkin C.
        • Kegg S.
        • et al.
        UK Collaborative HIV Cohort (UK CHIC) Study. Impact on life expectancy of HIV-1 positive individuals of CD4+ cell count and viral load response to antiretroviral therapy.
        AIDS. 2014; 28: 1193-1202
        • Minne L.
        • Abu-Hanna A.
        • de Jonge E.
        Evaluation of SOFA-based models for predicting mortality in the ICU: A systematic review.
        Critical Care. 2008; 12: R161
        • Nishijima T.
        • Inaba Y.
        • Kawasaki Y.
        • Tsukada K.
        • Teruya K.
        • Kikuchi Y.
        • et al.
        Mortality and causes of death in people living with HIV in the era of combination antiretroviral therapy compared with the general population in Japan.
        AIDS. 2020; 34: 913-921
        • Ortega M.
        • Almela M.
        • Soriano A.
        • Marco F.
        • Martínez J.A.
        • Muñoz A.
        • et al.
        Bloodstream infections among human immunodeficiency virus-infected adult patients: epidemiology and risk factors for mortality.
        European Journal of Clinical Microbiology & Infectious Diseases. 2008; 27: 969-976
        • Panis C.
        • Matsuo T.
        • Reiche E.M.V.
        Nosocomial infections in human immunodeficiency virus type 1 (HIV-1) infected and AIDS patients: major microorganisms and immunological profile.
        Braz. J. Microbiol. 2009; 40: 155-162
        • Pedro-Botet M.L.
        • Mòdol J.M.
        • Vallés X.
        • Romeu J.
        • Sopena N.
        • Giménez M.
        • et al.
        Changes in bloodstream infections in HIV-positive patients in a university hospital in Spain (1995-1997).
        Int J Infect Dis. 2002; 6: 17-22
        • Petrosillo N.
        • Pagani L.
        • Ippolito G.
        Gruppo HIV e Infezioni Ospedaliere. Nosocomial infections in HIV-positive patients: an overview.
        Infection. 2003; 2: 28-34
        • Petrosillo N.
        • Viale P.
        • Nicastri E.
        • Arici C.
        • Bombana E.
        • Casella A.
        • et al.
        Nosocomial bloodstream infections among human immunodeficiency virus-infected patients: incidence and risk factors.
        Clin Infect Dis. 2002; 34: 677-685
        • Poorolajal J.
        • Hooshmand E.
        • Mahjub H.
        • Esmailnasab N.
        • Jenabi E.
        Survival rate of AIDS disease and mortality in HIV-infected patients: a meta-analysis.
        Public Health. 2016; 139: 3-12
        • Rein S.M.
        • Smith C.J.
        • Chaloner C.
        • Stafford A.
        • Rodger A.J.
        • Johnson M.A.
        • et al.
        Prospective association of social circumstance, socioeconomic, lifestyle and mental health factors with subsequent hospitalisation over 6–7 year follow up in people living with HIV.
        Eclinical Medicine. 2020; (article 100665)
        • Rodríguez-Créixems M.
        • Alcalá L.
        • Muñoz P.
        • Cercenado E.
        • Vicente T.
        • Bouza E.
        Bloodstream infections: evolution and trends in the microbiology workload, incidence, and etiology, 1985-2006.
        Medicine (Baltimore). 2008; 87: 234-249
        • Sabin C.A.
        Do people with HIV infection have a normal life expectancy in the era of combination antiretroviral therapy?.
        BMC Med. 2013; 11: 251
        • Sadaka F.
        • Ethmane Abou El Maali C.
        • Cytron M.A.
        • Fowler K.
        • Javaux V.M.
        • O'Brien J.
        Predicting Mortality of Patients With Sepsis: A Comparison of APACHE II and APACHE III Scoring Systems.
        J Clin Med Res. 2017; 9: 907-910
        • Simões J.
        • Augusto G.
        • Fronteira I.
        • Hernández-Quevedo C.
        Portugal: Health System Review.
        Health Systems in Transition. 2017; 19: 1-184
        • Singer M.
        • Deutschman C.S.
        • Seymour C.W.
        • et al.
        The Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3).
        JAMA. 2016; 315: 801-810
        • Smit J.
        • Kaasch A.J.
        • Søgaard M.
        • Thomsen R.W.
        • Nielsen H.
        • Frøslev T.
        • et al.
        Use of Glucocorticoids and Risk of Community-Acquired Staphylococcus aureus Bacteremia: A Population-Based Case-Control Study.
        Mayo Clin Proc. 2016; 91: 873-880
        • Stroud L.
        • Srivastava P.
        • Culver D.
        • Bisno A.
        • Rimland D.
        • Simberkoff M.
        • et al.
        Nosocomial Infections in HIV-Infected Patients Preliminary Results from a Multicenter Surveillance System (1989-1995).
        Infection Control & Hospital Epidemiology. 1997; 18: 479-485
        • Suetens C.
        • Latour K.
        • Kärki T.
        • Ricchizzi E.
        • Kinross P.
        • Moro M.L.
        • et al.
        The Healthcare-Associated Infections Prevalence Study Group. Prevalence of healthcare-associated infections, estimated incidence and composite antimicrobial resistance index in acute care hospitals and long-term care facilities: results from two European point prevalence surveys, 2016 to 2017.
        Euro Surveill. 2018; 231800516
        • Tacconelli E.
        • Tumbarello M.
        • Donati KdG
        • Bertagnolio S.
        • Pittiruti M.
        • Leone F.
        • et al.
        Morbidity associated with central venous catheter-use in a cohort of 212 hospitalized subjects with HIV infection.
        Journal of Hospital Infection. 2000; 44: 186-192
        • Taramasso L.
        • Tatarelli P.
        • Di Biagio A.
        Bloodstream infections in HIV-infected patients.
        Virulence. 2016; 7: 320-328
        • The Antiretroviral Therapy Cohort Collaboration
        Survival of HIV-positive patients starting antiretroviral therapy between 1996 and 2013: a collaborative analysis of cohort studies.
        The Lancet HIV. 2017; 4: e349-e356
        • The Antiretroviral Therapy Cohort Collaboration
        Mortality of HIV-infected patients starting potent antiretroviral therapy: comparison with the general population in nine industrialized countries.
        Int J Epidemiol. 2009; 38: 1624-1633
        • Tumbarello M.
        • Tacconelli E.
        • Donati K.G.
        • Leone F.
        • Morace G.
        • Cauda R.
        • et al.
        Nosocomial bloodstream infections in HIV-infected patients: attributable mortality and extension of hospital stay.
        J Acquir Immune Defic Syndr Hum Retrovirol. 1998; 19: 490-497