Acute kidney injury and other factors associated with mortality in hiv-infected patients

1. School of Medicine, Public Health and Medical Sciences Post-Graduation Programs, Health Sciences Center, University of Fortaleza. Fortaleza, Ceará, Brasil. 2. Department of Internal Medicine, School of Medicine, Federal University of Ceará.Fortaleza, Ceará, Brasil. 3. School of Medicine, Medical Sciences and PharmacologyPost-Graduation Programs, Federal University of Ceará. Fortaleza, Ceará, Brasil. 4. Department of Community Health, Public Health Post-Graduation Program, School of Medicine, Federal University of Ceará. Fortaleza, Ceará, Brasil.


INTRODUCTION
According to the United Nations Program on HIV/ AIDS (UNAIDS) 1 , an estimated 36.7 million people worldwide were living with HIV in 2015, while 2.1 million people were infected in the same year.This alarming epidemiologic situation confirms the role of HIV infection as an undoubtedly severe public health issue, which leads to an estimated annual spending of 19 billion dollars and over a million deaths every year 1 .
In the late 90s, the association of multiple antiretroviral drugs was a milestone in the management of HIV infection, initiating the highly active antiretroviral therapy (HAART) era.The access to HAART has increased since then, reaching more than 17 million people in 2015 2 .Several studies have demonstrated the efficacy of such treatment in reducing viral RNA copies, preserving CD4 cells, and consequently decreasing morbidity and mortality 3 .On the other hand, HAART has been associated with significant adverse effects, such as liver dysfunction, skin rash, hypertension, diabetes, dyslipidemia, myocardial infarction and nephrotoxicity 4, 5 .
After the extensive introduction of HAART in developed countries of Western Europe and the United States in 1996, HIV-associated mortality has markedly decreased.In these countries, the leading causes of HIV-related death have been changing significantly, shifting from opportunistic infectious diseases to chronic conditions 6 .On the other hand, the national provision of HAART in Latin America started from 1997 to 2004 and Brazil was the first country in the region to establish policies to provide free and universal access to these medications.A study showed that only six out of eleven studied countries in the area presented a decline in HIV mortality from 1997 to 2007, probably due to this delay in implementing such free-access policies 7 .
Considering the importance of the theme, the aim of this study is to investigate factors associated with death in HIV-infected patients.

Study Design
This is a cross-sectional study with 200 HIV infected patients admitted to the São José Infectious Diseases Hospital, in Fortaleza, Northeast of Brazil, from January 2013 to December 2014.This hospital is a reference for all infectious diseases in the state of Ceará, in Northeast Brazil.All patients included were ≥ 16 years old, with a confirmed diagnosis of HIV infection.According to the protocol of the Brazilian Ministry of Health, at least two different tests must be used to confirm HIV infection, including enzyme-linked immunosorbent assay (ELISA), simple/rapid test devices, and western blot.All patients with previous estimated glomerular filtration rate (eGFR) <60ml/min/1.73m²,heart failure, nephrolithiasis, use of nephrotoxic drugs (except for antiretrovirals), or other conditions that may affect renal function were excluded.Patients were divided into two different groups for comparison: those who died after admission (non-survivors) and those who did not (survivors).

Definitions
Acute kidney injury (AKI) was defined and classified according to the Kidney Diseases Improving Global Outcomes (KDIGO) criteria 8 .Hypotension was defined as mean arterial blood pressure (MAP) <60mmHg, and therapy with vasopressors was initiated when the MAP remained <60mmHg, despite adequate fluid administration.Oliguria was defined as urine output <0.5ml/kg/h after adequate fluid replacement.Hemodialysis was indicated for those patients that remained oliguric after effective hydration, in cases with uremia-associated hemorrhagic phenomena or severe respiratory failure, as well as for patients with treatment-refractory hyperkalemia and metabolic acidosis.
Thrombocytopenia was defined as platelets count lower than 150,000/mm 3 and anemia as hemoglobin (Hb) <12g/dL.The occurrence of metabolic acidosis was evidenced when pH < 7.35 and serum bicarbonate <20mEq/L, and severe metabolic acidosis when pH<7.10.Liver dysfunction was defined as the elevation of aspartate aminotransferase (AST) or alanine aminotransferase (ALT) >120U/L (three times higher than the normal range).All associated infections were diagnosed by the attendant clinicians during hospitalization and were confirmed by serologic or in order to assess variable distribution.Variables with normal distribution were expressed through mean ± standard deviation (SD).Variables with non-normal distribution were expressed through median values.A comparison of categorical variables was executed using Pearson's Chi-square, while numerical variables were compared using Student's T-test (for variables with normal distribution) or Mann-Whitney U test (for variables with non-normal distribution).P values ≤0.05 were considered statistically significant.In order to evaluate risk factors for death, a logistic regression model was used for categorical variables.All variables with statistical significance in univariate analysis were included in multivariate analysis.Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were calculated.

Ethics
This study was reviewed and approved by the Ethics Committee of São José Infectious Disease Hospital.

RESULTS
A total of 200 patients were included in the studied, 69.5% male.The mean age was 39 years, ranging from 16 to 74 years.Fifteen patients (7.5%) died.The percentage of males was significantly higher in non-survivors than in survivors (93.3 vs. 67.3%,p=0.037).Almost half of the patients (49%) used laboratory tests, when possible.Disseminated histoplasmosis (DH), specifically, was diagnosed by identification of H. capsulatum in peripheral blood smear, bone marrow aspirate, blood and bone marrow culture, or histopathology of different tissues.

Studied parameters
Clinical and demographical parameters included age, gender, hospitalization time, main signs and symptoms on admission, the occurrence of associated infections, previous HAART use, AKI development, hemodialysis requirement, and death.All infectious diseases other than HIV presented by the patients, opportunistic or not, were considered "associated infections".Laboratory assessment included: mean HIV viral load, CD4 lymphocytes count, serum urea and creatinine, hemoglobin (Hb), hematocrit, leukocytes count, platelets count, sodium, potassium, aspartate amino transaminase (AST), alanine aminotransaminase (ALT), and lactate dehydrogenase (LDH).Arterial blood gas analysis included arterial pH, carbon dioxide partial pressure (pCO 2 ), and bicarbonate (HCO 3 ).

Statistics
Statistical analysis was executed using the SPSS software for Windows version 20.0 (IBM, USA) and its results were expressed through tables.The Kolmogorov-Smirnov test was used for numeric variables  as presented in Table 2.

DISCUSSION
Despite a remarkable decrease in HIV-related mortality after the introduction of HAART, its rates remain notably high, especially in developing countries.The leading causes of death in these patients have progressively changed, shifting from an infectious and AIDS-related disease majority to chronic and non-AIDS conditions predominance, such as hepatic, pulmonary, and cardiovascular illnesses 6 .In this context, the present work represents one of the main studies to investigate risk factors for death among HIV-infected patients in our region.
When analyzing demographic data, we noticed that the studied population included a majority of males, as well as young and middle-aged people (<40 years = 58%).Older age has frequently been linked to elevated mortality among HIV patients, due to a higher incidence of comorbidities and lesser adherence to treatment.On the other hand, in a large epidemiolog-HAART previously to the admission, but there was no significant difference between groups (50.3 vs. 33.3%,p=0.207).Non-survivors presented a significantly shorter hospital stay (8 ± 6 vs. 18 ± 15 days, p=0.005), developed more AKI (73.3 vs. 10.3%, p<0.001), had more severe forms of AKI (Stage 3 -53.3 vs. 4.5%, p<0.001) and needed more hemodialysis (46.7 vs. 2.7%, p<0.001) than survivors, as presented in Table 1.
Regarding laboratory evaluation, non-survivors presented lower levels of hematocrit ( 28 ical Chinese study, death was more common among those who were 20-49 years old, similarly to the majority of our patients 9 .
In the present study, the percentage of males was higher among non-survivors than in survivors.Higher mortality in males has been extensively reported, mostly due to late referencing of men to HIV care and poor adherence to treatment 10 .Studies have demonstrated that women tend to initiate HAART sooner, which can be linked to HIV screening in prenatal exams for all pregnant women, which is mandatory in Brazil since 1996.On the other hand, men tend to look for medical attention only when they manifest symptoms, leading to late diagnosis and treatment 11 .We also observed that non-survivors had a shorter hospitalization time, mostly because they frequently died a few days after admission due to an acute condition or a complication.
Interestingly, HAART use was not significantly different between groups in the present study.Many studies have associated antiretroviral therapy with reduction of mortality, and its introduction may be considered one of the most important advances in the history of HIV infection, since it was responsible for reducing several complications and poor outcomes, including death 12 .Considering the fact that HAART use in the studied population did not influence CD4 lymphocytes count and viral load in both groups, we hypothesize that the lack of mortality decrease secondary to HAART use derives from a combination of irregular administration and insufficient information from charts.When data were collected, only basic information of previous HAART use was available in the charts (yes or no), but it was not described if these medications were being adequately taken by the patients, or how long they have been taking them.Hence, we believe that some patients who were classified as using HAART were not taking it correctly, leading to no difference between groups regarding HAART use, CD4 lymphocytes count, and viral load.
According to several studies, the general incidence of AKI on HIV-infected individuals is about 2.8-5.9% per year in outpatients, 6-18% in hospitalized patients, and 47.4% in critically ill patients [13][14][15] .In the present study, non-survivors had a significantly higher incidence of AKI and oliguria, needed more renal replacement therapy, and presented higher levels of urea and creatinine.Among hospitalized patients, AKI is more common in HIV-infected individuals than in the general population, as expected.
In the pre-HAART era, pre-renal azotemia was the unquestionable main form of AKI in HIV-infected patients, caused by volume depletion, associated infection, and sepsis 16 .On the other hand, in the post-HAART era AKI etiology changed and became multifactorial.Currently, volume depletion and sepsis are of both short and long-term mortality in HIV-infected individuals, as well as its association with other poor outcomes, like heart failure and end-stage renal disease (ESRD) 14, 18 .In the present study, severe forms of AKI were more prevalent in non-survivors.As previously described, they were considered an important risk factor for in-hospital HIV mortality 19, 20 .Additionally, we noticed that patients in the non-survivors group had significantly lower levels of serum bicarbonate and arterial CO2 partial pressure, which represents a higher incidence of metabolic acidosis and hyperventilation.Acidosis has been recognized as a consequence of AKI in HIV patients and has been described as a risk factor for death in this population 21 .We believe that AKI was the primary cause of metabolic acidosis, but other factors may have contributed, such as drug toxicity by the nucleoside/nucleotide reverse transcriptase inhibitors (NRTI/NtRTIs) 22 .Several studies have demonstrated that drugs like stavudine, didanosine, zidovudine, lamivudine, abacavir, and tenofovir may cause acidosis through different mechanisms, such as mitochondrial dysfunction, Fanconi syndrome (proximal acidosis), and distal renal tubular acidosis 23, 24 .In a previous study from our group with pediatric HIV patients, we observed that metabolic acidosis also happened as a consequence of HAART use 25 .
Regarding liver involvement, non-survivors presented significantly higher levels of AST and ALT on admission.An extensive study by Weber et al. 26 showed that liver-related diseases caused 14.5% of all deaths and were the first cause of non-AIDS-related deaths.Many different factors may be linked to liver dysfunction in HIV patients, including hepatitis B (HBV) and C (HCV) coinfection, steatosis and non-alcoholic steatohepatitis (NASH), insulin resistance, drug toxicity, and direct effect of HIV 27 .It is difficult to define which factors are responsible for liver injury in our patients, but interestingly none of them tested positive for HBV or HCV.These infections share transmission routes with HIV and are observed in 5-25% of all HIV-infected patients worldwide but are usually overshadowed by opportunistic infections and hepatotoxic drugs as causes of liver injury 28 .
Pulmonary infections are widespread in immunocompromised hosts, even in the HAART era and nearly 70% of HIV patients will experience respiratory complications at some point in their disease 29 .Pulmonary complications are the leading cause of hospitalization 30 and the leading causes of respiratory involvement in developed countries are bacterial pneumonia, bronchitis ,and Pneumocystis jiroveci pneumonia.Tuberculosis is still a frequent complication of HIV-infection in developing countries, while it is becoming rare in the developed world 31 .In the present study, dyspnea on admission was a risk factor for death in the multivariate analysis, which reflects pulmonary involvement.Other studies have shown that pulmonary infections are risk factors for death even in the HAART era, and HIV-patients are at higher risk of these conditions.However, it was not assessed whether the clinical symptom of dyspnea on admission was associated with higher mortality 32 .
Additionally, disorientation was another important finding on admission.HIV-infection may cause disorientation for a myriad of reasons, including psychiatric comorbidities, immune dysfunction, HIV infection of the central nervous system (CNS), or progression of a systemic disease 33 .Although HAART treatment may prevent mental status alterations by precluding some of the above-mentioned complications, it can be the cause of the disorientation itself

34
. In the present study, disorientation was another risk factor for death, increasing mortality more than five-fold, in accordance with previous studies 35 .However, altered mental status has been demonstrated to increase mortality not only when there is ongoing HIV-associated CNS infection, but also in general wards 36, 37 .Such findings suggest that the increase in the risk of death derives primarily from the progression of the systemic disease, rather than from the infection of the CNS.Among the myriad of HIV-associated infections, histoplasmosis is an important and dangerous global systemic mycosis caused by H. capsulatum.It is endemic in the Americas and has a notably high incidence in the state of Ceará, where this study was conducted 38,39 .It is considered an opportunistic infection which usually affects HIV patients with low CD4 levels, occurring in up to 5% of HIV infected patients in endemic areas 40 .Recent studies have stated that Histoplasma infection is more widespread than previously thought and its morbidity and mortality have been hugely underestimated due to severe underdiagnosing and underreporting 41 .The clinical presentation of this harmful disease may vary from asymptomatic infection with fever, chills, headache and malaise, to a life-threatening disseminated condition with septic shock, renal dysfunction, hepatic failure, acute respiratory distress syndrome and coagulopathy 39 .In addition, disseminated histoplasmosis [DH] has been described as an AIDS-defining condition, which usually occurs early in the course of HIV untreated infection and could even be the first manifestation of AIDS in some cases 42 .
Unfortunately, poor outcomes are still very common in HIV patients with DH.Mortality rates among these patients can reach almost 50%, more prominently when severe manifestations like sepsis, renal failure, and pulmonary insufficiency are present 43 .In the studied population, DH was more common among non-survivors than survivors and it was a significant predictor of death in multivariate analysis.Histoplasmosis remains an important cause of death in HIV patients, and it has been described as the third most fatal mycosis among this population in Brazil 44 .Both early and long-term histoplasmosis mortality rates have been decreasing since the introduction of HAART, and antiretroviral use has been described as the most critical variable that influenced mortality reduction 45 .Several factors have been associated with mortality in patients with HIV-DH coinfection.Some authors demonstrated that high levels of creatinine and low levels of serum albumin were related to poor outcomes while others found that dyspnea, thrombocytopenia, and increased lactate dehydrogenase (LDH) levels were associated with mortality 40, 46 .In a previous study from our group, other factors like diarrhea, neurologic manifestations, low hemoglobin levels, increases serum urea, liver involvement, respiratory insufficiency, and AKI were also linked to death in this population 47 .
On the other hand, LDH levels were significantly lower in survivors group than in nonsurvivors, and LDH >1000U/L was a risk factor for death among our patients.In the present context, increased LDH levels are strongly related to histoplasmosis and may be used to differentiate this infection from Pneumocystis jiroveci pneumonia (PJP) 48 .In a previous study with HIV-infected patients from our group, we found higher levels of LDH in histoplasmosis than in other opportunistic infections 39 .Other authors have found that levels of LDH activity >2 times the upper limit of the normal range was a predictor of death 46 .
In recent years, it has been stated that the significant LDH increase in histoplasmosis was associated with the development of hemophagocytic syndrome.This syndrome, also known as hemophagocytic lymphohistiocytosis (HLH), results from unregulated macrophage activation with overproduction of inflammatory cytokines, leading to hemophagocytosis 49 .It may be primary or secondary, triggered by hematologic malignancies, autoimmune diseases, or several infections, such as DH and HIV itself 50 .There is still scarce information about histoplasmosis-associated HLH, but it is recognized as a high-mortality condition 50 .Considering the intimate relationship between LDH and DH in the HIV setting, we believe that the role of elevated LDH levels as a risk factor for death derives from its connection to higher DH incidence and consequently more HLH.Since DH has been recognized as an underestimated and underdiagnosed global burden, further studies are necessary to investigate the connection between LDH, HLH, and mortality in DH patients.
In conclusion, neurologic impairment, pulmonary and respiratory involvement and AKI development were considered risk factors for death in HIV-infected patients.Disseminated histoplasmosis and elevated levels of lactate dehydrogenase (LDH) were also predictors of death in this population.

FIGURE 1 -
FIGURE 1 -Comparison of clinical presentation on admission between survivors and non-survivors.

FIGURE 2 -
FIGURE 2 -Comparison of associated infections between survivors and non-survivors.

TABLE 1 -
COMPARISON OF DEMOGRAPHIC DATA, CLINICAL PARAMETERS, CO-MORBIDITIES, AND OUTCOMES BETWEEN SURVIVORS AND NON-SURVIVORS.
HVB -hepatitis virus B; HVC -hepatitis virus C. Chi-square test and Student's T-test were used.P values <0.05 were considered statistically significant.

TABLE 2 -
COMPARISON OF LABORATORY DATA ON ADMISSION BETWEEN SURVIVORS AND NON-SURVIVORS.

TABLE 3 -
RISK FACTORS FOR DEATH AMONG HIV INFECTED PATIENTS.