Potential Association of Hypernatremia With Mortality in Patients With Acute Kidney Injury and COVID-19

Background The outcome of acute kidney injury (AKI) in patients with COVID-19 and the factors associated with its outcome, including mortality, are understudied among the Indian population. Objective The objective of this study is to determine the outcome of AKI in a cohort of patients with COVID-19 admitted to medical wards and associated intensive care unit (ICU) and the factors associated with its outcome, including mortality. Method This is a retrospective study of patients with COVID-19 and AKI admitted to a tertiary care hospital. A total of 1765 patients were admitted to a hospital with COVID-19 between March 23, 2021, and June 30, 2021, during the second wave of the pandemic chiefly attributed to SARS-Co-V-2 lineage B.1.617. Patients with AKI for whom a nephrology call was sought for management (N=60) were included. Measurements carried out were the stage of AKI, co-morbidities, ICU admission, mechanical ventilation, lab parameters, and mortality. We classified AKI by comparing the highest to lowest recorded serum creatinine in hospital and staged AKI based on the Kidney Disease: Improving Global Outcomes (KDIGO) system. We further developed stepwise logistic regression models to find independent factors associated with mortality. Results Out of the 1765 patients hospitalized with COVID-19, a total of 60 (3.4%) patients with AKI were referred to nephrology for management. The observed mortality in this cohort was 41/60 (68.3%). AKI stage 3 was observed to be the most common (78.3%). Based on a univariate analysis of association, age, chronic kidney disease, admission to ICU, the requirement for vasopressor and ventilation, lactate dehydrogenase (LDH) DH, liver function tests (LFT), hypernatremia, and leucocytosis were associated with the mortality of patients (p<0.05) with AKI and COVID-19 infection. Multivariate analysis using logistic regression led to the identification of hypernatremia (OR 5.24 {0.95-42.31}) and multiple co-morbidities (OR 2.59 {1.03-8.75}, p<0.07) as potential factors independently associated with mortality. Conclusion The study indicates the potential association of hypernatremia with mortality in AKI, along with the simultaneous presence of multiple co-morbidities with COVID-19. As the statistical power of the association is weak, we are claiming the association as potential only. It needs to be confirmed in other larger studies.


Introduction
Acute kidney injury (AKI) in patients infected with COVID-19 has been found to be a serious illness; the mortality rate among patients with AKI has been reported to vary between 20 and 76% [1]. Kidney injury results from direct effects as well as from systemic effects of COVID-19 illness such as septic shock. Nephrotoxins also contribute to kidney injury [2,3]. Several electrolyte abnormalities (dysnatremias, hypoand hyperkalemia, and hypocalcemia) are reported in patients with COVID-19 infection and may be a manifestation of renal injury due to COVID-19 illness or may be a result of other pathophysiology, e.g., syndrome of inappropriate antidiuretic hormone (SIADH) [4,5]. The severity of AKI, age, and various scores of patient illness, such as Sequential Organ Failure Assessment (SOFA), Acute Physiology and Chronic Health Evaluation (APACHE II), D-dimers, bilirubin, oliguria in patients on renal replacement therapy (RRT), etc., have been reported to be predictors of mortality in the subgroup of patients with AKI [2,6,7]. Factors associated with mortality in patients with AKI and COVID-19 illness are understudied and under-reported in India [8,9]. Our objective was to determine which factors conferred an increased odds of death for patients admitted to a hospital with COVID-19 and AKI. In this article, we report the single-center outcome of an Indian cohort of patients with AKI and COVID-19 infection admitted to the tertiary care center during the second wave of the pandemic.

Study design and settings
The study was conducted at a tertiary care center, All India Institute of Medical Sciences in Bhopal, India, that was converted into a "COVID-only" hospital between March 23, 2021, and June 30, 2021, during the second wave of the pandemic to provide care to COVID-19 patients only. The most dominant strain during the second wave in India reported was SARS-Co-V-2 lineage B.1.617 [10]. We planned a retrospective medical record review to find the outcome of COVID-19 patients with AKI and the factors associated with mortality. The study was approved by the internal ethics committee of the All Indian Institute of Medical Sciences, Bhopal. The consent of the subjects was waived off by the Institutional Human Ethics Committee (IHEC), considering the retrospective nature of the study and the evaluation of data on a de-identified basis.

Study participants
The inclusion criteria were: 1) age between 18 and 80 years, 2) COVID-19 diagnosed by positive RT-PCR, (3) patient should have the presence of AKI either on admission or during stay for which a nephrology consultation was sought, and (4) patient should have been admitted to wards or peripheral ICUs between March 23, 2021, and June 30, 2021.

Study variables
AKI was defined according to the Kidney Disease: Improving Global Outcomes (KDIGO) clinical practice guideline, based on changes in serum creatinine [11]. We reported the stages as per the KDIGO framework as stages 1, 2, and 3. These correspond to 1.5-1.9 times their baseline creatinine, 2-2.5 times their baseline creatinine, and 3 or more times their baseline creatinine or newly required dialysis, respectively. Urine output was not considered due to the limited collection of relevant data in routine inpatient care. Particularly, ward patients were unlikely to have corrected timed input and output calculations. Change in creatinine was calculated by comparing the peak creatinine with the lowest creatinine during hospitalization, under the assumption that the lowest creatinine would best represent baseline kidney function. In addition, we assessed AKI on admission (pre-hospitalization AKI) by comparing the admission creatinine with the lowest creatinine during hospitalization. This was done to determine whether AKI was present at the time of admission. Pre-hospitalization AKI was defined as an increase of creatinine by greater than 0.3mg/dL. We calculated this by subtracting their admission creatinine from their lowest recorded creatinine during the hospital stay.
We used Acute Disease Quality Initiative (ADQI) 16 Workgroup [12] definitions of AKI recovery to classify AKI recovery, wherein persistent AKI is characterized by the continuance of AKI by serum creatinine or urine output criteria (as defined by KDIGO) beyond 48h from AKI onset. Complete reversal of AKI by KDIGO criteria within 48h of AKI onset characterizes rapid reversal of AKI and acute kidney disease (AKD) is defined as a condition wherein criteria for AKI stage 1 or greater persists for seven or more days after an exposure.
Other exposures included demographic factors, factors of COVID-19 infection severity (need for vasopressors, ventilator, need for ICU, length of admission), co-morbidities, and lab parameters including hypernatremia defined as serum sodium >145 mEq/Lt and severe hypernatremia as >152 mEq/Lt..

Outcome
Primary outcomes included mortality within the hospital for the studied hospitalization event due to COVID-19 patients with AKI.

Statistical methods
We described patient characteristics using median (25th-75th percentile) for numerical variables and frequency (percentage) for nominal variables. We stratified the various demographic, clinical, and biochemical factors across those who died or were discharged. The distribution of these variables was compared using a chi-squared test in case of nominal variables and the Mann-Whitney test in case of numerical variables. Then, looking at the results of these comparisons, we identified the factors for performing a multivariate analysis. We constructed stepwise logistic regression models, wherein variable selection was based on biological plausible previously described associations with AKI in both COVID-19 and general medical literature as well as factors selected from the univariate analysis. These included ages, multiple co-morbidities, chronic kidney disease (CKD), AKI stage, hypernatremia, white blood cell (WBC) counts, and deranged liver functions (serum bilirubin and serum glutamic-oxaloacetic transaminase (SGOT). We reported odds ratios (ORs) with 95% confidence intervals. Further, model diagnostics was performed for assessing model fit. Lastly, all analyses were done using R version 4.0 (R Foundation, Vienna, Austria).

Basic characteristics
The study included 60 patients, after the exclusion of five patients with missing data, admitted with COVID-19 infection and AKI in this analysis, for which a nephrology consultation was sought. All patients were reverse transcription polymerase chain reaction (RT-PCR) positive. Patients with AKI who did not survive were more likely to show the following conditions: higher age, male, multiple co-morbidities, diabetes, and hypertension, admission to ICU, stage 3 AKI, the requirement of vasopressors, ventilatory support, hypernatremia, lactate dehydrogenase (LDH), WBC counts, and deranged liver functions. They also had lower absolute lymphocyte counts(ALC), renal recovery, and diastolic blood pressure. Tables 1, 2 provide more details on baseline characteristics and univariately associated factors.

Outcomes of patients with AKI
There were 41 (68.3%) deaths in our cohort, whereas the remaining 19 patients with COVID-19 and AKI survived (discharged). Twenty-five (61%) of the patients who died were admitted to ICU, while 15 (79%) of the survived patients with AKI were from wards. In the mortality group, 33 (80%) of patients had stage 3 AKI, as opposed to 11 (58%) patients in the survived group. Only 19.6% had stage 1 and 2 AKI in the mortality group; in the discharge group, 42% had stage 1 or 2 AKI. In the mortality group, 22 (54%) required vasopressors, whereas 14 (74%) in the non-mortality group were not on vasopressors. In the mortality group, 33 (80%) patients required mechanical ventilation, as opposed to 13 (68%) of the survivors. Hypernatremia was present in 25 (61%) of the patients who died, whereas it was present in only 3 (16%) of the patients who survived. Most of the patients developed persistent AKI 44/60 (73.3%), whereas AKD developed in 13 (21.7%) cases. These findings are summarized in Table 3.

Discussion
The study reports the outcome of COVID-19 patients (n=60) with AKI for whom a nephrology consultation was sought; therefore, a true incidence of AKI cannot be commented upon because patients for whom a renal consult was not sent may also have AKI. However, the proportion of referred patients (3.4%) does not differ much from the incidences (3.5-7.1%) reported by other Indian studies [8,13,14]. The mortality found was high at 68.3%, as reported in other studies [1,8,13].
Association on univariate analysis with factors such as age, admission to ICU, lower blood pressure, a higher requirement of vasopressors and ventilation, deranged liver function tests (bilirubin, SGOT serum glutamic pyruvic transaminase (SGPT), and albumin), and higher LDH indicates the overall critical status of the patients as found by other researchers. Although stage 3 AKI was the most common stage of occurrence, we did not find an association with mortality; it may be related to the small cohort size of our study. Severity scores for all patients were not available, so they were not included in the analysis.
AKI on admission was also not associated with mortality; others have also reported the same [15]. This may be related to an event that can be explained by the fact that COVID-19 patients develop symptoms such as fever, cough, anorexia, diarrhea, and fatigue several days before going to the hospital [16], although no detailed analysis of the type of AKI -pre-renal, renal, or post-renal -was done in our cohort of patients. On the other hand, new onset AKI after admission to ICU in critically ill patients has been reported to be a strong indicator of increased hazard of death even in patients with lower SOFA and APACHE II severity scores on admission [17].
Hypernatremia and multiple co-morbidities, i.e., diabetes, hypertension, and cardiovascular disease, simultaneously were the only independent factor associated with higher odds of death with a trend of significance (p=0.07) in our cohort with AKI in COVID-19-infected individuals. Other studies have reported severity of AKI, age, and various scores of patient illness, i.e., SOFA, APACHE-II, D-dimers, bilirubin, and oliguria in patients on RRT, etc., to be a predictor of mortality in the subgroup of patients with AKI [2,6,7]. An Indian study found age, stage 3 AKI, and the need for mechanical ventilation to be associated with mortality on multivariate analysis [9].
One case series reported six out of 12 patients with COVID-19 and AKI had therapy-resistant hypernatremia, all but one patient had proteinuria, and one in each group was on RRT, indicating renal injury [19]. Another Indian study reported hypernatremia (>145 mEq/L) in 26/111 (23.4%) of patients with AKI and COVID-19 [8].
Hypernatremia generally results from either a deficit of total body water or an inappropriately high sodium input. In our study, a cumulative fluid balance was negatively associated with hypernatremia, which makes hypervolemic hypernatremia unlikely. The high-prevailing temperature (40-42 °C) during summers in India, when the second wave of the pandemic was in progress, with high insensible losses, may have contributed to hypernatremia in the studied cohort, but it also seems less likely because of reports from other parts of the world as well [2,19]. Unphysiological high reabsorption of sodium and chloride through proximal tubular cells mediated via increased angiotensin II activity due to the down-regulation of angiotensin-converting enzyme-2 (ACE-2) receptors by SARS-CoV-2 binding has been proposed as a mechanism [19]. Future studies focusing on hypernatremia evaluation, including measurement of free water clearance, urinary electrolytes, infused electrolyte load, and urine osmolality, may guide the exact pathophysiology of hypernatremia in these patients.
Hypernatremia has been associated with adverse outcomes. When sodium plasma concentrations exceeded 150 mmol/L, the associated mortality rate was as high as 48% [20]. Our study also supports these findings. The median serum sodium (mEq/Lt) (IQR) in our study was 147.0 (139.0, 150.0) in non-survivors. Severe hypernatremia (152 mEq/Lt) was also present in significantly higher numbers in non-survivors as compared to survivors ( Table 2).
Previous studies have found an independent effect of ICU-acquired hypernatremia on mortality after adjusting for other risk factors [21,22]. Hypernatremia and the following hyperosmolar state have a variety of well-known adverse effects on physiologic functions. Therefore, the association between hypernatremia and increased mortality could be causal [20].
Multiple co-morbidities have been associated with the severity of COVID-19 infection. A meta-analysis of 120 studies and 125,446 patients, pooled estimation of all studies, revealed that 28% of COVID-19 cases had multiple co-morbidities (95% CI: 18-40%). It also reported that COVID-19 severity remained similar in patients with at least one co-morbidity (41%) or multiple co-morbidities (44%), substantially higher than in COVID-19 cases without any co-morbidity (19%). However, no association between multiple co-morbidities and mortality was reported [23]. On the other hand, another recent meta-analysis did not report an association between multiple co-morbidities and AKI [24].