Lactate to Albumin Ratio for Predicting Clinical Outcomes after In-Hospital Cardiac Arrest

In-hospital cardiac arrest (IHCA) is associated with high mortality and poor neurological outcomes. Our objective was to assess whether the lactate-to-albumin ratio (LAR) can predict the outcomes in patients after IHCA. We retrospectively screened 75,987 hospitalised patients at a university hospital between 2015 and 2019. The primary endpoint was survival at 30-days. Neurological outcomes were assessed at 30 days using the cerebral performance category scale. 244 patients with IHCA and return of spontaneous circulation (ROSC) were included in this study and divided into quartiles of LAR. Overall, there were no differences in key baseline characteristics or rates of pre-existing comorbidities among the LAR quartiles. Patients with higher LAR had poorer survival after IHCA compared to patients with lower LAR: Q1, 70.4% of the patients; Q2, 50.8% of the patients; Q3, 26.2% of the patients; Q4, 6.6% of the patients (p = 0.001). Across increasing quartiles, the probability of a favourable neurological outcome in patients with ROSC after IHCA decreased: Q1: 49.2% of the patients; Q2: 32.8% of the patients; Q3: 14.7% of the patients; Q4: 3.2% of the patients (p = 0.001). The AUCs for predicting 30-days survival using the LAR were higher as compared to using a single measurement of lactate or albumin. The prognostic performance of LAR was superior to that of a single measurement of lactate or albumin for predicting survival after IHCA.


Introduction
In-hospital cardiac arrest (IHCA) is a major adverse event associated with high mortality and poor neurological outcomes [1][2][3]. Overall, the incidence is high with 1.5-2.8 cases per 1000 admissions in Europe [4]. A significant percentage of patients do not survive resuscitation attempts and fail to achieve stable spontaneous circulation. Moreover, there is also a notable proportion of patients who die during the post-resuscitation phase [5,6]. Only about 15-20% of patients with IHCA survive the event to discharge [6,7]. The postresuscitation phase is characterised by intricate pathophysiological changes that arise from preceding ischemia, reperfusion, and the underlying cause of cardiac arrest. These complex alterations have the potential to trigger multiorgan failure [8].
Accurate prediction of outcomes after IHCA could provide critical information for physicians and relatives and help with appropriate treatment decisions. The timely anticipation of neurological outcomes holds significance in guiding subsequent therapeutic approaches. Deciding whether to continue, restrict, or terminate intensive care therapy presents a significant issue with substantial ethical and socioeconomic implications in

Study Design
In this study, all data were analyzed retrospectively. At the time of submission to the ethics committee, all data had already been collected and were fully available in the hospital information system of the University Hospital Düsseldorf or the respective patient records.
This study was conducted in accordance with the guidelines of the Declaration of Helsinki and was approved by the University of Düsseldorf Committee on Human Research (Study number 2022-2273). The requirement for informed patient consent was waived by the University of Düsseldorf Committee on Human Research due to the use of anonymised data collected during critical care. Whenever it was feasible to reach out to them, every individual who survived provided written informed consent for the utilisation of their anonymous medical data relating to the defined hospitalisation.

Population
We retrospectively screened 75,987 patients who had been hospitalised in the internal medicine or neurological general ward from 2015 to 2019 at the University Hospital of Dusseldorf for IHCA occurrence. Patients who achieved ROSC and had lactate and albumin levels measured within one hour after ROSC were included in this study. After ROSC, the evaluation of lactate and albumin was standardised. Blood sampling was usually performed immediately after reaching ROSC in the intensive care unit, but the subsequent assessments of these variables during the clinical course were conducted at different time points. To eliminate the possibility of sampling time bias, we only included data from the standardised time point in our analysis. Patients who did not have a lactate and/or albumin value measured within the first hour were excluded from the study (Figure 1).
Dusseldorf for IHCA occurrence. Patients who achieved ROSC and had lactate and albumin levels measured within one hour after ROSC were included in this study. After ROSC, the evaluation of lactate and albumin was standardised. Blood sampling was usually performed immediately after reaching ROSC in the intensive care unit, but the subsequent assessments of these variables during the clinical course were conducted at different time points. To eliminate the possibility of sampling time bias, we only included data from the standardised time point in our analysis. Patients who did not have a lactate and/or albumin value measured within the first hour were excluded from the study (Figure 1). Patient characteristics, baseline data, and data related to the events were assessed using patients' charts and medical records.
LAR was calculated as the ratio of lactate (measured in mmol/L) to albumin (measured in g/dL). Depending on the LAR value, patients were divided into 4 quartiles. The first quartile, Q1, included patients with the lowest LAR, and the patients in the fourth quartile, Q4, had the patients with the highest LAR. Accordingly, the study population was divided into four groups (Q1-Q4).
The primary endpoint of this study was 30-day survival. The secondary endpoint was favourable neurological outcomes at 30 days.

Cerebral Performance Category
Neurological outcomes were assessed using the cerebral performance category (CPC) scale, a five-point scale that assesses brain recovery by evaluating functional and cognitive aspects of brain function [19]. The CPC scale was determined by a chart review of the neurological assessment performed by a physician at the time of discharge. In cases of missing chart reviews, a structured interview was conducted. A favourable neurological outcome was defined as a CPC of 1 or 2 at 30 days, and CPC 3 to 5 are considered poor neurologic outcomes or brain death. [19].

Statistical Analysis
Statistical analyses were performed using SPSS statistics, version 28 (IBM, Armonk, NY, USA), MedCalc (Version 20.11, Ostend, Belgium), and GraphPad Prism ® version 7.0 (Graph Pad Software, San Diego, CA, USA). Patient characteristics, baseline data, and data related to the events were assessed using patients' charts and medical records.
LAR was calculated as the ratio of lactate (measured in mmol/L) to albumin (measured in g/dL). Depending on the LAR value, patients were divided into 4 quartiles. The first quartile, Q1, included patients with the lowest LAR, and the patients in the fourth quartile, Q4, had the patients with the highest LAR. Accordingly, the study population was divided into four groups (Q1-Q4).
The primary endpoint of this study was 30-day survival. The secondary endpoint was favourable neurological outcomes at 30 days.

Cerebral Performance Category
Neurological outcomes were assessed using the cerebral performance category (CPC) scale, a five-point scale that assesses brain recovery by evaluating functional and cognitive aspects of brain function [19]. The CPC scale was determined by a chart review of the neurological assessment performed by a physician at the time of discharge. In cases of missing chart reviews, a structured interview was conducted. A favourable neurological outcome was defined as a CPC of 1 or 2 at 30 days, and CPC 3 to 5 are considered poor neurologic outcomes or brain death. [19].
Categorical variables were reported as absolute values and percentages. Continuous data were expressed as median (interquartile range [IQR]). The interquartile range was calculated as the measure of dispersion.
Categorical data were compared using the χ2 test. The D'Agostino & Pearson omnibus normality test was used to assess the normal distribution of continuous variables. Student's unpaired t-test (in the case of a normal distribution) or Mann-Whitney U test (in the case that continuous variables did not follow a normal distribution) were performed to compare the means between the two groups. The 30-day survival was analyzed using the Kaplan-Meier method. The survival rates of all four groups were then compared using logrank tests.
Logistic regression analysis was used to identify predictors of survival to discharge and favourable neurological outcomes after IHCA. Variables with a p-value < 0.10, in the univariate analysis and variables known or assumed to be associated with survival after IHCA were included in the multivariable model. Receiver operating characteristic (ROC) curve analysis was performed to determine the area under the curve (AUC) of the variables for predicting survival. The Delong test was performed to compare the AUC of different variables. The Youden's Index was used to balance the sensitivity and specificity of the prediction of 30-day survival and favourable neurological outcomes. Statistical significance was set at p < 0.05.

Patients' Characteristics
From 2015 to 2019, 75,987 patients were hospitalised in the internal medicine or neurology departments of the University Hospital Dusseldorf, Germany. During this period, 442 patients experienced IHCA. A total of 327 (74%) of the IHCA patients achieved ROSC. Of these patients, 83 were excluded from the study due to incomplete data sets.
A total of 244 patients were included in this study, divided into quartiles of LAR Overall, there were no differences in key baseline characteristics, rates of preexisting comorbidities, and Charlson comorbidity index among the LAR quartiles ( Table 1).
Most patients suffered from preexisting cardiovascular disease with no differences in all groups. CAD was present in 61% of the Q1 group, 77% of the Q2 group, 61% of the Q3 group, and 54% of the Q4 group (p = 0.057) Most patients suffered from arterial hypertension: Q1: 75% of the patients; Q2: 84% of the patients; Q3: 75% of the patients; Q4: 77% of the patients, p = 0.656. There was no difference in the prevalence of liver cirrhosis in all four groups: Q1: 2% of the patients; Q2: 7% of the patients; Q3: 5% of the patients; Q4: 2% of the patients, p = 0.374.
D-dimer, troponin T, and peak NSE levels were higher in the higher LAR quartiles ( Table 2).

The Outcome of Patients after IHCA
In the full cohort, 94 out of 244 patients (38.5%) survived for at least 30 days after IHCA. Patients with a higher LAR had poorer survival after IHCA than those with a lower LAR. The probability of 30-day survival in patients with ROSC after IHCA decreased in a stepwise manner across increasing quartiles: Q1, 43 out of 61 patients (70.4%); Q2, 31 out of 61 patients (50.8%); Q3, 16 out of 61 patients (26.2%); Q4, 4 out of 61 patients (6.6%), (p = 0.001 among the groups). The Kaplan-Meier curve and log-rank (Mantel-Cox) test confirmed lower survival of patients with higher LAR (Figure 2

The Outcome of Patients after IHCA
In the full cohort, 94 out of 244 patients (38.5%) survived for at least 30 days after IHCA. Patients with a higher LAR had poorer survival after IHCA than those with a lower LAR. The probability of 30-day survival in patients with ROSC after IHCA decreased in a stepwise manner across increasing quartiles: Q1, 43 out of 61 patients (70.4%); Q2, 31 out of 61 patients (50.8%); Q3, 16 out of 61 patients (26.2%); Q4, 4 out of 61 patients (6.6%), (p = 0.001 among the groups). The Kaplan-Meier curve and log-rank (Mantel-Cox) test confirmed lower survival of patients with higher LAR (Figure 2).
In addition, multivariate analysis demonstrated that a low LAR (OR 0.734, CI 0.540-0.997, p = 0.048) was an independent predictor of favourable neurological outcomes (Table  A2). In the analysis of multivariate regression for predicting favorable neurological outcomes, the OR for lactate was 0.871 (95% CI 0.765-0.992, p = 0.037) when it was included instead of LAR, while the OR for albumin was 1.262 (95% CI 0.613-2.598, p = 0.527).
The AUCs for predicting 30-day survival using the LAR were higher  A CPC scale of 1 or 2 indicates a good neurological outcome. Patients with a higher LAR had worse neurological outcomes than those with a lower LAR. IHCA = In-hospital cardiac arrest; CPC = Cerebral Performance Category; LAR = Lactate to albumin ratio We performed a regression analysis using the full cohort to identify the predictors of 30-day survival after IHCA. In multivariate analysis, the presence of a shockable primary rhythm, higher haemoglobin level, and lower LAR (Odds ratio [OR] 0.616, 95% confidence interval [CI] 0.422-0.899, p = 0.012) were independent predictors of survival (Table A1). The OR for lactate was 0.822 (95% CI 0.705-0.959, p = 0.013) when it was included instead of LAR. For albumin, the OR was 1.837 (95% CI 0.926-3.643, p = 0.082).
In addition, multivariate analysis demonstrated that a low LAR (OR 0.734, CI 0.540-0.997, p = 0.048) was an independent predictor of favourable neurological outcomes (Table A2). In the analysis of multivariate regression for predicting favorable neurological outcomes, the OR for lactate was 0.871 (95% CI 0.765-0.992, p = 0.037) when it was included instead of LAR, while the OR for albumin was 1.262 (95% CI 0.613-2.598, p = 0.527).
The AUCs for predicting 30-day survival using the LAR were higher  Figure 4B). The AUC for predicting a favorable neurologic outcome using the NSE peak level was 0.706 (CI 0.629-0.776).

Discussion
This study demonstrated that a low LAR immediately after ROSC was associated with increased survival and favourable neurological outcomes. The prognostic performance of the LAR might be superior to that of a single measurement of lactate or albumin for predicting survival or favourable neurologic outcomes after IHCA.
Outcomes after IHCA are affected by patient characteristics [3,[20][21][22], the time point of cardiac arrest, and the time to ROSC [3,21,23,24]. It has been previously demonstrated The AUC for predicting 30-day survival using the LAR was higher as compared to using a single measurement of lactate or albumin. (B) The AUC for predicting favourable neurologic outcomes using the LAR was higher as compared to using a single measurement of lactate or albumin. ROC = Receiver operating characteristics; IHCA = In-hospital cardiac arrest; AUC = Area under the curve; LAR = Lactate to albumin ratio.

Discussion
This study demonstrated that a low LAR immediately after ROSC was associated with increased survival and favourable neurological outcomes. The prognostic performance of the LAR might be superior to that of a single measurement of lactate or albumin for predicting survival or favourable neurologic outcomes after IHCA.
Outcomes after IHCA are affected by patient characteristics [3,[20][21][22], the time point of cardiac arrest, and the time to ROSC [3,21,23,24]. It has been previously demonstrated that patients who are witnessed or monitored at the time of a cardiac arrest have a significantly higher survival rate to hospital discharge than those who are neither monitored nor witnessed [24][25][26].
In the present study, the presence of a shockable primary rhythm, higher haemoglobin level, and lower LAR were independent predictors of 30d survival. The presence of a non-shockable rhythm may reflect the delayed detection of initial shockable rhythms that progress to non-shockable rhythms. There is some evidence obtained from other studies to suggest that lower haemoglobin levels following cardiac arrest are associated with poor outcomes [27,28]. Tissue oxygen delivery is dependent on hemoglobin level, and cerebral oxygenation can be compromised by anaemia, particularly following an acute brain injury [29]. Low hemoglobin levels could exacerbate the effects of oxygen deprivation during cardiac arrest and reperfusion injury. It is already known that hyperoxia after ROSC is associated with increased mortality in both OHCA and IHCA [30,31]. However, in our study, no variation in arterial oxygen partial pressure was observed across all four groups, allowing us to exclude this factor as a potential cause for the observed outcomes. In addition, utilizing capillary refill time (CRT) as a means to evaluate the perfusion status of critically ill patients at the bedside could serve as a valuable tool in guiding resuscitation efforts [32,33]. The impact of CRT on predicting the outcome following cardiac arrest remains uncertain. Unfortunately, our study lacks information regarding the CRT of the patients, which prevents us from making any assertions about a potential association between LAR and CRT.
Accurate prediction of survival and favourable neurological outcomes after IHCA could provide critical information for physicians and relatives and could help with appropriate treatment decisions. Prediction tools such as the Good Outcome Following Attempted Resuscitation (GO FAR 2) score provide clinicians with a prognostic estimate of the likelihood of a good neurological outcome after IHCA based on pre-arrest patient factors [34,35]. However, calculating the GO-FAR score requires more than ten variables, and it has shown the AUC to predict good or moderate neurological outcomes after an OHCA of 0.7 [34,36].
In contrast, the LAR is easy to determine because only two parameters are required. Kong et al. analysed the prognostic value of LAR in patients after out-of-hospital cardiac arrest. They demonstrated that LAR had a better predictive performance than lactate alone in terms of favourable neurological outcomes and survival after OHCA [16]. Here, we demonstrated, for the first time, the prognostic value of LAR for predicting clinical outcomes in patients with in-hospital cardiac arrest. In the present study, the AUC for predicting survival after IHCA was 0.83, and 0.77 for predicting good neurological outcomes. Thus, LAR was superior in predicting survival and favourable neurological outcomes after IHCA than using a single measurement of lactate or albumin.
Increased lactate levels are associated with mortality and are widely used for early diagnosis, management, and risk stratification in critically ill patients [37]. High levels of lactate are also associated with lower survival after IHCA [12,38] and, after a cardiac arrest, are related to tissue hypoperfusion and ischaemia-reperfusion [38]. However, lactate levels can be affected by several different conditions, including decreased lactate elimination due to hepatic or renal dysfunction as well as accelerated glycolysis, and the diagnostic value of the initial lactate level alone might be low [39][40][41].
Low serum albumin is significantly associated with mortality and morbidity in critically ill patients and after IHCA [15,42,43]. Human serum albumin possesses protective effects through anti-inflammatory properties and reduction of ischaemia-reperfusion injury, both of which might be important to counterbalance the post-resuscitation inflammatory response [44][45][46][47]. Similar to lactate, serum albumin levels are affected by multiple conditions, including inflammation, malnutrition, and liver cirrhosis [48]. The decreased level of albumin may be due to albumin loss in the urine in renal diseases/protein-losing gastropathy, decreased hepatic albumin synthesis, or protein malnutrition dysfunction [49][50][51].
The comprehensive combination of lactate and albumin analysis can provide potential benefits for predicting the prognosis of patients with ROSC after cardiac arrest using the ratio of inverse change due to the different mechanisms of the two independent predictors [52]. The combination of lactate as an expression for hypoxia and possibly the no-flow time after cardiac arrest and albumin as an expression for the status of the critically ill patient seems to have a predictive value for in-hospital mortality. Our study did not find any significant differences in the incidence of liver cirrhosis, renal function, or medical conditions assessed by the Charlson comorbidity score among the groups. Nevertheless, we believe that the patients varied in terms of their nutritional status and inflammation level, and the LAR measurement serves as an appropriate surrogate marker for these factors and can differentiate patients in terms of intrahospital survival and neurological outcome.
Routine monitoring of albumin and lactate levels after IHCA should be established in the clinical routine, as it can support risk stratification in post-resuscitation care.
There were limitations in this study. The present study was a single-centre study that included patients in an internal medicine or neurology non-intensive care unit ward; therefore, the interpretation of the results may not be generalised and transferred to other medical disciplines. In addition, the limited number of patients included in the study constrained the multivariate analysis.
In conclusion, the prognostic performance of LAR might be superior to that of a single measurement of lactate or albumin for predicting 30-day survival or favourable neurologic outcomes after IHCA.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of the Heinrich-Heine-University Düsseldorf (protocol code 2022-2273).

Informed Consent Statement:
The requirement for informed patient consent was waived by the University of Düsseldorf Committee on Human Research due to the use of anonymized data collected during critical care. Whenever it was feasible to reach out to them, every individual who survived provided written informed consent for the utilization of their anonymous medical data relating to the defined hospitalization.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to ethical reasons.

Conflicts of Interest:
The authors declare no conflict of interest.