Blood lactate level as a predictor of patients’ outcome at the Respiratory Intensive Care Unit of Zagazig University Hospitals

Background Many variables measured in critically ill patients have been used to predict patient outcomes; however, it is unlikely that one measurement can replace all, but lactate levels may come close. Aim This study aims to evaluate the role of blood lactate level as a predictor of patients’ outcome at Respiratory Intensive Care Unit at Zagazig University Hospitals (RICU-ZUH). Patients and methods A prospective cohort study was conducted on 52 patients recruited from RICU-ZUH. All patients’ functional conditions were assessed on admission by the Simplified Acute Physiology Score II scoring system, Glasgow Coma Scale (GCS), and Sequential Organ Failure Assessment scores as well as assessment of sepsis. The blood lactic acid level was measured at H0 (initial blood lactate level), H6, H12, H24, and H48 (in mmol/l). Patients were classified into two groups: (i) normal blood lactate level group, and (ii) hyperlactatemia group. Lactate clearance and lactime were also measured. Results Out of the 52 cases studied, hyperlactatemia was present in 30 (57.6%) patients, whereas a normal blood lactate level was found in 22 (42.4%) patients. H0 was significantly high (P<0.01) in the hyperlactatemia group (4.41±1.69 mmol/l), with lactime 42.4±10.5 h; also, lactate clearance at H6 was nonsignificantly high (P>0.05). The Simplified Acute Physiology Score II and Sequential Organ Failure Assessment scores were significantly positively correlated with H0 and lactime, but significantly negatively correlated with lactate clearance, whereas GCS was negatively correlated in a significant way with H0 and lactime and positively correlated with lactate clearance. The significant highest mortality risk of 2.86 was reported with lactime more than 48 h, followed by a 2.28 risk of mortality with H0 blood lactate level more than 3.9 mmol/l, and the least risk was reported with GCS less than 10.5. Conclusion Hyperlactatemia at admission and a prolonged lactime are valuable independent predictors of mortality of RICU patients.


Introduction
Many variables measured in critically ill patients have been used to predict patient outcomes; however, it is unlikely that one measurement can replace all, but lactate levels may come close [1]. It is frequently measured in critically ill patients to detect tissue hypoxia [2]. Furthermore, microcirculatory processes hampering oxygen utilization at the tissue level may also increase lactate level [2,3]. The correlation between lactate levels in arterial and venous blood was found to be acceptable [4]. Should we routinely monitor lactate in the critically ill? And if so, when should we measure it? What would be the best predictors of patient outcome? Accordingly, the research question was, would monitoring of blood lactate predict patients' outcome in the respiratory intensive care unit (RICU)? We hypothesize that blood lactate level, clearance, and lactime predict patients' outcome in the RICU.
The aim of this work was to assess the role of blood lactate level as a predictor of patients' outcome at Respiratory Intensive Care Unit at Zagazig University Hospitals (RICU-ZUH).

Objectives
(1) To identify blood lactate level at patients' admission. (2) To monitor lactime and lactate clearance through RICU stay. (3) To relate previous lactate measures to patients' outcome to determine best predictors of mortality.

Patients and methods
This was a prospective cohort study and was conducted at the RICU of the Chest Department of Zagazig University Hospitals during the period from October 2014 to May 2015 after being approved by the Zagazig University Institutional Review Board.

Study population
Fifty-two patients (selected sample) were enrolled in the study from cases that were admitted to the RICU.
They included 33 male and 19 female patients, with an age range of 23-76 years and a mean±SD of 52.4±14.6 years.

Inclusion criteria
Patients aged 18 years or more, who had primary respiratory diseases with or without respiratory failure were included in the study.

Exclusion criteria
Patients with known chronic organ failure − for example, liver cell failure, renal failure, heart failure, pregnancy, recent trauma, diabetic ketoacidosis, and patients on chemotherapy and/or radiotherapy were excluded.
Informed consent was obtained either from the patients or their relatives. All patients were subjected to the following: (1) Initial diagnosis which was based on the following: medical history, clinical examination (general and local chest examination), radiological chest assessment [plain radiograph posteroanterior view and chest computed tomography when needed], and laboratory assessment (complete blood count (CBC), random blood sugar (RBS), kidney function test (KFT), liver function test (LFT), and arterial blood gases). (2) Assessment of patients' functional condition at admission by calculating the following scores: (a) The Simplified Acute Physiology Score II (SAPS II) scoring system for assessing severity of the disease [5]. (b) The Glasgow Coma Scale (GCS) for assessing the neurological condition of the patients [6]. (c) The Sequential Organ Failure Assessment (SOFA) score for assessing organ failure [7]. (3) Assessment for the presence of sepsis: sepsis is suspected infection in the presence of two or more systemic inflammatory response syndrome criteria [8,9]: (a) Body temperature less than 36°C or more than 38°C. (b) Heart rate more than 90 beats/min. (c) Tachypnea with more than 20 breaths/min or PaCO 2 less than 32 mmHg. (d) White blood cell count less than 4000 cells/mm 3 or more than 12 000 cells/mm 3 , or the presence of more than 10% immature neutrophils (band forms). Band forms more than 3% is called bandemia or a 'left-shift'.  According to the type of data, the following tests were used to test differences in significance; differences between frequencies (qualitative variables) and percentages in groups were compared by the χ 2test. Differences between means (quantitative variables) in parametric multiple groups were compared by the analysis of variance test, in parametric two groups by the t-test and in nonparametric variables by the Mann-Whitney test.
Receiver operating characteristic curves were used for the assessment of sensitivity and specificity and cutoff, and logistic regression for independent predictors. P value was set at less than 0.05 for significant results, at less than 0.001 for highly significant results, and at less than 0.05 for nonsignificant results.
The most sensitive variables were the SOFA score, GCS, lactime, and lactate clearance at H6 (91.7%) and the most specific variables were the initial lactate level (H0) (95%) and SAPS II (92.5%); also, Table 6 showed the superiority of lactate over studied scores in mortality prediction such as area under the curve (AUC), which was 0.93 for initial lactate (H0) against 0.85 for SOFA, 0.91 for SAPS II, and 0.44 for GCS. The significant (P<0.05) highest mortality risk (odds ratio=2.869) was reported with lactime cutoff value more than 48, followed by the risk of mortality with initial blood lactate level (H0) cutoff value more than 3.9 mm/l (odds ratio=2.28), and the least risk reported was with GCS cutoff value less than 10.5 (odds ratio=0.109). Accordingly, the best independent predictors of mortality in RICU-ZUH were initial blood lactate level (H0) more than 3.9, lactime more than 48 h, and GCS less than 10.5.

Discussion
To understand the importance of an increased lactate level, it is important not only to consider anaerobic production, but also aerobic mechanisms and changes in lactate clearance. Despite this complex evaluation, increased lactate levels usually reflect increased morbidity and high mortality [1]. Physicians need accurate and very early indications concerning the prognosis, and both initial blood lactate levels and early lactate clearance were appropriate tools [12]. In trauma patients, alcohol or drug use, which are frequently encountered conditions, do not modify the predictive accuracy of initial blood lactate levels [13]. Blood lactate level can be measured using various devices (central laboratory, point-of-care blood gas analyzers, hand-held devices), and generally most devices used at the bed side have acceptable limits of agreement   compared with laboratory devices [14,15]. In addition, the sampling site of the blood (arterial, venous, capillary, etc.) also does not seem to affect the results much [16,17]. The current study revealed that the initial blood lactate level (H0) was highly significantly (P<0.01) elevated in the hyperlactatemia group, which concurs with the study of Régnier et al. [12] and van Beest et al. [18], who worked on a heterogeneous ICU population including cases with respiratory failure as well as Zhang et al. [19] who used initial lactate (H0) with normalization time (lactate clearance) to assess the predictive value of lactate in unselected critically ill patients. In contrast, the lactime in the hyperlactatemia group (P>0.05) as well as lactate clearance at H6 were nonsignificant. The current study selected H6 based on the study of Nguyen et al. [20] and van Beest et al. [18] who found that normalization of lactate less than 6 h (lactate clearance at H6) after ICU admission revealed better survival compared with normalization of lactate at greater than 6 h; moreover, Zhang et al. [19] stated that most studies defined lactate clearance as the relative reduction in serum lactate levels within 6 h. Also, the study of Marty et al. [11] showed that there is a significant difference between the H0 lactate value and H6, H12, or H24 lactate values in the survivor group. No significant difference was found in the nonsurvivor group between H6, H12, and H24 lactate values compared with H0. Hyperlactatemia in the current study was reported to be associated significantly with increased severity of disease as assessed by SAPS II, and was highly significantly positively correlated with H0 and significantly positively correlated with lactate time, but highly significantly negatively correlated with lactate clearance as well as SOFA scores, where it was highly significantly positively correlated with H0 and significantly positively correlated with lactate time, but highly significantly negatively correlated with lactate clearance, whereas GCS was negatively correlated with H0 and lactate time and positively correlated with lactate clearance; these findings concurred those of Zhang et al. [19] about the predictive value of lactate in critically ill patients, where nonsurvivors had higher SAPS II SOFA scores than survivors, and there was a significant correlation between H0 and lactime with the SOFA score and the SAPS II score, whereas Yolbaşet al. [21] showed that there is a considerably significance difference between lactate level and GCS in a study about the relationship between different lactate levels and mortality rate. The current study reported 100% survival rate in the normal lactate group, whereas it was 60% in the hyperlactatemia group, and this is nearly matched with the results of van Beest et al. [18] where hospital mortality of their population was 18% and was higher in patients with hyperlactatemia during ICU stay compared with those without hyperlactatemia. Also, in the Pasha et al.'s [22] study there were 25% nonsurvivors out of 125 patients, and was higher in patients with hyperlactatemia during ICU stay when compared with those without hyperlactatemia. The current study determined that the most sensitive variables as predictors of mortality were the SOFA score, GCS, lactime, and lactate clearance at H6 (91.7%) and the most specific variables were the initial lactate level (H0) (95%) and SAPS II (92.5%); also, the superiority of lactate over studied scores in mortality prediction such as AUC was 0.93 for initial lactate (H0) against 0.85 for SOFA, 0.91 for SAPS II, and 0.44 for GCS. This is matched with the results of Bakker et al. [10] where among patients with septic shock, the lactime was the best predictor of outcome in a multiple regression analysis. Similarly, in trauma patients, lactate normalization within 24 h was associated with 100% survival [2]. Further data showed that a decrease in lactate of at least 10% during the first 6 h of septic shock was associated with an improved outcome, and an 11% decrease in mortality was observed with each 10% increase in lactate clearance [20]. Badreldin et al. [23] compared the predictive value of lactate with complex physiological scores in a cohort of cardiothoracic surgery patients, where they found that the diagnostic performance of lactate was significantly superior to these scores, as reflected by an AUC of 0.88 for lactate against 0.83, 0.79, and 0.76 for SOFA, SAPS II, and APACHE II, respectively, which are in accordance with the current study. The current study reported that by multivariate logistic regression analysis, significant independent predictors of mortality were initial lactate (H0) more than 3.9 mmol/l with odds ratio 2.28, lactime (>48 h) with odds ratio 2.86, and GCS (<10.5) with odds ratio 0.109. In agreement with the current study, Maarslet et al. [24] found that an increased initial lactate (H0) more than 4.5 mmol/l resulted in an odds ratio of 8.4 (95% confidence interval: 1.5-46.1) for mortality and this is also confirmed in the study of Hajjar et al. [25] in which the lactate level measured at 6 h after ICU admission was also found to be an independent predictor of complications after major cardiothoracic surgery, whereas the study of Régnier et al. [12] showed that the best cutoffs for initial blood lactate (H0) level was 4.7 mm/l (95% confidence interval: 3.4-5.6). Although the normalization time used for predicting mortality commonly ranged from 6 to 72 h, Marty et al. [11], Puskarich et al. [26], and Zhang et al. [19] showed that a normalization time (lactate clearance) within 100 h was positively associated with the hazard. This difference between cutoffs values in different studies can be explained by the differences in sample size, excluded patients, and the specialty of the study. Our study showed that lactate clearance did not have the same value as H0 and lactime in predicting outcome. In contrast, Nguyen et al. [20] reported that lactate clearance was an independent predictor of mortality. According to a recent study on patients with hyperlactatemia in ICU admission, lactate-guided therapy significantly reduced hospital fatal outcomes, which suggests that initial lactate monitoring has much clinical benefit [27]. In conclusion, initial blood lactate levels and lactime are valuable prognostic tools for the prediction of RICU outcomes.
We can recommend the measurement of blood lactate level on RICU admission, and monitoring lactime and lactate clearance during ICU stay for (a) prediction of outcome and (b) application of lactate-guided therapy.

Financial support and sponsorship
Nil.

Conflicts of interest
There is no conflict of interest.