The Association Between Serum Lactate Dehydrogenase Level and in-hospital Death Due to Pulmonary Embolism


 Background: There are few studies evaluating the prognostic value of lactate dehydrogenase (LDH) in patients with pulmonary embolism (PE). We analyzed possible power of serum LDH level to predict in-hospital mortality. Methods: In this cross-sectional study 217 patients with confirmed PE diagnosis with CT angiography and available serum LDH level at first 24-hours upon admission were included. Results: The mean age of patients were 63.04±16.81 years old, 23 patients (10.6%) died during hospitalization. Multivariate analysis showed that only LDH, WBC were independent predictors of in-hospital mortality, however this association was not significant. Conclusions: In patients with pulmonary embolism, LDH can be a good prognostic marker for predicting in-hospital death.


Introduction
Pulmonary embolism (PE), together with deep vein thrombosis (DVT) termed as venous thromboembolism (VTE), can be a life-threatening condition. So, accurate and immediate diagnosis is critical. [1] Validated Wells [2] and revised Geneva rules [3], which categorize patients according to the pretest probability of PE are frequently used as clinical tools to help diagnosis of PE. In a non-high risk patient based on one of these criteria, D-dimer level below 500 µg/L can rule out the PE diagnosis with high con dence in about 20-30% of patients without the need for additional imaging. [4] On the other hand, because of limited number of intensive care unit (ICU) beds in many health centers and because of high cost of ICU admissions, nding high risk patients and risk strati cation of patients are very important. Although there are some scoring systems such as pulmonary embolism severity index (PESI) to de ne high risk PE patients, still we need more investigations to nd biomarkers which have association with higher mortality to extend our knowledge about pathophysiology of disease.
LDH is an enzyme found in almost all cells of body. It has ve different subtypes with different distribution in tissues and releases in the bloodstream with cell and/or tissue injuries. [5].
According to previous studies, signi cant association has been reported between serum LDH level and pump thrombosis in patients who have left ventricular assist device (LVAD). [6, 7] LDH-1 isoenzyme especially showed high relation to pump thrombosis.
[8] Signi cant correlation between elevated LDH level and some other thrombotic events has been demonstrated. In a study about predictors of splanchnic vein thrombosis, LDH < 500U/L was found to be associated with thrombosis. [9] In another study, an association was seen between elevated LDH level -as a marker of hemolysis-and thrombosis risk in paroxysmal nocturnal hemoglobinuria patients. [10,11] Because LDH has association with tissue damage and is involved in anaerobic metabolism of glucose in decreased oxygen supply, theoretically its increased levels could predict adverse outcome in patients with severe pulmonary thromboembolism. [12] Elevated serum LDH also was identi ed as an independent risk factor for venous thromboembolism in patients with testicular germ cell tumor undergoing chemotherapy. [11] Higher serum LDH level has been reported in PE patients. [13] Based on another study, no association was found between LDH level and bleeding or thrombosis, but higher LDH level on ICU admission was shown to be signi cantly associated with increased 7-day and 30-day mortality. [14] Importance of LDH as a marker of severity in PE has been stated in some studies [15], while others has been reported no signi cant association between PE and LDH level.
[16] Finding new biomarkers can help us to achieve better risk strati cation and treatment strategies in order to reduce mortality of PE patients.
In current study, we aimed to nd out the possible association of serum LDH level and in hospital mortality of PE patients.

Methods
In this cross-sectional study 217 patients admitted between 2018 till 2020 in two tertiary hospitals with acute PE were included. Our inclusion criteria included as hospitalized patients aged over 18 years, con rmed PE diagnosis with CT angiography and available serum LDH level at rst 24-hours upon admission.
Our exclusion criteria were hepatic and renal diseases, pregnancy, hemolytic disorders, left ventricular infarction, recent stroke, positive history of active cancer, acute and chronic infections and reticuloendothelial-related diseases were excluded. Segmental and sub-segmental PE was treated with anticoagulant while massive PE de ned as PE with systolic blood pressure less than 90 or/and existence of thrombus in left or right or main pulmonary artery took thrombolytic therapy. [17] Pulmonary computed tomography (CT) angiogram lms were reported by two expert radiologists.
Diagnosis of PE was made by CTPA (Siemens 32-slice computed tomography scanners). Two expert radiologists were investigated CTPA images as blinded fashion.
This study was approved by ethics committee of our University. All patients had signed informed consent form and patient anonymity was preserved in our study.
Any death during hospital course due to PE was de ned as in-hospital mortality. When death occurred due to non-PE causes (e.g. myocardial infarction, intracranial or gastrointestinal bleeding), patients were excluded from study. Overlay one patient had intracranial bleeding after brinolytic therapy and was excluded.
We measured simpli ed Pulmonary Embolism Severity Index (sPESI) value for all the patients. Factors including age over 80 years, positive history of cancer, heart rate below 110 beats/minute, chronic cardiopulmonary disease, systolic blood pressure less than 100 mm Hg, and oxyhemoglobin saturation less than 90% were assessed in this scoring system and each variable has one point. The patient will categorized as high risk even with presence of one point. [18] Information about demographic characteristics of the patients, past medical history as well as presenting vital sings, laboratory variables and oxygen saturation, were collected from their medical records.
Hypertension was de ned as SBP ≥ 140 mmHg or DBP ≥ 90 mmHg. [19] Diabetes mellitus was de ned as fasting plasma glucose levels of ≥ 126 mg/dl and HbA1c ≥ 6.5%. [20] Simpli ed pulmonary embolism severity index was calculated according to previous studies. [18] Right ventricular dysfunction was de ned as the presence of right ventricular dilatation and a TAPSE less than 16 mm in echocardiography ndings. [21] Every ECG was reported by two expert cardiologists to nd out right ventricular strain pattern (inverted T wave in V1-V3).
Statistical analysis IBM SPSS V.22 software was used for statistical analysis (IBM Corp., Armonk, NY, USA). We used t-test for quantitative values and chi-squire test for qualitative variables. Multiple linear regression and ROC (receiver operating characteristics) curve were used to nd cutoff value for LDH level and mortality. Univariate and multivariate analyses were employed to analyze risk factors for mortality

Results
In this cross-sectional study, we included 217 patients with de nite diagnosis of pulmonary embolism.
The mean age of patients was 63.04 ± 16.81 years, 98 patients (45.2%) were female. During hospital admission 23 patients (10.6%) died. Past medical history showed that 40 patients (18.4%) had diabetes mellitus, 78 patients (35.9%) had hypertension, 31 patients (14.3%) had history of smoking. Pulmonary embolism was con rmed in all cases by computed tomography (CT) angiography. Table 1 shows demographic, laboratory and physical exam ndings in patients with pulmonary embolism according to their in-hospital mortality. Table 2 shows association between LDH and other variables. Univariate analysis showed that among laboratory data ndings, higher levels of LDH, white blood cells (WBC), red distribution width (RDW) had signi cant association with in-hospital mortality. (P values < 0.05). (Table 2) only LDH, WBC were independent predictors of in-hospital mortality, however this association was not signi cant statistically (Table 2). ROC curve showed that an LDH cut-off value of 515 U/l had a sensitivity of 91.3% and speci city of 45.9% in predicting in-hospital mortality (95% con dence interval = 0.636-0.761, p = 0.0003) (Fig. 1). Table 3 shows the association between LDH and other variables.

Discussion
In this cross-sectional study, we evaluated the level of lactate dehydrogenase in 217 patients with a de nite diagnosis of pulmonary embolism. Our study showed that serum LDH was associated with higher risk of in-hospital death, but this association was not signi cant in multivariate analysis. By studying new biomarkers affecting outcome of patients and better understanding of pathophysiology of disease, early and more effective treatment with lesser cost could be achieved.
Increased LDH has been linked to higher risk of ARDS [22], in ICU complications [23], and death. [22,24] Pulmonary thromboembolism is one of the most dangerous complications involving cardiovascular system. [25] The importance of evaluating and predicting the course and outcome of the disease has been an era of interest for researchers. There are a few studies investigating the association between LDH level in PE patients and in-hospital mortality.
In line with our study, Leite et al in a retrospective study included 165 patients with acute PE. The main end point of this study were in-hospital and all-cause mortality. They showed that LDH had signi cant association with in-hospital and late all-cause mortality and LDH cut-off value of 310 U/l with a sensitivity of 54.5% and speci city of 71.3% could predict adverse outcome.
[26] Our study by a larger sample size, showed that a LDH cut-off value of 515 U/l had a sensitivity of 91.3% and speci city of 45.9% in predicting in-hospital mortality.
Serum LDH level has been reported to be higher in massive PTE compared to sub-massive and nonmassive PTE. [27,28] The increased LDH level was associated with higher pulmonary artery pressure, right ventricular dysfunction. [27] Lactate dehydrogenase is abundantly made in the human body.It has 5 types of isozymes, LDH-1 and LDH-3 isozymes are presented in cardiomyocytes and pneumocytes respectively. [29,30] Karlsson et al showed that LDH had signi cant correlation with hypoxic ischemic encephalopathy in newborn infants. [31] By catalyzing pyruvate to lactate, LDH is an important enzyme in anaerobic metabolism of glucose during hypoxia. [12] A recent study showed that patients with COVID-19 and high LDH levels are more susceptible to develop acute respiratory distress syndrome. [32] Increasing in cardiac, lung and hypoxic tissue damage makes LDH a suitable biomarker for predicting outcome of patients with pulmonary embolism. Ben et al suggested that using LDH-3 and D-dimer together could improve the diagnosis of PE.
[33] Our study showed that presence of massive embolism and higher sPESI were better predictors of in-hospital death, but higher LDH and WBC also could help in better differentiation of patients. Further studies by revealing pathophysiology of underlying causes of LDH related morbidity and mortality could improve patients' management and outcome.

Limitations of study
We didn't have autopsy for all deaths and pure PE related death could be misdiagnosed in few cases.
Although we used our exclusion criteria to decrease the effect of confounding factors, still other factors such as undiagnosed cancers and some medications could affect our results adversely.

Abbreviations
The study was designed by S.G. and R.H., data collection and manuscript written by T.M., K.M., M.M., S.G., H.K., and A.S. interpreted the data, and R.H. and A.S. revised the manuscript for important intellectual content. All authors read and approved the nal manuscript 33. Ben SQ, Ni SS, Shen HH, Shi YX, Huang SB, Xu JH, Huang JF: The dynamic changes of LDH isoenzyme 3 and D-dimer following pulmonary thromboembolism in canine. Thromb Res 2007, 120(4):575-583. Figure 1 A ROC curve (receiver operating characteristic curve) shows the best cut off point for LDH to predict hospital death.