Splenectomy is performed for treatment of the spleen which is mostly injured by a trauma among a variety of conditions [1-3]. Hematologic squeal following a splenectomy includes the appearance of nucleated red blood cells (RBCs), target cells, leukocytosis, and thrombocytosis [4-6].  Several studies are determining the hematologic characteristics in pre and post-splenectomy, but various aspects endure unwell comprehended. Especially, the preoperative and postoperative data considering the temporal behavior of differential white blood cell (WBC) lines, the counts and indices of erythrocyte are still limited [7].  There are some studies assessed the outcomes of splenectomy on the neutrophil-to-lymphocyte ratio (NLR), a simple marker of systemic inflammation in patients undergone splenectomy [1, 8-11]. Nevertheless, there aren’t any announced records measuring the function of the hematological indexes comparing with the clinical parameters, in predicting the outcomes of splenectomy in trauma patients. Patients following splenectomy have possibilities for pulmonary complications such as pneumonia or pleural effusion. Currently, the non-infectious complications such as thromboembolic incidents, pulmonary embolism or chronic thromboembolic pulmonary hypertension have gained an importance [12]. 

Since there is a limited number of studies showing the cardiac functions of the splenectomy patients during postoperative follow-up, the complex pathophysiology of those pulmonary settings hasn’t been obviously implied. Hence, in the current analysis, we intended to investigate the cardiac functions of the trauma patients underwent the emergency splenectomy by correlating the outcomes with the hematological variations to highlight the possible cardio-pathogenesis of these patients and to ascertain the need for echocardiographic screening during the management and follow-up.

Patients

A total of 112 patients who underwent an emergency splenectomy due to trauma (gun shut, stab wound, trauma, vehicle accident, etc.) and applied in XXXX Research and Training Hospital between January 2014 and December 2019 were involved in the study and analyzed retrospectively. Ethical approval from the institutional review of the board was obtained for this cohort study. All procedures were inconsistent by the ethics of the Helsinki Declaration. The patient consent could not be received from the patients because of the retrospective assignment of the study. Two cardiologists blinded to each other performed the echocardiographic screening on patients after discharge during the follow-up period of outpatient clinic. The demographical, radiological and per-operative clinical findings and hematological parameters at the admission and discharge were evaluated. Blood glucose, urea, amylase, hemoglobin (Hb), C-reactive protein (CRP), aspartate aminotransferase (AST), alanine aminotransferase (ALT), platelet count (PLT) levels, and neutrophil/lymphocyte ratio (NLR) and red cell distribution width (RDW) were routinely measured by laboratory tests at the admission and discharge time of patients. The reference range of glucose, urea, amylase, Hb, CRP, AST, ALT, PLT levels were 74-106 mg/dL, 17-43 mg/dL, 28-100 U/L, 11.9-15.4 g/dL, 0-5 mg/L, 1-50 U/L, 1-50 U/L and 151-304 x 10³/µL, respectively. The reference range of NLR and RDW was 0.78-3.58 and 12.3%-14.3% (AU680 ©Beckman Coulter, Inc., USA) [13]. 

The inclusion criteria were being older than 18 years old and having an emergency splenectomy such as car or motorcycle accident, gun shut injury, stab wound injury and blunt trauma. Patients who are younger than 18 or had elective surgery and diagnosed with any hematological disease such as Thalassemia major, idiopathic thrombocytopenic purpura and sickle cell disease or hereditary hemolytic anemia were eliminated from the study. Also, patients with missing data and patients without an outpatient clinic follow-up were excluded.

Echocardiography

During follow-up, echocardiography was performed on 25 patients who had a routine outpatient clinical check-up. Echocardiography was performed within an average 20.8 (min-max, 2-48) months of discharge by an Acuson Sequoia C256 ultrasound system. The parasternal long axis and the areas at end-diastole were used to calculate two-dimensional measurements regarding the apical 4- and 2-chamber views; thus, a modified Simpson’s method evaluated the left ventricular ejection fraction (LVEF). The mitral inflow pulsed-wave Doppler measurement evaluated the left ventricular diastolic filling features. The average values of transmitral parameters such as peak early (E) and late (A) transmitral filling velocities, and their ratio (E/A) were processed among the Doppler spectra of 3–5 consecutive cardiac cycles. Other measured parameters were left ventricular end dimension (LVED),  left atrium diameter (LA), interventricular septum thickness (IVSd), posterior wall thickness (PW), aortic root diameter (Ao), right ventricular diameter (RV), right ventricular outflow tract-diameter (RVOT-D), the velocity-time integral across the pulmonary artery (PA-TVI), tricuspid annular plane systolic excursion (TAPSE), right atrial size (RA), pulmonary artery pressure (PAP), right ventricular basal dimension (A4RVB) at the apical four-chamber view and right ventricular mid-level dimension at the apical four-chamber view (A4RVM).

Statistical Analysis

The statistical analysis was performed using GraphPad InStat (Version 3.06, 2003, San Diego, USA) software. The normality of the distribution of variables was examined by using the Kolmogorov Smirnov test. The comparisons of the variables between groups were performed by the Kruskal-Wallis Test (Nonparametric ANOVA), and Dunn's Multiple Comparisons Test as a post-hoc test. Parametric variables were correlated with the Pearson correlation test, and non-parametric variables were correlated with the Spearman correlation test. P <0.05 was considered statistically significant.

Demographics and Clinical Parameters

The mean age of 112 patients (78 male and 34 female) was 42.3 ± 17.7. 97 of the patients (86.6%) underwent a splenectomy while 15 of the patients (13.4%) underwent an additional surgery including adjacent organ resection (intestinal resection or urogynecological organ resection). The mean amount of packed red blood cell (PRBC) solution injected to the patients was 3.01 ± 3.63 lt, and fresh frozen plasma (FFP) was 232.1 ± 290.1 ml (Table 1).

Table 1: Demographics and clinical parameters of the patients underwent an emergency splenectomy

1 Adjacent organ resection including intestinal resection or uro-gynocological organ resection.  FFP: Fresh frozen plasma, CT: Computerized Tomography, GOS: Glasgow Outcome Scale, PRBC: packed red blood cell.

Mean ± SD [Min-Max], 2 For surviving patients

The most common GOS was Score 5 (62.5%) which referred to a good recovery, resumption of normal life despite minor deficits. There was no GOS of Score 2 which referred to persistent vegetative patients with minimal responsiveness. The least common GOS was Score 3 (1.8%) which referred to the severe disability with conscious but disabled and dependent for daily support (Table 1). 

CT-based investigation of splenic injury showed most of the patients (46.4%) had Grade III injury which included subcapsular hematoma covering >50% of surface area or expanding, intraparenchymal hematoma ruptured with ≥a 5-cm diameter or expanding, and laceration with >3-cm parenchymal depth or involving trabecular vessels. The rarest splenic injury (2.7%) was in Grade I which showed a subcapsular hematoma covering <10% of surface area and laceration including a capsular tear with <1-cm parenchymal depth (Table 1).

Forty-five of the patients (40.2%) had liver damage, 3 patients (2.7%) had kidney damage and 2 patients (1.8%) had both of the damages (Table 1).

The most common per-operative or postoperative complication was bleeding, observed in 62.5% of patients. 6 of them (6.3%) died during the splenectomy. The mean duration of hospitalization was 14.7 ± 22.7 days and of follow-up was 20.8 ± 10.98 months for surviving patients. The mortality rate was 21.4% (n=24) (Table 1).

Comparison of the biochemical parameters at admission and discharge

When the mean of biochemical parameters at admission was subtracted from the discharge values, the decreases were detected in the blood levels of NLR, Hb, glucose, and amylase while increases were detected in the blood levels of RDW, PLT, Urea, CRP, AST, and ALT. The increases in blood levels of RDW and urea were significant (p<0.01) and PLT and CRP were also significant compared with the levels in admission (p<0.001). Instead, the quantities of Hb and glucose at the discharge were more likely lower than the admission levels (p<0.001) (Table 2).

Table 2: Comparison of the biochemical parameters of the patients underwent an emergency splenectomy (N = 112)

**P < 0.01 and ***P < 0.001 were considered significant. Mann-Whitney Test

Hb: Hemoglobin NLR: Neutrophil/Lymphocyte ratio, RDW: Red cell distribution width, CRP: C-reactive protein, AST: Aspartate aminotransferase, ALT: Alanine aminotransferase, PLT: Platelet count

Association of the clinical factors with biochemical parameters at admission and difference levels

When the patients were groups according to their clinical parameters (Table 3), the patients given GOS 1 score showed a considerably higher ratio of ∆NLR compared with patients given GOS 5 (p < 0.05). The patients having a splenic injury of grade I showed dramatically higher NLR and ∆Hb than the other grades of splenic injury (p < 0.001 and < 0.05, respectively). The patients with additional organ damage showed significantly lower NLR and ∆NLR compared to the patients without any other organ damage (p < 0.05). The patients hospitalized longer had higher ∆RDW but lower NLR than the patients hospitalized shorter (p < 0.05). There was no significant difference between the patients with or without any morbidity for none of the biochemical parameters. However, ex-patients showed lower NLR (P < 0.05), higher ∆NLR and lower ∆PLT contrary to the survived patients (P<0.01) (Table 3).

Table 3: Comparison of the clinical parameters with biochemical parameters (admission and difference levels) in the patients underwent an emergency splenectomy (N = 112)

¹ P < 0.05 vs GOS 5. ² P < 0.01 and ³P < 0.05 vs Grade I; Non-parametric comparison of two groups: Mann-Whitney Test, parametric multiple comparison: ANOVA + Tukey-Kramer Multiple Comparisons Test, non-parametric multiple comparison: Kruskal-Wallis Test (Nonparametric ANOVA) + Dunn's Multiple Comparisons Test

Hb: Hemoglobin NLR: Neutrophil/Lymphocyte ratio, RDW: Red cell distribution width, PLT: Platelet count

Correlation between the echocardiographic outcomes and biochemical parameters at admission and difference levels 

Among 25 patients examined by echocardiography, three patients (12%) had hypertension, eight (32%) patients were smokers, and two patients (8%) had diabetes mellitus type 2. The mean LDL and GFR levels of all these patients were 138.3±36 mg/dL and 89.5±18 mL/min, respectively (Table 1).

When the echo outcomes of the 25 patients were compared with the gender, none of them showed significant difference regardless of the physiological finding for LVED (Table 4). 

Table 4: Echocardiographic outcomes of the patients (n=25) underwent an emergency splenectomy

*P<0.01 vs female patients

LA: Left atrium diameter, Ao: Aortic root diameter, IVSd: Interventricular septum thickness, LVED: Left ventricular end dimension, PW: Posterior wall thickness, RV: Right ventricular diameter, RVOT-D: Right ventricular outflow tract- diameter, PA-TVI: The velocity-time integral across the pulmonary artery, TAPSE: Tricuspid annular plane systolic excursion, RA: Right atrial size, E/A: the ratio of early filling to mitral peak velocity of late filling, PAP: Pulmonary artery pressure, A4RVB: Right ventricular basal dimension at apical four-chamber view, A4RVM: Right ventricular midlevel dimension at apical four-chamber view

Correlation between the echo outcomes and biochemical parameters showed that there was a substantial positive association, between Ao and ∆PLT and A4RVM and NLR1 (r_s: 0.44, 0.48, respectively; P < 0.05). There were more likely negative correlations among Ao and ∆Hb (r_s: -0.45, P < 0.05); between LVED and both NLR2 (discharge) and ∆NLR (r_s: -0.42, P < 0.05); between RV and both NLR2 and ∆NLR (r_s: -0.42, P < 0.05), between PA-TVI and both RDW2 and ∆RDW (r_s: -0.41 and -0.42, respectively; P < 0.05), PA-TVI and both NLR2 and ∆NLR (r_s: -0.43, P < 0.05); between RA and ∆RDW (r_s: -0.46; P < 0.05) (Table 4). Lastly the values of A4RVB and A4RVM were negatively correlated with ∆NLR (r_s: -0.4, -0.42, -0.45; P < 0.05). Any of the biochemical parameters showed an association for LA, IVSd, PWD, RVOT-D and E/A (Table 5).

Table 5: Correlation between the echocardiographic outcomes and biochemical parameters at admission (1), discharge (2) and odd levels (∆) of the patients (n=25) underwent an emergency splenectomy

Spearman Rank Correlation, P < 0.05. Hb: Hemoglobin NLR: Neutrophil/Lymphocyte ratio, RDW: Red cell distribution width, PLT: Platelet count, LA: Left atrium diameter, Ao: Aortic root diameter, IVSd: Interventricular septum thickness, LVED: Left ventricular end dimension, PW: Posterior wall thickness, RV: Right ventricular diameter, RVOT-D: Right ventricular outflow tract- diameter, PA-TVI: The velocity-time integral across the pulmonary artery, TAPSE: Tricuspid annular plane systolic excursion, RA: Right atrial size, E/A: the ratio of early filling to mitral peak velocity of late filling, PAP: Pulmonary artery pressure, A4RVB: Right ventricular basal dimension at apical four-chamber view, A4RVM: Right ventricular midlevel dimension at apical four-chamber view

Comparison of the biochemical parameters with the echocardiographic outcomes according to the reference levels

Since there were pathological outcomes (E/A, A4RVB, and A4RVM) in the echocardiography of some patients (Table 6), these substances were allocated into two sets as normal and pathological groups. The patients having a lower ratio of E/A showed a significantly higher level of RDW at admission and a higher ratio of NLR at discharge (p< 0.05). The patients with increased size of A4RVB revealed a dramatically larger ∆NLR (p < 0.05). The patients with increased size of A4RVM revealed a considerably higher ratio of NLR at admission and a significant difference in NLR, a lower level of PLT at admission and discharge (P < 0.05) (Table 6).

Table 6: Comparison of the biochemical parameters at admission (1), discharge (2) and the odd levels (∆) with the echocardiographic outcomes according to the reference levels in patients (n=25)

*P < 0.05; Non-parametric comparison of two groups: Mann-Whitney Test     

Hb: Hemoglobin, NLR: Neutrophil/Lymphocyte ratio, RDW: Red cell distribution width, PLT: Platelet count, E/A: the ratio of early filling to mitral peak velocity of late filling, A4RVB: Right ventricular basal dimension at apical four-chamber view, A4RVB: Right ventricular midlevel dimension at apical four-chamber view

Since a large number of trauma patients in Turkey have undergone a splenectomy operation, we currently concentrated on the trauma splenectomy cases. Splenectomy alleviates hypersplenism and the subsequent cardiovascular incidents involving myocardial infarction, stroke, pulmonary hypertension and splenic or portal vein thrombosis have been increased [12, 14-16].  Segraves et al., [17] reported the frequency of splenectomy in porto-pulmonary hypertension, and higher right ventricular systolic pressure (RVSP) was estimated by transthoracic echocardiography in splenectomy group. But splenectomy was associated with different pulmonary hemodynamics. They suggested earlier echocardiography screening in potential patients for porto-pulmonary hypertension earlier in patients following splenectomy [17].  Regarding the echocardiography as an advantageous in the follow-up of the patients for hypertension [16], we investigated the echocardiographs of 25 splenectomized patients by comparing them with the biochemical parameters. Although positive correlations of Ao and ∆PLT, and negative correlations between Ao and ∆Hb were present, the values of A4RVB and A4RVM were negatively correlated with ∆NLR and PLT (Table 5-6). Eventually, these findings have shown great value for future to suspect the patients in the setting of right-sided cardiac dysfunctions following the hematological parameters as a counter effect. 

RDW is one of these hematologic variables which is a quantitative value indicating variation in the size of circulating RBCs and correlated with coronary artery disease, acute myocardial infarction, acute and chronic heart failure, pulmonary hypertension [6].  RDW predicts pulmonary hypertension in patients with chronic obstructive pulmonary disease (COPD) as a prognostic marker [18] & [19].  The splenectomy is supposed to raise the RBC count which is also affected by some baseline levels in pre-splenic injury studies [6] & [20]. The splenectomized patients had notably greater RDW contrary to the patients without splenectomy or who experienced an embolization [6] & [21]. Consistently, we observed higher RDW at the discharge compared with the admission to hospital. The patients with higher grades of splenic injuries showed significant changes in RDW between the time of admission and discharge (Table 2-3). Echocardiographic mitral inflow E/A ratio of < 1 suggest an early diastolic dysfunction. Another important point in this study was that higher RDW1 values are associated with increased early diastolic dysfunction. Besides this, RDW2 and ∆RDW were negatively correlated with PA-TVI (Table 5-6). 

NLR, calculated from the white cell distribution, is a valuable diagnostic and predictive parameter for inflammation and infectious conditions. NLR is as an objective predictive aspect of morbidity and mortality in numerous settings, such cardiovascular diseases [8, 22, 23].  Although the NLR score showed no predictive value in overall survival after splenectomy for hypersplenism, it appeared to have a better predictive value in overall survival in patients who did not develop malignancy [8]. The ex-patients showed a lower score of NLR at admission compared to survived patients, prior to inflammatory process. We reported without any abnormal impact of NLR exceeded the range of 0.78-3.58 in patients based on the absence of any comorbidities and substance abuse. The negative correlations between NLR2 and ∆NLR with echo parameters of LVED, RV, PA-TVI are interesting findings (Table 4-6). Lower micro-inflammation indicated by NLR may allow adaptive enlargement of these chambers and increased flow regarding the platelet regulation and destruction after splenectomy¹¹. Most of the patients' PLT increased significantly following splenectomy and recovered to normal or above normal, respectively [24]. Consistently, PLT levels before and after the splenectomy increased significantly at the discharge from hospital up to approximately two times of admission levels. Moreover, the change in PLT was significantly correlated with the mortality rate, regardless of other clinical parameters such as GOS score, the grade of splenic injury, other organ damages (Table 2-3). Eventually, the patients with lower PLT1 or PLT2 count have larger dimensions of A4RVM which supposed to affect the ventricular functions (Table 5-6). Elevated cell-free hemoglobin and increased scavenging of nitric oxide in patients underwent splenectomy may contribute to pulmonary vasculopathy and pulmonary vascular resistance in pulmonary arterial hypertension [25, 26]. The grade of splenic injury was found to be correlated with the changes in hemoglobin levels at the discharge; the higher grades of splenic injury showed lower levels of Hb.

Nevertheless, this study had some limitations; low sample size, single-centered and the missing data of information for anticoagulant or fibrinolytic therapy. Despite these limitations, the strength of this study was that the cardiac parameters by echocardiography and correlation with hematological variables were evaluated regarding the emergency splenectomy.

Subsequently, the patients who underwent splenectomy remains at a relatively high risk of developing ventricular dysfunctions, and the changes in NLR, RDW, and PLT are associated with ventricular dysfunctions. NLR, RDW, and PLT might be a potential marker in the clinical management of splenectomy patients and echocardiography may be added in the follow-up of these patients to further therapeutic evaluation of cardiac dysfunction. There we need more attention to changes in patients’ outcomes clinically overtime following emergency splenectomy. We need prospective randomized clinical trials and large sample size studies to validate our diagnostic markers for the ventricular dysfunctions following emergency splenectomy.

The authors declare that they have no conflict of interest

No funding sources

The study was approved by the Bagcilar Training and Research Hospital University of Health Science, Istanbul. 

1. Rottenstreich, Amihai, et al. "Thromboembolic events following splenectomy: risk factors, prevention, management and outcomes." World journal of surgery 42 (2018): 675-681. doi:10.1007/s00268-017-4185-2.

2. Andraus, Wellington, et al. "The best approach for splenectomy in portal hypertension." Surgery 149.6 (2011): 853. doi:10.1016/j.surg.2011.02.014.

3. Petroianu, Andy, Vivian Resende, and Rodrigo Gomes Da Silva. "Late follow-up of patients submitted to subtotal splenectomy." International Journal of Surgery 4.3 (2006): 172-178. doi:10.1016/j.ijsu.2005.12.006.

4. De Porto, A. P. N. A., et al. "Assessment of splenic function." European journal of clinical microbiology & infectious diseases 29 (2010): 1465-1473. doi:10.1007/s10096-010-1049-1.

5. William, Basem M., and Gino R. Corazza. "Hyposplenism: a comprehensive review. Part I: basic concepts and causes." Hematology 12.1 (2007): 1-13. doi:10.1080/10245330600938422.

6. Wernick, B., et al. "Temporal changes in hematologic markers after splenectomy, splenic embolization, and observation for trauma." European Journal of Trauma and Emergency Surgery 43 (2017): 399-409. doi:10.1007/s00068-016-0679-0.

7. Khan, Palwasha N., et al. "Postsplenectomy reactive thrombocytosis." Baylor University Medical Center Proceedings. Vol. 22. No. 1. Taylor & Francis, 2009.  doi:10.1080/08998280.2009.11928458

8. Du, Zhaoqing, et al. "Predictive value of the preoperative neutrophil-to-lymphocyte ratio for the development of hepatocellular carcinoma in HBV-associated cirrhotic patients after splenectomy." PLoS One 13.4 (2018): e0195336. doi:10.1371/journal.pone.0195336.

9. Hirakawa, Yusuke, et al. "Immunological consequences following splenectomy in patients with liver cirrhosis." Experimental and therapeutic medicine 18.1 (2019): 848-856.  doi:10.3892/etm.2019.7640.

10. Petroianu, Andy, et al. "Haematological and biochemical characteristics of the splenic effluent blood in schistosomal patients undergoing splenectomy." Revista do Colégio Brasileiro de Cirurgiões 41 (2014): 176-180. doi:10.1590/s0100-69912014000300007.

11. Das, Sudipta Sekhar, Soumya Bhattacharya, and Subrata Sen. "Managing uncontrolled postsplenectomy reactive thrombocytosis in idiopathic thrombocytopenic purpura: role of thrombocytapheresis." Transfusion and Apheresis Science 49.2 (2013): 171-173.  doi:10.1016/j.transci.2012.09.009.

12. Palkar, Atul V., et al. "Post splenectomy related pulmonary hypertension." World journal of respirology 5.2 (2015): 69. doi:10.5320/wjr.v5.i2.69.

13. Forget, Patrice, et al. "What is the normal value of the neutrophil-to-lymphocyte ratio?." BMC research notes 10.1 (2017): 1-4. doi:10.1186/s13104-016-2335-5.

14. Boyer, Thomas D., and Shahid Habib. "Big spleens and hypersplenism: fix it or forget it?." Liver International 35.5 (2015): 1492-1498.  doi:10.1111/liv.12702.

15. Katz, S. C., & Pachter, H. L. (2006). Indications for splenectomy. The American Surgeon, 72(7), 565-580. PMID: 16875077.

16. Rajalingam, Rajesh, et al. "Management of hypersplenism in non-cirrhotic portal hypertension: a surgical series." Hepatobiliary & Pancreatic Diseases International 11.2 (2012): 165-171.  doi:10.1016/s1499-3872(12)60143-x.

17. Segraves, Justin M., et al. "Relationship between portopulmonary hypertension and splenectomy: Mayo Clinic experience and review of published works." Hepatology Research 48.3 (2018): E340-E346. doi:10.1111/hepr.12930.

18. Yang, Jie, et al. "Red blood cell distribution width predicts pulmonary hypertension secondary to chronic obstructive pulmonary disease." Canadian respiratory journal 2019 (2019). doi:10.1155/2019/3853454

19. Hampole, Chetan V., et al. "Usefulness of red cell distribution width as a prognostic marker in pulmonary hypertension." The American journal of cardiology 104.6 (2009): 868-872.  doi:10.1016/j.amjcard.2009.05.016.

20. Pratl, Barbara, et al. "Partial splenic embolization in children with hereditary spherocytosis." European journal of haematology 80.1 (2008): 76-80. doi:10.1111/j.1600-0609.2007.00979.x.

21. Corazza, G. R., et al. "Changes in peripheral blood lymphocytes and immune complexes in splenectomized patients: lack of correlation with residual splenic function." Journal of clinical & laboratory immunology 31.1 (1990): 33-38. https://europepmc.org/article/med/1966982

22. Kahramanca, Şahin, et al. "Neutrophil-to-lymphocyte ratio as a predictor of acute appendicitis." Turkish Journal of Trauma and Emergency Surgery 20.1 (2014): 19-22. doi:10.5505/tjtes.2014.20688.

23. Ishizuka, Mitsuru, Takayuki Shimizu, and Keiichi Kubota. "Neutrophil-to-lymphocyte ratio has a close association with gangrenous appendicitis in patients undergoing appendectomy." International surgery 97.4 (2013): 299-304. doi:10.9738/CC161.1.

24. Guo, Binbin, et al. "Platelet's Rule of Change and Clinical Significance before and after Splenectomy." The American Surgeon 85.11 (2019): 1288-1293. DOI: https://doi.org/10.1177/000313481908501138

25. Ghosh, Kanjaksha, V. Meera, and Farah Jijina. "Pulmonary hypertension in patients with hematological disorders following splenectomy." Indian Journal of Hematology and Blood Transfusion 25 (2009): 45-48. doi:10.1007/s12288-009-0012-8.

26. Brittain, Evan L., et al. "Elevation of plasma cell-free hemoglobin in pulmonary arterial hypertension." Chest 146.6 (2014): 1478-1485. doi:10.1378/chest.14-0809.