https://orcid.org/0000-0002-2145-5587
Figures
Abstract
Background
Previous studies have shown that both age and hemoglobin are closely associated with the prognosis of septic shock. A recent study found that hemoglobin may change with age. Hemoglobin-to-Age Ratio (HAR) takes both age and hemoglobin into consideration as essential factors. So far, the effect of HAR on the prognosis of septic shock is still unclear. This research aimed to investigate the association between the HAR and the prognosis of patients with septic shock.
Methods
Cox proportional hazards regression analysis, restricted cubic spline, Kaplan-Meier survivor analysis and stratified interaction analysis were used to elucidate the relationship between the HAR and prognosis of patients with septic shock.
Results
There is a nonlinear association between the HAR and mortality within 28 days after intensive care unit admission. When the HAR was lower than 0.13, mortality within 28 days after ICU admission decreased significantly as the HAR increased. When the HAR was higher than 0.13, the HAR was not a protective factor for mortality within 28 days after ICU admission. In patients with septic shock, the HAR was more effective in reducing the risk of death in patients with atrial fibrillation than in patients without atrial fibrillation.
Conclusion
There is a nonlinear association between the HAR and mortality within 28 days after intensive care unit admission. When the HAR was at a low level, mortality within 28 days after ICU admission decreased significantly as the HAR increased. When the HAR was at high levels, the HAR might not be a protective factor for mortality within 28 days after ICU admission. In patients with septic shock, the HAR was more effective in reducing the risk of death in patients with atrial fibrillation than in patients without atrial fibrillation.
Citation: Yang C, Xue Y, You Z, Huang T, He X, Jiang X, et al. (2024) Nonlinear relationship between Hemoglobin-to-Age Ratio and all-cause mortality in patients with septic shock: A retrospective cohort study in the MIMIC-IV database. PLoS ONE 19(12): e0313937. https://doi.org/10.1371/journal.pone.0313937
Editor: Ennio Polilli, Pescara General Hospital, ITALY
Received: July 9, 2024; Accepted: November 2, 2024; Published: December 6, 2024
Copyright: © 2024 Yang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This work was sponsored by Fund for youth scientific research, Fujian provincial health technology project (Grant number: 2023QNA001), Startup Fund for scientific research, Fujian Medical University (Grant number: 2020QH1149) and the General Program of Natural Science Foundation of Fujian Province (Grant Number: 2023J011179).
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: SOFA, Sequential Organ Failure Assessment; SIRS, Systemic Inflammatory Response Syndrome; NEWS, National Early warning Score; MEWS, Modified Early warning Score; HAR, Age-to-Hb Ratio; MIMIC-IV, Medical Information Mart for Intensive Care IV; ICU, Intensive care unit; ICD, International Classification of Diseases; SQL, Structured Query Language; AP, Acute pancreatitis; AF, Atrial fibrillation; AMI, Acute myocardial infarction; AKI, Acute kidney injury; WBC, White blood cell; Hb, Hemoglobin; Scr, Serum creatinine; BUN, Blood Urea Nitrogen; PTT, Partial thromboplastin time; INR, International Normalized Ratio; SAPS II, Simplified Acute Physiology Score II; TET, Total erythrocyte transfusion; HR, Hazard ratios; CI, confidence intervals; IQR, Interquartile range; LOS, Length of stays
1 Introduction
Sepsis is a life-threatening organ dysfunction caused by a dysregulated host response to infection [1]. Septic shock (SS) is the most serious complication of sepsis [2]. Epidemiologic findings have shown the mortality of SS remains as high as 30% [3,4]. Previous studies have shown that both age and hemoglobin (Hb) concentration are closely associated with the prognosis of SS [5,6]. A recent study found that hemoglobin concentration may change with age [7]. There are limitations in applying Hb alone to evaluate the prognosis of SS. The Hemoglobin-to-Age Ratio (HAR) takes age and Hb into consideration as essential factors and corrects for the effect of age on hemoglobin. So far, no studies have been reported on the association between the HAR and the prognosis of SS. Our research assessed the role of the HAR in predicting all-cause mortality in patients with SS and provided additional methods for evaluating the prognosis of SS.
2 Materials and methods
2.1 Study population
This study is a retrospective observational study. The data of the study population were all obtained from the MIMIC-IV (version 2.2) database. The MIMIC-IV database consists of high-quality medical records of patients admitted to the Intensive care unit (ICU) of Beth Israel Deaconess Medical Center [8]. One author (Chao Yang) complied with the requirements for access to the database and was responsible for the data extraction.
Inclusion criteria: (1) Patients diagnosed with SS according to the 9th and 10th revisions of the International Classification of Diseases (ICD); (2) Patients requiring ICU admission.
Exclusion criteria: (1) Patients whose age at the time of admission to the ICU was less than 18 years old; (2) Patients whose length of stay in the ICU was less than 24 hours; (3) Patients who lacked data of Hb within 24 hours after ICU admission; (4) For patients with multiple ICU admissions, we only extracted data from their first ICU admission. (5) Patients with bleeding complications during hospitalization. Finally, a total of 5180 patients were enrolled in this study (S1 Raw data). All patients were grouped into four groups based on the quartiles of the HAR, where the first quartile of the HAR was regarded as the reference group (Fig 1).
2.2 Data collection
Potential variables were divided into six main categories: (1) Demographics, including age, gender, and race. (2) Comorbidities, including asthma, diabetes, acute pancreatitis (AP), atrial fibrillation (AF), acute myocardial infarction (AMI), and acute kidney injury (AKI). (3) Laboratory tests within 24 hours after first admission to the ICU, including white blood cells (WBC), Hb, platelets, serum creatinine (Scr), blood urea nitrogen (BUN), International Normalized Ratio (INR), serum potassium, serum sodium, serum chloride and serum total calcium. (4) Disease severity scores within 24 hours after first admission to the ICU, including the Sequential Organ Failure Assessment (SOFA) score, the Simplified Acute Physiology Score II (SAPS II). (5) Total erythrocyte transfusion (TET) within 28 days after admission to the ICU. Variables were excluded if they had more than 20% missing data. Age and hemoglobin were used to calculate the HAR. The formula was HAR = hemoglobin (g/dl)/ Age(year). Follow-up started at the time of first ICU admission and ended on the 28th day after the first ICU admission or when the endpoint events occurred.
2.4 Statistical analysis
Continuous variables were presented as mean ± standard deviation or median (interquartile range). Categorical variables were presented as proportions. The Kolmogorov-Smirnov test was used to assess the normality of continuous variables. Continuous variables that were normally distributed were analyzed by t-test or ANOVA; Continuous variables that were not normally distributed were analyzed by the Kruskal-Wallis test. Qualitative variables were analyzed by chi-square test. The effects of the different levels of the HAR on endpoint events were evaluated using the Kaplan-Meier survivor analysis (log-rank test).
Hazard ratios (HR) and 95% confidence intervals (CI) for the risk of endpoint events that were associated with elevated HAR were estimated using Cox proportional hazards models with adjustment for covariates. Criteria for adjustment covariates: (1) In the univariate analysis, the P-value of the regression coefficient of the covariate on the mortality within 28 days after ICU admission is less than 0.1. (2) Inclusion of covariate in the basic model or exclusion of covariate from the complete model affects regression coefficients of the HAR by more than 10%. (3) We performed correlation analyses for continuous variables and constructed correlation matrix to exclude multicollinearity between variables (Fig 2). We applied scatterplots to illustrate the relationships between some variables (see S1 to S9 Figs). Finally, the covariates that were selected included sex, AF, AMI, AKI stage, BUN, SAPS II, TET. For the selected variables, we performed univariate and multivariate Cox regression.
Restricted cubic spline was used to analyze the association between the HAR and endpoint events. We applied segmented regression (also known as piece-wise regression), which uses a separate line segment to fit each interval. Log-likelihood ratio test comparing one-line (non-segmented) model to segmented regression model was used to determine whether threshold exists. The inflection point which connects the segments was based on maximum likelihood that the model gives, and it was determined using the two-steps recursive method.
Interaction and stratified analyses were conducted based on age (≤65 and >65 years), sex, and comorbidity (asthma, diabetes, AP, AF).
All analyses were performed using the R software (version 4.2.2) and Empower Stats (http://www.empowerstats.com, X&Y Solutions, Inc.). A two-sided significance level of 0.05 was used to evaluate statistical significance.
2.5 Management of missing data
The number of missing data for WBC is 14, the number of missing data for platelets is 23, the number of missing data for Scr is 1, the number of missing data for INR is 162, the number of missing data for serum potassium is 8, the number of missing data for serum sodium is 1, the number of missing data for serum chloride is 2 and the number of missing data for serum total calcium is 16. The number of samples with missing data in this study is 162, only 3.1% of the total sample (S1 File). The number of samples with missing data is very small. So multiple imputation of missing data is not required.
3 Results
The median HAR of all patients included was 0.144 (IQR: 0.045–0.709). Patients were grouped into four groups based on the quartiles of the HAR [quartiles Q1 (0.045–0.12), Q2 (0.12–0.144), Q3 (0.144–0.179), Q4 (0.179–0.709)]. Mortality within 28 days after ICU admission was 34.29% (Table 1).
3.1 Baseline characteristics
Compared with the other three groups, patients in the lowest quartile of the HAR had a higher prevalence of age, TET, Scr, BUN and AF and a lower prevalence of Hb and asthma. Patients with lowest quartiles of the HAR had highest mortality (Table 1).
Survivors had higher HAR, Hb and length of stays (LOS) in the hospital than non-survivors. Survivors had a lower prevalence of BUN, SOFA, TET, LOS in the ICU than non-survivors (Table 2).
3.2 Primary results
The results of Cox regression and Cox proportional hazards modeling showed that the HAR was a protective factor for death within 28 days after ICU admission in both unadjusted (model 1), partially adjusted (model 2) and completely adjusted models (model 3), HRs were 0.31 (95% CI: 0.14, 0.68), 0.31 (95% CI: 0.14, 0.69) and 1.125 (95% CI: 0.14, 0.72), respectively (S2 File, Table 3).
Restricted cubic spline showed a nonlinear association between the HAR and mortality within 28 days after ICU admission, after adjustment for covariates (Fig 3).
Heavy central lines represent the estimated adjusted hazard ratios, with shaded ribbons denoting 95% confidence intervals.
Threshold effect analysis showed that, after adjustment for covariates, when the HAR was lower than 0.13, mortality within 28 days after ICU admission decreased significantly as the HAR increased, and the HR was 0.0 (95% CI: 0.0, 0.1). However, when the HAR was higher than 0.13, the HAR might not be a protective factor for mortality within 28 days after ICU admission (P = 0.235) (Table 4).
The results of the Kaplan-Meier survivor analysis showed that, after adjustment for sex, AF, AMI, AKI stage, BUN, SAPS II, TET, patients with the HAR above 0.13 had a higher cumulative survival within 28 days after ICU admission, compared to patients with the HAR below 0.13 (Fig 4).
Spearman’s rank correlation analysis showed a negative correlation between HAR and mortality within 28 days after ICU admission in the overall patients with SS (P<0.001). The results of segmented correlation analysis showed that, when the HAR was lower than 0.13, a negative correlation between HAR and mortality within 28 days after ICU admission (P = 0.020). When the HAR was lower than 0.13, there was no correlation between HAR and mortality within 28 days after ICU admission (P = 0.9776) (Table 5).
3.3 Subgroup analysis
The interaction analysis revealed that there was a significant interaction between AF and HAR in predicting the risk of death within 28 days after ICU admission. After adjustment for covariates, the effect of the HAR on reducing death within 28 days after ICU admission remained more significant in patients with AF than in patients without AF [HR = 0.038 (95% CI: 0.005, 0.321) vs. HR = 0.464 (95% CI: 0.191, 1.127), P for interaction = 0.029] (Fig 5).
4 Discussion
In our study, we excluded SS patients with bleeding comorbidities and adjusted for the confounding variable of total erythrocyte transfusion (TET) within 28 days after ICU admission. The results of the study showed that the HAR might be a protective factor in reducing all-cause mortality in patients with SS within 28 days after ICU admission. Threshold effect analyses showed that when the HAR was lower than 0.13, mortality within 28 days after ICU admission decreased significantly with increasing HAR; however, when HAR was greater than 0.13, changes in HAR had no effect on mortality within 28 days after ICU admission. This phenomenon suggests that patients with SS of different ages have different hemoglobin requirements. In patients with SS whose HAR was lower than 0.13, the hemoglobin requirement was higher in older patients. However, in patients with SS whose HAR was higher than 0.13, the hemoglobin requirement may no longer be affected by age.
Hb concentration is closely associated with the prognosis of sepsis. Oxygen delivery depends mainly on Hb concentration, cardiac output, and oxygen saturation [9,10]. A decrease in Hb concentration leads to a reduction in oxygen delivery [11]. When the body’s demand for oxygen is greater than oxygen delivery, tissue cells undergo anaerobic glycolysis and produce lactic acid [12,13]. Persistent hyperlactatemia is strongly associated with an elevated risk of death in patients with sepsis [14]. Other studies have shown that anemia may impair cardiac function and induce heart failure [15,16]. Studies have shown that anemia increases mortality in patients with sepsis [11,17]. Previous studies have analyzed the effect of transfusion threshold on mortality in patients with SS. The results showed no significant difference in mortality between patients with SS whose transfusion threshold was 7 g/dL and those whose transfusion threshold was 9 g/dL. Blood transfusion may improve the prognosis of patients with SS only when the hemoglobin concentration is below 7 g/dL [1,12,18,19]. Recent studies have shown that hemoglobin concentration varies with age [7]. Therefore, basic hemoglobin and oxygen delivery levels may vary in patients with SS of different ages. Yang P et al. showed a nonlinear relationship between hemoglobin concentration and the 28-day mortality risk in elderly patients with sepsis. Both hemoglobin concentrations below 10 g/dL and above 15 g/dL increased the 28-day mortality risk in patients with sepsis [20]. Zhu YB et al. found that low levels of hemoglobin concentration were an independent risk factor for poor prognosis of sepsis in pediatric patients with sepsis [21]. The above findings may suggest that the requirement for hemoglobin concentration may not be the same in patients with sepsis of different ages.
Several studies have demonstrated that advanced age is an essential factor contributing to the elevated risk of death in sepsis and SS [5,22–24]. Loue S et al. observed T-cell depletion in septic-aged mice (aged between 20 and 22 months). Also, interleukin 6 was at persistently high levels in both elderly patients (age >65 years) and elderly mice (age between 20 months and 22 months) with sepsis [25]. Saito H et al. also observed significantly elevated levels of interleukin 6 in elderly mice with sepsis (age was 24 months) [26]. Michels EHA et al. showed that, compared to the younger patients with sepsis, there was significant dysfunction of endothelial cells in older patients with sepsis [27]. Epidemiological findings have shown that the prevalence of anemia is progressively increasing in the older people [28]. In elderly patients with anemia, especially those with HAR below 0.13, aggressive improvement of hemoglobin concentration may be able to compensate for the detrimental effects of advanced age on patients with SS. Although transfusion of erythrocyte can increase the hemoglobin concentration of patients. However, transfusion of large amounts of erythrocyte may induce transfusion-related acute lung injury [29–34], febrile nonhemolytic transfusion reaction, acute hemolytic transfusion reaction, allergic and anaphylactic transfusion reactions, transfusion-associated circulatory overload, septic transfusion reaction, hypotensive transfusion reaction, delayed hemolytic transfusion reactions, transfusion-associated graft-versus-host disease and so on [35–37]. Therefore, when the HAR of a patient with SS is higher than 0.13, the patient may not benefit from transfusion of erythrocyte.
Stratified analyses showed that the protective effect of HAR was more pronounced in patients with AF compared with those without AF. Previous studies have shown that AF is strongly associated with the development of heart failure [38–40]. Patients with AF have poorer basic cardiac function compared with those without AF. As mentioned earlier, oxygen delivery is affected by both Hb concentration and cardiac output. Therefore, among patients with SS of the same age group, patients with AF may have a higher Hb requirement than those without AF.
This study also has some limitations. Restricted cubic spline showed that when the HAR was below 0.13, mortality within 28 days after ICU admission decreased significantly with increasing HAR. Spearman’s rank correlation analysis showed that when HAR was below 0.13, there was a negative correlation between HAR and mortality within 28 days after ICU admission. However, the correlation was weak. The reason for these results may be that the prognosis of SS is influenced by a combination of many factors. The correlation analysis did not exclude the interference of other confounding variables on the prognosis of SS. Therefore, the relationship between HAR and the prognosis of SS still needs to be further validated in prospective randomized controlled studies.
5 Conclusion
In conclusion, there was a nonlinear association between the HAR and mortality within 28 days after ICU admission. When the HAR was at a low level, mortality within 28 days after ICU admission decreased significantly as the HAR increased. When the HAR was at high levels, the HAR might not be a protective factor for mortality within 28 days after ICU admission. In patients with SS, the HAR was more effective in reducing the risk of death in patients with AF than in patients without AF.
References
- 1. Evans L, Rhodes A, Alhazzani W, Antonelli M, Coopersmith CM, French C et al. Surviving Sepsis Campaign: International Guidelines for Management of Sepsis and Septic Shock 2021. Crit Care Med. 2021 Nov 1;49(11): e1063–e1143. pmid:34605781
- 2. Annane D, Bellissant E, Cavaillon JM. Septic shock. Lancet. 2005 Jan 1–7;365(9453):63–78. pmid:15639681
- 3. Foster DM, Kellum JA. Endotoxic Septic Shock: Diagnosis and Treatment. Int J Mol Sci. 2023 Nov 10;24(22):16185. pmid:38003374
- 4. Angus DC, van der Poll T. Severe sepsis and septic shock. N Engl J Med. 2013 Aug 29;369(9):840–51. Erratum in: N Engl J Med. 2013 Nov 21;369(21):2069. pmid:23984731.
- 5. Wardi G, Tainter CR, Ramnath VR, Brennan JJ, Tolia V, Castillo EM et al. Age-related incidence and outcomes of sepsis in California, 2008–2015. J Crit Care. 2021 Apr; 62:212–217. pmid:33429114
- 6. Holst LB, Haase N, Wetterslev J, Wernerman J, Guttormsen AB, Karlsson S et al. Lower versus higher hemoglobin threshold for transfusion in septic shock. N Engl J Med. 2014 Oct 9;371(15):1381–91. Epub 2014 Oct 1. pmid:25270275
- 7. Su F, Cao L, Ren X, Hu J, Tavengana G, Wu H, et al. Age and sex trend differences in hemoglobin levels in China: a cross-sectional study. BMC Endocr Disord. 2023 Jan 9;23(1):8. pmid:36624464
- 8. Johnson AEW, Bulgarelli L, Shen L, Gayles A, Shammout A, Horng S et al. MIMIC-IV, a freely accessible electronic health record dataset. Sci Data. 2023 Jan 3;10(1):1. pmid:36596836
- 9. Hameed SM, Aird WC, Cohn SM. Oxygen delivery. Crit Care Med. 2003 Dec;31(12 Suppl): S658–67. pmid:14724463
- 10. Huang YC. Monitoring oxygen delivery in the critically ill. Chest. 2005 Nov;128(5 Suppl 2):554S–560S. pmid:16306053
- 11. Qi D, Peng M. Early Hemoglobin Status as a Predictor of Long-Term Mortality for Sepsis Patients in Intensive Care Units. Shock. 2021 Feb 1;55(2):215–223. pmid:33433167
- 12. Nichols D, Nielsen ND. Oxygen delivery and consumption: a macrocirculatory perspective. Crit Care Clin. 2010 Apr;26(2):239–53, table of contents. pmid:20381718
- 13. Rabinowitz JD, Enerbäck S. Lactate: the ugly duckling of energy metabolism. Nat Metab. 2020 Jul;2(7):566–571. pmid:32694798
- 14. Garcia-Alvarez M, Marik P, Bellomo R. Sepsis-associated hyperlactatemia. Crit Care. 2014 Sep 9;18(5):503. pmid:25394679
- 15. Qian WL, Xu R, Shi R, Li Y, Guo YK, Fang H, et al. The worsening effect of anemia on left ventricular function and global strain in type 2 diabetes mellitus patients: a 3.0 T CMR feature tracking study. Cardiovasc Diabetol. 2023 Jan 24;22(1):15. pmid:36694151
- 16. Ying Y, Ye J, Yuan Z, Cai D. Association of anaemia on heart failure and left ventricular function: A bidirectional Mendelian randomization study. ESC Heart Fail. 2024 Feb;11(1):299–305. Epub 2023 Nov 20. pmid:37984882
- 17. Muady GF, Bitterman H, Laor A, Vardi M, Urin V, Ghanem-Zoubi N. Hemoglobin levels and blood transfusion in patients with sepsis in Internal Medicine Departments. BMC Infect Dis. 2016 Oct 13;16(1):569. pmid:27737630
- 18. Long B, Koyfman A. Red Blood Cell Transfusion in the Emergency Department. J Emerg Med. 2016 Aug;51(2):120–30. pmid:27262735
- 19. Cable CA, Razavi SA, Roback JD, Murphy DJ. RBC Transfusion Strategies in the ICU: A Concise Review. Crit Care Med. 2019 Nov;47(11):1637–1644. pmid:31449062
- 20. Yang P, Yuan J, Yu L, Yu J, Zhang Y, Yuan Z, et al. Clinical significance of hemoglobin level and blood transfusion therapy in elderly sepsis patients: A retrospective analysis. Am J Emerg Med. 2023 Nov; 73:27–33. Epub 2023 Aug 9. pmid:37579529
- 21. Zhu YB, Liu TL, Dai Q, Liu SF, Xiong P, Huang H, et al. Characteristics and Risk Factors for Pediatric Sepsis. Curr Med Sci. 2024 Jun;44(3):648–656. Epub 2024 May 15. pmid:38748371
- 22. Lakbar I, Munoz M, Pauly V, Orleans V, Fabre C, Fond G et al. Septic shock: incidence, mortality and hospital readmission rates in French intensive care units from 2014 to 2018. Anaesth Crit Care Pain Med. 2022 Jun;41(3):101082. pmid:35472583
- 23. Zhou J, Tian H, Du X, Xi X, An Y, Duan M et al. Population-Based Epidemiology of Sepsis in a Subdistrict of Beijing. Crit Care Med. 2017 Jul;45(7):1168–1176. pmid:28422777
- 24. Nasa P, Juneja D, Singh O, Dang R, Arora V. Severe sepsis and its impact on outcome in elderly and very elderly patients admitted in intensive care unit. J Intensive Care Med. 2012 May-Jun;27(3):179–83. pmid:21436163
- 25. Inoue S, Suzuki K, Komori Y, Morishita Y, Suzuki-Utsunomiya K, Hozumi K et al. Persistent inflammation and T cell exhaustion in severe sepsis in the elderly. Crit Care. 2014 Jun 24;18(3): R130. pmid:24962182
- 26. Saito H, Sherwood ER, Varma TK, Evers BM. Effects of aging on mortality, hypothermia, and cytokine induction in mice with endotoxemia or sepsis. Mech Ageing Dev. 2003 Dec;124(10–12):1047–58. pmid:14659593
- 27. Michels EHA, Butler JM, Reijnders TDY, Cremer OL, Scicluna BP, Uhel F et al. Association between age and the host response in critically ill patients with sepsis. Crit Care. 2022 Dec 13;26(1):385. pmid:36514130
- 28. Stauder R, Valent P, Theurl I. Anemia at older age: etiologies, clinical implications, and management. Blood. 2018 Feb 1;131(5):505–514. Epub 2017 Nov 15. pmid:29141943
- 29. Semple JW, Rebetz J, Kapur R. Transfusion-associated circulatory overload and transfusion-related acute lung injury. Blood. 2019 Apr 25;133(17):1840–1853. Epub 2019 Feb 26. pmid:30808638
- 30. Yu Y, Lian Z. Update on transfusion-related acute lung injury: an overview of its pathogenesis and management. Front Immunol. 2023 May 12; 14:1175387. pmid:37251400; PMCID: PMC10213666.
- 31. Tung JP, Chiaretti S, Dean MM, Sultana AJ, Reade MC, Fung YL. Transfusion-related acute lung injury (TRALI): Potential pathways of development, strategies for prevention and treatment, and future research directions. Blood Rev. 2022 May; 53:100926. Epub 2022 Jan 5. pmid:35065815
- 32. Roubinian N. TACO and TRALI: biology, risk factors, and prevention strategies. Hematology Am Soc Hematol Educ Program. 2018 Nov 30;2018(1):585–594. Epub 2018 Dec 14. pmid:30570487
- 33. Sihler KC, Napolitano LM. Complications of massive transfusion. Chest. 2010 Jan;137(1):209–20. Erratum in: Chest. 2010 Mar;137(3):744. pmid:20051407
- 34. Vlaar APJ, Toy P, Fung M, Looney MR, Juffermans NP, Bux J, et al. A consensus redefinition of transfusion-related acute lung injury. Transfusion. 2019 Jul;59(7):2465–2476. Epub 2019 Apr 16. pmid:30993745
- 35. Wang Y, Rao Q, Li X. Adverse transfusion reactions and what we can do. Expert Rev Hematol. 2022 Aug;15(8):711–726. Epub 2022 Aug 17. pmid:35950450
- 36. Savage WJ. Transfusion Reactions. Hematol Oncol Clin North Am. 2016 Jun;30(3):619–34. pmid:27113000
- 37. Bojanić I, Lukić M, Plenković F, Raos M, Medenjak M, Ćepulić BG. Changes in the incidence of transfusion reactions in hematological patients over the past 30 years. Transfusion. 2022 Mar;62(3):600–611. Epub 2022 Feb 1. pmid:35149994
- 38. Mehra MR, Butler J. Atrial fibrillation and heart failure: it takes two to tango. Foreword. Heart Fail Clin. 2013 Oct;9(4): xiii–xiv. Epub 2013 Aug 19. pmid:24054485
- 39. Chua W, Fabritz L. Heart failure in patients with atrial fibrillation: why it matters now more than ever. Heart. 2021 Aug;107(16):1278–1279. Epub 2021 May 13. pmid:33985988
- 40. Taniguchi N, Miyasaka Y, Suwa Y, Harada S, Nakai E, Shiojima I. Heart Failure in Atrial Fibrillation- An Update on Clinical and Echocardiographic Implications. Circ J. 2020 Jul 22;84(8):1212–1217. Epub 2020 Jul 8.