Risk Factors of Thrombocytopenia After Cardiac Surgery with Cardiopulmonary Bypass

Introduction Postoperative thrombocytopenia is common in cardiac surgery with cardiopulmonary bypass, and its risk factors are unclear. Methods This retrospective study enrolled 3,175 adult patients undergoing valve surgeries with cardiopulmonary bypass from January 1, 2017 to December 30, 2018 in our institute. Postoperative thrombocytopenia was defined as the first postoperative platelet count below the 10th quantile in all the enrolled patients. Outcomes between patients with and without postoperative thrombocytopenia were compared. The primary outcome was in-hospital mortality. Risk factors of postoperative thrombocytopenia were assessed by logistic regression analysis. Results The 10th quantile of all enrolled patients (75×109/L) was defined as the threshold for postoperative thrombocytopenia. In-hospital mortality was comparable between thrombocytopenia and non-thrombocytopenia groups (0.9% vs. 0.6%, P=0.434). Patients in the thrombocytopenia group had higher rate of postoperative blood transfusion (5.9% vs. 3.2%, P=0.014), more chest drainage volume (735 [550-1080] vs. 560 [430-730] ml, P<0.001), and higher incidence of acute kidney injury (12.3% vs. 4.2%, P<0.001). Age > 60 years (odds ratio [OR] 2.25, 95% confidence interval [CI] 1.345-3.765, P=0.002], preoperative thrombocytopenia (OR 18.671, 95% CI 13.649-25.542, P<0.001), and cardiopulmonary bypass time (OR 1.088, 95% CI 1.059-1.117, P<0.001) were positively independently associated with postoperative thrombocytopenia. Body surface area (BSA) (OR 0.247, 95% CI 0.114-0.538, P<0.001) and isolated mitral valve surgery (OR 0.475, 95% CI 0.294-0.77) were negatively independently associated with postoperative thrombocytopenia. Conclusion Positive predictors for thrombocytopenia after valve surgery included age > 60 years, small BSA, preoperative thrombocytopenia, and cardiopulmonary bypass time. BSA and isolated mitral valve surgery were negative predictors.


INTRODUCTION
Postoperative thrombocytopenia is common in cardiac surgery with cardiopulmonary bypass (CPB), leading to increased postoperative blood loss. The etiology of thrombocytopenia in cardiac surgery is related to blood destruction, hemodilution, and platelet activation during CPB [1] . The incidence of postoperative thrombocytopenia is about 10-40% according to different criteria of thrombocytopenia [2][3][4] . Postoperative thrombocytopenia was reported to be associated with postoperative adverse outcomes, including acute kidney injury (AKI), stroke, and death [5][6][7] . However, risk factors of postoperative thrombocytopenia are unclear. A study including 42 patients [5] identified preoperative platelet count, age, and intraoperative blood transfusion as independent predictors for postoperative thrombocytopenia. Given the small sample size, the conclusion was not convincing enough. The present retrospective study was aimed to describe the distribution of postoperative thrombocytopenia in patients undergoing cardiac surgery with CPB, investigate the relationship between postoperative thrombocytopenia and outcomes, and identify risk factors of postoperative thrombocytopenia.

Study Design and Population
This retrospective study was approved by our institutional review board (approval number 2020-1288), and the need for informed consent was waived. Data were extracted from the CPB database (2017)(2018). Only valve surgeries were included to exclude the impact of perioperative antiplatelet drugs on platelet counts in coronary artery bypass grafting (CABG) and minimize surgical heterogeneity. The inclusion criteria were: (1) age > 18 years and (2) undergoing valve replacement or repair surgery from January 1, 2017 to December 30, 2018. The exclusion criteria included: (1) combined CABG, aortic surgery, or other surgery; (2) preoperative antiplatelet treatment; (3) previous cardiac surgery within six months; and (4) emergency surgery.

Postoperative Thrombocytopenia
Platelet counts were extracted from blood routine tests at three time points: (1) last test before surgery (preoperative platelet count); (2) first test after surgery, immediately after admission to intensive care unit (ICU) (postoperative platelet count); and (3) last test before discharge from hospital (before-discharge platelet count). Postoperative thrombocytopenia was defined as a postoperative platelet count below the 10 th quantile of all the enrolled patients, referring to Kertai MD's method [5] . Platelet count recovery was defined as before-discharge platelet counts ≥125×109/L or the preoperative level.

Risk Factors
Risk factors of postoperative thrombocytopenia were assessed. Variables contained demographic information, body surface area (BSA), comorbidity, European System for Cardiac Operative Risk Evaluation (EuroSCORE), preoperative laboratory tests, surgical information (including valve position, valve numbers, valve replacement or repair, and mechanical or bioprosthetic valve), CPB circuit, CPB time, aortic cross-clamping time, CPB temperature, and total heparin dose.

Outcomes
Outcomes were compared between patients with and without postoperative thrombocytopenia. The primary outcome was in-hospital mortality. Secondary outcomes included chest drainage volume, postoperative blood transfusion, re-thoracotomy, platelet count recovery, cerebrovascular accidents (defined as the diagnosis of stroke or intracranial hemorrhage), AKI (defined as a postoperative increase in serum creatinine levels to more than two times baseline within a month or to > 354 mmol/l postoperatively), postoperative myocardial infarction (defined as an increase in creatine kinase to > 240 IU/l within 48 hours postoperatively), intraaortic balloon pump, and extracorporeal membrane oxygenation, mechanical ventilation time, and ICU length of stay. The postoperative red blood cell transfusion threshold was 8 g/L. Platelets were transfused when the level was < 50×10 9 /L. Fresh frozen plasma was transfused at excessive bleeding/chest drainage circumstances, with laboratory evidence suggesting coagulopathy.

Surgical Procedure
Cardiac surgery was performed with mild hypothermic CPB (32-34ºC) under general anesthesia. CPB flow was maintained at 2.2-2.6 L/min/BSA (m 2 ). A conventional or minimized CPB circuit was used according to the patient's preoperative hemoglobin level and the perfusionists' preference. A conventional CPB circuit consisted of a roller pump, a hollow-fiber oxygenator, an arterial filter, and an uncoated polyvinyl chloride circuit, with a priming volume of 1600 ml. A mini circuit [8] , with 900-ml priming volume, consisted of a hollow-fiber oxygenator integrated with arterial filter, smaller tubes, and vacuum-assisted venous reservoir drainage. Unfractionated heparin (400 U/kg) was given to achieve activated clotting time (ACT) ≥ 410 seconds before CPB. ACT was maintained ≥ 410 seconds during CPB. After CPB was weaned off, protamine sulfate was administered for heparin reversal (1 mg protamine vs. 100 units heparin).

Statistical Analysis
Variables were retrospectively collected from electric records. Continuous variables were presented as median (P25-P75) and were compared with the non-parametric test, according to its abnormal distribution. Categorical variables were presented as n (%) and compared with the x 2 test for equal proportion. Logistic regression of multiple risk factors was performed to identify risk factors of postoperative thrombocytopenia. Univariable associations with P<0.10 and clinical significance were evaluated by a forward stepwise method to establish a logistic regression model of multiple risk factors. Nonlinear continuous variables were transformed into categorical variables, including age, preoperative platelet count, blood urea nitrogen (BUN), and international normalized ratio (INR). The variables with strong collinearity were chosen according to clinical significance. For example, prosthetic valve type (mechanical or bioprosthetic valve) was not put into the logistic regression analysis due to its strong correlation with age. The fit of the logistic regression analysis was evaluated by the Hosmer-Lemeshow test. Two sensitivity analyses were also performed to determine if the risk factors of thrombocytopenia were affected by preoperative thrombocytopenia and intraoperative platelet transfusion. We excluded patients with preoperative platelet counts ≤ 125×109/L and patients who received intraoperative platelet transfusion, respectively. All statistical testing was two-sided, and a P-value < 0.05 was considered significant. All statistical analyses were performed using IBM Corp. Released 2015, IBM SPSS Statistics for Windows, Version 23.0, Armonk, NY: IBM Corp. The study was presented following the Strengthening the Reporting of Observational Studies in Epidemiology (or STROBE) guidelines.

Outcomes
Outcomes were compared between patients with and without postoperative thrombocytopenia. In-hospital mortality was   (Table 4). BSA (OR 0.247, 95% CI 0.114-0.538) and isolated mitral valve surgery (OR 0.475, 95% CI 0.294-0.77) were negative predictors for postoperative thrombocytopenia. These five risk factors were confirmed by sensitivity analyses excluding patients with preoperative platelet counts ≤ 125×109/L and patients who received intraoperative platelet transfusion (Table 5).

DISCUSSION
This retrospective study gave an overview of postoperative thrombocytopenia after cardiac surgery. In-hospital mortality was not increased in the thrombocytopenia group. Patients with postoperative thrombocytopenia tended to have more postoperative bleeding and required more blood transfusion. In addition, postoperative thrombocytopenia was accompanied by a series of complications, including AKI. Predictors for postoperative thrombocytopenia included age > 60 years, BSA, preoperative thrombocytopenia, CPB time, and isolated mitral valve surgery. There is no universally accepted definition of postoperative thrombocytopenia in cardiac surgery. Platelet count decline was typical in patients after cardiac surgery with CPB, and many patients' platelet counts were below the reference value in normal healthy individuals. But the threshold of platelet count with the most crucial clinical significance was undetermined. The definition of postoperative thrombocytopenia in the present study was < 75×10 9 /L, which was the 10 th quantile of the study population. Two previous large studies also took 75×109 as the threshold of postoperative thrombocytopenia [5,6] . They reported that postoperative platelet count < 75×10 9 was associated with adverse outcomes, including mortality, AKI, and prolonged ICU length of stay [5][6] . The present study did not find an association between postoperative thrombocytopenia and in-hospital mortality. But our study demonstrated that patients with postoperative thrombocytopenia had a higher incidence of AKI and longer mechanical ventilation time and ICU length of stay, which was consistent with the abovementioned two studies. We also found the relationship between postoperative thrombocytopenia  [9] . The mechanism under the relationship between thrombocytopenia and adverse outcomes is unclear. Thrombocytopenia has also been identified as a risk factor of multi-organ failure in critically ill patients [10][11][12][13] . Postoperative thrombocytopenia might be coupled with massive platelet-derived chemokines, which were associated with multi-organ injury [14,15] . Besides, postoperative thrombocytopenia and other adverse outcomes might be both the consequences of prolonged CPB and some underlying situations (i.e., microvascular thrombosis [16] ). Another possible assumption was that postoperative thrombocytopenia led to more bleeding and transfusions, which increased the risks of AKI and other adverse outcome [17,18] . Platelet consumption and hemodilution were considered as the leading causes of platelet count decline after cardiac surgery with CPB. Platelet was an essential modulator in inflammation and hemostasis. Once CPB was started, platelets were activated by mechanical shear stress and artificial surface contact, and platelet activation cascade was induced [1] . High shear stress could also induce platelet apoptosis [19] . In addition, hemolysis during CPB enhanced platelet aggregation [20] . Our study identified age > 60 years, BSA, preoperative thrombocytopenia, and CPB time as independent risk factors of postoperative thrombocytopenia. These factors were related to platelet consumption and hemodilution. Patients with small BSA had more hemodilution during CPB. Prolonged CPB indicated more platelet activation. Older age was reported to be associated with platelet hyperreactivity and enhanced aggregation [21] . Therefore, we guessed older patients might have more platelet activation during cardiac surgery. In addition, older individuals had fewer platelet counts than younger people [22] . In addition, the present study found a relationship between valve surgical information and postoperative thrombocytopenia. Isolated mitral valve surgery was a negative predictor for thrombocytopenia. The reasons might be stated as follows. Firstly, compared to multivalve surgeries, isolated mitral valve surgeries usually required shorter CPB time and therefore caused less Heparan dose (U) 32400 (28400-36800) 30000 (26800-34400) < 0.001 Intraoperative FFP transfusion (ml) 0 (0-0) 0 (0-400) < 0.001 Intraoperative platelet transfusion (U) 0 (0-0) 0 (0-0) < 0.001 "Data are presented as n (%) or median (interquartile range). *Valve type, surgeries with more valves. FFP=fresh frozen plasma; RBC=red blood cell; VAVD=vacuum-assisted venous drainage platelet consumption. Secondly, compared to isolated aortic valve surgeries, a larger proportion of isolated mitral valve surgeries was valve repair surgery. Prosthetic valves would lead to increased platelet activation [23] . For patients with high risks of postoperative thrombocytopenia, an active platelet preservation strategy might be applied. Heparinbonded cardiopulmonary circuit reduced platelet activation compared to standard uncoated circuit [24] . Several potential agents have been reported, including nitric oxide [25] . More research is required to obtain comprehensive recognition of platelet preservation during CPB. Heparin-induced thrombocytopenia (HIT) was also a big concern in postoperative thrombocytopenia. A biphasic platelet count pattern was typical of HIT, while early-onset thrombocytopenia was mainly related to CPB [26] . In the present study, postoperative thrombocytopenia was defined with the first platelet count test after surgery, and hence HIT was not the primary concern. We screened all the patients with postoperative thrombocytopenia  and those who received postoperative platelet transfusion retrospectively. Only three patients were suspected with HIT, but not confirmed as anti-PF4/heparin antibody test was not available in our institute.

Limitations
This study had several limitations, mainly due to its retrospective nature. Firstly, postoperative platelet count was affected by intraoperative platelet transfusion. However, a sensitivity analysis excluding patients with intraoperative platelet transfusion was performed and obtained similar results. Secondly, the result might be influenced by the definition of postoperative thrombocytopenia. Thirdly, platelet function and markers of platelet activation were not tested. Fourthly, other factors (i.e., redo surgery and valve brands), which might be associated with postoperative thrombocytopenia, were not included in risk factor analysis.

CONCLUSION
Thrombocytopenia was common after cardiac surgery. The risk factors of postoperative thrombocytopenia included age > 60 years, BSA, preoperative thrombocytopenia, and CPB time. Patients with postoperative thrombocytopenia tended to have more postoperative bleeding and required more blood transfusion, and post thrombocytopenia was associated with a series of complications.
Financial support: This study was supported by the Non-profit Central Research Institute Fund of Chinese Academy of Medical Sciences (2019XK320052). No conflict of interest.