Abstract
Background:
Thrombocytopenia is a risk factor for morbidity and mortality in critically ill patients, and is common following cardiopulmonary bypass (CPB). In this study, we evaluate whether thrombocytopenia following CPB is an independent risk factor for post-operative morbidity and mortality.
Methods:
We retrospectively evaluated 1,364 patients requiring CPB at University of Colorado Hospital between January 2011 and May 2016. Platelet nadir, absolute change in platelets, and percent change in platelets were modeled as continuous variables. Patients with post-operative thrombocytopenia (defined a nadir <75x103/μL within 72 hours) were also compared to patients without thrombocytopenia in a propensity-matched model. The primary outcome was in-hospital mortality, and secondary outcomes included post-operative infection, post-operative acute kidney injury (AKI), post-operative stroke, and prolonged intensive care unit (ICU) and hospital lengths of stay (LOS).
Results:
Post-operative thrombocytopenia occurred in 356 (26.0%) patients. In multivariable analysis platelet nadir was significantly inversely associated with mortality [0.955 (0.934-0.975), p<0.001], post-operative infection [0.992 (0.986-0.999), p=0.03], AKI (all stage) [0.993 (0.988-0.998), p = 0.01], AKI (stage 3) [0.966 (0.951-0.982), p <0.001], post-operative stroke [0.974 (0.956-0.992), p=0.006] prolonged ICU stay [.0986 (0.981-0.991), p<0.001], and hospital LOS [0.998 (0.997-0.999), p=0.001]. Percent change in platelets from baseline was also significantly associated with all primary and secondary outcomes.
Conclusions:
Post-operative thrombocytopenia is independently associated with post-operative mortality, AKI, infection, stroke, and prolonged ICU and hospital LOS. Serial platelet monitoring may help identify patients at higher risk of post-operative complications. Further studies investigating strategies to reduce post-operative thrombocytopenia, including reducing CPB time, are needed.
Keywords: Cardiopulmonary bypass/CPB, Platelets, Post-operative Mortality, Infection, Kidney/renal failure/dialysis, Surgery complications
Introduction
Thrombocytopenia is common following cardiopulmonary bypass (CPB), with a reported incidence >30%. The platelet nadir tends to occur 48-72 hours following surgery, and the average platelet count decreases by nearly 50%(1–3). Thrombocytopenia is an adverse prognostic marker in critically ill patients (4) and correlates with increased morbidity and mortality in patients following cardiac surgery(5, 6). Not only do platelet counts tend to decrease during cardiopulmonary bypass, but platelet function and morphology are also altered, as evidenced by decreased aggregation(7–9), changes in platelet surface markers(10), and decreased mean platelet volumes(11). While the exact mechanism behind these platelet changes requires further elucidation, one possibility is that blood-membrane interactions and blood shearing in the extracorporeal circuit during CPB cause activation and degranulation of platelets(12).
In addition to their role in thrombus formation, recent studies show that platelets play a vital role in response to infection(13, 14) by facilitating translocation of immune cells to areas of inflammation and by releasing cytokines and other molecular mediators. Post-CPB thrombocytopenia may inhibit platelet action against infection by decreasing both the number of platelets and their functionality, which could contribute to increased rates of post-surgical infection(15). Platelets also play a role in endothelial function, and thrombocytopenia has been associated with stroke and acute kidney injury (AKI) in the post-operative period(5, 16).
The purpose of this retrospective study was to evaluate the effect of post-CPB thrombocytopenia on post-operative mortality and morbidity, including post-operative infection, AKI, stroke, and ICU and hospital length of stay (LOS). We hypothesized that thrombocytopenia would be associated with increased mortality and complications even after adjustment for confounding factors.
Methods
Study Population:
We conducted a retrospective chart review from January 2011 to May 2016 utilizing the University of Colorado (UCH) Society of Thoracic Surgeons (STS) database. Institutional review board approval was obtained from UCH with a waiver of informed consent. Subjects were included if they were ≥18 years of age and underwent non-emergent cardiac surgery requiring CPB. Subjects were excluded for the following reasons: 1) intraoperative death, 2) received antibiotics within 2 weeks prior to surgery, 3) developed an infection within 48 hours of surgery, 4) unclear documentation of perioperative antimicrobial prophylaxis, 5) end stage renal disease, 6) procedures performed for infective endocarditis, aortic dissection, or heart transplant.
Predictor Variable:
The primary predictor variable was platelet nadir within 72 hours of CPB modeled as a continuous variable. Secondary predictors were an absolute change in platelets and percent change in platelets from baseline; both modeled as continuous variables. An additional propensity-score matched model compared patients with post-operative thrombocytopenia, defined as a platelet nadir <75x103/μL within 72 hours, to patients without were the thrombocytopenia. This cut-off was chosen because it represented the 25th percentile of platelet nadir following surgery. Finally, patients not included in the study due to lack of CPB were propensity-score matched to patients with CPB to determine the change in platelets related to the use of CPB.
Outcomes:
The primary outcome was in-hospital mortality. Other secondary outcomes were the post-operative infection, post-operative AKI (any stage), post-operative stage 3 AKI, post-operative stroke, intensive care unit (ICU) length of stay (LOS) ≥ 72 hours, and hospital LOS. AKI was defined using Kidney Disease Improving Global Outcomes (KDIGO) serum creatinine guidelines(17). Infection was defined as having one of the following: 1) documented surgical site infection which included mediastinitis, cellulitis of the sternal incision, or cellulitis related to vein harvest sites; 2) non-surgical site cellulitis; 3) positive blood culture, urine culture, or tracheal secretions culture; 4) imaging suggestive of new-onset pneumonia.
Adjustment Variables:
We adjusted our analyses for several variables that are shown to be associated with primary and secondary outcomes. These include the age, gender, body mass index, comorbidities including hypertension, diabetes mellitus, lung disease, smoking status, use of immunosuppressive medication preoperatively, previous cardiovascular interventions, pre-operative hematocrit, eGFR, ejection fraction as well as the type of surgery, CPB duration, and intraoperative blood transfusions.
Statistical Analysis:
We performed a multivariable logistic regression analysis to test for an association between platelet nadir at 72 hours, absolute change in platelets, and percent change in platelets and the binary primary and secondary outcomes listed above. For the propensity model patients were matched using a 1-to-1 greedy matching method in the comparison of post-operative thromboctyopenia to non-thrombocytopenia, and a 2-to-1 greedy matching method in the comparison of patients without CPB to those who underwent surgery with CPB. Matching was successful in both groups (Supplementary Tables 1+2). Generalized estimating equation (GEE) models were used to account for correlation within each matched pair. Statistical significance and confidence intervals were adjusted using the Bonferroni method. All statistical tests were considered significant at a 2-sided p < 0.05. All analyses were performed using SAS software version 9.4 (SAS Inc., Cary, NC).
Results
Baseline Data.
A total of 2,362 patients underwent cardiac surgery during the study period, of which 1,364 were included (Figure 1). The bivariate comparison of cohorts demonstrated that the patients who developed thrombocytopenia (platelets nadir <75x103/μL) were more often female, older, and had higher rates of congestive heart failure, immunosuppressive medication use, prior cardiovascular intervention, smoking, and IV drug abuse. They also had lower rates of diabetes, lower baseline platelets count and eGFR, and prolonged surgical and cardiopulmonary bypass times (Table 1). The post-operative thrombocytopenia group had significantly higher unadjusted rates of in-hospital mortality, post-operative infection, AKI (all stages), stage 3 AKI, post-operative stroke, ICU LOS, and hospital LOS. The perfusion times, surgical times, rates of post-operative thrombocytopenia, and in-hospital mortality by surgical type are listed in Table 2.
Figure 1.
Flowsheet depicting the number of patients screened, and reasons for patient exclusion. (AKI, acute kidney injury; ESRD, end-stagerenal disease; pre-op, preoperative.)
Table 1.
Comparison of baseline characteristics of patients with and without post-operative thrombocytopenia, defined as a platelet nadir within 72 hours of < 75×103/μL.
| Platelet Nadir < 75* (n=356) |
Platelet Nadir ≥ 75* (n=1,008) |
||
|---|---|---|---|
| Characteristics | N (%)† | N (%)† | P value‡ |
| Female | 126 (35.4) | 261 (25.9) | <.001 |
| Age, years, mean (SD) | 65.0 (13.6) | 59.2 (13.7) | <.0001 |
| Race/ethnicity | .01 | ||
| Black | 15 (4.2) | 73 (7.2) | |
| Caucasian | 282 (79.2) | 746 (74.0) | |
| Hispanic | 38 (10.7) | 152 (15.1) | |
| Other/unknown | 21 (5.9) | 37 (3.7) | |
| Body mass index, mean (SD) | 26.9 (5.2) | 29.1 (6.6) | <.0001 |
| Comorbidities | |||
| Sleep apnea | 66 (18.5) | 204 (20.2) | .49 |
| Hypertension | 239 (67.1) | 702 (69.6) | .38 |
| Diabetes Mellitus | 89 (25.0) | 334 (33.1) | .004 |
| Chronic lung disease | 69 (19.4) | 214 (21.2) | .46 |
| Liver disease | 13 (3.7) | 26 (2.6) | .30 |
| Congestive heart failure | 211 (29.3) | 422 (41.9) | <.0001 |
| Cancer with 5 years | 35 (9.8) | 90 (8.9) | .61 |
| Peripheral vascular disease | 24 (6.7) | 53 (5.3) | .30 |
| Cerebrovascular disease | 38 (10.7) | 102 (10.1) | .77 |
| Dyslipidemia | 203 (57.0) | 593 (58.8) | .55 |
| Baseline platelets, mean (SD) | 177.4 (57.8) | 232.0 (63.2) | <.0001 |
| Baseline creatinine, mean (SD) | 1.1 (0.3) | 1.0 (0.4) | .04 |
| Baseline eGFR, mean (SD) | 71.8 (20.9) | 79.5 (21.2) | <.0001 |
| Peak creatinine 72 hours, mean (SD) | 1.4 (0.7) | 1.2 (0.5) | <.0001 |
| Smoker at time of surgery | 73 (20.5) | 265 (26.3) | .03 |
| History of IV drug abuse | 8 (2.5) | 48 (4.8) | .04 |
| Last hematocrit, mean (SD) | 39.9 (6.3) | 41.2 (5.6) | <.001 |
| Last white blood cell count, mean (SD) | 6.7 (2.1) | 7.6 (2.5) | <.0001 |
| Cardiopulmonary bypass time, mean (SD) | 173.2 (78.6) | 131.2 (59.7) | <.0001 |
| Ejection fraction, mean (SD) | 50.8 (17.0) | 52.3 (15.6) | .13 |
| Surgical time, hours, mean (SD) | 7.6 (2.2) | 6.3 (1.7) | <.0001 |
| Cardiogenic shock | 7 (2.0) | 10 (1.0) | .17 |
| Intraoperative blood products | 217 (61.0) | 341 (33.8) | <.0001 |
| Red blood cell units, mean (SD) | 1.6 (3.1) | 0.4 (1.0) | <.0001 |
| Platelet units, mean (SD) | 1.1 (1.3 ) | 0.4 (0.8) | <.0001 |
| Fresh frozen plasma, mean (SD) | 2.5 (3.7) | 0.9 (2.0) | <.0001 |
| Cryoprecipitate units, mean (SD) | 0.1 (0.4) | 0.03 (0.2) | <.001 |
| Immunosuppressive medication prior to surgery | 20 (5.6) | 27 (2.7) | .01 |
| Steroid use | 15 (4.2) | 28 (2.8) | .18 |
| Previous cardiovascular intervention | 152 (42.7) | 302 (30.0) | <.0001 |
| Surgical type | <.0001 | ||
| Aneurysm repair | 34 (9.6) | 115 (11.4) | |
| Valve surgery | 135 (37.9) | 286 (28.4) | |
| Coronary artery bypass | 58 (16.3) | 361 (35.8) | |
| Left ventricular assist device | 27 (7.6) | 48 (4.8) | |
| Other | 16 (4.5) | 57 (5.7) | |
| Combined | 86 (24.2) | 141 (14.0) | |
| Family history cardiovascular disease | 247 (69.4) | 689 (68.4) | .72 |
| New York Heart Association class | |||
| None | 146 (41.0) | 587 (58.2) | <.0001 |
| Class I | 6 (1.7) | 17 (1.7) | |
| Class II | 57 (16.0) | 111 (11.0) | |
| Class III | 89 (25.0) | 191 (19.0) | |
| Class IV | 58 (16.3) | 102 (10.1) | |
| Adverse outcomes | |||
| In-hospital death | 21 (5.9) | 6 (0.6) | <.0001 |
| Post-operative infection | 61 (17.1) | 85 (8.4) | <.0001 |
| Post-operative deep sternal wound infection | 9 (2.5) | 11 (1.1) | .04 |
| Post-operative sepsis | 27 (7.6) | 28 (2.8) | <.0001 |
| Pneumonia | 28 (7.9) | 28 (2.8) | <.0001 |
| Acute kidney injury | 109 (30.6) | 199 (19.7) | <.0001 |
| Stage 3 acute kidney injury | 30 (8.4) | 16 (1.6) | <.0001 |
| Stroke | 15 (4.2) | 11 (1.1) | <.001 |
| Intensive care unit stay greater than 72 hours | 205 (57.6) | 374 (37.1) | <.0001 |
| Length of stay, median (IQR) | 10 (7-16) | 8 (6-12) | <.0001 |
Abbreviations: SD, standard deviation; eGRF, estimated glomerular filtration rate; IV, intravenous; IQR, interquartile range.
Platelet nadir of 75 x 103/μL.
Values are frequency and column percent unless otherwise stated.
P values are from chi-square and t-test and are bolded when less than .05.
Table 2.
Perfusion and surgical time, rates of acute kidney injury, infection, and in-hospital deaths by types of procedures performed.
| Procedure type | Total N (%)† | Perfusion time, minutes Mean (SD) | Surgical time, hours Mean (SD) | Platelet Nadir < 75 N (%)‡ | Infections N (%)‡ | AKI N (%) | Stage 3 AKI N (%)‡ | In-hospital deaths N (%)‡ |
|---|---|---|---|---|---|---|---|---|
| Aneurysm repair | 149 (10.9) | 166.3 (59.9) | 7.2 (1.9) | 34 (22.8) | 13 (8.7) | 31 (20.8) | 3 (2.0) | 4 (2.7) |
| Valve surgery | 421 (30.9) | 153.5 (67.3) | 6.3 (1.8) | 135 (32.1) | 33 (7.8) | 69 (16.4) | 12 (2.9) | 8 (1.9) |
| Coronary artery bypass | 419 (30.7) | 114.2 (50.6) | 6.6 (1.7) | 58 (13.8) | 55 (13.1) | 110 (26.) | 15 (3.4) | 4 (1.0) |
| Left ventricular assist device | 75 (5.5) | 110.7 (60.1) | 6.8 (2.0) | 27 (36.0) | 11 (14.7) | 20 (26.7) | 5 (6.7) | 2 (2.7) |
| Other/TAVR/TEVAR | 73 (5.4) | 108.5 (75.4) | 5.5 (2.1) | 16 (21.9) | 6 (8.2) | 13 (17.8) | 2 (2.7) | 3 (4.1) |
| Combined§ | 227(16.6)) | 178.3 (71.2) | 7.4 (1.9) | 86 (37.9) | 28 (12.3) | 65 (28.6) | 9 (4.0) | 6 (2.6) |
| Total | 1,364 | 142.2 (67.7) | 6.7 (1.9) | 356 (26.1) | 146 (10.7) | 308 (28.6) | 46 (3.4) | 27 (2.0) |
Abbreviations: AKI, acute kidney injury; SD, standard deviation; TAVR, transcatheter aortic valve replacement; TEVAR, thoracic endovascular aortic repair.
Values are frequency and column percentage.
Values are frequency and row percentage.
Combined includes all patients with procedures that fell into multiple categories.
Platelet decreases following cardiopulmonary bypass.
Thrombocytopenia following CPB was nearly universal, as demonstrated by the fact that 1,228 (90.2%) patients had a platelet nadir <150x103/μL within 72 hours of CPB. The number of patients with a platelet nadir <100x103/μL was 729 (53.5%), and the number with platelet nadir <75x103/μL was 356 (26.0%). The mean platelet decrease was 52.7%. Thrombocytopenia before surgery was uncommon, as only 173 (12.7%) had a pre-operative platelet count <150x103/μL and only 26 (1.9%) had a platelet count <100x103/μL. Notably, only six patients out of the 356 with post-operative thrombocytopenia were found to have heparin-induced thrombocytopenia (HIT).
Primary and secondary outcomes for continuous predictor variables.
Platelet nadir modeled as a continuous variable was inversely related to in-hospital mortality, post-operative infection, post-operative AKI (all stage and stage 3), post-operative stroke, ICU LOS, and hospital LOS. All associations were highly statistically significant, even after adjustments for covariates (Figure 2).
Figure 2.
Odds Ratios of adverse events in patients with and without post-operative thrombocytopenia, defined as < 75x103/μL following cardiac surgery. (AKI, acute kidney injury; ESRD, end-stage renal disease; ICU, intensive care unit; LOS, length of stay.)
Similarly, percent change in platelets as a continuous variable was also inversely associated with all primary and secondary outcomes, and all associations were statistically significant. However, absolute change in platelets did not achieve statistical significance even in unadjusted models, except for prolonged ICU LOS (Table 3).
Table 3.
Multivariable adjusted adverse outcomes using platelet nadir, absolute change in platelets, and percent change in platelets as continuous variables.
| Platelet Nadir | % Change in Platelets | Absolute Change in Platelets | ||||
|---|---|---|---|---|---|---|
| Adverse outcome | Odds Ratio (95% CI) | P value | Odds Ratio (95% CI) | P value | Odds Ratio (95% CI) | P value |
| In-Hospital death | ||||||
| Unadjusted | 0.954 (0.940-0.968) | <.0001 | 0.922 (0.896-0.950) | <.0001 | 0.994 (0.989-1.000) | .06 |
| Multivariable adjusted | 0.955 (0.934-0.975) | <.0001 | 0.935 (0.899-0.972) | <.001 | 0.995 (0.987-1.002) | .17 |
| Infection | ||||||
| Unadjusted | 0.992 (0.987-0.997) | .002 | 0.978 (0.966-0.991) | <.001 | 0.997 (0.994-1.000) | .06 |
| Multivariable adjusted | 0.992 (0.986-0.999) | .03 | 0.984 (0.970-0.999) | .03 | 0.998 (0.994-1.001 | .19 |
| Stroke | ||||||
| Unadjusted | 0.975 (0.963-0.988) | <.001 | 0.946 (0.919-0.973) | <.001 | 0.996 (0.989-1.002) | .16 |
| Multivariable adjusted | 0.974 (0.956-0.992) | .006 | 0.953 (0.918-0.987) | .01 | 0.996 (0.988-1.005) | .39 |
| Acute kidney injury | ||||||
| Unadjusted | 0.993 (0.990-0.997) | <.001 | 0.993 (0.984-1.002) | .12 | 1.000 (0.998-1.002) | .97 |
| Multivariable adjusted | 0.993 (0.988-0.998) | .01 | 0.991 (0.980-1.002) | .10 | 1.000 (0.997-1.002) | .72 |
| Stage 3 acute kidney injury | ||||||
| Unadjusted | 0.969 (0.959-0.980) | <.0001 | 0.955 (0.935-0.976) | <.0001 | 0.999 (0.994-1.005) | .83 |
| Multivariable adjusted | 0.966 (0.951-0.982) | <.0001 | 0.967 90.941-0.995) | .02 | 0.966 (0.951-0.982) | <.0001 |
| Intensive care unit stay > 72 hours | ||||||
| Unadjusted | 0.991 (0.988-0.994) | <.0001 | 0.971 (0.963-0.979) | <.0001 | 0.991 (0.988-0.994) | <.0001 |
| Multivariable adjusted | 0.986 (0.981-0.991) | <.0001 | 0.972 (0.962-0.982) | <.0001 | 0.986 (0.981-0.991) | <.0001 |
| Length of stay* | ||||||
| Unadjusted | 0.998 (0.997-0.998) | <.0001 | 0.995 (0.993-0.998) | <.0001 | 0.998 (0.997-0.998) | <.0001 |
| Multivariable adjusted | 0.998 (0.997-0.999) | .001 | 0.997 (0.995-0.999) | .01 | 0.998 (0.997-0.999) | .001 |
Abbreviation: CI, confidence interval.
Length of stay was fitted to a negative binomial model.
Primary and secondary outcomes for propensity-score matched analysis.
Post-operative thrombocytopenia was associated with increased rates of in-hospital mortality, post-operative infection, AKI, post-operative stroke, and prolonged ICU and hospital LOS in both unadjusted and multivariable analysis. These associations continued to be statistically significant in the propensity-score matched models, except in-hospital mortality, which barely missed significance with a p-value of 0.05 (Table 4).
Table 4.
Unadjusted, multivariable adjusted, and propensity matched adverse outcomes for thrombocytopenia, defined as a platelet nadir < 75 x 103/μL.
| Adverse outcome | Odds Ratio (95% CI) | P value |
|---|---|---|
| In-Hospital death | ||
| Unadjusted* | 10.469 (4.190-26.156) | <.0001 |
| Multivariable adjusted* | 6.192 (1.926-19.912) | .002 |
| Propensity matched† | 4.632 (0.980-21.887) | .05 |
| Infection | ||
| Unadjusted* | 2.245 (1.576-3.199) | <.0001 |
| Multivariable adjusted* | 2.075 (1.308-3.290) | .002 |
| Propensity matched† | 3.170 (1.618-6.211 ) | .001 |
| Stroke | ||
| Unadjusted* | 3.987 (1.814-8.764) | <.001 |
| Multivariable adjusted* | 3.182 (1.112-9.111) | .03 |
| Propensity matched† | 5.168 (1.109-24.088) | .04 |
| Acute kidney injury | ||
| Unadjusted* | 1.794 (1.365-2.359) | <.0001 |
| Multivariable adjusted* | 1.640 (1.152-2.335) | .01 |
| Propensity matched† | 1.486 (0.998-2.213) | .05 |
| Stage 3 acute kidney injury | ||
| Unadjusted* | 5.706 (3.071-10.601) | <.0001 |
| Multivariable adjusted* | 4.171 (1.693-10.278) | .002 |
| Propensity matched† | 5.709 (1.239-26.309) | .03 |
| Intensive care unit stay > 72 hours | ||
| Unadjusted* | 2.301 (1.800-2.943) | <.0001 |
| Multivariable adjusted* | 1.920 (1.389-2.655) | <.0001 |
| Propensity matched† | 1.550 (1.088-2.208) | .02 |
| Length of stay‡ | ||
| Unadjusted* | 1.384 (1.282-1.495) | <.0001 |
| Multivariable adjusted* | 1.176 (1.094-1.265) | <.0001 |
| Propensity matched† | 1.217 (1.062-1.393) | .005 |
Abbreviation: CI, confidence interval.
Sample size: 356 platelet nadir <75 & 1,008 platelet nadir ≥ 75.
Sample size: 248 platelet nadir <75 & 248 platelet nadir ≥ 75.
Length of stay was fitted to a negative binomial model.
Rates of thrombocytopenia in CPB compared to non-CPB.
In this propensity-score matched analysis, patients who were not on CPB had significantly lower rates of post-operative thrombocytopenia. The cut-off used to define thrombocytopenia did not change the association. Furthermore, these associations remained statistically significant even after adjustments for covariates in the multivariable and propensity-score matched models (Table 5).
Table 5.
Unadjusted, multivariable adjusted, and propensity matched outcomes for patients without and with cardiopulomanry bypass.
| Adverse outcome | Odds Ratio (95% CI) | P value |
|---|---|---|
| Platelet nadir < 75 | ||
| Unadjusted* | 0.508 (0.334-0.773) | .002 |
| Multivariable adjusted* | 0.318 (0.139-0.731) | .01 |
| Propensity matched† | 0.463 (0.235-0.910) | .03 |
| Platelet nadir < 100 | ||
| Unadjusted* | 0.472 (0.342-0.650) | <.0001 |
| Multivariable adjusted* | 0.246 (0.127-0.478) | <.0001 |
| Propensity matched† | 0.415 (0.247-0.698) | <.001 |
| Platelet nadir < 150 | ||
| Unadjusted* | 0.374 (0.254-0.551) | <.0001 |
| Multivariable adjusted* | 0.100 (0.039-0.251) | <.0001 |
| Propensity matched† | 0.276 (0.141-0.540) | <.001 |
Abbreviation: CI, confidence interval.
Sample size: 185 without perfusion time & 1,368 with a perfusion time.
Sample size: 105 without perfusion time & 159 with a perfusion time.
Length of stay was fitted to a negative binomial model.
Discussion
Our study demonstrates that post-operative thrombocytopenia is associated with a significantly increased risk of mortality, as well as increased risk of morbidity in the form of post-operative infection, development of AKI, post-operative stroke, and prolonged ICU and hospital LOS. Our results also demonstrate that CPB is a major risk factor for the development of post-operative thrombocytopenia. These findings suggest a role for serial platelet monitoring following CPB to identify patients at risk for adverse outcomes, and also provides a basis for further studies investigating the impact of interventions designed to reduce post-operative thrombocytopenia.
Thrombocytopenia is common following cardiopulmonary bypass surgery. The average platelet decrease in other studies ranges from 30-50%, with platelet nadirs occurring in the first 72 hours following surgery(7–9). Furthermore, this platelet decrease has been correlated with increased patient mortality(5, 6). In our study, at least some platelet decrease was observed in nearly every patient, and the majority of patients reached a nadir of <100x103/μL, with an average platelet decrease of 52.7%. Of 356 patients with platelets nadirs <75x103/μL following surgery, only 6 (1.7%) patients in this high-risk population were found to have HIT. These findings are consistent with previous investigations that have shown that HIT occurs in 1-2% of patients following CPB surgery(18). HIT therefore appears to have contributed minimally to the increased morbidity and mortality seen in our post-operative thrombocytopenia group.
Platelets have recently been shown to participate in both innate and adaptive immunity, suggesting they are an essential component of the immune system(17). It is therefore plausible that dysregulation of platelets as observed following CPB and reflected by thrombocytopenia would impair the body’s ability to resist infections in the post-surgical period. Our study suggests that platelet nadir and % platelet change within 72 hours of CPB were associated with the development of post-operative infection.
An association between thrombocytopenia following open-heart surgery and acute kidney injury was recently demonstrated in a population of pediatric patients undergoing cardiac surgery(19) as well as an adult population undergoing CABG(5). While a mechanism by which thrombocytopenia could cause AKI remains to be conclusively demonstrated, one possibility is that post-CPB thrombocytopenia is a marker of microvascular thrombus formation. Post-operative platelet counts were recently shown to correlate with decreased sublingual microcirculation(20). AKI may, therefore, result from the decreased microvascular flow within the kidneys. CPB patients have been shown in one study to have increased risk of stroke(16), which we again demonstrated in this study. Microvascular flow disturbances may be responsible for the increased risk of stroke in this population, but further research is needed to determine the role of thrombocytopenia in AKI and stroke following CPB.
The mechanism of platelet loss during and after cardiopulmonary bypass remains to be determined, but one hypothesis is that blood-membrane contact and shear forces within the extracorporeal circuit cause platelet activation, which leads to subsequent consumption in the peripheral circulation, especially within the spleen(12). Previous studies have shown changes not only in platelet number, but also in platelet function following CPB as demonstrated by changes in measures of aggregation, differential expression of cell-surface markers, and changes in platelet volume and shape(7–11). In our study, we compared non-CPB patients to CPB patients, and even after propensity score matching non-CPB had significantly lower rates of post-operative thrombocytopenia. This suggests that CPB plays an important contributory role in the loss of platelets following surgery. It is therefore plausible that the impacts of CPB time on outcomes are mediated in part through effects on platelets, and that a reduction in CPB time, or use of off-bypass techniques, will improve patient outcomes. Indeed, preventing platelet activation during CPB through pharmacologic methods (GP IIb/IIIa inhibition) has been shown to decrease post-operative thrombocytopenia and mortality(21), albeit at the expense of increased bleeding. Further studies should investigate strategies, including reduced CPB time, to reduce rates of thrombocytopenia development.
Our study has several significant limitations. Given the retrospective nature of this study, it is possible that other confounding factors may be present that were not accounted for in the analysis, such as known post-operative thrombus formation. Also, due to the retrospective nature of this study, we can neither discern the mechanisms by which thrombocytopenia occurs following surgery nor by which platelet loss and thrombocytopenia may lead to poor outcomes. Platelet surface markers and platelet function panels, as examples, were not available. While acuity scores were not available within our database, we did exclude emergency cases and cases with an active infection in an attempt to limit the known contributions of illness acuity to thrombocytopenia.
In conclusion, our study demonstrates that post-surgical thrombocytopenia is associated with post-operative mortality and morbidity in adult patients undergoing non-emergent cardiac surgery. Further studies should focus on strategies to decrease the cardiopulmonary bypass time during cardiac surgery to minimize the risk of post-operative thrombocytopenia and its impact on post-operative morbidity and mortality. Future studies should also assess the pathophysiological mechanism that may link thrombocytopenia and adverse outcomes.
Supplementary Material
Acknowledgments and Disclosures
We would like to acknowledge the M-TRAC (Multidisciplinary Translational Research in AKI Collaborative) Investigators at the University of Colorado including Sophia Ambruso, J. Pedro Texiera, Meghan SooHoo, John Kim, Nathan Clendenen,,and Jorge DiPaola, as well as Courtney Matter, Michael Wells, and Kimberley Marshall for their significant contribution to this manuscript.
Funding: Dr. Aftab is supported by a Division of Cardiothoracic Surgery Faculty Seed Grant, Anschutz Medical Campus, University of Colorado, Aurora, CO; Benjamin Griffin is supported by an NIH Grant, T32 DK 007135
Footnotes
Meeting Presentation: Presented at the STS Annual Meeting 2019, 27th-29th January, San Diego
Selected for the Best Critical Care Poster Award at the STS scientific poster session.
References
- 1.Mujtaba SS, Ledingham S, Shah AR, Schueler S, Clark S, Pillay T. Thrombocytopenia after aortic valve replacement: Comparison between sutureless perceval s valve and perimount magna ease bioprosthesis. Braz J Cardiovasc Surg 2018;33(2):169–175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hamid M, Akhtar MI, Naqvi HI, Ahsan K. Incidence and pattern of thrombocytopenia in cardiac surgery patients. J Pak Med Assoc. 2017;67(7):1019–1023. [PubMed] [Google Scholar]
- 3.Ranucci M, Carlucci C, Isgro G, Baryshnikova E. A prospective pilot study of platelet function and its relationship with postoperative bleeding in pediatric cardiac surgery. Minerva Anestesiol. 2012;78(5):556–563. [PubMed] [Google Scholar]
- 4.Griffin BR, Jovanovich A, You Z, et al. Effects of Baseline Thrombocytopenia and Platelet Decrease Following Renal Replacement Therapy Initiation in Patients With Severe Acute Kidney Injury. Crit Care Med. 2019. April;47(4):e325–e331 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Kertai MD, Zhou S, Karhausen JA et al. Platelet counts, acute kidney injury, and mortality after coronary artery bypass grafting surgery. Anesthesiology. 2016;124(2):339–352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Thielmann M, Bunschkowski M, Tossios P et al. Perioperative thrombocytopenia in cardiac surgical patients - incidence of heparin-induced thrombocytopenia, morbidities and mortality. Eur J Cardiothorac Surg. 2010;37(6):1391–1395. [DOI] [PubMed] [Google Scholar]
- 7.Agarwal S, Johnson RI, Kirmani BH. Pre- and post-bypass platelet function testing with multiple electrode aggregometry and teg platelet mapping in cardiac surgery. J Cardiothorac Vasc Anesth. 2015;29(5):1272–1276. [DOI] [PubMed] [Google Scholar]
- 8.Weitzel NS, Weitzel LB, Epperson LE, Karimpour-Ford A, Tran ZV, Seres T. Platelet mapping as part of modified thromboelastography (teg(r)) in patients undergoing cardiac surgery and cardiopulmonary bypass. Anaesthesia. 2012;67(10):1158–1165. [DOI] [PubMed] [Google Scholar]
- 9.Bonding Andreasen J, Hvas AM, Ravn HB. Marked changes in platelet count and function following pediatric congenital heart surgery. Paediatr Anaesth. 2014;24(4):386–392. [DOI] [PubMed] [Google Scholar]
- 10.Maquelin KN, Berckmans RJ, Nieuwland R et al. Disappearance of glycoprotein ib from the platelet surface in pericardial blood during cardiopulmonary bypass. J Thorac Cardiovasc Surg. 1998;115(5):1160–1165. [DOI] [PubMed] [Google Scholar]
- 11.Unal EU, Ozen A, Kocabeyoglu S et al. Mean platelet volume may predict early clinical outcome after coronary artery bypass grafting. Journal of cardiothoracic surgery 2013;8:91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Wahba A, Black G, Koksch M et al. Cardiopulmonary bypass leads to a preferential loss of activated platelets. A flow cytometric assay of platelet surface antigens. Eur J Cardiothorac Surg. 1996;10(9):768–773. [DOI] [PubMed] [Google Scholar]
- 13.Ali RA, Wuescher LM, Worth RG. Platelets: Essential components of the immune system. Curr Trends Immunol. 2015;16:65–78. [PMC free article] [PubMed] [Google Scholar]
- 14.Morrell CN, Aggrey AA, Chapman LM, Modjeski KL. Emerging roles for platelets as immune and inflammatory cells. Blood. 2014;123(18):2759–2767. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.SooHoo M, Griffin B, Jovanovich A et al. Acute kidney injury is associated with subsequent infection in neonates after the norwood procedure: A retrospective chart review. Pediatr Nephrol 2018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Karhausen JA, Smeltz AM, Akushevich I, et al. Platelet Counts and Postoperative Stroke After Coronary Artery Bypass Grafting Surgery. Anesth Analg. 2017. October;125(4):1129–1139 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kellum JA, Lameire N. Diagnosis, evaluation, and management of acute kidney injury: A kdigo summary (part 1). Crit Care. 2013;17(1):204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Pishko AM, Cuker A. Heparin-Induced Thrombocytopenia in Cardiac Surgery Patients. Semin Thromb Hemost. 2017. October; 43(7): 691–698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Tew S, Fontes ML, Greene NH et al. Natural history of nonimmune-mediated thrombocytopenia and acute kidney injury in pediatric open-heart surgery. Paediatr Anaesth. 2017;27(3):305–313. [DOI] [PubMed] [Google Scholar]
- 20.Di Dedda U, Ranucci M, Porta A et al. The combined effects of the microcirculatory status and cardiopulmonary bypass on platelet count and function during cardiac surgery. Clin Hemorheol Microcirc. 2018. [DOI] [PubMed] [Google Scholar]
- 21.Smith PK, Goodnough LT, Levy JH et al. Mortality benefit with prasugrel in the triton-timi 38 coronary artery bypass grafting cohort: Risk-adjusted retrospective data analysis. J Am Coll Cardiol. 2012;60(5):388–396. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.