Cardiopulmonary Bypass, Myocardial Management, and Support Techniques Hyperbilirubinemia after cardiac operation Incidence, risk factors, and clinical significance Three hundred and two consecutive patients who had undergone cardiac operation for various cardiac lesions were studied prospectively to evaluate the incidence, risk factors, and the associated mortality of postoperative hyperbilirubinemia after cardiopulmonary bypass. Concentrations of the serum total (conjugated and unconjugated) bilirubin, alanine aminotransferase, aspartate aminotransferase, lactate dehydrogenase, alkaline phosphatase, albumin, globulin, and serum haptoglobin were measured before the operation and again on the first, second, and seventh postoperative days. Postoperative hyperbilir- ubinemia was defined as occurrence of a serum total bilirubin concentration of more than 3 mg/ dl in any measurement during the postoperative period. Logistic regression was done to identify possible risk factors for postoperative hyperbilirubinemia. Overall incidence of postoperative hyperbilirubinemia was 35.1 %; the incidence of postoperative hyperbilirubinemia was higher in patients whose valves were replaced with mechanical prostheses than in those without prostheses (p < 0.00001). In patients with postoperative hyperbilirubinemia, 70% of the increase of total bilirubin on the first postoperative day came about from an increase in unconjugated bilirubin. Serum haptoglobin decreased significantly at the same time (p < 0.01). Development of the postoperative hyperbilirubinemia was associated with a higher mortality (5.6% versus 0.5%, p < 0.01) and higher frequency of use of intraaortic balloon counterpulsation, especially for patients in whom the highest postoperative total bilirubin occurred after the first 2 days. The numbers of valves replaced, preoperative right atrial pressure, and preoperative total bilirubin concentration are the significant risk factors that, in combination, correctly predict the occurrence of postoperative hyperbilirubinemia in 80 % of the patients. We concluded that postopera- tive hyperbilirubinemia results mainly from an increase in unconjugated bilirubin and is associated with higher mortality, especially for patients in whom highest postoperative total bilirubin occurred late after operation. Patients with the higher preoperative right atrial pressure and total bilirubin level who then underwent multiple valve replacement procedures are at greater risk for development of postop- erative hyperbilirubinemia. (J THoRAc CARDIOVASC SURG 1994;108:429-36) Ming-Jiuh Wang, MD,a Anne Chao, MD,a Chi-Hsiang Huang, MD,a Chang-Her Tsai, MD,b Fang-Yue Lin, MD,b Shoei-Shen Wang, MD,b Chien-Chiang Liu, MD,a and Shu-Hsun Chu, MD,b Taipei, Taiwan From the Departments of Anesthesiology" and Surgery," National Tai- wan University Hospital, Taipei, Taiwan. Received for publication Aug. 24, 1993. Accepted for publication Jan. 9, 1994. Address for reprints: Shu-Hsun Chu, MD, Department of Surgery, National Taiwan University Hospital, 7 Chung-Shan South Rd., Taipei, Taiwan. Copyright © 1994 by Mosby-Year Book, Inc. 0022-5223/94 $3.00 + 0 12/1/54442 429
Hyperbilirubinemia after cardiac operation
Apr 11, 2023
Three hundred and two consecutive patients who had undergone cardiac operation for various cardiac
lesions were studied prospectively to evaluate the incidence, risk factors, and the associated mortality
of postoperative hyperbilirubinemia after cardiopulmonary bypass. Concentrations of the serum total
(conjugated and unconjugated) bilirubin, alanine aminotransferase, aspartate aminotransferase, lactate
dehydrogenase, alkaline phosphatase, albumin, globulin, and serum haptoglobin were measured before
the operation and again on the first, second, and seventh postoperative days
Welcome message from author
Postoperative hyperbilirubinemia was defined as occurrence of a serum total bilirubin concentration of more than 3 mg/dl in
any measurement during the postoperative period. Logistic regression was done to identify possible risk factors for postoperative hyperbilirubinemia
Transcript
Hyperbilirubinemia after cardiac operationHyperbilirubinemia after cardiac operation
Three hundred and two consecutive patients who had undergone cardiac operation for various cardiac lesions were studied prospectively to evaluate the incidence, risk factors, and the associated mortality of postoperative hyperbilirubinemia after cardiopulmonary bypass. Concentrations of the serum total (conjugated and unconjugated) bilirubin, alanine aminotransferase, aspartate aminotransferase, lactate dehydrogenase, alkaline phosphatase, albumin, globulin, and serum haptoglobin were measured before the operation and again on the first, second, and seventh postoperative days. Postoperative hyperbilir ubinemia was defined as occurrence of a serum total bilirubin concentration of more than 3 mg/dl in any measurement during the postoperative period. Logistic regression was done to identify possible risk factors for postoperative hyperbilirubinemia. Overall incidence of postoperative hyperbilirubinemia was 35.1 %; the incidence of postoperative hyperbilirubinemia was higher in patients whose valves were replaced with mechanical prostheses than in those without prostheses (p < 0.00001). In patients with postoperative hyperbilirubinemia, 70% of the increase of total bilirubin on the first postoperative day came about from an increase in unconjugated bilirubin. Serum haptoglobin decreased significantly at the same time (p < 0.01). Development of the postoperative hyperbilirubinemia was associated with a higher mortality (5.6% versus 0.5%, p < 0.01) and higher frequency of use of intraaortic balloon counterpulsation, especially for patients in whom the highest postoperative total bilirubin occurred after the first 2 days. The numbers of valves replaced, preoperative right atrial pressure, and preoperative total bilirubin concentration are the significant risk factors that, in combination, correctly predict the occurrence of postoperative hyperbilirubinemia in 80 % of the patients. We concluded that postopera tive hyperbilirubinemia results mainly from an increase in unconjugated bilirubin and is associated with higher mortality, especially for patients in whom highest postoperative total bilirubin occurred late after operation. Patients with the higher preoperative right atrial pressure and total bilirubin level who then underwent multiple valve replacement procedures are at greater risk for development of postop erative hyperbilirubinemia. (J THoRAc CARDIOVASC SURG 1994;108:429-36)
Ming-Jiuh Wang, MD,a Anne Chao, MD,a Chi-Hsiang Huang, MD,a Chang-Her Tsai, MD,b Fang-Yue Lin, MD,b Shoei-Shen Wang, MD,b Chien-Chiang Liu, MD,a and Shu-Hsun Chu, MD,b Taipei, Taiwan
From the Departments of Anesthesiology" and Surgery," National Tai wan University Hospital, Taipei, Taiwan.
Received for publication Aug. 24, 1993.
Accepted for publication Jan. 9, 1994.
Address for reprints: Shu-Hsun Chu, MD, Department of Surgery, National Taiwan University Hospital, 7 Chung-Shan South Rd., Taipei, Taiwan.
Copyright © 1994 by Mosby-Year Book, Inc.
0022-5223/94 $3.00 + 0 12/1/54442
Hyperbilirubinemia was reported in the early studies1. 2 to
occur in about 10% of patients after cardiac operations with cardiopulmonary bypass (CPB). However, more
recent studies estimate the incidence of postoperative hyperbilirubinemia to be more than 20% or even as high as 40%.3-5 In addition, it has not been clear whether post operative hyperbilirubinemia resulted from increased
conjugated or unconjugated bilirubin and whether the incidence of postoperative hyperbilirubinemia differs among various cardiac operations.v? The occurrence of postoperative hyperbilirubinemia has been reported to be associated with mortality as high as 25% according to one study"; another study reported no association between development of postoperative hyperbilirubinemia and mortality or morbidity.' The purpose of this present study was ( I ) to examine the incidence and nature of postoper ative hyperbilirubinemia in patients undergoing different kinds of cardiac operations, (2) to identify possible preoperative risk factors for development of the postop erative hyperbilirubinemia, and (3) to determine the clinical significance of the postoperative hy perbilirubinemia with regard to the mortality and mor bidity.
Patients and methods
Patients older than 18 years of age who had undergone car diac operation with CPB from November 1991 to June 1993 were studied prospectively. Written informed consent was obtained from every patient, and the study protocol was approved by the Committee of Human Investigation of this institution. Patients were not selected with any predetermined criteria. Patients with preoperative hyperbilirubinemia, defined as total bilirubin concentrations of more than 2 ttig]dl, were also included. Patients were separated into fivegroups according to their surgical procedures, which included coronary artery by pass grafting (CABG), first-time valvular replacements, reop erations for valvular replacements, correction of congenital heart disease, and operations for combinations of CABG and valvular replacements.
Anesthesia was induced with fentanyl (30 to 100 ttg/kg), diazepam (0.2 to 0.3 mg/kg) or etodimate (0.3 to 0.4 rug/kg), and pancuronium or vecuronium (0.1 to 0.15 rng/kg). Halothane was not used because of its possible hepatotoxic effects. In every operation, the routine clinical monitors includ ed lead II and V5 electrocardiogram, the radial arterial line, the pulse oximeter, end-tidal carbon dioxide, nasopharyngeal and rectal temperatures, urine output via Foley catheter, the central venous pressure line, and the pulmonary artery catheters.
In all patients, moderate hypothermic CPB with lowest nasopharyngeal temperature around 25° C with pulsatile flow was instituted with the roller pump (Sarns 5000;Sarns 3M, Ann Arbor, Mich.) and either bubble (William Harvey H 1700; Bard, Billerica, Mass.) or membrane (Capiox E; Terumo, Tokyo, Japan) oxygenators were used. The perfusion flow was kept over 2.2 L/m2 during normothermia and over 1.8 L/m2
during hypothermia in every patient. The mean arterial pressure was kept between 50 to 100 mm Hg with phenylephrine and
The Journal of Thoracic and
Cardiovascular Surgery September 1994
phentolamine during CPB. Arterial blood gas was monitored routinely every hour or on any occasion when considered nec essary. The priming solution contained 1.5 to 2 L of lactated Ringer's solution, mannitol (300 mg/kg), heparin (2000 U/L), and Y2 to 1 unit of packed red cells if blood cardioplegic solution was indicated or the predicted hematocrit levelduring CPB was below 22%. Systemic heparin was given through the right atri um at a dose of 300 U/kg just before cannulation. For valvular replacements operations, either a Carbomedics (Carbomedics, Inc., Austin, Tex.) or aSt. Jude Medical (St. Jude Medical, Inc., St. Paul, Minn.) mechanical prosthesis was used. The operating time, CPB time, aortic crossclamp time, types of oxygenator, quantity of packed red cells primed during CPB, and the types and the numbers of mechanical valves replaced were recorded.
Right atrial pressure and pulmonary artery pressure were obtained from the preoperative cardiac catheterization record. Preoperative blood samples for biochemical analysis of the hepatobiliary function were collected within 2 days before operation by venipuncture. After the operation, blood samples were obtained from the central venous line or venipuncture on the first, second, and seventh postoperative days. Bloodsamples were analyzed for concentrations of albumin, globulin, aspar tate aminotransferase, alanine aminotransferase, lactate de hydrogenase, alkaline phosphatase, total bilirubin, conjugated bilirubin, and unconjugated bilirubin by an automated bio chemical analyzer (Hitachi 736; Hitachi, Tokyo, Japan), which was calibrated and quality controlled periodically by this insti tution's clinical laboratory. Postoperative hyperbilirubinemia was defined as total bilirubin concentrations over 3 mg/dl (50 ttmol/dl) in anyone of the three postoperative measurements. Serum concentrations of haptoglobin were measured with poly clonal antibody and an automated machine (Beckman array 3600; Beckman, Brea, Calif.) of the four time points stated pre viously. After the operation, the days of hospitalization in the intensive care unit (lCU), the use of intraaortic balloon coun terpulsation (lABP), and the number of patients who died dur ing hospitalization were registered.
Data are shown as mean ± standard error of the mean. Stu dent's t test was used to compare the means between those patients with or without postoperative hyperbilirubinemia. One-way analysis of variance with post hoc Student-Newman Keuls multiple comparisons test compared the means among three or more groups of patients. x 2 Test was used to analyze the proportions between patients with and patients without postop erative hyperbilirubinemia, and Bonferroni's correction was done for comparisons among more than two groups. Fisher's exact test was used if the expected frequency was less than 5. The Wilcoxon-matched pair rank-sum test was used to compare the haptoglobin measurement data, which included undetect able values in postoperative samples. The preoperative variables including types of oxygenator, amount of blood priming, num bers of valves replaced, types of operation, CPB time, operation time, aortic clamp time, and the results of preoperative liver function tests were evaluated with the logistic regression model to identify predictors for the development of postoperative hyperbilirubinemia by the forward-likelihood ratio method with the SPSS software (SPSS Inc., Chicago, 111.).
Results
Incidence and severity. Among the 302 patients enrolled in this study, 29 had preoperative hyperbiliru-
The Journal of Thoracic and Cardiovascular Surgery
Volume 108, Number 3 Wang et at. 4 3 I
binemia. Demographic data for the patients among dif ferent disease categories are shown in Table I. The inci dence of the postoperative hyperbilirubinemia in these patients is shown in Fig. I. The overall incidence of post operative hyperbilirubinemia was 35.1% and, if the 29 patients with preoperative hyperbilirubinemia were excluded, the postoperative incidence was 28.9% (79 of 273). For patients with preoperative hyperbilirubinemia, the incidence ofthe postoperative hyperbilirubinemia was 93.1% (Fig. I, B). Among the patients with preoperative
Fig. 1. The incidenceof the postoperativehyperbilirubinemia among different disease categories (A), in patients with and without valves replaced with mechanical prostheses (B, left), and in patients with and without preoperativehyperbilirubine mia (B, right). CHD, Operation for congenital heart disease; Complex, combination of CABG and valvular replacement procedures; PH, postoperative hyperbilirubinemia; Prehigh, patientswith preoperativebilirubinconcentrationover2 mgjdl; Redo, reoperations for valve replacements; Valve, first-time valve replacement; VR, valves replaced with mechanical pros theses.*p < 0.0001 as compared with the patients with CABG; #p < 0.0001 as compared with the patients with CHD; @p < 0.001 between the patients with and without valves replacedor between the Prehigh and non-Prehigh patients.
Day 7Day 2
Preop
2
3
4
o
7
6
rn c: o ~ CO ~-c: Q) o c: 8 c: :.0 :::J ~
..-... -l 5 "'0-0> E--
Fig. 2. The perioperativechanges of serum bilirubin concen trations. The asterisks indicate the significant differences between the values of the preoperative and first postoperative day. BC, Conjugated bilirubin; BT, total bilirubin; BU, uncon jugated bilirubin;PH, postoperativehyperbilirubinemia;Preop, preoperative;Day 1, 2, 7, first, second, and seventh postopera tive days, respectively.
hyperbilirubinemia, 70.4% had severe postoperative hyperbilirubinemia with highest total bilirubin concen tration greater than 6 mg/dl, whereas in other patients only 19% had a highest total bilirubin concentration greater than 6 mg/dl (p < 0.0001). The highest total bil irubin concentration was significantly greater in patients with preoperative hyperbilirubinemia than patients with out (l2.98 ± 15.94 versus 2.99 ± 4.01,p<0.001).
@
_ PH patients
'"' eft. 60 '-"
432 Wang et al. The Journal of Thoracic and
Cardiovascular Surgery September 1994
CABG Valve Redo CHD Complex Total
No. of patients 90 (I) 99 (II) 74 (16) 19 (0) 20 (I) 302 Age (yr) 60.3 ± 0.9 48.7 ± 1.3 48.8 ± 1.5 34.2 ± 2.3 60.35 ± 3.0 51.87 ± 0.8 Sex (M/F) 69:21 49:50 31:43 7:12 14:6 170:132 Body surface area (rn-) 1.75 ± 0.01 1.59 ± 0.02 1.56 ± 0.01 1.55 ± 0.04 1.65 ± 0.04 1.63 ± 0.01
Values are shown as mean ± standard error of the mean; the number in parentheses indicates number of patients with preoperative elevated bilirubin level. Valve, First-time valvular replacements; Redo, reoperations for valvular replacements; CHD, congenital heart disease; Complex, combination of coronary artery bypass grafting and valvular replacement.
Table II. Perioperative changes in bilirubin concentration
Table IV. Mortality and morbidity and the time of the highest bilirubin level
Values were expressed as mean ± standard error of the mean.
*p < 0.001 between patients with nonpostoperative hyperbilirubinemia and those with postoperative hyperbilirubinemia.
Patients with Patients with nonpostoperative postoperative
hyperbilirubinemia hyperbilirubinemia
Values are shown as mean ± standard error of the mean.
*p < 0.01 between patients with nonpostoperative hyperbilirubinemia and those with postoperative hyperbilrubinemia.
tp < 0.05 between patients with nonpostoperative hyperbilirubinemia and those with postoperative hyperbilirubinemia.
50% (5/10)* 40% (4/10)* 15.6 ± 7.63 17.8 ± 5.0* 8.58 ± 2.73*
Peak after day 2
6.3 ± 0.5* 1.57 ± 0.197*
Mortality IABP frequency ICU stay (days) Peak bilirubin (mg/dl) Day 7 unconjugated bilirubin
(mg/dl) Day 7 unconjugated bilirubin
proportion
Values are shown as means ± standard error of the mean. Unconjugated biliru bin proportion, Proportion of unconjugated bilirubin in the total bilirubin concen tration; day 2 and 7, second and seventh postoperative days, respectively.
*p <0.001 between the patients with peak before day 2 and those with peak after day 2.
Nature of the hyperbilirubinemia. Perioperative changes of the total, unconjugated and conjugated, bili rubin concentrations are shown in Fig, 2 and Table II. On the first postoperative day, the total and unconjugated bilirubin concentrations increased in both the patients with postoperative hyperbilirubinemia and those with nonpostoperative hyperbilirubinemia compared with pre operative levels (p < 0.001). For patients with postoper ative hyperbilirubinemia, 70% of the increased total bil irubin was from an increase of unconjugated bilirubin. Serum haptoglobin concentrations decreased significant lyon the first postoperative day in both groups of patients (p < 0.0001). In patients with postoperative hyperbiliru binemia, 60.3% reached peak total bilirubin concentra tion on the first postoperative day, 30.1% on the second day, and 9.4% on the seventh day. Proportions of uncon jugated bilirubin from total bilirubin at the peak level were 0.73 ± 0.01,0.62 ± 0.03, and 0.53 ± 0.Q3 for the patients who reached their peak total bilirubin levelat the first, second, and seventh postoperative days (p < 0.05), respectively. The time at which the peak bilirubin level was reached did not differ between the patients with and without preoperative hyperbilirubinemia.
Mortality and morbidity. Mortality and morbidity of these patients are shown in Table III. The only case that ultimately resulted in death in the nonpostoperative
5.6%* 8.6%
7.3 ± 14.5t
2.10 ± 1.50*
3.1 ± 0.23*
7.94 ± 10.47*
0.70 ± 0.05
Total bilirubin 0.84 ± 0.12* increase on first postoperative day
Unconjugated bilirubin 0.68 ± 0.70* increase on first postoperative day
Proportion of unconjugated 0.64 ± 0.03 bilirubin increase
Postoperative highest 1.84 ± 0.87* bilirubin
were excluded. Seventeen patients underwent mitral reconstructive operations, and the incidence of postoper ative hyperbilirubinemia for these patients was similar to that of patients who underwent valvular replacement procedures (41.2% versus 47.6%, p > 0.6). For the 106 patients with postoperative hyperbilirubinemia, 84% had the onset of hyperbilirubinemia on the first postoperative day, 11.3% on the second day, and 4.7% on the seventh day.
The Journal of Thoracic and Cardiovascular Surgery Volume 108, Number 3
Table V. Analysis ofpossible risk factors for postoperative hyperbilirubinemia
Wang et al. 4 3 3
Possible risk factors
Preoperative total bilirubin (rng/dl) Conjugated bilirubin (rng/dl) Ratio of total bilirubin versus conjugated bilirubin Age (yr) Body surface area (m-) Right atrial pressure (mm Hg) Pulmonary artery pressure (mm Hg) Blood primed in CPB (Unit) Operation time (min) CPB time (min) Ischemic time (min) Preoperative albumin level (gm/dl) Preoperative globulin level (gm/dl) Alkaline phosphatase (U/d!) Lactate dehydrogenase (U Idl) Alanine aminotransferase (U Id!) Aspartate aminotransferase (U Idl) Lowest nasopharyngeal temperature (0 C)
Nonpostoperative hyperbilirubinemia
22.5 ± 1.2 0.92 ± 0.04
268.9 ± 6.0 116.2 ± 3.1 66.6 ± 2.2 3.82 ± 0.03 2.77 ± 0.47
120.6 ± 3.2 477.7 ± 18.8
Postoperative hyperbilirubinemia
1.66 ± 0.11 0.58 ± 0.05 3.47 ± 0.1 51.6 ± 1.24 1.60 ± 0.02 10.4 ± 0.8
31 ± 1.4 0.91 ± 0.07
284.9 ± 7.9 130.3 ± 4.2 77.1 ± 3.2 3.82 ± 0.07 2.94 ± 0.71
156.6 ± 8.9 657.2 ± 68.7
p Value
<0.001 <0.005
NS NS
NS NS
Values are shown as mean ± standard error of the mean. NS, Not significant.
hyperbilirubinemia group involved reoperations for dys functional aortic porcine valves. The patient's serum total bilirubin concentration increased from a preoperative level of 0.6 mg/dl to 1.7 mg/dl on the first postoperative morning and decreased to 0.9 tug]dl on the seventh post operative day. He died 24 days after the operation of res piratory failure. Among six operative mortality cases in the postoperative hyperbilirubinemia group, five involved an increase in total bilirubin immediately after operation (four on the first postoperative day), and in three cases the highest level (over 25 mg/dl) was reached on the seventh postoperative day. Four of these six patients underwent reoperations for valvular replacements, whereas the oth er two patients received CABG. IABP was used in the first or second postoperative day in five of these patients, and inotropic agents including dobutamine (6.78 ± 1.45 f.Lg/kg per minute) and dopamine (5.64 ± 0.72 f.Lg/kg per minute) were given to all patients during their stay in the ICU.
The onset time of postoperative hyperbilirubinemia was not associated with mortality, frequency ofIABP use, or days of ICU hospitalization. However, mortality and frequency of IABP use were significantly lower in patients whose highest total bilirubin occurred immediately after operation than in patients whose highest total bilirubin level was reached on the seventh postoperative day. A later appearance of peak total bilirubin level is associated with higher peak total bilirubin concentration (Table IV). Of the ten patients whose peak bilirubin concentration developedon the seventh postoperative day, seven patients had a steady increase in bilirubin concentrations from the
first postoperative day. Five of these seven patients died with clinical evidence of cardiac failure and high biliru bin concentration on the seventh postoperative day (mean 24.4 mg/dl, see previous data). Aortic crossclamp time, CPB time, preoperative total, unconjugated and conju gated bilirubin concentrations, and the values of other liver function tests did not differ in patients with postop erative hyperbilirubinemia whether their peak bilirubin level was reached before or after the second postoperative day (data not shown).
Possible risk factors for postoperative hyperbiliru binemia. No association of gender or types of oxygenators used was found with the occurrence of postoperative hyperbilirubinemia. No hypoxemic episodes occurred be fore or during the surgical procedures for any patient. Analysis of other possible risk factors is shown in Table V. The results oflogistic regression are shown in Table VI. Right atrial pressure, the numbers of valves replaced, and preoperative total bilirubin concentrations were identified as the important predictors for development of postoper ative hyperbilirubinemia. Combinations of these three preoperative risk factors could correctly predict develop ment of postoperative hyperbilirubinemia in 80.61% of the patients.
Changes of other liver function tests. The results of other liver function tests are shown in Fig. 3. The patterns of changes were similar in patients with and without postoperative hyperbilirubinemia. Large standard errors on the second postoperative day (Fig. 3, E and F) are attributable to one female patient whose alanine ami notransferase and aspartate aminotransferase reached
434 Wangetal. The Journal of Thoracic and
Cardiovascular Surgery September 1994
Table VI. The results of logistic regression for the risk factors of the postoperative hyperbilirubinemia
Variables {3 SE Significance R
Valve numbers replaced 0.6658 0.2963 0.024 0.1487 Right atrial pressure 0.1093 0.0339 OJJ061 0.1479 Preoperative total bilirubin 1.3752 0.3707 0.0002 0.2161 Constant - 3.4525 0.4872 0.0001
{3, Regression coefficient; SE, standard error; R, partial correlation coefficient.
more than 10,000 and 3,000 Uj dl with serum total bili rubin as high as 45 mgjdl on the second postoperative day. Her severe jaundice resolved within 2 weeks, and she was discharged 3 weeks after operation with the alanine aminotransferase and aspartate aminotransferase levels lower than 50 Ujdl and a total bilirubin concentration of 1.3 mgjdl.
Discussion
This prospective study showed that the incidence of postoperative hyperbilirubinemia is 35.1% (a figure sim ilar to the data reported by Chu' and Collins4 and their associates and Kleptko, Base, and Muller") despite progress in techniques of cardiac operations, CPB, and cardiac anesthesia in the past 10 years. Some arguments exist about whether incidence differs among different disease categories.': 5 This study clearly showed that the incidence of postoperative hyperbilirubinemia is signifi cantly higher in patients with valvular heart disease, regardless of whether the valve was replaced or…
Three hundred and two consecutive patients who had undergone cardiac operation for various cardiac lesions were studied prospectively to evaluate the incidence, risk factors, and the associated mortality of postoperative hyperbilirubinemia after cardiopulmonary bypass. Concentrations of the serum total (conjugated and unconjugated) bilirubin, alanine aminotransferase, aspartate aminotransferase, lactate dehydrogenase, alkaline phosphatase, albumin, globulin, and serum haptoglobin were measured before the operation and again on the first, second, and seventh postoperative days. Postoperative hyperbilir ubinemia was defined as occurrence of a serum total bilirubin concentration of more than 3 mg/dl in any measurement during the postoperative period. Logistic regression was done to identify possible risk factors for postoperative hyperbilirubinemia. Overall incidence of postoperative hyperbilirubinemia was 35.1 %; the incidence of postoperative hyperbilirubinemia was higher in patients whose valves were replaced with mechanical prostheses than in those without prostheses (p < 0.00001). In patients with postoperative hyperbilirubinemia, 70% of the increase of total bilirubin on the first postoperative day came about from an increase in unconjugated bilirubin. Serum haptoglobin decreased significantly at the same time (p < 0.01). Development of the postoperative hyperbilirubinemia was associated with a higher mortality (5.6% versus 0.5%, p < 0.01) and higher frequency of use of intraaortic balloon counterpulsation, especially for patients in whom the highest postoperative total bilirubin occurred after the first 2 days. The numbers of valves replaced, preoperative right atrial pressure, and preoperative total bilirubin concentration are the significant risk factors that, in combination, correctly predict the occurrence of postoperative hyperbilirubinemia in 80 % of the patients. We concluded that postopera tive hyperbilirubinemia results mainly from an increase in unconjugated bilirubin and is associated with higher mortality, especially for patients in whom highest postoperative total bilirubin occurred late after operation. Patients with the higher preoperative right atrial pressure and total bilirubin level who then underwent multiple valve replacement procedures are at greater risk for development of postop erative hyperbilirubinemia. (J THoRAc CARDIOVASC SURG 1994;108:429-36)
Ming-Jiuh Wang, MD,a Anne Chao, MD,a Chi-Hsiang Huang, MD,a Chang-Her Tsai, MD,b Fang-Yue Lin, MD,b Shoei-Shen Wang, MD,b Chien-Chiang Liu, MD,a and Shu-Hsun Chu, MD,b Taipei, Taiwan
From the Departments of Anesthesiology" and Surgery," National Tai wan University Hospital, Taipei, Taiwan.
Received for publication Aug. 24, 1993.
Accepted for publication Jan. 9, 1994.
Address for reprints: Shu-Hsun Chu, MD, Department of Surgery, National Taiwan University Hospital, 7 Chung-Shan South Rd., Taipei, Taiwan.
Copyright © 1994 by Mosby-Year Book, Inc.
0022-5223/94 $3.00 + 0 12/1/54442
Hyperbilirubinemia was reported in the early studies1. 2 to
occur in about 10% of patients after cardiac operations with cardiopulmonary bypass (CPB). However, more
recent studies estimate the incidence of postoperative hyperbilirubinemia to be more than 20% or even as high as 40%.3-5 In addition, it has not been clear whether post operative hyperbilirubinemia resulted from increased
conjugated or unconjugated bilirubin and whether the incidence of postoperative hyperbilirubinemia differs among various cardiac operations.v? The occurrence of postoperative hyperbilirubinemia has been reported to be associated with mortality as high as 25% according to one study"; another study reported no association between development of postoperative hyperbilirubinemia and mortality or morbidity.' The purpose of this present study was ( I ) to examine the incidence and nature of postoper ative hyperbilirubinemia in patients undergoing different kinds of cardiac operations, (2) to identify possible preoperative risk factors for development of the postop erative hyperbilirubinemia, and (3) to determine the clinical significance of the postoperative hy perbilirubinemia with regard to the mortality and mor bidity.
Patients and methods
Patients older than 18 years of age who had undergone car diac operation with CPB from November 1991 to June 1993 were studied prospectively. Written informed consent was obtained from every patient, and the study protocol was approved by the Committee of Human Investigation of this institution. Patients were not selected with any predetermined criteria. Patients with preoperative hyperbilirubinemia, defined as total bilirubin concentrations of more than 2 ttig]dl, were also included. Patients were separated into fivegroups according to their surgical procedures, which included coronary artery by pass grafting (CABG), first-time valvular replacements, reop erations for valvular replacements, correction of congenital heart disease, and operations for combinations of CABG and valvular replacements.
Anesthesia was induced with fentanyl (30 to 100 ttg/kg), diazepam (0.2 to 0.3 mg/kg) or etodimate (0.3 to 0.4 rug/kg), and pancuronium or vecuronium (0.1 to 0.15 rng/kg). Halothane was not used because of its possible hepatotoxic effects. In every operation, the routine clinical monitors includ ed lead II and V5 electrocardiogram, the radial arterial line, the pulse oximeter, end-tidal carbon dioxide, nasopharyngeal and rectal temperatures, urine output via Foley catheter, the central venous pressure line, and the pulmonary artery catheters.
In all patients, moderate hypothermic CPB with lowest nasopharyngeal temperature around 25° C with pulsatile flow was instituted with the roller pump (Sarns 5000;Sarns 3M, Ann Arbor, Mich.) and either bubble (William Harvey H 1700; Bard, Billerica, Mass.) or membrane (Capiox E; Terumo, Tokyo, Japan) oxygenators were used. The perfusion flow was kept over 2.2 L/m2 during normothermia and over 1.8 L/m2
during hypothermia in every patient. The mean arterial pressure was kept between 50 to 100 mm Hg with phenylephrine and
The Journal of Thoracic and
Cardiovascular Surgery September 1994
phentolamine during CPB. Arterial blood gas was monitored routinely every hour or on any occasion when considered nec essary. The priming solution contained 1.5 to 2 L of lactated Ringer's solution, mannitol (300 mg/kg), heparin (2000 U/L), and Y2 to 1 unit of packed red cells if blood cardioplegic solution was indicated or the predicted hematocrit levelduring CPB was below 22%. Systemic heparin was given through the right atri um at a dose of 300 U/kg just before cannulation. For valvular replacements operations, either a Carbomedics (Carbomedics, Inc., Austin, Tex.) or aSt. Jude Medical (St. Jude Medical, Inc., St. Paul, Minn.) mechanical prosthesis was used. The operating time, CPB time, aortic crossclamp time, types of oxygenator, quantity of packed red cells primed during CPB, and the types and the numbers of mechanical valves replaced were recorded.
Right atrial pressure and pulmonary artery pressure were obtained from the preoperative cardiac catheterization record. Preoperative blood samples for biochemical analysis of the hepatobiliary function were collected within 2 days before operation by venipuncture. After the operation, blood samples were obtained from the central venous line or venipuncture on the first, second, and seventh postoperative days. Bloodsamples were analyzed for concentrations of albumin, globulin, aspar tate aminotransferase, alanine aminotransferase, lactate de hydrogenase, alkaline phosphatase, total bilirubin, conjugated bilirubin, and unconjugated bilirubin by an automated bio chemical analyzer (Hitachi 736; Hitachi, Tokyo, Japan), which was calibrated and quality controlled periodically by this insti tution's clinical laboratory. Postoperative hyperbilirubinemia was defined as total bilirubin concentrations over 3 mg/dl (50 ttmol/dl) in anyone of the three postoperative measurements. Serum concentrations of haptoglobin were measured with poly clonal antibody and an automated machine (Beckman array 3600; Beckman, Brea, Calif.) of the four time points stated pre viously. After the operation, the days of hospitalization in the intensive care unit (lCU), the use of intraaortic balloon coun terpulsation (lABP), and the number of patients who died dur ing hospitalization were registered.
Data are shown as mean ± standard error of the mean. Stu dent's t test was used to compare the means between those patients with or without postoperative hyperbilirubinemia. One-way analysis of variance with post hoc Student-Newman Keuls multiple comparisons test compared the means among three or more groups of patients. x 2 Test was used to analyze the proportions between patients with and patients without postop erative hyperbilirubinemia, and Bonferroni's correction was done for comparisons among more than two groups. Fisher's exact test was used if the expected frequency was less than 5. The Wilcoxon-matched pair rank-sum test was used to compare the haptoglobin measurement data, which included undetect able values in postoperative samples. The preoperative variables including types of oxygenator, amount of blood priming, num bers of valves replaced, types of operation, CPB time, operation time, aortic clamp time, and the results of preoperative liver function tests were evaluated with the logistic regression model to identify predictors for the development of postoperative hyperbilirubinemia by the forward-likelihood ratio method with the SPSS software (SPSS Inc., Chicago, 111.).
Results
Incidence and severity. Among the 302 patients enrolled in this study, 29 had preoperative hyperbiliru-
The Journal of Thoracic and Cardiovascular Surgery
Volume 108, Number 3 Wang et at. 4 3 I
binemia. Demographic data for the patients among dif ferent disease categories are shown in Table I. The inci dence of the postoperative hyperbilirubinemia in these patients is shown in Fig. I. The overall incidence of post operative hyperbilirubinemia was 35.1% and, if the 29 patients with preoperative hyperbilirubinemia were excluded, the postoperative incidence was 28.9% (79 of 273). For patients with preoperative hyperbilirubinemia, the incidence ofthe postoperative hyperbilirubinemia was 93.1% (Fig. I, B). Among the patients with preoperative
Fig. 1. The incidenceof the postoperativehyperbilirubinemia among different disease categories (A), in patients with and without valves replaced with mechanical prostheses (B, left), and in patients with and without preoperativehyperbilirubine mia (B, right). CHD, Operation for congenital heart disease; Complex, combination of CABG and valvular replacement procedures; PH, postoperative hyperbilirubinemia; Prehigh, patientswith preoperativebilirubinconcentrationover2 mgjdl; Redo, reoperations for valve replacements; Valve, first-time valve replacement; VR, valves replaced with mechanical pros theses.*p < 0.0001 as compared with the patients with CABG; #p < 0.0001 as compared with the patients with CHD; @p < 0.001 between the patients with and without valves replacedor between the Prehigh and non-Prehigh patients.
Day 7Day 2
Preop
2
3
4
o
7
6
rn c: o ~ CO ~-c: Q) o c: 8 c: :.0 :::J ~
..-... -l 5 "'0-0> E--
Fig. 2. The perioperativechanges of serum bilirubin concen trations. The asterisks indicate the significant differences between the values of the preoperative and first postoperative day. BC, Conjugated bilirubin; BT, total bilirubin; BU, uncon jugated bilirubin;PH, postoperativehyperbilirubinemia;Preop, preoperative;Day 1, 2, 7, first, second, and seventh postopera tive days, respectively.
hyperbilirubinemia, 70.4% had severe postoperative hyperbilirubinemia with highest total bilirubin concen tration greater than 6 mg/dl, whereas in other patients only 19% had a highest total bilirubin concentration greater than 6 mg/dl (p < 0.0001). The highest total bil irubin concentration was significantly greater in patients with preoperative hyperbilirubinemia than patients with out (l2.98 ± 15.94 versus 2.99 ± 4.01,p<0.001).
@
_ PH patients
'"' eft. 60 '-"
432 Wang et al. The Journal of Thoracic and
Cardiovascular Surgery September 1994
CABG Valve Redo CHD Complex Total
No. of patients 90 (I) 99 (II) 74 (16) 19 (0) 20 (I) 302 Age (yr) 60.3 ± 0.9 48.7 ± 1.3 48.8 ± 1.5 34.2 ± 2.3 60.35 ± 3.0 51.87 ± 0.8 Sex (M/F) 69:21 49:50 31:43 7:12 14:6 170:132 Body surface area (rn-) 1.75 ± 0.01 1.59 ± 0.02 1.56 ± 0.01 1.55 ± 0.04 1.65 ± 0.04 1.63 ± 0.01
Values are shown as mean ± standard error of the mean; the number in parentheses indicates number of patients with preoperative elevated bilirubin level. Valve, First-time valvular replacements; Redo, reoperations for valvular replacements; CHD, congenital heart disease; Complex, combination of coronary artery bypass grafting and valvular replacement.
Table II. Perioperative changes in bilirubin concentration
Table IV. Mortality and morbidity and the time of the highest bilirubin level
Values were expressed as mean ± standard error of the mean.
*p < 0.001 between patients with nonpostoperative hyperbilirubinemia and those with postoperative hyperbilirubinemia.
Patients with Patients with nonpostoperative postoperative
hyperbilirubinemia hyperbilirubinemia
Values are shown as mean ± standard error of the mean.
*p < 0.01 between patients with nonpostoperative hyperbilirubinemia and those with postoperative hyperbilrubinemia.
tp < 0.05 between patients with nonpostoperative hyperbilirubinemia and those with postoperative hyperbilirubinemia.
50% (5/10)* 40% (4/10)* 15.6 ± 7.63 17.8 ± 5.0* 8.58 ± 2.73*
Peak after day 2
6.3 ± 0.5* 1.57 ± 0.197*
Mortality IABP frequency ICU stay (days) Peak bilirubin (mg/dl) Day 7 unconjugated bilirubin
(mg/dl) Day 7 unconjugated bilirubin
proportion
Values are shown as means ± standard error of the mean. Unconjugated biliru bin proportion, Proportion of unconjugated bilirubin in the total bilirubin concen tration; day 2 and 7, second and seventh postoperative days, respectively.
*p <0.001 between the patients with peak before day 2 and those with peak after day 2.
Nature of the hyperbilirubinemia. Perioperative changes of the total, unconjugated and conjugated, bili rubin concentrations are shown in Fig, 2 and Table II. On the first postoperative day, the total and unconjugated bilirubin concentrations increased in both the patients with postoperative hyperbilirubinemia and those with nonpostoperative hyperbilirubinemia compared with pre operative levels (p < 0.001). For patients with postoper ative hyperbilirubinemia, 70% of the increased total bil irubin was from an increase of unconjugated bilirubin. Serum haptoglobin concentrations decreased significant lyon the first postoperative day in both groups of patients (p < 0.0001). In patients with postoperative hyperbiliru binemia, 60.3% reached peak total bilirubin concentra tion on the first postoperative day, 30.1% on the second day, and 9.4% on the seventh day. Proportions of uncon jugated bilirubin from total bilirubin at the peak level were 0.73 ± 0.01,0.62 ± 0.03, and 0.53 ± 0.Q3 for the patients who reached their peak total bilirubin levelat the first, second, and seventh postoperative days (p < 0.05), respectively. The time at which the peak bilirubin level was reached did not differ between the patients with and without preoperative hyperbilirubinemia.
Mortality and morbidity. Mortality and morbidity of these patients are shown in Table III. The only case that ultimately resulted in death in the nonpostoperative
5.6%* 8.6%
7.3 ± 14.5t
2.10 ± 1.50*
3.1 ± 0.23*
7.94 ± 10.47*
0.70 ± 0.05
Total bilirubin 0.84 ± 0.12* increase on first postoperative day
Unconjugated bilirubin 0.68 ± 0.70* increase on first postoperative day
Proportion of unconjugated 0.64 ± 0.03 bilirubin increase
Postoperative highest 1.84 ± 0.87* bilirubin
were excluded. Seventeen patients underwent mitral reconstructive operations, and the incidence of postoper ative hyperbilirubinemia for these patients was similar to that of patients who underwent valvular replacement procedures (41.2% versus 47.6%, p > 0.6). For the 106 patients with postoperative hyperbilirubinemia, 84% had the onset of hyperbilirubinemia on the first postoperative day, 11.3% on the second day, and 4.7% on the seventh day.
The Journal of Thoracic and Cardiovascular Surgery Volume 108, Number 3
Table V. Analysis ofpossible risk factors for postoperative hyperbilirubinemia
Wang et al. 4 3 3
Possible risk factors
Preoperative total bilirubin (rng/dl) Conjugated bilirubin (rng/dl) Ratio of total bilirubin versus conjugated bilirubin Age (yr) Body surface area (m-) Right atrial pressure (mm Hg) Pulmonary artery pressure (mm Hg) Blood primed in CPB (Unit) Operation time (min) CPB time (min) Ischemic time (min) Preoperative albumin level (gm/dl) Preoperative globulin level (gm/dl) Alkaline phosphatase (U/d!) Lactate dehydrogenase (U Idl) Alanine aminotransferase (U Id!) Aspartate aminotransferase (U Idl) Lowest nasopharyngeal temperature (0 C)
Nonpostoperative hyperbilirubinemia
22.5 ± 1.2 0.92 ± 0.04
268.9 ± 6.0 116.2 ± 3.1 66.6 ± 2.2 3.82 ± 0.03 2.77 ± 0.47
120.6 ± 3.2 477.7 ± 18.8
Postoperative hyperbilirubinemia
1.66 ± 0.11 0.58 ± 0.05 3.47 ± 0.1 51.6 ± 1.24 1.60 ± 0.02 10.4 ± 0.8
31 ± 1.4 0.91 ± 0.07
284.9 ± 7.9 130.3 ± 4.2 77.1 ± 3.2 3.82 ± 0.07 2.94 ± 0.71
156.6 ± 8.9 657.2 ± 68.7
p Value
<0.001 <0.005
NS NS
NS NS
Values are shown as mean ± standard error of the mean. NS, Not significant.
hyperbilirubinemia group involved reoperations for dys functional aortic porcine valves. The patient's serum total bilirubin concentration increased from a preoperative level of 0.6 mg/dl to 1.7 mg/dl on the first postoperative morning and decreased to 0.9 tug]dl on the seventh post operative day. He died 24 days after the operation of res piratory failure. Among six operative mortality cases in the postoperative hyperbilirubinemia group, five involved an increase in total bilirubin immediately after operation (four on the first postoperative day), and in three cases the highest level (over 25 mg/dl) was reached on the seventh postoperative day. Four of these six patients underwent reoperations for valvular replacements, whereas the oth er two patients received CABG. IABP was used in the first or second postoperative day in five of these patients, and inotropic agents including dobutamine (6.78 ± 1.45 f.Lg/kg per minute) and dopamine (5.64 ± 0.72 f.Lg/kg per minute) were given to all patients during their stay in the ICU.
The onset time of postoperative hyperbilirubinemia was not associated with mortality, frequency ofIABP use, or days of ICU hospitalization. However, mortality and frequency of IABP use were significantly lower in patients whose highest total bilirubin occurred immediately after operation than in patients whose highest total bilirubin level was reached on the seventh postoperative day. A later appearance of peak total bilirubin level is associated with higher peak total bilirubin concentration (Table IV). Of the ten patients whose peak bilirubin concentration developedon the seventh postoperative day, seven patients had a steady increase in bilirubin concentrations from the
first postoperative day. Five of these seven patients died with clinical evidence of cardiac failure and high biliru bin concentration on the seventh postoperative day (mean 24.4 mg/dl, see previous data). Aortic crossclamp time, CPB time, preoperative total, unconjugated and conju gated bilirubin concentrations, and the values of other liver function tests did not differ in patients with postop erative hyperbilirubinemia whether their peak bilirubin level was reached before or after the second postoperative day (data not shown).
Possible risk factors for postoperative hyperbiliru binemia. No association of gender or types of oxygenators used was found with the occurrence of postoperative hyperbilirubinemia. No hypoxemic episodes occurred be fore or during the surgical procedures for any patient. Analysis of other possible risk factors is shown in Table V. The results oflogistic regression are shown in Table VI. Right atrial pressure, the numbers of valves replaced, and preoperative total bilirubin concentrations were identified as the important predictors for development of postoper ative hyperbilirubinemia. Combinations of these three preoperative risk factors could correctly predict develop ment of postoperative hyperbilirubinemia in 80.61% of the patients.
Changes of other liver function tests. The results of other liver function tests are shown in Fig. 3. The patterns of changes were similar in patients with and without postoperative hyperbilirubinemia. Large standard errors on the second postoperative day (Fig. 3, E and F) are attributable to one female patient whose alanine ami notransferase and aspartate aminotransferase reached
434 Wangetal. The Journal of Thoracic and
Cardiovascular Surgery September 1994
Table VI. The results of logistic regression for the risk factors of the postoperative hyperbilirubinemia
Variables {3 SE Significance R
Valve numbers replaced 0.6658 0.2963 0.024 0.1487 Right atrial pressure 0.1093 0.0339 OJJ061 0.1479 Preoperative total bilirubin 1.3752 0.3707 0.0002 0.2161 Constant - 3.4525 0.4872 0.0001
{3, Regression coefficient; SE, standard error; R, partial correlation coefficient.
more than 10,000 and 3,000 Uj dl with serum total bili rubin as high as 45 mgjdl on the second postoperative day. Her severe jaundice resolved within 2 weeks, and she was discharged 3 weeks after operation with the alanine aminotransferase and aspartate aminotransferase levels lower than 50 Ujdl and a total bilirubin concentration of 1.3 mgjdl.
Discussion
This prospective study showed that the incidence of postoperative hyperbilirubinemia is 35.1% (a figure sim ilar to the data reported by Chu' and Collins4 and their associates and Kleptko, Base, and Muller") despite progress in techniques of cardiac operations, CPB, and cardiac anesthesia in the past 10 years. Some arguments exist about whether incidence differs among different disease categories.': 5 This study clearly showed that the incidence of postoperative hyperbilirubinemia is signifi cantly higher in patients with valvular heart disease, regardless of whether the valve was replaced or…
Related Documents
