| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Eur J Cardiothorac Surg 2005;28:229-233
© 2005 Elsevier Science NL
Original articles |
a Department of Cardiothoracic Surgery, Thoraxcenter, Bd 156, Erasmus MC, P.O. Box 2040, 3000 CA Rotterdam, The Netherlands
b Department of Pediatric Cardiology, Erasmus M C, Rotterdam, The Netherlands
Received 22 December 2004; received in revised form 29 March 2005; accepted 25 April 2005.
* Corresponding author. Tel.: +31 10 4635411; fax: +31 10 4633993. (Email: r.vanthiel{at}erasmusmc.nl).
Abstract
Objective: To determine occurrence, causes and associated mortality of postoperative metabolic alkalosis in pediatric cardiac surgery. Methods: We retrospectively analyzed clinical and biochemical variables of 186 consecutive cardiac operations other than ductal ligations on children less than 2 years old during the years 1999 and 2000. Metabolic alkalosis was defined as a pH>7.48 corrected for PCO2, with a base excess 
5 on two or more consecutive measurements during an 8h period. Results: Median age was 15 weeks [range 2 days–95 weeks] and median weight 4.5kg [range 2.1–15.7kg]. In 157 cases, cardiopulmonary bypass was used. In 92 [49%] procedures, metabolic alkalosis occurred with the highest corrected pH 24.3h after operation. Multivariate regression analysis associated age [P<0.001], cardiopulmonary bypass [P<0.001] and preoperative ductal dependency [P=0.04] with postoperative metabolic alkalosis. Of the surgical procedures the arterial switch for transposition of the great arteries [n=19] was strongly associated with metabolic alkalosis [100%, P<0.001]. Hemodilution appeared to enhance the development of alkalosis: those who experienced alkalosis had been hemodiluted to a greater extent [P=0.007]. Nearly 95% of patients experienced some increase in bicarbonate, but patients with metabolic alkalosis experienced more than those without [5.9 versus 3.5mmol/l, P<0.001]. There were four postoperative deaths, only one coincidental with metabolic alkalosis. Conclusions: Metabolic alkalosis has a high incidence after pediatric cardiac surgery, strongly associated with younger age, cardiopulmonary bypass, preoperative ductal dependency and perioperative hemodilution. Early recognition allows for timely therapeutic intervention.
Key Words: Alkalosis • Pediatrics • Cardiac surgery • Cardiopulmonary bypass • Ductus arteriosus
1. Introduction
Metabolic alkalosis (MA) is the most common acid–base disorder in hospitalized patients [1,2]. MA may lead to various clinical disturbances [3] and in different ICU settings, it has been found to be associated with a significant mortality [1,4].
In our experience, the occurrence of this metabolic disorder after cardiac surgery is much more frequent in neonates and infants compared to adults. Reports on the incidence and potential risk factors for MA are rare, especially for pediatric patients after cardiac surgery. Wong and Chundu studied 56 pediatric patients and concluded that younger age and chloride depletion are important risk factors in the development of MA in infants and young children [5]. We previously reported on two patients with postoperative acute cardiac failure while experiencing severe MA [6]. In order to determine a relation between such complications and MA in pediatric cardiac surgery in our institution, we retrospectively evaluated the incidence of MA and potential risk factors, and the postoperative outcome.
2. Materials and methods
From January 1999 to December 2000, 228 cardiothoracic operations were performed on infants younger than 2 years of age in the Erasmus MC in Rotterdam, The Netherlands. All simple ductal ligations in premature patients (N=30) were excluded, as such patients follow an entirely different clinical course. Twelve additional patients were excluded from the study for the following reasons: seven children died during surgery (not due to acid–base balance disturbances) and of five otherwise uneventful procedures, there were not enough acid–base data available for adequate analysis. The 186 remaining procedures were performed on 167 patients. Thirteen children had two operations; three patients underwent three procedures. In 157 procedures, cardiopulmonary bypass was used, with an additional circulatory arrest in 17. In our institution, alpha-stat pH-management is used. Cardiopulmonary bypass was applied with a flow of 2.4l/m2. Priming additives included bicarbonate, 0–15mmol (median 5.0±2.2mmol), varying according to priming volume, erythrocyte volume and patient size. In cases necessitating circulatory arrest, we aimed for a central body temperature of 18°C; in other cases our goal was a central temperature between 28 and 32°C, depending on the type and duration of the procedure. Target hematocrits on bypass were 25–28%.
The following preoperative data were collected: age (weeks) and weight (kg) at operation, ductal dependency (yes/no), use of loopdiuretics (yes/no) and serum creatinin level (µmol/l), preoperative hematocrits and acid–base data, if available within 1 week before operation.
The following intra- and postoperative data were collected: type of procedure performed, first hematocrits after induction of anesthesia, duration of extracorporeal circulation (min), cross-clamp time (min) and circulatory arrest during surgery (min), bicarbonate (mmol/kg) and citrate blood (ml/kg) administered during surgery, postoperative serum creatinin level (µmol/l), maximum amount of inotropes (µg/kg per min) received up to 48h postoperatively, total amount of ACE-inhibitor (mg/kg) received up to 48h postoperatively, potassium supplementation (mg/kg per day) received on the day of operation and up to 48h postoperatively, urine production (ml/kg per h) on the day of operation and up to 48h postoperatively, total calcium (mmol/l) and albumin (g/l) levels measured on the first day postoperative, highest and lowest chloride and ionized calcium levels (mmol/l) within 48h postoperative, and the time after surgery they were measured (h), acid–base data to assess the presence of metabolic alkalosis within 48h postoperative and blood lactates in the same period.
Of the postoperative period, the day of operation and the following 48h were considered for evaluation as this is in our own experience the period with the highest occurrence of MA. We reviewed only the laboratory acid–base data of arterial blood samples obtained within 48h after surgery. Metabolic alkalosis was defined as an arterial pH greater than 7.48 with a base excess of 5 at the same time, both in the setting of a normal PCO2. Each patient with these criteria on two or more consecutive measurements within 8h was considered to have MA. This is in accordance with the definition Wong and Chundu used in their study [5]. The pH was corrected for PCO2 by using the Henderson-Hasselbach equation [7,8]. Standard bicarbonate value was defined as the concentration of bicarbonate measured in a blood sample with a PCO2 of 40Torr (5.3kPa) [9]. In our laboratory, the standard bicarbonate is calculated.
Treatment of significant MA, while on artificial ventilation, consisted of respiratory compensation to keep pH within normal range, with a PCO2 up to 54Torr (7.2kPa). Chloride supplementation was given with sodium-, potassium-, or calcium chloride when indicated by low values of the respective cations. No NH4Cl or HCl solutions were used. In addition, there were very few instances in which carbonic anhydrase inhibitors were administered.
All data were analyzed using SPSS 11.0 for Windows (SPSS, Chicago, IL). Data are reported as mean±1SD. Means were compared by the unpaired T-test. The 
2-test or Fisher's Exact test was used to compare categorical variables. All tests were two-sided, with an 
-level of 0.05. Logistic regression was used to study potential determinants for MA. Multiple logistic regression was used to study independent determinants for MA. The final model was obtained using the stepwise backward method with criteria for entry P<0.05 and removal P>0.10. An additional logistic regression analysis was done for those procedures that included the use of CPB.
3. Results
The median age of the children at the time of operation was 15 weeks (range 0–95) and the median weight 4.5kg (range 2.1–15.7). As age and weight bear a strong relation (confirmed in our study: Pearson correlation value: 0.88, P<0.001), obviously weight can be assumed to be similarly associated with MA. Therefore, we did not include weight as a separate factor into the multivariate analysis of the CPB subgroup.
Forty-two patients were dependent on the arterial duct before surgery. Preoperatively 118 patients received furosemide with a mean of 2.3 (±1.3) mg/kg per day (range 0.3–11). The preoperative creatinine level (n=181) was 32 (±14) µmol/l (range 11–118). Patient characteristics are displayed in Table 1 .
|
Within 30 days postoperatively, four children died (mean after 13 days). Two died of circulatory failure related to pulmonary hypertension, respectively, after receiving Norwood palliation for a hypoplastic left heart syndrome (immediately after operation) and a correction of pulmonary venous stenosis in a case of pulmonary artery atresia (after 19 days), while one received a modified Blaälock–Taussig shunt for a hypoplastic right ventricle (18 days postoperatively) and another was operated on for congenital mitral insufficiency and stenosis (after 15 days). Only the latter had developed MA within 48h after surgery, while dying 2 weeks later. Therefore, no relation between MA and mortality could be determined.
Analysis of patient and operative characteristics with respect to development of MA revealed certain relations depicted in Table 2 . The use of CPB, preoperative dependency on the duct, younger age and lower weight were associated with the development of MA and were included in multivariate logistic regression analysis. This showed that the use of CPB was the most important predictor of the occurrence of MA (Table 3 ). Intraoperatively administered bicarbonate (N=76, 1.2±1.1mmol) and citrate blood, use of inotropes and ACE-inhibitor, potassium supplementation, urine production and levels of calcium, albumin and chloride postoperatively showed no significance in univariate analysis. No bicarbonate was administered postoperatively. Too few postoperative (14.5%) lactates were measured, predominantly in the last period and obviously with too large a variation (1.8±5.2mmol/l), to relate to the development of alkalosis.
|
|
|
|
For 35% of the study population (N=63), there were also preoperative acid–base data to evaluate. This subpopulation was hospitalized, and significantly younger and more alkalotic than the rest. Still, for these patients the change in Ht during the induction of anesthesia was not significantly different from the rest of the population; their preoperative bicarbonate did not significantly differ from the first intraoperative value, and the changes in pH, base excess and bicarbonate after operation regarded from the preoperative measurement correlated strongly with the first intraoperative values (respective Pearson correlation coefficients 0.763, 0.803, and 0.822, P<0.001 for all). This supports our notion that the first intraoperative acid–base measurement, taken after induction of anesthesia, in spite of the hemodilution, can be viewed as representative of the preoperative acid–base status. Nevertheless, there was a positive correlation between the change in Ht and the occurrence of MA: those who were to develop MA proved to have been hemodilated significantly more than those who were not: 
Ht (MA versus non-MA): –0.068 versus –0.052 (–17 versus –13%) (P=0.007).
Furthermore, comparison of the first intraoperative acid–base values of the entire group (N=180) with postoperative data demonstrated a significant rise in mean pH, bicarbonate and base excess postoperatively, suggesting that all patients become alkalotic after surgery. Actually, for 95% (N=174) the standard bicarbonate increased. This was not due to bicarbonate administration or citrate blood given during the procedure. Table 6 depicts the acid–base status during and changes in between the sample periods.
|
Untreated MA is reported to be associated with increased morbidity and mortality especially in intensive care settings [1,3,4,10]. Insight into the prevalence and the mechanisms of MA, however, is incomplete, especially in the pediatric age group after cardiac surgery [5]. Metabolic alkalosis in adults after heart surgery with cardiopulmonary bypass used to be common in the 1960s mainly related to the use of citrate anticoagulated blood [11,12]. Patients, in whom CPB was used, more often developed MA compared to those in whom CPB was not used. Since then very few reports on this phenomenon have appeared.
MA, however, turns out to be multifactorial in cause and still occurs frequently. Acid–base management during hypothermia on CPB and during deep hypothermic cardiac arrest has been implicated as a factor in the development of acid–base disorders [13–15]. The optimal approach to pH-management during hypothermic CPB and circulatory arrest in newborns remains controversial. One study showed gradual development of MA with alpha-stat [14]. In another study, it was shown that the occurrence of pH>7.6 postoperatively in pH-stat and alpha-stat was not significantly different (16 and 17%, respectively) [15]. In our institution alpha-stat is used. It remains unclear to what extent this exerts an influence on the postoperative pH equilibrium.
Our study shows that postoperative MA was present in nearly 50% of the neonates and infants under 2 years of age within 48h after cardiac surgery. Furthermore, younger patient age, preoperative dependency on the arterial duct, the use of cardiopulmonary bypass were predictive of the development of MA after cardiac surgery. The duration of circulatory arrest was strongly correlated to the duration of CPB. The association we found between age and development of MA is in accordance with the study of Wong and Chundu [5]. The splanchnic circulation, albeit in adults, has been shown to produce an alkalinating effect during CPB [16]; studies in infants have not yet been carried out.
The tendency of virtually all patients to become more alkalotic implies that cardiac surgery or the management during such surgery increases total body bicarbonate, apparently related to age. Hemodilution may play a role, as it occurred in virtually all patients. The fluids used in these cases were minute amounts of sodium chloride for the insertion of an intravenous drip and larger amounts of fresh frozen plasma whenever necessary for hemodynamic stability.
Some bias may originate from the fact that patients who developed MA were shown to have more preoperative measurements done [49%] than those which did not develop MA (20%). Alternatively, one could speculate that younger, more severely ill children are more prone to develop MA.
Laudignon et al. studied risk factors for the occurrence of MA in newborns and infants [10] and concluded that renal immaturity is a risk factor related to MA. In newborns, tubular mechanisms may be more easily overwhelmed [17], and circulating aldosterone is abundant [18], which may explain why in younger children the kidney does not sufficiently excrete excess bicarbonate, thus leading to more frequent development of MA.
Volume depletion and chloride depletion are well-known effects of diuretics and causes of MA. Wong and Chundu concluded that chloride depletion is the predominant factor in the pathogenesis of MA [5]. In this regard we expected to find an independent relation between the use of diuretics and development of MA. However, this correlation was found only in univariate analysis. Because no routine measurements of chloride concentrations were available further statistical analysis of this variable was not possible.
In order to determine the influence of the nature of the preoperative cardiac diagnosis, we divided the patients according to major diagnosis groups. Although all children who underwent arterial switch for correction of transposition of the great arteries developed MA, this relation was less obvious for other diagnoses. As the mean age of patients in all but one group was lower in the cases of MA, though not consistently statistically significant (Table 5), this may to some extent explain the other relations found.
Although we were unable to relate MA to mortality, it might be implicated in circulatory disturbances as alkalosis influences ionic concentrations of various solutes, including calcium [19]. It is conceivable that our clinical protocol prevented worse outcome. As a high PCO2 in alkalosis appears not to be harmful [20], we mainly use respiratory compensation by means of mechanical ventilation. Additionally, in case of sodium, potassium, calcium or magnesium deficiency, the respective chloride salts are supplied. Finally, aldosterone antagonists and carbonic anhydrase inhibitors may be administered. We have never resorted to the use of ammonium chloride.
Based on the arguments above the reported two successful resuscitations of postoperative infants in the setting of a severe MA appear to be incidents [6]. Still, we consider it beneficial, on the basis of the predictive factors found, to be prepared for the development of MA, and act accordingly.
5. Conclusions
A younger age, preoperative dependency on the arterial duct and the use of cardiopulmonary bypass during surgery are factors that increase the chance of the occurrence of metabolic alkalosis in infants after heart surgery. Hemodilution adds to these risk factors. Children operated on for transposition of the great arteries have a very high chance of developing MA (in our study up to 100%). The acid–base status of children who are subjected to several of these factors should be closely monitored to prevent severe alkalosis.
Acknowledgments
The authors recognize the scientific and practical contributions of A. Struijs, MD, PhD.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
