- Email: [email protected]
Clinical Outcome Score Predicts Adverse Neurodevelopmental Outcome After Infant Heart Surgery
CONGENITAL HEART
Clinical Outcome Score Predicts Adverse Neurodevelopmental Outcome After Infant Heart Surgery Andrew S. Mackie, MD, SM, Shabnam Vatanpour, MS, Gwen Y. Alton, RN, MN, Irina A. Dinu, PhD, Lindsay Ryerson, MD, Diane M. Moddemann, MD, MEd, and Julie Thomas Petrie, PhD, for the Western Canadian Complex Pediatric Therapies Program Follow-Up Group Department of Pediatrics and School of Public Health, University of Alberta, and Stollery Children’s Hospital and Glenrose Rehabilitation Hospital, Edmonton, Alberta; Department of Pediatrics and Child Health, University of Manitoba, Winnipeg, Manitoba; and Department of Psychology, Children’s and Women’s Health Centre of British Columbia, Vancouver, British Columbia, Canada
Background. The purpose of this study was to determine whether a clinical outcome score derived from early postoperative events is associated with Bayley-III scores at 18 to 24 months among infants undergoing cardiopulmonary bypass surgery. Methods. Included were infants aged 6 weeks or less who underwent surgery between 2005 and 2009, all of whom were referred for neurodevelopmental evaluation at 18 to 24 months. We excluded children with chromosomal abnormalities. The prespecified clinical outcome score had a range of 0 to 7. Lower scores indicated a more rapid postoperative recovery. Patients requiring extracorporeal life support were assigned a score of 7. Results. One hundred and ninety-nine subjects were included. Surgical procedures were arterial switch (72), Norwood (60), repair of total anomalous pulmonary venous connection (29), and other (38). Nine subjects had postoperative extracorporeal life support. Mean clinical outcome score in the Norwood group was 4.0 ± 1.4 versus the
arterial switch group (2.6 ± 1.5, p < 0.001), total anomalous pulmonary venous connection group (2.8 ± 1.8, p < 0.01), and other group (4.0 ± 1.8, p [ not significant). Among children who had a clinical outcome score of 4 or greater, there was a decrease in Bayley-III cognitive score of 5.7 (95% confidence interval: 1.5 to 9.9, p [ 0.009), a decrease in language score of 10.0 (95% confidence interval: 4.9 to 15.1, p < 0.001), and a decrease in motor score of 9.7 (95% confidence interval: 4.8 to 14.5, p < 0.001). Time until lactate of 2.0 mmol/L or less and highest 24-hour inotrope score increased with increasing clinical outcome score (p < 0.0001). Conclusions. Clinical outcome scores of 4 or greater were associated with significantly lower Bayley-III scores at 18 to 24 months. This score may be valuable as an endpoint when evaluating novel potential therapies for this high-risk population.
I
unit (ICU) at no cost. However, the Bayley Scales of Infant Development-II has been replaced by the Bayley Scales of Infant and Toddler Development, third edition (BayleyIII) [4]. Therefore, the objective of the current study was to evaluate the correlation between the clinical outcome score, as measured in the early postoperative period, with the Bayley-III scores as measured at 18 to 24 months. A secondary objective was to determine the correlation between this clinical outcome score and time to normalization of postoperative lactate, postoperative inotrope score, postoperative ICU length of stay (LOS), and total hospital LOS.
nfants with complex congenital heart disease are at risk of neurodevelopmental delay [1, 2], although such outcomes are difficult to predict in early childhood. Many innovative therapies for infants undergoing cardiopulmonary bypass (CPB) surgery have been evaluated in recent years and will continue to evolve, yet selection of primary endpoints for clinical trials remains challenging. These issues speak to the potential value of surrogate outcomes that reliably predict neurodevelopmental sequelae and can be measured early in the postoperative course. A novel clinical outcome score has been developed that predicts adverse psychomotor development on the Bayley Scales of Infant Development-II [3]. This clinical outcome score is objective and measurable in the intensive care
(Ann Thorac Surg 2015;99:2124–33) Ó 2015 by The Society of Thoracic Surgeons
Patients and Methods Design
Accepted for publication Feb 10, 2015. Address correspondence to Dr Mackie, Division of Cardiology, Stollery Children’s Hospital, 4C2 Walter C. Mackenzie Center, 8440-112th St NW, Edmonton, Alberta T6G 2B7, Canada; e-mail: [email protected].
Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier
This study is part of a larger multiprovince inception cohort that was established to evaluate neurodevelopmental outcomes of infants aged 6 weeks or less who undergo CPB surgery in Western Canada [5]. 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2015.02.029
Ann Thorac Surg 2015;99:2124–33
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
Preoperative, intraoperative, and postoperative variables were prospectively collected, as previously described [6]. The follow-up study and database received Institutional Review Board approval, and written consent was provided by a parent or guardian for participation in the cohort.
Subjects
Early Childhood Assessments Children were evaluated at 18 to 24 months of age. Certified pediatric-experienced psychometricians administered the Bayley-III, yielding separate cognitive, language, and motor composite scores, each with a mean of 100 and standard deviation of 15. Maternal education was determined by years of schooling. CLINICAL OUTCOME SCORE. The clinical outcome score (Table 1) was derived using clinically meaningful variables that can be recorded objectively and early in the postoperative course (eg, time to first extubation) and that reflect hemodynamic status. Possible scores range from 0 (reflecting least morbidity) to 7 (reflecting use of postoperative mechanical circulatory support). The score was calculated by adding the subscore of each individual component (range, 0 to 2 each). Subjects requiring postoperative extracorporeal life support were assigned a score of 7, the highest possible score, to reflect the most severe morbidity. INOTROPE SCORE. The postoperative inotrope score was calculated as per Wernovsky and colleagues [7]. The highest 24-hour postoperative inotrope score for a given patient was recorded for the purpose of correlation with the clinical outcome score. LACTATE MEASUREMENTS. Serum lactate was measured postoperatively in the ICU at least every 4 hours during Table 1. Clinical Outcome Score Score Clinical Variable Time until first negative fluid balance, days Time until sternal closure, days Time until first extubation, days
0
1
2
1 1 4
2 2–4 5–8
3 5 9
Use of extracorporeal life support was assigned a score of 7.
the first postoperative day, as previously described [8]. The interval between the patient’s admission to ICU and first lactate measurement of 2.0 mmol/L or less was measured in hours.
Statistical Analysis Categoric data are presented as proportions. Continuous data are expressed as mean SD or median with interquartile range. One-way analysis of variance with Tamhane’s multiple comparisons and Fisher’s exact test were used to compare surgical groups; Bonferroni correction was applied. Descriptive variables for outcomes were analyzed with Student’s t test, c2 test, and Fisher’s exact test. Correlations between clinical outcome score and other outcome variables were examined using Spearman’s rank order correlation coefficient. Subjects with a clinical outcome score of 0 or 1 were grouped together owing to small sample size in the 0 group. Exploratory data analyses supported the use of multiple indicators for the clinical outcome score, with the lowest score being the reference category. Multiple linear regression analysis included clinical outcome score and any other potential predictor variables having p less than 0.10 on univariate linear regression analysis, and after checking for multicolinearity. Potential predictor variables considered were birth weight, gestational age at birth, age at surgery (in days), year of surgery, number of preoperative days of ventilation, duration of CPB, duration of deep hypothermic circulatory arrest, lowest flow rate on CPB, mother’s years of schooling, and antenatal diagnosis (yes/no). All analyses were two-sided. All p values less than 0.05 were considered statistically significant. Statistical analyses were performed using SPSS version 17 (SPSS, Chicago, IL), SAS version 9.3 (SAS Institute, Cary, NC), and R version 2.30.
Results Two hundred and seventeen infants met inclusion criteria. Of these, 18 were excluded owing to chromosomal abnormalities. The remaining 199 subjects were included in this analysis. Surgical procedures performed were arterial switch (72), Norwood stage 1 (60), repair of total anomalous pulmonary venous connection (29), and other (38). The other group consisted of tetralogy of Fallot with pulmonary atresia (11), truncus arteriosus (7), hypoplastic aortic arch (6), interrupted aortic arch (4), double outlet right ventricle (3), tetralogy of Fallot (3), pulmonary atresia with intact ventricular septum (2), double inlet left ventricle (1), and left atrial aneurysm (1). Preoperative, intraoperative, and postoperative variables are provided in Table 2. Prematurity (gestational age less than 37 weeks) was present in 15 of 199 subjects (8%), with 13 of 199 (7%) having a birth weight less than 2.5 kg. The distribution of clinical outcome scores among these 199 subjects is shown in Figure 1. Nine subjects had postoperative extracorporeal life support and therefore were assigned a score of 7. Mean clinical outcome score in the Norwood group was 4.0 1.4, compared with a mean of 2.6 1.5 (p < 0.001) in the arterial switch group and a
CONGENITAL HEART
Inclusion criteria were infants aged 6 weeks or less who underwent CPB surgery at the Stollery Children’s Hospital in Edmonton, Alberta, between 2005 and 2009. All survivors received multidisciplinary neurodevelopmental assessments including the Bayley-III in a developmental clinic, with the exception of 6 children who were lost to follow-up. The developmental clinics were located in Edmonton and Calgary, Alberta; Regina and Saskatoon, Saskatchewan; Vancouver, British Columbia; and Winnipeg, Manitoba. We excluded children with known chromosomal abnormalities (eg, deletion 22q11.2). Children who died at any age before Bayley-III testing were not excluded and contributed to results until their death.
2125
2126
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
Ann Thorac Surg 2015;99:2124–33
Table 2. Preoperative, Intraoperative, and Postoperative Variables by Type of Infant Cardiac Surgery
CONGENITAL HEART
Variable Preoperative Age at surgery, days Mean (SD) Median (IQR) Gestational age, weeks Mean (SD) Median (IQR) Birth weight, kg Mean (SD) Median (IQR) Duration preoperative ventilation, days Mean (SD) Median (IQR) Mother’s years of schooling Mean (SD) Median (IQR) Intraoperative Duration CPB, minutes Mean (SD) Median (IQR) Duration aortic cross-clamping, minutes Mean (SD) Median (IQR) Duration DHCA, minutes (n ¼ 148) Mean (SD) Median (IQR) Postoperative Time until lactate 2.0 mmol/L, hours Mean (SD) Median (IQR) Highest 24-hour inotrope score Mean (SD) Median (IQR) Time until first negative fluid balance, days Mean (SD) Median (IQR) Delayed sternal closure (n ¼ 124) Time until sternal closure, days Mean (SD) Median (IQR) Time until first extubation, days Mean (SD) Median (IQR) Total postoperative ventilation,a days Mean (SD) Median (IQR) Hospital LOS, days, median (IQR)b Clinical outcome score, mean (SD)
Norwood (n ¼ 60)
Arterial Switch (n ¼ 72)
TAPVC Repair (n ¼ 29)
Other (n ¼ 38)
12.5 (6.1) 11 (7)
10.0 (5.3) 9 (4)
13.2 (11.1) 10 (11)
15.8 (8.9) 14 (11)
38.8 (1.6) 39 (2)
39.1 (1.8) 40 (2)
39.0 (1.6) 39 (2)
38.4 (1.8) 39 (2)
3.26 (0.52) 3.19 (0.59)
3.42 (0.56) 3.47 (0.75)
3.45 (0.64) 3.62 (0.86)
3.09 (0.51) 3.05 (0.77)
8.5 (6.8) 8 (8)
6.2 (5.1) 5 (4)
2.9 (3.8) 1 (4)
7.3 (8.7) 5 (12)
13.7 (2.0) 13 (3)
13.7 (3.2) 14 (4)
12.3 (3.0) 12 (1)
13.5 (3.2) 13 (4)
109 (39) 92 (60)
123 (45)d 108 (57.5)
66 (3) 56 (25)
129 (83) 115.5 (53)
48 (17) 44 (24)
69 (22) 65 (19)
34 (10) 33 (12)
55 (33) 49 (28)
23 (10) 22 (16)
12 (12) 8 (6.5)
29 (10) 28.5 (15)
20 (11) 20.5 (15)
20.1 (18.5) 16 (12.5)
11.4 (8.4) 11.5 (14.75)
17.3 (16.1) 14 (8)
11.5 (8.2) 12.5 (9.5)
18.4 (8.8) 17 (10.75)
14.1 (7.6) 12.5 (7.5)
16.1 (7.3) 15.5 (7.5)
13.6 (5.8) 15 (6.5)
2.6 (1.0) 3 (1) 59
2.5 (0.9) 2.5 (1) 25
2.4 (0.7) 2 (1) 11
3.2 (1.4) 3 (2) 29
4.6 (4.5) 4 (2)
1.4 (2.4)c 0 (3)
1.8 (2.5)c 0 (4)
4.7 (4.7) 3 (4)
10.5 (10.6) 7.5 (4)
6.2 (5.2) 5 (4)
6.0 (3.7) 5 (5)
10.9 (10.6) 8 (7)
13.6 9 33 4.0
8.6 6 19 2.6
7.1 6 16 2.8
13.2 9 28 4.0
(14.1) (8) (23) (1.4)d
(12.7) (4.5) (12) (1.5)c
(3.8) (5) (10) (1.8)
(13.5) (8) (24) (1.8)e
Thirty-two patients required reintubation and therefore had a total duration of postoperative ventilation that exceeded time until first extubab c d tion. Includes preoperative and postoperative days. Denotes difference from Norwood (p < 0.001). Denotes difference from TAPVC (p < e 0.01). Denotes difference from arterial switch (p ¼ 0.001).
a
CPB ¼ cardiopulmonary bypass; DHCA ¼ deep hypothermic circulatory arrest; anomalous pulmonary venous connection.
IQR ¼ interquartile range;
LOS ¼ length of stay;
TAPVC ¼ total
Ann Thorac Surg 2015;99:2124–33
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
50 45
Number of Patients
40 35 30 25
20 15 10 5
0
1
2
3
4
5
6
7
Clinical Outcome Score
Fig 1. Distribution of clinical outcome scores among study subjects.
mean of 2.8 1.8 (p < 0.01) in the total anomalous pulmonary venous connection group. Scores in the other group were 4.0 1.8, also statistically significantly higher than the arterial switch group (p ¼ 0.001). Mean age at Bayley-III testing was 22.8 5.1 months. Mean cognitive Bayley-III score was 94 14, mean language score was 90 17, and mean motor score was 93 16. On multiple regression analysis, Bayley-III cognitive, language, and motor scores were all lower among patients with a clinical outcome score of 4 or greater. After adjusting for birth weight and mother’s education, a decrease in Bayley-III cognitive score of 5.7 (95% confidence interval [CI]: 1.5 to 9.9, p ¼ 0.009) was found among subjects with clinical outcome score of 4 or greater compared with subjects with a score less than 4. There was a statistically significant change of 4 in the percent of variation (R2) in Bayley-III cognitive score, explained by the clinical outcome score when added to a model consisting of birth weight and mother’s education (p ¼ 0.009). After adjusting for mother’s education, a decrease in Bayley-III language score of 10.0 (95% CI: 4.9 to 15.1, p < 0.001) was found among subjects with clinical outcome score of 4 or greater compared with subjects having a score less than 4. There was a statistically significant change of 7.8 in the percent of variation (R2) in Bayley-III language score explained by the clinical outcome score when added to a model consisting of mother’s education (p < 0.001). After adjusting for preoperative ventilation time and cross-clamp time, a decrease in Bayley-III motor score of 9.7 (95% CI: 4.8 to 14.5, p < 0.001) was found among subjects with clinical outcome score of 4 or more compared with subjects having a score less than 4. There was a statistically significant change of 8 in the percent of variation (R2) in Bayley-III motor score explained by the clinical outcome score when added to a model consisting of preoperative ventilation time and cross-clamp time (p < 0.001). The positive predictive values (PPV) of a clinical outcome score of 4 or more for Bayley-III score less than 85 were 26%, 48%, and 43% for cognitive, language, and motor scores, respectively. The negative predictive values of a clinical outcome score less than 4 for Bayley-III score
of 85 or greater were 89%, 74%, and 81%, respectively. The sensitivities of a clinical outcome score of 4 or greater for detection of a Bayley-III score less than 85 were 57%, 52%, and 57%, respectively, whereas the specificities were 67%, 71%, and 71%, respectively. Receiver-operating characteristic curves are illustrated in Figure 2. It was not possible to include ICU LOS or total hospital LOS in the multiple regression analysis, because these variables were highly correlated with the clinical outcome score. On univariate analysis, the clinical outcome score had a similar or greater effect on Bayley-III scores as did total hospital LOS, reflecting an advantage of the clinical outcome score over hospital LOS. The percentage of variance in cognitive score explained by total hospital LOS was 9.8%, compared with that explained by the clinical outcome score (4.9%). However, each additional 30 days in hospital was associated with a mean decrease in Bayley-III cognitive score of 6.3, similar to the mean decrease of 6.2 with a clinical outcome score of 4 or higher. The percentage of variance in language score explained by total hospital LOS was 4.5%, less than that explained by the clinical outcome score (8.4%). Each additional 30 days in hospital was associated with a mean decrease in Bayley-III language score of 5.4, compared with a mean decrease of 10.4 with a clinical outcome score of 4 or higher. The percentage of variance in motor score explained by total hospital LOS was 8.7%, compared with that explained by the clinical outcome score (9.3%). Each additional 30 days in hospital was associated with a mean decrease in Bayley-III motor of 7.2, compared with a mean decrease of 10.2 with clinical outcome score of 4 or higher. Twenty-eight subjects (14%) died during follow-up, 9 during the postoperative period and 19 after initial discharge. Among the 9 subjects requiring postoperative extracorporeal life support, 3 died in hospital and 2 died after initial postoperative discharge. The clinical outcome score was associated with death, either in hospital or after discharge (p < 0.0001); the risk of death was 4.85 times higher in the patient group with clinical outcome score of 4 or greater (95% CI: 2.05 to 11.49). Only 5 subjects underwent transplantation during follow-up (age at transplantation, 25 days to 13 months), all of whom had a clinical outcome score of 4 or more (p ¼ 0.03). The risk of death or transplant was 5.77 times higher in the patient group with clinical outcome score of 4 or more (95% CI: 2.48 to 13.44). Longer time to first lactate value of 2.0 mmol/L or less and highest postoperative inotrope score both had positive correlation with the clinical outcome score (both p < 0.0001; Figs 3 and 4, respectively). The percent of variance (R2) for the regression model containing only clinical outcome score on Bayley-III cognitive, language, and motor scores was 10%, 9.5%, and 14%, respectively; by comparison, the R2 for lactate less than 2.0 mmol/L was 0.95%, 0.2%, and 0.7%, respectively, and the R2 for inotrope score was 0.87%, 1.6%, and 0.7%, respectively. Combining both lactate and inotrope resulted in an R2 of 1.7%, 1.7%, and 2.8%, respectively. Postoperative length of ICU stay also had a positive correlation with clinical
CONGENITAL HEART
0
2127
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
Ann Thorac Surg 2015;99:2124–33
CONGENITAL HEART
2128
Fig 2. Receiver-operating characteristic curve for (A) cognitive score, (B) language score, and (C) motor score. For all domains, a clinical outcome score of 4 or more was associated with the best combination of sensitivity and specificity for identifying patients having a Bayley-III score of less than 85.
outcome score (p < 0.0001), as did total hospital LOS (p < 0.0001; Figs 5 and 6, respectively).
Comment We found that a clinical outcome score derived from the timing of postoperative events among infants aged 6 weeks or less undergoing CPB surgery was associated with Bayley-III scores at 18 to 24 months, time to normalization of serum lactate, highest 24-hour postoperative inotrope score, and with ICU and total hospital LOS. Clinical outcome scores were higher among subjects
undergoing Norwood stage 1 palliation compared with subjects undergoing the arterial switch procedure or repair of total anomalous pulmonary venous connection. The full range of potential scores was observed among the study population. Bayley-III scores across all areas of neurodevelopment were significantly lower among subjects having a clinical outcome score of 4 or greater, compared with a score of less than 4. A half standard deviation change in score is considered clinically significant [9]. Because one standard deviation is 15 points, a half standard deviation is 7.5, meaning that the mean drop in language and motor
Ann Thorac Surg 2015;99:2124–33
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
2129
Fig 3. Relationship between postoperative time to lactate 2.0 mmol/L or less and clinical outcome score (rho ¼ 0.36, p < 0.0001). Boxes represent 25th to 75th percentiles; the line within each box represents the median; whiskers represent 10th and 90th percentiles.
CONGENITAL HEART
scores (10 and 9.7, respectively) associated with a clinical outcome score of 4 or higher is substantial and can potentially change a child’s classification from a qualitative descriptor of “average” (eg, 95, at the 37th percentile), to “low average” (85, at the 16th percentile). While the disadvantages associated with a small (3- to 5-point) decrease may be controversial, in the case of the cognitive score decrease of 5.7, which constitutes a decrease of one third of a standard deviation, the mean score of 94 would decrease to 88, shifting from the 34th percentile (average range) to the 23rd percentile (low average range) and would potentially be associated with lower educational achievement, occupational achievement [10, 11], and social adjustment [12]. The relationship between clinical outcome score and Bayley-III scores was not linear, suggesting that there is a threshold effect of perioperative hemodynamic status on cerebral injury. It is not possible to conclude to what degree congenital brain abnormalities or preoperative insults contributed to lower Bayley-III scores, relative to a difficult postoperative clinical course reflected by a high clinical outcome score. The Bayley-III is an instrument that measures a wide range of skills and concepts and is designed to identify current status and delay. Although there are distinct differences between developmental milestones and later intelligence testing [13], there are moderate to high correlations between the Bayley-III cognitive and language scores and preschool intelligence scores [14], as well as preschool language skills, and moderate correlations between Bayley-III motor score and motor quotients up to 6 years of age [15]. Thus,
high clinical outcome scores may identify patients most at risk for neurodevelopmental outcome difficulties. The correlation between clinical outcome score and ICU LOS was anticipated, given that the events determining the clinical outcome score (Table 1) must occur in the ICU rather than on the ward. However, the clinical outcome score also correlated well with total hospital LOS, which reflects the sum of preoperative and both early and late postoperative morbidity. This finding is relevant as LOS correlates with neurodevelopmental outcomes among children after cardiopulmonary bypass surgery [16]. Newburger and colleagues [16] reported that among 8-year-old children who had undergone the arterial switch operation in infancy, both longer ICU LOS and longer hospital LOS were associated with lower fullscale IQ and verbal IQ even after adjustment for perioperative and sociodemographic variables. In the current study, we found that each additional 30 days spent in the hospital was associated with a mean decrease in cognitive, language, and motor scores of 6.3, 5.4, and 7.2, respectively. By comparison, a clinical outcome score of 4 or higher corresponded to a drop in cognitive score of 6.2, similar to that with an additional 30 days in hospital, but corresponded with a more dramatic drop in mean language and motor scores of 10.4 and 10.2, respectively. Clinical scoring systems have been previously evaluated in children undergoing CPB surgery. Zobel and colleagues [17] applied the Acute Physiologic Score for Children, Pediatric Risk of Mortality, Therapeutic Intervention Scoring System, and Organ System Failure score. All four scores differed significantly between survivors
2130
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
Ann Thorac Surg 2015;99:2124–33
CONGENITAL HEART
Fig 4. Relationship between highest postoperative inotrope score and clinical outcome score (rho ¼ 0.36, p < 0.0001). Boxes represent 25th to 75th percentiles; the line within each box represents the median; whiskers represent 10th and 90th percentiles.
and nonsurvivors [17]. The Risk Adjustment for Congenital Heart Surgery-1 category has also been validated as a risk factor for mortality in pediatric open-heart surgery [18], as has the Modified Sequential Organ Failure Assessment score [19]. The Aristotle Basic Complexity score predicts infants less than 6 months of age having longer ICU LOS [20]. However, we are not aware of scoring systems that predict neurodevelopmental outcomes in this population. The ideal surrogate outcome is inexpensive, easy to observe, can be measured independent of treatment assignment, and is free of measurement error. The clinical outcome score that we developed does not incorporate any laboratory investigations, and therefore, is not associated with any cost. The components of the score (eg, time to first extubation) apply to all infants undergoing CPB surgery. The score is easy to calculate, and can be done before a patient’s discharge from intensive care. Posting the derivation of the score (Table 1) in clinical settings (ie, ICU) would allow clinicians to quickly and accurately calculate the score for a given patient. This study has several limitations. Patients were operated on at a single surgical center, and the validity of our findings should be confirmed in a multicenter study. The
association with Bayley-III might be different at institutions with practice patterns, such as delayed sternal closure, that are different from our own. For example, decision for delayed sternal closure and timing of sternal closure varies among centers. Delayed sternal closure is routinely done in our institution after the Norwood procedure, but less frequently after other operations; that, in part, accounts for between-group differences in the clinical outcome score (Table 2). The “other” group (n ¼ 38) represents a heterogeneous group of lesions and surgical interventions. The low PPV of the clinical outcome score must be interpreted within the context that the PPV is influenced by the prevalence of the outcome in the at-risk population. Fortunately, the prevalence of Bayley-III scores less than 85 was low; hence, the PPV was low. The timing of postoperative lactate monitoring was not prespecified but, rather, was at the discretion of the responsible intensivist. It is possible that patients who were more hemodynamically stable may have had less frequent lactate monitoring, and therefore may have taken relatively longer to achieve a lactate of 2.0 mmol/L or less compared with patients who were deemed more hemodynamically unstable. Nonetheless, we demonstrated a strong correlation between clinical outcome
Ann Thorac Surg 2015;99:2124–33
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
2131
Fig 5. Relationship between postoperative intensive care unit (ICU) length of stay and clinical outcome score (rho ¼ 0.62, p < 0.0001). Boxes represent 25th to 75th percentiles; the line within each box represents the median; whiskers represent 10th and 90th percentiles.
CONGENITAL HEART
Fig 6. Relationship between total hospital length of stay and clinical outcome score (rho ¼ 0.51, p < 0.0001). Boxes represent 25th to 75th percentiles; the line within each box represents the median; whiskers represent 10th and 90th percentiles.
CONGENITAL HEART
2132
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
score and time to normalization of lactate. The long-term predictive ability of the Bayley-III scores measured at 18 to 24 months is unclear [21]. This is an observational study and, therefore, we can only establish an association between clinical score and Bayley-III scores, not a cause and effect relationship. The clinical outcome score may not distinguish moderate from severely impaired brain function. In the multiple regression model, it was not possible to adjust for ICU LOS or total hospital LOS as these variables are highly correlated with the clinical outcome score. We have not demonstrated that interventions aimed at lowering the clinical outcome score in any class of patients are associated with improved neurodevelopmental outcomes. In summary, the clinical outcome score is associated with adverse Bayley-III scores at 18 to 24 months of age and correlated strongly with other markers of postoperative events, including time to normalization of serum lactate and ICU and hospital LOS. This score is easy to calculate, has no cost, and can be used to identify children who may benefit from early and comprehensive neurodevelopmental intervention. The score may be valuable as an outcome variable in future studies aimed at evaluating novel therapies in this high-risk population. The members of the Western Canadian Complex Pediatric Therapies Program Follow-Up Group are Charlene M. T. Robertson, Ari R. Joffe, David B. Ross, and Ivan M. Rebeyka, Edmonton, Alberta; Reg Sauve, Calgary, Alberta; Patricia Blakley, Saskatoon, Saskatchewan; Anne Synnes, Vancouver, British Columbia; and Jaya Bodani, Regina, Saskatchewan, Canada.
References 1. Bellinger DC, Wypij D, duPlessis AJ, et al. Neurodevelopmental status at eight years in children with dextrotransposition of the great arteries: the Boston Circulatory Arrest Trial. J Thorac Cardiovasc Surg 2003;126:1385–96. 2. Shillingford AJ, Glanzman MM, Ittenbach RF, Clancy RR, Gaynor JW, Wernovsky G. Inattention, hyperactivity, and school performance in a population of school-age children with complex congenital heart disease. Pediatrics 2008;121:e759–67. 3. Mackie AS, Alton GY, Dinu IA, et al. Clinical outcome score predicts the need for neurodevelopmental intervention after infant heart surgery. J Thorac Cardiovasc Surg 2013;145: 1248–54.e2. 4. Bayley N. Bayley scales of infant and toddler development. 3rd ed. San Antonio, TX: Harcourt Assessment; 2006.
Ann Thorac Surg 2015;99:2124–33
5. Robertson CMT, Joffe AR, Sauve RS, et al. Outcomes from an interprovincial program of newborn open heart surgery. J Pediatr 2004;144:86–92. 6. Robertson CMT, Sauve RS, Joffe AR, et al. The registry and follow-up of complex pediatric therapies program of western Canada: a mechanism for service, audit, and research after life-saving therapies for young children. Cardiol Res Pract 2011;2011:965740. 7. Wernovsky G, Wypij D, Jonas RA, et al. Postoperative course and hemodynamic profile after the arterial switch operation in neonates and infants. A comparison of low-flow cardiopulmonary bypass and circulatory arrest. Circulation 1995;92:2226–35. 8. Cheung PY, Chui N, Joffe AR, Rebeyka IM, Robertson CM. Postoperative lactate concentrations predict the outcome of infants aged 6 weeks or less after intracardiac surgery: a cohort follow-up to 18 months. J Thorac Cardiovasc Surg 2005;130:837–43. 9. Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health-related quality of life: the remarkable universality of half a standard deviation. Med Care 2003;41: 582–92. 10. Hunter JE. Cognitive ability, cognitive aptitudes, job knowledge, and job performance. J Vocat Behav 1986;29:340–62. 11. McCall RB. Childhood IQ’s as predictors of adult educational and occupational status. Science 1977;197:482–3. 12. Moffitt TE, Gabrielli WF, Mednick SA, Schulsinger F. Socioeconomic status, IQ, and delinquency. J Abnorm Psychol 1981;90:152–6. 13. Fagan J. Intelligence in infancy. In: Sternberg RJ, Kaufman SB (eds). New York: Cambridge Press; 2011:130–43. 14. Wechsler D. Wechsler preschool and primary scale of intelligence-fourth edition (WPPSI-IV). San Antonio, TX: Pearson; 2012. 15. Bayley N (ed). Bayley scales of infant and toddler development-third edition. Technical manual. San Antonio, TX: Harcourt Assessment; 2006. 16. Newburger JW, Wypij D, Bellinger DC, et al. Length of stay after infant heart surgery is related to cognitive outcome at age 8 years. J Pediatr 2003;143:67–73. 17. Zobel G, Rodl S, Rigler B, et al. Prospective evaluation of clinical scoring systems in infants and children with cardiopulmonary insufficiency after cardiac surgery. J Cardiovasc Surg 1993;34:333–7. 18. Kang N, Cole T, Tsang V, Elliott M, De Leval M. Risk stratification in paediatric open-heart surgery. Eur J Cardiothorac Surg 2004;26:3–11. 19. Shime N, Kageyama K, Ashida H, Tanaka Y. Application of modified sequential organ failure assessment score in children after cardiac surgery. J Cardiothorac Vasc Anesth 2001;15:463–8. 20. Gillespie M, Kuijpers M, Van Rossem M, et al. Determinants of intensive care unit length of stay for infants undergoing cardiac surgery. Congenit Heart Dis 2006;1:152–60. 21. Aylward GP. Continuing issues with the Bayley-III: where to go from here. J Develop Behav Pediatr 2013;34:697–701.
INVITED COMMENTARY Neurodevelopmental outcome is one of the most important factors that parents of children born with congenital heart defects are concerned about once the initial anxieties about the diagnosis and operation have settled. Long-term studies have shown that developmental and functional abnormalities across multiple domains are present at school entry in these children. Among the many predictors of later developmental
Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier
difficulties (microcephaly, genetic predisposition, age at operation, palliation, circulatory arrest time, abnormal postoperative neurologic features, and maternal education), very few are modifiable. Recent advances in the understanding of the fetal mechanisms of impaired cerebral blood flow in the context of congenital heart disease suggest that immediate postoperative events may be only an indicator and may not explain causality
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2015.02.072
Clinical Outcome Score Predicts Adverse Neurodevelopmental Outcome After Infant Heart Surgery Andrew S. Mackie, MD, SM, Shabnam Vatanpour, MS, Gwen Y. Alton, RN, MN, Irina A. Dinu, PhD, Lindsay Ryerson, MD, Diane M. Moddemann, MD, MEd, and Julie Thomas Petrie, PhD, for the Western Canadian Complex Pediatric Therapies Program Follow-Up Group Department of Pediatrics and School of Public Health, University of Alberta, and Stollery Children’s Hospital and Glenrose Rehabilitation Hospital, Edmonton, Alberta; Department of Pediatrics and Child Health, University of Manitoba, Winnipeg, Manitoba; and Department of Psychology, Children’s and Women’s Health Centre of British Columbia, Vancouver, British Columbia, Canada
Background. The purpose of this study was to determine whether a clinical outcome score derived from early postoperative events is associated with Bayley-III scores at 18 to 24 months among infants undergoing cardiopulmonary bypass surgery. Methods. Included were infants aged 6 weeks or less who underwent surgery between 2005 and 2009, all of whom were referred for neurodevelopmental evaluation at 18 to 24 months. We excluded children with chromosomal abnormalities. The prespecified clinical outcome score had a range of 0 to 7. Lower scores indicated a more rapid postoperative recovery. Patients requiring extracorporeal life support were assigned a score of 7. Results. One hundred and ninety-nine subjects were included. Surgical procedures were arterial switch (72), Norwood (60), repair of total anomalous pulmonary venous connection (29), and other (38). Nine subjects had postoperative extracorporeal life support. Mean clinical outcome score in the Norwood group was 4.0 ± 1.4 versus the
arterial switch group (2.6 ± 1.5, p < 0.001), total anomalous pulmonary venous connection group (2.8 ± 1.8, p < 0.01), and other group (4.0 ± 1.8, p [ not significant). Among children who had a clinical outcome score of 4 or greater, there was a decrease in Bayley-III cognitive score of 5.7 (95% confidence interval: 1.5 to 9.9, p [ 0.009), a decrease in language score of 10.0 (95% confidence interval: 4.9 to 15.1, p < 0.001), and a decrease in motor score of 9.7 (95% confidence interval: 4.8 to 14.5, p < 0.001). Time until lactate of 2.0 mmol/L or less and highest 24-hour inotrope score increased with increasing clinical outcome score (p < 0.0001). Conclusions. Clinical outcome scores of 4 or greater were associated with significantly lower Bayley-III scores at 18 to 24 months. This score may be valuable as an endpoint when evaluating novel potential therapies for this high-risk population.
I
unit (ICU) at no cost. However, the Bayley Scales of Infant Development-II has been replaced by the Bayley Scales of Infant and Toddler Development, third edition (BayleyIII) [4]. Therefore, the objective of the current study was to evaluate the correlation between the clinical outcome score, as measured in the early postoperative period, with the Bayley-III scores as measured at 18 to 24 months. A secondary objective was to determine the correlation between this clinical outcome score and time to normalization of postoperative lactate, postoperative inotrope score, postoperative ICU length of stay (LOS), and total hospital LOS.
nfants with complex congenital heart disease are at risk of neurodevelopmental delay [1, 2], although such outcomes are difficult to predict in early childhood. Many innovative therapies for infants undergoing cardiopulmonary bypass (CPB) surgery have been evaluated in recent years and will continue to evolve, yet selection of primary endpoints for clinical trials remains challenging. These issues speak to the potential value of surrogate outcomes that reliably predict neurodevelopmental sequelae and can be measured early in the postoperative course. A novel clinical outcome score has been developed that predicts adverse psychomotor development on the Bayley Scales of Infant Development-II [3]. This clinical outcome score is objective and measurable in the intensive care
(Ann Thorac Surg 2015;99:2124–33) Ó 2015 by The Society of Thoracic Surgeons
Patients and Methods Design
Accepted for publication Feb 10, 2015. Address correspondence to Dr Mackie, Division of Cardiology, Stollery Children’s Hospital, 4C2 Walter C. Mackenzie Center, 8440-112th St NW, Edmonton, Alberta T6G 2B7, Canada; e-mail: [email protected].
Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier
This study is part of a larger multiprovince inception cohort that was established to evaluate neurodevelopmental outcomes of infants aged 6 weeks or less who undergo CPB surgery in Western Canada [5]. 0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2015.02.029
Ann Thorac Surg 2015;99:2124–33
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
Preoperative, intraoperative, and postoperative variables were prospectively collected, as previously described [6]. The follow-up study and database received Institutional Review Board approval, and written consent was provided by a parent or guardian for participation in the cohort.
Subjects
Early Childhood Assessments Children were evaluated at 18 to 24 months of age. Certified pediatric-experienced psychometricians administered the Bayley-III, yielding separate cognitive, language, and motor composite scores, each with a mean of 100 and standard deviation of 15. Maternal education was determined by years of schooling. CLINICAL OUTCOME SCORE. The clinical outcome score (Table 1) was derived using clinically meaningful variables that can be recorded objectively and early in the postoperative course (eg, time to first extubation) and that reflect hemodynamic status. Possible scores range from 0 (reflecting least morbidity) to 7 (reflecting use of postoperative mechanical circulatory support). The score was calculated by adding the subscore of each individual component (range, 0 to 2 each). Subjects requiring postoperative extracorporeal life support were assigned a score of 7, the highest possible score, to reflect the most severe morbidity. INOTROPE SCORE. The postoperative inotrope score was calculated as per Wernovsky and colleagues [7]. The highest 24-hour postoperative inotrope score for a given patient was recorded for the purpose of correlation with the clinical outcome score. LACTATE MEASUREMENTS. Serum lactate was measured postoperatively in the ICU at least every 4 hours during Table 1. Clinical Outcome Score Score Clinical Variable Time until first negative fluid balance, days Time until sternal closure, days Time until first extubation, days
0
1
2
1 1 4
2 2–4 5–8
3 5 9
Use of extracorporeal life support was assigned a score of 7.
the first postoperative day, as previously described [8]. The interval between the patient’s admission to ICU and first lactate measurement of 2.0 mmol/L or less was measured in hours.
Statistical Analysis Categoric data are presented as proportions. Continuous data are expressed as mean SD or median with interquartile range. One-way analysis of variance with Tamhane’s multiple comparisons and Fisher’s exact test were used to compare surgical groups; Bonferroni correction was applied. Descriptive variables for outcomes were analyzed with Student’s t test, c2 test, and Fisher’s exact test. Correlations between clinical outcome score and other outcome variables were examined using Spearman’s rank order correlation coefficient. Subjects with a clinical outcome score of 0 or 1 were grouped together owing to small sample size in the 0 group. Exploratory data analyses supported the use of multiple indicators for the clinical outcome score, with the lowest score being the reference category. Multiple linear regression analysis included clinical outcome score and any other potential predictor variables having p less than 0.10 on univariate linear regression analysis, and after checking for multicolinearity. Potential predictor variables considered were birth weight, gestational age at birth, age at surgery (in days), year of surgery, number of preoperative days of ventilation, duration of CPB, duration of deep hypothermic circulatory arrest, lowest flow rate on CPB, mother’s years of schooling, and antenatal diagnosis (yes/no). All analyses were two-sided. All p values less than 0.05 were considered statistically significant. Statistical analyses were performed using SPSS version 17 (SPSS, Chicago, IL), SAS version 9.3 (SAS Institute, Cary, NC), and R version 2.30.
Results Two hundred and seventeen infants met inclusion criteria. Of these, 18 were excluded owing to chromosomal abnormalities. The remaining 199 subjects were included in this analysis. Surgical procedures performed were arterial switch (72), Norwood stage 1 (60), repair of total anomalous pulmonary venous connection (29), and other (38). The other group consisted of tetralogy of Fallot with pulmonary atresia (11), truncus arteriosus (7), hypoplastic aortic arch (6), interrupted aortic arch (4), double outlet right ventricle (3), tetralogy of Fallot (3), pulmonary atresia with intact ventricular septum (2), double inlet left ventricle (1), and left atrial aneurysm (1). Preoperative, intraoperative, and postoperative variables are provided in Table 2. Prematurity (gestational age less than 37 weeks) was present in 15 of 199 subjects (8%), with 13 of 199 (7%) having a birth weight less than 2.5 kg. The distribution of clinical outcome scores among these 199 subjects is shown in Figure 1. Nine subjects had postoperative extracorporeal life support and therefore were assigned a score of 7. Mean clinical outcome score in the Norwood group was 4.0 1.4, compared with a mean of 2.6 1.5 (p < 0.001) in the arterial switch group and a
CONGENITAL HEART
Inclusion criteria were infants aged 6 weeks or less who underwent CPB surgery at the Stollery Children’s Hospital in Edmonton, Alberta, between 2005 and 2009. All survivors received multidisciplinary neurodevelopmental assessments including the Bayley-III in a developmental clinic, with the exception of 6 children who were lost to follow-up. The developmental clinics were located in Edmonton and Calgary, Alberta; Regina and Saskatoon, Saskatchewan; Vancouver, British Columbia; and Winnipeg, Manitoba. We excluded children with known chromosomal abnormalities (eg, deletion 22q11.2). Children who died at any age before Bayley-III testing were not excluded and contributed to results until their death.
2125
2126
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
Ann Thorac Surg 2015;99:2124–33
Table 2. Preoperative, Intraoperative, and Postoperative Variables by Type of Infant Cardiac Surgery
CONGENITAL HEART
Variable Preoperative Age at surgery, days Mean (SD) Median (IQR) Gestational age, weeks Mean (SD) Median (IQR) Birth weight, kg Mean (SD) Median (IQR) Duration preoperative ventilation, days Mean (SD) Median (IQR) Mother’s years of schooling Mean (SD) Median (IQR) Intraoperative Duration CPB, minutes Mean (SD) Median (IQR) Duration aortic cross-clamping, minutes Mean (SD) Median (IQR) Duration DHCA, minutes (n ¼ 148) Mean (SD) Median (IQR) Postoperative Time until lactate 2.0 mmol/L, hours Mean (SD) Median (IQR) Highest 24-hour inotrope score Mean (SD) Median (IQR) Time until first negative fluid balance, days Mean (SD) Median (IQR) Delayed sternal closure (n ¼ 124) Time until sternal closure, days Mean (SD) Median (IQR) Time until first extubation, days Mean (SD) Median (IQR) Total postoperative ventilation,a days Mean (SD) Median (IQR) Hospital LOS, days, median (IQR)b Clinical outcome score, mean (SD)
Norwood (n ¼ 60)
Arterial Switch (n ¼ 72)
TAPVC Repair (n ¼ 29)
Other (n ¼ 38)
12.5 (6.1) 11 (7)
10.0 (5.3) 9 (4)
13.2 (11.1) 10 (11)
15.8 (8.9) 14 (11)
38.8 (1.6) 39 (2)
39.1 (1.8) 40 (2)
39.0 (1.6) 39 (2)
38.4 (1.8) 39 (2)
3.26 (0.52) 3.19 (0.59)
3.42 (0.56) 3.47 (0.75)
3.45 (0.64) 3.62 (0.86)
3.09 (0.51) 3.05 (0.77)
8.5 (6.8) 8 (8)
6.2 (5.1) 5 (4)
2.9 (3.8) 1 (4)
7.3 (8.7) 5 (12)
13.7 (2.0) 13 (3)
13.7 (3.2) 14 (4)
12.3 (3.0) 12 (1)
13.5 (3.2) 13 (4)
109 (39) 92 (60)
123 (45)d 108 (57.5)
66 (3) 56 (25)
129 (83) 115.5 (53)
48 (17) 44 (24)
69 (22) 65 (19)
34 (10) 33 (12)
55 (33) 49 (28)
23 (10) 22 (16)
12 (12) 8 (6.5)
29 (10) 28.5 (15)
20 (11) 20.5 (15)
20.1 (18.5) 16 (12.5)
11.4 (8.4) 11.5 (14.75)
17.3 (16.1) 14 (8)
11.5 (8.2) 12.5 (9.5)
18.4 (8.8) 17 (10.75)
14.1 (7.6) 12.5 (7.5)
16.1 (7.3) 15.5 (7.5)
13.6 (5.8) 15 (6.5)
2.6 (1.0) 3 (1) 59
2.5 (0.9) 2.5 (1) 25
2.4 (0.7) 2 (1) 11
3.2 (1.4) 3 (2) 29
4.6 (4.5) 4 (2)
1.4 (2.4)c 0 (3)
1.8 (2.5)c 0 (4)
4.7 (4.7) 3 (4)
10.5 (10.6) 7.5 (4)
6.2 (5.2) 5 (4)
6.0 (3.7) 5 (5)
10.9 (10.6) 8 (7)
13.6 9 33 4.0
8.6 6 19 2.6
7.1 6 16 2.8
13.2 9 28 4.0
(14.1) (8) (23) (1.4)d
(12.7) (4.5) (12) (1.5)c
(3.8) (5) (10) (1.8)
(13.5) (8) (24) (1.8)e
Thirty-two patients required reintubation and therefore had a total duration of postoperative ventilation that exceeded time until first extubab c d tion. Includes preoperative and postoperative days. Denotes difference from Norwood (p < 0.001). Denotes difference from TAPVC (p < e 0.01). Denotes difference from arterial switch (p ¼ 0.001).
a
CPB ¼ cardiopulmonary bypass; DHCA ¼ deep hypothermic circulatory arrest; anomalous pulmonary venous connection.
IQR ¼ interquartile range;
LOS ¼ length of stay;
TAPVC ¼ total
Ann Thorac Surg 2015;99:2124–33
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
50 45
Number of Patients
40 35 30 25
20 15 10 5
0
1
2
3
4
5
6
7
Clinical Outcome Score
Fig 1. Distribution of clinical outcome scores among study subjects.
mean of 2.8 1.8 (p < 0.01) in the total anomalous pulmonary venous connection group. Scores in the other group were 4.0 1.8, also statistically significantly higher than the arterial switch group (p ¼ 0.001). Mean age at Bayley-III testing was 22.8 5.1 months. Mean cognitive Bayley-III score was 94 14, mean language score was 90 17, and mean motor score was 93 16. On multiple regression analysis, Bayley-III cognitive, language, and motor scores were all lower among patients with a clinical outcome score of 4 or greater. After adjusting for birth weight and mother’s education, a decrease in Bayley-III cognitive score of 5.7 (95% confidence interval [CI]: 1.5 to 9.9, p ¼ 0.009) was found among subjects with clinical outcome score of 4 or greater compared with subjects with a score less than 4. There was a statistically significant change of 4 in the percent of variation (R2) in Bayley-III cognitive score, explained by the clinical outcome score when added to a model consisting of birth weight and mother’s education (p ¼ 0.009). After adjusting for mother’s education, a decrease in Bayley-III language score of 10.0 (95% CI: 4.9 to 15.1, p < 0.001) was found among subjects with clinical outcome score of 4 or greater compared with subjects having a score less than 4. There was a statistically significant change of 7.8 in the percent of variation (R2) in Bayley-III language score explained by the clinical outcome score when added to a model consisting of mother’s education (p < 0.001). After adjusting for preoperative ventilation time and cross-clamp time, a decrease in Bayley-III motor score of 9.7 (95% CI: 4.8 to 14.5, p < 0.001) was found among subjects with clinical outcome score of 4 or more compared with subjects having a score less than 4. There was a statistically significant change of 8 in the percent of variation (R2) in Bayley-III motor score explained by the clinical outcome score when added to a model consisting of preoperative ventilation time and cross-clamp time (p < 0.001). The positive predictive values (PPV) of a clinical outcome score of 4 or more for Bayley-III score less than 85 were 26%, 48%, and 43% for cognitive, language, and motor scores, respectively. The negative predictive values of a clinical outcome score less than 4 for Bayley-III score
of 85 or greater were 89%, 74%, and 81%, respectively. The sensitivities of a clinical outcome score of 4 or greater for detection of a Bayley-III score less than 85 were 57%, 52%, and 57%, respectively, whereas the specificities were 67%, 71%, and 71%, respectively. Receiver-operating characteristic curves are illustrated in Figure 2. It was not possible to include ICU LOS or total hospital LOS in the multiple regression analysis, because these variables were highly correlated with the clinical outcome score. On univariate analysis, the clinical outcome score had a similar or greater effect on Bayley-III scores as did total hospital LOS, reflecting an advantage of the clinical outcome score over hospital LOS. The percentage of variance in cognitive score explained by total hospital LOS was 9.8%, compared with that explained by the clinical outcome score (4.9%). However, each additional 30 days in hospital was associated with a mean decrease in Bayley-III cognitive score of 6.3, similar to the mean decrease of 6.2 with a clinical outcome score of 4 or higher. The percentage of variance in language score explained by total hospital LOS was 4.5%, less than that explained by the clinical outcome score (8.4%). Each additional 30 days in hospital was associated with a mean decrease in Bayley-III language score of 5.4, compared with a mean decrease of 10.4 with a clinical outcome score of 4 or higher. The percentage of variance in motor score explained by total hospital LOS was 8.7%, compared with that explained by the clinical outcome score (9.3%). Each additional 30 days in hospital was associated with a mean decrease in Bayley-III motor of 7.2, compared with a mean decrease of 10.2 with clinical outcome score of 4 or higher. Twenty-eight subjects (14%) died during follow-up, 9 during the postoperative period and 19 after initial discharge. Among the 9 subjects requiring postoperative extracorporeal life support, 3 died in hospital and 2 died after initial postoperative discharge. The clinical outcome score was associated with death, either in hospital or after discharge (p < 0.0001); the risk of death was 4.85 times higher in the patient group with clinical outcome score of 4 or greater (95% CI: 2.05 to 11.49). Only 5 subjects underwent transplantation during follow-up (age at transplantation, 25 days to 13 months), all of whom had a clinical outcome score of 4 or more (p ¼ 0.03). The risk of death or transplant was 5.77 times higher in the patient group with clinical outcome score of 4 or more (95% CI: 2.48 to 13.44). Longer time to first lactate value of 2.0 mmol/L or less and highest postoperative inotrope score both had positive correlation with the clinical outcome score (both p < 0.0001; Figs 3 and 4, respectively). The percent of variance (R2) for the regression model containing only clinical outcome score on Bayley-III cognitive, language, and motor scores was 10%, 9.5%, and 14%, respectively; by comparison, the R2 for lactate less than 2.0 mmol/L was 0.95%, 0.2%, and 0.7%, respectively, and the R2 for inotrope score was 0.87%, 1.6%, and 0.7%, respectively. Combining both lactate and inotrope resulted in an R2 of 1.7%, 1.7%, and 2.8%, respectively. Postoperative length of ICU stay also had a positive correlation with clinical
CONGENITAL HEART
0
2127
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
Ann Thorac Surg 2015;99:2124–33
CONGENITAL HEART
2128
Fig 2. Receiver-operating characteristic curve for (A) cognitive score, (B) language score, and (C) motor score. For all domains, a clinical outcome score of 4 or more was associated with the best combination of sensitivity and specificity for identifying patients having a Bayley-III score of less than 85.
outcome score (p < 0.0001), as did total hospital LOS (p < 0.0001; Figs 5 and 6, respectively).
Comment We found that a clinical outcome score derived from the timing of postoperative events among infants aged 6 weeks or less undergoing CPB surgery was associated with Bayley-III scores at 18 to 24 months, time to normalization of serum lactate, highest 24-hour postoperative inotrope score, and with ICU and total hospital LOS. Clinical outcome scores were higher among subjects
undergoing Norwood stage 1 palliation compared with subjects undergoing the arterial switch procedure or repair of total anomalous pulmonary venous connection. The full range of potential scores was observed among the study population. Bayley-III scores across all areas of neurodevelopment were significantly lower among subjects having a clinical outcome score of 4 or greater, compared with a score of less than 4. A half standard deviation change in score is considered clinically significant [9]. Because one standard deviation is 15 points, a half standard deviation is 7.5, meaning that the mean drop in language and motor
Ann Thorac Surg 2015;99:2124–33
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
2129
Fig 3. Relationship between postoperative time to lactate 2.0 mmol/L or less and clinical outcome score (rho ¼ 0.36, p < 0.0001). Boxes represent 25th to 75th percentiles; the line within each box represents the median; whiskers represent 10th and 90th percentiles.
CONGENITAL HEART
scores (10 and 9.7, respectively) associated with a clinical outcome score of 4 or higher is substantial and can potentially change a child’s classification from a qualitative descriptor of “average” (eg, 95, at the 37th percentile), to “low average” (85, at the 16th percentile). While the disadvantages associated with a small (3- to 5-point) decrease may be controversial, in the case of the cognitive score decrease of 5.7, which constitutes a decrease of one third of a standard deviation, the mean score of 94 would decrease to 88, shifting from the 34th percentile (average range) to the 23rd percentile (low average range) and would potentially be associated with lower educational achievement, occupational achievement [10, 11], and social adjustment [12]. The relationship between clinical outcome score and Bayley-III scores was not linear, suggesting that there is a threshold effect of perioperative hemodynamic status on cerebral injury. It is not possible to conclude to what degree congenital brain abnormalities or preoperative insults contributed to lower Bayley-III scores, relative to a difficult postoperative clinical course reflected by a high clinical outcome score. The Bayley-III is an instrument that measures a wide range of skills and concepts and is designed to identify current status and delay. Although there are distinct differences between developmental milestones and later intelligence testing [13], there are moderate to high correlations between the Bayley-III cognitive and language scores and preschool intelligence scores [14], as well as preschool language skills, and moderate correlations between Bayley-III motor score and motor quotients up to 6 years of age [15]. Thus,
high clinical outcome scores may identify patients most at risk for neurodevelopmental outcome difficulties. The correlation between clinical outcome score and ICU LOS was anticipated, given that the events determining the clinical outcome score (Table 1) must occur in the ICU rather than on the ward. However, the clinical outcome score also correlated well with total hospital LOS, which reflects the sum of preoperative and both early and late postoperative morbidity. This finding is relevant as LOS correlates with neurodevelopmental outcomes among children after cardiopulmonary bypass surgery [16]. Newburger and colleagues [16] reported that among 8-year-old children who had undergone the arterial switch operation in infancy, both longer ICU LOS and longer hospital LOS were associated with lower fullscale IQ and verbal IQ even after adjustment for perioperative and sociodemographic variables. In the current study, we found that each additional 30 days spent in the hospital was associated with a mean decrease in cognitive, language, and motor scores of 6.3, 5.4, and 7.2, respectively. By comparison, a clinical outcome score of 4 or higher corresponded to a drop in cognitive score of 6.2, similar to that with an additional 30 days in hospital, but corresponded with a more dramatic drop in mean language and motor scores of 10.4 and 10.2, respectively. Clinical scoring systems have been previously evaluated in children undergoing CPB surgery. Zobel and colleagues [17] applied the Acute Physiologic Score for Children, Pediatric Risk of Mortality, Therapeutic Intervention Scoring System, and Organ System Failure score. All four scores differed significantly between survivors
2130
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
Ann Thorac Surg 2015;99:2124–33
CONGENITAL HEART
Fig 4. Relationship between highest postoperative inotrope score and clinical outcome score (rho ¼ 0.36, p < 0.0001). Boxes represent 25th to 75th percentiles; the line within each box represents the median; whiskers represent 10th and 90th percentiles.
and nonsurvivors [17]. The Risk Adjustment for Congenital Heart Surgery-1 category has also been validated as a risk factor for mortality in pediatric open-heart surgery [18], as has the Modified Sequential Organ Failure Assessment score [19]. The Aristotle Basic Complexity score predicts infants less than 6 months of age having longer ICU LOS [20]. However, we are not aware of scoring systems that predict neurodevelopmental outcomes in this population. The ideal surrogate outcome is inexpensive, easy to observe, can be measured independent of treatment assignment, and is free of measurement error. The clinical outcome score that we developed does not incorporate any laboratory investigations, and therefore, is not associated with any cost. The components of the score (eg, time to first extubation) apply to all infants undergoing CPB surgery. The score is easy to calculate, and can be done before a patient’s discharge from intensive care. Posting the derivation of the score (Table 1) in clinical settings (ie, ICU) would allow clinicians to quickly and accurately calculate the score for a given patient. This study has several limitations. Patients were operated on at a single surgical center, and the validity of our findings should be confirmed in a multicenter study. The
association with Bayley-III might be different at institutions with practice patterns, such as delayed sternal closure, that are different from our own. For example, decision for delayed sternal closure and timing of sternal closure varies among centers. Delayed sternal closure is routinely done in our institution after the Norwood procedure, but less frequently after other operations; that, in part, accounts for between-group differences in the clinical outcome score (Table 2). The “other” group (n ¼ 38) represents a heterogeneous group of lesions and surgical interventions. The low PPV of the clinical outcome score must be interpreted within the context that the PPV is influenced by the prevalence of the outcome in the at-risk population. Fortunately, the prevalence of Bayley-III scores less than 85 was low; hence, the PPV was low. The timing of postoperative lactate monitoring was not prespecified but, rather, was at the discretion of the responsible intensivist. It is possible that patients who were more hemodynamically stable may have had less frequent lactate monitoring, and therefore may have taken relatively longer to achieve a lactate of 2.0 mmol/L or less compared with patients who were deemed more hemodynamically unstable. Nonetheless, we demonstrated a strong correlation between clinical outcome
Ann Thorac Surg 2015;99:2124–33
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
2131
Fig 5. Relationship between postoperative intensive care unit (ICU) length of stay and clinical outcome score (rho ¼ 0.62, p < 0.0001). Boxes represent 25th to 75th percentiles; the line within each box represents the median; whiskers represent 10th and 90th percentiles.
CONGENITAL HEART
Fig 6. Relationship between total hospital length of stay and clinical outcome score (rho ¼ 0.51, p < 0.0001). Boxes represent 25th to 75th percentiles; the line within each box represents the median; whiskers represent 10th and 90th percentiles.
CONGENITAL HEART
2132
MACKIE ET AL NEURODEVELOPMENTAL OUTCOME POST INFANT CHD SURGERY
score and time to normalization of lactate. The long-term predictive ability of the Bayley-III scores measured at 18 to 24 months is unclear [21]. This is an observational study and, therefore, we can only establish an association between clinical score and Bayley-III scores, not a cause and effect relationship. The clinical outcome score may not distinguish moderate from severely impaired brain function. In the multiple regression model, it was not possible to adjust for ICU LOS or total hospital LOS as these variables are highly correlated with the clinical outcome score. We have not demonstrated that interventions aimed at lowering the clinical outcome score in any class of patients are associated with improved neurodevelopmental outcomes. In summary, the clinical outcome score is associated with adverse Bayley-III scores at 18 to 24 months of age and correlated strongly with other markers of postoperative events, including time to normalization of serum lactate and ICU and hospital LOS. This score is easy to calculate, has no cost, and can be used to identify children who may benefit from early and comprehensive neurodevelopmental intervention. The score may be valuable as an outcome variable in future studies aimed at evaluating novel therapies in this high-risk population. The members of the Western Canadian Complex Pediatric Therapies Program Follow-Up Group are Charlene M. T. Robertson, Ari R. Joffe, David B. Ross, and Ivan M. Rebeyka, Edmonton, Alberta; Reg Sauve, Calgary, Alberta; Patricia Blakley, Saskatoon, Saskatchewan; Anne Synnes, Vancouver, British Columbia; and Jaya Bodani, Regina, Saskatchewan, Canada.
References 1. Bellinger DC, Wypij D, duPlessis AJ, et al. Neurodevelopmental status at eight years in children with dextrotransposition of the great arteries: the Boston Circulatory Arrest Trial. J Thorac Cardiovasc Surg 2003;126:1385–96. 2. Shillingford AJ, Glanzman MM, Ittenbach RF, Clancy RR, Gaynor JW, Wernovsky G. Inattention, hyperactivity, and school performance in a population of school-age children with complex congenital heart disease. Pediatrics 2008;121:e759–67. 3. Mackie AS, Alton GY, Dinu IA, et al. Clinical outcome score predicts the need for neurodevelopmental intervention after infant heart surgery. J Thorac Cardiovasc Surg 2013;145: 1248–54.e2. 4. Bayley N. Bayley scales of infant and toddler development. 3rd ed. San Antonio, TX: Harcourt Assessment; 2006.
Ann Thorac Surg 2015;99:2124–33
5. Robertson CMT, Joffe AR, Sauve RS, et al. Outcomes from an interprovincial program of newborn open heart surgery. J Pediatr 2004;144:86–92. 6. Robertson CMT, Sauve RS, Joffe AR, et al. The registry and follow-up of complex pediatric therapies program of western Canada: a mechanism for service, audit, and research after life-saving therapies for young children. Cardiol Res Pract 2011;2011:965740. 7. Wernovsky G, Wypij D, Jonas RA, et al. Postoperative course and hemodynamic profile after the arterial switch operation in neonates and infants. A comparison of low-flow cardiopulmonary bypass and circulatory arrest. Circulation 1995;92:2226–35. 8. Cheung PY, Chui N, Joffe AR, Rebeyka IM, Robertson CM. Postoperative lactate concentrations predict the outcome of infants aged 6 weeks or less after intracardiac surgery: a cohort follow-up to 18 months. J Thorac Cardiovasc Surg 2005;130:837–43. 9. Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health-related quality of life: the remarkable universality of half a standard deviation. Med Care 2003;41: 582–92. 10. Hunter JE. Cognitive ability, cognitive aptitudes, job knowledge, and job performance. J Vocat Behav 1986;29:340–62. 11. McCall RB. Childhood IQ’s as predictors of adult educational and occupational status. Science 1977;197:482–3. 12. Moffitt TE, Gabrielli WF, Mednick SA, Schulsinger F. Socioeconomic status, IQ, and delinquency. J Abnorm Psychol 1981;90:152–6. 13. Fagan J. Intelligence in infancy. In: Sternberg RJ, Kaufman SB (eds). New York: Cambridge Press; 2011:130–43. 14. Wechsler D. Wechsler preschool and primary scale of intelligence-fourth edition (WPPSI-IV). San Antonio, TX: Pearson; 2012. 15. Bayley N (ed). Bayley scales of infant and toddler development-third edition. Technical manual. San Antonio, TX: Harcourt Assessment; 2006. 16. Newburger JW, Wypij D, Bellinger DC, et al. Length of stay after infant heart surgery is related to cognitive outcome at age 8 years. J Pediatr 2003;143:67–73. 17. Zobel G, Rodl S, Rigler B, et al. Prospective evaluation of clinical scoring systems in infants and children with cardiopulmonary insufficiency after cardiac surgery. J Cardiovasc Surg 1993;34:333–7. 18. Kang N, Cole T, Tsang V, Elliott M, De Leval M. Risk stratification in paediatric open-heart surgery. Eur J Cardiothorac Surg 2004;26:3–11. 19. Shime N, Kageyama K, Ashida H, Tanaka Y. Application of modified sequential organ failure assessment score in children after cardiac surgery. J Cardiothorac Vasc Anesth 2001;15:463–8. 20. Gillespie M, Kuijpers M, Van Rossem M, et al. Determinants of intensive care unit length of stay for infants undergoing cardiac surgery. Congenit Heart Dis 2006;1:152–60. 21. Aylward GP. Continuing issues with the Bayley-III: where to go from here. J Develop Behav Pediatr 2013;34:697–701.
INVITED COMMENTARY Neurodevelopmental outcome is one of the most important factors that parents of children born with congenital heart defects are concerned about once the initial anxieties about the diagnosis and operation have settled. Long-term studies have shown that developmental and functional abnormalities across multiple domains are present at school entry in these children. Among the many predictors of later developmental
Ó 2015 by The Society of Thoracic Surgeons Published by Elsevier
difficulties (microcephaly, genetic predisposition, age at operation, palliation, circulatory arrest time, abnormal postoperative neurologic features, and maternal education), very few are modifiable. Recent advances in the understanding of the fetal mechanisms of impaired cerebral blood flow in the context of congenital heart disease suggest that immediate postoperative events may be only an indicator and may not explain causality
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2015.02.072