Objective. To evaluate the incremental risk of congenital heart disease on mortality after noncardiac surgery in children.
Design. We reviewed the clinical information network database of the University Hospital Consortium for the period January 1, 1993, through December 31, 1996, and identified 191 261 patients <18 years old having 1 or more noncardiovascular surgical procedures. Of these patients, 6.5% had a diagnosis of congenital heart disease. Patients having ambulatory surgery are excluded from this database. Short-term (1-, 2-, and 3-day) and 30-day mortality were compared, as well as mortality for neonates, infants (31 days to 1 year), and older children (1–17 years); mortality for the 100 most common surgical procedures, mortality for 10 relatively minor surgical procedures, and mortality in subgroups of patients with minor versus severe cardiac diagnoses.
Results. Short-term and 30-day mortality was increased in the patients with congenital heart disease patients (30-day mortality odds ratio 3.5; 95% confidence limit, 3.2–3.9). Mortality was also increased in patients with congenital heart disease in the 2 youngest age groups, for the 100 most common operations, and for 10 relatively minor operations. Children with more severe heart disease diagnoses had higher mortality than did children carrying less serious cardiac diagnoses.
Conclusion. A diagnosis of congenital heart disease adds significant incremental risk of mortality in children requiring inpatient noncardiovascular surgery. This outcome difference is present for both minor and major surgical procedures, and regardless of whether mortality is measured at 1, 3, or 30 days. The incremental risk is greatest in neonates and infants where the presence of congenital heart disease is associated with a 2-fold increase in mortality from noncardiac surgery.
Although the presence of ischemic coronary heart disease increases the risk of noncardiac surgery in adults,1–3these data are not completely relevant to children with structural congenital heart disease (CHD) because of differences in anatomy and physiology. There is little information regarding the incremental risk of co-existent CHD in infants and children undergoing noncardiac surgery.4 This is particularly relevant because of the frequent association of cardiac disease with other congenital anomalies requiring surgical therapy.
To determine the effect of CHD on mortality in young patients having noncardiac surgery, we used the clinical information network database of the University Hospital Consortium (UHC), a group of >60 university hospitals in the United States that shares diagnostic, demographic, procedural, and outcome data on all hospital admissions.
We queried the UHC database for patients <18 years old who had any of 3136 International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) procedure codes during the period January 1, 1993, through December 31, 1996. Patients were excluded if the surgical procedure was related to repair or palliation of a cardiovascular anomaly, or if the sole procedure codes for that patient were related to a diagnostic procedure rather than a surgical procedure. In addition, patients undergoing circumcision as the only surgical procedure were also excluded. In general, surgical procedures were included if it was felt that they would require a significant degree of sedation or anesthesia in a pediatric population. Of the 3136 codes, 2221 included patients fulfilling the admission criteria for the study and were included for further analysis.
Patients were considered to have CHD if their discharge diagnoses included any of 171 codes referable to congenital cardiovascular conditions (ICD-9-CM codes 35.00 through 38.04). Data from the UHC database for our hospital were compared with internally generated data to confirm validity of the UHC database information.
In addition to the common definition of operative mortality as any death within 30 days of the surgical procedure, we also determined mortality at 1, 2, and 3 days after surgery. To determine the effect of age on outcome, we compared mortality in patients with and without CHD undergoing surgery at <31 days old, 31 days to 1 year old, and >1 year old. Mortality was also determined for the 100 most commonly performed surgical procedures in the database, and for a group of 10 commonly performed, relatively minor surgical procedures. Data were corrected to account for possible multiple procedures done on any one patient.
Because the severity and complexity of a congenital cardiac defect could also significantly affect surgical risk, we compared the 30-day survival of a subset of children with relatively minor cardiac diagnoses (patent ductus arteriosus, ventricular septal defect, or atrial septal defect) with a subset of children with major cardiac anomalies (truncus arteriosus, D-transposition of the great arteries, double outlet right ventricle, tetralogy of Fallot, and endocardial cushion defect).
Statistical analysis was by χ2 with the Yates correction for continuity as needed. Statistical significance was assumed for P < .05. Odds ratios (ORs) were calculated and are given with 95% confidence limits.
CHD was listed in the diagnostic codes of 3.1% of the 191 261 patients <18 years old satisfying these criteria of a noncardiovascular surgical procedure within the UHC during the 4-year study period. Overall 30-day mortality was 4.0%. Mortality was 3.8% in the 185 285 patients without CHD and 6.0% in the 5976 patients with heart disease (P < .001; OR, 3.53 [3.15–3.95]) (Table 1).
Early postoperative mortality was likewise elevated in patients with CHD. Overall 1-day mortality was 1.6%. Mortality in the patients with normal hearts was 1.6%, and was 2.0% in the group with CHD (P < .001; OR, 2.84 [2.36–3.43]). Mortality on the second postoperative day was 0.38% in patients without CHD and 0.49% in patients with CHD. Mortality on the third postoperative day was 0.21% and 0.35%. Irrespective of the presence of CHD, approximately one third of all mortality occurred within 1 day of surgery (35% and 29%). An additional 7% to 9% on the second day, and 5% to 6% on the third day.
Among patients both with and without CHD operative mortality was affected by age. In infants <31 days old, 30-day mortality was 6.6% with no heart disease versus 13.3% for those infants with heart disease (OR, 2.18 [1.89–2.50]). Thirty-day mortality for those 31 days to 1 year was 1.4% versus 3.1% (OR, 2.23 [1.6–3.12]) and for those >1 year old 1.2% versus 1.5% (OR, 1.19 [0.86–1.64]). Thus, after 1 year of age, there is no incremental risk of CHD.
Among patients who had 1 or more of the 100 most common procedures, 30-day mortality was 7.4% in children with CHD (293 deaths of 3947) compared with 2.3% for patients without heart disease (2841 of 127 652) (P < .001). Among the 10 selected minor surgical procedures, 30-day mortality was 1.5% in patients with CHD (11 of 719) and only 0.29% in patients without heart disease (75 of 25 527) (P < .001; Table 2). Although overall mortality figures may be instructive, it should be noted that for any single surgical procedure, the mortality is so low that the mortality incidence may be heavily influenced by small numbers of deaths.
The nature of the underlying cardiac anomaly clearly had an effect on surgical mortality. For patients with major cardiac abnormalities, the 30-day mortality was 11.3% (63 of 556), compared with 5.9% (121 of 2043) (P < .001) in patients with less severe anomalies.
By examining a large multiinstitutional database, we have shown that children with CHD undergoing inpatient, noncardiac surgery have significantly greater mortality than children without CHD. This outcome difference is present for a large number of both minor and major surgical procedures, for children <1 year old, and regardless of whether mortality is measured at 1, 3, or 30 days.
The information within the UHC database, and similar discharge databases, does not allow determination of the actual cause of death. It is possible that the presence of CHD was simply a surrogate for multiple comorbidities and that the increased mortality, especially at 30 days, may have been because of factors other than cardiac disease. However, even 1-day mortality was higher among patients with heart disease, suggesting that deaths were related to cardiopulmonary instability rather than chronic problems such as infection, renal failure, or malnutrition.
Intuitively, the presence of CHD might be expected to increase the risk of anesthesia and surgery in children. Although there are published guidelines for the treatment of children with CHD during noncardiac surgery,5,,6 none specifically addresses the magnitude of incremental risk conferred by the presence of heart disease. Since the landmark review of anesthetic mortality by Beecher and Todd in 1954,7 there have been other reviews of perioperative mortality. However, few studies have focused primarily on children, particularly after the major advances in perioperative care of the past 2 decades.8–10 Two large clinical series reporting pediatric postoperative mortality8,,11 had only a single death between them. These studies included anesthetics for diagnostic as well as for surgical procedures, anesthetics for ambulatory surgery, were limited to the immediate perioperative period, and did not address children with CHD. The only publication dealing specifically with operative morbidity and mortality of children with CHD having noncardiac surgery was fairly small, and evaluated only intraoperative complications.4 The major finding of that study was that there is a significant incidence of nonfatal intraoperative airway complications in these children. The closed claims project of the American Society of Anesthesiologists has reviewed closed malpractice claims and contrasted the morbid events of pediatric and adult cases,10 and this has given rise to the Pediatric Perioperative Cardiac Arrest Registry, a multicenter prospective data gathering group. This multiinstitutional registry is currently collecting information regarding pediatric perioperative cardiac arrests to define the etiologies and outcomes. However, this registry is also limited to the immediate perioperative period.
There is information concerning mortality of children with CHD having cardiac surgery. For example, mortality among 7169 infants and children undergoing cardiac surgery in New York State from 1992 to 1995 was 6.75%, ranging from 46.15% in procedures for repair of single ventricle to <1% for resection of subvalvular aortic stenosis or closure of atrial septal defect.12 It is noteworthy that the overall mortality for children having cardiac surgery in New York State was similar to the mortality rate for noncardiac surgery in patients with CHD in the UHC database (6.0%).
The mortality rates for children with and without CHD in this study exceeded that of previous reports.8,,11 There are several reasons for what seems like apparent excessive mortality reported in our study: 1) our study includes only hospitalized patients, ranging in age from premature infants to adolescents. The majority of pediatric surgical procedures in the United States are currently performed in an outpatient setting. Procedures performed in outpatient facilities or on an ambulatory, day-surgery basis in a hospital are not included in this database. Thus, this very large cohort of healthy children with low perioperative risk is excluded. 2) It is likely the database contains patients whose surgery had been planned or performed as an ambulatory procedure but subsequently required hospital admission because of complications or perceived risk. Thus, the present data are undoubtedly skewed toward patients with more serious conditions. 3) The discharge data did not allow determination of the cause of death, and death may have been unrelated to surgery or to CHD. 4) It may be that in a certain number of children the presence of CHD serves as a marker for a multiorgan genetic or dysmorphic syndrome.
When compared for disease-specific mortality, our data are in keeping with previously reported information. For example, for repair of diaphragmatic hernias in infants <1 year old operated on between 1989 and 1997, Weber et al13 reported overall survival of 79%, similar to reports of others.14 Survival in similar patients in our series, operated on more recently, was 83% (CHD) and 93% (no CHD).
In summary, we have identified the presence of CHD as a major risk factor in pediatric patients undergoing inpatient noncardiac surgery. The magnitude of this problem is evident by the presence of CHD in >3% of children undergoing surgery at UHC, and by the incremental risk encountered by these children. This incremental risk persisted in children <1 year old, and was present for both major and minor surgical procedures as well as commonly performed operations. The additional risk should be considered before performance of surgical procedures in children with CHD, and must be considered when evaluating surgical outcomes for both individual patients and for populations. Additional studies are necessary to determine the precise cause of death in these patients and to develop strategies to minimize the increased risks imposed by the presence of CHD.
- Received June 8, 1999.
- Accepted August 6, 1999.
Reprint requests to (V.C.B.) Department of Anesthesiology, University of Virginia Medical Center, Box 800710, Charlottesville, VA 22908-0710. E-mail:
- CHD =
- congenital heart disease •
- UHC =
- University Hospital Consortium •
- ICD-9-CM =
- International Classification of Diseases, Ninth Revision, Clinical Modification •
- OR =
- odds ratio
- Eagle KA,
- Rihal CS,
- Mickel MC,
- Holmes DR,
- Foster ED,
- Gersh BJ
- Fleisher LA
- ↵Teske DW, Cooney DR. Noncardiac surgery in patients with heart disease. In: Emmanouilides GC, Riemenschneider TA, Allen HA, Gutgesell HP, eds. Moss and Adams Heart Disease in Infants, Children, and Adolescents, Including the Fetus and Young Adult. 5th ed. Baltimore, MD: Williams & Wilkins; 1995:517–521
- ↵Kovalchin JP, Cooper JR, Vargo TA. Noncardiac surgery in patients with congenital heart disease. In: Garson A, Bricker JT, Fisher DJ, Neish SR, eds. The Science and Practice of Pediatric Cardiology. 2nd ed. Baltimore, MD: Williams & Wilkins; 1998:2273–2280
- Tiret L,
- Nivoche Y,
- Hatton F,
- Desmonts JM,
- Vourc'h G
- Hannan EL,
- Racz M,
- Kavey RE,
- Quagebeur JM,
- Williams R
- Copyright © 2000 American Academy of Pediatrics