Objectives. To evaluate changes in the clinical management of infants with hypoplastic left heart syndrome (HLHS) over a 10-year period.
Background. Orthotopic heart transplantation (OHT) and the Norwood procedure have emerged as the treatment options for HLHS over the last 2 decades.
Methods. We used 1988–1997 hospital discharge data from the National Inpatient Sample dataset. Patients ≤30 days of age with a principal diagnosis of HLHS were identified. Clinical management included the Norwood procedure, OHT, in-hospital death without surgery, discharge home without surgery, and transfer to another hospital. Multivariate logistic regression was used to evaluate variables associated with the choice of management.
Results. There were 1986 cases of HLHS with 812 in-hospital deaths, yielding a mortality rate 40.9%. The in-hospital mortality rate decreased from 54.4% in 1988 to 38.1% in 1997. The proportion of patients treated with the Norwood procedure increased from 8% in 1988 to 34% in 1997. The proportion of patients who died in the hospital without surgery decreased over time while the percentage discharged from the hospital without surgery or transferred to another hospital remained relatively unchanged. The in-hospital mortality rate was significantly lower in the OHT group compared with the Norwood group (26.2% vs 46.0%). We found no differences in gender, race, type of insurance, or home income between patients treated with the Norwood procedure compared with those who received comfort care. Patients from a later era, in the South, and in teaching hospitals were more likely to undergo the Norwood procedure.
Conclusions. Between 1988 and 1997, the proportion of infants with HLHS treated with the Norwood procedure increased while the use of comfort care decreased. Gender, race/ethnicity, type of medical insurance, and home income did not correlate with treatment choices.
Hypoplastic left heart syndrome (HLHS) is one of the most severe forms of congenital heart defect and carries a poor prognosis. The incidence of HLHS is 1 in 4000 to 6000 live births.1 Without surgical treatment, the mortality rate of infants with HLHS in the first year of life exceeds 90%.1 Until the early 1980s, there was no surgical treatment for HLHS, and families of infants with HLHS were offered comfort care as the only treatment option. In 1980, Norwood et al2 reported the first successful surgical palliation of HLHS. The so-called “Norwood procedure” was an innovative, experimental procedure in the 1980s but was gradually adopted as the treatment of choice by a limited number of centers. In the mid-1980s, Bailey et al3 proposed orthotopic heart transplant (OHT) as an alternative to the Norwood procedure for treating HLHS. In the past decade, the Norwood procedure, OHT, and comfort care have become the 3 treatment options that most pediatric cardiologists and neonatologists offer to families of infants with HLHS in the United States.4–6 However, in many European countries and Japan, comfort care is used frequently and may be the exclusive treatment option for these infants.7–9
Advances in diagnostic techniques, preoperative management, surgical techniques, and postoperative care at increasingly smaller cardiac centers have decreased surgical mortality for the Norwood procedure, making it the primary treatment for HLHS by some centers. At the same time, donor shortages have limited the availability of OHT as a treatment for HLHS. The debate continues as to whether the Norwood procedure or OHT provides better survival and long-term outcomes for children with HLHS.10 Limited information is available regarding historical trends and current practice patterns among pediatric cardiac centers treating infants with HLHS.
Gutgesell and Massaro11 used a database of 636 neonates with HLHS admitted to 40 university hospitals from 1989 to 1993 to evaluate management strategies and outcomes. These authors reported that 35% of neonates with HLHS were treated with the Norwood procedure, 6% were treated with OHT, and 59% had no surgical intervention. This study included only patients from a consortium of university hospitals, which consisted of almost exclusively tertiary care, teaching hospitals. Thus, aggressive surgical strategies may have been overrepresented in this study. In addition, the authors did not seek to explore the factors associated with the choice of treatment strategy.
The goal of this study was to investigate historical and current practice paradigms for the clinical management of infants with HLHS. Specifically, the objectives were: 1) to evaluate the historical evolution of clinical management and outcomes of infants with HLHS; and 2) to assess the effects of patient and hospital characteristics on the choice of clinical management.
This study used abstracted hospital discharge data from a national representative data sample, the Nationwide Inpatient Sample (NIS). The NIS is a subsample of data collected for the Health care Cost and Utilization Project, a multistate database project maintained by the Agency for Healthcare Research and Quality. The NIS contains data on ∼7 million hospital discharges per year from over 900 hospitals located in 22 states. It is designed to represent 20% of all discharges from nonfederal, general, and specialty hospitals, including children’s hospitals, in the United States.12 The NIS data include >100 clinical and nonclinical variables for each hospital stay, including primary and secondary diagnoses, primary and secondary procedures, admission and discharge status, patient demographics (eg, gender, age, race, median income for zip code), expected payment source, total charges, and length of stay. Hospital characteristics, such as location (census region, urban, or rural), size (number of beds), and teaching status are also specified in the NIS data. Our study used 10 years of data (1988–1997) from NIS Release 1, 2, 3, 4, 5, and 6.
The NIS data contain International Classification of Diseases, Ninth Revision, Clinical Modification (ICD9-CM) discharge diagnosis and procedure codes assigned by hospitals in participating states to each individual discharge. For this study, we selected a study population of infants with HLHS using age ≤30 days old at hospital admission and a diagnosis code of 746.7, indicating HLHS. Although many infants with HLHS are diagnosed prenatally or soon after birth, some infants may present with cyanosis, heart failure, or shock in the first few weeks of life. Therefore, we used 30 days as the upper age limit to include the majority of patients with HLHS. Each patient in the NIS database is assigned a unique sequence number allowing identification of patients with multiple admissions. Patients who were treated with balloon valvuloplasty without surgery were excluded because these patients probably represent a group with aortic stenosis who were miscoded as HLHS. Furthermore, we examined all diagnosis codes of each selected patient for anatomic information including transposition or malposition of great arteries, anomalous pulmonary venous drainage, and atrioventricular canal defect.
Because there is no ICD9-CM code for the Norwood procedure, we identified infants undergoing the Norwood procedure using a procedure code indicating cardiopulmonary bypass (39.61) together with 1 or more of the following codes for cardiac surgery: surgical creation of septal defect (34.42), repair of heart or pericardium (37.4), incision, excision, or occlusion of aorta (38.14), and systemic to pulmonary shunt (39.0). ICD9-CM procedure code 37.5 was used to identify patients treated with OHT. For patients not treated with the Norwood procedure or OHT, we examined the “discharge status” to determine whether the patient died in the hospital, was discharged from the hospital alive, or was transferred to another hospital. Each patient was entered in the analysis only once. All patients were assigned to 1 of the following 5 groups based on the patient’s treatment and/or discharge status: Norwood procedure, OHT, die in hospital without surgery, discharge alive without surgery, and transfer to another hospital.
For patients with multiple hospitalizations, which could be identified by the unique sequence number assigned to each patient, we examined all hospitalizations of the patient to determine whether a Norwood procedure or OHT was performed at any time during any of the hospitalizations. If neither of these procedures was performed, the patient’s discharge status (alive or in-hospital death) was identified at the last hospital discharge. If a patient’s discharge status was “transfer to another hospital” and the patient did not have subsequent hospital discharge from another hospital included in the NIS data, we assumed this patient was transferred to a hospital not included in the NIS data and assigned the patient to the “transfer to another hospital” group. The assignment of a patient who was transferred to another hospital in the NIS data or had multiple hospitalizations depended on the outcome of the final hospitalization.
Descriptive continuous data, such as hospital charges and length of stay, are presented as mean ± SD and median when appropriate. Continuous variables were compared using a Student t test. Categorical variables were compared using a χ2 or Fisher exact test. A P value <.05 was considered statistically significant. Statistical analyses were performed using SAS System for Windows, Release 6.12 (SAS Institute, Cary, NC).
To further compare changes in practice patterns over the study period, we divided the 10-year period into 2 eras: Era 1, 1988–1992; and Era 2, 1993–1997. In addition, data were analyzed on a yearly basis to determine trends in management paradigms from year to year. Linear regression models were used to characterize these changes in management over the study period.
We used weighted linear regression to determine the relation between case volume of the hospital at which the Norwood procedure was performed and in-hospital mortality rate. In this analysis, case volume of the hospital was the independent variable and in-hospital mortality rate for the Norwood procedure was the dependent variable. Case volume of the hospital was used as the weighting factor for each regression point.
Multivariate logistic regression was conducted to determine factors associated with treatment choice. The Norwood procedure or comfort care were the dependent variables in these regression models. Because of the small number of patients treated by OHT, OHT was not analyzed separately in this regression analysis. Independent variables included age, gender, race and ethnicity, type of insurance, median income of home zip code, region (West, South, Midwest, and Northeast), hospital teaching status, and hospital size by the number of beds.
There were 1986 patients ≤30 days old with HLHS admitted to 326 hospitals in the NIS database over the 10-year study period. The number of patients for each year, the number of hospitals with HLHS patients, and the states included in the NIS data are listed in Table 1.
There was a male predominance with a male to female ratio of 3:2. Table 2 lists demographic and hospital stay information of the study population. Over the study period, there were 812 in-hospital deaths, yielding an overall in-hospital mortality rate 40.9%. The mortality rate decreased from 54.4% in 1988 to 38.1% in 1997, or from 43.8% in Era 1 to 38.5 in Era 2 (P = .02). Of patients in this study, 22 (1.1%) also had anomalous pulmonary venous return, 45 (2.3%) had transposition or malposition of great arteries, and 24 (1.2%) had atrioventricular canal defect.
As demonstrated in Fig 1, 407 patients (20.5%) underwent surgery, including 346 patients who were treated with the Norwood procedure (17.4%) and 61 treated with OHT (3.1%). Among the 1579 patients (79.5%) who did not undergo surgery, 637 died in hospital (32.1%), 422 were discharged from the hospital alive (21.2%), and 520 were transferred to other hospitals (26.2%).
Fig 2 shows the major trends of HLHS management over the 10-year study period. An obvious trend was seen in the increasing percentage of patients treated with the Norwood procedure, from <8% in 1988 to 34% in 1997. Linear regression of the percentage of patients treated with the Norwood procedure versus time yields an r = 0.93 (P < .01). If this trend persists, by the year 2004 over 50% of infants with HLHS will be treated with the Norwood procedure. In addition, the proportion of patients who died in the hospital without surgery decreased consistently over time, from close to 50% in 1988 to 20% to 25% in 1995 to 1997 (r = −0.90; P < .01).
As demonstrated in Fig 3, the proportion of patients who were transferred to other hospitals remained relatively constant at ∼20% to 30% throughout the 10-year study period. The percentage of patients who were treated with OHT increased from 1% to 5% in Era 1 and remained at ∼5% range throughout Era 2. The proportion of patients who were discharged without surgery (comfort care group) fluctuated between 15% and 25% throughout the 10 years. No statistically significant trends were found in the rate of OHT, comfort care, or the percentage of patients transferred to other hospitals over the study period.
The overall in-hospital mortality rate for patients who underwent the Norwood procedure was 46.0%. The in-hospital mortality rate for patients undergoing the Norwood procedure demonstrated a trend toward lower mortality, from 64% in 1988 to close to 30% in 1996–1997 (Fig 4). The mortality for the Norwood procedure decreased significantly from 58% in Era 1 to 43% in Era 2 (P < .01). Using weighted linear regression, we found a negative correlation between the case volume of hospitals and in-hospital mortality rates (r = −0.20 for Era 1 and r = −0.31 for Era 2, P < .01, respectively).
Among the 637 patients who did not undergo surgical intervention and died in-hospital, the median survival was 3 days. Ninety patients (14.1%) died in the first day of hospitalization, 379 (59.4%) died in the first 3 days of hospitalization, and 505 (79.2%) died in the first week of hospitalization. The longest survival in the hospital was 139 days.
As listed in Table 3, the median length of hospital stay for patients undergoing the Norwood procedure was 18.5 days, and the total hospital charge was $105 724. In contrast, patients undergoing OHT had approximately a 3 times longer length of stay (55 days; P < .01) and hospital charges ($305 057; P < .01). The overall in-hospital mortality rate was significantly lower in the OHT group than the Norwood procedure group (26.2% vs 46.0%; P < .01). Patients who underwent surgery (Norwood operation or OHT) had significantly longer hospitalization and higher hospital charges than patients who were discharged from the hospital without surgery or died in the hospital.
Among 326 hospitals with HLHS patients, the case volume of the hospitals ranged from 1 to 78 with a median of 2. There were 47 hospitals performing the Norwood procedure: 25 hospitals (13.0% of all hospitals) in Era 1 and 33 hospitals (11.8% of all hospitals) in Era 2. Among these hospitals, 13 hospitals also performed OHT. There were 2 hospitals that performed more OHT cases than Norwood procedures. All but 6 hospitals performing the Norwood procedure had some patients discharged from the hospital without surgery. Among patients admitted to teaching hospitals, 22.7% underwent the Norwood procedure. In contrast, 13.0% patients underwent the Norwood procedure in nonteaching hospitals (P < .01).
In a multiple regression analysis, we found that patients admitted to teaching hospitals (odds ratio [OR]: 2.0; 95% confidence interval [CI]: 1.5–2.5), hospitals in the South (OR: 1.6; CI: 1.1–2.4), and patients admitted in Era 2 (OR: 1.3; CI: 1.1–1.6) were more likely to undergo the Norwood procedure. We found no statistical difference in age, gender, race/ethnicity, home income, and type of insurance in determining the management of HLHS.
When we examined patients who were discharged from the hospital without surgery to determine the factors associated with the choice of comfort care, we found no statistically significant differences in the choice of comfort care by age, gender, race/ethnicity, insurance, home income, region, hospital size, or hospital teaching status.
In the present study, we have delineated changes in the management of neonates and infants with HLHS over 10 years. Although comfort care was the most common management strategy for patients with HLHS, over time, more patients in an increasing number of hospitals were palliated with the Norwood procedure. Although OHT has been advocated by several investigators as a treatment strategy that provides a better survival and possibly superior quality of life, no apparent increase in its use to treat patients with HLHS was seen. In exploring the factors associated with the treatment choice, we found patients at teaching hospitals, at hospitals in the South and from Era 2 were more likely to be treated with the Norwood procedure. Patient’s gender, race, insurance, and home income were not important determinants of the type of treatment.
In comparing our study to the study by Gutgesell and Massaro,11 which found 41% of patients with HLHS underwent surgery, we found only 8.7% patients underwent surgery in Era 1 (similar to the period of 1989–1993 used in Gutgesell’s study). This difference is primarily because of the different data sources and types of institutions represented in the 2 studies. Gutgesell’s study used data from a consortium of 40 university hospitals which were mostly teaching hospitals and tertiary referral centers. Therefore, more aggressive treatment strategies were included in this study. This interpretation is supported by our results showing that patients in teaching hospitals are more likely to undergo the Norwood procedure. Importantly, our study includes data derived from a much broader and more representative range of hospitals. Thus, our study captured a sample which may be more representative of the overall treatment alternatives used in US hospitals in general during the study period.
The debate continues whether OHT is a better surgical treatment for HLHS than the Norwood procedure. It is not surprising that centers which perform more OHTs in infants report better patient survival with OHT.13,14 However, the short- and medium-term survival of patients undergoing the Norwood procedure in the most experienced pediatric cardiac centers matches the outcomes of patients treated with OHT.15,16 The management and outcomes of patients with HLHS depend not only on a hospital’s experience with the Norwood procedure or OHT, but also on organ availability and parent as well as physician attitudes toward comfort care.17 Despite of the ongoing controversy, our study found a continuing increase in the proportion of patients treated with the Norwood procedure. This proportion is projected to exceed 50% of all patients with HLHS by 2004. The use of OHT, probably limited by the availability of appropriate donors, remained in the range of 5% without significant increase over time.
There were 637 patients (32.1%) who died in the hospital without surgery. It is possible that these patients may have been scheduled for surgery but died before surgery could be performed. It is also possible that comfort care was intended for these patients, and they died in the hospital after withdrawal of support. In a series of 242 patients with HLHS intended for surgical treatment reported by Jenkins et al,10 2 (1.7%) of 118 patients who were intended for the Norwood procedure died before surgery was performed, and 30 (24.2%) of 124 patients for whom OHT was planned died before surgery. If the proportions of patients who were intended for surgery and died before surgery was performed in our series were similar to Jenkins’ study, we calculated that 20 of the 637 patients died before a proposed surgical intervention. That is, the remaining 617 of the 637 patients died in hospital as the result of planned comfort care. This speculation is supported by the fact that more than half of the patients who died in the hospital without surgical intervention died in the first 3 days of the hospitalization and 79.2% died in the first week of the hospitalization. If this assumption is correct, we estimate a total of 1039 (52.3%) patients in our series were treated with comfort care (patients who had no surgery and either died in the hospital or was discharged from the hospital).
The question remains, is comfort care still a valid option to offer to parents of infants with HLHS at the present time? In our series of patients, it is estimated that >50% infants with HLHS in the study period were treated with comfort care. In a survey of 108 neonatal intensive care unit directors, Caplan et al4 found 76% of respondents discussed comfort care as an option and 24% recommended comfort care only. Our data show that the treatment strategies for HLHS continue to evolve and we see a continuing trend toward increasing use of the Norwood procedure. In fact, we project that the number of infants with HLHS who undergo the Norwood procedure will surpass the number of cases treated by comfort care in 2001. It is, however, important to note that in many other countries such as Japan and the United Kingdom, comfort care is still the predominant option offered to families of infants with HLHS. In a recent report from a university hospital in Canada, more than half of parents opted for comfort care.18
The finding that patients from teaching hospitals and in Era 2 were more likely to undergo the Norwood procedure is not surprising. However, the slight but statistically significant increased use of the Norwood Procedure in the South cannot be easily explained. Analysis of the NIS database revealed many hospitals in the South with a large number of Norwood procedure cases. Whether the finding of a higher incidence of the use of the Norwood procedure in the South is attributable to geographic differences in clinical practice or hospital selection bias in the NIS database requires additional investigation. Importantly, we found that the patient’s gender, race and ethnicity, insurance type, and home income level are not significant determinants of treatment strategies.
As has been shown in many other studies, our data indicate that higher case volume is associated with lower in-hospital mortality. As a result, the overall mortality for HLHS as well as the mortality for the Norwood procedure is declining over time. This is consistent with the finding that although mortality for the Norwood procedure in many centers remains high, some large, experienced centers are able to achieve a 5-year survival of 70%.15,16
Because of the availability of a multistate hospital discharge database from the Health care Cost and Utilization Project, we are able to report on the management of HLHS across the nation over a period of 10 years. However, a major limitation of this study results from potential errors related to miscoding and missing data in an administrative database.19–21 For example, the proportion of patients with HLHS who also had transposition of great arteries, anomalous pulmonary drainage, or atrioventricular canal defect seemed to be low in our data. This could be attributable to underreporting of clinical information that was not listed in the principal diagnosis. We attempted to minimize the effect of errors from miscoding by careful selection of study cases and elimination of multiple admissions of the same patients. Nonetheless, in interpreting our results, it should be remembered that a potential bias of our findings may result from data underreporting or miscoding.
Another limitation of the study is our inability to track the management and outcomes of patients who were transferred to hospitals not included in the NIS database. Because the NIS database uses a stratified sampling technique, it is designed to represent all hospitals in the United States.12 Therefore, patients who were transferred to hospitals not in the NIS database theoretically received the same types of care of patients in the database. In addition, patients who were transferred from hospitals not in the NIS data to hospitals in the NIS data might be considered to offset the effect of patients who were transferred out of NIS data. Nevertheless, the 26.2% of patients who were transferred to non-NIS hospitals could potentially affect of the results of management and outcomes.
The mortality rates reported in our study were calculated from deaths that occurred during hospitalization. The overall mortality rate of HLHS, including deaths that occurred out of hospital, should be higher than reported in our study. For instance, if we consider the 21.2% of patients discharged from the hospital without surgery were out-of-hospital mortality, the overall mortality would be higher than 60%.
Clinical management of neonates and infants with HLHS continues to evolve. Among the 3 management options for HLHS used by most hospitals in the United States, patients treated by the Norwood procedure was increasing, cases of comfort care were decreasing, and there was no apparent change in the use of OHT. Outcomes of patients treated by the Norwood procedure had improved significantly over the study period. Gender, race/ethnicity, insurance, and home income are not important determinants of treatment choice, whereas patients in teaching hospitals, in the South, and in later years were more likely to undergo the Norwood procedure.
Dr Chang was a postdoctoral fellow of the Agency for Healthcare Research and Quality and received an institutional research grant from the Harbor-UCLA Research and Education Institute. Dr Chen was a Robert Wood Johnson Clinical Scholar at the University of California, Los Angeles.
- Received January 28, 2002.
- Accepted March 28, 2002.
- Reprint requests to (R-K.R.C.) Division of Cardiology, Department of Pediatrics, Harbor-UCLA Medical Center, 1000 W Carson St, Torrance, CA 90509. E- mail:
- ↵Fyler DC. Report of the New England Regional Infant Cardiac Program. Pediatrics.1980;65 :436– 439
- ↵O’Kelly SW, Bove EL. Hypoplastic left heart syndrome. Terminal care is not the only option. BMJ.1997;3147074 :87– 88
- ↵Agency for Healthcare Research and Quality. Introduction to the HCUP Nationwide Inpatient Sample (NIS). Available at: http://ahrq.gov/data/hcup/nisintro. htm#over
- ↵Fortuna RS, Chinnock RE, Bailey LL. Heart transplantation among 233 infants during the first six months of life: the Loma Linda experience. Loma Linda Pediatric Heart Transplant Group. Clin Transpl.1999:263– 272
- ↵Mahle WT, Spray TL, Wernovsky G, Gaynor JW, Clark BJ 3rd. Survival after reconstructive surgery for hypoplastic left heart syndrome: a 15-year experience from a single institution. Circulation.2000;102(suppl 3) :136– 141
- Copyright © 2002 by the American Academy of Pediatrics