| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2010;139:119-127
© 2010 The American Association for Thoracic Surgery
Congenital Heart Disease |
a Department of Surgery, Oregon Health & Science University, Portland, Ore
b Department of Cardiothoracic Surgery, University of Michigan, Ann Arbor, Mich
c Division of Pediatric Cardiac Surgery, Doernbecher Children's Hospital, Oregon Health & Science University, Portland, Ore
Received for publication June 24, 2008; revisions received February 27, 2009; accepted for publication April 27, 2009. * Address for reprints: Tara Karamlou, MD, Department of Surgery, L 223, Oregon Health & Science University, 3181 Sam Jackson Park Rd, Portland, OR 97239. (Email: karamlou{at}comcast.net).
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionLimitationsConclusionAppendix 1References |
|---|
Methods: Neonates with hypoplastic left heart syndrome were retrospectively identified by use of the Nationwide Inpatient Sample 1988–2005. Treatment was categorized as (1) transplantation, (2) Norwood operation (as defined by Risk Adjustment in Congenital Heart Surgery), (3) transfer to another facility, or (4) no surgical intervention (comfort care).
Results: A total of 3286 neonates were identified, yielding a national estimate of 16,781 ± 586 cases. Of these, 2% (348 ± 47) underwent transplantation, 16% (2767 ± 286) had Norwood operations, 25% (4143 ± 156) were transferred to another facility, and 57% (9523 ± 436) had comfort care. Changes in practice patterns occurred over time, with an increasing number of neonates undergoing Norwood, concomitant with decreasing numbers undergoing transplantation (P < .001). Bias toward the Norwood operation over time paralleled a significant, nearly linear decrease in the in-hospital mortality rate for the Norwood operation, from 86% in the earliest sextile to 24% in the most recent sextile (P < .001). Prevalence of transfer to definitive care hospitals remained constant over time, as did the number of infants (approximately half) who received no surgery (comfort care).
Conclusions: Despite improved surgical outcomes, the majority of infants continue to receive no surgical care. There has been an increase in the number of infants offered the Norwood operation for hypoplastic left heart syndrome over the past 2 decades, which seems to have come mostly owing to a decrease of transplants. The advent of prenatal diagnosis has not decreased the proportion of neonates born at institutions unequipped to provide definitive care.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionLimitationsConclusionAppendix 1References |
|---|
It is unclear what factors motivate an institutional choice toward surgical treatment versus no intervention, especially considering the lack of consensus among physicians, 40% of whom recommended a disparate treatment strategy than they would themselves choose.1,4-9 Additionally, given the comparatively poor long-term prognosis of HLHS, especially among certain subsets of patients with concomitant syndromes (Turner syndrome, 22q deletion), the option for compassionate care without intervention (comfort care) may be underused in the United States,10,11 although data elucidating the incidence of comfort care are lacking.
There have been three major reports1,4,5 chronicling the national practice patterns for infants with HLHS. However, none of these included the most recent era with the introduction of newer management strategies, fetal intervention, and widespread prenatal diagnosis, which may have major impact on treatment selection. Additionally, prior reports have not included the entire spectrum of United States hospitals treating infants with HLHS, nor used a validated and widely accepted algorithm to define the Norwood operation.12
We therefore undertook this study to determine current practice national patterns for HLHS and identify determinants associated with a decision toward a particular management strategy. We also focused on national in-hospital mortality and morbidity trends over the course of an 18-year period.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionLimitationsConclusionAppendix 1References |
|---|
The database is a stratified, cross-sectional sample that includes approximately 20% of all non-federal hospital discharges in the United States. The sampling protocol is such that when a hospital is chosen, all discharges from that hospital for the selected time period are included. To ensure the representative nature of the database, the NIS is stratified by geographic region, urban versus rural location, teaching status, hospital ownership, and hospital bed size. This study's data were derived from the combination of the NIS databases from 1988 to 2005, and our analysis uses sampling weights provided within the NIS to derive national estimates. The sampling frame of the NIS has been nearly constant over time except for a modification in 1998, which excluded short-term rehabilitation hospitals from the sampling frame. To account for this small change, we have used the revised NIS trend weights, published by the Agency for Healthcare Research and Quality,13 which allows one to compare across years with the same effective sampling scheme. The NIS is not a self-selecting, and potentially growing, set of hospitals reporting their data. Rather, it is a national sample of hospitals representing 20% of all discharges. Additionally, it is designed to provide national estimates, not a total of participating hospitals. NIS data are available from 1988 to 2005 over which time the number of states participating in the NIS has grown from 8 to 37. In 2005, the database contained discharge data on approximately 8 million hospital stays at 1054 hospitals in 37 states.
Patients and Hospitals
We sought to identify the most homogeneous population of infants with HLHS, whose morphologic features would be amenable to either Norwood operation or transplantation (OHTx), and therefore excluded those infants with other complex cardiac anomalies (eg, total anomalous pulmonary veins, heterotaxy syndrome). The NIS database was searched for the individual years 1988 to 2005, selecting out hospital discharges that met the following criteria: (1) infants 30 days of age or younger; (2) possessed a primary International Classification of Diseases, ninth revision, clinical modification (ICD-9-CM) diagnosis code of 746.7 (HLHS) without concomitant cardiac diagnoses; (3) underwent a surgical procedure (Norwood operation or OHTx), were transferred to another facility, or were discharged from the hospital or died in-hospital without any intervention (comfort care). Discharges with additional cardiac procedure codes were excluded. Because there is no ICD-9-CM code for the Norwood operation, we used the Risk Adjustment in Congenital Heart Surgery (RACHS-1) definition with slight modification to identify these patients as described in Appendix 1. The ICD-9-CM procedure code 37.5 was used to identify patients undergoing OHTx.
To minimize an important source of error, we also determined, to the extent possible, the final disposition of transferred neonates. The NIS contains admission codes identifying neonates who were received in-transfer from another hospital. We therefore were able to determine the disposition of neonates who were transferred from a hospital included in the NIS sample that year to a hospital that was also included within the NIS sample that year. We incorporated this admission code and then reanalyzed our data set to investigate trends in final treatment for neonates who were transferred over time. We then compared the disposition of the transferred neonates at the receiving hospital to the disposition of neonates who were not transferred.
Statistical Analysis
Overall descriptive statistics were computed. Treatment groups were compared by the Rao–Scott 
2 for categorical variables or t test for continuous variables. Years were grouped into sextiles and annual RACHS-1 volume for each unique hospital was calculated. The probability for each management option was investigated by multivariable ordinal regression models (PROC SURVEYLOGISTIC), adjusted for patient and hospital level characteristics. Risk-adjusted in-hospital mortality was also compared among surgical treatment strategies (OHTx or Norwood) with multivariable ordinal regression using PROC SURVEYLOGISTIC. C-indices were generated for all logistic regression models to provide insight into model discrimination. SAS software, version 9.1 (SAS Institute, Inc, Cary, NC) was used to fit the models to account for the survey design of the NIS, the potential clustering of outcomes within a hospital, and the decrease in clustering occurring over increasing time intervals. Institutional review board approval was obtained, but given the de-identified nature of the NIS, patient consent was waived for this study.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionLimitationsConclusionAppendix 1References |
|---|
30 days of age). Overall, there was a significant increase in the proportion of neonates undergoing the Norwood operation over time (weighted frequency 39.9 ± 14.6 [2% of total cases] in the first sextile [1988–1990] versus 923.8 ± 164.2 [24% of cases] in the most recent sextile [2003–2005]; P < .001). In contradistinction, OHTx prevalence declined steadily from a peak of 120 ± 13 (5% of cases) in the third sextile (1994–1996) to 14 ± 6 (<1% of cases) in the most recent sextile (P < .001). Despite the increased national popularity of the Norwood operation, however, the proportion of neonates having nonoperative therapy, either comfort care or transfer (P = .03), also rose modestly over time (Figure 1
).
|
|
Mortality
There were 975 total deaths out of 3286 neonates (national estimate of 5002 ± 215) during the study period, yielding an overall mortality rate of 30%. Overall mortality was higher for the Norwood operation (n = 178/975; 18%) than for OHTx (n = 21/975; 2%). However, Norwood-specific mortality decreased nearly linearly over time, from 86% (6/7) in the earliest sextile to 24% (47/191) in the most recent sextile, whereas in-hospital mortality for OHTx remained constant (Figure 2
).
|
|
|
|
|
|
Influence of Hospital Volume on Management Strategy
The median annual number of RACHS-1 coded cases performed among included hospitals was 107 (range, 0–723). Higher annual RACHS-1 volume was significantly associated with a decision toward Norwood operation (odds ratio, 1.004; 95% CI, 1.004–1.005; P < .001). Figure 5 is a graphic description of the proportion of neonates managed with a particular treatment strategy within each hospital as a function of increasing percent intervention, excluding those hospitals whose annual RACHS-1 volume was less than 3 cases, and segregating OHTx and non-OHTx hospitals. There are clear treatment biases, with the majority of hospitals choosing to transfer neonates with HLHS or provide comfort care. A minority of hospitals uniformly performed the Norwood operation as the only surgical therapy and a small number provide only OHTx. Importantly, hospitals providing both the Norwood operation and OHTx had the highest percentage of intervention in neonates with HLHS and used transfer with decreasing frequency.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionLimitationsConclusionAppendix 1References |
|---|
Three earlier reports by Gutsegell and Massaro,5 Gutsegell and Gibson,4 and Chang, Chen, and Klitzner1 have elucidated similar trends albeit in an earlier time period before the widespread use of prenatal diagnosis and fetal intervention. These earlier reports, though, showed a more precipitous increase in Norwood prevalence, up to 60% in one study,4 coupled with a lower prevalence of nonoperative management. On the basis of their data, Chang, Chen, and Klitzner1 predicted that the Norwood operation would be undertaken in 50% of neonates by 2004. Our study, in contradistinction, suggests that although the Norwood operation has increased nationwide, it is still performed in the minority (25%) of infants in the recent sextile). Moreover, we found that the average percentage of neonates undergoing no intervention over the entire 18-year period was higher than that reported by Chang, Chen, and Klitzner1 at nearly 58% and declined only slightly over time. Discrepant findings may be due in part to differences in the sampling population. We used the NIS, which may be more representative of the overall treatment patterns within all US hospitals (including community and nonteaching institutions) as opposed to a consortium of university hospitals as was used in the studies by Gutsegell and Gibson4 and Gutsegell and Massaro,5 which may have biased their data toward more aggressive treatment protocols.
We also found that an important number of neonates were transferred to other hospitals after initial admission, presumably because they required a higher level of care than available at the admitting institution. The proportion on neonates transferred after initial admission to rural or urban nonteaching hospitals (38%) underscores this contention. The cost of care for these infants, over 10,000 USD, was not inconsequential, and it is likely that definitive care was delayed in these infants. Infants with certain morphologic features requiring urgent intervention, including an intact or highly restrictive atrial septum, are particularly fragile and thereby prenatal triage to a facility equipped to provide definitive care would be advantageous to improve outcome.14-16 Studies by Vlahos,14 Rychik,17 and their colleagues have reported nearly 50% 30-day mortality for this subset compared with 9% for age-matched control patients with HLHS. We attempted to identify infants within this category by selecting those in whom a preoperative balloon atrial septostomy was performed, but the number (n = 10) identified was insufficient to determine whether an important proportion required transfer or received comfort care. However, this limitation notwithstanding, our study highlights an enduring shortcoming of care for infants with HLHS. Our data suggest that even with widespread availability of prenatal diagnosis, an important number of neonates, especially in the rural setting, are still admitted to hospitals not able to provide definitive care. Elucidation of antenatal surveillance and treatment paradigms focused on families residing in rural areas may lead to an overall survival benefit.
Over the 18-year study period, use of the Norwood operation increased nearly linearly concomitant with steadily decreasing national rates of OHTx from a peak in 1996 of nearly 5% of total HLHS cases to less than 1% of total cases in 2005. Although we cannot determine the precise causes of this decline from our retrospective analysis, we agree with Chrisant and colleagues8 that decreasing national OHTx rates reflect a trend toward Norwood palliation. The bias toward the Norwood operation is compounded by improved short-term outcomes coupled with the fact that infants with HLHS have particularly high pretransplant attrition, with nearly 25% dying before receiving an organ despite relatively short wait-list times of less than 18 months.6
A traditional perception has been that the United States is considerably more aggressive in treating infants with HLHS surgically than other developed countries,10 a dichotomy articulated by Elliott as "life at any cost" in the United States compared with a European approach of "quality is more important than quantity." It is unlikely that incidence rates of HLHS have important global variation, but it certainly is true that termination rates in the United States are lower (near 17%), thus increasing the potential number of neonates born with HLHS.18 Brackley and colleagues19 reported a 44% termination rate in an active center in the treatment of HLHS in Birmingham, and another unit in London reported a 63% elective rate of termination in prenatally diagnosed HLHS.20 Surprisingly, though, despite the emerging popularity of the Norwood operation in US hospitals, we also found that "no treatment" (comfort care) still represents the most common treatment option for infants with HLHS nationwide. Furthermore, our study found that comfort care was used in over 50% of neonates in the most recent sextile, suggesting that physicians may be adopting a more balanced approach to infants with HLHS.
Motivation for selection of a treatment strategy is complex, involving the desires and belief systems of both physician and parents. Vandvik and Forde21 reviewed the characteristics of 20 mothers of neonates with HLHS, 10 of whom had children who survived surgery and 10 of whom had children who died after the mothers opted for comfort care. Mothers choosing comfort care (a choice motivated by a desire to "prevent suffering") were more educated, more likely to be employed in health care, and reported a better childhood environment than those mothers choosing surgery (a choice perceived as the only acceptable option). Physician influences were examined in a recent provocative article by Kon, Ackerson, and Lo.7 These authors examined how 454 physicians in the United States made treatment recommendations to parents of infants with HLHS and what factors motivated these recommendations. These investigators found that 26% of physicians did not discuss comfort care, and 25% of those at nontransplant centers do not discuss OHTx with parents, indicating that physicians recommend the treatment preferred at their own institution even when they predict better outcomes with another approach.7
|
Limitations |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionLimitationsConclusionAppendix 1References |
|---|
|
Conclusion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionLimitationsConclusionAppendix 1References |
|---|
|
Appendix 1 |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionLimitationsConclusionAppendix 1References |
|---|
Dx 746.7 Hypoplastic left heart syndrome
Proc 35.41 Enlargement of existing atrial septal defect
or 35.42 Creation of septal defect in heart
Proc 39.0 Systemic–pulmonary artery shunt
or 35.92 Creation of conduit between right ventricle and pulmonary artery
Proc 38.35 or 38.45 Resection of thoracic vessel
or 38.34 or 38.44 Resection of abdominal aorta
or 38.64 or 38.65 Other excision of vessel/aorta
or 38.84 or 38.85 Other surgical occlusion of vessel/aorta
or 39.56, 39.57, 39.58 Repair of blood vessel
or 39.59 or 36.99 Other operation on vessel of heart
Cannot have:
Proc 35.94 Creation of conduit between atrium and pulmonary artery
Proc 35.95 Revision corrective procedure on heart
Proc 39.21 Cavo–pulmonary artery anastomosis
Allow:
Dx 745.5 Ostium secundum atrial septal defect
Dx 746.3 Congenital stenosis of aortic valve
Dx 746.5 Congenital mitral stenosis
Dx 747.0 Patent ductus arteriosus
Dx 747.10 Coarctation of aorta
Dx 747.22 Atresia and stenosis of aorta
Dx 747.89 Other anomalies of great veins
Proc 37.33 Excision of other lesion/tissue of heart
Cannot have:
Any other cardiac diagnosis (Dx)
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionLimitationsConclusionAppendix 1References |
|---|
This article has been cited by other articles:
|
|
 |
|
  B. Murtuza, N. Dedieu, A. Vazquez, M. Fenton, M. Burch, T.-Y. Hsia, V. T. Tsang, and M. Kostolny Results of orthotopic heart transplantation for failed palliation of hypoplastic left heart Eur J Cardiothorac Surg, March 1, 2013; 43(3): 597 - 603. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
