Hypoplastic Left Heart Syndrome, Extreme Prematurity, Comfort Care Only, and the Principle of Justice

Abstract

At our institution and at many other pediatric centers, there have been ongoing debates about the management of 2 distinct medical problems: hypoplastic left heart syndrome (HLHS) and extreme prematurity (EP). For each problem, physicians sometimes struggle with the question of when it is appropriate to offer parents the option of providing comfort care only (CCO).HLHS and EP are very different diagnoses, and the debates about management have usually been separate and unrelated discussions. However, they share much in common. Both diagnoses are fatal in the absence of aggressive medical intervention. Even with aggressive management, there is significant mortality associated with both of them. All survivors face a difficult and expensive hospital course. For each condition, there is significant morbidity, including neurologic morbidity, among many of the survivors. Finally, for both diagnoses, some pediatricians have felt that aggressive medical or surgical care should be optional rather than obligatory, and parents should be given the option of providing CCO.Given these features common to both HLHS and EP, it may well be that our approach to one should be informed, at least in part, by our approach to the other. The approach to both conditions should be guided by basic ethical principles. In particular, the principle of justice should be kept very much in mind.Informed consent, a fundamental tenet of modern medical ethics, requires that the patient give consent for any proposed treatment, such as surgery or mechanical ventilation, after having been presented the relevant information in an unbiased and noncoercive manner. Anticipated consequences of all treatment options, as well as their likelihood of success, should be reviewed.1 The right to refuse aggressive medical or surgical intervention seems, at first, to be consistent with the doctrine of informed consent. If a patient refuses to give consent for a procedure, it should not be performed, even at the cost of the patient's life. Exceptions are made for individuals who are not competent to decide or speak for themselves, including infants. For them, surrogate decision-makers give or withhold consent on their behalf. For infants, this task falls to the parents in nearly all cases.It is important to understand, however, that the parents' right to refuse a recommended treatment on behalf of their child is not as strong as a competent adult's right to refuse on his or her own behalf. As pediatricians, we recognize a parent's right to refuse only until that refusal is clearly opposed to the child's best interest.2 In other words, the doctrine of informed consent (and the right to refuse treatment) may be trumped, in the case of children, by the patient's best-interest standard. This standard directs parents and physicians to decide for children on the basis of an assessment of the benefits and burdens of the proposed treatment to the patient. At some threshold, the chance of a successful outcome becomes so high that it becomes inappropriate to withhold the treatment. Of course, the location of the threshold at which to override parental refusal is subjective.How does the parental right to refuse treatment, and the limitations of that right, apply to the diagnoses in question? When trying to determine if aggressive treatment should be offered, or perhaps even be obligatory, it is important to know the likelihood that the treatment will be successful. Success can be defined in terms of survival and subsequent disability.There are 2 primary surgical treatments available for newborns with HLHS: cardiac transplantation and staged surgical repair (3 separate surgeries, the first being some variation of the Norwood procedure). A recent review of data from 20 centers by the Pediatric Heart Transplant Study Group showed that, for patients with HLHS who underwent cardiac transplant, the survival rate at 5 years was 72%. If one also takes into account those with HLHS who were listed for transplant but died while waiting for a heart, the survival rate at 5 years was 54%.3The limited supply of donor hearts, combined with the continued improvement in outcome for staged surgical repair, has resulted in an increasing preference for the latter among those who choose surgical intervention. Staged surgical repair is associated with a risk of mortality for each of the 3 surgical procedures, and for the months in between those procedures, but most deaths occur in association with the first (Norwood) procedure. A large database review has determined the short-term outcome for stage I repair of 801 patients with HLHS treated at 29 centers from 1998 to 2001. The rate of survival to hospital discharge after the initial (Norwood) procedure was 68%. Survival varied widely between centers. The 10 centers with the highest surgical volume achieved a better survival rate (71%) than those with the lowest surgical volume (48%).4Regarding long-term survival, a Canadian center recently reported a 54% survival rate at 5 years for infants with HLHS born from 1996 to 1999.5 Patients treated more recently lack follow-up of that duration. It is worth noting, however, that 1 of the busiest US centers for HLHS surgery calculated a “predicted three year survival estimate” of 70% for those who underwent stage I surgical repair in 2002–2004 on the basis of analysis of follow-up data at a mean of 17 months.6 Another major US center recently reported a hospital survival rate for stage I of 93%, and up to stage II of 81%,7 which suggests that 3-year survival rates even higher than 70% may be attainable.Survivors of staged surgical repair of HLHS show a variety of neurologic problems, including motor delays, behavioral abnormalities, learning disabilities, mental retardation, and other problems. A major US cardiac center recently published neurodevelopmental outcome at 1 year of age for 88 patients who underwent surgical repair for HLHS from 1998 to 2003. Patients with recognizable chromosomal or phenotypic syndromes at birth were excluded from the study. At 1 year, abnormal or suspect neuromuscular examination results were found in 64% of the subjects. The Mental Developmental Index was <70 for 11% of the patients, and the Psychomotor Developmental Index was <70 for 48% of them. Many of these children (35%) had known or suspected genetic syndromes at 1 year that were not recognized at birth.8Neurodevelopmental follow-up performed at as early as 1 year should be viewed with caution, because it may overestimate subsequent disability, as has been found for EP (see below). One study of school-aged children who previously underwent staged surgical repair for HLHS revealed that 18% of these children were classified as having mental retardation. Although many survivors of HLHS repair have some degree of neurologic impairment, it would seem that most do not have a major disability and, by their and their parents' assessment, have been left with a good quality of life.9The likelihood of survival would seem to be a straightforward question, but in the case of EP it is not. Survival data at borderline gestational ages are available, but these data can be misleading when trying to determine the odds of survival with aggressive treatment.10 Survival statistics for EP are typically based on the percentage of all infants born at a given gestational age who survive to hospital discharge. Unlike the reported survival statistics for patients with HLHS, preterm newborns who received CCO are included in the denominator. Obviously, this will lead to underestimation of the effectiveness of treatment.Perhaps a more valuable statistic when assessing the effectiveness of a treatment would be the “potential for survival,” which would reflect the percent survival among those for whom all efforts were made. These numbers are hard to find for EP, but perhaps 1 proxy is survival among those for whom mechanical ventilation was initiated. Among 14 major US centers in the National Institute of Child Health and Human Development Neonatal Network, overall the survival rate to discharge was 26% for those born between 1998 and 2003 at 23 completed weeks' gestation. However, for those who were born at 23 completed weeks' gestation and provided with at least a trial of mechanical ventilation, the survival rate was 37%. Among those who were born at 24 and 25 completed weeks' gestation and provided a trial of mechanical ventilation, survival rates were 60% and 77%, respectively.11 Potential for survival may be even greater than these figures suggest if mechanical ventilation was initiated but subsequently withdrawn from some infants who might ultimately have survived, perhaps out of concern for subsequent disability.Morbidity is also difficult to predict because of inherent flaws in the available data. The network data demonstrate that for infants born at 24 weeks' gestation, 50% of survivors had neurodevelopmental impairment at 18 to 22 months' corrected age, defined as a Mental Developmental Index or Psychomotor Developmental Index of <70, moderate or severe cerebral palsy, bilateral blindness, and/or bilateral hearing loss that requires amplification. Of those impaired, just under half were felt to be profoundly impaired.11 It should be noted, however, that recent data suggest that assessment of preterm infants at 18 to 22 months may significantly overestimate subsequent disability.12,13 Also, most survivors with a birth weight of <1000 g who have reached young adulthood have a self-reported quality of life similar to that of peers who were born at normal birth weight.14For extremely preterm newborns and for those with HLHS, physicians need to decide if CCO is an ethically acceptable option. For newborns with each disorder, parents have a right to know that there is a risk of death despite treatment, a risk of permanent disability should the child survive, and the certainty that treatment will entail a great deal of intensive medical care, undoubtedly with some pain. Whether parents should also have the right to refuse treatment (ie, request CCO) may ultimately depend on the magnitude of these risks. For example, physicians might be rightfully reluctant to offer CCO if the risk of mortality with treatment were 10% and the risk of disability among survivors were also 10%. Similarly, it seems likely that most physicians would clearly consider CCO an option if those risks were both 90%. The difficulty, both intellectually and ethically, lies in between, where the current data for EP and HLHS are found. At some point, the predicted outcome will become so good that aggressive medical care should be required. How do we locate that point? Ultimately, that will be a subjective judgment, but a few simple guidelines may prove helpful.First, when determining whether to offer CCO as an option, we must remember that the question at hand is not whether it is the preferable choice but, rather, whether it is an ethically acceptable one. This should be an easier standard to reach. Unless physicians believe that it is ethically unacceptable, not just inadvisable, parents should be offered the option of CCO. One possible test of the acceptability of CCO is our willingness to take the parents to court in an effort to override their refusal of treatment. If we would be unwilling to do so, perhaps this suggests that the strength of our judgment in favor of treatment is not strong enough to deny parents the choice.Second, honest communication requires that parents be told of all their options in an unbiased and noncoercive manner. It is not reasonable for us to fail to discuss an option simply because we would recommend against it unless we feel strongly that it would be ethically unacceptable. When considering CCO, if we feel strongly that parents should decide a certain way, it is reasonable for us to say so and to provide our reasons, but only after the relevant information, anticipated results, and all options have been presented in an objective manner. Full disclosure also requires us to inform parents if other medical centers are willing to provide an option, including CCO, that we are not willing to provide.The third guideline for determining when CCO should be offered is based on the principle of justice. Discussions of justice in the medical setting frequently focus on social justice. Important issues, such as access to care and fair allocation of resources, fall within that domain. Here, however, consideration of the principle of justice refers to one of its most basic requirements: the requirement to treat equals equally.15 This should pertain even to patients with different diagnoses that require different therapies. If 2 newborns have similar prognoses for survival and morbidity with aggressive care, it does not seem appropriate to offer parents CCO in one case but not the other.Possible justifications for an inconsistent offer of CCO might include a significant difference in predicted disability, a significant difference in the pain and suffering the child would have to undergo as part of the treatment, or, perhaps more controversially, a significant difference in impact on the family or society. Absent any of these, it seems fundamentally unjust to treat 2 patients with similar potentials for survival differently in terms of offering CCO.When comparing data from the same recent years, it would seem that the potential for long-term survival for HLHS is similar to that for EP at 24 weeks, and although hard to compare, disability rates may also be roughly similar. What is the current practice regarding CCO for these 2 conditions? In the past, CCO was commonly chosen for neonates with HLHS, but with improvements in surgical outcome, this seems to be the case less often. Indeed, the impressive results for staged surgical repair in recent years have led some physicians to conclude that CCO should no longer be an option.16Neonatologists generally agree that below some minimal gestational age parents should have the right to refuse resuscitation and instead choose CCO for their child. The majority also agree, however, that above some threshold gestational age parents should no longer be given the option of CCO. The American Academy of Pediatrics currently recommends that at 23 to 24 weeks' gestation physicians should defer to the wishes of informed parents regarding whether to attempt resuscitation and mechanical ventilation.17 The Nuffield Council of the United Kingdom recommends deferring to parents' wishes regarding intensive care versus CCO at 23 weeks' gestation and, if the parents and clinicians agree that intensive care is not in the child's best interest, at 24 weeks as well.18 Neither group recommends offering the option of CCO by 25 weeks' gestation. Recent data suggest, however, that many American neonatologists no longer give parents the option of CCO even by 24 weeks' gestation.19Given similar outcomes, the principle of justice suggests that a neonate with HLHS should be treated similarly to a neonate with EP at 24 weeks' gestation. Specifically, if parents of neonates born at 24 weeks are offered the option of providing CCO, as recommended by the American Academy of Pediatrics, it becomes difficult to justify denying parents of a child with HLHS that same right. Conversely, if we are convinced that children with HLHS deserve aggressive treatment even in the face of parental refusal, it is unclear why the child born at 24 weeks' gestation should not warrant the same treatment. The point here is not that parents should or should not have the option of CCO in either situation, but that if we are going to treat one group of patients (and parents) differently than the other, we need to identify a morally relevant difference between them.As noted earlier, an apparent disparity in suffering between the 2 groups as a result of treatment could potentially serve as that morally relevant difference. It may be that one group or the other suffers more overall, but it is not inherently obvious, and it would be difficult to quantify and compare. At present, there is not sufficient evidence for a disparity in suffering between these 2 groups to justify different approaches to offering CCO.Perhaps, when determining whether to offer CCO with either diagnosis, a closer look at the predicted mortality and morbidity for that particular child would be more appropriate. It has been shown, for example, that the predicted survival for a child born at 24 weeks' gestation varies widely depending on such things as birth weight, antenatal steroid exposure, gender, and whether it was a singleton or multiple gestation.11 Similarly, predicted survival for a child with HLHS is influenced by birth weight and certain cardiac and noncardiac characteristics.20–22 A better application of the principle of justice might be to say that 2 infants, regardless of diagnosis, who are roughly equal in terms of predicted survival and morbidity, should be treated equally. That is, parents of any infant with a predicted long-term chance of survival, or survival without severe disability, below some threshold should be given the option of CCO. Of course, we would still need to determine what that threshold should be. By this view, we may well determine that parents of some (but not all) preterm neonates at 24 weeks' gestation should be offered CCO, and parents of some (but not all) infants with HLHS should similarly be offered CCO.HLHS and EP are both fatal in the absence of aggressive intervention. For both conditions, there is significant mortality and morbidity among those who receive intervention. For both, pediatricians have long debated if and when it is appropriate to offer parents the option of CCO. Guidelines to assist in making that determination have been presented, with an emphasis on the principle of justice, which suggests that infants with similar prognoses should be given similar options. If there are inconsistencies in the way these 2 groups (or any 2 groups) of patients are treated with regard to the option of CCO, that inconsistency should be justified by identifying a morally relevant difference between the groups. If such a difference cannot be identified, the inconsistency should be eliminated.Finally, we need to remember that therapies are continuously evolving, and outcome data should be continuously updated and refined. We particularly need to avoid making judgments on the basis of obsolete prognostic data. Situations in which CCO was ethically acceptable in the past may no longer be so, and options deemed acceptable today may not be so in the future. The decision of whether CCO can ethically be offered for any diagnosis is not one we can make and then set aside for years. As the data evolve and predicted outcomes, we hope, continue to improve, our judgments need to be revisited frequently.The idea of our willingness to take a case to court as one possible measure of our own judgment of ethical acceptability was based on a proposal by Harris Jacobs, MD.