4
NEWBORN SCREENING FOR HIV INFECTION
Proposals for newborn HIV screening have cited several possible goals: (1) identification of potential HIV infection in the neonate in order to provide early therapeutic intervention and more aggressive medical management, (2) modification of routine pediatric immunization practices, (3) prevention of horizontal HIV transmission (from infant to caretaker or medical provider), (4) enhancement of epidemiological research, including surveillance and natural history studies, and (5) identification of maternal infection. The committee carefully examined these goals in the context of the current state of the art in HIV diagnostic technology and medical therapy for HIV-infected infants and children. The following section presents the committee's assessment of and conclusions regarding whether the above goals could reasonably be achieved through neonatal HIV screening
Early Therapeutic Intervention for Asymptomatic HIV-Infected Children
The major goal of screening newborns for HIV infection is improvement, through early medical intervention, of the duration or quality of HIV-infected infants' lives. Considerable medical and scientific debate has centered on whether this objective can be realized. The controversy stems from difficulties in quantifying the level of medical benefit associated with early therapeutic intervention (both antiretroviral therapy and prophylaxis against opportunistic infection) for asymptomatic HIV-infected children. Research has shown that zidovudine (AZT) therapy in children with symptomatic HIV infection and AIDS may lead to improvements in weight and growth parameters and in neurological and immunologic function, and
may even prolong life (McKinney et al., 1990; Pizzo, 1990).1 Zidovudine now constitutes standard therapy for children with AIDS. In addition, clinical trials of zidovudine among asymptomatic HIV-infected adults (with CD4+ cell counts below 500) suggest that early intervention delays the progression of disease (Volberding et al., 1990). Whether such benefits will be found among asymptomatic HIV-infected children, especially newborns, is unknown. The short-or long-term therapeutic risks for infants and children may also be different from those for adults. Similar uncertainty surrounds prophylactic therapy for Pneumocystis carinii pneumonia (PCP), one of the most common and devastating opportunistic infections affecting perinatally infected children (Scott, 1989). PCP prophylaxis has been effective for HIV-infected adults (CDC, 1989b) and for pediatric cancer patients (Hughes et al., 1977, 1987); there is reason to believe that HIV-infected children may derive comparable benefits, although this assumption is unconfirmed at this point.
Limitations in HIV Diagnostic Technology for Infants
Even if early therapeutic intervention confers some benefit to HIV-infected infants, deciding who to treat in early infancy is problematic, HIV-antibody tests are currently the most appropriate instrument for screening purposes, but these tests do not differentiate truly infected neonates from those who are uninfected. (All babies born to HIV-infected mothers carry passively acquired maternal antibodies, which may persist for up to 15 months of age; only about one-third of such infants are actually infected.) Both HIV p24 antigen tests and virus culture have been used to diagnose HIV infection in adults and children; however, their relative insensitivity in young infants makes them inappropriate for use as screening tools.2 Several other tests are under investigation (e.g., assays for IgM and IgA HIV antibodies, which do not cross the placenta [Weiblen et al., 1990]; in vitro HIV antibody production [Amadori et al., 1988]; polymerase chain reaction [Rogers et al., 1989]) that may eventually be useful in the early diagnosis of HIV infection in infants, but these tests require more extensive development and evaluation before they are ready
for wide-scale use. In particular, their reliability and validity need to be established, as do guidelines for their interpretation. There are also practical considerations (e.g., cost, specialized laboratory expertise, time required to perform the test) in the wider application of such assays.
In view of the limitations of current serological tests in distinguishing infected infants from those who merely carry maternal antibody, a logical question is, should all HIV-seropositive infants (approximately two-thirds of whom are likely to be uninfected) be treated prophylactically to ensure that those who actually are infected receive early treatment? Many pediatric experts believe that, in the absence of known HIV infection, HIV-antibody-positive infants should not receive antiretroviral therapy or PCP prophylaxis as a matter of course. These therapies are not without toxicity, and prescribing them for all seropositive infants means that uninfected children would be exposed to substantial toxicity without deriving any medical benefits. Moreover, the potential long-term toxicity of such exposure is still uncertain, particularly for infants.
Aggressive Medical Management of HIV-Seropositive Infants
Some experts argue that knowing a child is "potentially" infected (i.e., HIV seropositive) at birth would lead to more aggressive medical management and possibly earlier detection of disease. Part of the support for this view comes from an understanding that subtle indications of progressive immunologic impairment and clinical illness may be present well before more serious symptoms of HIV disease are manifest. For example, although the common wisdom has been that PCP presents as a fulminant, full-blown pneumonia in HIV-infected infants (which may be fatal), recent anecdotal evidence suggests that, in retrospect, many infants who initially developed PCP had early, subtle manifestations of HIV infection, such as neurodevelopmental abnormalities, failure to thrive, and persistent thrush, well before they succumbed to pneumonia (Connor et al., 1990). If seropositive infants are carefully monitored from birth, it is argued, signs and symptoms that may herald the onset of more severe opportunistic infections can be identified early and treated more vigorously. The magnitude of benefit derived from such intensive primary care and early medical intervention is still uncertain. Quality of life may well be improved, but it is unclear whether careful follow-up and aggressive medical management of seropositive children would actually lead to better prognosis for those children who are truly infected. Many, if not most, of these infants are likely to be classified as at risk for reasons other than exposure to HIV (e.g., maternal drug use, socially disadvantaged circumstances); consequently, they should be candidates for intensive primary
care regardless of their HIV-antibody status. It is unclear, therefore, what would be done differently for these infants if HIV infection were suspected.
Given present uncertainty regarding the benefits and risks of early therapeutic intervention for asymptomatic HIV-infected infants and the difficulty in distinguishing infants with only internal HIV antibody from those who are truly infected, the committee concludes that, at present, insufficient medical benefits have been demonstrated from newborn HIV screening to justify its implementation. All infants at risk of adverse health outcomes because of poverty, social circumstances, or parental risk factors would benefit from comprehensive primary care. The rationale that HIV screening will identify infants for intensive primary care is not sufficient by itself to warrant screening of all newborns. Nevertheless, the committee encourages providers and medical centers to develop an aggressive primary care system for all infants at increased risk of adverse health outcomes.
The committee also opposes newborn HIV screening because it is tantamount to involuntary maternal screening in that testing newborns for HIV antibodies reveals their mothers' infection status. Using newborn HIV screening to identify HIV-infected mothers would also mean that postpartum women currently would be the only civilian, noninstitutionalized adult population not given the opportunity to consent to or refuse HIV testing, an outcome that is ethically unacceptable.
Other Considerations for Newborn HIV Screening
Three additional arguments for newborn IV screening (secondary to the potential for early therapeutic intervention) have been proposed. First, knowing that a child is "potentially" infected with HIV may trigger modifications in pediatric immunization schedules—for example, administering inactivated poliovirus vaccine rather than oral poliovirus vaccine (OPV). As a justification for screening, this rationale lacks strength: there is little empirical evidence to indicate that the adverse neurological sequelae theoretically associated with the administration of OPV in HIV-immunocompromised children actually occur (CDC, 1986; McLaughlin et al., 1988).3 Second, identifying neonates at risk of HIV infection may prevent horizontal transmission. Yet transmission of HIV from young children to caretakers is extremely rare (Rogers et al., 1990), and prudent
use of infection control techniques in handling blood and body fluids from all patients (CDC, 1987b) would further minimize the risk of transmission in the health care setting. Finally, neonatal HIV screening could enhance the conduct of natural history studies of perinatally acquired HIV infection and the follow-up of seropositive but uninfected infants for evidence of any long-term adverse effects of perinatal HIV exposure. None of these reasons are sufficient to justify screening newborns for HIV infection at this time, particularly in the absence of clearly safe, effective early therapeutic intervention for HIV-infected infants.
Yet despite the committee's objections to newborn HIV screening at this time, it recognizes that there may be circumstances warranting individual HIV testing of a newborn for medical reasons. For example, when a health care provider has reason to suspect that an infant is at increased risk of infection but the infant's mother is not competent or available to consent to testing for herself, HIV-antibody testing of the infant may be indicated to manage the infant's care more effectively. In this case, the decision to test should be based on the provider's clinical judgment and assessment of the infant's overall medical condition. As is generally the case, someone who is authorized to speak for the child must consent to testing or treatment, unless there is a medical emergency. If a parent is unable to consent to necessary medical care, an alternate guardian should be sought.
Finally, the committee endorses the continuation of anonymous newborn HIV screening for surveillance purposes. This approach provides unbiased epidemiological data for monitoring national and local trends in the distribution of HIV infection, particularly among childbearing women. These data are also useful in planning and evaluating public health interventions, targeting community outreach and prevention campaigns, and anticipating health care resource needs.
Reevaluating Newborn HIV Screening Policy
In light of evolving HIV diagnostic technology and medical therapy, it is clear that current recommendations regarding newborn HIV screening should be periodically reevaluated. In anticipation of such a need, the committee considered several possible developments that might lead to modification of its present conclusions.
If safe, effective antiretroviral therapy or prophylactic treatment for opportunistic infection and a definitive diagnostic screening tool for newborns were available, the argument for voluntary newborn HIV screening with "fight of refusal," to ensure that all infants who would benefit from early intervention were identified and treated, would be
compelling. In such circumstances, the parent or legal guardian would be informed that, unless he or she expressly refused, the newborn would be tested; effective treatment for the infant would also be offered. This scenario is consistent with the long tradition of voluntary screening with ''right of refusal'' for devastating neonatal conditions for which effective therapy is available (e.g., phenylketonuria). 4
A definitive diagnostic test for neonatal HIV infection, in the absence of additional therapeutic advances, might also change the current situation dramatically. A valid, reliable screening test would make the requirements for therapeutic efficacy less stringent because the risk of exposing uninfected infants to potentially toxic therapy would be diminished. For instance, even if available drug therapy proved only moderately effective, voluntary newborn HIV screening (with specific informed consent from the parent or legal guardian) might be appropriate because infants who would benefit from early treatment could be precisely identified in the first few days or weeks of life.
In the absence of a definitive diagnostic screening tool, however, voluntary newborn HIV screening might also be justified by the availability of safe, effective treatment that cured infection. In eliciting parental informed consent for newborn testing, it would be important to discuss the inherent limitations in neonatal diagnosis (i.e., that testing the newborn reveals maternal infection but does not confirm neonatal infection). Moreover, whether to expose potentially uninfected infants to unknown therapeutic risks in order to derive therapeutic benefits for truly infected infants is a question that must be weighed carefully. For example, if such therapy were relatively benign, more benefit might be derived from administering it uniformly to all seropositive infants to ensure that those who were truly infected received early treatment. Above all, the providers decision to treat a seropositive infant (without known infection) should be made in concert with the parent or parents and include complete disclosure of the known risks and benefits of such therapy.
Without a precise diagnostic screening tool, voluntary newborn HIV screening might also be warranted if early therapeutic intervention (e.g., antiretroviral therapy that delayed the progression of infection or extended the lives of asymptomatic children, or effective prophylaxis against opportunistic infection) was shown to provide clear medical benefit to HIV-infected infants. The same information described above regarding the limitations in neonatal diagnosis should be conveyed to the parents. In addition, because it is likely that any effective drug therapy (whether curative, as in the earlier scenario, or ameliorative) would be sufficiently
toxic to warrant caution in prescribing it for all HIV-seropositive infants, the risks and benefits of treating versus not treating should be weighed individually in consultation with the parent or parents. The potential unknown long-term toxicity of such therapy should also be considered. Finally, decisions regarding initiation of therapy should take into account the heterogeneity of the clinical course of HIV disease in perinatally infected infants—for instance, most infants will develop symptoms within the first year or two of life, whereas some will remain asymptomatic for a number of years.