Risk Assessment and Neurodevelopmental Outcomes

Motor Function

Motor impairments including cerebral palsy (CP) are among the most frequently reported neurodevelopmental outcomes for high-risk infants. Motor difficulties may become evident over months or years, yet timely identification of motor difficulties may allow for interventions to improve outcomes, thereby reinforcing the critical importance of vigilant long-term follow-up. CP is defined as a disorder of movement and posture that involves abnormalities in tone, reflexes, coordination, and movement, delays in motor milestone achievement, and aberration in primitive reflexes that is permanent but not unchanging and is caused by a nonprogressive interference, lesion, or abnormality of the developing immature brain. CP is also categorized by type (spastic, dyskinetic, or dystonic); topography (limbs involvement); and descriptors of extent and pattern of involvement (monoplegia, diplegia, hemiplegia, and quadriplegia). Previously, it was considered that CP could not be diagnosed prior to approximately 18 to 24 months corrected age. However, it is now clear that a child may receive an accurate diagnosis of CP or high-risk for CP before 6 months of corrected age using evidence-based and standardized clinical assessment tools and imaging appropriate for age. In a systematic review utilizing Grading of Recommendations Assessment, Development, and Evaluation ( GRADE) framework, Novak and colleagues presented specific essential and additional criteria, and detection pathway options based on level of risk. In terms of standardized assessment tools, they found that in combination with medical history and depending on age of assessment, magnetic resonance imaging, the Prechtl Qualitative Assessment of General Movements, the Hammersmith Infant Neurological Examination, and the Developmental Assessment of Young Children were most predictive for detecting risk for CP. Of note, successful deployment of these guidelines for early diagnosis and risk stratification has been demonstrated across diverse high-risk infant follow-up clinic populations. The goal of early identification of children with CP or high-risk for CP is to allow for early and targeted intervention, thereby maximizing neuroplasticity in the developing brain, and minimizing adverse impact of CP to muscle and bone growth and development. There is evidence that involving and supporting parents as a critical part of this intervention process also leads to improved parent outcomes as well as parent-child interactions. International clinical practice guidelines have recently been published to guide early intervention approaches for CP across the domains of motor function, cognitive skills, communication, eating and drinking, vision, sleep, managing muscle tone, musculoskeletal health, and parental support.

The Gross Motor Function Classification System (GMFCS) provides a valid and reliable system to classify the extent of activity limitation in CP. It is a five-level system (I to V) used to categorize children up to 18 years of age based on their usual performance, with a focus on functional capabilities, including sitting, mobilizing, walking, and need for assistive devices. Such classification helps clinicians communicate information about severity, choice of interventions, and prognosis in a standardized, easy-to-use, valid, and reliable way. Higher level on the GMFCS is associated with increasing functional difficulty, but, of note, distinctions between levels I and II are not as significant as differences between other levels, particularly for younger children. Children classified as level I overall are able to walk without restrictions but may have difficulty with the speed, balance, and coordination required for higher-level skills; between 2 and 4 years, they are able to floor sit with both hands free to manipulate objects, move in and out of floor sitting and standing without adult assistance, and they walk as the preferred method of mobility without the need for any assistive mobility device. In contrast, children classified as level V are profoundly impaired with no means for independent mobility; between 2 and 4 years, this is described as physical impairments that restrict voluntary control of movement and the ability to maintain antigravity head and trunk postures and functional limitations in sitting and standing that are not fully compensated with the use of adaptive equipment. GMFCS categorization in children less than 2 years old depends predominantly on the amount of support required for the child to sit and also considers more advanced skills such as crawling and walking. Although some children diagnosed before 2 years of age will require reclassification, most are not reclassified by more than one level. The positive predictive value (PPV) of a classification of GMFCS level I, II, or III (child will walk with or without aids) as compared with level IV or V (child will probably need a wheelchair for mobility) is very high (0.96). Thus the GMFCS provides a sound approximation, rather than a definitive final categorization, in this age group.

Although CP with severe functional limitations is of great concern, it is relatively rare. An Australian CP registry review from the 1970s to 2004 showed increasing prevalence of CP in the 1970s and 1980s attributed to the increasing survival of EPT infants; however, CP rates stabilized or decreased between the early 1990s and 2004. Analysis of the Danish nationwide CP registry revealed a significant decline in the rate of CP in birth years 2005–2007 explained primarily by fewer cases of severe bilateral spastic CP among term children between these periods, although the rate of CP among premature infants had been shown to dramatically decline in previous periods. Among infants born < 27 weeks’ GA in the National Institute of Child Health and Human Development (NICHD) Neonatal Research Network (NRN) and assessed at 18 to 26 months’ corrected age, there was a decrease in the rate of any CP (16% in 2011 compared with 12% in 2014), with a decrease in severe CP (26% to 16% of all CP cases) during the study period.

There are numerous motor and coordination challenges apart from CP that are reported in high-risk children, including those born preterm. Developmental coordination disorder (DCD) is diagnosed when (1) coordinated motor skills are substantially below those expected given the individual’s chronological age; (2) the motor skills deficit significantly and persistently interferes with the activities of everyday living appropriate to chronological age and impacts school, play, and leisure activities; (3) the motor skills deficits are not better accounted for by any other medical, neurodevelopmental, psychological, social condition, or cultural background; and (4) the onset of symptoms was during childhood. Current recommendations for diagnosis of DCD include a score less than the 16th percentile on the Movement Assessment Battery for Children or equivalent test such as the Bruininks–Oseretsky Test of Motor Proficiency. A recent longitudinal study of children born <30 weeks’ GA demonstrated that at 5 and 7 years of age, nearly 1 in 3 had DCD, with an approximately fourfold increased odds for DCD compared with term-born children. Although the motor difficulties associated with DCD are usually considered “minor” motor impairments, particularly in comparison with disabling CP, nonetheless they can have significant impact on the child. These difficulties may include important functional skills such as fine motor skills, speed and accuracy in motor planning, balance, and coordination. Children with DCD or probable DCD have been shown to be at increased risk for executive function and cognitive challenges, social–emotional and behavior problems, speech and language impairment, and other issues.

Cognitive Assessment

A central component of a high-risk infant neurodevelopmental follow-up visit has been administration of a standardized developmental test. These tests are intended to provide a measure of “cognitive” function, although there are widely acknowledged limitations including the evolution of test versions making it difficult to compare across cohorts, preclusion of extrapolation of 2- to 3-year results to intelligent quotient (IQ) at later time points, and challenges to interpretation of results in the preterm population using standardized “cut points” alone and in the absence of a contemporaneous normal birthweight (NBW) term control group (see later section Limitations and Challenges to Interpreting Early Neurodevelopmental Outcomes Studies).

The Bayley Scales of Infant and Toddler Development (BSID), with a test age range of 1 to 42 months, is now the most widely used developmental test for high-risk infants across the United States and Europe. The original version, released in 1969, was revised in 1993. The BSID-II had two developmental scores: the Mental Developmental Index (MDI), a composite of cognitive and language tasks, and the Psychomotor Developmental Index (PDI), a composite of fine and gross motor skills. This perceived drawback, as well as the usual drive to revise editions due to the “Flynn effect,” contributed to the development of BSID-III, which contains three main domains, (1) a cognitive composite score, (2) a language composite score (with receptive and expressive subscores), and (3) a motor composite score (with gross and fine motor subscores), in addition to social–emotional and adaptive behavior domains. The goal of the BSID-III was to allow identification of delays, as well as relative strengths and challenges, in specific developmental domains, and to target interventions to areas of need. However, in part because of a change in approach to norming the BSID-III, and also possibly because of separation of the cognitive and language scales, cognitive scores on the BSID-III are substantially higher than anticipated among both preterm high-risk children and term control groups. These findings also led to concern that the BSID-III underestimates developmental delay if utilizing normative test means alone, which has serious implications for both clinical and research endeavors. Previously, commonly used “cut points” for categorization of “moderate” and “severe” developmental delay or disability were 2 to 3 standard deviations (SD) and greater than 3 SD below the normative mean, respectively; thus for BSID-II, MDI 55 to 70 was considered a moderate delay whereas an MDI less than 55 was considered a severe delay. However, in the absence of contemporaneous term control groups, commonly used cut points shifted in the era of the BSID-III. Some have recommended that BSID-III cognitive and language scores less than 85 or “combined BSID-III” scores less than 80 provide the best definition of “moderate-to-severe” delay for equivalence with BSID-II MDI less than 70, whereas others have modified the threshold for cognitive delay categorization to define moderate delay as 70 to 84, severe delay 55 to 69, and profound delay as less than 54. The Bayley Scales of Infant and Toddler Development, Fourth Edition (Bayley-4) was published in 2019. The Bayley-4 has more contemporary normative data; notably, this sample does not include the 10% of at-risk children that had been included in the BSID-III normative sample.

The Parent Report of Children’s Abilities-Revised (PARCA-R) is a parent questionnaire assessing a child’s cognitive and language development, but is limited to evaluation at 24 to 27 months of age or corrected for prematurity. The PARCA-R is available for free ( https://www2.le.ac.uk/partnership/parca-r ), in multiple translations, and has concurrent validity with examiner-administered developmental tests, and excellent test-retest reliability. Age- and sex-adjusted standard scores were developed using data from over 6,000 children, with a normative mean of 100 and SD of 15 and allow an assessment of development ranging from < −3SD to > +3 SD. Norms and percentile ranks can therefore now be used to aid in identifying children with either advanced or delayed cognitive and/or language development. The PARCA-R has been used as an outcome measure in observational studies and clinical trials and has been recommended to screen children born preterm in NICE (United Kingdom) developmental follow-up guidelines. Of note, this instrument has also been used widely in telehealth-based assessments.

There are numerous cognitive development tests available to provide full-scale IQ or equivalent that have been used for later childhood assessments, including the Wechsler Preschool and Primary Scale of Intelligence (age range: 2 years 6 months to 7 years 7 months), now in the fourth edition (WPPSI-IV), the Differential Ability Scales (age range: 2 years 6 months to 17 years 11 months), now in the second edition (DAS-II), and the Wechsler Intelligence Scale for Children (age range: 6 years 0 months to 16 years 11 months), now in the fifth edition (WISC-V). The Wechsler Individual Achievement Test (age range: 4 years 0 months to 50 years 11 months), now in the fourth edition (WIAT-4), identifies academic strengths and weaknesses, and has been used in clinical, research, and educational settings.

Hearing and Vision Outcomes

Severe neurosensory impairments, including profound hearing and vision impairment, among preterm infants are now low in incidence but have important long-term consequences. Rates of blindness and significant hearing impairment are inversely related to gestational age. Both moderate-to-severe vision and hearing impairment are more common among high-risk infants and with neonatal morbidities including bronchopulmonary dysplasia (BPD), brain injury, and seizures.

Early detection of hearing impairment is vital for optimizing speech and language development, and guidelines and recommendations reflect the importance of recognizing potential problems as early as possible. As stated in the AAP Joint Committee on Infant Hearing position statement, infants admitted to the NICU for more than 5 days are to have auditory brainstem response included as part of their predischarge screening so that neural hearing loss will not be missed; for those who fail, referral should be made directly to an audiologist for rescreening and, when indicated, comprehensive evaluation for hearing loss. Reevaluation should occur, regardless of initial evaluation results, based on individual risk factors and readmissions. All infants, including well infants, should have a hearing screening by 1 month of age, with rescreening and referral for audiology evaluation by 3 months of age for those who do not pass the initial screen. In addition, a validated global screening tool is administered to all infants at 9, 18, and 24 to 30 months or sooner when there is concern about hearing or language.

Prematurely born children have an increased risk of various ophthalmic and visual dysfunctions and abnormalities, in particular those children with a history of severe or treatment-requiring retinopathy of prematurity (ROP) and those with severe brain injury. These functional visual challenges include strabismus, problems with acuity, convergence and visual fields, and retinal morphology. It is important to recognize that a short-term outcome of severe ROP in the NICU may not result in the most severe functional vision outcomes in early childhood. Even children born preterm without a history of ROP or with only mild ROP also have an increased risk of problems. Among EPT children in the population-based Extremely Preterm Infants in Sweden Study (EXPRESS) at 6.5 years of age, 55% had poor visual-motor integration (VMI) performance. Compared with term-born controls, VMI scores of EPT-born children were lower, and were associated with challenges in everyday school activities. In addition, VMI was most impaired in the children with the lowest. Recommendations from the AAP and the American Association of Pediatric Ophthalmology outlines a detailed construct for the termination of acute retinal examinations based on age and retinal ophthalmoscopic findings; therefore, if hospital discharge is considered before appropriate retinal development, follow-up must be arranged with an ophthalmologist trained in ROP care prior to discharge. Furthermore, regardless of whether infants have required treatment for ROP, ophthalmologic follow-up is indicated within 4 to 6 months after discharge because they are at increased risk for other seemingly unrelated visual disorders, such as strabismus, amblyopia, high refractive errors, cataracts, and glaucoma.

For early neurodevelopmental outcomes studies among preterm infants at 18 to 30 months’ corrected age, criteria for “profound” or “severe” disability in the hearing domain have generally included some definition of “no useful hearing” even with aids or “some hearing but loss not corrected by aids,” whether accompanied by specific decibel hearing loss (dBHL) audiologic evaluation cut points (profound >90 dBHL; severe 70 to 90 dBHL) or by examination and observation of bilateral hearing loss not correctable by amplification. Similarly, “severe” disability in the visual domain has generally been defined as functional bilateral blindness, including examination consistent with presumed visual acuity less than 20/200, or inability to perceive light, or only able to perceive light or reflecting objects. Definitions of “moderate” hearing and vision impairment differ among studies.

Neurodevelopmental Impairment—Difficulties and Realities of a Composite Outcome

Neurodevelopmental impairment is a composite outcome, combining criteria and cut points from several domains as presented above, including neuromotor, cognitive, hearing, and vision. The relative prevalence of each of these outcomes is not consistent, and rates of each outcome may respond differently to an experimental therapy. This necessarily leads to a number of difficulties in overall interpretation, as well as challenges in generalizability and counseling. Use of composite outcomes during counseling in the prenatal or postnatal setting may not be meaningful to parents and families. The value of each of the components may vary broadly for each individual family, outcomes may be conceived differently, and statistics are probably difficult to grasp.

In addition, significant center variation in 18- to 22-month neurodevelopmental outcomes has been demonstrated even after adjustment for demographic variables, prenatal interventions, and neonatal clinical factors, thus presenting challenges to accurate counseling from multicenter datasets. Unfortunately, robust data from a single center on outcomes of specific high-risk groups are generally unavailable. Furthermore, the largest contributing component to the composite outcome is that of “cognitive” delay or impairment. As explored previously and discussed in more detail later in the section School-Age Outcomes After Prematurity, developmental tests at 18 to 30 months are intended to assess cognitive abilities. However, in children born at extremely low birth weight (ELBW) and EPT, scores on these early administered tests predict cognitive scores at school age poorly.

Death or NDI at 2 to 3 years is a primary or main secondary outcome in many clinical trials and prospective studies in neonatal medicine, particularly for EPT infants. This combination is understandable because death and NDI are competing outcomes; it is assumed in this context that adverse outcomes (death and NDI) will not be influenced in opposite directions by an intervention and ideally that components of the combined outcome will carry similar value; these tenets may not always be true. Because the incidence of the composite outcome is greater than any individual component, death or NDI may also be a logical statistical choice for powering a trial; however, it may not be the most biologically plausible target. Moreover, although neurodevelopmental outcome at 2 to 3 years is critically important for any prospective observational study or clinical trial of high-risk infants, it may not be the most appropriate primary outcome for every trial depending on the intervention under evaluation.

Limitations and Challenges to Interpreting Early Neurodevelopmental Outcomes Studies

As described previously, early childhood follow-up visits for high-risk infants at 18 months to 3 years typically assess outcomes in multiple domains. Although these evaluations are important and informative, interpretation and comparison of studies are difficult. Many studies report outcomes by categories and frequently include any adverse finding to yield an “impaired” or “unimpaired” status. The definitions of “adverse” outcomes may not be consistent across studies; indeed, the definitions of the individual components of “impairment” such as CP, blindness, deafness, and developmental delay often differ across studies. A literature review highlighted the impact of varying definitions of NDI, applying definitions from several international neonatal follow-up networks to infants 23 to 28 weeks’ GA in the Canadian Neonatal Network from 2009 to 2011 with follow-up at 18 to 22 months’ corrected age. A fourfold difference in severe NDI incidence was noted, ranging from 3.5% to 14.9%. In addition, differences in risk factors associated with severe NDI were found depending on the definition applied. Furthermore, not all prospective studies have enrolled contemporaneous term, NBW controls. Comparing test scores from a VPT or EPT study group with standardized norms instead of scores from a peer, term-born control group has been demonstrated to substantively limit the relevance and veracity of the findings and may have important public policy and resource implications.

Very early childhood developmental and neurologic outcomes evaluations should only be considered as a first step to comprehensive follow-up; assessments into school age, adolescence, and adulthood are critically important to understand the lifetime trajectory and functional and societal outcomes of former high-risk infants. Concerns regarding changes in cognitive abilities over time, and an increasing recognition that physical and environmental effects as well as early intervention approaches may modify recovery, underscore the need for later assessments. Some neurocognitive, executive function, and behavioral challenges may only be detected at school age; even recognizing that such learning and attention problems may occur in preterm infants is a critical step to ensuring adequate support and services for families and teachers to help children achieve their best possible outcomes. Evaluation of neuromotor outcomes throughout childhood is also critical. Although most toddler-age and very early childhood outcome studies focus narrowly on the diagnosis of CP, later neuromotor and coordination problems, such as DCD, are prevalent among school-age children born EPT compared with term and can be associated with other functional challenges and academic difficulties. Furthermore, parental perceptions of DCD are not reliable, yet interventions may be able to remediate the functional limitations of DCD, reinforcing the importance of ongoing clinical assessments throughout childhood.

Similarly, there are substantial concerns regarding the ability of developmental or cognitive tests at toddler age, particularly the BSID, to detect developmental delay when using standardized test norms. In a group of EPT and ELBW infants at 2 years’ corrected age, Anderson et al. found mean BSID-III cognitive scores of 96.9 and motor scores of 100.4 but also substantially higher than expected scores among a term-born control group. If normative BSID-III cut-point criteria alone were applied to the scores of this cohort, it would severely underestimate moderate-to-severe cognitive and motor delay relative to the control group. Vohr et al. compared mean BSID-II versus BSID-III scores at 18 to 22 months’ corrected age for prematurity among infants less than 27 weeks' EGA born in National Institute of Child Health and Human Development (NICHD) Neonatal Research Network (NRN) hospitals over two adjacent periods and found mean cognitive composite scores on the BSID-III to be 11 points higher than mean BSID-II MDI scores. Even after adjusting for differences between groups, BSID-III was a significant factor in a perceived decrease in neurodevelopmental impairment in the more recent time period. Dilemmas also exist with regard to meaningful utilization of BSID-III motor composite and fine motor scaled score normative data and cut points in the context of other gross motor assessments at 18 to 24 months corrected age and the value of these early evaluations to predict later outcomes and intervene appropriately. In a prospective, longitudinal EPT cohort from the Victoria Infant Collaborative Study Group (VICS Group), the BSID-III motor scale normative cut points for impairment at 2 years seriously underestimated rates of motor impairment at 4 years. In addition, although the BSID-III cognitive and language scales at 2 years were associated with cognitive functioning at 4 years as assessed by DAS-II, developmental delay at 2 years as determined by BSID-III reference data and normative cut points had low sensitivity in predicting future cognitive, verbal, and nonverbal reasoning impairments at 4 years on the DAS-II. All of these findings have important implications for resource availability and public policy. In countries, states, or regions without specific policies advocating for ongoing, longitudinal assessments and services for children born EPT, the 2-year or 3-year evaluation may be the final opportunity to identify challenges before transitioning to the school system. If ongoing needs are determined only by scoring below a normative cut point on the BSID-III, many children at significant risk for future impairments could be left behind.

Despite these many challenges and provisos, a substantial body of literature exists on early neurologic and cognitive outcomes. An understanding of the range of initial neurodevelopmental outcomes, and the factors associated with adverse outcomes, is crucial for both the family and medical care team. In later sections of this chapter, we review early neurodevelopmental outcomes of recent high-risk cohorts, including those born EPT and selected groups of high-risk late preterm and term infants.

The Victoria Infant Collaborative Study Group

The VICS Group has reported on a series of birth cohorts of EPT neonates born in the state of Victoria (Australia) from 1991 to 1992, in 1997 and 2005, and most recently from April 1, 2016, to March 31, 2017. Neonates born alive at 22 to 27 completed gestational weeks and surviving to follow-up received neurodevelopmental assessment at 2 years of corrected age (1991–1992, n = 219, 97% of survivors; 1997, n = 149, 99% of survivors; 2005, n = 163, 95% of survivors; 2016–2017, n = 174, 81% of survivors). Importantly, contemporaneous term controls were also enrolled and evaluated for each cohort. Children were evaluated for blindness, deafness (hearing loss requiring amplification or worse), and developmental delay. To account for the different versions of the Bayley Scales used, developmental delay was classified relative to the MDI or combined cognitive and language composite scores for the contemporaneous term controls, rather than the normative test scores and cut points alone. Developmental delay was defined as less than −1 SD relative to the mean of the controls and categorized as mild (−2 SDs to < −1 SD), moderate (−3 SDs to < −2 SDs), or severe (< −3 SDs). Neurologic examination for CP was also performed, utilizing the GMFCS, describing “severe” CP as unlikely ever to walk and “moderate” CP as unable to walk at 2 years but likely to walk. Overall disability was considered “severe” for children with severe CP, blindness, or severe developmental delay; “moderate” with moderate CP, deafness, or moderate developmental delay; and “mild” with mild CP or mild developmental delay. “Major” disability comprised either the moderate or severe category.

Neurodevelopmental outcomes of the 2016–2017 VICS cohort at 2 years’ corrected age are shown in Table 24.1 . The birth cohorts, rates of any disability, and moderate-severe disability at 2 years were not significantly different across eras. However, survival with no major disability increased over time from 42% in the 1991–1992 cohort to 62% in the 2016–2017 cohort (odds ratio [OR] 1.3 per decade, 95% confidence interval [CI] 1.15 to 1.48, p<0.001). Of note, for the 1991–1992, 1997, and 2005 cohorts, rates of any CP were 11%, 12%, and 10% respectively, whereas for the 2016–2017 cohort the rate of any CP had decreased to 6%. Moderate or severe developmental delay were similar overall across cohorts, with a decrease noted between 1997 and 2005 but an increase in 2016. In terms of 2-year outcomes by gestational ages, mortality and survival with no major disability appeared stable across birth cohorts for those born at 22 and 23 weeks. Mortality improved consistently for those born at 24 weeks from 1991–1992 to 2016–2017. A trend toward improving rates of survival with no major disability was observed across cohorts particularly for those in the more advanced gestational age week groups, with approximately 80% of those born at 26 weeks and 27 weeks in the 2016–2017 cohort having survived with no major disability at 2 years’ corrected age. Although the follow-up rate was lower for the most recent birth cohort compared with previous cohorts, multiple imputation and sensitivity analysis was performed, which resulted in the same results and conclusions.