Tinka Broring

We suggest that underresponsive and overresponsive behavior in very preterm children is (mis)labeled as symptoms of ADHD and/or ASD. Possibly the absent-minded ADHD-like and the aloof ASD-like behavior is primarily a reflection of underresponsiveness. Likewise, the sensory seeking behavior such as repetitive play and touching materials may be interpreted as restrictive and strange behavior in ASD and the fidgety or on the go behavior to seek sensory stimuli may be interpreted as hyperactive and distractible behavior consistent with ADHD.33 Conversely, fearful and cautious overresponsive reactions to sensory stimuli may be interpreted as ASD-like behaviors and negative and defiant overresponsive reactions as ADHD-like behavior.33 Understanding behavioral difficulties as part of the “preterm behavioral phenotype” from a sensory processing perspective is further supported by evidence in our studies on associations between ADHD and compromised somatosensory registration and tactile overresponsiveness and by extensive evidence on sensory processing difficulties from other studies in both ADHD and ASD.33–40

With respect to the different domains of sensory processing (registration, integration and modulation), it is the domain of modulation in particular, that is pertinent in the relation between very preterm birth and symptom levels of ADHD and ASD. Although the very preterm group showed impaired somatosensory registration, these difficulties did not relate to ADHD and ASD symptom levels. A possible explanation for this finding is that it is not so much the lower-order registration of somatosensory stimuli, but rather the higher-order modulation of responses to sensory stimuli, that is important in understanding symptoms of ADHD and ASD and subsequent adaptation to the environment. Brain circuits involved in higher-order modulation are far more complex, using extensive brain networks, both top-down and bottom-up, and are therefore more vulnerable to subtle white matter damage than lower-order registration circuits.12,41 This is bolstered by a claim of Wallace and Stevenson, that low-level sensory functioning is unaffected in children with autism, since local cortical organization is preserved, while more extensive brain networks are impaired.39

In contrast to our findings that girls with ADHD showed higher levels of tactile overresponsiveness, we found no significant sex differences in sensory modulation in very preterm children. Although in our sample of very preterm children, sex was not a relevant predictor for sensory modulation difficulties, further studies are needed to assess these sex differences, as it is known that male sex is a risk factor for white matter injury42 as well as more severe neurodevelopmental sequelae43,44 in very preterm children.

According to the multiple-hit hypothesis,45 being born very preterm in conjunction with neonatal complications may particularly compromise normal brain development and may enhance the risk for white matter brain abnormalities with subsequent neurodevelopmental problems,42 such as the sensory processing difficulties and ADHD and ASD symptoms found in our studies.

We argue that the white matter abnormalities in very preterm children11,46,47 show similarities to those found in children with a primary sensory processing disorder,12 and that the impact of the sensory challenging NICU stay48 is evident on both sensory processing and behavioral difficulties. In our review, we found some evidence that neonatal complications, including white (and grey) matter abnormalities and length of NICU stay, showed a dose-response relationship with sensory processing difficulties. This is supported by our finding that the presence of neonatal complications (i.e. being born small for gestational age, PVL, infections) is pertinent in the relation between very preterm birth, symptom levels of ADHD and ASD and sensory modulation, as only in very preterm children who had at least one additional neonatal complication, sensory modulation difficulties impacted on ADHD and ASD symptom levels. In addition, the finding that extended hospital stay is important in the relation between very preterm birth, symptom levels of ADHD and ASD and sensory modulation, not only acknowledges that longer hospital stay is associated with greater exposure to detrimental neonatal complications, but may also fit the idea that the sensory challenging NICU itself, largely independent of other neonatal complications, may negatively affect long term neurodevelopmental outcome in very preterm children.49

Strengths and limitations

The studies presented in this thesis have both strengths and limitations. The very preterm children included in our studies are part of a large and representative sample of Dutch children born before 32 weeks of gestation (Study Towards the Effects of Postdischarge nutrition on growth and body composition of infants born ≤ 32 weeks of gestation and/or ≤ 1500 gram birth weight [STEP study]). Where the incidence of IVH, PVL and infections is relatively low in this group, other baseline characteristics including IUGR, SGA, BPD, length of NICU stay and SES are representative for the very preterm population. Moreover, recruitment of an equally sized full-term born control group, matched on sex, age and parental education, allowed meaningful comparison between very preterm children and full-term born children on behavioral and sensory processing measures. Another strength of our studies is that sensory processing was evaluated at multiple levels. Sensory processing included registration, integration and modulation, which allowed us to thoroughly differentiate between these three levels. Additionally, behavior was evaluated extensively, with multiple informants (parents and teachers) reporting on ADHD and ASD symptoms. Moreover, we used multiple questionnaires, a screening instrument and an interview, tapping into the same domain (ADHD and ASD). Finally, our studies in a group of children with a diagnosis of ADHD allowed us to provide a strong rationale for our finding that symptoms of ADHD are mediated by sensory modulation difficulties in very preterm children.

The presented studies also have some limitations. Since all very preterm children initially participated in a RCT on a postdischarge feeding intervention, it is possible that this has interfered with our results, as optimal feeding strategies aim to reduce long term growth deficits and risks for adverse developmental consequences later in life.50 We analyzed the potential intervention effects by analyses of variance and found no meaningful effects of the intervention on any of our measures, except for a small but beneficial effect of postdischarge formula over standard term formula on sensory modulation. However, the observed differences between very preterm and full-term children on sensory modulation persisted, despite positive effects in the group of very preterm children receiving enhanced postdischarge formula. This suggests a robust difference between the very preterm and full-term group. Another concern is that our empirical studies on very preterm children are all performed in the STEP cohort, thereby capitalizing on the same group of children, so findings might be related to the idiosyncratic characteristics of our sample. For example, our STEP sample showed a relatively low prevalence of risk factors for developing white matter abnormalities, including PVL, subependymal hemorrhage and infections. Furthermore, our relatively small sample size of very preterm children prevented us from robustly studying the relation between neonatal risk factors and sensory processing difficulties and behavioral problems. Another limitation, unfortunately very common in follow-up studies in very preterm children, is the substantial attrition, with only half of the initial cohort willing to participate in this follow-up study. Of the 152 infants included in the original RCT, 112 children were still available for follow-up at 8–10 years of age, of which 57 (51%) agreed to participate in the current studies. However, no differences were found between the groups of participants and non-participants on sex, parental education, gestational age (GA), birth weight, PVL, and the presence of perinatal infections. Furthermore, regarding our choice of measures we consider the use of a screening instrument (SCQ) instead of a diagnostic interview on ASD as an inconsistency and shortcoming, since we did include a diagnostic interview on ADHD. However, the SCQ is a well validated alternative for the gold standard, but time-consuming, Autism Diagnostic Interview-Revised, and is widely used in studies on very preterm birth.51–54 In addition, we have measured the domains of sensory processing with very different measures, in terms of parent report (modulation) versus child-administered tasks (registration and integration) and clinical measures (registration) versus a computerized task (integration). Yet, outside clinical diagnostic procedures on sensory modulation and a validated test for the infant age only (TSFI), no child-administered test is available for the domain of sensory modulation. Moreover, the three domains are very different in the demands they place on a child. Therefore some variety in measures will be inevitable.

With respect to the studies in the ADHD sample, larger sample sizes, especially for the group of girls with an ADHD diagnosis, would have benefited the statistical power of the studies to detect alterations in somatosensory functioning, tactile overresponsiveness and pain experience. Another limitation of this ADHD sample is that some levels of sensory processing were not studied as thoroughly, with sensory integration missing in the measures and sensory modulation only represented by tactile overresponsiveness.

Clinical implications

Our findings show that very preterm children are at risk for developing sensory processing difficulties and elevated symptom levels of ADHD and ASD. Moreover, we suggest that symptoms of ADHD and ASD, at least partly, originate from sensory modulation difficulties. In terms of long-term consequences of very preterm birth, sensory processing and behavioral difficulties are deemed as minor impairments. Yet, the impact of these minor impairments on adaptive functioning and quality of life in very preterm children may be substantial.55,56 For instance, sensory processing and behavioral difficulties have been described to hamper normal development by interfering with social activities, play and leisure.25,56–58 Although our studies found low to moderate effect sizes and the findings of the studies in this thesis await more research and replication, we believe that interpreting behavioral difficulties from a sensory processing perspective is useful in the follow-up care for very preterm infants.

In the last three decades the NICU stay of a very preterm infant has changed tremendously, starting with developmental care interventions from NIDCAP (Newborn Individualized Development Care and Assessment Program),59,60 progressing to kangaroo care and (multi) sensory stimulation61 and more recently to strong parent involvement in family integrated care for very preterm infants with promising results.62 As the sensory system is powerfully shaped by the number and types of sensory experiences directly after birth,63,64 interventions during NICU and High Care stay are crucial. Both developmental care interventions and family integrated care may mitigate sensory overstimulation and understimulation.61,65–68 The use of proven effective analgesia diminishes procedural pain (overstimulation), especially when combined with parental holding of the infant or, if possible, with breastfeeding.69 Kangaroo care, preventing tactile understimulation, has additional positive effects on both the infant and the parent, including better growth of the infant, decreased stress in mothers, and better mother-infant interaction.61,67 Fine-tuned sensory stimulation, for instance by intensive parent involvement in the care for their preterm infant, may break down tactile (i.e. holding) and vestibular (i.e. handling/rocking) understimulation and normalize auditory stimulation (i.e. voices) and has shown positive effects on both infant (weight gain) and parent (stress levels).68

Yet, the sensory system continues to be shaped throughout the course of life.35 Therefore, signaling sensory processing difficulties may be advisable across the full childhood age range in very preterm children. In the Dutch follow-up care for children born very preterm, screening for sensory processing difficulties is not yet standard. The results of this thesis prompt the consideration of screening for sensory processing difficulties, at least at the level of modulation (overresponsiveness and underresponsiveness), and more in-depth screening of ADHD and ASD symptoms, in particular assessment of attention problems, social impairment and communication problems (rather than screening for a diagnosis of ADHD and/or ASD) in the follow-up care for very preterm children. Children with the “preterm behavioral phenotype” may not show difficulties on all symptom dimensions and therefore may fail to meet criteria for a full diagnosis of ADHD or ASD, yet the impact on daily functioning, especially in conjunction with sensory modulation problems, may be distinct.15,70 Moreover, even if children qualify for an ADHD or ASD diagnosis, this diagnosis alone may not fully capture the whole clinical presentation nor lead to a tailored treatment indication.71

Understanding ADHD and ASD symptoms from a sensory processing perspective may provide additional leads for intervention and treatment in very preterm children.35,72 Tailored interventions including counseling of parents and teachers by child psychologists on the expression of sensory processing and behavioral difficulties in the home and school environment and/or referral of the very preterm child to occupational therapy, may be pivotal to downsize behavioral difficulties in very preterm children.33,35,57,73 Although counseling of parents is not extensively studied in the context of sensory processing difficulties, it is regarded as good clinical practice. Moreover, our experience in clinical practice suggests that if parents better understand the origin of the behavioral difficulties of their child, they will be more flexible and understanding in their parenting style; preventing conflicts, stress and miscommunication. The understanding of underresponsive and overresponsive behavioral patterns, as well as specific sensory processing problems across sensory modalities offers opportunities for interventions to lessen the impact on both the school and home environment. These interventions may include, but not be limited to, different seating arrangements in the classroom, selective use of headphones with or without music, addition of activating tactile materials during listening, shorter periods of working on one task, increase or decrease of distraction in a child’s room (visual, auditory), specific use of materials and fitting for a child’s clothes, and selective addition of types of new food.33,74 Occupational therapy is an intervention that aims to improve the child’s sensory responsivity across sensory modalities, social behavior, motor competence, and participation in daily life by stimulating the child to interact with sensory materials in an active, meaningful, and joyful manner in close collaboration with parents. Effectiveness of occupational therapy, although far from rigorously studied, has shown positive results in clinical practice and has also recently been proven effective in a small randomized controlled trial in children with sensory processing difficulties.73

Future research

Future research on sensory processing in preterm children is needed to replicate and extend the available results of this thesis. On both the registration and integration level, more studies are needed on the different sensory modalities. On the modulation level research should also include questionnaires on behavioral problems. Measures should best be a combination of child-administered tests and questionnaires. Parents should be considered as the primary informant to judge sensory processing in daily life of a very preterm child, and at the age of four also school teachers may be included as informants. Gathering information on self-report would be helpful in children above eight years of age. With respect to child-administered tests, both clinical and computerized measures may be used, complemented by quantitative sensory testing with brain evoked potentials, for instance to further explore sensory detection thresholds. Future studies would preferably be term-born controlled longitudinal studies combining sensory processing measures with behavioral measures tapping into ADHD and ASD to reveal crucial underpinnings for the “preterm behavioral phenotype”. Imaging studies, including diffusion tensor imaging (DTI), are recommended to better understand the underlying brain abnormalities of sensory processing difficulties, linking white matter integrity and connectivity to the different levels of sensory registration, integration and modulation in very preterm children. Additionally, impact of NICU stay may be taken into account more thoroughly within prospective research designs, mapping for instance number of invasive and/or skin-breaking procedures, hours of kangaroo care, and neonatal pain measurements during hospitalization. Finally, although tailored interventions, such as parental counseling and referral to occupational therapy, are considered good clinical care, effectiveness of these interventions should be carefully evaluated. Scarcity of empirical research on therapeutic approaches targeting the sensory system, but also on parental sensitivity and parenting skills, illustrates the importance of collaborative, translational research. In the near future, we hope to evaluate effectiveness of parental counseling on sensory processing difficulties and regulatory problems in young infants by joining forces with clinicians (medical psychologists) and embedded scientists in the area of child development.

Concluding remarks

The results of this thesis confirm the presence of sensory processing difficulties in the domains of registration and modulation, as well as behavioral difficulties in terms of elevated symptom levels of ADHD and ASD symptoms in very preterm children. Moreover, sensory modulation in particular may be related to symptoms of ADHD and ASD in very preterm children and might be considered as one of the pathways that lead to adverse behavioral outcomes observed in very preterm children. Understanding ADHD and ASD symptoms from a sensory processing perspective may provide additional leads for intervention and treatment in very preterm children. Screening for sensory processing difficulties and symptoms of ADHD and ASD should therefore be considered to be included in the follow-up care in very preterm children.