Do Correlates of Preterm Infants' Attention Problems Have Clinical Utility?
T
he history of attention deficit hyperactivity disorder Many prior studies have cited clinical correlates of ADHD (ADHD), the most prevalent of childhood neurodeveand proposed that such factors may be of diagnostic signifilopmental disorders, is replete with studies searching cance and value. Select examples include the search for minor for cause or causes, as well as for correlates that can inform physical anomalies and evidence of neurologic immaturity diagnosis, intervention, and even prevention.1 As debate as and neuromaturational delay. Frequently, correlational findto the existence of a discrete clinical disorings from a given study are challenged or See related article, p 20 der characterized by inattention, impulrefuted by subsequent research, notwithsivity, and overactivity seems less prominent now than in standing the stability of certain findings related to such biopast decades,2 the relentless pursuit of possible etiologic faclogical measures as heritability, pathophysiology, and tors accelerates, facilitated by such advances as the applicaneuropsychology.7 For example, Waldrop et al noted an intion of neuroimaging techniques (eg, functional magnetic crease in minor physical anomalies (eg, “electric” hair, epiresonance imaging),3 and increases in our understanding of canthal folds, low-set ears, high-arched palate, and clinodactyly) among boys with hyperactive behavior.8 Acsuch fields as molecular genetics4 and epigenetics.5 cardo et al found lower dysmorphology scores among boys In this issue of The Journal, Downey et al examine antewith a diagnosis of hyperactivity or attention deficit disornatal and early postnatal antecedents of parent-reported der.9 Similarly, even though some investigators have sugattention problems in a cohort of 826 preschool children 6 born between 23 and 27 weeks gestation. The investigators gested that findings of neuromaturational delay such as an increase in “neurologic soft signs” (eg, overflow movements) find an intriguing array of antecedents to be associated are useful in suggesting ADHD,10 others note their lack of with parent report of attention problems at 2 years of age such as being born to a woman who had no formal education utility in predicting a diagnosis.11 beyond high school and/or a woman who was exposed to secondhand smoke, recovery of a single organism from the “Seek and Ye Shall Find” placenta, fetal stem vessel thrombosis and recovery of Mycoplasma species, and the recovery of bacteria from a tracheal Spurious correlations are a feature of so-called “noisy” data aspirate. The authors speculate as to the implications for presets that examine a large number of variables.12 This study vention of potentially modifiable antenatal and perinatal facis part of a larger research initiative, the Extremely Low tors. In this editorial, my goal is to explore the implications of Gestational Age Newborn Study, a multisite investigation these findings for general pediatricians engaged in the diagof the causes of medical problems and developmental diffinosis and treatment of children with ADHD. The article raiculties among infants born extremely premature. The broad ses a number of issues that, although suggesting the need for scope of the study includes gathering and analyzing a wide caution in deriving future implications for prevention, offer array of variables reflecting antenatal and early postnatal cirpotential utility in current clinical practice. cumstances and conditions. Hendrickson and Balzer warn of the perils of a “seek and ye shall find” approach when the large number of factors subject to analyses increases the likeCorrelation and Causation lihood of identifying apparent but clinically meaningless associations.13 The current study’s statistical analyses are Statisticians have long acknowledged that correlation does not imply causation and emphasize the need for caution in designed to minimize spurious results. The authors employ assuming that a mathematical relationship between 2 variunivariable analyses to select variables as potential conables implies that one causes the other. The authors of the founders and logistic regression models so that earliest occurcurrent study carefully and appropriately refrain from sugring predictors are not displaced by later occurring gesting that their identified correlates are causes of ADHD. covariates. Evidence in support of the nonspuriousness of For example, in discussing the finding that placental thromfindings includes prior research demonstrating a relationship bosis places a newborn at increased risk of an attention probbetween apparent correlates. The authors cite earlier studies lem and ADHD, they speculate that abnormal placental suggesting associations between neurodevelopmental probimplantation may contribute to or be a surrogate for other lems and such factors as low socioeconomic status, maternal processes that lead to attention problems among extremely preterm infants. The author declares no conflicts of interest.
ADHD
Attention deficit hyperactivity disorder
0022-3476//$ - see front matter. Copyright ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2014.08.052
6
Vol. 166, No. 1 January 2015 antenatal smoking, placental abnormalities, early postnatal systemic inflammation, and antecedents of chronic lung disease. Ultimately, however, no correlational explanation is adequate without the identification of a plausible causal mechanism. I encourage the authors and others to suggest and explore testable hypotheses as to possible causal mechanisms for the array of correlations between prenatal and antenatal antecedents and attention problems.
Inattention and Young Children Notwithstanding variations among children based on their temperament, behavior during infancy and the preschool years is typically noteworthy for children’s short attention span, lack of impulse control, and high activity level. Numerous studies have documented the limited stability and predictive validity of such behaviors as activity level and inattention during the early childhood years. Palfrey et al found that from birth to kindergarten, some 40% of preschool children were noted to have some attentional indicator, but many of the findings were minor and transient.14 Attention span and activity level during the preschool years only weakly predict such behaviors during the school-age years, although preschool children with extreme levels of hyperactivity and inattention are at increased risk for a subsequent diagnosis of ADHD.15 Lahey et al found that ADHD subtypes are not stable over the years, noting that although most young children diagnosed with ADHD persist in fulfilling diagnostic criteria, ratings of hyperactivity, impulsivity, and inattention change over time.16 As the current study’s authors acknowledge, follow-up is necessary to determine the later significance of findings at age 2 years and the stability of parent ratings. Whether children with attention problems at age 2 years as defined by the study’s methodology are at increased risk for a subsequent diagnosis of ADHD is unknown.
Diagnostic Challenges of ADHD Evolving Diagnostic and Statistical Manual criteria reflect the challenges of diagnosing ADHD. The most recent revision, Diagnostic and Statistical Manual-Fifth Edition, persists in requiring documentation of difficulties in multiple settings (eg, school and home) by multiple observers.17 The American Academy of Pediatrics recommends that children be at least 4 years of age to be considered to meet diagnostic criteria.18 Parent reporting of inattention at 24 months of age should not be conflated with the diagnostic criteria for ADHD.
Implications This study’s interesting correlations deserve further consideration. Confirmation of the antecedents of parent-reported attention issues should encourage exploration of possible explanatory mechanisms and inform subsequent research design. Although the diagnostic and prevention implications of findings are speculative at this time, clinicians may, none-
theless, find the results useful. In my discussions with the families whose children I diagnose with ADHD, I am inclined to share, when present, such findings as minor congenital anomalies and evidence of neuromaturational delay. Despite their equivocal correlations with ADHD, I cite such findings as possible evidence of the biological basis for their child’s behavioral difficulties with the hope of alleviating inappropriate parental concern as to their culpability in causing their child’s issues. Similarly, citing prenatal and antenatal antecedents as correlates of attention issues of preterm infants may help to promote greater understanding of the biological basis for ADHD and set the stage for discussion of the complexities of assessment, diagnosis, and multimodal treatment. n Paul H. Dworkin, MD Office for Community Child Health and Help Me Grow National Center Connecticut Children’s Medical Center University of Connecticut School of Medicine Hartford, Connecticut Reprint requests: Paul H. Dworkin, MD, Office for Community Child Health, Connecticut Children’s Medical Center, 282 Washington St, Hartford, CT 06106. E-mail: [email protected]
References 1. Dworkin PH. Learning and behavioral problems of schoolchildren. Philadelphia: WB Saunders; 1985. 2. Schechter NL. The baby and the bathwater: hyperactivity and the medicalization of child rearing. Perspect Biol Med 1982;25:406-16. 3. Paloyelis Y, Mehta MA, Kuntsi J, Asherson P. Functional MRI in ADHD: a systematic literature review. Expert Rev Neurother 2007;7: 1337-56. 4. Faraone SV, Perlis RH, Doyle AE, Smoller JW, Goralnick JJ, Holmgren MA, et al. Molecular genetics of attention-deficit/ hyperactivity disorder. Biol Psychiatry 2005;57:1313-23. 5. Archer T, Berman MO, Blum K. Epigenetics in developmental disorder: ADHD and endophenotypes. J Genet Syndr Gene Ther 2011;2:104. 6. Downey LC, O’Shea TM, Allred EN, Kuban K, McElrath TF, Warner DD, et al. Antenatal and early postnatal antecedents of parentreported attention problems at 2 years of age. J Pediatr 2015;166:20-5. 7. Coghill D. The value and limitations of the concepts of ADHD and hyperkinetic disorder in guiding treatment - A clinician’s perspective. NICE Clinical Guidelines, No. 72. National Collaborating Centre for Mental Health (UK). Leicester (UK): British Psychological Society; 2009, http://www.ncbi.nlm.nih.gov/books/NBK53648/. Accessed August 16, 2014. 8. Waldrop MF, Pedersen FA, Bell RQ. Minor physical anomalies and behavior in preschool children. Child Dev 1968;39:391-400. 9. Accardo PJ, Tomazic T, Morrow J, Haake C, Whitman BY. Minor malformations, hyperactivity, and learning disabilities. Am J Dis Child 1991; 145:1184-7. 10. MacNeil LK, Xavier P, Garvey MA, Gilbert DL, Ranta ME, Denckla MB, et al. Quantifying excessive mirror overflow in children with attentiondeficit/hyperactivity disorder. Neurology 2011;76:622-8. 11. Fellick JM, Thomson APJ, Sills J, Hart CA. Neurological soft signs in mainstream pupils. Arch Dis Child 2001;85:371-4. 12. Aldrich J. Correlations genuine and spurious in Pearson and Yule. Stat Sci 1995;10:364-76. 13. Hendrickson M, Balzer B. Evaluating oncopathological studies: the need to evaluate the internal and external validity of study results. In: Marchevsky AM, Wick M, eds. Evidence based pathology and laboratory medicine. New York: Springer; 2011. 7
THE JOURNAL OF PEDIATRICS
www.jpeds.com
14. Palfrey JS, Levine MD, Walker DK, Sullivan M. The emergence of attention deficits in early childhood: a prospective study. J Dev Behav Pediatr 1985;6:339-48. 15. McGee R, Partridge F, Williams S, Silva A. A twelve-year follow-up of preschool hyperactive children. J Am Acad Child Adolesc Psychiatry 1991;30:224-32. 16. Lahey BB, Pelham WE, Loney J, Lee SS, Willcutt E. Instability of the DSM-IV subtypes of ADHD from preschool through elementary school. Arch Gen Psychiatry 2005;62:896-902.
Vol. 166, No. 1 17. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 5th ed. Arlington, VA: American Psychiatric Association; 2013. 18. American Academy of Pediatrics Subcommittee on Attention-Deficit/ Hyperactivity Disorder Steering Committee on Quality Improvement and Management. Implementing the key action statements: an algorithm and explanation for process of care for the evaluation, diagnosis, treatment, and monitoring of ADHD in children and adolescents. Pediatrics 2011;128(Suppl 1):SI1-21.
Pitfalls in Food Allergy Diagnosis: Serum IgE Testing
I
and 100% of foods removed based on serum IgE testing n this issue of The Journal, Bird et al report the results of an were able to be returned to the diet after OFCs.2 interesting retrospective chart review of 797 new patients referred to the Food Allergy Center at Southwestern MediThe misdiagnosis of food allergy due to misinterpretation cal Center between September 2011 and December 2012.1 Of of serum IgE testing, though, is not unique to the population with atopic dermatitis. Misdiagnosis is partially due to the these, 284 (35%) had serum IgE testing performed to foods; low positive predictive values of serum IgE 10 patients were excluded due to eosinoSee related article, p 97 immunoassays, which are considered to be philic esophagitis, leaving 274 for further less than the predictive values of skin prick tests, for which analysis. Only 90 of the 274 (32.8%) had a clinical history they are
T
he history of attention deficit hyperactivity disorder Many prior studies have cited clinical correlates of ADHD (ADHD), the most prevalent of childhood neurodeveand proposed that such factors may be of diagnostic signifilopmental disorders, is replete with studies searching cance and value. Select examples include the search for minor for cause or causes, as well as for correlates that can inform physical anomalies and evidence of neurologic immaturity diagnosis, intervention, and even prevention.1 As debate as and neuromaturational delay. Frequently, correlational findto the existence of a discrete clinical disorings from a given study are challenged or See related article, p 20 der characterized by inattention, impulrefuted by subsequent research, notwithsivity, and overactivity seems less prominent now than in standing the stability of certain findings related to such biopast decades,2 the relentless pursuit of possible etiologic faclogical measures as heritability, pathophysiology, and tors accelerates, facilitated by such advances as the applicaneuropsychology.7 For example, Waldrop et al noted an intion of neuroimaging techniques (eg, functional magnetic crease in minor physical anomalies (eg, “electric” hair, epiresonance imaging),3 and increases in our understanding of canthal folds, low-set ears, high-arched palate, and clinodactyly) among boys with hyperactive behavior.8 Acsuch fields as molecular genetics4 and epigenetics.5 cardo et al found lower dysmorphology scores among boys In this issue of The Journal, Downey et al examine antewith a diagnosis of hyperactivity or attention deficit disornatal and early postnatal antecedents of parent-reported der.9 Similarly, even though some investigators have sugattention problems in a cohort of 826 preschool children 6 born between 23 and 27 weeks gestation. The investigators gested that findings of neuromaturational delay such as an increase in “neurologic soft signs” (eg, overflow movements) find an intriguing array of antecedents to be associated are useful in suggesting ADHD,10 others note their lack of with parent report of attention problems at 2 years of age such as being born to a woman who had no formal education utility in predicting a diagnosis.11 beyond high school and/or a woman who was exposed to secondhand smoke, recovery of a single organism from the “Seek and Ye Shall Find” placenta, fetal stem vessel thrombosis and recovery of Mycoplasma species, and the recovery of bacteria from a tracheal Spurious correlations are a feature of so-called “noisy” data aspirate. The authors speculate as to the implications for presets that examine a large number of variables.12 This study vention of potentially modifiable antenatal and perinatal facis part of a larger research initiative, the Extremely Low tors. In this editorial, my goal is to explore the implications of Gestational Age Newborn Study, a multisite investigation these findings for general pediatricians engaged in the diagof the causes of medical problems and developmental diffinosis and treatment of children with ADHD. The article raiculties among infants born extremely premature. The broad ses a number of issues that, although suggesting the need for scope of the study includes gathering and analyzing a wide caution in deriving future implications for prevention, offer array of variables reflecting antenatal and early postnatal cirpotential utility in current clinical practice. cumstances and conditions. Hendrickson and Balzer warn of the perils of a “seek and ye shall find” approach when the large number of factors subject to analyses increases the likeCorrelation and Causation lihood of identifying apparent but clinically meaningless associations.13 The current study’s statistical analyses are Statisticians have long acknowledged that correlation does not imply causation and emphasize the need for caution in designed to minimize spurious results. The authors employ assuming that a mathematical relationship between 2 variunivariable analyses to select variables as potential conables implies that one causes the other. The authors of the founders and logistic regression models so that earliest occurcurrent study carefully and appropriately refrain from sugring predictors are not displaced by later occurring gesting that their identified correlates are causes of ADHD. covariates. Evidence in support of the nonspuriousness of For example, in discussing the finding that placental thromfindings includes prior research demonstrating a relationship bosis places a newborn at increased risk of an attention probbetween apparent correlates. The authors cite earlier studies lem and ADHD, they speculate that abnormal placental suggesting associations between neurodevelopmental probimplantation may contribute to or be a surrogate for other lems and such factors as low socioeconomic status, maternal processes that lead to attention problems among extremely preterm infants. The author declares no conflicts of interest.
ADHD
Attention deficit hyperactivity disorder
0022-3476//$ - see front matter. Copyright ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2014.08.052
6
Vol. 166, No. 1 January 2015 antenatal smoking, placental abnormalities, early postnatal systemic inflammation, and antecedents of chronic lung disease. Ultimately, however, no correlational explanation is adequate without the identification of a plausible causal mechanism. I encourage the authors and others to suggest and explore testable hypotheses as to possible causal mechanisms for the array of correlations between prenatal and antenatal antecedents and attention problems.
Inattention and Young Children Notwithstanding variations among children based on their temperament, behavior during infancy and the preschool years is typically noteworthy for children’s short attention span, lack of impulse control, and high activity level. Numerous studies have documented the limited stability and predictive validity of such behaviors as activity level and inattention during the early childhood years. Palfrey et al found that from birth to kindergarten, some 40% of preschool children were noted to have some attentional indicator, but many of the findings were minor and transient.14 Attention span and activity level during the preschool years only weakly predict such behaviors during the school-age years, although preschool children with extreme levels of hyperactivity and inattention are at increased risk for a subsequent diagnosis of ADHD.15 Lahey et al found that ADHD subtypes are not stable over the years, noting that although most young children diagnosed with ADHD persist in fulfilling diagnostic criteria, ratings of hyperactivity, impulsivity, and inattention change over time.16 As the current study’s authors acknowledge, follow-up is necessary to determine the later significance of findings at age 2 years and the stability of parent ratings. Whether children with attention problems at age 2 years as defined by the study’s methodology are at increased risk for a subsequent diagnosis of ADHD is unknown.
Diagnostic Challenges of ADHD Evolving Diagnostic and Statistical Manual criteria reflect the challenges of diagnosing ADHD. The most recent revision, Diagnostic and Statistical Manual-Fifth Edition, persists in requiring documentation of difficulties in multiple settings (eg, school and home) by multiple observers.17 The American Academy of Pediatrics recommends that children be at least 4 years of age to be considered to meet diagnostic criteria.18 Parent reporting of inattention at 24 months of age should not be conflated with the diagnostic criteria for ADHD.
Implications This study’s interesting correlations deserve further consideration. Confirmation of the antecedents of parent-reported attention issues should encourage exploration of possible explanatory mechanisms and inform subsequent research design. Although the diagnostic and prevention implications of findings are speculative at this time, clinicians may, none-
theless, find the results useful. In my discussions with the families whose children I diagnose with ADHD, I am inclined to share, when present, such findings as minor congenital anomalies and evidence of neuromaturational delay. Despite their equivocal correlations with ADHD, I cite such findings as possible evidence of the biological basis for their child’s behavioral difficulties with the hope of alleviating inappropriate parental concern as to their culpability in causing their child’s issues. Similarly, citing prenatal and antenatal antecedents as correlates of attention issues of preterm infants may help to promote greater understanding of the biological basis for ADHD and set the stage for discussion of the complexities of assessment, diagnosis, and multimodal treatment. n Paul H. Dworkin, MD Office for Community Child Health and Help Me Grow National Center Connecticut Children’s Medical Center University of Connecticut School of Medicine Hartford, Connecticut Reprint requests: Paul H. Dworkin, MD, Office for Community Child Health, Connecticut Children’s Medical Center, 282 Washington St, Hartford, CT 06106. E-mail: [email protected]
References 1. Dworkin PH. Learning and behavioral problems of schoolchildren. Philadelphia: WB Saunders; 1985. 2. Schechter NL. The baby and the bathwater: hyperactivity and the medicalization of child rearing. Perspect Biol Med 1982;25:406-16. 3. Paloyelis Y, Mehta MA, Kuntsi J, Asherson P. Functional MRI in ADHD: a systematic literature review. Expert Rev Neurother 2007;7: 1337-56. 4. Faraone SV, Perlis RH, Doyle AE, Smoller JW, Goralnick JJ, Holmgren MA, et al. Molecular genetics of attention-deficit/ hyperactivity disorder. Biol Psychiatry 2005;57:1313-23. 5. Archer T, Berman MO, Blum K. Epigenetics in developmental disorder: ADHD and endophenotypes. J Genet Syndr Gene Ther 2011;2:104. 6. Downey LC, O’Shea TM, Allred EN, Kuban K, McElrath TF, Warner DD, et al. Antenatal and early postnatal antecedents of parentreported attention problems at 2 years of age. J Pediatr 2015;166:20-5. 7. Coghill D. The value and limitations of the concepts of ADHD and hyperkinetic disorder in guiding treatment - A clinician’s perspective. NICE Clinical Guidelines, No. 72. National Collaborating Centre for Mental Health (UK). Leicester (UK): British Psychological Society; 2009, http://www.ncbi.nlm.nih.gov/books/NBK53648/. Accessed August 16, 2014. 8. Waldrop MF, Pedersen FA, Bell RQ. Minor physical anomalies and behavior in preschool children. Child Dev 1968;39:391-400. 9. Accardo PJ, Tomazic T, Morrow J, Haake C, Whitman BY. Minor malformations, hyperactivity, and learning disabilities. Am J Dis Child 1991; 145:1184-7. 10. MacNeil LK, Xavier P, Garvey MA, Gilbert DL, Ranta ME, Denckla MB, et al. Quantifying excessive mirror overflow in children with attentiondeficit/hyperactivity disorder. Neurology 2011;76:622-8. 11. Fellick JM, Thomson APJ, Sills J, Hart CA. Neurological soft signs in mainstream pupils. Arch Dis Child 2001;85:371-4. 12. Aldrich J. Correlations genuine and spurious in Pearson and Yule. Stat Sci 1995;10:364-76. 13. Hendrickson M, Balzer B. Evaluating oncopathological studies: the need to evaluate the internal and external validity of study results. In: Marchevsky AM, Wick M, eds. Evidence based pathology and laboratory medicine. New York: Springer; 2011. 7
THE JOURNAL OF PEDIATRICS
www.jpeds.com
14. Palfrey JS, Levine MD, Walker DK, Sullivan M. The emergence of attention deficits in early childhood: a prospective study. J Dev Behav Pediatr 1985;6:339-48. 15. McGee R, Partridge F, Williams S, Silva A. A twelve-year follow-up of preschool hyperactive children. J Am Acad Child Adolesc Psychiatry 1991;30:224-32. 16. Lahey BB, Pelham WE, Loney J, Lee SS, Willcutt E. Instability of the DSM-IV subtypes of ADHD from preschool through elementary school. Arch Gen Psychiatry 2005;62:896-902.
Vol. 166, No. 1 17. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 5th ed. Arlington, VA: American Psychiatric Association; 2013. 18. American Academy of Pediatrics Subcommittee on Attention-Deficit/ Hyperactivity Disorder Steering Committee on Quality Improvement and Management. Implementing the key action statements: an algorithm and explanation for process of care for the evaluation, diagnosis, treatment, and monitoring of ADHD in children and adolescents. Pediatrics 2011;128(Suppl 1):SI1-21.
Pitfalls in Food Allergy Diagnosis: Serum IgE Testing
I
and 100% of foods removed based on serum IgE testing n this issue of The Journal, Bird et al report the results of an were able to be returned to the diet after OFCs.2 interesting retrospective chart review of 797 new patients referred to the Food Allergy Center at Southwestern MediThe misdiagnosis of food allergy due to misinterpretation cal Center between September 2011 and December 2012.1 Of of serum IgE testing, though, is not unique to the population with atopic dermatitis. Misdiagnosis is partially due to the these, 284 (35%) had serum IgE testing performed to foods; low positive predictive values of serum IgE 10 patients were excluded due to eosinoSee related article, p 97 immunoassays, which are considered to be philic esophagitis, leaving 274 for further less than the predictive values of skin prick tests, for which analysis. Only 90 of the 274 (32.8%) had a clinical history they are
