OBJECTIVE: Previous studies have demonstrated an increased risk for attention-deficit/hyperactivity disorder (ADHD) in follow-up studies of preterm survivors from NICUs. In this study we analyzed the effect of moderate as well as extreme preterm birth on the risk for ADHD in school age, taking into account genetic, perinatal, and socioeconomic confounders.
METHODS: Register study in a Swedish national cohort of 1 180 616 children born between 1987 and 2000, followed up for ADHD medication in 2006 at the age of 6 to 19 years. Logistic regression was used to test hypotheses. A within-mother-between-pregnancy design was used to estimate the importance of genetic confounding in a subpopulation of offspring (N = 34 334) of mothers who had given birth to preterm (≤34 weeks) as well as term infants.
RESULTS: There was a stepwise increase in odds ratios for ADHD medication with increasing degree of immaturity at birth; from 2.1 (1.4–2.7) for 23 to 28 weeks' gestation, to 1.6 (1.4–1.7) for 29 to 32 weeks', 1.4 (1.2–1.7) for 33 to 34 weeks', 1.3 (1.1–1.4) for 35 to 36 weeks', and 1.1 (1.1–1.2) for 37 to 38 weeks' gestation compared with infants born at 39 to 41 weeks' gestation in the fully adjusted model. The odds ratios for the within-mother-between-pregnancy analysis were very similar. Low maternal education increased the effect of moderate, but not extreme, preterm birth on the risk for ADHD.
CONCLUSION: Preterm and early term birth increases the risk of ADHD by degree of immaturity. This main effect is not explained by genetic, perinatal, or socioeconomic confounding, but socioeconomic context modifies the risk of ADHD in moderately preterm births.
WHAT'S KNOWN ON THIS SUBJECT:
Attention problems have been described in survivors of neonatal intensive units in school age.
WHAT THIS STUDY ADDS:
This study demonstrates that not only extremely preterm birth, but also moderately preterm birth, increases the risk of ADHD by degree of immaturity at birth. Social adversity, as expressed by low maternal education, modifies the risk of ADHD in moderately preterm birth.
Increasing numbers of infants born preterm are reaching adulthood as a consequence of advances in perinatal care. A negative effect of preterm birth on the neurologic and psychological development of children and youth is well documented.1,–,3 Severe neurologic impairment has been reported in 10% to 12% of school-aged children born very preterm.4,5 Much higher rates have been reported for more subtle neurocognitive impairment with reports of lower IQ and poor academic achievement.2,3,6,–,9
Attention-deficit/hyperactivity disorder (ADHD) is the most common neurodevelopmental disorder in Western countries,10 with a prevalence of 3% to 5% in Swedish school children.11 Several studies have indicated that attention problems are more common in children born preterm. In previous Swedish studies, Farooqi et al12 found that 11-year-old children born after 23 to 25 gestational weeks had 3 to 4 times as many attention problems compared with term infants, while Stjernqvist and Svenningsen,9 in their study of Swedish 10-year-old children born after <29 weeks, found a more modest twofold increase. A recent French study of 1102 5-year-old children, born after 22 to 32 gestational weeks, demonstrated a twofold risk for hyperactivity/inattention problems compared with term controls.13
Follow-up studies of ADHD in children born preterm have, with very few exceptions, focused on children born extremely preterm and cared for in NICUs.2 However, studies that include the moderately preterm have indicated that this much larger group of infants also is at risk for negative outcomes in school age and young adulthood,14,–,16 including ADHD.3
There is agreement about the importance of genetic factors for ADHD even if specific genes have not been identified.17 A variety of environmental factors are also influential. Exposure, especially fetal exposure, to toxic substances such as lead, alcohol, and PCB have been reported to increase the risk of ADHD (for a review, see Refs 18 and 19). The psychosocial environment, manifested as maltreatment20 or unfavorable socioeconomic conditions,21,22 also is associated with an increased risk of ADHD.
Several perinatal complications have been associated with an increased risk for short and long-term neurologic sequeale. A link between intrauterine growth retardation and ADHD has been suggested, even if results are not conclusive.18,23,24 Perinatal asphyxia may cause hypoxic-ischemic encephalopathy in the neonate with widespread brain injuries as a consequence, and periventricular hemorrhage is a well known risk factor for cerebral palsy in preterm infants.25,–,27 However, these perinatal risk factors have not been investigated much in relation to ADHD.
The interplay between potential etiologic factors is intricate, and confounding may obscure seemingly well established associations. For instance, genetic confounding, was found to explain most of the well established association between fetal exposure to nicotine and ADHD in a recent Swedish study.28 Thus, in studies on preterm birth and ADHD, contributions of other potentially influential factors need to be estimated.2,29 Swedish national registers offers good prerequisites for studies of this kind. They include comprehensive coverage of the whole population with good quality data about perinatal and socioeconomic risk factors for children born at all gestational ages. Large study groups can be created, which makes it possible to control for genetic influence by comparing offspring from 2 pregnancies of the same mother and father, when only 1 child has been exposed to preterm birth.28,30 Thus, the aim of this study was to study the effect of different levels of preterm birth on ADHD, taking genetic, environmental and perinatal risk factors into consideration.
This study was based on Swedish national registers held by the National Board of Health and Welfare and Statistics Sweden. All Swedish residents are assigned a unique 10-digit identification number at birth or immigration. This identification number was used to link information from different register sources.
The study population was created from the 1,242,459 children with a Swedish-born mother in the birth cohorts of 1987 to 2000 who survived infancy, according to the Swedish Medical Birth Register. From this total population, we identified 1,227,096 (98.8%) who were still residents in Sweden on December 31, 2005, according to the Register of the Total Population. We excluded 35 404 children because of at least 1 significant malformation at birth diagnosed by the attending pediatrician and 11 076 because of a registered birth weight above 3 SD or less than −6 SD, according to the growth chart developed by Marsal et al,31 which indicated a probable coding error.32 The remaining 1,180,616 were included in the study population and followed up during 2006 at the age of 6 to 19 years.
Perinatal and Socioeconomic Indicators
Perinatal variables were collected from the Swedish Medical Birth Register, a register that contains information about all births in Sweden (Table 1). Gestational age was categorized according to national Swedish perinatal statistics as extremely preterm (23–28 weeks), very preterm (29–32 weeks), and moderately preterm (33–36 weeks). The latter group was divided in 2 (33–34 and 35–36 weeks) when the numbers allowed us to do this. Term was divided into 2 groups: early term (37–38 weeks) and term (39–41 weeks).
Sociodemographic and parental morbidity variables were obtained through linkage via the Multigeneration Register to the biological mother and father (see Table 1). Information on inpatient care in the Patient Discharge Register was used to create the parental morbidity variables, and information on inpatient and outpatient care from the Patient Discharge Register was used to create the cerebral palsy variable (Table 1). Evaluations of the Swedish Medical Birth Register and the Patient Discharge Register demonstrate coverage of 98% to 99% of all hospital admissions and births in Sweden in public as well as private care. Evaluations of the register have found the quality of the main variables to be generally high.34,–,36
National guidelines for medication of ADHD, issued by the National Board of Health and Welfare in 2002, stated that medication should be reserved for cases in which other supportive interventions have failed, thus reserving medication for the more severe cases. The right to prescribe stimulants for ADHD in Sweden is restricted to specialists with particular familiarity with treatment of this disorder.
The Swedish Prescribed Drug Register contains data with unique patient identifiers for all drugs prescribed and dispensed to the whole population of Sweden (more than 9 million inhabitants) since July 2005.20 In this study we used the purchase of at least 1 prescription of a stimulant (see Table 1 for definition) during the calendar year 2006 as our outcome variable. There were considerable regional differences in the consumption of ADHD medication. Because these differences did not follow any obvious demographical or geographical pattern, we assumed that they mirrored varying prescription patterns in different counties rather than variations in the prevalence of ADHD. The counties were classified in 4 categories according to the proportion of children having purchased ADHD medication during 2006: high prescription rates (>0.8%); high average prescription rates (0.7–0.8%); low average prescription rates (0.5–0.6%); and low prescription rates (<0.5%).
Logistic regression was used to calculate odds ratios (OR) with 95% confidence intervals as estimates of effects, with ADHD medication (see Table 1) as the outcome variable.
We used 3 models to investigate the effects of preterm birth on ADHD medication. Model 1 included gender, a 3-category variable for age (6–9, 10–15, 16–19), and county of residence in 4 categories according to level of ADHD medication. In model 2 we added birth order, maternal age, maternal education, single parenthood, public welfare, maternal smoking, and maternal and paternal psychiatric/addictive disorder (see Table 1 for definitions) as possible confounders. In the final model 3 we added low Apgar score and being small for gestational age as possible perinatal mediators/confounders of preterm birth.
To adjust the analysis for potential influence of genetic confounding, we analyzed the within-subjects variation in the subpopulation (N = 34 334) of offspring of mothers who had given birth to preterm as well as term infants. A generalized linear model with the binomial distribution was used to create a conditional logistic regression in which the effects were compared between pregnancies in the same mother. This within-mother-between-pregnancy model, apart from maternal identification, was adjusted for gender and birth order of each child included, age of the mother at the birth of each child included, and age of each child included in 2006.
Interaction analyses were made in a logistic regression model with adjustment for age, gender, and county only. In this analysis we investigated possible effect modifications of cerebral palsy and preterm birth, small for gestational age and preterm birth, low Apgar score and preterm birth to ADHD as potential pathways for the effects of preterm birth on ADHD. Low education was investigated as a marker of social adversity. In these analyses, a simplified categorization of gestational age was applied ([39–41] vs preterm birth [22–36 weeks] or moderately preterm birth [33–36 weeks])because of the comparatively small numbers.
All statistical analyses were performed using SPSS 18.0 (SPSS Inc, Chicago, IL) for Windows.
In all, 7 506 children in the study population had a record of ADHD medication in the register, corresponding to 1.05% of the boys and 0.29% of the girls. The most commonly purchased drug was methylphenidate (87.8%), followed by atomoxetine (9.2%) and amphetamine (3%).
In Table 2, sociodemographic indicators, parental psychiatric morbidity, and perinatal variables by gestational age and ADHD medication are presented. Such medication was more common in the presence of the following variables: teenage mother; single parent; public welfare; low maternal education; maternal as well as paternal addictive/ psychiatric disorder; low gestational age; small for gestational age; low Apgar score; maternal smoking during pregnancy; and cerebral palsy.
In Table 3, the multivariate analysis of ADHD medication is presented. In the analysis of ADHD medication in model 1, that was adjusted for age, gender, and county of residence, the odds ratios (ORs) for ADHD medication were 2.5 for 23 to 28 weeks' gestation, 1.9 for 29 to 32 weeks, 1.6 for 33 to 34 weeks, 1.4 for 35 to 36 weeks and 1.2 for 37 to 38 weeks' gestation compared with being born at term (39–41 weeks). The ORs decreased slightly to 2.1, 1.6, 1.4, 1.3, and 1.1 in the fully adjusted model. Infants born postterm did not have any increased risk for ADHD.
In Table 4, 2 different regression analyses of 34 344 children who are offspring of mothers who had given birth preterm (≤34 weeks) as well as term (39–41 weeks) within our study population, are presented. The first analysis repeats model 1 in Table 3 with similar results. In a second analysis, we compared different offspring of the same mother in a within-mother-between-pregnancy analysis, with ORs of 2.1 for being born at 23 to 28 weeks' gestation, 1.7 at 29 to 32 weeks and 1.4 at 33 to 34 weeks, compared with term births (39–41 weeks), thus similar to those in the fully adjusted model of Table 3.
In an interaction analysis, the effect of moderate (week 33–36) preterm birth on ADHD medication was higher (P < .01) in mothers with a low education (see Fig 1). The effects of gestational age on ADHD medication were similar in boys and girls.
A low Apgar score had a marginal effect on the risk of ADHD medication and did not modify the effect of preterm birth on ADHD medication. Being small for gestational age in children born term increased the OR for ADHD medication by 1.4 (1.2–1.6) after adjustment for sociodemographic variables. Being small for gestational age did not modify the effect of gestational age on ADHD medication.
Having an indication of cerebral palsy increased the OR of ADHD medication by 2.5 (1.8–3.3) in the whole study population. This effect was lower in all classes of preterm birth compared with term birth (ORs of 0.3–0.4 in interaction analyses, with P < .05 for a dichotomized variable of 22–36 weeks).
In this register study of whole national cohorts of more than 1 million school children, the degree of immaturity at birth, as measured by gestational age, was associated with risk of ADHD medication in school age. This effect was largely independent of socioeconomic and perinatal confounders, and, as confirmed by the within-mother-between-pregnancy analysis, also independent of genetic confounding. However, social adversity, as defined by low maternal education, increased the effect of moderately preterm birth on the risk of ADHD.
The risk for ADHD was doubled for children born at 23 to 28 weeks' gestation in the fully adjusted models, which is on the same level as in the meta-analysis of Bhutta et al.2 The gradual risk increase for ADHD by increasing degree of immaturity, and the lack of a connection with cerebral palsy and growth restriction mentioned above, seems to point to the role of brain development for explaining the main mechanisms linking preterm birth with ADHD. The refinement of cortical connection peaks at gestational weeks 24 to 28 and the number of neurons around week 28.36 There is evidence of a maturational lag in children with ADHD.37 This lag has also been neuroanatomically verified by using neuroimaging techniques38; children with ADHD attained peak cortical thinness 3 years later than healthy controls (10.5 vs 7.5 years). It has been hypothesized that the effects of perinatal factors unfold across development.39 It seems reasonable to hypothesize that such factors are involved in the development of ADHD symptoms in children born preterm. The risk decreased gradually in more mature preterm children but was still 30% higher in children born at 35 to 36 weeks of gestation. This is congruent with a hypothesis of disturbed maturation processes, diminishing over time.
Our results confirm the previous follow-up studies from NICUs that have demonstrated an association between extremely preterm birth and ADHD, using more precise measures of ADHD than in this study.13,40,41 They also support findings from 1 study that indicated a link between moderately preterm birth and behavioral and attention problems.42
Cerebral palsy increased the risk for ADHD medication in term infants, but did not increase the risk for ADHD medication in preterm infants,27 which indicates that the leukomalaciae caused by periventricular hemorrhage, which is so closely associated with the cerebral palsy of children born preterm, is not a common causal mechanism for ADHD associated with preterm birth. It is also noteworthy that growth retardation, as expressed by being small for gestational age, did not modify the risk of preterm birth, which indicates that these 2 perinatal risk factors for ADHD are more or less independent of each other.
The importance of childhood socioeconomic factors in studies on psychological outcomes of children born prematurely has been demonstrated in previous Swedish follow-up studies of preterm infants. In these studies, low socioeconomic status modified the incremental association of cognitive competence with gestational level at all levels of preterm birth,14 and low socioeconomic status increased the risk of psychiatric hospitalization associated with degree of preterm birth.15 In this study, low maternal education is most probably a marker for a complex web of risk factors associated with social adversity, such as relative poverty, living in low status housing areas and family discord.14,15 Thus, it is not surprising that the risk for ADHD caused by moderately preterm birth was modified by the socioeconomic context of the family in which the child had been brought up.
There are limitations to this study. ADHD is a difficult concept to define in a reliable way in epidemiologic studies because of the subjective and context-bound nature of the impairment criteria built into existing diagnostic classifications,43 although successful attempts to operationalize diagnostics have been made.44 We believe, however, that ADHD medication is a quite valid indicator of the more severe cases of ADHD in a Swedish context, considering the national guidelines and the restriction to prescribe stimulants described above.
The Swedish health care system offers child psychiatric and pediatric care free of charge to all Swedish residents. However, there are significant differences in access to ADHD medication between counties in Sweden, as demonstrated in this study. Counties are free to make their own priorities in health care and have different potential in attracting qualified pediatricians and child psychiatrists. However, this variation was adjusted for to a certain extent in the analysis, and moreover it seems quite improbable that such bias would be a major problem in the within-mother-between-pregnancy analysis, which confirmed the association between preterm birth and ADHD in this study. A more important source of bias in this study is probably the special access to care in follow-up programs for the most vulnerable patients from NICUs. It seems likely that this may lead to a better access to qualified care for ADHD-symptoms for the most preterm, but probably not the moderately preterm, compared with the general population.
The crude indicators of parental psychiatric morbidity in this study cannot be expected to be very efficient as indicators of genetic ADHD trait. This was the rationale behind adding a within-mother-between-pregnancy analysis (Table 4). This analysis gave no indication that genetic factors are major confounders in the association between preterm birth and ADHD, although the number of individuals in this analysis was too small to completely exclude this possibility.
In this study it is demonstrated that the association of preterm birth to ADHD is graded by degree of immaturity with significant increased risks also for children born moderately preterm. This study adds to the growing body of evidence that indicates that more attention is needed toward the advancement of care and follow-up for infants born moderately preterm.
This study was funded by the Swedish National Board of Health and Welfare and the Swedish Council for Working Life and Social Research.
- Accepted February 1, 2011.
- Address correspondence to Anders Hjern, MD, PhD, Centre for Health Equity Studies, Karolinska Insitutet/Stockholm University, 106 91 Stockholm, Sweden. E-mail:
Dr Lindström performed analyses and made the first draft of the manuscript; Dr Hjern came up with the idea of the study, designed the study, performed some analyses, and contributed to the writing of the manuscript; Dr Lindblad contributed to the interpretation of the data and the writing of the manuscript; and all authors have seen approved the final version of the manuscript.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- ADHD =
- attention-deficit/hyperactivity disorder •
- OR =
- odds ratio
- Aarnoudse-Moens CS,
- Weisglas-Kuperus N,
- van Goudoever JB,
- Oosterlaan J
- Saigal S,
- den Ouden L,
- Wolke D,
- et al
- Ericson A,
- Kallen B
- Saigal S,
- Hoult LA,
- Streiner DL,
- Stoskopf BL,
- Rosenbaum PL
- Farooqi A,
- Hagglof B,
- Sedin G,
- Gothefors L,
- Serenius F
- Delobel-Ayoub M,
- Arnaud C,
- White-Koning M,
- et al
- Ekeus C,
- Lindstrom K,
- Lindblad F,
- Rasmussen F,
- Hjern A
- Lindstrom K,
- Lindblad F,
- Hjern A
- Lindstrom K,
- Winbladh B,
- Haglund B,
- Hjern A
- Wallis D,
- Russell H,
- Muenke M
- Millichap JG
- O'Keeffe MJ,
- O'Callaghan M,
- Williams GM,
- Najman JM,
- Bor W
- Lindblad F,
- Hjern A
- 34.↵The Swedish Medical Birth Registry. A Summary of Content and Quality. Stockholm, Sweden: National Board of Health and Welfare; 2003. Report No. 2003-112-3
- 35.↵Socialstyrelsen. Kvalitet och innehåll i patientregistret. Utskrivningar från slutenvården 1964–2007 och besök i specialiserad öppenvård (exclusive primärvårdsbesök)1997–2007
- Shaw P,
- Eckstrand K,
- Sharp W,
- et al
- Elgen I,
- Sommerfelt K,
- Markestad T
- van Baar AL,
- Vermaas J,
- Knots E,
- de Kleine MJ,
- Soons P
- Shaffer D,
- Fisher P,
- Lucas CP,
- Dulcan MK,
- Schwab-Stone ME
- Copyright © 2011 by the American Academy of Pediatrics