SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Aim  To examine the incidence of paroxysmal epileptic and non-epileptic disorders and the associated prenatal and perinatal factors that might predict them in the first year of life in a population-based cohort.

Method  This study was embedded in the Generation R Study, a population-based prospective cohort study from early fetal life onwards. Information about the occurrence of paroxysmal events, defined as suddenly occurring episodes with an altered consciousness, altered behaviour, involuntary movements, altered muscle tone, and/or a changed breathing pattern, was collected by questionnaires at the ages of 2, 6, and 12 months. Information on possible prenatal and perinatal determinants was obtained by measurements and questionnaires during pregnancy and after birth.

Results  Information about paroxysmal events in the first year of life was available in 2860 participants (1410 males, 1450 females). We found an incidence of paroxysmal disorders of 8.9% (n=255) in the first year of life. Of these participants, 17 were diagnosed with febrile seizures and two with epilepsy. Non-epileptic events included physiological events, apnoeic spells, loss of consciousness by causes other than epileptic seizures or apnoeic spells, parasomnias, and other events. Preterm birth (p<0.001) and low Apgar score at 1 minute (p<0.05) were significantly associated with paroxysmal disorders in the first year of life. Continued maternal smoking during pregnancy and preterm birth were significantly associated with febrile seizures in the first year of life (p<0.05).

Interpretation  Paroxysmal disorders are frequent in infancy. They are associated with preterm birth and a low Apgar score. Epileptic seizures only form a minority of the paroxysmal events in infancy. In this study, children whose mothers continued smoking during pregnancy had a higher reported incidence of febrile seizures in the first year of life. These findings may generate various hypotheses for further investigations.


What this paper adds

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • • 
    Paroxysmal disorders are frequent in infancy.
  • • 
    They are associated with preterm birth and a low Apgar score.
  • • 
    Epilepsy and seizures only form a minority of these events.
  • • 
    Children who were born preterm or whose mothers continued smoking during pregnancy had a higher reported incidence of febrile seizures in the first year of life.

Seizures and epilepsy are a common problem in childhood.1 There are also many conditions that can mimic them.2–4 It is important to distinguish non-epileptic paroxysmal events from epileptic seizures. In epileptic seizures, the suddenly occurring motor, sensory, autonomic, or psychic symptoms are caused by a transient dysfunction of the brain or part of the brain, caused by a sudden discharge of a group of hypersensitive neurons. Different types of seizures and epilepsy syndromes are distinguished.5,6 Non-epileptic paroxysmal events can be due to physiological or exaggerated physiological responses, parasomnias, movement disorders, behavioural or psychiatric disturbances, or to haemodynamic, respiratory or gastrointestinal dysfunction.2,3 They seem to be frequent in early childhood.4,7

Several epidemiological studies have examined the incidence of epilepsy and seizures.1 Few data exist on the incidence of non-epileptic paroxysmal events and the relative frequency of the different types of event in the general population. In one study, they were reported in 25% of the children in the first 2 years of life. Most of these events were innocent events often related to feeding.7 In the UK National Child Development Study, 6.7% of all children had experienced at least one episode of altered consciousness at the age of 11 years.8

As physicians are often confronted with paroxysmal disorders in young children, we wanted to determine how often these disorders occur in the first year of life in an unselected population in the Netherlands, and what part is accounted for by epilepsy and seizures. We hypothesized that the incidence of various types of paroxysmal disorders would be considerable, but that epileptic seizures would only account for a small minority of them. We also examined whether predefined prenatal and perinatal variables were associated with any or all of these disorders.

Method

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Design

This study was embedded in the Generation R Study, a population-based prospective cohort study from early fetal life until young adulthood. This study is being conducted in Rotterdam, the second largest city in the Netherlands. It is designed to identify early environmental and genetic determinants of growth, development, and health in fetal life, childhood, and adulthood and has been described previously in detail.9,10 Enrolment took place in early pregnancy. Assessments during pregnancy included questionnaires, physical examinations, and fetal ultrasound examinations. Postnatal information on growth and development of the participating children was obtained by questionnaires and from information obtained routinely in young children in Dutch child health centres.

This study was approved by the medical ethics committee of the Erasmus Medical Center, Rotterdam. Written informed consent was obtained from the legal representatives of all participants.

Assessment of paroxysmal events

We defined paroxysmal events as suddenly occurring, mostly short-lasting events, with altered consciousness, altered behaviour, involuntary movements, altered muscle tone, and/or a changed breathing pattern alternating with periods during which the symptoms did not occur. With this definition we aimed to include all epileptic seizures as well as conditions that could mimic them. Information about the occurrence of paroxysmal events was collected by questionnaires at the ages of 2, 6, and 12 months. Each of these questionnaires contained direct questions asking whether any epileptic attack, epilepsy, or febrile seizure had occurred. In addition, the parents were asked about symptoms that could have been caused by or associated with a seizure, or a non-epileptic paroxysmal event with symptoms in common with a seizure, according to our definition (Table I). These questions were adapted from previously used screening instruments for epilepsy,8,11–13 adjusted for this young age with some questions added that seemed relevant according to the opinion of the last author (WMFA). If one or more of these questions were answered positively, an extended follow-up questionnaire about this episode was sent to the parents to gather detailed information about the event. In this follow-up questionnaire the parents were also asked to describe the event in their own words. If questions remained, the parents were telephoned for additional information. When a physician was consulted, the medical record of this visit and results of supplementary investigations were obtained. Based on this information, the events were classified into one of eight mutually exclusive categories of paroxysmal events (Table II). When participants reported at least one of a certain paroxysmal event, they were classified as having this paroxysmal disorder independent of how often the event had occurred. The events were classified independently by the first (AMV) and last author (WFMA). Direct consensus was present in 95%; in the other 5%, discussion led to agreement in about half of the events. If consensus could not be reached, the event was classified as ‘unknown’.

Table I.   Screening questions
1. Has your child had an epileptic attack during this period or has your child been diagnosed as having epilepsy?
2. Has your child had feverish fits or convulsions during this period?
3. Has your child sometimes suddenly lost consciousness, either partially or completely, during this period and failed to react when spoken to?
4. Has your child had periods during these months in which he or she was suddenly distracted, and you could no longer make contact with him or her?
5. Has your child ever had an attack of muscular spasms or trembling that started suddenly or any other involuntary movements of the entire body or part thereof (arms, legs, hands, face, torso, tongue) during this period?
6. During this period has your child become completely rigid or tense or contrary to this, completely limp?
7. Has your child had a period during these months when it suddenly stopped breathing or had a different breathing pattern?
8. Have there been occasions during this period when your child suddenly started to show strange behaviour or made strange movements while asleep?
Table II.   Classification of events
  1. aDefinite, diagnosis confirmed by a physician; probable, diagnosis ‘probable seizure’ according to a physician or history characteristic for a seizure; possible, diagnosis ‘possible seizure’ according to a physician or history that could point to a seizure but not characteristic. Possible seizures were categorized in group 8, ‘unknown’.

1. No paroxysmal eventDoes not fulfil our definition of a paroxysmal event
2. Physiological eventBehaviour that can be normally seen in children of this age group
3. Seizure disorderaIncluding epilepsy, neonatal seizures, febrile seizures, other seizures
4. Apnoeic spellCyanotic spells or reflex anoxic spells
5. Loss of consciousnessNot caused by a seizure or a breath-holding spell
6. ParasomniaBenign neonatal myoclonus, repetitive movements, arousal disorder
7. OtherParoxysmal events not belonging to one of the other categories
8. UnknownInsufficient data to classify in one of the aforementioned groups

Events that did not correspond to our definition of a paroxysmal event were classified as ‘no paroxysmal event’. Physiological events were defined as those that were consistent with the normal pattern of behaviour of children that age; most often reported were a short apnoea during sleep in newborn infants or sudden stiffening in reaction to startle or pain.

For the epileptic seizures, the classification of seizures by the International League Against Epilepsy was used.5,6 Febrile seizures were defined as a condition in childhood characterized by seizures occurring during an acute febrile episode but without evidence of intracranial infections or other defined causes.14 Epilepsy was defined as a disorder characterized by recurrent unprovoked seizures. Neonatal seizures were defined as seizures within the first 4 weeks of life. All other seizures were categorized as ‘seizure, other’. All seizures were given the annotation ‘definite’, ‘probable’, or ‘possible’.

Apnoeic spells included cyanotic spells, provoked by anger and crying, and reflex anoxic spells (cardiovagal syncope), provoked by sudden pain or fright. Episodes of losing consciousness due to causes other than an epileptic seizure or apnoeic spell (for example vasovagal syncope) were placed in category 5. The parasomnias included arousal disorders, sleep-wake transition disorders, and other parasomnias (mainly benign neonatal sleep myoclonus).15

All events that did not fit into one of the aforementioned categories were defined as ‘other’. Most of these were related to the respiratory or gastrointestinal system (choking, pseudo-croup, reflux). In some cases there was insufficient information to classify the event into one of the aforementioned groups, or the event remained unclear even after thorough examination. These events were classified as ‘unknown’. Because of the degree of uncertainty, episodes that had been classified as ‘possible seizures’ were reclassified into the ‘unknown’ group.

When participants reported the same event in different questionnaires, we registered this event only once at the youngest age at occurrence.

Prenatal and perinatal risk factors

Information about ethnicity and parity as well as the medical history of the mother were obtained by the first questionnaire at enrolment in the study. Maternal smoking habits were assessed in each questionnaire. Smoking during pregnancy was defined as continued smoking after the pregnancy was known. Gestational age was established by fetal ultrasound examination. At the ultrasound examination in the third trimester of pregnancy (after 25wks) fetal weight was estimated and head circumference was measured. A low estimated fetal weight or a small head circumference were defined as an estimated fetal weight or head circumference under the 10th centile of the whole group, adjusted for gestational age. Date of birth, birthweight, Apgar scores, and sex were obtained from midwife and hospital registries. Preterm birth was defined as birth before 37 completed weeks of gestation. A low Apgar score was defined as a score below 7. Head circumference was measured prenatally at 30 weeks’ gestation and postnatally at about 8 weeks of age. Head circumference at birth was not available in all children.

Study population

In total, 6559 of the prenatally included children were enrolled for the postnatal phase (Fig. 1). Owing to logistic constraints not all participants were sent all questionnaires. In total, 6015 participants were sent the 2-month questionnaire, 5576 participants the 6-month questionnaire, and 6492 the 12-month questionnaire. The response on these questionnaires was about 80%. There were 650 (13.6%) screen-positives in the 2-month questionnaire, 341 (8.6%) in the 6-month questionnaire, and 288 (6.4%) in the 12-month questionnaire. Altogether, 4770, 3984, and 4522 participants returned the 2-, 6-, and 12-month questionnaires respectively, as well as the follow-up questionnaire for these sampling periods for screen-positives. When the three questionnaires were combined, 409 participants reported a paroxysmal event in at least one of these sampling periods. Of all participants, 4820 were sent the 2-, 6-, as well as the 12-month questionnaire. Of these, 2860 returned all three questionnaires. In total, 2292 participants were screen-negative in all sampling periods and 568 were screen-positive at least once. Of those, 104 (18%) were not sent the follow-up questionnaire owing to logistic constraints; of those who did receive the follow-up questionnaire, 70 (15%) did not return it and 139 (24%) did not fulfil our criteria of a paroxysmal event. Therefore, 255 of the 2680 participants returning all three screening questionnaires reported at least one paroxysmal event in the first year of life.

image

Figure 1.  Flow diagram of participants who completed all three questionnaires in the first year of life.

Download figure to PowerPoint

Subgroup analysis showed that participants whose parents or carers did not return their questionnaires were more often of another origin than Dutch, their mothers were not as well-educated, and more often had continued smoking during pregnancy.

Statistical analysis

The incidence of paroxysmal disorders, that is the percentage of children with a paroxysmal disorder in the first year of life, was calculated. For this calculation only participants who returned all three questionnaires in the first year of life were used (n=2860). As the profile of positively answered screening questions of the responders and of the non-responders on the follow-up questionnaire was similar (data not shown), we also calculated the estimated incidence assuming that the distribution of events in the non-responders was similar as in the responders. For each sampling period the distribution of the different disorders was calculated. For this calculation the participants who answered the questionnaires for this sampling period were used, irrespective of whether they answered the other questionnaires (0–2mo, n=4770; 3–6mo, n=3984; 7–12mo, n=4522). To calculate the distribution of the different disorders in the entire first year only participants who answered all questionnaires were used (n=2860). Exploratory analyses examined whether prenatal or perinatal factors were associated with the occurrence of paroxysmal disorders in the first year of life. For this calculation we used participants who reported at least one paroxysmal event in one of the sampling periods as cases, irrespective of whether they answered the other questionnaires in the first year of life (n=409). The participants who were screen-negative in all questionnaires were used as controls (n=2292). First unadjusted odds ratios (ORs) were calculated. Afterwards we included all variables that were associated with the risk of paroxysmal events in univariable models (p<0.1) for multivariable logistic regression analysis. This was done to determine whether any apparent associations between explanatory variables and the risk of paroxysmal events might have instead been caused by correlations with other associated explanatory variables. All statistical analyses were performed using the Statistical Package of Social Sciences version 15.0 for Windows (SPSS, Chicago, IL, USA).

Results

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Paroxysmal events

Table III shows the distribution of the different types of paroxysmal disorder in the first year of life. Paroxysmal events were reported most frequently in the first 2 months of life, declining over the 3- to 6-month and the 7- to 12-month periods. The incidence of paroxysmal disorders, that is the percentage of children with a paroxysmal disorder in the first year of life, was 255 out of 2860 (8.9%) of which 133 of 255 (52%) were physiological events. Probably this value of 8.9% is an underestimate as not all screen-positives returned the follow-up questionnaire. Extrapolating the distribution of events in the responders to the non-responders on the additional questionnaire would result in a percentage of paroxysmal disorders in the first year of life of 12.9%.

Table III.   Paroxysmal disorders in the first year of life
 0–2mo n=4770 n (%)3–6mo n=3984 n (%)7–12mo n=4522 n (%)First year of life n=2860 n (%)
  1. aSeizure associated with fever. bNumber represents event types (259 event types in 255 participants). Numbers represent the number of participants with the specific event at least once in this period. Calculation of the distribution of different paroxysmal disorders in the three sampling periods was based on participants who answered at least the 2- (n=4770), 6- (n=3984), and 12-month (n=4522) questionnaires respectively. Calculation of the distribution of paroxysmal disorders in the first year of life was based on participants who answered all questionnaires in the first year of life (n=2860).

Physiological events150 (62)40 (45)15 (18)133 (51)
Seizures7 (3)7 (8)23 (27)24 (9)
 Febrile seizure1a72017
 Epilepsy122
 Neonatal seizure54
 Other 11
Loss of consciousness18 (8)16 (18)7 (8)26 (10)
Apnoeic spells5 (2)6 (7)11 (13)12 (5)
Parasomnias8 (3)5 (5)3 (4)10 (4)
Paroxysmal events, other40 (17)11 (12)11 (13)36 (14)
Unknown12 (5)5 (5)14 (17)18 (7)
Total2409084259b

In the first 6 months of life most events consisted of physiological events, whereas in the 7 to 12 months age category physiological events constituted less than 20% of the reported paroxysmal events. In the first year of life 25 children had at least one seizure and most of these (n=17) were febrile. They occurred especially between 6 and 12 months of age. Neonatal seizures occurred in four children. Epilepsy was diagnosed in two children in the first year of life.

Risk factors

Table IV shows that children who are born preterm (OR 2.44; 95% confidence interval [CI] 1.61–3.70) or have a low Apgar score at 1 minute (OR 1.98; 95% CI 1.29–3.04) have a higher incidence of paroxysmal disorders in the first year of life. Physiological events are reported more often in first-born children (OR 1.93; 95% CI 1.40–2.66). As these events form a large part of the paroxysmal disorders in the first year of life, this remains a significant factor when all disorders are taken together (OR 1.50; 95% CI 1.20–1.88). In multivariable logistic regression analysis, preterm birth and a low estimated fetal weight at ultrasound examination in late pregnancy are significantly associated with non-physiological paroxysmal disorders in the first year of life.

Table IV.   Relations between pre- and perinatal variables and the occurrence of a paroxysmal disorder in the first year of life
 Paroxysmal disorders, all (n=409) OR (95% CI)Non–physiological events (n=174) OR (95% CI)Physiological events (n=205) OR (95% CI)Febrile seizures (n=28) OR (95% CI)Apnoeic spells (n=22) OR (95% CI)Loss of consciousness (n=41) OR (95% CI)
  1. ap<0.05 in multivariable analyses; bp<0.05 in univariable analyses; cp<0.10 in univariable analyses. Total number of participants was 2701; participants who reported a paroxysmal event in the first year of life (n=409) or who were screen-negative at all questionnaires in the first year of life (n=2292). All variables associated with the risk of paroxysmal events in univariate models (p<0.10) were included for multivariable analysis. EFW, estimated fetal weight; HC, head circumference; preterm, gestational age <37wks at birth.

Maternal determinants
 Smoking during pregnancy1.15 (0.83–1.59)1.06 (0.65–1.73)1.08 (0.69–1.69)2.72 (1.05–7.00)a1.13 (0.33–3.84)0.44 (0.11–1.86)
 Primiparous1.50 (1.20–1.88)a1.08 (0.79–1.49)1.93 (1.40–2.66)a1.19 (0.55–2.58)1.42 (0.58–3.49)1.28 (0.67–2.44)
 History of febrile seizures1.31 (0.67–2.54)1.03 (0.37–2.87)1.45 (0.62–3.43)1.89 (0.25–14.35)2.12 (0.28–15.2)1.12 (0.15–8.30)
Child determinants
 Male1.01 (0.82–1.24)1.10 (0.81–1.49)0.99 (0.74–1.32)1.62 (0.75–3.46)1.04 (0.45–2.41)1.34 (0.72–2.49)
 Ethnicity, Dutch1.07 (0.85–1.34)1.17 (0.83–1.65)1.00 (0.74–1.36)1.67 (0.68–4.15)0.79 (0.33–1.90)1.24 (0.62–2.49)
 EFW 30wks pregnancy SDa less than tenth centile1.21 (0.87–1.70)1.53 (0.98–2.41)a0.94 (0.58–1.54)2.04 (0.77–5.40)2.08 (0.70–6.18)1.29 (0.50–3.32)
 HC 30wks pregnancy SD less than tenth centile1.08 (0.77–1.54)1.31 (0.81–2.10)0.95 (0.58–1.55)2.59 (1.04–6.44)b0.93 (0.22–4.02)1.01 (0.36–2.86)
 Preterm birth2.44 (1.61–3.70)a2.27 (1.29–3.99)a2.03 (1.17–3.51)a2.72 (0.81–9.14)a0.55 (0.08–4.06)
 Birthweight SD less than tenth centile1.13 (0.80–1.59)1.21 (0.75–1.97)1.00 (0.62–1.61)1.18 (0.35–3.95)0.43 (0.06–3.19)1.61 (0.67–3.86)
 Apgar score at 1min <71.98 (1.29–3.04)a2.04 (1.14–3.64)b1.69 (0.95–3.00)c1.70 (0.40–7.23)1.04 (0.14–7.82)3.13 (1.21–8.13)b
 Apgar score at 5min <72.33 (0.96–5.65)c2.95 (1.00–8.73)c1.11 (0.26–4.75)2.87 (0.38–21.74)

Table IV also shows the results for some subgroups of paroxysmal disorders. Children with a small head circumference on the ultrasound examination at 30 weeks’ gestation had a higher incidence of febrile seizures in the first year of life (OR 2.59; 95% CI 1.04–6.44); the same was true for children whose mother continued smoking during pregnancy (OR 2.72; 95% CI 1.05–7.00) or who were born preterm. Also, children whose mothers had a history of febrile seizures themselves tended to have febrile seizures more often; however, this was not significant. In multivariable regression analysis the association with head circumference was no longer statistically significant (OR 1.89; 95% CI 0.63–5.67). None of the risk factors studied predisposed for apnoeic spells in the first year of life. Children with a low Apgar score 1 minute after birth more often had an episode with loss of consciousness.

Discussion

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

In this prospective population-based study we found that paroxysmal disorders, defined as disorders characterized by suddenly occurring episodes with an altered consciousness, altered behaviour, involuntary movements, altered muscle tone, and/or a changed breathing pattern, are common in the first year of life, with an estimated incidence of 8.9 to 12.9%. The largest group of these events are innocent, physiological events. Epileptic seizures only form a small minority of these events. Preterm birth and low Apgar scores were associated with paroxysmal disorders in the first year of life. Physiological events were reported more often in first-born children. Children with a small head circumference at 30 weeks’ gestational age, who were born preterm, or whose mothers continued smoking during pregnancy had a higher incidence of febrile seizures in the first year of life.

We found an incidence of paroxysmal disorders, i.e. the percentage of children with a paroxysmal disorder in the first year of life, of 8.9 to 12.9%. The only other study on different types of paroxysmal event in the general population that we are aware of reported a prevalence of 25% in the first 2 years of life.7 That study also found preterm birth and a low Apgar score at birth to be associated with paroxysmal events in early life. This may partly be a result of more anxiety and attention among parents of children with a difficult start, but these children may also be more vulnerable for some of these events. It is remarkable that low Apgar scores are associated with paroxysmal disorders in the first year of life, as it is well established that early Apgar scores at 1 and 5 minutes have no substantial predictive value on neurological outcome.16 That physiological events are reported more for first-born children is probably explained by parents with their first child not yet being familiar with certain events in young children.

We found that children who had a small head circumference on the ultrasound examination in late pregnancy had a higher incidence of febrile convulsions in the first year of life. This could point to negative influences on brain development during pregnancy rendering the child more susceptible to febrile convulsions later in life. Possibly the smoking behaviour of the mother might play a role in this as it is known that smoking during pregnancy has a negative impact on brain development.17 We found that mothers of children with febrile convulsions more often had continued smoking during pregnancy. After adjusting for smoking during pregnancy the association between head circumference and febrile seizures became smaller, suggesting that smoking might be part of the causal pathway. The literature on the association of smoking during pregnancy and febrile seizures in the offspring shows conflicting results. In the largest study a small increase in the incidence of febrile convulsions was found in the offspring of mothers smoking more than 10 cigarettes a day during pregnancy.18,19 The relation between maternal smoking, fetal brain development, and febrile convulsions deserves further investigation.

We found that epileptic seizures only form a small subgroup of paroxysmal events in the first year of life. This is important to realise as an erroneous diagnosis of epilepsy can have negative consequences, both because of unnecessarily prescribed medication with possible adverse effects and because of the negative implications this diagnosis might have for daily life.20

The strength of this study is the population-based cohort with a large number of participants studied from early pregnancy. Our data were collected prospectively and are, therefore, less dependent on parental recall. Our study also has some limitations. In the first place, no validated screening instrument for epilepsy or seizures exists for this young age. In the past, different screening questionnaires have been used for seizure-disorders, few of which had been validated before.8,11–13 Our screening questions were partly based on these existing questionnaires, adjusted for this low age group, and extended with some questions that appeared relevant according to an experienced child neurologist. We found only two children diagnosed with epilepsy in the first year. However, in this relatively small population, this is compatible with what might be expected for a relatively infrequent disorder like epilepsy.1 The specificity of our questions for epileptic events was low and most of the paroxysmal events did not concern seizures. We used these events to estimate the incidence of non-epileptic paroxysmal disorders in the general population.

Information on paroxysmal events was missing in about 20% of the participants in the three sampling periods (0–2, 3–6, and 7–12mo) and only 2860 participants received and returned all screening questionnaires in the first year of life. About 25% of all screen-positives, did not receive or return the supplementary questionnaire. Subgroup analysis showed that participants whose parents or carers did not return the questionnaires were more often of another origin than Dutch, their mothers were less educated, and more often had continued smoking during pregnancy. Therefore, our results should be only cautiously extrapolated to this group. The estimated incidence of paroxysmal disorders in the first year of life might be affected. Some parents might not report a physiological event as they already feel the event was innocent or they did not pay attention to it. This could lead to an underestimation of the incidence of physiological events. Using the complete cases only for the first year analysis might have introduced bias in the subdivision of paroxysmal disorders. However, as the subdivision of paroxysmal disorders in the three sampling periods was similar for participants who participated in the full year analysis and for those who did not, this seems unlikely. Our effect estimates for the associations of prenatal and perinatal factors with paroxysmal disorders would be biased if these associations differed between those with and without complete data. This cannot be excluded but seems unlikely.

In conclusion, we found that paroxysmal events are seen frequently in early childhood. Many of these events are innocent physiological events, especially in the first 6 months of life. Epileptic seizures only form a minority of the paroxysmal events in the first year of life. This information seems to us to be very relevant. It may be of help in developing rational investigation strategies in infants who present with paroxysmal events. Moreover, we found preterm birth, low birthweight, and low Apgar scores to be associated with paroxysmal disorders in the first year of life. Children with a small head circumference in the third trimester of pregnancy have a higher incidence of febrile seizures in the first year of life. Maternal smoking during pregnancy might be partly responsible for this finding. However, as there was multiple testing, our results should be considered as hypothesis generating rather than suggesting direct causal associations. Further studies are needed to explain these associations.

Acknowledgements

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The Generation R Study is conducted by the Erasmus Medical Center Rotterdam, in close collaboration with the Faculty of Social Sciences of the Erasmus University, Rotterdam, the Municipal Health Service Rotterdam area, the Rotterdam Homecare Foundation, and the Stichting Trombosedienst en Artsenlaboratorium Rijnmond, Rotterdam. We acknowledge the contribution of general practitioners, hospitals, midwives, and pharmacies in Rotterdam. The first phase of the Generation R Study was made possible by financial support from the Erasmus Medical Center Rotterdam, the Erasmus University Rotterdam, and the Netherlands Organization for Health Research and Development. The present study was supported by an additional grant from the Dutch National Epilepsy Foundation.

References

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References