The subjects were recruited from the population-based TCHAD-study ,
and comprised of twins born from May 1985 to December 1986, living in
Stockholm. Of 271 twin pairs, 156 twin pairs i.e. 312 individuals (135
boys and 177 girls), and at least one parent participated in the
interview procedure. Mean age was 16 years, ranging from 14.6 to 16.7
years. The sample consisted of 17 pairs of monozygotic boys, 16 pairs
of dizygotic boys, 34 pairs of monozygotic girls, 17 pairs of dizygotic
girls, and 72 pairs of mixed-sex dizygotic twins. Blood sample was
obtained from 247 adolescent individuals (106 boys and 141 girls).
Sixty-seven families refused to participate and 48 families did not
answer by phone and letters. (Fig 1). Four pairs of twins (8 individuals) were of non-Caucasian origin.
Figure 1. Sample selection procedure.
The twin pairs and parents were invited to an assessment including a
structured clinical interview with Kiddie Schedule for Affective
Disorders and Schizophrenia for School-Age Children-Present and
Lifetime Version (K-SADS-PL). The K-SADS-PL is a widely used semi
structured diagnostic interview designed to assess current and past
episodes of psychopathology in children and adolescent according to
Diagnostic and Statistical Manual of Mental Disorders, 3rd revised
edition (DSM-III-R) and DSM-IV criteria. K-SADS-PL has been shown to be
a reliable and valid diagnostic instrument for child and adolescent
psychiatric diagnoses .
In this study, the present version of the Swedish K-SADS-PL was used
to assess the symptoms according to DSM-IV. In the K-SADS-PL interview
procedure, the interview was first performed with the child alone and
thereafter with at least one parent. The clinical interviewer then
summarized the information from the parent (about the child) with the
information from the child and classified the symptoms as "not present"
(0), "possible" (1) or "certain" (2).
For the purpose of this study, the symptoms included in the criteria
for ADHD subtypes, ODD and CD were assessed according to K-SADS-PL and
compiled in three ways:
1. Dimensional scales of the symptoms, that is, the summary scores
of symptoms included in the criteria for the following diagnoses
(scaled 0–2) were calculated: ADHD inattentive type, ADHD hyperactive
type, ADHD combined type, ODD, CD and a combined scale for CD and ODD.
2. DSM-IV diagnostic criteria where applied using the information
from K-SADS-PL. Each symptom was counted if the item was assessed as
"certain" by the interviewer giving the following diagnoses: ADHD
inattentive type, ADHD hyperactive type, ADHD combined type, ODD, CD
and the combination (CD or ODD) (Table 1).
Attention-deficit and disruptive behavior disorders derived from
Kiddie-SADS – PL interviews in a population-based sample of adolescents
3. The symptoms of the diagnoses described in 1 and 2
were also dichotomized. DSM-IV diagnostic criteria were applied
according to the following; the individual child was regarded as having
"high score of the ADHD phenotype" if the symptom was assessed as
"possible" or "certain" using the same diagnostic criteria as in 2. In
order to avoid confusion with clinical diagnoses, we have chosen to use
"high/low scores of phenotype" instead of the term "sub threshold" or
"probable" diagnosis (Table 1).
To study the validity of the high/low dichotomization we used the
Children Global Assessment Scale (CGAS) which is a scale for children
very similar with the Global Assessment of Functioning in DSM-IV .
Genotyping and MAO-B activity measurement
Blood samples were obtained from 247 individuals (123 twin pairs and
one individual), 106 boys and 141 girls, and genomic DNA was isolated
by standard methods. For MAO-B activity measurement, platelet rich
plasma was prepared by low-speed centrifugation, 200 × g for 10
minutes. Platelet concentration of the plasma samples were estimated
electronically and the plasma was stored at -80°C.
PCR-based genotyping was performed as described in  (MAO-A VNTR) and 
(5-HTT LPR). The MAO-A and 5-HTT PCR products were analyzed by
electrophoresis on 2% agarose gels and visualized under UV light by
ethidium bromide staining. Genotypes were called in two separate
Because the MAO-A gene is X-linked, only boys were included in the
analysis of the MAO-A VNTR. Girls, having two X chromosomes, can be
heterozygous and cannot be functionally characterized with certainty
because it is not possible to know which of the two alleles is
Catalytic activity of platelet MAO-B was analyzed with C14-labelled 2-phenylethylamine (β-PEA)
as substrate. Before analysis, the samples of platelet rich plasma were
thawed and sonicated at 0°C during 5 × 10 seconds with intervals of 5
seconds for lysis of the platelets. 50 μl of the plasma was added to 50 μl of 0.1 mM 14C-β-PEA (0.5 μCi/ml)
in 0.1 M sodium phosphate buffer. The reaction mixture was incubated at
37°C for 4 minutes, and the reaction terminated by the addition of 30 μl 1 M HCL. Thereafter, the radioactive aldehyde product formed was extracted by shaking for 30 seconds into 750 μl
toluene: ethylacetate (1:1). The samples were then centrifuged at room
temperature for 5 minutes at 1000 × g. The organic phase (500 μl),
containing the aldehyde product, was pipetted into vials with 8 ml
scintillation fluid and the amount of radioactive aldehyde product
subsequently quantified by scintillation analysis. Enzyme activity is
expressed as nmol of substrate oxidized per 1010 platelets per minute. All samples were analyzed blindly and in duplicate.
There are compounds in cigarette smoke that exert inhibitory effect
on MAO activity, however only in quantities exceeding 300 cigarettes
per month .
Information about cigarette smoking was obtained by asking the
subjects, at the time for the blood sampling, whether they had smoked
in the past 24 hours (10% boys, 11% girls). This information about
smoking was used as a covariate in the statistical analyses of the
relationships between behavior and platelet MAO-B activity. The
K-SADS-PL inventory contains questions about smoking habits as well. We
used smoking more than 2 cigarettes a day as cutoff (11% boys, 12%
girls). In order to further explore the possible effects of smoking on
our results, subjects with positive smoking information from K-SADS-PL
were, for the statistical analysis, combined with the subjects that had
been smoking the past 24 hours before the blood test.
Each of the main behavioral symptoms ADHD, ODD and CD, included
either as categorical variables (high/low scores of ADHD phenotype) or
dimensional scales, were analyzed with respect to their association
with MAO-B activity in platelets and genotype of the MAO-A VNTR and the
5-HTT LPR. Separate analyses were performed for boys and girls. For
these analyses the general linear model (GLM) in Stata statistical
software package was used .
Standard errors were adjusted for clustering within twin pairs by
increasing the estimated standard errors giving robust estimates of for
example p-values. The method is based on the sandwich or Huber/White
variance estimator, a method available in Stata 9.0. The descriptive
analysis of the data presented in the tables includes the entire
sample. However, in the genetic association analyses one of the
children in each pair of the MZ twins was randomly excluded from the
statistical analyses (n = 51). This is because the MZ twins are
supposed to share their genetic background which makes it somewhat
uncertain whether the adjustment of standard errors, as described
above, is correct for MZ twins.
The dimensional measures of ADHD, ODD and CD problems had skewed
distributions. Therefore, for the dimensional measures, non-parametric
bootstrap tests were performed using the GLM model in Stata. In the GLM
analyses, when calculating the relationships between behavioral
problems and 5-HTT genotype, the nominal scales included three variants
(long-long, long-short, short-short alleles). This nominal scale was
transformed into three index variables and used in the GLM analyses
using the long-long variant as the reference variable (omitting this
variable). In each of the GLM analyses of platelet MAO-B activity, the
information about cigarette smoking, at the time for the blood
sampling, whether they had smoked in the past 24 hours, and the smoking
information from K-SADS-PL were included as a covariate; smoking
(yes/no). For the dimensional ADHD/disruptive behavior scales
interactions between sex and polymorphisms in the 5-HTT genotypes were
analyzed using GLM. In the same way interactions between sex and
platelet MAO-B activity were studied.
Children with missing values for two or more symptoms included in
the criteria for each of the diagnoses in this study (the ADHD
subtypes, ODD and CD respectively), were excluded from the analyses. If
there were missing values for only one of the symptoms in each of the
diagnoses this missing value was recorded as "no symptom."
The study was approved by the ethical committee at Karolinska
Hospital, Stockholm, Sweden. For all subjects the participating parent
and the teenagers gave written informed consent.