RESEARCH LETTERS
tube or its coiling in the mouth. By using the swallowing
reflex the tube was successfully placed at the first or second
attempt in 19 of 20 trials. The only failure occurred in a
patient with a left MCA infarction in whom the swallowing
reflex could not be provoked. No serious complications,
such as severe coughing—with the potential risk of
inducing an increase of the intracranial pressure—or
respiratory distress were noted. The procedure was well
tolerated by all patients.
Our study confirms that the placing of nasogastric tubes
by inducing the swallowing reflex is a useful alternative if
the conventional method fails. Our new method is much
less distressing for the patients than direct placement of the
nasogastric tube (with the help of a laryngoscope and
Magil forceps), which requires sedation of the patient and
which is usually recommended for such patients.5
However, our procedure may not work in patients who
have had a brainstem infarction because they are prone to
lack a sufficient swallowing reflex.
1 Mann G, Dip PG, Hankey GJ, Cameron D. Swallowing function
after stroke—prognosis and prognostic factors at 6 months. Stroke
1999; 30: 744–48.
2 Smithard DG, O’Neill PA, Park C, et al. Complications and
outcome after acute stroke—does dysphagia matter? Stroke 1996; 27:
1200–04.
3 Nakajoh K, Nakagawa T, Sekizawa K, Matsui T, Arai H, Sasaki H.
Relation between incidence of pneumoniae and protective reflexes in
post-stroke patients with oral or tube feeding. J Int Med 2000; 247:
39–42.
4 Teramoto S, Matsuse T, Fukuchi Y, Ouchi Y. Simple two-step
swallowing provocation test for elderly patients with aspiration
pneumonia. Lancet 1999; 353: 1243.
5 Dyer I, Ashton WB. How to pass a nasogastric tube. Br J Hosp Med
1991; 45: 45–46.
Department of Neurology, Universitätsklinikum Münster, 48129
Münster, Germany (Rainer Dziewas MD, Peter Lüdemann MD,
Carsten Konrad MD, Florian Stögbauer MD)
Correspondence to: Dr Rainer Dziewas
(e-mail: dziewas@uni-muenster.de)
Association between idiopathic
infantile macrocephaly and autism
spectrum disorders
Patrick F Bolton, Melanie Roobol, Lucy Allsopp, Andrew Pickles
We conducted a case-controlled, catch-up study of a cohort of
boys born with macrocephaly in order to determine
whether infantile macrocephaly is a risk marker for the
later development of autism spectrum disorders. Our results
show that infantile macrocephaly was associated with an
increased risk of developing autism spectrum disorders
(odds ratio 5·44, 95% CI 1·11–52·15; p=0·03). These findings
suggest that neurobiological differences during infancy may
predict behavioural manifestations of autism spectrum
disorders.
Lancet 2001 358: 726–27
Autism is a disorder that becomes manifest in the first
3 years of life, but it remains unclear when the
neurobiological abnormalities that may underpin it begin.
Genetic factors are important in its aetiology and create a
liability for a broad range of social and communication
impairments that includes Asperger’s syndrome and other
forms of autistic-like pervasive developmental disorder.1
726
Copyright © 2001 All Rights Reserved
For personal use. Only reproduce with permission from The Lancet Publishing Group.
Group Boys (n) Cases of ASD (%)
General population5 ~7746 73 (0·94)
Normocephalic 362 2 (0·55)
Macrocephalic 306 9 (2·94)*†
*General population rate versus macrocephaly (p=0·003).5 †Normocephalic
versus macrocephalic (p=0·03).
Frequency of autism spectrum disorders (ASD) in boys
These conditions are collectively referred to as autism
spectrum disorders. Macrocephaly (head circumference
>97th centile) occurs in 25–30% of autism spectrum
disorder cases and is considered to be a consequence of
megalencephaly.2 The risk of a macrocephalic child
developing an autism spectrum disorder has not yet been
established. We compared the rates of autism spectrum
disorders in an epidemiological sample of boys with and
without infantile macrocephaly.
The Cambridge and Huntingdon Health Authority
child-health surveillance programme has systematically
collected data from GP and health-visitor checks on all
children born since 1991 and living in the health-authority
area. From 1993 onwards all contacts with community
health services (paediatric, child psychiatry, and speech
therapy services) have also been recorded. We identified all
boys born between May 1, 1991, and December 31, 1993,
and selected those that had macrocephaly at the GP and
health-visitor’s 8-month check, usually conducted between
20 and 52 weeks of age (n=307). In addition, we identified
all normocephalic boys (head circumference 25th–75th
centiles) (n=1239) and selected children listed sequentially
in the database (excluding those with medical disorders
known to give rise to developmental problems) as
age-matched controls (n=362). The macrocephalic
children were more likely than their normocephalic
counterparts to come from higher social grades (social
grading was derived from postcodes with the ACORN
classification system, collapsing grades 1+2, 3+4, 5+6).3
The number of macrocephalic and normocephalic children
in each social grade was 159 versus 131 grade 1+2, 99
versus 156 grade 2+3, 49 versus 75 grade 4+5 (p<0·0001
by Fisher’s exact test).
The autism screening questionnaire4 was posted to parents
of all the children. 308 (46·4%) parents replied and
confirmed that their child had no known handicapping
medical disorder. Seven children had scores above threshold
that indicated a possible diagnosis of an autism spectrum
disorder. Macrocephalic children were more likely to have
suprathreshold scores than normocephalic children (7/162 vs
0/146, respectively; Fisher’s exact test p=0·02). To take
account of any response biases, further analyses were
undertaken with complex survey analysis methods
(STATACorp, 1999). These confirmed the higher rates of
suspected autism spectrum disorders in the macrocephalic
group (p=0·004; data weighted for non-response by strata
defined by social grade, macrocephaly status, and age).
Clinical records of all children referred to paediatric or
child psychiatry services were also examined (158/669
children). Nine had a diagnosis of idiopathic autism
spectrum disorder according to ICD-10 criteria (one autism,
three atypical autism, and five Asperger’s syndrome). Four
of these nine cases had also been identified by the
questionnaire survey. There remained, therefore, three
children with high autism screening questionnaire scores
who had not been clinically diagnosed with an autism
spectrum disorders. Parents of two of these children agreed
to further assessments. Both were confirmed as having an
autism spectrum disorder (one had atypical autism and the
other Asperger’s syndrome) with the Autism Diagnostic
Interview—Revised and Observation Schedule. The third
THE LANCET • Vol 358 • September 1, 2001

undiagnosed child with a suspected autism spectrum
disorder had been noted to have developmental problems
and macrocephaly by paediatric services, but the family was
not available for further evaluation and so he was excluded
from the study. Overall, there were more macrocephalic
children (9/306) than normocephalic children (2/362) with a
confirmed ICD-10 diagnosis of an autism spectrum disorder
(odds ratio [OR] 5·44, 95% CI 1·11–52·15; p=0·03).
Moreover, the rate of idiopathic autism spectrum disorders
in the macrocephalic children was also higher than that
expected in the general population (binomial test, p=0·003)
based on a recent UK prevalence study of autism spectrum
disorders in children in this age range (general population
prevalence rate=62·6/10 000; male to female ratio=4:1.5 By
contrast, the rate of autism spectrum disorders among the
normocephalic children was no greater than expected
(binomial test, p=0·8). The higher rate of autism spectrum
disorders in the macrocephalic children remained after
stratifying by social grade (OR 5·61, 95% CI 1·13–54·55;
p=0·03). Seven of the nine children with infantile
macrocephaly and a diagnosis of an autism spectrum
disorder were still macrocephalic in later childhood. The
remaining two children had head circumferences on the 75th
and 85th centiles. Computed tomography or magnetic-
resonance imaging brain scans in five macrocephalic children
with confirmed autism spectrum disorders did not reveal an
identifiable medical cause for the macrocephaly.
Many of the children in the survey (39%) had been in
contact with speech therapy, paediatric, or child psychiatry
services and so most (68·5%) had been screened for autism
spectrum disorders either through contact with clinical
services or by our questionnaire survey. Most cases of autism
spectrum disorders in these cohorts were likely, therefore, to
have been identified.
This study reports the rate of autism spectrum disorders in
a population-based sample of children with infantile
macrocephaly. We have found that the association between
macrocephaly and the development of an autism spectrum
disorder is evident in infancy, well before the emergence of
clear signs of the disorder. Our findings suggest that in some
children with autism spectrum disorders neurobiological
differences are present before exposure to putative causal
environmental factors. The results also indicate that infantile
macrocephaly may be an early marker of risk and could be
of use in prospective longitudinal studies of the early
development of autism spectrum disorders. Further research
will be required to clarify when the neurobiological
differences first emerge and how brain growth proceeds in
children with autism spectrum disorders.
1 Bailey A, Luthert P, Dean A, et al. A clinicopathological study of
autism. Brain 1998; 121: 889–905.
2 Fombonne E, Roge B, Claverie J, Courty S, Fremolle J.
Microcephaly and macrocephaly in autism. J Autism Dev Disord 1999;
29: 113–19.
3 Morgan M, Chinn S. ACORN group, social class, and child health.
J Epidemiol Community Health 1983; 37: 196–203.
4 Berument S, Rutter M, Lord C, Pickles A, Bailey A. Autism screening
questionnaire: diagnostic validity. Br J Psychiatry 1999; 175: 444–51.
5 Chakrabarti S, Fombonne E. Pervasive developmental disorders in
preschool children. JAMA 2001; 285: 3093–99
Autism Research Centre and Developmental Psychiatry Section,
University of Cambridge, Cambridge CB2 2AH, UK
(P F Bolton FRCPsych, M Roobol MRCPsych, L Allsopp MRCPsych); and School
of Epidemiology and Health Science and Centre for Census and
Survey Research, University of Manchester, Manchester
(A Pickles PhD)
Correspondence to: Dr Patrick Bolton
(e-mail: pfb1000@cam.ac.uk)
THE LANCET • Vol 358 • September 1, 2001
Copyright © 2001 All Rights Reserved
For personal use. Only reproduce with permission from The Lancet Publishing Group.
RESEARCH LETTERS
Positron-emission tomography of
vector-mediated gene expression in
gene therapy for gliomas
A Jacobs, J Voges, R Reszka, M Lercher, A Gossmann, L Kracht, Ch
Kaestle, R Wagner, K Wienhard, W D Heiss
In clinical gene-therapy trials for recurrent glioblastomas,
transduction of the herpes simplex virus type-1 thymidine
kinase (HSV-1-tk) gene with subsequent prodrug activation by
ganciclovir was found to be safe, but clinical response was poor.
We used positron-emission tomography (PET) with I-124-
labelled 2’-fluoro-2’-deoxy-1-D-arabino-furanosyl-5-iodo-uracil
([124I]-FIAU)—a specific marker substrate for gene expression of
HSV-1-tk—to identify the location, magnitude, and extent of
vector-mediated HSV-1-tk gene expression in a phase I/II
clinical trial of gene therapy for recurrent glioblastoma in five
patients. The extent of HSV-1-tk gene expression seemed to
predict the therapeutic response. The expression of an
exogenous gene introduced by gene therapy into patients with
gliomas can be monitored non-invasively by PET.
Lancet 2001; 358: 727–29
The design of effective gene therapy strategies relies on
concerted research to define the genetic and
pathophysiological alterations causing the disease so that the
biological characteristics of the target tissue can be
understood. A safe vector and expression system can then be
developed to achieve efficient, targeted, and regulated
alteration of specific therapeutic gene expression. In clinical
gene-therapy trials for recurrent glioblastomas, transduction
of the herpes simplex virus type-1 thymidine kinase (HSV-1-
tk) gene with subsequent prodrug activation by ganciclovir
was found to be safe, however, clinical responses were
observed only in a few patients with small brain tumours.
One of the most important issues for making gene therapy
widely applicable to human beings is the development of
technology for non-invasive monitoring of the location,
magnitude, and duration of vector-mediated gene expression
in vivo.1–4 Prompted by the death of a patient with liver-
directed adenovirus vector-mediated gene therapy, the
Recombinant DNA Advisory Committee of the National
Institutes of Health, USA, called for better assays for
measuring transgene expression in cells and tissues.
Non-invasive localisation of gene expression (molecular
imaging) relies on the transduction of marker genes that
encode enzymes or receptors that lead to a regional
accumulation of radiolabelled or paramagnetic marker
substrates or receptor binding compounds. These substrates
or compounds can then be detected by radionuclide or
magnetic-resonance imaging (MRI).1–4 I-124-labelled 2-
fluoro-2-deoxy-1-D-arabino-furanosyl-5-iodo-uracil ([124I]-
FIAU), a specific substrate for HSV-1 thymidine kinase
(HSV-1-tk), and positron-emission tomography (PET) have
been successfully used for the non-invasive localisation of
retroviral,1 adenoviral,2 and herpes viral4 vector mediated
HSV-1-tk gene expression in rodent models. We used the
same method in a prospective gene-therapy trial to investigate
the safety of intratumourally infused liposome–gene complex
(LIPO-HSV-1-tk) followed by ganciclovir administration.
Vector-mediated HSV-1-tk gene expression was followed
by PET and [124I]-FIAU in five patients (age range 49–67
years) with recurrent glioblastomas. After biopsy sampling of
the tumour, one or two catheters were stereotactically placed
within the tumour and subcutaneously connected to a port.
This allowed targeted intratumoural infusion of the cationic
liposomal vector DAC-30 (3  (N-N, N-dimethyl-
aminoethane-carbamoxyl-cholesterol/1,2 dioleoylphospha-
727