For the full versions of these articles and the references see bmj.

com
EDITORIALS

Physical activity and health

Even low intensity exercise such as walking is associated with better health
Evidence that physical activity improves health is con- tions is the lack of assessment of health gains; however,
vincing,1 but we lack knowledge about how to increase epidemiological studies suggest that health benefits
physical activity in individuals and populations. Taking of active transport are substantial. The nurses health
part in sport may improve health, but sport is only study found that women who increased both walking
taken up by a small proportion of the adult population, distance and speed had a lower risk of cardiovascular
and mainly by the better educated. disease, type 2 diabetes, and all cause mortality.8 9 The
����
comstockcomplete.com

In this week’s BMJ, a systematic review by Ogilvie risk in the upper quintile of walking was around half
and colleagues assesses the effect of interventions to that seen in the sedentary group. Similarly, another
improve walking on how much people walk, physical study found a 30% lower mortality rate in participants
activity, fitness, disease risk factors, and wellbeing.2 who cycled to work than in non-cyclists after adjust-
It found that interventions tailored to people’s needs, ing for general physical activity level, socioeconomic
which targeted the most sedentary or those motivated background, and smoking.10
Research, p 1204 to change, can increase walking by up to 30-60 minutes Ogilvie and colleagues’ study shows that interven-
each week. Few studies included in the review assessed tions can increase the amount of walking. It has not yet
Lars Bo Andersen professor clinical benefits from the increased walking, and this been proved that the lower rates of disease and mortal-
Norwegian School of Sport
Sciences, Department of Sports remains to be shown in randomised controlled trials. ity seen in people who walk is caused by walking itself,
Medicine, Box
���� ������
4014, ������
0806, ������
Oslo, So what is the evidence so far on the effects of inter- but even this low intensity type of exercise probably
Norway ventions on other types of physical activity? A recent improves metabolic control and other health para‑
Lars.bo.andersen@nih.no
Competing interests: None
Cochrane review of randomised controlled trials found meters. The challenge now is to make politicians work
declared. that trials promoting physical activity in general sig- for an environment that promotes walking, and to call
Provenance and peer review: nificantly increased self reported physical activity on doctors to encourage patients to walk, especially
Commissioned; not externally peer (standardised mean increase of 0.31, 95% confidence those with disorders such as hypertension, metabolic
reviewed.
interval 0.12 to 0.50), and fitness (0.40, 0.0.9 to 0.70).3 syndrome, or raised fasting insulin.11
BMJ 2007;334:1173 The review by Ogilvie and colleagues also included
doi: 10.1136/bmj.39225.414537.80 non-randomised studies, which, although considered 1  US Department of Health and Human Services. Physical activity and
health: a report of the surgeon general. Atlanta: Centers for Disease
weaker forms of evidence, are necessary to assess the Control and Prevention, National Center for Chronic Disease Prevention
effect of population level interventions such as bike and Health Promotion, 1997.
2  Ogilvie D, Foster CE, Rothnie H, Cavill N, Hamilton V, Fitzsimons CF, et
lanes, walking paths, and recreational areas. al; on behalf of the Scottish Physical Activity Research Collaboration
One non-randomised community intervention in (SPARColl). Interventions to promote walking: systematic review. BMJ
Odense, Denmark, promoted bicycling through many 2007 doi: 10.1136/bmj.39198.722720.BE.
3  Hillsdon M, Foster C, Thorogood M. Interventions for promoting
initiatives and increased the number of bicycle trips by physical activity. Cochrane Database Syst Rev 2007;(2):CD003180.
more than 20% over five years.4 At the same time, the 4  Troelsen J, Jensen SU, Andersen T. Evaluering af Odense—Danmarks
nationale cykelby. Andersen T, Edrén K. Odense: Kerteminde Tryk,
number of accidents involving cyclists was 20% lower 2004.
than in the rest of the country. 5  Cooper AR, Wedderkopp N, Wang H, Andersen LB, Froberg K, Page AS.
Another study found that children who cycled to Active travel to school and cardiovascular fitness in Danish children
and adolescents. Med Sci Sports Exerc 2006;38:1724-31.
school were 8% more fit than children who used other 6  Cooper AR, Page AS, Foster LJ, Qahwaji D. Commuting to school: are
modes of transport including walking.5 It concluded children who walk more physically active? Am J Prev Med 2003;25:273-
6.
that a 10-15 minute session of cycling twice a day would 7  Tudor-Locke C, Ainsworth BE, Adair LS, Popkin BM. Objective physical
be enough to increase aerobic fitness in children.5 activity of Filipino youth stratified for commuting mode to school. Med
Observational studies
�������������������������������������
have consistently shown that Sci Sports Exerc 2003;����������
35:465-71.
8  Manson JE, Hu FB, Rich-Edwards JW, Colditz GA, Stampfer MJ, Willett
children who walk or cycle to school engage in more WC, et al. A prospective study of walking as compared with vigorous
physical activity (other than the travel activity) than exercise in the prevention of coronary heart disease in women. N Engl J
Med 1999;341:650-8.
those who travel by other means.6 7 This
��������������������
extra activity 9  Hu FB, Sigal RJ, Rich-Edwards JW, Colditz GA, Solomon CG, Willett WC, et
may reflect selection (children who are generally more al. Walking compared with vigorous physical activity and risk of type 2
active choose active transport) or it may be that children diabetes in women. JAMA 1999;282:1433-9.
10  Andersen LB, Schnohr P, Schroll M, Hein HO. ��������������������
All-cause mortality
who are encouraged to take up active transport go on associated with physical activity during leisure time, work, sports, and
to engage in other activities. However, because of the cycling to work. Arch Intern Med 2000;�����������
160:1621-8.
lack of cycle lanes in many countries it may be difficult 11  Andersen LB, Boreham CA, Young IS, Davey SG, Gallagher AM, Murray
L, et al. Insulin sensitivity and clustering of coronary heart disease risk
to promote increased cycling for safety reasons. factors in young adults. The Northern Ireland young hearts study. Prev
A weakness in many of the trials of walking interven- Med 2006;42:73-7.

BMJ | 9 june 2007 | Volume 334 1173

 EDITORIALS

Teaching children basic life support skills

Improve outcomes but implementation needs to be earlier and more widespread
Research, p 1201 Basic life support performed by bystanders improves who received such training with those who did not
outcomes in cardiorespiratory collapse, yet less than showed that after five months the trained children were
Ian Maconochie 1% of the general population can perform it effectively. more willing to undertake emergency life saving proce-
consultant in paediatric
emergency medicine, St Mary’s It has been estimated that if 15-20% of the population dures and conducted resuscitation significantly better.3
Hospital, London W2 1NY could perform basic life support, out of hospital mortality Despite these promising results some caveats exist. In
ian.maconochie@st-marys. could be significantly reduced.1 The most effective way both adults and children the skills decline over time, so
nhs.uk
Bob Bingham of achieving this is to teach this technique in schools, refresher courses are needed. A study that repeated the
consultant paediatric anaesthetist making it a “life skill.” training after six months in school aged children found
The Hospital for Sick Children, Great In this week’s BMJ, a study by Jones and colleagues that knowledge was maintained and that the children’s
Ormond Street, London WC1N 3JH
Sheila Simpson assesses the effect of a basic life support programme on resuscitation skills improved.4
senior resuscitation training officer the ability of children to administer effective chest com- “Hands-on” practice is needed to maintain the motor
pressions on a manikin.2 Of the three age groups com- skills required to perform basic life support. Although
Competing interests: None
declared.
pared (9-10, 11-12, 13-14 years), only children aged over additional teaching aids such as online resuscitation train-
Provenance and peer review: 13 years could perform chest compressions to the rec- ing may help with the child’s knowledge, they do not
Commissioned; not externally peer ommended depth of 38-51 mm as effectively as adults. improve the child’s skills.5
reviewed. However, younger children could place their hands in Effective skills can only be attained through high qual-
BMJ 2007;334:1174
the correct position on the chest to perform basic life ity training.6 Poor performance arises from inadequate
doi: 10.1136/bmj.39218.422650.80 support. The authors suggest that younger children could instruction and not allowing sufficient time for the child
use this knowledge to instruct an adult on the appropri- to learn the technique. This includes time for the method
ate technique, despite not being able to do it themselves. to be demonstrated and for the child to practise the tech-
Also, young children could be taught how to assess the nique under adequate supervision.
need for basic life support and activate the emergency High quality teaching can only improve outcomes if
medical services. These conclusions support the teaching uptake is adequate. In countries where teaching basic life
of basic life support to children. support in schools is optional, the uptake of training is
Structured courses such as the “Injury minimisation low. Barriers include funding and time constraints in the
programme for schools” (www.impsweb.co.uk), which “overfull” school curriculum.7 8 Compulsory training is
started in 1994, have integrated the teaching of basic probably necessary to obtain the levels of skill required
life support into the school curriculum (with the support to improve outcomes.
of local hospitals) and have trained more than 114 000 The final barrier to implementing basic life support
children in the United Kingdom. training is lack of resources. Head teachers in Barcelona,
Courses are also taught by the British Red Cross, St sampled in a questionnaire survey, thought that school
John’s Ambulance Service, St Andrew’s Ambulance was the most appropriate setting for teaching these
Service, Heartstart, and Opportunities for Resuscita- skills, and that such training would increase children’s
tion and Citizen Safety (ORCS). In Northern Ireland, self esteem and could potentially save lives. However,
the “ABC for life” programme was set up in 2005 by they identified funding as a potential problem, estimating
the Queen’s University Belfast, with the aim of teaching that the cost would be between €5 (£3.40; $6.80) and
25 000 primary schoolchildren each year. Most of these €10 per child— although this seems a small price to pay
courses focus on teaching children aged 10 years and for improving survival.9
older. At this age children are more likely to be develop- Out of hospital survival from cardiorespiratory col-
ing “abstract thinking” and may be physically capable of lapse could be improved if basic life support was rou-
performing chest compression. tinely taught to all schoolchildren. Introducing it as early
Basic life support courses can change children’s atti- as possible in the school curriculum, perhaps in story
tudes and behaviour. A large study comparing children and online learning formats, would be non-threatening
to young children, who are usually keen to learn and
able to absorb new information. Once they are physi-
cally able, the transition from theoretical knowledge to
practical skills should be relatively easy.
1  Connolly M, Toner P, Connolly D, McCluskey D. The “ABC for life”
programme—teaching basic life support in schools. Resuscitation
2007;72:270-9.
2  Jones I, Whitfield R, Colquhoun M, Chamberlain D, Vetter N,
Newcombe R. At what age can schoolchildren provide effective chest
compressions? An observational study from the Heartstart UK schools
training programme. BMJ 2007 doi: 10.1136/bmj.39167.459028.DE.
3  Frederick K, Bixby E, Orzel M, Stewart-Brown S, Willett K. An evaluation
heartstart uk

of the effectiveness of the injury minimization programme for schools.

Inj Prev 2000;6:92-5.
4  Van Kerschaver E, Delooz H, Moens G. The effectiveness of repeated
cardiopulmonary resuscitation training in a school population.
Resuscitation 1989;3:211-22.

1174 BMJ | 9 june 2007 | Volume 334

 EDITORIALS

Treatment of epilepsy in developing countries

Cheap and effective drugs exist but are not accessible to most patients
Of the 35 million people with epilepsy who live in assigned to phenobarbital. Drug concentrations were
developing countries, around 85% receive no treatment not reported. The two groups were not well balanced
at all.1 2 As a consequence, they experience morbidity for some characteristics; girls were under-represented
related to seizures and the psychosocial consequences in the phenobarbital group, a potentially important
of stigma and discrimination. Regrettably, most of factor because behavioural problems were more
these people—many of whom are children—could have frequent in girls than in boys.
their seizures completely controlled and they could Despite these limitations, the study shows that most
bonnie carlson/usaid

return to a normal life by taking a single daily dose children tolerated phenobarbital well and behaviour
of a drug that costs less than $3 (£1.50; €2.20) each even improved in many. This supports other find-
year.3 In this week’s BMJ, a randomised controlled trial ings in similar settings. In a randomised study of 302
conducted in Bangladesh by Banu and colleagues com- children and adults with epilepsy in rural Kenya, side
pares the effects of carbamazepine and phenobarbital effects were reported more frequently with pheno-
Research, p 1207 on seizure control and behavioural side effects in 108 barbital than with carbamazepine, but the number of
Emilio Perucca professor
children with epilepsy.4 patients with side effects did not differ significantly
Clinical Pharmacology Unit, The World Health Organization recommends phe- between drugs; 3% of patients on phenobarbital
Institute of Neurology, IRCCS C nobarbital as the treatment of choice for partial and were withdrawn for adverse effects and 5% on car-
Mondino Foundation, University
of Pavia, I-27100 Pavia, Italy
tonic clonic seizures in resource restricted countries,5 bamazepine.9 When 73 children with newly diag-
perucca@unipv.it but this policy has been questioned because pheno- nosed epilepsy were randomised to phenobarbital,
Competing interests: EP has barbital is thought to be less well tolerated than other carbamazepine, or valproate in Taiwan, no significant
received speaker’s or consultancy antiepileptic drugs.6 Concerns apply particularly to differences in psychometric scores were found between
fees or research grants from the
manufacturers of carbamazepine children, who are especially vulnerable to this drug’s groups.10 Similarly, no treatment related differences
and oxcarbazepine (Novartis); adverse cognitive and behavioural effects.7 Differ- in behaviour rating scores were found in 94 children
gabapentin, phenytoin, and ences in tolerability between phenobarbital and other with epilepsy randomised to phenobarbital or pheny-
pregabalin (Pfizer); lamotrigine
(GlaxoSmithKline); levetiracetam anticonvulsants are probably less prominent than toin in rural India.11 Observational studies support the
(UCB Pharma); tiagabine, generally thought, however, and they were detected conclusion that phenobarbital is relatively well toler-
valproate, and vigabatrin (Sanofi mostly in trials where the assessment of outcomes ated in developing countries.2 Apart from its low cost,
Aventis); topiramate (Johnson
and Johnson); rufinamide and may have been affected by doctor or patient bias.3 8 phenobarbital has other merits such as efficacy against
zonisamide (Eisai) Importantly, most studies in developing countries did all seizures other than absences, seizure freedom rates
not show excess neuropsychological toxicity of phe- comparable to those associated with modern drugs, a
BMJ 2007;334:1175-6
nobarbital compared with other anticonvulsants,9-11 starting dose within the effective range, a low risk of
doi: 10.1136/bmj.39065.460208.80
possibly because dosages in these studies tended to life threatening adverse effects, linear pharmacokinet-
be lower than those used in developed countries, or ics, once daily dosing, and availability of a parenteral
because lack of options make people less willing to formulation.8
report side effects. Most controlled trials of phenobarbital in epilepsy
The study by Banu and colleagues found no sig- have methodological shortcomings, including an open
nificant difference in behavioural problems such as label design, small sample size, and, at times, ques-
restlessness and hyperactivity between phenobarbital tionable choice of dosing regimens.3 Although larger
and carbamazepine (7% v 11%), and no significant dif- double blind randomised studies are needed for a
ference in psychological and behavioural assessments better assessment of the role of phenobarbital in the
after one year.11 Of those children who completed a treatment of epilepsy,8 Banu and colleagues deserve
12 month follow-up, 47.5% of those on phenobarbital praise for providing more evidence supporting its use
and 60% of those on carbamazepine were seizure-free in resource restricted settings.
for the last six months. The burden of untreated epilepsy in terms of human
Conducting clinical trials in resource restricted coun- suffering and social costs is enormous. Governments
tries is difficult. As with previous similar studies, the and non-governmental organisations in developing
trial by Banu and colleagues has limitations, includ- countries need to ensure that effective treatment is
ing an open label design and low power to detect available for all. Even in these settings, drug choice
potentially important differences in seizure outcome should be tailored to the individual, and phenobarbital
and behavioural test scores. More children were lost will not be the best option for all. In fact, the price of
to follow-up in the phenobarbital group (22%) than drugs is a small part of the cost of ensuring a mini-
in the carbamazepine group (9%). Therefore, on an mum standard of epilepsy care. Dispensing facilities
intention to treat basis, the proportion of children who are often unavailable in remote rural areas, and even
were seizure free in the last six months was consider- when available they often fail to provide a continu-
ably higher in the carbamazepine group than in the ous supply of drugs,12 which has potentially serious
phenobarbital group (50% v 35%), which raises ques- consequences. Seven children in Banu’s study discon-
tions about potentially lower compliance in children tinued treatment for more than seven days for various

BMJ | 9 june 2007 | Volume 334 1175

 EDITORIALS

reasons; four developed convulsive status epilepticus 5  World Health Organization. Initiative of support to people with
epilepsy. Geneva: WHO, 1990 (unpublished document WHO/MNH/
while not taking their drugs. An efficient epilepsy man- MND/90.3).
agement programme will work only if fully integrated 6  Michelucci R, Tassinari CA. Phenobarbital, primidone and other
within a community healthcare delivery system,2 which barbiturates. In: Shorvon S, Perucca E, Fish D, Dodson E, eds. The
treatment of epilepsy. 2nd ed. Oxford: Blackwell, 2004:461-74.
should provide not only reliable supplies of drugs, with 7  Wallace SJ. A comparative review of the adverse effects of
adequate facilities for storage and dispensing, but also anticonvulsants in children with epilepsy. Drug Saf 1996;15:378-93.
8  Kale R, Perucca E. Revisiting phenobarbital for epilepsy. BMJ
educational programmes for health practitioners and 2004;329:1199-200.
the general population. 9  Feksi AT, Kaamugisha J, Sander JW, Gatiti S, Shorvon SD.
. Comprehensive primary health care antiepileptic drug treatment
1  Kale R. The treatment gap. Epilepsia 2002;43(suppl 6):S31-3. programme in rural and semi-urban Kenya. Lancet 1991;337:406-9.
2  World Health Organization, International Epilepsy Bureau and 10  Chen Y-J, Kang W-M, So WCM. ����������������������������������
Comparison of antiepileptic drugs
International League against Epilepsy. Atlas. Epilepsy Care in the on cognitive function in newly diagnosed epileptic children: a
World 2005. Geneva: WHO, 2005. psychometric and neurophysiological study. Epilepsia 1996;37:81-6.
3  Kwan P, Brodie MJ. Phenobarbital for the treatment of epilepsy in the 11  Pal DK, Das T, Chaudhury G, Johnson AL, Neville BG. �����������
Randomised
21st century: a critical review. Epilepsia 2004;45:1141-9. controlled trial to assess acceptability of phenobarbital for childhood
4  Banu SH, Jahan M, Koli UK, Ferdousi S, Khan NZ, Neville B. epilepsy in rural India. Lancet 1998;351:19-23.
Side effects of phenobarbital and carbamazepine in childhood 12  Mac TL, Le VT, Vu AN, Preux PM, Ratsimbazafy V. Antiepileptic drugs
epilepsy: randomised controlled trial. BMJ 2007 doi: 10.1136/ availability and professional practices in delivery outlets in a city
bmj.39022.436389.BE. center in southern Vietnam. Epilepsia 2006;47:330-4.

Control of methamphetamine misuse

Policy should target substance misuse as a whole, rather than
single substances
Methamphetamine is a highly addictive substance that intoxication and heavy metal exposure as a result of
has caused serious public health problems globally.1 mercuric chloride and lead acetate used in the illicit
As it is relatively easy to manufacture from precursor production of methamphetamine.4 5
substances, regulation of precursors has taken centre According to the World Drug Report issued in 2006
stage in global strategies for drug control. Recently, by the United Nations Office of Drugs and Crime,6
the UK Medicines and Healthcare Products Regu‑ around 200 million people used illicit drugs. Ampheta-
latory Agency announced that the precursors pseudo‑ mine-type substances ranked second, after cannabis,
ephedrine and ephedrine, also used in flu remedies with an estimated 35 million users. Of these, 25 mil-
sold over the counter, may in future be available on lion used amphetamines (including methamphetamine)
prescription only.2 while the remaining 10 million used ecstasy. When all
Methamphetamine was first synthesised in Japan indicators of amphetamine production and use were
in 1919 and has been manufactured legally in the combined, the overall global trend was towards a stable
DEA

United States since the 1950s. Use declined dur- to mildly increasing amphetamine market after years
feature, p 1190
ing the 1970s when the public became aware of of annual increases. However, the results of specific
Tracy D Gunter the harms of amphetamines and practitioners were market indicators were mixed, and trends for specific
assistant professor of psychiatry inhibited from prescribing them by the Controlled geographical regions varied.
Psychiatry Research, University of
Iowa Carver College of Medicine, Substance Act (1970).http://www.answers.com/topic/ The report also found that seizures of substances
Iowa City, Iowa 52242, US single-convention-on-narcotic-drugs. However, when diverted for manufacturing amphetamine-type sub-
tracy-gunter@uiowa.edu methamphetamine re-emerged in the 1980s, it had stances reached record levels and exceeded seizures
been transformed into “ice,” a smoked form of high of the end product in 2005.6 The main methampheta-
Competing interests: None
declared. purity that produces sustained intoxication. As it exists mine precursors seized were pseudoephedrine and
Provenance and peer review: today, illicit methamphetamine is manufactured in ephedrine. The main amphetamine precursors seized
Commissioned; not externally peer many forms and may be used in many ways (inhaled, were phenyl-2-propanone and phenylacetic acid.6
reviewed.
ingested, smoked, or injected). Although the rate of dismantling laboratories that
BMJ 2007;334:1176-7 Many definitions of which substances are included produce amphetamines has increased, dismantling
doi: 10.1136/bmj.39225.469630.80 in the class of synthetic stimulants or amphetamine- of large volume international laboratories (so called
type substances exist,3 but generally the class includes superlabs) has not.
amphetamine, methamphetamine, and 3,4 methyl‑ The relation between the regulation of precursor
enedioxymethamphetamine (MDMA or ecstasy). substances and outcomes in drug users, such as hospi-
They cause increased energy, decreased appetite, tal admissions and arrests, has been reported by two
and a heightened sense of wellbeing. The onset studies in the US.7 8 They concluded that regulations
and duration of action vary by specific compound, limiting access to bulk powder and single ingredient
dose, purity, and route of administration. Complica- ephedrine and pseudoephedrine products reduced hos-
tions of use vary greatly and include cardiovascular, pital admissions and arrests. However, regulations tar-
neurological, and psychiatric effects. Other possible geting mixed agent cold remedies used by small scale
complications include risk taking behaviour during manufacturers did not result in similar decreases.

1176 BMJ | 9 june 2007 | Volume 334

 EDITORIALS

So what is the most effective strategy to reduce 1  Office on Drugs and Crime. Ecstasy and amphetamines: global survey,
2003. New York: ODC, 2003. www.unodc.org/pdf/publications/report_
harm from amphetamine-type substances? Although ats_2003-09-23_1.pdf.
the manufacture and misuse of synthetic stimulants 2  BBC News. Crystal meth fears over medicine. 25 April 2007. http://
news.bbc.co.uk/go/pr/fr/-/1/hi/health/6590513.stm .
contribute greatly to morbidity and mortality in sub- 3  Maxwell JC. Methamphetamine: a constantly changing epidemic.
stance users worldwide, the global disease burden EpiLink Online Bull 2007;64:22-8.
of this class of substances is much lower than that 4  Burton BT. Heavy metal and organic contaminants associated with illicit
methamphetamine production. In: Miller MA, ed. Methamphetamine
of tobacco, alcohol, and marijuana.6 9 10 Also, most abuse: epidemiologic issues and implications. NIDA research
people who use amphetamine-type substances take monograph series. Rockville, MD: US Department of Health and Human
Services, Public Health Service, Alcohol, Drug Abuse, and Mental Health
multiple substances.11 Services, 1991:47-59. www.nida.nih.gov/pdf/monographs/115.pdf.
Even if the pattern of drug use is stable over time, 5  National Institute on Drug Abuse. Methamphetamine abuse and
drug markets are dynamic. Efforts to prevent the addiction. Washington, DC: NIDA, 2002 (revised 2006). www.nida.nih.
gov/ResearchReports/Methamph/Methamph.html.
manufacture and use of amphetamine-type substances 6  United Nations Office on Drugs and Crime. World drug report 2006.
should, therefore, be integrated into a rational scheme Vienna: UNODC, 2006. www.unodc.org/unodc/world_drug_report.
html.
to reduce overall substance use that is designed to 7  Cunningham JK, Liu LM. Impacts of federal ephedrine and
tackle existing and emerging drug threats. Over‑ pseudoephedrine regulations on methamphetamine-related hospital
investing resources in the control of one drug, or one admissions. Addiction 2003;98:1229-37.
8  Cunningham JK, Liu LM. Impacts of federal precursor chemical
precursor, carries with it the risk of failing to appre- regulations on methamphetamine arrests. Addiction 2005;100:479-88.
ciate emerging threats. For instance, many people 9  Thomas G, Davis CG. Comparing the perceived seriousness and actual
costs of substance abuse in Canada. Ottawa: Canadian Centre on
fear the “meth crisis,” but fewer seem aware of the Substance Abuse, 2007. http://www.ccsa.ca/NR/rdonlyres/98CA9F87-
recent warnings issued by the UN Office of Drugs 1BE2-40EB-B345-90984F994BFD/0/ccsa0113502007.pdf.
10  WHO. The tobacco health toll. Cairo: WHO, 2005. www.emro.who.int/
and Crime about the resurgence of cocaine in Western TFI/PDF/TobaccoHealthToll.pdf.
Europe.12 11  Gunter TD, Arndt S, Wenman G. Characteristics of admissions for primary
Responses to this crisis should include limiting stimulant dependence during 2001. Subst Use Misuse 2006;41: 
1277-86.
supply and distribution,13 educating the public about 12  United Nations Office on Drugs and Crime. Annual report 2007. Vienna:
harms, screening for early use, and aggressively treat- UNODC, 2007. www.unodc.org/unodc/annual_report_2007.html.
13  Browenstein HH, Taylor BG. Measuring the stability of illicit drug
ing addiction in an integrated approach that tackles markets: why does it matter? Drug Alcohol Depend 14 Dec 2006. Epub
addiction in its many forms. ahead of print.

Clinical trial registration

Looking back and moving ahead
Christine Laine and Richard In 2005, the International Committee of Medical the largest trial registry at the time, contained 13 153
Horton on behalf of the ICMJE Journal Editors (ICMJE) initiated a policy requiring trials; this number climbed to 22 714 one month
working group investigators to deposit information about trial design after the policy came into effect.3 In April 2007, the
claine@mail.acponline.org
into an accepted clinical trials registry before the onset registry contained over 40 000 trials, with more than
See bmj.com for full details of the of patient enrolment.1 This policy aimed to ensure 200 new trial registrations occurring weekly (D Zarin,
ICMJE working group that information about the existence and design of personal communication). The four other registries
Competing interests: Employment: clinically directive trials was publicly available, an that meet the ICMJE criteria have also grown as
FG was previously editorial director ideal that leaders in evidence based medicine have scores of journals have adopted the ICMJE clinical
of Current Controlled Trials, which advocated for decades.2 The policy precipitated much trials registration policy. In response to burgeoning
owned the ISRCTN (International
Standard Randomised Controlled angst among research investigators and sponsors, registration, many investigators, sponsors, and
Trial Number) trials register; SK who feared that registration would be burdensome government agencies have asked the ICMJE to
is employed by the National and would stifle competition. Yet, the response to this recognise their local registries as databases that
Library of Medicine, which pro-
duces ClinicalTrials.gov, but he is not policy has been overwhelming. The ICMJE prom- meet the policy. Fortunately, the World Health
responsible for activities or policies ised to re-evaluate the policy two years after imple- Organization’s (WHO) International Clinical Tri-
regarding ClinicalTrials.gov. Expert mentation. Here, we summarise that re-evaluation, als Registry Platform (ICTRP), which was nascent
testimony: FG. Other: RH is co-chair
and JMD and HCS are members of specifically commenting on registries that meet the when the ICMJE began to require trial registration,
the WHO ICTRP scientific advisory policy requirements, the types of studies that require has matured rapidly and provides options for those
group; MBVDW is a member of the registration, and the registration of trial results. As is who desire a wider array of registries. The ICTRP has
government advisory committee
for the Australian and New Zealand always the case, the ICMJE establishes policy only taken the first steps towards developing a network of
Clinical Trials Registry. for the 12 member journals (a detailed description of primary and partner registers that meet WHO speci-
PS’s affiliation as a representative the ICMJE and its purpose is available at www.icmje. fied criteria.4 Primary registers are WHO selected
and past president of the World org), but many other journals have adopted our initial registers managed by not-for-profit entities that will
Association of Medical Editors
(WAME) does not imply endorse- trial registration recommendations, and we hope that accept registrations for any interventional trials, delete
ment by WAME member journals they will also adopt the modifications discussed in duplicate entries from their own register, and provide
that are not part of the ICMJE. this update. data directly to WHO. Partner registers, which will be
BMJ 2007;334:1177-8 The research community has embraced trial regis- more numerous, will include registers that submit data
doi: 10.1136/bmj.39233.510810.80 tration. Before the ICMJE policy, ClinicalTrials.gov, to primary registers but limit their own register to trials

BMJ | 9 june 2007 | Volume 334 1177

EDITORIALS

process of care changes). Health outcomes include

Summary
any biomedical or health related measures obtained
• In addition to accepting registration in any of the five
in patients or participants, including pharmacokinetic
existing registries, the International Committee of
Medical Journal Editors (ICMJE) will accept registration measures and adverse events. As previously, purely
of clinical trials in any of the primary registers that observational studies (those in which the assignment
participate in WHO’s International Clinical Trials Registry of the medical intervention is not at the discretion
Platform (ICTRP). Registration in a partner register only is of the investigator) will not require registration. The
insufficient ICMJE member journals will start to implement the
• The ICMJE will begin to implement the WHO definition expanded definition of clinically directive trials for
of clinical trials for all trials that begin enrollment on or all trials that begin enrollment on or after 1 July 2008.
after 1 July 2008. This definition states that a clinical trial Those who are uncertain whether their trial meets the
is “any research study that prospectively assigns human
expanded ICMJE definition should err on the side
participants or groups of humans to one or more health-
related interventions to evaluate the effects on health of registration if they wish to seek publication in an
outcomes” ICMJE journal.
• The ICMJE will not consider results posted in the same Over the time during which registration of trial
clinical trials registry in which the primary registration methods has become common practice, several forces
resides to be a previous publication if the results are have begun advocating for registration of trial results.
presented in the form of a brief, structured (<500 words) We recognise that the climate for results registration
abstract or table. will probably change dramatically and unpredictably
over coming years. For the present, the ICMJE will
in a restricted area (such as a specific disease, com- not consider results posted in the same primary clini-
pany, academic institution, or geographic region). cal trials register in which the initial registration resides
The ICMJE strongly supports WHO’s efforts, as previous publications if the results are presented in
through the ICTRP, to develop a coordinated process the form of a brief, structured (<500 words) abstract or
for identifying, gathering, deduplicating, and search- table. The ICMJE favours a standard abstract format
ing trials from registries around the world, thus even- for results reporting, and the CONSORT (Consoli-
tually providing a one stop search portal for those dated Standards for the Reporting of Trials) group’s
seeking information about clinical trials. In addition forthcoming guidelines for abstracts related to trials
to the five existing registries, the ICMJE will now also may be one such option. The ICMJE believes that
accept registration in any of the primary registers that parties interested in results registration should con-
participate in WHO ICTRP. Because it is crucial that sider requiring the deposition of such an abstract in
trial registries are independent of for-profit interests, the registry 24 months after closure of data collection
the ICMJE policy requires registration in a WHO pri- if results are not published in a peer reviewed venue
mary register rather than solely in a partner register, by that time. The registered abstract should either cite
since for-profit entities manage some partner regis- any related full, peer reviewed publications or include
ters. As previously, trial registration with missing or a statement that indicates that the report has not yet
uninformative fields for the minimum data elements been published in a peer reviewed journal. Research-
is inadequate.1 ers should be aware that editors may consider more
Initially, the ICMJE required registration of all clini- detailed deposition of trial results in publicly available
cally directive trials, which it defined as “any research registries to be prior publication. When submitting
project that prospectively assigns human subjects to a paper, authors should fully disclose to editors all
intervention or comparison groups to study the cause posting in registries of results of the same or closely
and effect relationship between a medical intervention related work.
and a health outcome.”1 In May 2005, the ICMJE Three years ago, trial registration was the exception;
clarified this definition to exclude preliminary trials now it is the rule. Registration facilitates the dissemi-
designed to study pharmacokinetics or major unknown nation of information among clinicians, researchers,
toxicity (phase I trials).5 However, the ICMJE rec- and patients, and it helps to assure trial participants
ognises the potential benefit of having information that the information that accrues as a result of their
about preliminary trials in the public domain, because altruism will become part of the public record. WHO’s
these studies can guide future research or signal safety global efforts towards comprehensive trial registration
concerns. Consequently, the ICMJE is expanding the and the ICMJE’s requirements for registration aim to
definition of the types of trials that must be registered increase public trust in medical science.
to include these preliminary trials and adopts WHO’s
1  Clinical trial registration: a statement from the International
definition of a clinical trial, “any research study that Committee of Medical Journal Editors. www.icmje.org/clin_trial.pdf.
prospectively assigns human participants or groups of 2  Simes RJ. Publication bias: the case for an international registry of
humans to one or more health-related interventions clinical trials. J Clin Oncol 1986;4:1529-41.
3  Zarin DA, Tse T, Ide NC. �������������������������������������������������
Trial registration at ClinicalTrials.gov between
to evaluate the effects on health outcomes.”4 Health May and October 2005. N Engl J Med 2005;353:2779-87.
related interventions include any intervention used 4 World Health Organization. International Clinical Trials Registry
to modify a biomedical or health related outcome Platform. www.who.int/ictrp/about/details/en/index.html.
5  International Committee of Medical Journal Editors. Is this clinical
(for example, drugs, surgical procedures, devices, trial fully registered? A statement from the International Committee of
behavioural treatments, dietary interventions, and Medical Journal Editors. www.icmje.org/clin_trialup.htm.

1178 BMJ | 9 june 2007 | Volume 334