Table 1

Characteristics of the 58 included studies

Study year (reference)Adolescents (n)Age of adolescents (years)Source of sampleNOS score for qualityStudy designConfounders allowed for*
Asbridge et al37340013–19School-based (30 metropolitan schools), Toronto Youth Crime and Victimisation Survey, Canada7Cross-sectionalA, B, C, E
Bauman et al38827313–17School-based, National Longitudinal Study of Adolescent Health, USA7LongitudinalA
Bergamaschi et al39269116School-based (2nd year high school students), Romagna, Italy6Cross-sectionalA
Bricker et al4055208School-based (40 schools),Washington State, USA6LongitudinalA
Cornelius et al4156714Family-based survey (from prenatal clinic), Magee Women's Hospital, Pittsburgh, USA8LongitudinalA, J
Den Exter Blockland et al42220610–14School-based (Dutch youth), The Netherlands7Cross-sectionalA, B, C, E
Elder et al4366011–16School-based, Hispanic adolescents, San Diego, California, USA6Cross-sectionalA, B, J
Fidler et al4465011–16School-based, (36 schools), HABITS Study, South London, UK8LongitudinalA, B, C, D
Fisher et al4510 59312–18Family study, Growing Up Today Study (GUTS), NHS 2 Study, USA6Cross-sectionalA, B, J
Forrester et al46364112–17Community study (rural area), Oregon, USA7LongitudinalA, B, E, J
Foster et al47215314Family study, Dartmouth, North England6LongitudinalA, C, J, S
Gilman et al4855912–17Family study, New England, USA7LongitudinalA, B, D
Griesbach et al710 50015WHO cross-national study of health behaviours, health and its social context in children and adolescents, Europe and North America7Cross-sectionalA, B, C, S
Griesler et al49116514–21National Longitudinal Survey of Youth (NLSY), African, Hispanic and white mothers and children7LongitudinalA, C, D, E, J, S
Gritz et al50100410–18School-based (6 schools), Houston-Galveston, Texas, USA6LongitudinalC, D, J
Hesketh et al51667413–18School-based, Zhejiang Province, Eastern China3Cross-sectional
Hollis et al52252614–17Clinic-based, adolescent primary care patients in medical office waiting rooms, Oregon, USA6Cross-sectionalA
Jackson53122011–14School-based (4 middle schools), Central Carolina, USA6Cross-sectionalA, B, D
Jackson and Dickinson5459417.1 (mean)School-based (1 element school out of 12), Child-to-Adolescent Panel Study (CAPS), North Carolina, USA6LongitudinalA, B, D
Johnson et al5536548–14School-based (24 public schools), The Child and Adolescent Trial for Cardiovascular Health (CATCH), from 4 geographical regions, USA7LongitudinalA, B, D
Johnson et al56476314–15School-based (22 schools), Acadiana Coalition of Teens against Tobacco (ACTT), 6 Louisiana parishes, USA7Cross-sectionalA
Kalesan et al5737 22411–18School-based (315 schools), Maryland, USA6Cross-sectionalB
Kelishadi et al5842712–20Isfanah, Iran7Case-ControlB, C, E, S
Komro et al59128213–16Community study, Minnesota, Tobacco Free Future Project (TFFP), USA6Cross-sectionalA
Kristajanson et al60743014–16School-based, youth in Icelandic secondary schools, Iceland6Cross-sectionalC, E, J, S
Malcon et al61118710–19School-based (urban area), Pellet, Southern Brazil6Cross-sectionalA, B
Martini and Sulistyowati62163013–18School-based, two regions and two cities, East Java Province, Indonesia4Cross-sectional
Menezes et al63445210–12Community-based, a cohort of children born in 1993 and followed and later sought 2004–57LongitudinalA, B, C, D, E
Milton et al642479–11School-based (6 primary schools), Liverpool, UK7LongitudinalA, B, D, J
Molyneux et al65652211–16School-based (10 secondary schools, 50 state secondary schools), Nottinghamshire, UK6Cross-sectionalA, B, C, J
Molyneux et al10210913–16School-based (10 secondary schools), from 50 state public secondary schools, Nottingham, UK5Cross-sectionalB, J
Moore et al66124615–16School-based (55 secondary schools), Wales, UK6Cross-sectionalA, B, E, J
Nichols et al6785812–15School-based, (30 schools), New York, USA6LongitudinalA, D
O'Byrne et al6881615.1 (mean)School-based (2 high schools and 2 junior high schools), Kansas City, Metropolitan area, USA7Cross-sectionalA, B, D
O'Callaghan et al6945415–14Mater-University, Queensland Study of pregnancy (MUSP): a cohort of women at first antenatal visit6LongitudinalA, C, S
Ogwell et al70113012–17School-based (10 schools from 209 primary schools), Nairobi Province, Kenya7Cross-sectionalA, B, E
Otten et al71549511–16School-based (33 schools), The Netherlands7LongitudinalA, B, E
Ozawa et al72201215–18Schools-based, senior high school, Fukuoka City, Japan7Cross-sectionalA, E
Pärna et al73404913–18School-based, Tallinn, Moscow7Cross-sectionalA, C, S
Pust et al5129812–17Study-Drug Affinity of Young People in Federal Republic of Germany7Cross-sectionalA, E
Rajan et al7439628–9School-based (20 schools), Washington State, USA7LongitudinalA
Rozi et al7577214.8 (mean)School-based, government and private schools, Karachi, Pakistan6Cross-sectionalA, D
Sargent and Dalton763728–17School-based (3 rural Vermont), Dartmouth, USA7LongitudinalA, C, E
Sasco et al77365011–17School-based, public urban schools, France6LongitudinalA, B, S
Scragg et al7829 27114–15School-based (4th form), New Zealand5Cross-sectionalA, B, C
Scragg and Laugesen7928 68914–15School-based (year 10 students), National Surveys of Year 10 Students, New Zealand7Cross-sectionalA, B, D
Shamsuddin and Haris8024415–16School-based (6 Secondary Schools), Kota Bharu, Kelantan, Malaysia7Cross-sectionalA, E, J
Simons-Morton et al81345111–14School-based (7 middle schools), Maryland School district, Washington DC, USA6Cross-sectionalA, B, D
Siziya et al84183813–15School-based, Zambia Global Youth Tobacco Survey (GYTS) 2002 data, Chongwe district, Zambia6Cross-sectionalA, J
Siziya et al8323236–16School-based, Global Youth Tobacco Survey (GYTS) 2003 data, Kilimanjaro region, Northern Tanzania6Cross-sectionalA, C, J
Siziya, et al82179013–15School-based (public and private junior high school), data from East Timor Global Youth Tobacco Survey (GYTS)6Cross-sectionalA, B, C, J
Szabo et al85412512–17School-based (68 schools), Victoria, Australia7Cross-sectionalA, B, E
Tyc et al8623712–18School-based (1 large junior school and 2 senior high schools), Memphis, USA7Cross-sectionalA, B, C
Vink et al87359812–15Netherlands Twin Register (NTR)5Longitudinal
Wen et al88395711–17School-based (6 secondary schools out of 18 in cluster sampling), Huangpu district, Guangzhou, South China7Cross-sectionalA, B
Withers et al89215014–16Family study: Family Health Services Authority (FHSA) and the Office of Population Consensus Surveys (OPC)5Cross-sectionalA, S
Yorulmaz et al9088311–19School-based (28 middle and high school), Edirne, Turkey6Cross-sectionalA, B, C
Zhang et al91276313.6 (mean)School-based (4 junior high schools), Henan Province, China7Cross-sectionalA, B, C
  • * Potentially important confounders allowed for: A, age; B, gender; C, measure of socioeconomic status; D, ethnicity/race; E, educational achievement; J, peer smoking; S, measure of family structure; –, performed no adjustment.

  • NOS, Newcastle–Ottawa Scale.