You are here

Article Text

Article menu

Download PDFPDF

Epidemiology
Association between smoking and genital warts: longitudinal analysis
  1. Bo Terning Hansen1,
  2. Maria Hagerup-Jenssen1,
  3. Susanne Krüger Kjær2,3,
  4. Christian Munk2,
  5. Laufey Tryggvadottir4,
  6. Pär Sparén5,
  7. Kai-Li Liaw6,
  8. Mari Nygård1
  1. 1Department of Screening-based Research, Cancer Registry of Norway, Oslo, Norway
  2. 2Institute of Cancer Epidemiology, Danish Cancer Society, Copenhagen, Denmark
  3. 3Department of Obstetrics and Gynaecology, Rigshospitalet, University of Copenhagen, Denmark
  4. 4Icelandic Cancer Society, Reykjavik, Iceland
  5. 5Department of Medical Epidemiology and Biostatistics, Karolinska Institute, Stockholm, Sweden
  6. 6Merck Research Laboratories, Merck & Co, Inc, Upper Gwynedd, Pennsylvania, USA
  1. Correspondence to Dr Bo Terning Hansen, Cancer Registry of Norway, PO Box 5313 Majorstuen, N-0304 Oslo, Norway; bo.terning.hansen{at}kreftregisteret.no

Abstract

Objectives To assess the association between smoking and the reported clinical diagnosis of genital warts.

Methods A sample of 58 094 women (aged 18–45) randomly drawn from the general female population of Denmark, Iceland, Norway and Sweden answered a questionnaire on lifestyle and health. Longitudinal data were reconstructed based on self report of age-specific events. In a Cox regression model, women who reported having been clinically diagnosed with genital warts were followed up until the age at first diagnosis, while women who reported never having been diagnosed with genital warts were censored at the age of interview. Age-specific smoking doses and ages at onset of smoking, sexual intercourse, condom use, hormonal contraceptive use, first pregnancy and alcohol drinking were included in the model as time-dependent covariates. The model also included lifetime number of coital partners and country of origin as fixed covariates.

Results Ever-smokers reported a lower age at first intercourse and more coital partners than never-smokers. The adjusted model showed that sexual behaviour strongly influenced the risk of being diagnosed with genital warts, and that smokers in addition had an increased risk compared with non-smokers (adjusted HR=1.27, 95% CI 1.17 to 1.37). There was also a modest additional dose–response effect of smoking, with smokers experiencing a 0.6% increased risk of being diagnosed with genital warts for each additional cigarette smoked daily (adjusted HR=1.006, 95% CI 1.001 to 1.012).

Conclusions Smokers experienced a moderately increased risk of being diagnosed with genital warts. This finding could be explained by the immunosuppressive effects of nicotine, or by confounding not accounted for in the adjusted model.

  • Condyloma acuminata
  • epidemiology
  • human papillomavirus
  • HPV
  • STI
  • risk factors
  • STD

https://doi.org/10.1136/sti.2009.038273

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Genital warts affect some 11% of Nordic women1 and are the most common clinical manifestation of human papillomavirus (HPV) infection of the genitalia. Genital warts are sexually transmitted, and they are caused by low-risk HPV types 6 and 11,2 which do not cause cancer. In women, genital warts are mainly found on the external genitalia and adjacent anal and perianal areas, but may also occur on the cervix and in the vagina. Genital warts are detrimental to quality of life.3 At the population level, there are considerable costs associated with the diagnosis and treatment of genital warts.4

    Numerous studies show that smokers have an increased risk of developing precancerous and cancerous cervical lesions caused by infection with high-risk HPV types.5 Less is known about the influence of smoking on the risk of acquiring genital warts. Some studies report no significant association between smoking and acquisition of genital warts among women,6–8 whereas others report an increased risk of genital warts among female smokers,9–11 ranging up to a fivefold increased risk for current smokers compared with non-smokers.12 Investigations of smoking and clearance/remission of genital warts similarly report mixed results.13–15 The studies examining a potential dose–response effect of cigarette smoking on the risk of acquiring genital warts have used rather crude interval categorisations of current smoking dose or smoking duration and have not taken into account the fact that individual smoking status and intensity often change with time. Such categorisation may mask variation which might be of importance if smoking is associated with the disease. More studies are thus needed to examine the role of smoking for the acquisition of genital warts. Smoking may influence the susceptibility to genital warts through a causal immunosuppressive effect,16 and/or through an association with risk taking that extends to sexual behaviour.17

    The aim of this study is to assess the association between smoking and genital warts. Our analysis takes into account the fact that smoking behaviour may vary with time within subjects. We use a prospective analysis on reconstructed longitudinal data that was collected cross-sectionally.

    Methods

    A random sample of women aged 18–45 from Denmark, Iceland, Norway and Sweden were asked to participate in a survey on lifestyle and health during 2004–5. Data on enrolment and data collection have previously been provided.1 Briefly, in total 69 486 women participated in the study by responding to a self-administered questionnaire (participation rate 66.9%). The women were asked whether they ever had been told by medical personnel that they had genital warts, and if so, the age at first diagnosis. They were also asked about smoking status (ie, current, former or never-smoker) and age-specific smoking doses measured by the number of cigarettes smoked daily in the age intervals <13, 13–16, 17–20, 21–24, 25–29, 30–34, 35–39 and 40–45 years. Further, they were asked about age at first pregnancy, and debut age of smoking, drinking alcohol, sexual intercourse, condom use and hormonal contraceptive use. A total of 11 392 participants did not respond to at least one of these questions, and were hence excluded from the analyses, leaving a total of 58 094 women to be included.

    Since the timing of the first onset of genital warts and several lifestyle parameters were available, we could reconstruct longitudinal data on the reported diagnosis of genital warts and some of its potential predictors. We used a Cox regression model with time-dependent covariates to this end. Integer age was treated as the temporal unit, thus the baseline hazard function describes the effect of age on the risk of being diagnosed with genital warts. Subjects with genital warts changed status from not having to having genital warts at the age when they were first diagnosed. Events subsequent to the first diagnosis of genital warts were not included in the model. Subjects never diagnosed with genital warts were censored at the age of interview. Sexual intercourse, condom use, hormonal contraceptive use, pregnancy and alcohol drinking were each dichotomous and time-dependent covariates. Subjects who experienced these events changed status permanently at the age of onset of the event. Smoking status was also dichotomous and hence changed at the age of onset for smokers, but subjects who stopped smoking before a diagnosis of genital warts or censoring changed status twice. Smoking dose was separately included in the model as the number of cigarettes smoked daily by each individual at each age. A positive dose identifies a smoker at a particular age. A dose of zero could pertain to never-smokers as well as ever-smokers who did not smoke at that particular age, and smokers smoking less than one cigarette a day at that particular age. Lifetime number of coital partners, country of origin and level of education were included as fixed covariates in all multivariate models. Level of education was dropped from the models and is hence not reported since it did not influence model estimates. Modelling was done with R 2.7.2.

    Results

    In total, 6077 of the 58 094 women included in the analyses reported having been clinically diagnosed with genital warts (ie, 10.5%). Ever-smokers and current smokers at the onset of genital warts or censoring constituted 45.2% and 29.3% of the sample, respectively. Mean age of smoking initiation was 16.4 (SD 3.12) years. Ever-smokers generally preceded never-smokers in experiences related to sexual behaviour, as well as in debut age for drinking alcohol. A higher percentage of smokers ever experienced genital warts, sexual intercourse and pregnancy, and ever used condoms, hormonal contraceptives and alcohol. In addition, ever-smokers reported more lifetime coital partners than never-smokers (table 1).

    View this table:
    Table 1

    Characteristics of never- and ever-smoking Nordic women aged 18–45

    The association between smoking and the other behaviours was adjusted for by the inclusion of various time-dependent and fixed covariates in a Cox regression model. This multiple model showed that smokers had an adjusted hazards ratio (HR) of 1.27 (95% CI 1.17 to 1.37) for being diagnosed with genital warts compared with non-smokers (table 2). There was a modest additional dose–response effect of smoking; the risk of being diagnosed with genital warts increased by 0.6% per additional cigarette smoked daily (HR=1.006, 95% CI 1.001 to 1.012).

    View this table:
    Table 2

    Risk of being diagnosed with genital warts among 58094 Nordic women aged 18–45

    The strongest predictors for the risk of being diagnosed with genital warts were associated with sexual behaviour. Having had sexual intercourse and having used condoms and hormonal contraceptives increased the risk substantially. Moreover, the lifetime number of coital partners (a fixed covariate) was positively associated with an increased risk of being diagnosed with genital warts. This effect was strong, as indicated by a very narrow CI. Having drunk alcohol was also associated with an increased risk. In contrast, having been pregnant had a protective effect. Finally, the adjusted model showed that the HR for being diagnosed with genital warts was somewhat higher among women in Iceland, Norway and Sweden than among women in Denmark.

    Discussion

    Smoking was associated with an increased risk of being diagnosed with genital warts in our large randomly selected study population of Nordic women, with smokers experiencing a 27% higher risk than non-smokers. There was a modest additional dose–response relationship between smoking and the risk of being diagnosed with genital warts, the risk increasing by 0.6% for each additional cigarette smoked daily. Hence, a woman in our population-based sample who smoked, for example, 20 cigarettes a day was overall at a 39% greater risk of being diagnosed with genital warts than a non-smoker.

    From our model and from previous studies,6 8 it is obvious that the risk of genital warts is strongly influenced by sexual behaviour. Our data also confirm that smoking is associated with various aspects of sexual behaviour as well as the initiation of alcohol use. Since smoking may be associated with other types of risk-taking behaviour, such as substance use and risky sex,17 it may be difficult to identify the individual contributions of these linked behaviours to the risk of disease. In our adjusted model we included a range of covariates related to sexual behaviour, as well as the onset of alcohol drinking and country of origin, in an attempt to control for potential confounders that may relate both to smoking and to the reported diagnosis of genital warts. This model showed that smoking explained some variation in the risk of being diagnosed with genital warts that was additional to that explained by the other covariates. However, this could result from incomplete control of the confounders of the association between smoking and genital warts.18 Although we included several covariates related to sexual behaviour, our model does not capture all aspects of sexual behaviour that may be of importance for the acquisition of genital warts. For instance, we do not have information on the sexual risk profile of the partners of the women included in this study, nor do we have information on age-specific frequencies of alcohol, condom or hormonal contraceptive use until the time of disease or censoring. Moreover, we include the lifetime number of coital partners as a fixed effect, which is suboptimal since only the partners who were acquired before the onset of genital warts are relevant for the outcome. Reliable information on such detailed aspects of behaviour is hard to obtain from a large sample of participants. If some of this missing information is related to the risk of being diagnosed with genital warts and also is differently distributed between smokers and non-smokers, the effects attributed to smoking in our analyses may in fact be due to differences in other behaviours.

    Alternatively, our documentation of an increased risk for being diagnosed with genital warts among smokers could result from a causal effect of smoking itself. A recent study suggests an independent effect of smoking in the development of HPV-induced invasive cervical cancer.19 Smoking may damage the cervical epithelium20 and impact on the humoral and cell-mediated immune response.16 It changes the density of Langerhans cells,21 T lymphocytes22 and immunomodulating agents23 in the cervical epithelium, and may thus increase the susceptibility to infection and act synergistically with HPV to increase the risk of acquisition of genital warts. The modest dose–response effect documented in this study lends some support to a causal link between smoking and the risk of being diagnosed with genital warts. However, we cannot exclude the possibility that an increase in smoking dose coincides with changes in risk factors not accounted for by our adjusted model.

    The association between smoking and genital warts has not been extensively investigated, and results from different studies vary.6–12 The inconsistencies may be explained by differences in sampling, sample size and analytical approach. The studies reporting the highest risks for smokers have relatively low sample sizes. There does not appear to be a trend for prospective studies to be more or less likely to report an effect of smoking than retrospective studies. Our study adds to the evidence that there is a relationship between smoking and genital warts after adjustment for potential confounders. A previous cross-sectional analysis from the same survey showed a weaker association between smoking and genital warts.1 Like the other studies that have examined this association, it did not take the time-varying nature of smoking behaviour into account. Moreover, it employed a crude lifetime measure of smoking dose rather than the dose acquired before the onset (or not) of disease.

    Women who had had sexual intercourse were at an increased risk of being diagnosed with genital warts, confirming the sexually transmitted nature of the infection. In addition, oral contraceptive use, drinking alcohol and condom use were associated with an increased risk of being diagnosed with genital warts, perhaps owing to more risky and/or a higher sexual activity among those who ever experienced these events than among those who never did so. Moreover, women using condoms may have had partners with a higher risk profile, which could have increased the risk of these women being diagnosed with genital warts, especially since any genital contact is sufficient for the transmission of genital warts and condoms hence offer limited protection. Hormonal contraceptive users may also have been more exposed to HPV if the use of hormonal contraceptives reduced the frequency of condom use. Alcohol consumption24 and oral contraceptive use25 may also have reduced individual immunocompetence, and hence increased the risk of being diagnosed with genital warts. Women who had been pregnant had a reduced risk of being diagnosed with genital warts compared with those who never had been pregnant. Possibly, ever-pregnant women overall had a reduced number of sex partners subsequent to the pregnancy, and/or partners with a lower risk profile, and hence experienced a somewhat reduced exposure to HPV. We also found that the HR for being diagnosed with genital warts differed somewhat between the countries, and that it was highest among Icelandic women. Some of this variation might be explained by differences in sexual behaviour, since Icelandic women report more coital partners.1

    This study has a large sample size and a high participation rate considering the sensitive nature of the questions asked. A further strength of this study is the registration of complete individual smoking histories. This gives a detailed account of smoking behaviour that allows for inclusion of temporal aspects in the consideration of smoking as a potential predictor for genital warts. Using lifetime or current measures of smoking status or intensity, as has been the rule, may obscure real effects with respect to the event of acquiring/not acquiring genital warts, since individual smoking status and intensity often vary over time. Here, we specifically examine time-dependent behaviour until the onset of disease (or age at interview for women who reported never having been diagnosed with genital warts).

    In order to minimise misclassification of disease, we asked whether participants had been diagnosed with genital warts by medical personnel, thereby avoiding potential imprecision introduced by self diagnosis. However, some of the women may have had undiagnosed cases of genital warts. Vaginal and cervical warts may be particularly likely to have escaped notice among women who have rarely had gynaecological examinations. Since smoking may be more common among individuals with a relatively low socioeconomic status,26 and low socioeconomic status could be associated with reduced access to certain healthcare services,27 there is a possibility that undiagnosed cases of genital warts may have occurred more frequently among smokers than among non-smokers in our sample. If so, we may have underestimated the effect of smoking on the risk of being diagnosed with genital warts. However, since socioeconomic status in a Nordic setting does not appear to be a strong predictor of the propensity to visit a general practitioner,28 or to attend screening for cervical cancer,29 the potential for a bias in undiagnosed cases of genital warts between smokers and non-smokers may be limited in our study.

    The main limitation of this study is that it is based on self report of retrospective events. Although collected cross-sectionally, most of the data used in this investigation are longitudinal, reconstructed by the participants' reports of the timing of events. The validity of this approach hinges on the respondents' ability to recall the retrospective events of interest with reasonable accuracy. Moreover, self report of topics that implicitly relate to social norms, like the occurrence of genital warts, may be erroneous. Recall of age-specific smoking doses is perhaps particularly susceptible to imprecision, but information on individual smoking histories is very hard to obtain by other means than self report. Some studies show that recall of smoking history,30 including age at smoking initiation,31 is reliably accounted for by self report. Moreover, current smoking status generally seems to be reliably reported.32 33 Mistakes in the relative ordering of smoking initiation and the onset of genital warts is unlikely to represent a major problem here, since the average age of smoking initiation and the first onset of genital warts were widely separated. Smoking initiation is largely a teenage phenomenon while the first onset of genital warts tends to occur somewhat later. However, an ideal design to examine the association between genital warts and smoking would be a prospective study in which smoking, sexual behaviour and disease status were closely tracked to minimise recall error and misclassification of disease.

    We conclude that smokers experienced a moderately increased risk of being diagnosed with genital warts. This finding could be explained by the immunosuppressive effects of nicotine, or by confounding not accounted for in the adjusted model.

    Key messages

    • The risk of being diagnosed with genital warts was strongly influenced by sexual behaviour among women from the general population in Denmark, Iceland, Norway and Sweden.

    • Smokers experienced an additionally increased risk of being diagnosed with genital warts.

    • The risk of being diagnosed with genital warts was modestly affected by smoking dose.

    • The associations between smoking and the risk of being diagnosed with genital warts may be due to biological effects of smoking, or residual confounding with sexual behaviour.

    Acknowledgments

    We thank Björn Hagmar and Ole Klungsøyr for comments on a previous draft of this paper.

    References

    1. ↵
      1. Kjær SK,
      2. Tran TN,
      3. Sparen P,
      4. et al
      . The burden of genital warts: a study of nearly 70.000 women from the general female population in the 4 Nordic countries. J Infect Dis 2007;196:1447–54.
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Aubin F,
      2. Pretet JL,
      3. Jacquard AC,
      4. et al
      . Human papillomavirus genotype distribution in external acuminata condylomata: a large French national study (EDiTH IV). Clin Infect Dis 2008;47:610–15.
      OpenUrlAbstract/FREE Full Text
    3. ↵
      1. Woodhall S,
      2. Ramsey T,
      3. Cai C,
      4. et al
      . Estimation of the impact of genital warts on health-related quality of life. Sex Transm Infect 2008;84:161–6.
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Hillemanns P,
      2. Breugelmans JG,
      3. Gieseking F,
      4. et al
      . Estimation of the incidence of genital warts and the cost of illness in Germany: a cross-sectional study. BMC Infect Dis 2008;8:76.
      OpenUrlCrossRefPubMed
    5. ↵
      1. Rajkumar T,
      2. Appleby P,
      3. Beral V,
      4. et al
      . Carcinoma of the cervix and tobacco smoking: collaborative reanalysis of individual data on 13,541 women with carcinoma of the cervix and 23,017 women without carcinoma of the cervix from 23 epidemiological studies – international collaboration of epidemiological studies of cervical cancer. Int J Cancer 2006;118:1481–95.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Habel LA,
      2. Van den Eeden SK,
      3. Sherman KJ,
      4. et al
      . Risk factors for incident and recurrent condylomata acuminata among women – a population-based study. Sex Transm Dis 1998;25:285–92.
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Wen LM,
      2. Estcourt CS,
      3. Simpson JM,
      4. et al
      . Risk factors for the acquisition of genital warts: are condoms protective? Sex Transm Infect 2008;75:312–16.
      OpenUrl
    8. ↵
      1. Garland SM,
      2. Steben M,
      3. Sings HL,
      4. et al
      . Natural history of genital warts: analysis of the placebo arm of 2 randomized phase III trials of a quadrivalent human papillomavirus (types 6, 11, 16, and 18) vaccine. J Infect Dis 2009;199:805–14.
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Daling JR,
      2. Sherman KJ,
      3. Weiss NS
      . Risk-factors for condyloma acuminatum in women. Sex Transm Dis 1986;13:16–18.
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Brisson J,
      2. Roy M,
      3. Fortier M,
      4. et al
      . Condyloma and intraepithelial neoplasia of the uterine cervix – a case-control study. Am J Epidemiol 1988;128:337–42.
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Munk C,
      2. Svare EI,
      3. Poll P,
      4. et al
      . History of genital warts in 10.838 women 20 to 29 years of age from the general population: risk factors and association with Papanicolaou smear history. Sex Transm Dis 1997;24:567–72.
      OpenUrlPubMedWeb of Science
    12. ↵
      1. Feldman JG,
      2. Chirgwin K,
      3. Dehovitz JA,
      4. et al
      . The association of smoking and risk of condyloma acuminatum in women. Obstet Gynecol 1997;89:346–50.
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Wilson JD,
      2. Brown CB,
      3. Walker PP
      . Factors involved in clearance of genital warts. Int J STD AIDS 2001;12:789–92.
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. Coremans G,
      2. Margaritis V,
      3. Snoeck R,
      4. et al
      . Topical cidofovir (HPMPC) is an effective adjuvant to surgical treatment of anogenital condylomata acuminata. Dis Colon Rectum 2003;46:1103–8.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Lacey CJN,
      2. Goodall RL,
      3. Tennvall GR,
      4. et al
      . Randomised controlled trial and economic evaluation of podophyllotoxin solution, podophyllotoxin cream, and podophyllin in the treatment of genital warts. Sex Transm Infect 2003;79:270–5.
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Arcavi L,
      2. Benowitz NL
      . Cigarette smoking and infection. Arch Intern Med 2004;164:2206–16.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Zweig JM,
      2. Lindberg LD,
      3. McGinley KA
      . Adolescent health risk profiles: the co-occurrence of health risks among females and males. J Youth Adolesc 2001;30:707–28.
      OpenUrlCrossRefWeb of Science
    18. ↵
      1. Phillips AN,
      2. Smith GD
      . Cigarette smoking as a potential cause of cervical cancer: has confounding been controlled? Int J Epidemiol 1994;23:42–9.
      OpenUrlAbstract/FREE Full Text
    19. ↵
      1. Kapeu AS,
      2. Luostarinen T,
      3. Jellum E,
      4. et al
      . Is smoking an independent risk factor for invasive cervical cancer? A nested case-control study within nordic biobanks. Am J Epidemiol 2009;169:480–8.
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Simons AM,
      2. Vanherckenrode CM,
      3. Rodriguez JA,
      4. et al
      . Demonstration of smoking-related DNA damage in cervical epithelium and correlation with human papillomavirus type 16, using exfoliated cervical cells. Br J Cancer 1995;71:246–9.
      OpenUrlPubMedWeb of Science
    21. ↵
      1. Barton SE,
      2. Jenkins D,
      3. Cuzick J,
      4. et al
      . Effect of cigarette smoking on cervical epithelial immunity: a mechanism for neoplastic change? Lancet 1988;2:652–4.
      OpenUrlPubMedWeb of Science
    22. ↵
      1. Szarewski A,
      2. Maddox P,
      3. Royston P,
      4. et al
      . The effect of stopping smoking on cervical Langerhans' cells and lymphocytes. Br J Obstet Gynaecol 2001;108:295–303.
      OpenUrlCrossRefWeb of Science
    23. ↵
      1. Scott ME,
      2. Ma YF,
      3. Farhat S,
      4. et al
      . Covariates of cervical cytokine mRNA expression by real-time PCR in adolescents and young women: effects of Chlamydia trachomatis infection, hormonal contraception, and smoking. J Clin Immunol 2006;26:222–32.
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      1. Szabo G
      . Alcohol's contribution to compromised immunity. Alcohol Health Res World 1997;21:30–8.
      OpenUrlPubMedWeb of Science
    25. ↵
      1. Cherpes TL,
      2. Marrazzo JM,
      3. Cosentino LA,
      4. et al
      . Hormonal contraceptive use modulates the local inflammatory response to bacterial vaginosis. Sex Transm Infect 2007;84:57–61.
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Barbeau EM,
      2. Krieger N,
      3. Soobader MJ
      . Working class matters: socioeconomic disadvantage, race/ethnicity, gender, and smoking in NHIS 2000. Am J Public Health 2004;94:269–78.
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Sørensen TH,
      2. Olsen KR,
      3. Vedsted P
      . Association between general practice referral rates and patients' socioeconomic status and access to specialised health care a population-based nationwide study. Health Policy 2009;92:180–6.
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Grøholt EK,
      2. Stigum H,
      3. Nordhagen R,
      4. et al
      . Health service utilization in the Nordic countries in 1996: influence of socio-economic factors among children with and without chronic health conditions. Eur J Public Health 2003;13:30–7.
      OpenUrlAbstract/FREE Full Text
    29. ↵
      1. Eaker S,
      2. Adami HO,
      3. Sparen P
      . Reasons women do not attend screening for cervical cancer: a population-based study in Sweden. Prev Med 2001;32:482–91.
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Brigham J,
      2. Lessov-Schlaggar CN,
      3. Javitz HS,
      4. et al
      . Reliability of adult retrospective recall of lifetime tobacco use. Nicotine Tob Res 2008;10:287–99.
      OpenUrlAbstract/FREE Full Text
    31. ↵
      1. Kenkel D,
      2. Lillard DR,
      3. Mathios A
      . Smoke or fog? The usefulness of retrospectively reported information about smoking. Addiction 2003;98:1307–13.
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Patrick DL,
      2. Cheadle A,
      3. Thompson DC,
      4. et al
      . The validity of self-reported smoking – a review and metaanalysis. Am J Public Health 1994;84:1086–93.
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      1. Vartiainen E,
      2. Seppala T,
      3. Lillsunde P,
      4. et al
      . Validation of self reported smoking by serum cotinine measurement in a community-based study. J Epidemiol Community Health 2002;56:167–70.
      OpenUrlAbstract/FREE Full Text

    Footnotes

    • Funding Merck & Co Inc (grant number: EPO 8014.016).

    • Competing interests None.

    • Ethics approval This study was conducted with the approval of the national data protection boards and ethics committees in Denmark, Iceland, Norway and Sweden.

    • Provenance and peer review Not commissioned; externally peer reviewed.