Role of snus (oral moist snuff ) in smoking cessation and smoking reduction in Sweden

Authors

  • Hans Gilljam,

    Corresponding author
    1. Department of Public Health Sciences, The Karolinska Institute, Stockholm, Sweden

      Hans Gilljam
      Centre for Tobacco Prevention
      Box 17533
      118 91 Stockholm
      Sweden
      Tel: + 46 8 51 77 80 56
      Fax: + 46 8 51 77 80 72
      E-mail: hans.gilljam@smd.sll.se
    Search for more papers by this author
  • M. Rosaria Galanti

    1. Department of Public Health Sciences, The Karolinska Institute, Stockholm, Sweden
    Search for more papers by this author


Hans Gilljam
Centre for Tobacco Prevention
Box 17533
118 91 Stockholm
Sweden
Tel: + 46 8 51 77 80 56
Fax: + 46 8 51 77 80 72
E-mail: hans.gilljam@smd.sll.se

ABSTRACT

Aims  To assess to what extent snus has been used as an aid to stop smoking among Swedish smokers.

Design  A random telephone retrospective survey of Swedish smokers and ex-smokers.

Setting  Survey conducted in November–December 2000.

Participants  A national sample of 1000 former and 985 current daily smokers aged 25–55 years.

Measurements  Smoking status, date and method of quitting by self-report.

Findings  Thirty-three per cent of former smokers and 27% of current smokers had ever used snus. The difference was larger among men (55% versus 45%, P = 0.003). Current smokers who made use of snus smoked on average fewer cigarettes per day than non-users of snus. The mean duration of abstinence among former smokers was not influenced by snus use. Conditionally on age, education and use of nicotine replacement therapy there was an increased probability of being a former rather than a current smoker with ever use (OR 1.72, 95% CI = 1.30–2.28) or current use (OR 1.81, 95% CI = 1.31–2.53) of snus. Having used snus at the latest quit attempt increased the probability of being abstinent by about 50% (OR 1.54, 95% CI = 1.09–2.20).

Conclusions  Our study suggests that by using snus, Swedish male smokers may increase their overall chances of abstinence. However, 71% of the men in this sample who quit smoking did so without using snus and the duration of abstinence was not affected by snus use. This suggests that snus is not a necessary component of smoking cessation at the population level. Snus use was very rare among women.

INTRODUCTION

Long-term cigarette smoking is a threat to human health. Among smokers, nicotine addiction and psychological factors are potent behaviour maintaining forces, beyond any rational consideration and even the desire to quit (APA 1994). In fact, the proportion of long-term abstinent smokers is still limited in most clinical settings and many smokers find it impossible to stop completely (Piasecki & Baker 2001). This has raised an interest in strategies of harm reduction which has been defined as: ‘A product is harm-reducing if it lowers total tobacco-related mortality and morbidity even though use of that product may involve continued exposure to tobacco-related toxicants’ (Stratton et al. 2001). It has been argued that the amount of risk reduction either might not be quantifiable or, in the absence of a proxy outcome, would require a very large study (thousands of subjects) and at least 8 years from inception to results (Hughes 2000). The discussion on harm reduction has evolved further (Kozlowski et al. 2001; Stratton et al. 2001) and it has been concluded that more research is needed to establish whether modified cigarettes, snus or other nicotine-containing products can improve public health.

Modifications to cigarettes, such as adding filters and introducing brands with less tar and nicotine, has not been shown to reduce harm (Stratton et al. 2001). Because most tobacco-related harm is caused by inhaled components of tobacco smoke, unburned tobacco products are considered relatively safer. An intuitive approach to harm reduction would therefore be to substitute inhaled nicotine with nicotine from other sources. Nicotine replacement therapy (NRT) involves a highly regulated pharmaceutical product, which can secure a stable reduction of the amount smoked among some regular smokers (Bolliger 2000). Little is known about the harm-reducing potential of the unregulated products, such as the many varieties of chewing tobacco, snus, snuff, etc.

A real-life long-term test of the concept of harm reduction has been ongoing in Sweden with little or no scientific monitoring (Henningfield & Fagerström 2001). Snus (oral moist snuff) has been produced in Sweden since the early 19th century. It was exclusive to Sweden and emerged as the preferred tobacco product for working-class men. During the 20th century cigarette smoking grew in popularity, and snus use declined. As a consequence, the dominant manufacturer considered terminating production. From the early 1970s, however, Swedish Match targeted young males and athletes as a new group of snus consumers and marketed the product intensively. Since then the sales have increased and snus is now used daily by about 20% of Swedish adult males, across all social groups. The levels of consumption are typically high (on average, ½−1 box a day). This is equivalent to the nicotine content of 35–75 cigarettes per day with high and sustained levels of blood nicotine (Holm et al. 1992). Due to the lack of consistent associations with major diseases, snus is considered to be less harmful than cigarettes (Huhtasaari et al. 1992; Bolinder et al. 1994; Lewin et al. 1998; Schildt et al. 1998; Huhtasaari et al. 1999; Ye et al. 1999; Lagergren et al. 2000), and few users seek help for their addiction. Furthermore, snus is less expensive than cigarettes and an average user will spend about half of the 36 SEK (3.50 USD) paid for a pack of cigarettes on a box of snus.

Sweden was the only industrialized country in the world to reach the goal set by WHO for the year 2000 of less than 20% adult smokers (Statistics Sweden 2001). It has been suggested that this low prevalence has been achieved through the widespread use of snus by Swedish men, but very little evidence has been presented to substantiate this assumption (Henningfield & Fagerstrom 2001; Ramstrom 2000).

Because of this historical background Sweden appears to be a natural setting to address questions on harm reduction at the population level (Henningfield & Fagerstrom 2001). If snus use would indeed be associated with an increase in smoking quit rates and/or with a reduction of smoking among current smokers, this would lend further strength to the suggested role of snus in a strategy of ‘harm reduction’. In the present study, we analysed the reports of snus use in a national sample of current and former daily smokers.

METHODS

On behalf of the Swedish Cancer Society and Pharmacia Inc., a telephone survey was conducted between 20 November and 17 December 2000 in order to recruit two equal-sized national samples of at least 1000 former and current smokers. All residents in Sweden between the age of 25 and 55 years, who currently smoked tobacco daily, or had done so in the past, were eligible for an in-depth interview. The latter aimed to describe social and behavioural characteristics of the two groups, with the purpose of identifying cues for prevention campaigns. In total 7883 individuals were screened, of whom 2715 fulfilled the criteria for inclusion. A total of 2002 individuals (1000 former and 1002 current smokers) were interviewed successfully, corresponding to 74% of the eligible sample. Reasons for non-participation among the 713 subjects who were not interviewed were: poor health (2.7%), non-Swedish speaking (10.3%) and refusal (87.0%).

Current smokers in the survey were defined as subjects either smoking at least one cigarette/other tobacco product a day, or smoking on at least seven occasions per week and using oral snuff. This analysis is based on 985 daily smokers. Former smokers (n = 1000) were defined as subjects having smoked daily for at least 6 months in their lives and having been continuously abstinent for at least 4 weeks by the time of the interview. Exclusive users of snus were identified, but not interviewed.

A structured interview was conducted, encompassing slightly different questions in the two groups. Smokers were interviewed about their consumption of cigarettes and/or other smoked tobacco, previous attempts to quit smoking, intention to quit in the future, motivations to quit, opinions on cigarettes price, past and current use of snus and of NRT and corresponding motivations for initial and continued use.

Former smokers were asked about previous quit attempts and duration of the present period of abstinence, but not on previous smoking behaviour. Use of snus and NRT was also investigated, with the same set of questions as for current smokers. Both groups were asked to rate the support received from different sources (e.g. social network, health professionals) during their previous or current abstinence periods, as well as their degree of agreement with tobacco regulation and policy.

Statistical analysis

In univariate analyses of categorical variables we compared the proportion using snus (ever, current and last quit attempt use), as well as other relevant characteristics among current and former smokers. The χ2 statistic was used to examine the departure of the observed distribution from that expected under the null hypothesis, with 5% level for statistical significance (P < 0.05). The means of continuous numerical variables were compared between two groups assuming a normal distribution (Student's t-test). When comparing the means of continuous variables in subgroups defined by several explanatory factors we used the multiple factors analysis of variance (ANOVA) for unbalanced design (Kleinbaum & Kupper 1978). Multiple regression analyses were also conducted to study the association between snus use and the odds of being a former, rather than a current smoker, conditionally on other potential confounding factors. As the study dealt with a dichotomous outcome variable, we derived the association estimates (odds ratios, OR) and their corresponding confidence limits (95% CI) by means of logistic regression based on maximum likelihood equations (Breslow & Day 1980).

RESULTS

The study population

Among 1002 current smokers, 985 (98.3%) were daily smokers. Of these, 927 (94%) smoked cigarettes exclusively, while virtually all the remaining 6% smoked cigarettes as well as other products (three subjects did not provide information on type of tobacco).

There were some significant socio-demographic differences between former and current smokers (Table 1). Former smokers were on average older, better educated and with higher income than current smokers. In addition, more former than current smokers lived in stable relations (spouse or cohabitant). Current smokers smoked on average 13.2 (SD 7.8) cigarettes or other tobacco products per day. The mean duration of abstinence among former smokers was 139.5 months (SD = 104.1, range 1–432).

Table 1.  Socio-demographic characteristics of current and former smokers in a national sample, Sweden 2000.
 Current
smokers
Former
smokers
n%n%
Gender
 Men42443.048848.8
 Women56157.051251.2
 χ21 d.f. = 6.616, P = 0.01
Age (years)
 25–29 78 7.9 39 3.9
 30–3411812.0 73 7.3
 35–3912512.713713.7
 40–4416316.617017.0
 45–4919319.624224.2
 50–5530831.333933.9
 χ25 d.f. = 31.192, P < 0.0001
Education
 Compulsory school26727.321922.1
 Senior high school49050.045045.3
 College or above22222.732432.6
 χ22 d.f. = 25.400, P < 0.0001
Living in couple
 No67368.381981.9
 Yes31231.718118.1
 χ21 d.f. = 48.986, P < 0.001
Annual income (USDth)
 <1940140.725825.8
 19–3837738.341641.6
 >3820721.032632.6
 χ22 d.f. = 59.407, P < 0.0001

Use of snus

In total, more former than current smokers had ever used snus and used it currently (Table 2). This difference was particularly pronounced and statistically significant only among men. Among women, only 141 subjects (13%) had ever used snus, and 31 (3%) were current users at the time of the interview. Among subjects who made at least one attempt to quit smoking, slightly more former than current smokers used snus during the latest attempt (for ex-smokers this corresponded to the current abstinence period). This difference, however, did not attain the conventional statistical significance among men (Table 2). Large gender differences were also recorded concerning current use and use at the latest attempt to quit smoking. Among current snus users, former smokers tended to report higher snus consumption than did current smokers (data not shown).

Table 2.  Number and proportion (%) of current and former smokers reporting use of snus.
 All subjectsMen
n%n%
  • a

    Only smokers with previous quit attempts included.

Ever use
 Current smokers26927.319044.8
 Former smokes32932.926754.7
χ2 1 d.f. = 7.37, P = 0.007χ2 1 d.f. = 8.90, P = 0.003
Current use
 Current smokers 9810.0 8419.8
 Former smokers15815.814128.9
χ2 1 d.f. = 15.12, P < 0.001χ2 1 d.f. = 10.07, P = 0.002
Use at the latest quit attempt
 Current smokersa 9712.3 7523.0
 Former smokers16316.314028.7
χ2 1 d.f. = 5.84, P < 0.016χ2 1 d.f. = 3.25, P = 0.072

An overview of the main reason for starting using snus, reported by 598 ever-users, is displayed in Table 3. The most common single citations included direct or indirect reference to harm reduction, without significant differences between the two groups. The intention to reduce/quit smoking to motivate the use of snus was given more often by women than by men (P = 0.03), especially among current smokers.

Table 3.  Main reason for starting using snus reported by current and former daily smokers who ever used snus.
ReasonaCurrent smokers (%)Former smokers (%)
  • a

    Only reasons given by at least 1% of the respondents in each group are listed.

  • b

    The most frequent answers to the open-ended question fell into three groups: curiosity (28% of CS and 41% of FS); aid to quit/substitute for smoking (17% of CS and 8% of FS); social pressure (15% of CS and 14% of FS).

 1. Help to stop smoking19.721.3
 2. Smoking reduction 7.1 4.6
 3. Health concern: snus less dangerous than smoking 1.5 2.4
  1 + 2 + 328.328.3
 4. Suggested by relatives or friends 2.6 2.4
 5. Offered/suggested during the military service 1.9 0.6
 6. Offered/suggested within sport teams/activities 1.1 0.3
 7. Alternative to NRT 0.7 1.5
 8. Snus cheaper than smoking 1.5 0.9
 9. Smoking forbidden on the work-place 3.4 1.8
10. Smoking forbidden in other public places 2.2 0.9
11. Smoking became uncomfortable 1.1 1.2
12. Was offered to try, found it pleasant, social circumstances13.417.6
13. Otherb27.928.3

Smoking cessation or reduction

The mean number of cigarettes smoked per day was lower among smokers using snus than among smokers who did not, particularly among men (Table 4). The proportion of current smokers smoking less than 10 cigarettes/day was nearly twice as high among users of snus than among non-users (44% versus 24%, respectively, data not shown). Moreover, among smokers of less than 20 cigarettes a day there was an inverse relation between amount of snus consumed in a week and number of cigarettes smoked in a day (data not shown).

Table 4.  Amount of daily smoking among current smokers and duration of abstinence among former smokers according to use of snus.
Current smokersMean no. of cigarettes/daySDSignificance
  • a

    Multiple factors ANOVA, including age, gender and NRT use.

  • b

    Multiple factors ANOVA, including age and NRT use.

All subjects
 Current use of snus 11.1 10.1P < 0.05a
 No current use of snus 13.4  7.4 
Men
 Current use of snus 11.0  9.8P < 0.052
 No current use of snus 15.1  8.7 
Former smokersMean duration of abstinence (months)
All subjects
 Ever used snus131.4103.5P > 0.05a
 Never used snus143.5104.3 
 Snus in the current abstinence period127.5103.7P > 0.05a
 No snus in the current abstinence period142.0104.1 
Men
 Ever used snus138.2105.7P > 0.05b
 Never used snus157.1108.3 
 Snus in the current abstinence period134.6107.3P > 0.05b
 No snus in the current abstinence period151.6106.9 

The mean duration of abstinence among former smokers was, if anything, longer among never users of snus than among ever users, although this difference did not attain the statistical significance. The same pattern was seen with use of snus during the latest (current) abstinence period (Table 4). Among 140 former male smokers who used snus at the beginning of the current abstinence period (latest attempt) 69% were current users (median duration of abstinence 72 months).

Table 5 reports the adjusted odds ratios of being a former, rather than a current, daily smoker according to use of snus. Conditionally on age, education and use of NRT there was a 70–80% increased probability to be a former rather than a current smoker with past or current use of snus. Having used snus at the latest attempt to quit smoking increased the probability of being currently abstinent by about 50%. These estimates were not modified appreciably after further adjustment for income and family status.

Table 5.  Odds ratio (OR) and confidence interval (95% CI) of being a former smoker among men, according to use of snus.
Use of snusORa95% CI
  • a

    Adjusted for categories of age, education and NRT use (past, current, none).

  • b

    Reference category.

  • c

    Including all former smokers and smokers with previous quit attempts.

Ever use
 NoRef b
 Yes1.721.30–2.28
Current use
 NoRef b
 Yes1.811.31–2.53
At the latest quit attemptc
 NoRef b
 Yes1.541.09–2.20

DISCUSSION

To date only one report of non-medicinal nicotine as an aid to stop smoking has been described in the scientific literature (Tilashalski et al. 1998). Our study is the first study where current smoking status has been investigated in relation to smokeless tobacco (snus) use in a large population sample.

Our understanding of the dynamics of tobacco use in Sweden is incomplete. The spontaneous shift from smoking to smokeless tobacco use was studied in the 1980s in a national panel interviewed during two household surveys (Tillgren et al. 1996). Over 8 years one in four of the 27% daily smokers identified by the time of the first interview had quit tobacco altogether, whereas one in 20 had shifted to snus, and only one in 50 had added snus to an already existing smoking habit. It appears that smokers who quit smoking in the 1980s turned to snus less often compared to the ones interviewed in the present study. Comparisons are, however, hampered by different recall strategies, calendar periods and age groups on which the two studies are based.

The present study explores another aspect of the complex patterns of tobacco use in Sweden, namely whether the use of snus among smokers is linked to the probability to quit smoking. To this end, current and former daily smokers at a given point in time may be seen as the cross-section of a cohort of smokers, some of which have achieved successfully a prolonged abstinence from smoking. First, it was evident in this sample that use of snus with the intention to quit smoking is widespread, although almost exclusively among men. Nearly one-third of male ex-smokers used snus regularly and as many used it during the process of quitting smoking. Among current smokers with a history of relapse more than one-fifth had used snus during their latest attempt to quit.

Secondly, significantly more male ex-smokers had ever used snus and still used it regularly compared to current smokers. After controlling for socio-demographic characteristics and use of NRT, this corresponded to a modest, but significantly increased probability of sustained smoking abstinence for users of snus compared to non-users. Thirdly, persistent smokers who used snus reported lower smoking levels than exclusive smokers.

Our study also confirms the large gender differences in the use of smokeless tobacco seen in population surveys, both in Sweden and the United States (Horn et al. 2000). In this sample women used snus to a much lower extent than men (although at a much higher rate than the general female population), and their current smoking status was apparently not associated with snus use. The determinants of this gender difference are unclear. The motives behind snus use are an unlikely explanation as in this study women were, if anything, more prone than men to interpret their initial motivation to start using snus in terms of harm reduction. It is possible that men and women who add snus use as a further recreational drug belong to very different groups of smokers with regard to nicotine dependence, motivation to quit, social circumstances, smoking-related morbidity, etc. Further studies are warranted to clarify these issues.

An additional, unexpected finding in this study was the agreement between current and former smokers in the alleged motives for their initial use of snus. In particular, the proportions referring, directly or indirectly, to the willingness to quit or cut down smoking for health purposes were strikingly similar. This indicates that a consciously instrumental use of snus is neither the main nor the only mechanism through which the process of quitting smoking is facilitated. In fact, a larger proportion of quitters than of smokers reported having started using snus because of curiosity, prompted by social circumstances.

Finally, it is worth noting that the overwhelming majority of successful quitters of both genders had never used snus. The duration of abstinence among these never users was, if anything, longer than among users. This lack of correlation between use of snus and duration of abstinence suggests that a majority of Swedish smokers attempting to quit is able to achieve long-lasting abstinence without the help of any nicotine-delivering product. Societal norms and increased awareness of the dangers associated with smoking are clearly at play. This is supported by experiences from other nations. In the 2001 Australian national survey, the prevalence of daily smokers was fairly comparable to that in Sweden (21.1% among men) without availability of other major sources of non-medicinal nicotine (AIHW 2001). Our study, on the other hand, indicates that using snus as an aid to quit smoking may be conducive to prolonged nicotine dependence. For instance, a large majority of former smokers who used snus as aid to cessation in the beginning of the current abstinence period were still using it years later.

This study had several limitations. Selection processes might have affected current and former smokers differently. For instance, there might have been a selection of former smokers with an a priori higher propensity to quit (e.g. more motivated or less dependent). Further, it is possible that the early quitters who did not turn to snus were better motivated, smoked less or were otherwise different from the ones who used snus to aid their successful quit attempt. Similar differences in outcomes between groups were observed in the surveys conducted in England and California and in both surveys the authors concluded that subgroups of more motivated and/or less addicted smokers do better with less help (Zhu et al. 2000; Buck & Morgan 2001) .

As we had no information on former smokers’ smoking levels or motivation to quit prior to the current quit attempt, we could not control for these aspects in the analysis. However, in order to inflate the association between snus use and successful abstinence spuriously, this selection should also have affected differently snus users and non-users, which seems unlikely.

Biases due to differential recall of snus use, of duration of abstinence among former smokers or of cigarette consumption among current smokers also seem improbable.

Reports of smoking or of abstinence in this sample were not verified with biochemical methods. However, in population surveys, misclassification of smoking status because of unreliable reports does not appear to be a major problem (Patrick et al. 1994). In addition, if this misclassification were not differential across subgroups of snus use it would bias the estimated association towards the null.

One of five Swedish men use snus on a daily basis, making this tobacco product more common than cigarette smoking among males. Whether the increasing consumption of snus in Sweden has been a decisive factor in the reduction of smoking, and the declining trends in the incidence of lung cancer and other tobacco-related diseases, remains to be elucidated (Stenbeck & Rosén 1995). Our study suggests that by using snus, Swedish male smokers may have increased their overall chance of abstinence by 70–80%. However, smokers using snus in order to quit smoking are probably a minority, even disregarding the gender imbalance. In the present sample, three of four Swedish males who quit smoking in the 1980s and 1990s reported doing so without using snus. Also, the duration of abstinence did not increase by the use of snus. This suggests that snus, however appealing as a tool for risk reduction, is certainly not a necessary component of smoking cessation at the population level.

ACKNOWLEDGEMENTS

The data collection for this study was conducted by TEMO on behalf of the Swedish Cancer Society and Pharmacia AB. We thank all three organizations and especially Mrs Britt-Marie Lindblad and Mrs Maria Lohm for making the data available.

Ancillary