IJE Advance Access originally published online on June 25, 2007
International Journal of Epidemiology 2007 36(4):789-804; doi:10.1093/ije/dym039
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Circulatory disease and smokeless tobacco in Western populations: a review of the evidence
PN Lee Statistics and Computing Ltd., 17 Cedar Road, Sutton, Surrey, SM2 5DA, UK. E-mail: PeterLee{at}pnlee.co.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
Background Use of oral snuff or ‘snus’ has risen in Sweden. Sales of snuff in the US have also risen, overtaking sales of chewing tobacco. There is some evidence that nicotine contributes to circulatory disease (CID) from smoking. We therefore reviewed the evidence relating smokeless tobacco (ST) to CID and related risk factors.
Methods Publications that described relevant cohort, case-control and cross-sectional studies were identified from in-house files, a Medline search in December 2005 and reference lists. Relative risks (RRs) and odds ratios (ORs) for ischaemic heart disease, stroke and all CID for ST use, stratified by smoking habit, were estimated and combined by meta-analysis to provide an overall RR estimate. For diabetes, increased blood pressure, and other risk factors, evidence was qualitatively reviewed, with results from clinical studies also considered.
Results ST use in non-smokers was associated with an increased risk of heart disease (RR 1.12, 95% CI 0.99–1.27, n = 8), stroke (1.42, 1.29–1.57, n = 5) and CID (1.25, 1.14–1.37, n = 3). The increases mainly derived from two large US studies. The Swedish studies provided little evidence of an increase for heart disease (1.06, 0.83–1.37, n = 5) or stroke (1.17, 0.80–1.70, n = 2), although the estimates by country are not notably heterogeneous, even for stroke (P = 0.29). No dose–response was evident. No increase was seen in former users of ST, or in ST users who also smoked. No clear relationship to diabetes was seen. In the US, an acute blood pressure rise following ST use was consistently reported, and isolated reports linked specific risk factors to ST. In Sweden, though one study reported that snuff acutely increased blood pressure, and two linked snuff to Raynaud-type symptoms, the overall evidence for an effect was inconclusive. Swedish studies generally showed no chronic effect of snuff on blood pressure or various risk factors.
Conclusions Any CID risk from ST appears to be substantially less than from smoking, and no clear risk from Swedish snuff is seen. However, the overall evidence is limited.
Keywords Heart diseases, cerebrovascular disorders, tobacco, smokeless, diabetes mellitus, blood pressure
Accepted 14 February 2007
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
Smokeless tobacco (ST) is mainly used orally, and nasal use has become rare.1–3 The two major products used in North America and Europe are chewing tobacco and snuff. There are several types of chewing tobacco and snuff, differing in their formulation and how the tobacco is treated.
In the United States, ST has formed an important part of total tobacco consumption for many years, e.g. 11.2% of all tobacco products by weight in 1950, 9.6% in 1980 and 12.9% in 2000.4 For many years sales of chewing tobacco were two or three times those of snuff, but since the early 1980s sales of snuff have risen sharply and have recently overtaken sales of chewing tobacco.4 Partly this may be because some types of fine-cut ST classified as chewing tobacco prior to 1981 are now categorized as moist/fine-cut snuff.1
In most other economically developed countries, ST forms only an unimportant part of the tobacco market.4 Though ST forms a few per cent of the market in Canada, Iceland and Norway, the most notable exception is Sweden where snuff (‘snus’) has always formed a large proportion of total sales of tobacco by weight (70% in 1920, 31% in 1950, 29% in 1980, 53% in 2000), though sales of chewing tobacco there are negligible. As in the US, the use of snuff in Sweden has increased sharply in recent decades.
ST is also widely used in parts of Central and South-East Asia. Use occurs in various ways, with the tobacco used alone or in combination with other products, such as betel nut quid, slaked lime, areca nut and even snail shells.1–3
Studies conducted in Sweden5–12 and the US13–17 and also in the UK18 provide evidence that the use of snuff and chewing tobacco involves an exposure to nicotine that is fairly equivalent to that from cigarette smoking. Indeed some of the studies in Sweden5–10 reported higher plasma cotinine levels in users of snuff. Since nicotine has been implicated in several processes related to cardiovascular disease,19 the possibility that ST, like cigarette smoking, might result in an increased risk, therefore deserves consideration. In this context, however, it should be noted that pipe smoking, which also results in nicotine exposure similar to that from cigarette smoking, is associated with an increase in heart disease risk that is substantially lower than that for cigarette smoking.20,21
This review is concerned with evidence relating ST use to circulatory disease (CID). Attention is restricted to chewing tobacco and snuff, with no consideration of the limited data on nicotine chewing gum. Attention is also restricted to studies in Western populations (in practice this means US and Sweden), omitting studies in Central and South-East Asia where, as noted above, products used are variable and often involve the inclusion of non-tobacco material.
The review concentrates on evidence from epidemiological studies linking use of ST to mortality from or onset or prevalence of ischaemic heart disease (IHD), acute myocardial infarction (AMI), stroke, or all CID. A briefer summary is also given of evidence linking use of ST to risk factors for heart disease such as diabetes, hypertension, cholesterol, fibrinogen, platelet function and ultrasonographic measurements of atherosclerosis.
An earlier review by Critchley and Unal22 which concluded that ‘there may be an association between ST use and cardiovascular disease’ was limited, partly because it did not include meta-analyses, partly because studies of stroke were not included and partly because it only included three23–25 of the eight epidemiological studies now available on IHD or AMI. Four were published in 200526–28 and one earlier.29
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
In December 2005, publications describing results of epidemiological case-control, cohort and cross-sectional studies relating ST use to risk of CID were sought from our in-house files, supplemented by a MEDLINE search (using the terms smokeless tobacco and cardiovascular disease) and from reference lists of articles retrieved. Publications relating to risk factors relevant to CID were similarly sought, with clinical studies also considered.
From the studies of CID risk, details were extracted of the study design, location and timing and of the potential confounding variables considered. Where available, estimates of the relative risk (RR) or odds ratio (OR) associated with ST use were extracted together with their associated 95% confidence interval (CI). The RRs and ORs extracted related to AMI (international classification of diseases (ICD) 9th revision 410), IHD (ICD 410–414, equivalent to coronary heart disease), stroke ICD 430–438 = cerebrovascular disease], all CID (390–459 = cardiovascular disease), or other CID. Attention was limited to estimates where the groups being compared differed only in exposure to ST use and not in exposure to smoking. Thus, if for example the source article presented RRs for each combination of ST use (ever, never) and smoking (current, former, never) relative to a totally unexposed denominator (never ST, never smoking), the RRs were recalculated, using standard methods,30–32 to compare ever and never ST users separately for current, former and never smokers. Fixed-effects meta-analysis33 would then be used to combine these estimates to produce a smoking-adjusted overall RR estimate. Fixed-effects meta-analysis was also used to combine RRs reported separately by age. Where RRs for a given comparison were available with varying degrees of confounder adjustment that adjusted for the most variables was selected for analysis.
In order to obtain an overall estimate of the risk associated with ST use, RR and OR estimates were combined by meta-analysis.33 Since an OR is an approximation to an RR, the combined estimate is referred to as being a RR estimate. Results are presented in the tables for both fixed-effects and random-effects models, with the random-effects estimates being cited in the text where appropriate.
The studies of risk factors related to CID were reviewed in a less detailed manner, mainly to summarize the number of studies which did and did not report associations of ST use with the various endpoints considered.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
CID
The studies
Table 1 summarizes details of five cohort studies, two in Sweden23,27 and three in the US,26,29 for which results have been reported relating ST use to risk of CID mortality or incidence. For the study of Swedish construction workers,23 some earlier results had been reported34 based on a cross-sectional analysis of the baseline sample. All of the studies presented results for males, with only one29 also presenting results for females. All of the studies involved follow-up for at least 10 years. The number of ST users considered in the analytical cohort followed up varied from 437 to 7745. Only one of the cohort studies, CPS-II,35 separated current and former use of ST, with CPS-I35 and the two Swedish studies23,27 only considering current ST use and the NHANES follow-up study29 only considering ever ST use. Smokers were excluded in the analyses of CPS-I and CPS-II,35 with the analyses of the Swedish construction workers study23 excluding those who had used pipes or cigars and those with mixed use of ST and cigarettes. All of the cohort studies took a range of potential confounding variables into account, though in one study23 detailed analyses were only presented adjusted for age and region, it being noted that additional adjustments had little effect on the risk estimates.
|
Table 2 similarly summarizes details of the five case-control studies that have been carried out, all in Sweden.24,25,28,36,37 Four of the five studies were conducted in the two northernmost counties, Norrbotten and Västerbotten. Two of these24,25 were non-overlapping case-control studies of AMI, while the other two36,37 were overlapping nested case-control studies of stroke, both based on new cases occurring in a combined sample from the same two (MONICA and VIP) studies, one based on cases occurring between 1985 and 1996, the other also including cases up to 2000. The other case-control study,28 of AMI, was conducted in two other Swedish counties, Stockholm and Västernorrland. With the exception of the earlier study of stroke, all of the case-control studies restricted attention to males. All of the cases were first occurrences of the index disease (AMI or stroke), with controls representative of the population studied. While four of the studies matched controls to cases on a 1:1 or 2:1 basis on age, region and other factors, the earliest study of AMI24 did not, selecting controls in the same age range and population. The three studies of AMI24,25,28 included fatal and non-fatal cases, relying on surrogate interviews for the fatal cases, but not for their matched controls. The number of cases in the studies varied between 589 and 1432 for AMI and between 123 and 276 for stroke. All the case-control studies reported results relating to current and former snuff use, with the exception of the earlier study of stroke,36 which considered only current use. Consideration of potential confounding variables was variable, with one study36 only presenting crude results for sexes combined, and two studies25,37 reporting results adjusted for a variety of factors.
|
|
Risk associated with smokeless tobacco use
Virtually all the evidence relating ST use to risk of CID is for males. There are only two exceptions to this:
- In females in the First NHANES follow-up study29 there was little evidence that ST increased risk of death from all CID (hazard ratio 1.2, 95% CI 0.7–1.9 in never smokers) after adjustment for age, race and poverty, or that ST increased risk of IHD (1.4, 0.8–2.2 in never smokers; 1.1, 0.4–3.2 in ever smokers) or stroke (1.0, 0.3–2.9 in never smokers; 1.7, 0.4–7.0 in ever smokers) after adjustment for a longer list of potential confounding variables (Table 1);
- In males and females combined in the first MONICA/VIP study36 the crude OR for stroke was not increased (0.56, 95% CI 0.25–1.27).
This evidence is clearly too sparse to allow a reliable conclusion, so, in the following, attention is restricted to males.
Table 3 presents RR or OR estimates relating to various indices of ST use, smoking groups and vascular endpoints. The first estimate listed is an OR deriving from a cross-sectional study34 which formed the baseline of a cohort study,23 while estimates 2–25 are RRs from cohort studies and estimates 26–43 are ORs from case-control studies. Of the 43 estimates, the lowest is 0.56 and the highest 1.89, with 23 estimates >1.00 and 19 < 1.00. Considered in isolation, the strongest statistical evidence of a positive relationship comes from estimates 1, 2 and 4 in the Swedish construction workers study, 14, 15 and 17 in CPS-I and 18, 19 and 21 in CPS-II.
Meta-analyses
Table 4 presents the results of various meta-analyses based on the estimates in Table 3. The first three of these are presented as forest plots in Figure 1.
|
|
For IHD or AMI, the eight RR/OR estimates for current ST use among never smokers (or the nearest equivalent available) showed some heterogeneity (
2 = 13.11 on 7 df, P = 0.07), varying from 0.60 to 1.41. The random-effects estimate was not clearly increased overall (1.12, 95% CI 0.99–1.27), or when results were considered separately for the US (1.14, 0.96–1.34) or Sweden (1.06, 0.83–1.37). Excluding estimate 9, which was for ever rather than current use, the random-effects estimate for Sweden and US combined changed slightly to 1.15 (1.03–1.29). Estimates were increased for cohort studies but not for case-control studies (heterogeneity 2 = 4.58 on 1 df, P = 0.03). Although the combined estimates in Table 4 provide some indication of a weak association of IHD or AMI with current ST use in never smokers, they provided no indication of any increase in risk associated with current ST use in current smokers or in all individuals, or of any increase in risk associated with former ST use.
For stroke, the five estimates for current ST use among never smokers showed no heterogeneity (2 = 3.53 on 4 df, P = 0.47) with the meta-analysis showing an increased risk (1.42, 1.29–1.57). The increase seemed less evident for Sweden (1.17, 0.80–1.70) than for the US (1.41, 1.17–1.71), but there was no clear heterogeneity (2 = 1.12 on 1 df, P = 0.29).
For all CID, the three estimates for current ST use among never smokers showed no marked heterogeneity (2 = 4.58 on 2 df, P = 0.10), and the random-effects estimate was increased (1.25, 1.13–1.37).
The analyses of risk associated with current ST use in never smokers were heavily dependent on the contributions of the two large US CPS studies. Thus, for IHD or AMI, CPS-I provided 62.3% and CPS-II 17.0% of the total weight of the combined estimate. Corresponding percentages were 76.3% and 16.2% for stroke and 68.2% and 18.5% for all CID, respectively. Excluding CPS-I and CPS-II the random-effects estimates showed little evidence of an association for IHD or AMI (0.99, 0.76–1.30) or for stroke (1.11, 0.78–1.59).
Risk by age
Although there is clear evidence that the association of smoking with IHD is age-dependent, with an increasing RR with decreasing age, only two of the 10 studies presented RRs or ORs for ST by age. In the Swedish construction workers study,23 the risk of IHD associated with current ST use in never smokers was higher (P = 0.02) in men aged 35–54 at entry into the study (RR 2.0, 95% CI 1.4–2.9) than in men aged 55–65 at entry (RR 1.2, 95% CI 1.0–1.5). In contrast, in the first MONICA study,24 the risk of AMI was somewhat lower (P = 0.31) in men aged 35–54 (OR 0.96, 95% CI 0.56–1.67) than in men aged 55–64 (OR 1.24, 95% CI 0.78–2.30).
Dose response
Only two studies presented any dose–response results for ST use. In CPS-II,26 frequency and duration of use among current ST users was unrelated to risk of IHD, stroke or all CID, despite there being over 200 cases of CID in ST users. In the first MONICA study,24 users of 2+ cans of snuff a day had no increased risk of AMI, though power to detect a dose–response was limited.
Separate associations with chewing tobacco and snuff
In CPS-II,26 RRs for IHD, stroke and all CID did not differ notably between those who chewed tobacco and never used snuff and those who used snuff and never chewed. However, the number of deaths in the second category was relatively low.
Effect of smoking
All of the seven studies that reported little or no evidence of an increased risk of CID associated with ST use24,25,27–29,36,38 reported a clear association of CID with smoking, with RRs typically exceeding 2. Similarly in the Swedish construction workers study,23 the RRs for IHD, stroke and all CID were always higher for smokers than for ST users. The source article26 for the ST results in CPS-I and CPS-II did not give results for smoking, but other publications39–41 made it clear that in these studies the RRs for smoking exceeded those seen for ST use.
Joint effects of smokeless tobacco use and smoking
Five studies24,25,27–29 provided estimates of the RR or OR of IHD or AMI associated with current (or ever) ST use separately for never (or non-current) smokers and for current smokers. As shown in Table 5, the smoker/non-smoker ratio of these estimates was close to 1 and provided no indication of a multiplicative interaction between the effects of smoking and the possible effects of ST. Thus, the data did not indicate any special hazard in relation to the dual use of smoking and ST.
|
Other outcomes
Diabetes
Ten publications9,10,26–28,42–46 have reported results relating the use of ST to diabetes or related endpoints such as glucose intolerance, glucose levels, insulin levels or insulin resistance. Except for the publications reporting results from the large American CPS-I and CPS-II studies,26 all the studies described were conducted in Sweden.
Of the six studies that provided results for diabetes incidence, prevalence or mortality (Table 6), the only notable association was reported by Persson et al.44 That study reported an OR of 3.9 (95% CI 1.1–14.3) comparing current exclusive snuff users and those who had never used tobacco and an OR of 2.7 (95% CI 1.3–5.5) in current heavy snuff users, regardless of tobacco consumption. These ORs did not compare diabetic patients with individuals with normal glucose tolerance. All the studies involved a small number of cases of diabetes in ST users, at most 20 in the CPS-I study. For the combined data, no clear association was seen, random-effects meta-analysis for current ST use giving estimates of 1.16 (95% CI 0.72–1.87, n = 5) for never smokers and 1.39 (95% CI 0.95–2.02, n = 3) for studies reporting all individuals regardless of smoking habit. The inconclusive results contrast with those for smoking where a clear association with diabetes risk has been reported in a number of large cohort studies.47–49
|
Two44,46 of the three studies of diabetes also presented further results for glucose intolerance, but neither reported any notable relationship with either snuff use or smoking.
No effect of snuff use or smoking was seen on fasting glucose levels in three studies9,10,45 or on glucose levels during a euglycaemic clamp in another study.42 One study10 reported no effect of snuff or smoking on post-load glucose levels in men, but did observe some decrease with smoking in women. The clamp study42 reported that basal glucose utilization was reduced by smoking but unaffected by snuff use.
The evidence relating to insulin levels following fasting is unclear. Evidence of increases relating to snuff and smoking was seen in two studies.9,45 In contrast, no increase was seen in an experimental study42 or in either sex in another study.10 Indeed, women who smoked had a clearly lower fasting insulin than non-users in that study, which also reported somewhat lower post-load insulin levels in male snuff users and smokers of either sex. Insulin levels during a euglycaemic clamp were unaffected by snuff use or smoking.42
In an experimental study42 smoking, but not snuff use, was found to impair insulin action, mainly due to a lower peripheral glucose uptake. A cross-sectional study44 found that, in men with impaired glucose tolerance, the odds of a low insulin response were increased in smokers and in former but not current snuff users. No such association was seen in individuals with normal glucose tolerance.
Blood pressure
In the US, nine experimental studies,13,14,50–56 mainly of crossover design, consistently reported an acute rise in blood pressure following use of ST. They also consistently reported an acute rise in heart rate. The studies13,14 comparing acute effects of smoking with ST found them to be similar in respect of effects on heart rate and blood pressure. An acute rise in blood pressure following ST use has also been noted in some case reports emanating from the US.57–60 Some authors14,60–62 have suggested that sodium or liquorice in ST may contribute to the blood pressure increase, though a direct effect of nicotine is more commonly referred to. Two studies,55,56 in which the controls were given, as a placebo, snuffs that did not contain tobacco and were nicotine-free, observed no acute rise in blood pressure.
In Sweden, only one experimental study,63 of crossover design, has investigated acute effects of oral snuff. Clear rises in blood pressure and heart rate were seen during rest, but not during exercise. Smoking was not investigated in this study.
A number of cross-sectional studies have been carried out comparing blood pressure in ST users and non-users. Six such studies have been conducted in the US29,50,64–67 and nine in Sweden.8–10,12,27,28,34,45,68 Four of the six US studies29,50,64,65 provided evidence of an increased blood pressure in ST users. However, in two of these studies,29,65 no age adjustment was carried out despite the ST users being markedly older than the non-tobacco users, and in one64 the increase related only to men aged 18 to 25, with no results reported for older men, or for women. Only one of the nine studies in Sweden34 reported increased blood pressure in ST users. This was the study of construction workers, noted earlier to have reported an association with cardiovascular mortality and incidence of MI.
Other CID risk factors
Eight publications from the US29,56,66,69–73 and 11 from Sweden6,7,9,10,12,28,34 provided information relating ST use to a range of other CID-related endpoints.
The material from the US included two case reports of Buerger's disease associated with the use of chewing tobacco,69,70 one study which reported that smoking but not ST was associated with reduced exercise performance,71 one study which reported an increased plasma lactate concentration and decreased stroke volume associated with ST use73 and one study56 which reported that snuff use was associated with an increase in mean arterial pressure and epinephrine, but not peripheral vascular resistance, muscle sympathetic nerve activity or norepinephrine. Two studies reported results relating to lipid levels. One of these studies72 reported that risk of hypercholesterolaemia was increased in both smokers and ST users. The RR in ST users implausibly increased from 1.08 to 2.51 after adjustment for potential confounding factors. A second study66 found no notable association of ST use with total cholesterol or HDL cholesterol levels, but did find some evidence of a reduction in HDL levels with increasing serum cotinine. One US study29 reported a number of differences (including lower blood cholesterol and higher dietary fat intake) between exclusive ST users and those who had never used tobacco. However, these analyses, though based on a large sample, were not adjusted even for age and sex.
The studies in Sweden concerned a range of endpoints, but the results generally showed a consistent picture. Apart from two reports34,74 of a similar increase in prevalence of Raynaud-type symptoms in both ST users and smokers, a report of an increase in triglycerides45 in ST users that was less marked than that seen in smokers, and a report28 that current snuff users were more likely to be overweight, the studies all showed little evidence of an adverse effect of ST use for endpoints that are virtually all associated with smoking. This conclusion applies to endpoints such as atherosclerosis,7,45 response to exercise,6,12 cholesterol levels,7,9,28,45,68 triglycerides,7,9,68 fibrinogen,7,9,10,68 markers of platelet activity12 and levels of antioxidant vitamins.75 Exceptionally, one study10 reported no effects of either ST use or smoking on cholesterol levels and triglycerides.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
This review was carried out to investigate in detail the epidemiological and clinical evidence relating CID and risk factors for CID to use of ST in Western populations. Since the use of ST, whether in the US or Sweden, involves an exposure to nicotine that is quite comparable with that from cigarette smoking, since cigarette smoking is associated with an increased risk of CID, and since nicotine has been implicated in several processes related to the disease, there is concern that the use of ST also might increase risk. Although, in view of the much lower heart disease risk in pipe smokers, such concern might not be fully justified, it was decided to review the available evidence relevant to this concern.
CID
Based on 10 epidemiological studies of mortality or disease onset (Tables 1 and 2), three conducted in the US and seven in Sweden, meta-analyses have been conducted based on the results for men, the available data for women being extremely limited. Among never smokers, the combined evidence (Table 4) gave some indication of a weak association of IHD or AMI with current ST use in never smokers, but no indication of any increase in risk in current smokers or with former ST use. The associations with stroke and with all CID among never smokers were somewhat more clearly seen.
The analyses have a number of strengths. These include quite a substantial total number of cases in the meta-analyses, availability of estimates adjusted for a wide range of relevant potential confounding variables in the majority of the studies and long-term follow-up in all the cohort studies. Limitations include variations in the definitions of exposure and of disease endpoints between study, small numbers of individuals and lack of confounder control in a few studies, the fact that five of the seven Swedish studies were conducted in the same area of Sweden and particularly the lack of good recent data from the US.
The appropriateness of carrying out meta-analysis at all deserves some comment. While recognizing that there is an active discussion about the use of meta-analysis to pool data from observational studies,76 and that the uncertainty about the meta-analytic estimates may not be fully expressed by the conventional statistical indices, failure to combine estimates might have produced conclusions that are more uncertain than the data warrant. The meta-analyses were conducted by combining RRs from cohort studies and ORs from case-control studies. ORs somewhat overestimate RRs where, as here, the disease is not rare,77 but the overestimation is of little practical importance for ORs in the range considered78 (e.g. for a disease with an incidence of 10% an OR of 1.20 is equivalent to an RR of 1.18) and has been ignored.
The first meta-analysis shown in Table 4 combines results for fatal and non-fatal AMI. In theory, effects of ST may differ for the two types, but in practice only two of the studies investigated this possibility. The two counties’ study28 reported that such analysis did not change its conclusion of a lack of increased risk of AMI from ST, while though the second MONICA study25 reported a higher OR with snuff for fatal cases (1.50, 95% CI 0.45–5.03) than it did for fatal and non-fatal cases combined (0.58, 0.35–0.94), the two estimates were not clearly heterogeneous (2 = 2.04 on 1 df, P = 0.15). More data are needed here.
Although the clearest way to evaluate possible effects of ST is to compare risks in current (and former) users who have never smoked with those who have never used tobacco, the limited data on risks of ST use in smokers or in the whole population (with adjustment for smoking) have also been considered. Given the large effect of smoking and the scope for imprecise classification, there are inevitably uncertainties about smoking-adjusted risks, but joint use of ST and smoking is very far from uncommon and it is important to gain some insight into whether it involves unexpected risks. In fact, the data provided no evidence of an association of ST use with increased risk of IHD or AMI in current smokers or in all individuals. Although the apparent lack of effect of ST use in smokers, taken together with the evidence of a possible effect in never smokers, might at first suggest the existence of some interaction of ST and smoking on risk, in fact, formal analysis (Table 5) did not confirm this.
Evidence of an association of current ST use with risk was only apparent in three studies, the two US CPS studies for IHD, stroke and all CID and the Swedish construction workers study for IHD and all CID (Table 3). When attempting to interpret the data it is important to note that, of those seven studies which found no evidence of an increased risk of IHD or AMI associated with ST use, none reported any results for all CID and only two did so for stroke.
The overall evidence is dependent particularly on the two CPS studies, which together provide 76% of the total studied cases of IHD or AMI and 90% of those of stroke. Although these studies are large, of cohort design and adjust for quite a wide range of confounding variables, a number of their features require comment. First, the questions on ST use were asked in 1959 (CPS-I) and 1982 (CPS-II), and it is unclear how relevant the products used at that time are to the ST currently on the market. Second, the question used in CPS-I—‘Do you chew tobacco or use snuff?’ with answers recorded as ‘Never,’ ‘Occasionally’ or ‘Regularly’—made no specific provision for former use, though those respondents who happened to volunteer such information were excluded from analysis. Third, all the multivariate-adjusted estimates shown in Table 3 were lower than the age-adjusted estimates given by Henley et al.26 reflecting the generally poorer lifestyle characteristics of ST users than non-tobacco users, and bringing into question the possibility of some residual confounding. Finally, Table 5 of Henley et al.26 showed, based on over 200 cases of cardiovascular disease in ST users in CPS-II, that there was no indication of a dose–response relationship with IHD, stroke or all CID, such a relationship usually being considered a prerequisite for demonstrating causality. Evidence on dose–response from other studies was extremely limited, and more information is clearly needed here.
The meta-analyses provide no clear evidence from studies in Sweden of an association of current ST use with IHD or AMI or with stroke. While there is no significant heterogeneity by country, and the evidence, particularly for stroke, is quite limited, the data are consistent with Swedish snus having little or no true effect on risk of CID. Given the ST products used in Sweden and in the US have historically been very different, there is no compelling reason to expect any effects of ST to be the same in the two countries.
Although the analyses are mainly concerned with quantifying risks for ST use in comparison with no ST use, comparisons were also made with risks associated with smoking. In both Sweden and the USA, risks for IHD/AMI, stroke and all CID were always clearly higher in relation to smoking than in relation to ST.
One of the problems in obtaining reliable evidence on the risk of CID associated with ST as currently used is the lack of recent data, particularly in the US, where the nature of the product has changed over time,79 and the only three epidemiological studies providing information are cohort studies where the baseline information was collected 25 years or more ago (Table 1). There is a need for a substantial case-control study of CID to be conducted in the US (and perhaps also in Sweden) which would obtain detailed data on use of ST, smoking history and potential confounding factors.
Other outcomes
The approach taken here was to review the evidence qualitatively. Because of limitations in how the data were reported, useful meta-analyses could not practically be carried out.
For the risk factors associated with CIDs, which have been discussed in this report, there is little clear evidence of any effect of snuff as used in Sweden. One study reported an association with diabetes,44 but other studies in Sweden did not (Table 6), and no effect on insulin levels or on glucose levels or uptake has been consistently reported. In Sweden, blood pressure levels have generally been found to be similar in snuff users and non-users, though one experimental study63 reported an acute rise in blood pressure at rest from oral snuff, a finding which needs replication. Use of snuff has been repeatedly reported in Swedish studies to have no evident effect on a variety of other endpoints, including atherosclerosis, response to exercise, cholesterol levels, triglycerides and fibrinogen. However, there have been two reports of an increase in prevalence of Raynaud-type symptoms in snuff users.34,74 While it is possible that larger and better designed studies in Sweden may detect other circulatory effects of snuff, it seems clear for many risk factors that if there is an effect it is less than that of smoking.
As used in the US, ST does have effects on risk factors relevant to CID. The most commonly reported is an acute rise of blood pressure and heart rate, seen in many experimental studies. There may also be a chronic effect on blood pressure, though some of the cross-sectional studies reporting an increase suffer from evident weaknesses. Whether cross-sectional studies can be regarded as providing evidence on chronic effects, as suggested by Westman,62 seems dubious, because while people are unlikely to be using ST while their blood pressure is being taken, it seems eminently possible that, unless instructions have been given to the contrary, some people may have been using it very recently, when its acute effects are still occurring. While four of the nine cross-sectional studies conducted in Sweden appear to have attempted to avoid acute effects, by insisting on abstention from tobacco overnight,9 for at least 8 h,8 for an hour10 or in the waiting room34, none of the six US studies reported that it did so. No notable effect on diabetes mortality was seen in the CPS-I and CPS-II studies.26 There are also two case reports of Buerger's disease associated with the use of chewing tobacco69,70 and isolated reports of a variety of endpoints relative to CID. As for mortality or disease onset, the findings for ST as used in the US are more suggestive of a possible effect on the circulatory system than is the case for the findings for snuff as used in Sweden.
|
Conclusion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
The overall evidence on use of snuff taken from a substantial number of studies in Sweden does not demonstrate any increase in the risk of CID, any chronic effect on blood pressure or any increased risk of a range of other risk factors relevant to CID. More evidence is needed to confirm whether Swedish oral snuff causes an acute rise in blood pressure. It may increase risk of Raynaud-type symptoms.
The evidence of a possible effect of ST as used in the US is more compelling. Two large cohort studies have reported finding evidence of an increased risk of IHD and stroke, with the increase substantially less than seen in smokers. However, these studies have limitations and were conducted many years ago, and more up-to-date evidence is needed. An acute effect of ST on blood pressure and heart rate is well documented, but a chronic effect has not been clearly shown. The evidence on circulatory risk factors is difficult to generalize, but shows more associations with ST than seen in the Swedish studies.
Risk of CID associated with ST use is clearly substantially less than that associated with smoking and it has in fact not been clearly demonstrated that Swedish oral snuff is associated with any adverse effect at all in either non-smokers or smokers. However, the number of studies available is still limited and more evidence is needed.
Although smokers who wish to reduce their risk of CID would clearly do best to reduce their risk of CID by giving up smoking, many find great difficulty in doing so and may find that switching to ST is easier. The evidence suggests strongly that switching would indeed substantially reduce their risk, though studies which compared risks of those who switched and those who continued to smoke would be needed to confirm this.
CID is clearly only one of the possible consequences of ST use, and reliable data are needed on a variety of other endpoints before any overall judgement can be made as to its health effects in non-smokers and the benefits of switching to smokers.
[Subsequent to our searches, the INTERHEART case-control study of tobacco use and risk of AMI conducted in 52 countries reported that chewing tobacco alone was associated with an OR of 2.23 (1.41–3.52), but numbers of snuff users were too small to draw conclusions. However, since the majority of ST users in this study were from South Asia, with no separate results reported for Western populations, these findings do not affect our conclusions.]
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
I thank Jan Hamling for checking the details of this article and her and Barbara Forey for helpful comments. I also thank Pauline Wassell and Diana Morris for their patient and accurate word-processing of the various drafts of this review. I am grateful to Philip Morris International, who gave financial support to the publication of this review and to British American Tobacco who supported an earlier version of this review. Although the sponsors suggested the review, its execution and the interpretation of the findings were carried out independently of them.
Conflicts of interest: None declared.
KEY MESSAGES
|
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
1 International Agency for Research on Cancer. Tobacco smoking. (1986) Volume 38. Lyon, France: IARC. (IARC Monographs on the evaluation of the carcinogenic risk of chemicals to humans.)
2 Critchley JA, Unal B. Health effects associated with smokeless tobacco: a systematic review. Thorax (2003) 58:435–43.
3 Bolinder G. Smokeless tobacco-a less harmful alternative? The Tobacco Epidemic. Progress in Respiratory Research.—Bolliger CT, Fagerström KO, eds. (1997) Vol. 28. Basel: Karger. 199–212.[CrossRef]
4 Forey B, Hamling J, Lee P, Wald N. International Smoking Statistics. A Collection of Historical Data from 30 Economically Developed Countries. (2002) 2nd edn. London and Oxford: Wolfson Institute of Preventive Medicine and Oxford University Press. Errata available at www.pnlee.co.uk/ISS2.htm. Data after 1995 by personal communication from the authors.
5 Holm H, Jarvis MJ, Russell MAH, Feyerabend C. Nicotine intake and dependence in Swedish snuff takers. Psychopharmacology (1992) 108:507–11.[CrossRef][Medline]
6 Bolinder G, Norén A, Wahren J, de Faire U. Long-term use of smokeless tobacco and physical performance in middle-aged men. Eur J Clin Invest (1997) 27:427–33.[CrossRef][Web of Science][Medline]
7 Bolinder G, Norén A, de Faire U, Wahren J. Smokeless tobacco use and atherosclerosis: an ultrasonographic investigation of carotid intima media thickness in healthy middle-aged men. Atherosclerosis (1997) 132:95–103.[CrossRef][Web of Science][Medline]
8 Bolinder G, de Faire U. Ambulatory 24-h blood pressure monitoring in healthy, middle-aged smokeless tobacco users, smokers, and nontobacco users. Am J Hypertens (1998) 11:1153–63.[CrossRef][Web of Science][Medline]
9 Eliasson M, Lundblad D, Hägg E. Cardiovascular risk factors in young snuff-users and cigarette smokers. J Intern Med (1991) 230:17–22.[Web of Science][Medline]
10 Eliasson M, Asplund K, Evrin P-E, Lundblad D. Relationship of cigarette smoking and snuff dipping to plasma fibrinogen, fibrinolytic variables and serum insulin. The Northern Sweden MONICA study. Atherosclerosis (1995) 113:41–53.[CrossRef][Web of Science][Medline]
11 Agewall S, Persson B, Lindstedt G, Fagerberg B. Smoking and use of smokeless tobacco in treated hypertensive men at high coronary risk: utility of urinary cotinine determination. Br J Biomed Sci (2002) 59:145–49.[Web of Science][Medline]
12 Wennmalm A, Benthin G, Granström EF, Persson L, Petersson A-S, Winell S. Relation between tobacco use and urinary excretion of thromboxane A2 and prostacyclin metabolites in young men. Circulation (1991) 83:1698–704.
13 Benowitz NL, Porchet H, Sheiner L, Jacob P III. Nicotine absorption and cardiovascular effects with smokeless tobacco use: comparison with cigarettes and nicotine gum. Clin Pharmacol Ther (1988) 44:23–28.[Web of Science][Medline]
14 Benowitz NL, Jacob P III, Yu L. Daily use of smokeless tobacco: systemic effects. Ann Intern Med (1989) 111:112–16.
15 Gritz ER. Smoking and smoking cessation in cancer patients. Br J Addict (1991) 86:549–54.[CrossRef][Web of Science][Medline]
16 Holiday DB, McLarty JW, Yanagihara RH, Rilay L, Shepherd SB. Two biochemical markers effectively used to separate smokeless tobacco users from smokers and nonusers. South Med J (1995) 88:1107–13.[Web of Science][Medline]
17 Jacob P III, Yu L, Liang G, Shulgin AT, Benowitz NL. Gas chromatographic-mass spectrometric method for determination of anabasine, anatabine and other tobacco alkaloids in urine of smokers and smokeless tobacco users. J Chromatogr (1993) 619:49–61.[Web of Science][Medline]
18 Russell MAH, Jarvis MJ, Devitt G, Feyerabend C. Nicotine intake by snuff users. Br Med J (1981) 283:814–17.
19 US Surgeon General. The health benefits of smoking cessation, a report of the Surgeon General. (1990) Rockville, Maryland: US Department of Health and Human Services. 90–8416. DHHS Publication No. (CDC).
20 Shaper AG, Wannamethee SG, Walker M. Pipe and cigar smoking and major cardiovascular events, cancer incidence and all-cause mortality in middle-aged British men. Int J Epidemiol (2003) 32:802–8.
21 Henley SJ, Thun MJ, Chao A, Calle EE. Association between exclusive pipe smoking and mortality from cancer and other diseases. J Natl Cancer Inst (2004) 96:853–61.
22 Critchley JA, Unal B. Is smokeless tobacco a risk factor for coronary heart disease? A systematic review of epidemiological studies. Eur J Cardiovasc Prev Rehabil (2004) 11:101–12.[CrossRef][Web of Science][Medline]
23 Bolinder G, Alfredsson L, Englund A, de Faire U. Smokeless tobacco use and increased cardiovascular mortality among Swedish construction workers. Am J Public Health (1994) 84:399–404.
24 Huhtasaari F, Asplund K, Lundberg V, Stegmayr B, Wester PO. Tobacco and myocardial infarction: is snuff less dangerous than cigarettes? BMJ (1992) 305:1252–56.
25 Huhtasaari F, Lundberg V, Eliasson M, Janlert U, Asplund K. Smokeless tobacco as a possible risk factor for myocardial infarction: a population-based study in middle-aged men. J Am Coll Cardiol (1999) 34:1784–90.
26 Henley SJ, Thun MJ, Connell C, Calle EE. Two large prospective studies of mortality among men who use snuff or chewing tobacco (United States). Cancer Causes Control (2005) 16:347–58.[CrossRef][Web of Science][Medline]
27 Johansson S-E, Sundquist K, Qvist J, Sundquist J. Smokeless tobacco and coronary heart disease a 12-year follow-up study. Eur J Cardiovasc Prev Rehabil (2005) 12:387–92.[CrossRef][Web of Science][Medline]
28 Hergens M-P, Ahlbom A, Andersson T, Pershagen G. Swedish moist snuff and myocardial infarction among men. Epidemiology (2005) 16:12–16.[CrossRef][Web of Science][Medline]
29 Accortt NA, Waterbor JW, Beall C, Howard G. Chronic disease mortality in a cohort of smokeless tobacco users. Am J Epidemiol (2002) 156:730–7.
30 Breslow NE, Day NE. The analysis of case-control studies. (1980) Vol. 1. Lyon: IARC. (Davis W, editor. Statistical methods in cancer research.) IARC Scientific Publication No. 32.
31 Greenland S, Longnecker MP. Methods for trend estimation from summarized dose–response data, with applications to meta-analysis. Am J Epidemiol (1992) 135:1301–09.
32 Fry JS, Lee PN. Revisiting the association between environmental tobacco smoke exposure and lung cancer risk. I. The dose–response relationship with amount and duration of smoking by the husband. Indoor Built Environ (2000) 9:303–16.
33 Fleiss JL, Gross AJ. Meta-analysis in epidemiology, with special reference to studies of the association between exposure to environmental tobacco smoke and lung cancer: a critique. J Clin Epidemiol (1991) 44:127–39.[CrossRef][Web of Science][Medline]
34 Bolinder GM, Ahlborg BO, Lindell JH. Use of smokeless tobacco: blood pressure elevation and other health hazards found in a large-scale population survey. J Intern Med (1992) 232:327–34.[Web of Science][Medline]
35 Henley SJ, Flanders WD, Manatunga A, Thun MJ. Leanness and lung cancer risk: fact or artifact? Epidemiology (2002) 13:268–76.[CrossRef][Web of Science][Medline]
36 Ahmed E, Stegmayr B, Trifunovic J, Weinehall L, Hallmans G, Lefvert AK. Anticardiolipin antibodies are not an independent risk factor for stroke: an incident case-referent study nested within the MONICA and Västerbotten cohort project. Stroke (2000) 31:1289–93.
37 Asplund K, Nasic S, Janlert U, Stegmayr B. Smokeless tobacco as a possible risk factor for stroke in men – a nested case-control study. Stroke (2003) 34:1754–59.
38 Asplund K. Smokeless tobacco and cardiovascular disease. Prog Cardiovasc Dis (2003) 45:383–94.[Web of Science][Medline]
39 US Surgeon General. Reducing the health consequences of smoking. 25 years of progress. A report of the Surgeon General. (1989) Rockville, Maryland: US Department of Health and Human Services; Public Health Services. DHHS Publication No. (CDC) 89-8411. Available at: http://www.cdc.gov/tobacco/sgr/index.htm.
40 US Surgeon General. The health consequences of smoking: a report of the Surgeon General. (2004) Atlanta, Georgia: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. Available at: http://www.cdc.gov/tobacco/sgr/sgr_2004/index.htm.
41 Yusuf S, Hawken S, Ôunpuu S, et al. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): case-control study. Lancet (2004) 364:937–52.[CrossRef][Web of Science][Medline]
42 Attvall S, Fowelin J, Lager I, Von Schenck H, Smith U. Smoking induces insulin resistance – a potential link with the insulin resistance syndrome. J Intern Med (1993) 233:327–32.[Web of Science][Medline]
43 Eliasson B, Taskinen M-R, Smith U. Long-term use of nicotine gum is associated with hyperinsulinemia and insulin resistance. Circulation (1996) 94:878–81.
44 Persson PG, Carlsson S, Svanström L, Östenson CG, Efendic S, Grill V. Cigarette smoking, oral moist snuff use and glucose intolerance. J Intern Med (2000) 248:103–10.[CrossRef][Web of Science][Medline]
45 Wallenfeldt K, Hulthe J, Bokemark L, Wikstrand J, Fagerberg B. Carotid and femoral atherosclerosis, cardiovascular risk factors and C-reactive protein in relation to smokeless tobacco use or smoking in 58-year-old men. J Intern Med (2001) 250:492–501.[CrossRef][Web of Science][Medline]
46 Eliasson M, Asplund K, Nasic S, Rodu B. Influence of smoking and snus on the prevalence and incidence of type 2 diabetes amongst men: the northern Sweden MONICA study. J Intern Med (2004) 256:101–10.[CrossRef][Web of Science][Medline]
47 Will JC, Galuska DA, Ford ES, Mokdad A, Calle EE. Cigarette smoking and diabetes mellitus: evidence of a positive association from a large prospective cohort study. Int J Epidemiol (2001) 30:540–46.
48 Rimm EB, Manson JE, Stampfer MJ, et al. Cigarette smoking and the risk of diabetes in women. Am J Public Health (1993) 83:211–14.
49 Rimm EB, Chan J, Stampfer MJ, Colditz GA, Willett WC. Prospective study of cigarette smoking, alcohol use, and the risk of diabetes in men. BMJ (1995) 310:555–59.
50 Squires WG Jr., Branton TA, Zinkgraf S, et al. Hemodynamic effects of oral smokeless tobacco in dogs and young adults. Prev Med (1984) 13:195–206.[CrossRef][Web of Science][Medline]
51 Simon DL, Iglauer A. The acute effect of chewing tobacco and smoking in habitual users. Ann N Y Acad Sci (1960) 90:119–32.[CrossRef][Web of Science]
52 Glover E, Edwards SW, Christen AG, Finnicum P. Smokeless tobacco reseach: an interdisciplinary approach. Health Values (1984) 8:21–25.[Medline]
53 Ksir C, Shank M, Kraemer W, Noble B. Effects of chewing tobacco on heart rate and blood pressure during exercise. J Sports Med (1986) 26:384–89.[Web of Science]
54 Gapter AL, Noble BJ. The effect of chewing tobacco on heart rate, blood pressure, and percieved exertion in male college athletes [Abstract]. Med Sci Sports Exerc (1986) 18:S72.
55 Fant RV, Henningfield JE, Nelson RA, Pickworth WB. Pharmacokinetics and pharmacodynamics of moist snuff in humans. Tob Control (1999) 8:387–92.
56 Wolk R, Shamsuzzaman ASM, Svatikova A, et al. Hemodynamic and autonomic effects of smokeless tobacco in healthy young men. J Am Coll Cardiol (2005) 45:910–14.
57 Blachley JD, Knochel JP. Tobacco chewer's hypokalemia: licorice revisited. N Engl J Med (1980) 302:784–85.[Web of Science][Medline]
58 McPhaul M, Punzi HA, Sandy A, Borganelli M, Rude R, Kaplan NM. Snuff-induced hypertension in pheochromocytoma. JAMA (1984) 252:2860–62.
59 Wells DG, Rustick JM. Hypertension from smokeless tobacco [Letter]. Anesthesiology (1986) 65:339.[CrossRef][Web of Science][Medline]
60 Adelman RD. Smokeless tobacco and hypertension in an adolescent [Letter]. Pediatrics (1987) 79:837–38.
61 Allen SS, Hatsukami D, Jensen J, Grillo M, Bliss R. Effects of treatment on cardiovascular risk among smokeless tobacco users. Prev Med (1995) 24:357–62.[CrossRef][Web of Science][Medline]
62 Westman EC. Does smokeless tobacco cause hypertension? South Med J (1995) 88:716–20.[Web of Science][Medline]
63 Hirsch J-M, Hedner J, Wernstedt L, Lundberg J, Hedner T. Hemodynamic effects of the use of snuff. Clin Pharmacol Ther (1992) 52:394–401.[Web of Science][Medline]
64 Schroeder KL, Chen MS Jr. Smokeless tobacco and blood pressure [Letter]. N Engl J Med (1985) 312:919.[Web of Science][Medline]
65 Westman EC, Guthrie GP Jr. Licorice, tobacco chewing, and hypertension [Letter]. N Engl J Med (1990) 322:850.[Medline]
66 Ernster VL, Grady DS, Greene JC, et al. Smokeless tobacco use and health effects among baseball players. JAMA (1990) 264:218–24.
67 Siegel D, Benowitz N, Ernster VL, Grady DG, Hauck WW. Smokeless tobacco, cardiovascular risk factors, and nicotine and cotinine levels in professional baseball players. Am J Public Health (1992) 82:417–21.
68 Ahlbom A, Olsson UA, Pershagen G. Health risks associated with Swedish snus. (1997) Symposium on Snus Arranged by the National Board of Health and Welfare, Stockholm, 19-20 September 1996: The National Board of Health and Welfare, Stockholm. 27.
69 O’Dell JR, Linder J, Markin RS, Moore GF. Thromboangiitis obliterans (Buerger's disease) and smokeless tobacco. Arthritis Rheum (1987) 30:1054–56.[Web of Science][Medline]
70 Lie JT. Thromboangiitis obliterans (Buerger's disease) and smokeless tobacco [Letter]. Arthritis Rheum (1988) 31:812–13.[Web of Science][Medline]
71 Bahrke MS, Baur TS, Poland DF, Connors DF. Tobacco use and performance on the U.S. army physical fitness test. Mil Med (1988) 153:229–35.[Web of Science][Medline]
72 Tucker LA. Use of smokeless tobacco, cigarette smoking, and hypercholesterolemia. Am J Public Health (1989) 79:1048–50.
73 van Duser BL, Raven PB. The effects of oral smokeless tobacco on the cardiorespiratory response to exercise. Med Sci Sports Exerc (1992) 24:389–95.
74 Ekenvall L, Lindblad LE. Vibrationsutlösta Raynaudfenomen och nikotinkonsumtion - en preliminär rapport. (Vibration induced white fingers and nicotine - a preliminary report). Opusc Med (1985) 30:28–31.
75 Stegmayr B, Johansson I, Huhtasaari F, Moser U, Asplund K. Use of smokeless tobacco and cigarettes – effects on plasma levels of antioxidant vitamins. Int J Vitam Nutr Res (1993) 63:195–200.[Web of Science][Medline]
76 US Surgeon General. The Health Consequences of Involuntary Exposure to Tobacco Smoke: A Report of the Surgeon General. (2006) Atlanta, Georgia: US Department of Health and Human Services, Centers for Disease Control and Prevention, Coordinating Center for Health Promotion, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. Available at: http://www.surgeongeneral.gov/library/secondhandsmoke/report.
77 Rodrigues L, Kirkwood BR. Case-control designs in the study of common diseases: updates on the demise of the rare disease assumption and the choice of sampling scheme for controls. Int J Epidemiol (1990) 19:205–13.
78 Zhang J, Yu KF. What's the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA (1998) 280:1690–91.
79 Rodu B, Jansson C. Smokeless tobacco and oral cancer: a review of the risks and determinants. Crit Rev Oral Biol Med (2004) 15:252–63.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Ebrahim Shock and awe: waking the dead Int. J. Epidemiol., August 1, 2007; 36(4): 701 - 702. [Full Text] [PDF]
|
|
|
|
|
|
M. McKee and A. Gilmore Commentary: Smokeless tobacco: seeing the whole picture Int. J. Epidemiol., August 1, 2007; 36(4): 805 - 808. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||