Introduction.

2.3 Relative roles of genetic and environmental factors in smoking behaviour and nicotine dependence Many diseases and related behaviours aggregate in families, i.e. family members resemble each other more than expected. Aggregation of diseases and behaviours in families can be due to environmental factors shared by family members as well as to shared genes. Genetic epidemiology aims to investigate the role of familial aggregation and genetic factors in human disease and such related risk factors as smoking. Th e fi rst twin studies of smoking were reported in the late 1950s and 1960s as part of a debate on the causal role of smoking in lung cancer. Th ese studies reviewed by Kaprio (1984) demonstrated that concordance for smoking was higher in monozygotic (MZ) than dizygotic (DZ) pairs, but these early studies did not provide quantitative estimates of heritability. Early twin studies on pairs reared apart supported the general conclusion of some genetic eff ects (Faber 1981, Kaprio, Koskenvuo & Langinvainio 1984, Shields 1962). Since then, quantitative genetic analyses have developed considerably (Boomsma, Busjahn & Peltonen 2002, Neale et al. 2003, Posthuma et al. 2003) and permitted more accurate estimates of the contribution of genetic factors. Smoking behaviour and nicotine dependence can be examined in family studies, particularly twin and adoption studies (Boomsma, Busjahn & Peltonen 2002, Posthuma et al. 2003). Results have revealed that MZ pairs are more similar than DZ pairs with respect to the trait being evaluated. Further evidence for genetic factors can come from including information on other relatives, such as siblings. Th e next step is characterizing genetic trait localization by linkage using genome-wide scans in families or genome-wide association in cases and controls. In addition, association analyses permit the identifi cation of specifi c genes responsible for genetic variation within the phenotype investigated. Heritability can be defi ned as the part of phenotypic variation explained by genetic diff erences in a specifi c population (Th omas 2004). 2.3.1 Genetics of smoking behaviour A considerable number of twin studies have provided evidence for the heritability of cigarette smoking and nicotine dependence (Lerman, Berrettini 2003, Li et al. 2003a). Earlier family and adoption studies support the fi nding of a genetic infl uence on smoking behaviour between biological siblings (Goode et al. 2003, Osler et al. 2001), but not between children and adoptive or biological parents (Osler et al. 2001). Finding genes that contribute to smoking behaviour and nicotine dependence has proven to be as challenging as one would expect for a complex trait 26 infl uenced by multiple genes and environmental factors as well as by their interactions. Th e genetic architecture of many features of smoking behaviour as single phenotypes is fairly well characterized based on twin and family studies (Li et al. 2003a). However, most of these studies have examined smoking initiation and smoking cessation, while quantitative genetic studies on other smoking behaviour phenotypes and interrelations between traits are less abundant. Th e estimates of heritability of diff erent components of smoking behaviour vary substantially. Gynther et al. (1999) suggest one explanation for the inconsistent results, speculating that genetic infl uences may promote a general disposition to smoke, while environmental experiences may lead to specifi c patterns of behaviour. Koopmans et al. (1999) propose that diff erent genetic and environmental factors infl uence diff erent smoking behaviour phenotypes. Smoking shows substantial familial aggregation, partly to genetic similarity of family members and partly due to social learning and other shared environmental factors. 2.3.2 Smoking initiation Earlier studies (Table 5) have revealed a considerable genetic contribution to risk of smoking initiation. Th e heritability estimates in these studies vary signifi cantly, the range being from 0.32 to 0.78 (Edwards, Austin & Jarvik 1995, Hamilton et al. 2006, Hardie, Moss & Lynch 2006, Heath et al. 1993, Heath et al. 1999, Heath, Madden & Martin 1998, Kendler et al. 1999, Madden et al. 2004, Maes et al. 2004, True et al. 1997, Vink et al. 2004). Th is variation is not surprising given that the role of genetic factors probably varies with time and place of investigation, as well as between the various populations studied (Kendler et al. 1999). A meta-analysis by Li et al. (2003a) showed that on average heritability of smoking initiation appears to be higher in women (55%) than in men (37%). Th is notion is supported by the studies of Madden (Madden et al. 1999) and Heath (Heath et al. 2002). However, most of these studies have examined becoming a smoker as a phenotype, whereas only a few have investigated age at initiation (Hardie, Moss & Lynch 2006, Heath et al. 1999, Morley et al. 2007, Vink et al. 2006) . 

Conclusions.  

Scientific conclusions The results confirmed that genetic factors are important in amount smoked and smoking cessation, and these genetic factors are largely independent of genetic influences on age at initiation. Starting to smoke is not a risk as such, but is an indicator of risk to develop future nicotine dependence. Further, it may not be a causal effect because the correlation is genetic, not environmental. Thus, some of the same genetic factors may underlie nicotine dependence and age at initiation. The NDSS correlated moderately highly with FTND and DSM-IV. The sum score of the NDSS is highly associated with FTND and DSM-IV-defined nicotine dependence, and expands upon DSM-IV substance dependence criteria to address the components of nicotine dependence more specifically. The NDSS measure can be used to assess nicotine dependence. A special strength in some situations may be that it is independent of information concerning the number of cigarettes smoked per day. This study confirms that genetic factors are important in nicotine dependence and smoking cessation. However, environmental factors are at least as important and can interact with one’s behaviour in smoking prevention and cessation and nicotine dependence. Social support plays a critical role in smoking cessation, not only in adolescence but also in adulthood, especially for men. Postponing smoking onset to adulthood appears to have some advantages. Age at initiation and smoking cessation have some genetic background, although environmental factors are important. Even if smoking initiation should and could be postponed to a later age, potential vulnerability to nicotine dependence cannot be completely inhibited. Practical implications The multidimensionality of nicotine dependence can now be measured with a validated scale. Based on these results and earlier literature, it seems evident that quitting smoking is really difficult for some people and large variation exists between individuals. Although the results confirm the role of genetic factors in smoking and nicotine dependence, we should not conclude that there is nothing to do in the face of genetic vulnerability to smoking and nicotine dependence. Concerning interventions, the main focus should be on environmental factors and how they interact with genetic vulnerabilities because genetic factors are expressed differently in different environments. This has now been shown empirically also for smoking in adolescents, with the genetic variance being dependent on rearing characteristics in the family (Dick et al. 2007). 65 Smoking behaviour and nicotine dependence are multidimensional phenomena, with some known but also relevant new dimensions. This information helps us to better understand variation in smoking behaviour and nicotine dependence. Nicotine dependence measurements (NDSS, FTND, DSM-IV) should be subjected to further critical evaluation. Only aft er thorough validation and experience will it bee possible to recommend the NDSS scale as a new measure for clinical use. Transition in marital status seems to be an important life event with regard to smoking cessation, especially among men. Further resources should be devoted to optimizing social support to motivate both sexes to quit smoking. This study may also help to clarify which risk groups should be directed towards future prevention programmes and what kinds of cessation programmes should be developed for the most vulnerable groups. Nicotine dependence has not earlier been studied in Finland in large data sets. The proportion of smokers who were nicotine-dependent ranged from less than 20% among those smoking less than 10 cigarettes per day to more than 80% among those smoking 30+ cigarettes a day. It was important to provide some heritability estimates for Finland; these were quite consistent with those found in studies from Australia, USA and the Netherlands. Informative genetic samples for smoking behaviours have thus far existed only in a handful of countries. Results suggest that postponing experiments with the first cigarette to a later age is beneficial. Not trying cigarettes at all would, of course, be ideal since nicotine dependence can develop relatively fast after some experiments (DiFranza et al. 2007a). Future research Earlier genome-wide scan results were replicated in this study and further genome wide, fi ne-mapping and candidate gene analyses are ongoing. More co-morbid phenotypes, such as nicotine dependence and depression, should be investigated to uncover potential candidate genes. Hardly surprisingly, the validation study of the novel nicotine dependence measure did not reveal a similar factor structure as in other studies because of the different data sets used. Further studies in randomly selected population samples of smokers are needed to develop and validate nicotine dependence measurements, and different dimensions of nicotine dependence should also be examined. In a comparison of three measures of nicotine dependence, which was not reported in the substudies, were found that almost one quarter of participants were not dependent by any measure. This might indicate that these individuals are really not dependent or that even these three measures do not assess all aspects of dependence. Further analyses are needed to characterize these individuals in more detail. Furthermore, quantitative trait linkage studies of NDSS and other nicotine dependence scales would be worthwhile. 66 Although this study contributed significantly to development of phenotypes suitable for genetic research, there are numerous challenges in this field. One strategy for elucidating the genetic dimension would be identifi cation of endophenotypes (i.e. heritable traits associated with disease susceptibility) (Flint, Munafò 2007). Such an endophenotype should more closely represent a biological process contributing to a trait of interest. Another interesting theme might be subjective reactions to the very first cigarette, which may help index heritable individual differences in reactions to nicotine. Further, to maximize power in genetic research, identifying quantitative traits would be valuable (Pomerleau et al. 2007). Overall, nicotine dependence and smoking behaviour demonstrate genetic liability but also substantial environmental background. If smoking initiation cannot be completely prevented, its postponement to adulthood would be beneficial to reduce the length of exposure to the harmful and toxic substances in cigarette smoke. However, even if starting could be postponed to later years, potential vulnerability to nicotine dependence probably cannot be inhibited. Starting to smoke is not a risk as such, but it is an indicator of vulnerability to nicotine dependence. Th e same genetic factors may underlie vulnerability to both nicotine dependence and age at initiation. More research is needed to test alternative mechanisms. Nicotine dependence is a complex issue. Many components and their interactions should be investigated to shed light on the multidimensionality, aetiology and mechanisms of nicotine dependence.