Early Onset of Drinking and Risk of Heavy Drinking in Young Adulthood—A 13-Year Prospective Study Ingeborg Rossow and Emmanuel Kuntsche Background: Prevention programs often aim at preventing early onset of drinking (EOD) on the grounds that this may curb heavy drinking in adulthood. While many studies have shown an associa- tion between EOD and adult alcohol use disorders, these findings could be inflated by retrospective reports or insufficient control for confounders. This study examined the association between EOD behavior assessed in early adolescence and heavy drinking in adulthood, controlling for deviant behav- ior and parental heavy drinking. Methods: Data were collected prospectively over a 13-year period from 1,311 Norwegian school students. At t1 (ages 13 to 14 years), onset of drinking behavior (any drinking and heavy episodic drinking), conduct problems (CP), other problem behaviors, and parental heavy drinking were assessed. At t2 (ages 26 to 27 years), heavy drinking behavior was assessed in terms of heavy episodic drinking frequency and AUDIT score. Results: EOD behavior was associated with CP, other problem behaviors, and parental heavy drinking in early adolescence. A higher risk of heavy drinking in adulthood was found among those with EOD behavior, yet after control for CP, this association became small and statistically nonsignifi- cant. Among low-risk individuals (i.e., those with no CP at t1), there was no association between EOD behavior and heavy drinking in adulthood, while there was a significant association among those with CP. Conclusions: EOD behavior appears not per se responsible for heavy drinking in adulthood unless being part of a broader array of problem behaviors. Key Words: Early Drinking Onset, Adult Heavy Drinking, Prospective Study. A N IMPORTANT AIM in many prevention programs targeting young people is to delay the onset of drinking behavior as long as possible, to prevent alcohol use disorders (AUD) later in life (Buddy, 2011; Ward et al., 2010). This claim is based on numerous studies that have demonstrated a correlation between early age at first drink (AFD) and later AUD (Dawson et al., 2008). Some authors have implied that the association is causal and suggested that by delaying onset of drinking, one may prevent heavy drinking or AUD in adulthood (Bonomo et al., 2004; DeWit et al., 2000; Guo et al., 2000; Hingson and Zha, 2009; Pedersen and Skrondal, 1998). However, the interpretations and implications of this correlation differ significantly between authors (Dawson et al., 2008; Harford, 2003; Rossow, 2006), and there are those who argue that early onset of drinking (EOD) is non- causal and that attempts to prevent the development of AUD by delaying drinking onset are unlikely to be successful (Prescott and Kendler, 1999). There are at least 2 reasons to suspect that many or most of the reported significant associations between EOD and later problem drinking are an overestimate of a possible cau- sal effect. First, the majority of studies on this topic have assessed EOD or AFD as reported retrospectively by adult drinkers. Such reports are only moderately reliable and with increasing age respondents tend to report older ages of drink- ing onset (Engels et al., 1997; Parra et al., 2003). If such recall bias is related to later problem drinking, it has signifi- cant impact on how to interpret the results (King and Chassin, 2007). A recent longitudinal study by Sartor and colleagues (2011) did indeed show that an upward shift in reported age of onset was more pronounced among lighter drinkers. They concluded that heavy drinkers may be over- represented among early onset users in retrospective studies, leading to inflated estimates of the association between early initiation and alcohol misuse (Sartor et al., 2011). It is there- fore argued that by using prospective studies following par- ticipants from early adolescence and into adulthood, such bias can be avoided (King and Chassin, 2007). A second reason is that EOD and later drinking problems share common characteristics that may account for some or all of the correlation between early onset and adult drinking problems. Already in 1965, Kettil Bruun noted that the observed association between EOD and later heavy drinking From the Norwegian Institute for Alcohol and Drug Research (IR), Oslo, Norway; Sucht Schweiz (EK), Research Institute, Lausanne, Switzerland; and Behavioural Science Institute (EK), Radboud Univer- sity Nijmegen, Nijmegen, The Netherlands. Received for publication March 21, 2012; accepted June 15, 2012. Reprint requests: Ingeborg Rossow, PhD, Norwegian Institute for Alcohol and Drug Research, PO Box 565 Sentrum, N-0105 Oslo, Norway; Tel.: +47 40626039; Fax: +47 22340401; E-mail: ir@sirus.no Copyright © 2012 by the Research Society on Alcoholism. DOI: 10.1111/j.1530-0277.2012.01924.x Alcohol Clin Exp Res, Vol 37, No S1, 2013: pp E297–E304 E297 ALCOHOLISM:CLINICAL AND EXPERIMENTAL RESEARCH Vol. 37, No. S1 January 2013