Experimental and Clinical Psychopharmacology The Role of First Use of Inhalants Within Sequencing Pattern of First Use of Drugs Among Brazilian University Students João Maurício Castaldelli-Maia, Sérgio Nicastri, Lúcio Garcia de Oliveira, Arthur Guerra de Andrade, and Silvia S. Martins Online First Publication, August 25, 2014. http://dx.doi.org/10.1037/a0037794 CITATION Castaldelli-Maia, J. M., Nicastri, S., de Oliveira, L. G., de Andrade, A. G., & Martins, S. S. (2014, August 25). The Role of First Use of Inhalants Within Sequencing Pattern of First Use of Drugs Among Brazilian University Students . Experimental and Clinical Psychopharmacology. Advance online publication. http://dx.doi.org/10.1037/a0037794 Experimental and Clinical Psychopharmacology 2014, Vol. 22, No. 5, 000 © 2014 American Psychological Association 1064-1297/14/$12.00 http://dx.doi.org/10.1037/a0037794 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. The Role of First Use of Inhalants Within Sequencing Pattern of First Use of Drugs Among Brazilian University Students João Maurício Castaldelli-Maia Sérgio Nicastri University of São Paulo and Fundação do ABC University of São Paulo Lúcio Garcia de Oliveira Arthur Guerra de Andrade University of São Paulo University of São Paulo and Fundação do ABC Silvia S. Martins Columbia University The present study investigated the role of first use of inhalants within a first drug sequencing pattern. In a representative sample of university students from 27 Brazilian capitals (n ⫽ 12,711), we analyzed the patterns of transition from/to first use of inhalants to/from the first use of alcohol, tobacco, cannabis, cocaine, hallucinogens, ecstasy, amphetamines, prescription opioids, and tranquilizers. Cox proportional hazards models were used to analyze data. Drugs that were not specified as the pair of drugs tested in each model were included as time-varying covariates in all models. In this sample, first use of inhalants was preceded only by the first use of alcohol and tobacco. However, first use of inhalants preceded first use of cannabis, amphetamines, cocaine, and tranquilizers. First use of inhalants preceded the first use of prescription opioids, and vice versa. This study highlights the need to intervene early with youths who are at risk of or just beginning to use inhalants, because this class of drugs seems to be the first illegal drug in Brazil to be experimented by respondents in our sample. There is also a call for attention to individuals who have already first used inhalants because of their higher chance to experiment with other drugs such as cannabis, cocaine, and prescription drugs. All these findings show an in-transition culture of drug use, which should be tracked through time, because some classical models (i.e., gateway model) might be outdated and might also not fit within different settings. Keywords: Brazil, experimentation, gateway, inhalants, transition, university students major consequences for abusers, their families, and society (Balster et al., 2009), yet it is one of the least-understood and poorly studied substance abuse problems (Balster et al., 2009). Despite the substantial prevalence and serious toxicities of inhalant use, it has been termed “the forgotten epidemic,” and remains as the least studied form of substance abuse, although research on its epidemiology, neurobiology, treatment, and prevention has accelerated in recent years (Howard et al., 2011). The high prevalence of lifetime use of inhalants is also evident in the general population of Brazil, as inhalants are the fourth most widely used drug (Fonseca, Galduroz, Noto, & Carlini, 2010). Considering these statements, the present study sought to investigate the role of first use of inhalants within sequencing pattern of first use of drugs among Brazilian university students. Kandel (1975) and Kandel and Faust (1975) proposed a gateway model that attempts to explain the passage from the first contact with a drug to the first contact with a second drug. In this model, the user would go through four stages of involvement with psychoactive substances. There was no special attention to inhalants in this proposed model despite their high prevalence of use. At first, usually individuals use beer and wine. Then, individuals could use tobacco and/or strong drinks (liquor, spirits). From the second stage the person could go to a third stage, in which use of cannabis can happen. The gateway hypothesis implicitly assumes a causal Inhalants are one of the most popular drug classes throughout the world. Inhalant abuse, which is also referred to as volatile substance abuse or solvent abuse, is a worldwide problem with João Maurício Castaldelli-Maia, Department of Psychiatry, Medical School, University of São Paulo, and Department of Neuroscience, Medical School, Fundação do ABC, Brazil; Sérgio Nicastri, Department of Psychiatry, Medical School, University of São Paulo; Lúcio Garcia de Oliveira, Department of Psychiatry, Medical School, University of São Paulo; Arthur Guerra de Andrade, Department of Psychiatry, Medical School, University of São Paulo, and Department of Neuroscience, Medical School, Fundação do ABC; Silvia S. Martins, Department of Epidemiology, Columbia University Mailman School of Public Health. This work was supported by National Secretariat for Drug Policies (SENAD), Brazil and São Paulo State Research Support Foundation FAPESP, Brazil. Dr. Martins receives research support from the National Institute on Drug Abuse (NIDA) Grant DA023434 and from the National Institute of Child and Human Development (NICHD) Grant HD060072, while working on this article. Dr. Oliveira receives research support from FAPESP Grant 2008/55550-7. Correspondence concerning this article should be addressed to João Maurício Castaldelli-Maia, Rua Dr Ovídio Pires de Campos, 785 - GREA, 05403-903, São Paulo, SP, Brazil. E-mail: jmcmaia2@gmail.com 1 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 2 CASTALDELLI-MAIA ET AL. chain sequence in which (a) alcohol and/or tobacco use precede cannabis use, (b) cannabis is used before the onset of other illegal drugs (including inhalants), and (c) the use of cannabis increases the likelihood of using other illegal drugs (Fergusson et al., 2006). Alcohol and tobacco are indeed the first drugs that adolescents experiment with according to several international studies (Van Etten et al., 1997; Wagner & Anthony, 2002; Herrera-Vázquez et al., 2004; Wells & McGee, 2008; Caris et al., 2009; Mayet et al., 2010), which may be attributable to the fact that these two drugs have the highest prevalence of use throughout the life span in most countries. However, there are a considerable number of individuals who experiment with cannabis before alcohol and/or tobacco depending on neighborhood (Tarter et al., 2006) and ethnic factors (Vaughn et al., 2008). Regarding sequence of first use of drugs other than alcohol and tobacco, data are more controversial (Castaldelli-Maia, Martins, de Oliveira, de Andrade, & Nicastri, 2014). In countries with a high prevalence of cannabis use, such as Germany, Ukraine, and New Zealand, the first use of cannabis precedes the first use of drugs other than alcohol and tobacco (Degenhardt et al., 2010). This pattern was indeed the case for Brazilian university students in a previous study (Castaldelli-Maia, Martins, de Oliveira, van Laar, de Andrade, & Nicastri, 2014). However, there are few studies (Ding et al., 2009; Sanchez, Noto, & Anthony, 2013) focusing on inhalants within drug first use sequencing pattern. In the United States, Ding et al. (2009) found that inhalants are more commonly used before other illicit drugs in a large sample of inhalant users. Within the Brazilian university student population, the prevalence of lifetime use of inhalants was relatively high (24.6%) according to a survey conducted during this decade (Wagner et al., 2007). Based on these data of relatively high prevalence of lifetime use of inhalants, one can hypothesize that inhalants could be commonly used before illegal drugs, occupying a similar role to cannabis. Sanchez, Ribeiro, Moura, Noto, & Martins (2013) analyzed data from a representative sample of 8th to 12th grade students (middle school ⫽ 8th to 9th grade; high school ⫽ 10th to 12th grade) from private schools (n ⫽ 5,226) in São Paulo, Brazil, to test for sequencing patterns among alcohol, tobacco, cannabis, and inhalants. In their study, inhalants were characterized as an intermediate drug class between legal (alcohol and tobacco) and illegal (cannabis) drugs. But this previous study (Sanchez, Ribeiro, et al., 2013) did not investigate the following: (a) the sequencing patterns involving inhalants and other important drugs to the Brazilian context (Castaldelli-Maia, Bhugra, de Andrade, & Lotufo-Neto, 2013), such as crack/cocaine (Moreira et al., 2014), prescription drugs (Opaleye et al., 2014), hallucinogens (Castaldelli-Maia, de Oliveira, de Andrade, Lotufo-Neto, & Bhugra, 2012), and ecstasy (Remy et al., 2013); and (b) older students from college years in private and public settings. Considering that adolescence is a period of experimentation with various drugs (Malta et al., 2010), the subsequent period (university or college years) may be of interest to access the sequence of experimentation of with these drugs. In a previous analysis using the sample of the present study, the average age of first use of each of the nine drugs analyzed in this study was between 15.2 and 21.2 years (Castaldelli-Maia, Martins, de Oliveira, van Laar, de Andrade, & Nicastri, 2014). This study investigated the role of first use of inhalants within a drug use sequencing pattern in a large representative sample of Brazilian university students. It is also important to consider the first use of drugs other than inhalants as time-dependent variables, as they may have a mediating role in the sequence of drug use. Method Data comes from an epidemiological study conducted in 27 Brazilian state capitals. The general objective of this epidemiological study was to evaluate the sociodemographic characteristics, drug use and mental health aspects of a nationally representative sample of university students (n ⫽ 12,711). The project was previously evaluated and approved by the Ethics Committee for the Analysis of Research Projects at the School of Medicine of the University of São Paulo. Data collection was completed between May and December of 2009. Sample The target population of this study were university students enrolled in undergraduate programs in Higher Education Institutions (HEIs), both public and private, in the 27 Brazilian state capitals (Brazil is composed by 27 states). The undergraduate programs can last from 4 to 6 years in Brazil. A random sample was stratified and recruited by unequal size clusters. Sampling was conducted in two stages; a sample of HEIs was selected, followed by the selection of classes of students was chosen from this sample. Because the size of the HEIs and the classes (in terms of number of universities) were not always the same, these conglomerates were of unequal sizes. Because all 27 state capitals had to be covered, including public and private HEIs, sample stratification was carried out based on these two variables (public and private), namely, by capital and by type of institution, a total of 54 levels. However, this stratification was used only for operating purposes. At the data analysis stage, only the five administrative regions (including the 27 state capitals) and the two HEI types were considered for stratification, generating a total of 10 levels. To make the field work economically feasible, it was decided to select a sample of HEIs and, within each one of these, select a sample of classes: the primary sampling unit considered for this study was the HEI, and the secondary sampling unit was the class. Thus, sampling selection was made in two stages, in which the clusters considered included both HEIs and student groups. Because the sizes (in relation to the number of students) of the HEIs and the groups are not always the same, these clusters are, in fact, of different sizes. Thus, in general, the sampling design consisted of a random sample, stratified by clusters of different sizes, selected in two stages. However, this sampling selection method posed a challenge because of the structure of the population under review. In conventional sampling, the clusters form a disconnected division, that is, the population elements are related to a single cluster, which is not true for higher education. Within a HEI, student groups are not disconnected, and there may be an overlap between groups of students, because a single student can be enrolled in more than one subject. To get around this issue, we used the multiplicity-based sampling method, enabling the individuals to be included in more than one cluster. However, it was necessary to obtain information about the number of clusters to which the elements selected for the sample were related, for later use in the statistical analysis stage. At the end of the data collection, 100 of 114 HEIs accepted to take part in this study This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. FIRST USE OF INHALANTS AND FIRST USE OF DRUGS (88% of the estimated size), as well as 654 student classes (70.6% of the estimated size), with a total participation of 12,721 college students throughout Brazil. Although the response rate of participation was 95.6% among the college students who were inside classes at the time of the interview, the final response rate for this study was around 72.1% when the estimated size of the college student sample was taken into consideration (12,721 of 17,651). Finally, of these students (12,721), 10 were excluded because they claimed to use Relevin (fictitious drug); thus, the data from 12,711 nationwide college students were analyzed. More details about this survey can be found in Andrade et al. (2012). Part of the 12,711 individuals was excluded from the present study. Respondents older than 40 years (n ⫽ 548) were excluded because of the high probability of memory bias. In addition, those with missing data on the variables of interest—age (n ⫽ 141), whether they used some of the drugs tested (n ⫽ 149 [cannabis]; 183 [inhalants]; 228 [opioid analgesics]; 174 [tranquilizers]; 206 [ecstasy]; 191 [amphetamines]; 190 [cocaine]; 165 [hallucinogens]), alcohol (n ⫽ 3203) or tobacco (n ⫽ 0), those who did not answer the age of first use of some of the drugs tested (n ⫽ 476 [cannabis]; 625 [inhalants]; 670 [opioid analgesics]; 602 [tranquilizers]; 316 [ecstasy]; 430 [amphetamines]; 284 [cocaine]; 298 [hallucinogens]); alcohol (n ⫽ 3,203) or tobacco (n ⫽ 985)—were excluded. Sample size varied according to the transition considered. 3 tionnaire was based on the World Health Organization research questionnaire, which was previously adapted by Andrade et al. (1997) and Stempliuk et al. (2005) for use with undergraduate students. The present study used part of these data, which focused on drug use. This section was collected in a series of questions with multiple response options. In the first part, the students reported whether they had ever used the following drugs (Have you ever tried NAME OF THE DRUG without a doctor’s prescription? Answer: Yes or No): inhalants, alcohol, tobacco, cannabis, cocaine, “merla” (a cocaine byproduct), crack, hallucinogens, ketamine, ayahuasca (Santo Daime), ecstasy (3– 4 methylenodyoximethamphtamine [MDMA]), anabolic steroids, tranquilizers, prescription opioids, anticholinergics, heroin, amphetamines, and synthetic drugs (methamphetamines and GHB). In the second part of the questionnaire, we asked for the age of the initial use of each of these drugs (How old were you when you first tried NAME OF THE DRUG? Answer: XX years). For this study, cocaine, “merla,” and crack were combined in the same group (named cocaine). Our survey instrument included questions that asked all respondents about the use of ‘inhalants or solvents,’ giving the following as examples: “loló,” glue, paint thinner, benzene, enamel, gasoline, and “lança-perfume.” All these examples of inhalants or solvents cover the majority of types of substances of this class of drugs used in Brazil (Mesquita et al., 1998; Hynes-Dowell et al., 2011; Sanchez, Noto, et al., 2013; Silva-Oliveira et al., 2014) Measures A structured, self-administered, and anonymous questionnaire of 98 closed questions was developed with an emphasis on drug use and related disorders, risky behaviors and the existence of psychiatric comorbidity (depressive symptoms and psychotic and nonspecific psychological complaints). The content of this ques- Statistical Analysis Time-dependent variables. Because the aim of this study was to assess whether the transition from the first use of one drug to the subsequent use of another drug occurred or not—always involving inhalants in one of the two roles—these comparisons Figure 1. Fluxogram presenting the division of the individuals between users and nonusers, among Brazilian university students of the 27 state capitals, 2009. See the online article for the color version of this figure. CASTALDELLI-MAIA ET AL. 4 Table 1 Sociodemographic Data (14 Variables Included in the Cox Survival Models) of Brazilian University Students From 27 Capitals, 2009 Table 1 (continued) Variable % 95% CI SE 28.67 26.42 17.00 14.33 9.64 2.13 21.83, 36.65 18.19, 36.70 11.22, 24.92 9.93, 20.23 3.75, 22.61 1.37, 3.31 0.03 0.04 0.03 0.02 0.04 0.01 15.05 84.95 8.27, 25.82 74.18, 91.73 0.04 0.04 92.03 7.53 88.58, 94.50 5.20, 10.80 0.01 0.01 c This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Variable Gender Female Male Agea ⬍22 years ⬎21 years Religion None Catholic Spiritualist Umbanda/Candomblé Jew Reformed/Evangelical Buddhist/Hindu Santo Daime/Vegetal Union Other Practice of the religion Special occasions only Monthly or more No Socioeconomic statusa Low Low-average Average High-average High Head of the family education Zero Up to 4 years 5–8 years 9–11 years 12 years or more Ethnic group White African-American Mulato Oriental Indian Otherb Marital status Single Married/living together Separated/divorced Widow Children Yes No Employment No Up to 20 hr/week ⬎20 hr/week Driver license Yes No Area of undergraduate studies Biological/health Mathematics Humanities % 95% CI SE 56.87 43.13 53.67, 60.00 40.00, 46.33 0.01 0.01 38.93 61.07 25.69, 54.04 45096, 74.31 0.07 0.07 14.85 50.00 8.90 1.77 0.67 17.42 0.64 0.20 4.97 12.25, 17.90 45.79, 54.22 7.40, 10.68 0.77, 4.04 0.37, 1.20 14.17, 21.24 0.40, 1.02 0.01, 0.44 3.39, 7.24 0.01 0.02 0.01 0.01 0.01 0.01 0.01 0.01 0.01 35.26 45.71 18.12 33.27, 37.30 43.26, 48.19 16.16, 20.26 0.01 0.01 0.01 0.01 1.75 25.23 62.84 10.09 0.01, 0.29 0.01, 2.37 21.64, 29.19 58.85, 66.66 7.76, 13.02 0.01 0.01 0.01 0.01 0.01 7.06 7.60 6.02 35.10 42.44 5.45, 9.10 6.01, 9.57 4.69, 7.69 31.66, 38.71 37.39, 47.66 0.01 0.01 0.01 0.01 0.02 62.27 6.50 24.71 2.45 0.11 2.96 66.05, 70.53 5.31, 7.94 21.44, 28.31 1.41, 4.22 0.57, 2.11 2.37, 3.71 0.01 0.01 0.01 0.01 0.01 0.01 75.68 20.10 3.57 0.65 63.89, 84.56 13.50, 28.84 2.03, 6.21 0.14, 2.87 0.05 0.03 0.01 0.01 19.78 79.56 11.82, 31.19 68.38, 87.51 0.04 0.04 39.65 17.85 41.85 34.95, 44.54 15.12, 20.94 35.50, 48.49 0.02 0.01 0.03 61.11 38.23 57.62, 64.49 34.97, 41.60 0.01 0.01 17.67 17.01 65.33 10.87, 27.40 11.60, 24.25 55.96, 73.64 0.04 0.03 0.04 were always performed between users and nonusers groups (e.g., investigating the transition from inhalants to cocaine—we tested inhalants users vs. inhalants nonusers—these groups were defined based on three options available in these two parts of the following Current year in college/university First Second Third Fourth Fifth Sixth All-day classes Yes No Happy with the undergraduate choice Yes No a Included in the Cox survival regression model as continuous measures. b The person does not include himself/herself within any of the ethnic groups listed above. c We did not ask whether students had failed in courses during their studies. question: if the subject had already used inhalants, if the subject had already used cocaine, and whether the subject had used the inhalants before, in the same year, or after cocaine. Inhalants users—if s/he just used cannabis or cocaine after inhalants. Inhalants nonusers—if s/he did not use both drugs, or used just cocaine, or inhalants before cocaine. If s/he used both in the same year, s/he was classified as a ‘drug seeker’). Users and nonusers were assembled for each transition to perform analyses of the transitions from the first use of a drug to the first use of another drug. Analyses focused on transitions of first use of inhalants with first use of alcohol, amphetamines, cannabis, cocaine, ecstasy, hallucinogens, prescription opioids, tobacco, and tranquilizers. Figure 1 depicts the division between users and nonusers. The drug seekers were considered drug users with the response (the use of the subsequent drug tested in the model) of event censored. With this we have tried to avoid loss of important data on other sociodemographic variables included in the regression models. Time-dependent covariates. Drugs other than the pair of drugs tested in each survival model were included as time-varying dependent covariates in all models. (e.g., in the model testing the pathway from cannabis to inhalants—if the individual first used alcohol after first using cannabis and before first trying inhalants, s/he was classified as alcohol user in the alcohol use time dependent variable of this model). All analyses were performed using the survey settings commands with weights, strata, and primary sampling units. The analysis of the survey data complied with the following characteristics of the sampling plan: (a) a complex sample, (b) the use of stratification, (c) clustering, and (d) dissimilar selection probabilities. Survival analysis methods were used to compare between one type of drug users and nonusers the cumulative prevalence of other type of drug use. We chose to use Cox Regression Survival Models (Cox, 1972), following an important framework (Singer & Willett, 1994). We used the statistical package STATA version 11 to run the models with time dependent covariates (as described above) and non–time-dependent covariates (these 14 sociodemographic covariates are presented in Table 1) adjustment. The Kaplan– Meier curves were plotted without any adjustment for other vari- FIRST USE OF INHALANTS AND FIRST USE OF DRUGS ables, and without the exclusion of individuals aged more than 40 years. We chose to use a link test to test for proportional-hazards assumption (Cleves et al., 2010). These re-estimation tests were carried out to test for the validity of each Cox model. It searches for variables to add to the model. Under the assumption that the Cox model is correctly specified, the added variables will add little or no explanatory power, so it tests whether these variables are insignificant. The Cox model reaches validity with p ⬎ .05 in this link test. Cox models with p ⬍ .05 in link test were not considered for discussion. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Results Table 1 depicts the sociodemographic information of the sample. Table 2 presents the results of the transitions analyzed, and Figure 2 highlights all the transitions that did not violate the Linktest of proportional hazard-assumptions (LPHA) and could be taken into account. Figure 3 depicts all Kaplan–Meier curves of the transitions that had significant differences that did not violate the LPHA. Overall, the first use of inhalants was subsequent to first use of alcohol (n ⫽ 1,181), tobacco (n ⫽ 655), and prescription opioids (n ⫽ 64), and preceded the subsequent first use of prescription opioids (n ⫽ 181), cannabis (n ⫽ 409), cocaine (n ⫽ 271), tranquilizers (n ⫽ 165), and amphetamines (n ⫽ 363). Transitions From First Use of Inhalants to First Use of Other Drugs Seven of nine models did not violate the LPHA. Within the valid models, the cumulative prevalence of first use of amphetamines (adjusted hazard ratios [aHR] ⫽ 2.59, 95% CI ⫽ 1.65– 4.07, p ⬍ .001), cannabis (aHR ⫽ 1.56, 95% CI ⫽ 1.26 –1.93, p ⬍ .001), cocaine (aHR ⫽ 4.33, 95% CI ⫽ 3.08 – 6.09, p ⬍ .001), prescrip- 5 tion opioids (aHR ⫽ 29.39, 95% CI ⫽ 7.17–120.54, p ⬍ .01), and tranquilizers (aHR ⫽ 2.10, 95% CI ⫽ 1.05– 4.18, p ⬍ .05) were significantly higher among those that had first initiated inhalant use as compared to inhalants nonusers. Regarding gender, there was higher cumulative prevalence of alcohol and cannabis first use among men as compared with women (aHR ⫽ 0.70, 95% CI ⫽ 0.53– 0.93, p ⬍ .05, and aHR ⫽ 0.56, 95% CI ⫽ 0.39 – 0.80, p ⬍ .01, respectively), adjusting for previous use of inhalants. However, regarding amphetamine and tranquilizer first use, an opposite situation occurred. The cumulative prevalence of first use in these cases was higher among women compared with men (aHR ⫽ 5.51, 95%CI ⫽ 3.72– 8.15, p ⬍ .001, and aHR ⫽ 2.28, 95%CI ⫽ 1.48 –3.50, p ⬍ .001, respectively), adjusting for previous use of inhalants. There were no gender differences with regard to cocaine, prescription opioid, and tobacco first use. Transitions From First Use of Other Drugs to First Use of Inhalants All models did not violate the LPHA. The cumulative prevalence of first use of inhalants was significantly higher among alcohol (aHR ⫽ 1.41, 95% CI ⫽ 1.09 –1.81, p ⬍ .01), prescription opioids (aHR ⫽ 3.51, 95% CI ⫽ 1.14 –10.82, p ⬍ .05), and tobacco (aHR ⫽ 2.70, 95% CI ⫽ 1.92–3.79, p ⬍ .001) first users as compared with nonusers of these drugs. In addition, the cumulative prevalence of first use of inhalants was significantly lower in cocaine (aHR ⫽ 0.06, 95% CI ⫽ 0.01– 0.23, p ⬍ .001) first users as compared with cocaine nonusers. Within all the transitions, there was a higher cumulative prevalence of inhalant first use among men as compared with women, adjusting for other drugs previous use. Table 2 Cox Survival Regression Models Results of Transitions From/to Inhalants Among University Students of 27 Brazilian Capitals, 2009 Transitiona Genderd From To Usersb aHR 95% CI p aHR 95% CI p p (TPHA)c Alcohol Amphetamines Cocaine Ecstasy Hallucinogens Cannabis Prescription opioids Tobacco Tranquilizers Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Inhalants Alcohol Amphetamines Cocaine Ecstasy Hallucinogens Cannabis Prescription Opioids Tobacco Tranquilizers 1181 29 20 23 24 283 64 655 37 115 363 271 417 338 409 181 274 165 1.41 0.88 0.06 0.63 0.96 0.95 3.51 2.70 1.07 0.78 2.59 4.33 2.12 6.88 1.56 29.39 1.30 2.10 1.09, 1.81 0.47, 1.64 0.01, 0.23 0.32, 1.23 0.59, 1.56 0.72, 1.24 1.14, 10.82 1.92, 3.79 0.37, 3.10 0.43, 1.43 1.65, 4.07 3.08, 6.09 0.93, 4.84 4.14, 11.42 1.26, 1.93 7.17, 120.54 0.84, 1.99 1.05, 4.18 .007 .700 <.001 .181 .881 .729 .029 <.001 .892 .426 <.001 <.001 .071 ⬍.001 <.001 <.001 .224 .035 0.58 0.60 0.67 0.61 0.61 0.56 0.61 0.66 0.67 0.70 5.51 0.69 0.51 0.52 0.56 1.59 0.85 2.28 0.45, 0.75 0.47, 0.77 0.56, 0.81 0.48, 0.78 0.48, 0.78 0.44, 0.72 0.48, 0.78 0.52, 0.83 2.55, 13.15 0.53, 0.93 3.72, 8.15 0.43, 1.11 0.26, 1.01 0.30, 0.93 0.39, 0.80 0.96, 1.86 0.69, 1.04 1.48, 3.50 <.001 <.001 <.001 <.001 <.001 <.001 <.001 .001 <.001 .017 <.001 .131 .052 .028 .002 .068 .130 <.001 .236 .177 .077 .169 .159 .141 .160 .070 .080 .308 .354 .461 .005 .006 .304 .246 .616 .700 a All the transitions were adjusted for the intermediate use of the other eight drugs and to 14 sociodemographic variables. b Individuals who used the drug in the second column before the drug in the first column. c Pearson coefficient of Test of Proportional-Hazard Assumption (link test). In bold are all the transitions that did not violate the test. d Analysis of the association between the variable gender and the second drug use prevalence, adjusted for each transition. The reference category was male. CASTALDELLI-MAIA ET AL. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 6 Figure 2. The scheme highlights all the transitions that did not violate the linktest of proportional hazardassumptions (LPHA) and could be taken into account (see details in Table 2). The arrows were drawn every time the Cox survival regression model reached significance (ⴱ p ⬍ .05). If the arrow is blue (gray), the cumulative prevalence of first use of the drug in the lefthead of the arrow was higher in users than nonusers of the drug in the righthead of the arrow (aHR ⬎ 1). If the arrow is red (gray), the cumulative prevalence of first use of the drug in the lefthead of the arrow was lower in users than nonusers of the drug in the righthead of the arrow (aHR ⬍ 1). See the online article for the color version of this figure. Discussion The main aim of the present study was to conduct an analysis of the role of first use of inhalants within the sequence of first drug use, and we found that it was preceded by the first use of alcohol, tobacco, and prescription opioids. However, the first use of inhalants preceded the first use of amphetamines, cannabis, cocaine, prescription opioids, and tranquilizers. These findings place inhalants into the position that is typically ascribed to cannabis, as its first use is more common after the first use of legal drugs and before the first use of illegal drugs. Additionally, cannabis first use did not precede inhalants first use; rather, the opposite pattern was evident. A global debate on the validity of studying the transition of various drug use (Van Etten et al., 1997; Wagner & Anthony, 2002; Herrera-Vázquez et al., 2004; Wells & McGee, 2008; Caris et al., 2009; Mayet et al., 2010) was revived because of the results of a study investigating epigenetic alterations related to drug experimentation in laboratory animals (Levine et al., 2011). Pretreatment of mice with nicotine increased their responses to cocaine, as assessed by addiction-related behaviors and synaptic plasticity in the striatum, a brain region critical for addiction-related rewards. The responses to cocaine were altered only when nicotine was administered first, and then nicotine and cocaine were administered concurrently. Reversing the order of the drug administration was ineffective, as cocaine had no effect on nicotine-induced behaviors and synaptic plasticity (Levine et al., 2011). There is also fresh epidemiologic data on this field. A recent multicenter study revealed that the sequence of drug experimentation may be linked to particular issues related to alcohol and drug use in a specific country (Degenhardt et al., 2010) and that violations of the classical sequence of drug experimentation (i.e., alcohol/tobacco ¡ cannabis ¡ other drugs) seem to be more common in countries with low lifetime use of alcohol, tobacco, and cannabis. Thus, it is important to study first use sequencing patterns in other less developed settings, such as Brazil, where there is different prevalence of lifetime drug use (e.g., higher levels of lifetime prevalence of inhalant use). Ding et al. (2009) studied the position of inhalants in the sequence of drug use in United States using data from the 2003 National Survey on Drug Use and Health (NSDUH). The age of first inhalant use was compared with the age of onset of other drug use among 6,466 inhalant users who also used at least one of 14 other drugs. The findings indicated that only 4.2% of multiple drug users had used inhalants before using other drugs, in particular, alcohol, tobacco, and cannabis. Thus, the theory that first use of inhalants is more common before the first use of illegal drugs was not supported in this study. Comparing this study (Ding et al., 2009) design with the current one, we have found many differences. In the present study, we used survival analysis methods to compare drug users and nonusers while taking into account the cumulative prevalence of other drug use across time. We chose to use Cox regression survival models following an important framework (Singer & Willett, 1994; 7 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. FIRST USE OF INHALANTS AND FIRST USE OF DRUGS Figure 3. Kaplan–Meier curves of all transitions which involved inhalants and did not violate linktest of proportional hazard-assumptions (LPHA). See the online article for the color version of this figure. Wagner & Anthony, 2002). Another explanation for the differences between the two studies could be trends over time. That is, inhalants could be more commonly used in 2009 in Brazil as compared with the United States in 2003. However, recent U.S. data show that 17% of youths initiated inhalants during middle school (Ober et al., 2013)—a considerably high rate of initiation. Our findings are in line with those from Sanchez, Ribeiro, et al. (2013) in their study with a younger sample of private school high- and middle-school students in São Paulo, Brazil. We also found that inhalants use preceded the cannabis use. In addition to that transition, our study expand this role of inhalant use— preceding the use of other illicit drugs—in a large and repre- This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 8 CASTALDELLI-MAIA ET AL. sentative sample of Brazilian urban university students. Considering that, a number of classical statements should be reconsidered in light of findings such that in our country. First, if there is a more common sequence of drug experimentation, then inhalant use should be placed in the same step as (or even before) cannabis—after legal drug use and before illegal drug use. There are few studies regarding drug transitions that take the position of inhalants in this sequence into account. The first use of inhalants preceded the first use of amphetamines, cannabis, cocaine, prescription opioids, and tranquilizers. However, the magnitude of the association between prior use of inhalants and subsequent use of these drugs varied a lot, which leads us to believe that the chance of using some drugs may be greater than others after using inhalants. Surely this is the case of prescription opioids and cocaine, with the highest aHRs. However, the absolute number of individuals who first used a drug after using inhalants was highest for cannabis and amphetamines, which should also call for the attention of policy makers. The transition to/from inhalants use seems to be a confounding factor in the association between gender and drug use. In all transitions tested from other drugs first use to inhalants first use, there was higher prevalence of inhalants first use in men compared with women, which is in line with previous findings of the prevalence of inhalants use in this sample (Andrade et al., 2012). Regarding the use of other drugs, there was higher cumulative prevalence of alcohol and cannabis first use among men, and of amphetamines and tranquilizers first use among women, which is also in line with previous findings (Andrade et al., 2012). However, the present study did not confirm higher prevalence of tobacco and cocaine first use among men, and first use of tranquilizers among women, as previously found in this sample (Andrade et al., 2012). This leads us to believe that adjusting for previous inhalant first use made the gender differences in cocaine and tranquilizers first use disappear, positing first use of inhalants as a confounding factor. Other studies could investigate the role of inhalants previous use as a confounding factor for associations between gender and drug use. There is a consensus among the few studies regarding inhalants from across all of the areas in the addiction field (Howard et al., 2011; Konghom et al., 2010), and our results support the implication that more attention to inhalant use and in its consequences and correlates are needed, especially in Brazil. Nonnemaker et al. (2011) suggested a strong relationship between inhalant use, other problem behaviors, and sensation-seeking. These results highlight the need to intervene early with youths who are at risk of or just beginning to engage in inhalant use, given that low-frequency inhalant users report predominately hedonic experiences during inhalant intoxication that are not aversive (Garland & Howard, 2010). Even nonpersistent use of inhalants is associated with a greater risk for nonmedical opioid and stimulant use and is a strong predictor of nonpersistent opioid and stimulant use (Nakawaki & Crano, 2012). Limitations An important limitation of this study is the recall bias, which is inherent in cross-sectional studies (Raphael, 1987). In addi- tion, it is also difficult to extrapolate from data obtained from university students to the Brazilian general population. College students in this country are not representative of the general population, as only 13.9% of young adults have access to higher education. Almost 50% of university students study in private institutions, which sets them apart from the general population. However, we did use a representative sample from all of the Brazilian capitals, thus providing balanced data from this large country, with various social, cultural, and economic differences (Victora et al., 2010; Parahyba & Veras, 2008). In addition, 8 of 36 transitions tested violated the LPHA), which implies the absence of important transitions (i.e., between cannabis and tobacco). Another limitation is that if a respondent did not provide the age of first use of a drug, we could not use this individual’s data to perform any analysis of a transition that involved this drug. In view of the large number of people who did not remember at least one of the ages of experimenting with some of the drugs, we decided to carry out a comparative analysis between individuals who remembered all the ages of experimentation with drugs and those who did not remember at least one of the ages (missing group). To this end, we analyzed the subsample of those individuals who used drugs at least 1 drug of all the following three groups: (i) alcohol and/or tobacco; (ii) marijuana and/or inhalants; (iii) cocaine and/or hallucinogens and/or ecstasy and/or amphetamine and/or prescription opioids and/or tranquilizers. From this analysis (see the Appendix), individuals who were older, had a religion, not practiced their religion, used hallucinogens, used ketamine, did not use tranquilizers, did not use barbiturates, and did not use codeine syrup were associated with the missing group. Therefore, one should take care when extrapolating our findings to these subpopulations. Future analyses on first drug use sequencing patterns involving hallucinogens and tranquilizers would be interesting to better evaluate these transitions. Conclusion This study highlights the need to intervene early with youths who are at risk of or are just beginning to engage in inhalant use, because this class of drugs seems to be the first to be experimented with in our sample. There is also a call for attention to individuals who have already first used inhalants because of their higher chance to experiment with other drugs such as cannabis, cocaine, and prescription drugs. It was also interesting to find that cannabis use was subsequent compared with inhalants within first use sequencing pattern. All these findings show an in-transition culture of drug use among young adults in Brazil. References Andrade, A. G., Duarte, P. do C., Barroso, L. P., Nishimura, R., Alberghini, D. G., & Oliveira, L. G. (2012). Use of alcohol and other drugs among Brazilian college students: Effects of gender and age. Revista Brasileira de Psiquiatria, 34, 294 –305. doi:10.1016/j.rbp.2012.02.002 Andrade, A. G., Queiroz, S., Villaboim, R. C. M., César, C. L. G., Alves, M. C. G. P., & Bassit, A. Z. (1997). Uso de álcool e drogas entre alunos de graduação da Universidade de São Paulo. Revista ABP-APAL, 19, 53–59. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. FIRST USE OF INHALANTS AND FIRST USE OF DRUGS Balster, R. L., Cruz, S. L., Howard, M. O., Dell, C. A., & Cottler, L. B. (2009). Classification of abused inhalants. Addiction, 104, 878 – 882. doi:10.1111/j.1360-0443.2008.02494.x Caris, L., Wagner, F. A., Rios-Bedoya, C. F., & Anthony, J. C. (2009). Opportunities to use drugs and stages of drug involvement outside the United States: Evidence from Republic of Chile. Drug and Alcohol Dependence, 102, 30 –34. doi:10.1016/j.drugalcdep.2008.12.004 Castaldelli-Maia, J. M., Bhugra, D., de Andrade, A. G., & Lotufo-Neto, F. (2013). Substance use and misuse in Brazilian Movies (2000 –2008). Substance Use & Misuse, 48, 248 –257. doi:10.3109/10826084.2012 .753549 Castaldelli-Maia, J. M., Martins, S. S., de Oliveira, L. G., de Andrade, A. G., & Nicastri, S. (2014). The role of drug use sequencing pattern in further problematic use of alcohol, tobacco, cannabis, and other drugs. Journal of Mental Health [in press]. Castaldelli-Maia, J. M., Martins, S. S., de Oliveira, L. G., van Laar, M., de Andrade, A. G., & Nicastri, S. (2014). Use transitions between illegal drugs among Brazilian university students. Social Psychiatry and Psychiatric Epidemiology, 49, 385–394. doi:10.1007/s00127-013-0748-1 Castaldelli-Maia, J. M., Oliveira, H. P., Andrade, A. G., Lotufo-Neto, F., & Bhugra, D. (2012). Using selected scenes from Brazilian films to teach about substance use disorders, within medical education. Sao Paulo Medical Journal, 130, 380 –391. doi:10.1590/S1516-318020 12000600005 Cleves, M., Gutierrez, R. G., Gould, W., & Marchenko, Y. V. (2010). An introduction to survival analysis using Stata (3rd ed.). College Station, TX: Stata Press. Cox, D. R. (1972). Regression models and life time tables. Journal of the Royal Statistical Society (Series B), 34, 187–202. Degenhardt, L., Dierker, L., Chiu, W. T., Medina-Mora, M. E., Neumark, Y., Sampson, N., . . . Kessler, R. C. (2010). Evaluating the drug use “gateway” theory using cross- national data: Consistency and associations of the order of initiation of drug use among participants in the WHO World Mental Health Surveys. Drug and Alcohol Dependence, 108, 84 –97. doi:10.1016/j.drugalcdep.2009.12.001 Ding, K., Chang, G. A., & Southerland, R. (2009). Age of inhalant first time use and its association to the use of other drugs. Journal of Drug Education, 39, 261–272. doi:10.2190/DE.39.3.c Fergusson, D. M., Boden, J. M., & Horwood, L. J. (2006). Cannabis use and other illicit drug use: Testing the cannabis gateway hypothesis. Addiction, 101, 556 –569. doi:10.1111/j.1360-0443.2005.01322.x Fonseca, A. M., Galduroz, J. C., Noto, A. R., & Carlini, E. L. (2010). Comparison between two household surveys on psychotropic drug use in Brazil: 2001 and 2004. Ciencia e Saude Coletiva, 15, 663– 670. Garland, E. L., & Howard, M. O. (2010). Phenomenology of adolescent inhalant intoxication. Experimental and Clinical Psychopharmacology, 18, 498 –509. doi:10.1037/a0021737 Herrera-Vázquez, M., Wagner, F. A., Velasco-Mondragon, E., Borges, G., & Lazcano-Ponce, E. (2004). Inicio en el consumo de alcohol y tabaco y transición a otras drogas en estudiantes de Morelos, México. Salud Publica de México, 46, 132–140. doi:10.1590/S0036-36342004000200007 Howard, M. O., Bowen, S. E., Garland, E. L., Perron, B. E., & Vaughn, M. G. (2011). Inhalant use and inhalant use disorders in the United States. Addiction Science and Clinical Practice, 6, 18 –31. Hynes-Dowell, M., Mateu-Gelabert, P., Barros, H. M. T., & Delva, J. (2011). Volatile substance misuse among high school students in South America. Substance Use & Misuse, 46, 27–34. doi:10.3109/10826084 .2011.580192 Kandel, D. B. (1975). Stages in adolescent involvement in drug use. Science, 190, 912–914. doi:10.1126/science.1188374 Kandel, D., & Faust, R. (1975). Sequence and stages in patterns of adolescent drug use. Archives of General Psychiatry, 32, 923–932. doi:10.1001/archpsyc.1975.01760250115013 9 Levine, A., Huang, Y., Drisaldi, B., Griffin, E. A.Jr., Pollak, D. D., Xu, S., . . . Kandel, E. R. (2011). Molecular mechanism for a gateway drug: Epigenetic changes initiated by nicotine prime gene expression by cocaine. Science Translational Medicine, 3, 107–109. doi:10.1126/ scitranslmed.3003062 Malta, D. C., Sardinha, L. M., Mendes, I., Barreto, S. M., Giatti, L., Castro, I. R., . . . Crespo, C. (2010). Prevalence of risk health behavior among adolescents: Results from the 2009 National Adolescent School-based Health Survey (PeNSE). Ciencia e Saude Coletiva, 15, 3009 –3019. doi:10.1590/S1413-81232010000800002 Mayet, A., Lagleye, S., Chau, N., & Falissard, B. (2010). The mediation role of licit drugs in the influence of socializing on cannabis use among adolescents: A quantitative approach. Addictive Behaviors, 35, 890 – 895. doi:10.1016/j.addbeh.2010.06.001 Mesquita, A. M., de Andrade, A. G., & Anthony, J. C. (1998). Use of the inhalant lança by Brazilian medical students. Substance Use & Misuse, 33, 1667–1680. doi:10.3109/10826089809058949 Moreira, M. M., Barbosa, G. L., Laranjeira, R., & Mitsuhiro, S. S. (2014). Alcohol and crack/cocaine use in women: A 14-year cross-sectional study. Journal of Addiction Diseases. [Epub ahead of print] Nakawaki, B., & Crano, W. D. (2012). Predicting adolescents’ persistence, non-persistence, and recent onset of nonmedical use of opioids and stimulants. Addictive Behaviors, 37, 716 –721. doi:10.1016/j.addbeh .2012.02.011 Nonnemaker, J. M., Crankshaw, E. C., Shive, D. R., Hussin, A. H., & Farrelly, M. C. (2011). Inhalant use initiation among U.S. adolescents: Evidence from the National Survey of Parents and Youth using discretetime survival analysis. Addictive Behaviors, 36, 878 – 881. doi:10.1016/ j.addbeh.2011.03.009 Ober, A. J., Miles, J. N., Ewing, B., Tucker, J. S., & D’Amico, E. J. (2013). Risk for inhalant initiation among middle school students: Understanding individual, family, and peer risk and protective factors. Journal of Studies on Alcohol and Drugs, 74, 835– 840. Opaleye, E. S., Ferri, C. P., Locatelli, D. P., Amato, T. C., & Noto, A. R. (2014). Nonprescribed use of tranquilizers and use of other drugs among Brazilian students. Revista Brasileira de Psiquiatria, 36, 16 –23. doi: 10.1590/1516-4446-2013-1180 Parahyba, M. I., & Veras, R. (2008). Socio-demographic differentials in the functional decline among the elderly in Brazil. Ciencia & Saude Coletiva, 13, 1257–1264. doi:10.1590/S1413-81232008000400022 Raphael, K. (1987). Recall bias: A proposal for assessment and control. International Journal of Epidemiology, 16, 167–170. doi:10.1093/ije/16 .2.167 Remy, L., Narvaez, J., Sordi, A., Guimarães, L. S., Von Diemen, L., Surratt, H., . . . Pechansky, F. (2013). Correlates of unprotected sex in a sample of young club drug users. Clinics (Sao Paulo), 68, 1384 –1391. doi:10.6061/clinics/2013(11)01 Sanchez, Z. M., Noto, A. R., & Anthony, J. C. (2013). Social rank and inhalant drug use: The case of lança perfume use in São Paulo, Brazil. Drug and Alcohol Dependence, 131, 92–99. doi:10.1016/j.drugalcdep .2012.12.001 Sanchez, Z. M., Ribeiro, L. A., Moura, Y. G., Noto, A. R., & Martins, S. S. (2013). Inhalants as intermediate drugs between legal and illegal drugs among middle and high school students. Journal of Addictive Diseases, 32, 217–226. doi:10.1080/10550887.2013.795472 Silva-Oliveira, F., Jorge, K. O., Ferreira e Ferreira, E., Vale, M. P., Kawachi, I., & Zarzar, P. M. (2014). The prevalence of inhalant use and associated factors among adolescents in Belo Horizonte, Brazil. Ciencia e Saude Coletiva, 19, 881– 890. doi:10.1590/1413-81232014193 .07542013 Singer, J. D., & Willett, J. B. (1994). Designing and analyzing studies of onset, cessation, and relapse: Using survival analysis in drug abuse research. NIDA Research Monograph, 142, 196 –263. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 10 CASTALDELLI-MAIA ET AL. Stempliuk, V. A., Barroso, L. P., Andrade, A. G., Nicastri, S., & Malbergier, A. (2005). Comparative study of drug use among undergraduate students at the University of São Paulo-São Paulo campus in 1996 and 2001. Revista Brasileira de Psiquiatria, 27, 185–193. doi:10.1590/ S1516-44462005000300006 Tarter, R. E., Vanyukov, M., Kirisci, L., Reynolds, M., & Clark, D. B. (2006). Predictors of marijuana use in adolescents before and after licit drug use: Examination of the gateway hypothesis. The American Journal of Psychiatry, 163, 2134 –2140. doi:10.1176/appi.ajp.163.12.2134 Van Etten, M. L., Neumark, Y. D., & Anthony, J. C. (1997). Initial opportunity to use marijuana and the transition to first use: United States, 1979 –1994. Drug and Alcohol Dependence, 49, 1–7. doi:10.1016/ S0376-8716(97)00127-0 Vaughn, M. G., Wallace, J., Perron, B. E., Copeland, V., & Howard, M. O. (2008). Does marijuana use serve as a gateway to cigarette use for African-Americans? The American Journal of Drug and Alcohol Abuse, 34, 782–791. doi:10.1080/00952990802455477 Victora, C. G., Matijasevich, A., Silveira, M., Santos, I., Barros, A. J., & Barros, F. C. (2010). Socio-economic and ethnic group inequities in antenatal care quality in the public and private sector in Brazil. Health Policy and Planning, 25, 253–261. doi:10.1093/heapol/czp065 Wagner, F. A., & Anthony, J. C. (2002). Into the world of illegal drug use: Exposure opportunity and other mechanisms linking the use of alcohol, tobacco, marijuana, and cocaine. American Journal of Epidemiology, 155, 918 –925. doi:10.1093/aje/155.10.918 Wagner, G. A., Stempliuk, V. A., Zilberman, M. L., Barroso, L. P., & Andrade, A. G. (2007). Alcohol and drug use among university students: Gender differences. Revista Brasileira de Psiquiatria, 29, 123–129. doi:10.1590/S1516-44462006005000033 Wells, J. E., & McGee, M. A. (2008). Violations of the usual sequence of drug initiation: Prevalence and associations with the development of dependence in the New Zealand Mental Health Survey. Journal of Studies on Alcohol and Drugs, 69, 789 –795. (Appendix follows) FIRST USE OF INHALANTS AND FIRST USE OF DRUGS 11 Appendix This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Results of Logistic Regression Comparing Individuals Who Used at Least 3 Drugs and Were Reminded of All the Ages of Experimentation of These Drugs and Those Who Any Age of Experimentation Was Missing Among Brazilian University Students, 2009 Variable aOR SE p Age Gender Type of religion Practicing the religion Scholarity of the head of the family Socioeconomic status Ethnic group Civil status Children Currently working Driver’s license University course area Semester of the course All-day course Happy with the course choice Alcohol use Tobacco use Cannabis use Inhalants use Cocaine use Hallucinogens use Ketamine use Ayahuasca use Ecstasy use Steroids use Tranquilizers use Sedative/barbiturates use Opioid analgesics use Codeine syrup use Anticholinergics use Heroin use Amphetamine use Synthetic drugs use 1.92 1.01 1.27 0.64 0.81 1.04 0.69 0.78 0.56 1.22 0.56 1.18 0.96 0.98 0.81 0.90 0.83 1.19 0.99 1.23 2.59 20.85 1.49 0.62 0.59 0.54 0.17 0.45 0.05 2.00 6.63 0.69 1.97 0.51 0.23 0.11 0.12 0.12 0.03 0.16 0.20 0.27 0.17 0.16 0.17 0.09 0.31 0.33 0.86 0.15 0.38 0.33 0.38 1.20 25.32 0.85 0.23 0.27 0.11 0.13 0.19 0.04 1.83 11.47 0.28 1.93 .018 .936 .008 .023 .177 .238 .125 .364 .250 .166 .056 .244 .735 .954 .607 .918 .330 .591 .979 .498 .044 .014 .486 .217 .259 .005 .028 .070 .002 .452 .277 .391 .490 [95% CI] 1.12, 0.64, 1.06, 0.44, 0.60, 0.97, 0.43, 0.46, 0.21, 0.91, 0.31, 0.88, 0.79, 0.51, 0.36, 0.13, 0.57, 0.62, 0.50, 0.66, 1.02, 1.86, 0.47, 0.29, 0.24, 0.35, 0.03, 0.18, 0.01, 0.32, 0.21, 0.30, 0.28, 3.28 1.61 1.52 0.94 1.09 1.11 1.10 1.32 1.50 1.63 1.01 1.58 1.17 1.86 1.82 6.06 1.20 2.27 1.95 2.30 6.55 233.25 4.65 1.32 1.46 0.83 0.82 1.06 0.33 1.23 206.65 1.59 1.39 Received April 2, 2014 Revision received July 10, 2014 Accepted July 11, 2014 䡲