Towards the characterization and validation of alcohol use disorder subtypes: integrating consumption and symptom data.

BACKGROUND: There is evidence that measures of alcohol consumption, dependence and abuse are valid indicators of qualitatively different subtypes of alcohol involvement yet also fall along a continuum. The present study attempts to resolve the extent to which variations in alcohol involvement reflect a difference in kind versus a difference in degree. METHOD: Data were taken from the 2001-2002 National Epidemiologic Survey of Alcohol and Related Conditions. The sample (51% male; 72% white/non-Hispanic) included respondents reporting past 12-month drinking at both waves (wave 1: n = 33644; wave 2: n = 25186). We compared factor mixture models (FMMs), a hybrid of common factor analysis (FA) and latent class analysis (LCA), against FA and LCA models using past 12-month alcohol use disorder (AUD) criteria and five indicators of alcohol consumption reflecting frequency and heaviness of drinking. RESULTS: Model comparison revealed that the best-fitting model at wave 1 was a one-factor four-class FMM, with classes primarily varying across dependence and consumption indices. The model was replicated using wave 2 data, and validated against AUD and dependence diagnoses. Class stability from waves 1 to 2 was moderate, with greatest agreement for the infrequent drinking class. Within-class associations in the underlying latent factor also revealed modest agreement over time. CONCLUSIONS: There is evidence that alcohol involvement can be considered both categorical and continuous, with responses reduced to four patterns that quantitatively vary along a single dimension. Nosologists may consider hybrid approaches involving groups that vary in pattern of consumption and dependence symptomatology as well as variation of severity within group.

Transitioning into and out of large-effect drinking in young adulthood.

As individuals age beyond the college years into young adulthood, many exhibit a tendency to moderate or "mature out of" alcohol involvement. The current study classified effect-drinking statuses in young adults and examined transitions among statuses using latent transition analysis, a latent variable state-sequential model for longitudinal data. At 3 occasions over 7 years (Years 1, 4, and 7), 443 men (47%) and women (mean age of both at baseline = 18.5 years; 51% with family history of alcoholism) responded to 3 past-30-day items assessing drinking and subjective effects of drinking: whether the respondent drank alcohol, felt high, and felt drunk. Latent statuses included abstainers (14% at Year 1), limited-effect drinkers (8%), moderate-effect drinkers (23%), and large-effect drinkers (54%). Respondents with family history of alcoholism were less likely to transition out of large-effect drinking than those without family history. Men exhibited more severe initial effect-drinking statuses and lower transition probabilities into less severe effect-drinking statuses than women.

Trajectories of concurrent substance use disorders: a developmental, typological approach to comorbidity.

BACKGROUND: Although alcohol and tobacco use disorders are highly comorbid, little is known regarding the combined course of these disorders. METHODS: The current study utilized latent class analysis to examine longitudinal patterns of alcohol-tobacco use disorder comorbidity. Participants were 449 young adults (baseline age, 18.5 years; 48% male; 51% with paternal history of alcoholism) assessed five times over 7 years. RESULTS: Five longitudinal types of alcohol or tobacco use disorder over time were identified: nondiagnosing; developmentally limited alcohol use disorder; chronic alcohol use disorder; chronic tobacco use disorder; and comorbid alcohol and tobacco use disorder. These typologies were distinguishable on the basis of family history of alcoholism status and sex. Etiologically important third variables (alcohol expectancies, behavioral undercontrol, childhood stressors) mediated the relation between family history and the latent classes. CONCLUSIONS: Characterizations of alcohol use disorders typically fail to consider important sources of heterogeneity such as course or comorbidity. By simultaneously modeling developmental course and comorbidity with tobacco dependence, we were able to identify distinct trajectories of single and concurrent substance use disorders. Such multi-substance trajectories represent meaningful subtypes that, although sharing substantial common influences, have unique etiologic correlates. Additionally, these subtypes might represent distinct groups from the perspective of intervention strategies.

Prospective analysis of comorbidity: tobacco and alcohol use disorders.

Alcohol use disorders (AUD) and tobacco use disorders (TD) frequently co-occur. The authors examined AUD-TD comorbidity over time using a state-trait (ST) model. The ST model represents variance in AUD/TD as a traitlike factor that spans measurement occasion and identifies distinct sources of variance in AUD-TD comorbidity. The ST model was evaluated on 450 young adults (baseline age = 18.5 years; 51% with family history of alcoholism) assessed 5 times over 7 years. The ST model demonstrated superior fit over a first-order autoregressive model. The tendency to diagnose with AUD and TD was partially explained by family history of alcoholism; this relationship was mediated by childhood stressors, alcohol expectancies, and behavioral undercontrol. Results supported a common third-variable influence (vs. directional) model of comorbidity. The ST model is an important conceptual and methodological approach to the prospective study of comorbidity in general.

Two separate verbal processing rates contributing to short-term memory span.

Previous research indicates that verbal memory span, the number of words people can remember and immediately repeat, is related to the fastest rate at which they can pronounce the words. This relation, in turn, has been attributed to a general or global rate of information processing that differs among individuals and changes with age. However, the experiments described in this article showed that the rates of 2 processes (rapid articulation and the retrieval of words from short-term memory) are related to memory span but not to each other. Memory span depends on a profile of processing rates in the brain, not only a global rate. Moreover, there appears to be only a partial overlap between the rate variables that change with age and those that differ among individuals.

Predicting stability and change in frequency of intoxication from the college years to beyond: individual-difference and role transition variables.

The authors examined whether individual-difference variables (e.g., family history of alcoholism, sex, personality traits, positive alcohol expectancies) and role transition-related variables (full-time work status, marital status, parenthood) moderate the "maturing-out" process whereby young adults who drink heavily during college decrease their drinking in the following years. Analyses were based on 288 young adults, assessed as full-time students (mostly college seniors, Year 4 of a larger study) and 3 years later (Year 7) when all had earned bachelor's degrees, and the analyses showed that frequency of intoxication (per week) decreased significantly (p < .0001). Entering the workforce full time, being male, and being less open to experience were associated with decreased postcollege drinking. Furthermore, relatively extraverted individuals were more likely to continue a pattern of frequent intoxication from Year 4 to year 7. The findings stress the importance of studying how individual-difference variables predict behavior across role transitions.

Alcohol use disorders and neuropsychological functioning in first-year undergraduates.

The relation between alcohol use disorders and neuropsychological functioning was examined in 489 first-year undergraduates, approximately half of whom had a history of alcoholism in their biological fathers. Factor analyses of 17 neuropsychological tests and subtests produced the following 5 factors that were the basis of subsequent analyses: Language/Verbal Memory, Visuospatial Ability, Motor Speed, Booklet Category Performance, and Attention. Participants with alcohol use disorders showed deficits in visuospatial ability. Those who had alcohol dependence showed deficits in both visuospatial ability and motor speed relative to participants who abused alcohol. The differences in neuropsychological functioning remained even after several potential confounding variables were controlled statistically. Results suggest that alcohol use disorders in first-year college students are associated with deficits in neuropsychological measures that are not attributable to several potential third-variable explanations.

Predicting academic problems in college from freshman alcohol involvement.

OBJECTIVE: The present article examines the relation of problematic alcohol use to collegiate academic problems based on a systematic assessment of problematic alcohol use and college transcript data. The degree to which this prospective association can be explained by reference to third variables is also explored. These third variables include: students' high school academic achievement and aptitude, concurrent drug use, participation in deviant behaviors and students' investment or participation in the college experience. METHOD: A sample of 444 (240 female) college freshman recruited for a longitudinal study of alcohol use was followed for 6 years. Alcohol and drug involvement, general deviance, academic investment, campus involvement and several background variables were assessed during the freshman year. Additional measures of high school aptitude and achievement as well as collegiate performance were calculated based on college transcript data from all institutions attended. RESULTS: A latent variable structural equation model revealed that problematic alcohol use during the freshman year correlated +.32 with collegiate academic problems. No evidence was found for a unique association between the two constructs when additional constructs were included in the model. Specifically, the association was substantially reduced when preexisting student differences traditionally associated with academic failure in college were taken into account. The inclusion of concurrent drug use and deviance also resulted in a significant reduction in the magnitude of the association. CONCLUSIONS: Although a substantial bivariate association exists between problematic alcohol use and academic problems during college, much of this association appears attributable to preexisting student differences on admission to college.

A prospective, high-risk study of the relationship between tobacco dependence and alcohol use disorders.

This study examined the extent to which tobacco dependence (TD) and alcohol use disorders (AUDs) reciprocally influenced each other in a mixed-gender sample of 452 individuals (n = 232 biological family history of paternal alcoholism, n = 220 no first- or second-degree family history of alcoholism) who were assessed once early in their freshman year of college, approximately 3 years later when many were college seniors, and approximately 3 years later when many had entered or were entering the work force. AUDs were more prevalent in men than women, in individuals with a family history of alcoholism, and decreased overall with time. TD was more prevalent in those with a family history of alcoholism, showed increasing rates of use over time, and was less prevalent but more stable than AUDs. Transitional probabilities indicated that although a previous AUD or TD diagnosis increased the likelihood of being diagnosed with the other disorder at a later time, comorbid AUDs and TD did not significantly affect the likelihood of recovery from either disorder. Finally, path analysis revealed significant reciprocal relationships between AUDs and TD diagnoses (each predicting the other over time), and significant prediction of AUDs and TD by family history of alcoholism at the first and third times of assessment. Findings supported two general models of AUD/TD comorbidity: a shared vulnerability model and a reciprocal influence model.

Anxiety and drinking behavior: moderating effects of tension-reduction alcohol outcome expectancies.

We evaluated whether alcohol outcome expectancies moderate the association between measures of anxiety and alcohol use. Student subjects completed questionnaires related to their level of anxiety, recent alcohol-use patterns, and outcome expectancies for alcohol to be tension reducing. Interviews were used to determine the presence or absence of alcohol dependence in subjects and in their first- and second-degree relatives. Consistent with predictions, male subjects with high tension-reduction alcohol outcome expectancies showed a stronger positive correlation between measures of anxiety and drinking behavior than did male subjects with low tension-reduction outcome expectancies. However, this effect was not found for female subjects. We note past studies showing similar gender effects, and relate the overall study findings to the tension-reduction hypothesis of stress-induced drinking.

Characteristics of children of alcoholics: putative risk factors, substance use and abuse, and psychopathology.

A sample of 253 children of alcoholics (COAs) and 237 children of nonalcoholics (non-COAs) were compared on alcohol and drug use, psychopathology, cognitive ability, and personality. COAs reported more alcohol and drug problems, stronger alcohol expectancies, higher levels of behavioral undercontrol and neuroticism, and more psychiatric distress in relation to non-COAs. They also evidenced lower academic achievement and less verbal ability than non-COAs. COAs were given Diagnostic Interview Schedule alcohol diagnoses more frequently than non-COAs. The relation between paternal alcoholism and offspring alcohol involvement was mediated by behavioral undercontrol and alcohol expectancies. Although gender differences were found, there were few Gender X Family History interactions; the effects of family history of alcoholism were similar for men and women. When gender effects were found, they showed greater family history effects for women.

Identification of a conjunctivitis-associated gene locus from the virulence plasmid of Yersinia enterocolitica.

The virulence plasmid (pYV) of Yersinia enterocolitica is necessary for production of conjunctivitis in guinea pigs and for mouse lethality. To identify the genes responsible for production of conjunctivitis in guinea pigs, we subcloned the BamHI and SalI restriction fragments of the virulence plasmid of Y. enterocolitica A2635 (serotype O:8) into derivatives of the broad-host-range plasmid pRK290 and introduced the constructions into plasmid-negative Y. enterocolitica strains. A mild, transient conjunctivitis was evident 24 h after inoculation with strains containing a 2.8-kilobase (kb) BamHI fragment of pYV. These strains were cytotoxic to HEp-2 cells but did not cause death in iron-loaded adult mice. When the 2.8- and adjacent 0.5-kb BamHI fragments were deleted from the virulence plasmid of a fully virulent Y. enterocolitica isolate, the resultant strain did not cause conjunctivitis in guinea pigs and was not cytotoxic to HEp-2 cells. However, the strain with the deletion appeared to be more virulent for mice, with more rapid dissemination after orogastric inoculation, compared with that of the parent strain. When the deletion was complemented by introduction of a plasmid containing the 2.8-kb BamHI fragment, the strain again caused conjunctivitis but had decreased virulence for mice.

Activation of c-Ki-ras in human gastrointestinal dysplasias determined by direct sequencing of polymerase chain reaction products.

Activation of c-Ki-ras by point mutation within exon 1 was studied in 33 specimens of dysplastic gastrointestinal lesions or of cancers presumed to arise from dysplasia. Samples were obtained from patients with underlying ulcerative colitis or Barrett's esophagus, two diseases associated with dysplasia and increased rates of colonic or esophageal adenocarcinoma, respectively. Genomic DNA was amplified using primers bounding this exon in the polymerase chain reaction. Polymerase chain reaction products were analyzed by direct dideoxy sequencing. Three point mutations in codon 13 of c-Ki-ras were found, all in colonic specimens (two high-grade dysplasias and one adenocarcinoma arising in ulcerative colitis). No point mutations were observed in the second exon of c-Ki-ras or in and around codons 12, 13, and 61 of c-N-ras and C-Ha-ras in a partial sampling of the specimens. These data indicate that ras family protooncogene activation is an uncommon event at this level of malignant progression in these disease states. Carcinogenesis in ulcerative colitis and Barrett's esophagus may proceed via different pathways than in sporadic colon cancer, perhaps involving loss or inactivation of suppressor genes.

Sequencing products of the polymerase chain reaction directly, without purification.

An improvement over current protocols for sequencing products of the polymerase chain reaction is described. This method allows sequencing products of the reaction without performing costly, time-consuming purification steps which often result in unacceptable loss of product. Conservation of small amounts of polymerase chain reaction products which can be obtained from limited DNA sources, such as tissue biopsies, is achieved. Clarity of autoradiograms obtained utilizing this adaptation is comparable to that obtained with the original method. In addition to streamlining the amplification-sequencing procedure, this procedure can potentially be subjected to total automation.

Rationale, Detection, and Implications of Interactions Between Independent Variables and Unmeasured Variables in Linear Models.

Conceptual rationales from five research contexts are presented which share the similarity that all posit unmeasured variables which interact with observed independent variables to produce a complete model of the dependent variable. For the class of linear underspecified models involving only observed effects, it is shown that estimated effects of the observed independent variables are overestimated, that squared error of the misspecified model overestimates true error, and that this misspecified squared error is a function of at least the square of the measured independent variable(s). A statistical test for this last relationship is presented (Cook & Weisberg, 1983). The effect of this phenomenon on statistical power is discussed. Possible analytic and design strategies to overcome the effect of such unmeasured interactions include use of the rank transform (Conover & Iman, 1981), use of several longitudinal assessments, and oversampling of levels of independent variables where variance heterogeneity is large relative to other levels. Limitations of these strategies are discussed.

Investigation of an outbreak of Salmonella enteritidis gastroenteritis associated with consumption of eggs in a restaurant chain in Maryland.

Salmonella enteritidis ser. enteritidis was isolated from patrons and employees of three restaurants in a restaurant chain in Maryland during August and September 1985. Isolates from all three restaurants had identical plasmid profiles; this profile was present in 13 of 40 randomly selected S. enteritidis isolates received by the Maryland state health department laboratory during a comparable time period. The outbreak in one restaurant resulted in at least 71 illnesses, with 17 persons known to have been hospitalized. Scrambled eggs served on a "breakfast bar" were implicated as the vehicle of transmission in this restaurant, with eggs a possible vehicle in another of the three restaurants. The data point out the risks associated with improper handling of eggs in food service establishments, provide further evidence for the observed association between S. enteritidis and eggs in the northeastern United States, and demonstrate the utility of plasmid analysis in investigation of outbreaks involving common Salmonella serotypes.

DNA probes for identification of enteroinvasive Escherichia coli.

Eighty-one Escherichia coli strains belonging to all known invasive O serogroups were tested with two distinct invasiveness probes (pMR17 and pSF55). All 54 Sereny test-positive strains and 5 strains that lost Sereny positivity during storage hybridized with both probes. Probe-positive strains carried a 120- to 140-megadalton plasmid, did not produce lysine decarboxylase, and, with the exception of certain serotypes, were nonmotile. Motile strains of serotype O144:H25 were for the first time characterized as invasive by hybridization with the probes.

Association between specific plasmids and relapse in typhoid fever.

We studied isolates from 73 patients hospitalized with typhoid fever in Lima, Peru. Of these 73 patients, 11 (15%) suffered a clinical relapse, with fever and positive blood cultures, within 3 months of their original illness. Using plasmids as epidemiologic markers, we found that three patients who subsequently relapsed were initially infected with more than one strain of Salmonella typhi. There was a highly significant association between relapse and isolation of a strain containing either a 24- or a 38-kilobase plasmid at the time of the original infection; however, we were unable to show any evidence of homology between these two plasmids. Our data indicate that infection with multiple strains is not uncommon in this endemic area and suggest that relapse may be partly strain dependent.

Identification of environmental Vibrio vulnificus isolates with a DNA probe for the cytotoxin-hemolysin gene.

We screened 44 lactose-positive Vibrio strains isolated from the marine environment for homology with a 3.2-kilobase DNA fragment encoding the Vibrio vulnificus cytotoxin-hemolysin gene. All 29 marine isolates identified as V. vulnificus on the basis of numerical taxonomy and DNA-DNA hybridization studies hybridized with the cytotoxin gene probe, as did all V. vulnificus reference strains. Homologous gene sequences were identified in no other lactose-positive marine vibrio isolates nor in 10 other Vibrio species.