This issue features a review article on costs of eating disorders, two papers on bulimia, groups of papers on genetics, family studies in obsessive–compulsive disorder, treatment, and an additional paper on catatonia.

Background. The economic burden and health service use of eating disorders have received little attention, although such data are necessary to estimate the implications of any changes in clinical practice for patient care and health care resource requirements. This systematic review reports the current international evidence on the resource use and cost of eating disorders.

Method. Relevant literature (1980–2002) was identified from searches of electronic databases and expert contacts.

Results. Two cost-of-illness studies from the UK and Germany, one burden-of-disease study from Australia and 14 other publications with relevant data from the UK, USA, Austria, Denmark and The Netherlands could be identified. In the UK, the health care cost of anorexia nervosa was estimated to be £4·2 million in 1990. In Germany, the health care cost was €65 million for anorexia nervosa and €10 million for bulimia nervosa during 1998. The Australian study reported the health care costs of eating disorders to be Aus$22 million for year 1993/1994. Other costing studies focused mostly on in-patient care reporting highly variable estimates. There is a dearth of research on non-health care costs.

Conclusions. The limited available evidence reflects a general under-detection and under-treatment of eating disorders. Although both cost-of-illness studies may significantly underestimate the costs of eating disorders because of important omitted cost items, other evidence suggests that the economic burden is likely to be substantial. Comprehensive data on the resource use of patients with eating disorders are urgently needed for better estimations, and to be able to determine cost-effective treatment options.

Background. In bulimic syndromes, binge episodes are thought to be caused by dietary restraint and negative moods. However, as central serotonin (5-hydroxytryptamine: 5-HT) mechanisms regulate appetite and mood, the 5-HT system could be implicated in diet- and mood-based binge antecedents.

Method. We used hand-held computers to obtain repeated ‘online’ measurements of eating behaviors, moods, and self-concepts in 21 women with bulimic syndromes, and modeled 5-HT system activity with a measure of platelet [3H]paroxetine-binding density.

Results. Mood and self-concept ratings were found to be worse before binge episodes (than at other moments), and cognitive restraint was increased. After binges, mood and self-concept deteriorated further, and thoughts of dieting became more intense. Intriguingly, lower paroxetine-binding density predicted poorer mood and self-concept before a binge, larger post-binge decrements in mood and self-concept, and larger post-binge increases in dietary restraint.

Conclusions. Paroxetine binding thus seemed to reflect processes that impacted upon mood-related antecedents to binge episodes, and consequences implicating mood and dietary restraint.

Background. There is an ongoing debate about the best way to classify eating disorders. This study examined potential subtypes of bulimia nervosa.

Method. Latent class analysis (LCA) was used to identify subtypes of bulimic symptomatology, utilizing data from 234 respondents in a cohort of black and white young women (n=2054). Participants were administered gated screening questions from the Eating Disorders Examination to determine DSM-IV diagnoses of eating disorders.

Results. A 3-class solution was judged best. Class 1, the ‘purger subtype’ (n=116), was characterized by vomiting, the use of fasting/diet pills, and relatively little bingeing. Class 2, the ‘binger subtype’ (n=97) comprised women who engaged in bingeing but minimal compensatory behaviors. Class 3, the ‘binge-purger subtype’ (n=21) had relatively high rates of all symptoms. Findings of differences between the three subtypes on validator variables and differences between the three subtypes compared to non-eating disorder groups suggest validity of the three bulimic subtypes identified in our analyses. Ethnicity and class membership were associated [χ2(3)=21·89, p<0·0001], reflecting a greater percentage of white women than black women in Class 1 and a greater percentage of black women than white women in Class 2.

Conclusions. LCA revealed one subtype that was similar to bulimia nervosa and two subtypes of bulimic symptomatology that did not meet criteria for bulimia nervosa yet appear to be clinically significant. Further study of the psychological correlates, course, and treatment response of these classes would be of clinical interest.

Background. In many biomedical disorders, early age at onset (AAO) is an index of high liability to illness which is manifest by an increased risk of illness in relatives. Most but not all prior studies report such a pattern for major depression (MD).

Method. Lifetime MD and AAO were assessed at personal interview using modified DSM-III-R criteria in 13864 twin pairs, including 4229 onsets of MD, from the Swedish National Twin Registry. Analyses were conducted using Cox proportional hazards models.

Results. Controlling for year of birth, gender, zygosity, co-twin history of MD and the interaction of zygosity and co-twin history, the best-fit model showed a significant main effect and a quadratic effect of AAO of MD in the co-twin on the log hazard ratio for MD in the index twin. When examined together, these effects predicted that from the ages of 15 to ~35 years, AAO of MD is moderately negatively related to risk of illness in relatives. However, past age 35, the function flattens out, with little change of risk in relatives with further increases of AAO. Even when the co-twin had a late AAO, the risk in the index twin substantially exceeded that seen when the co-twin had no history of MD.

Conclusion. In this large sample, AAO is a meaningful, albeit modest, index of familial liability to MD. The relationship is nonlinear and results largely from an increased liability in individuals with an early AAO. These results should be interpreted in the context of the limitations of long-term recall.

Background. Subjective well-being (SWB) can be partitioned into the components life satisfaction and affect. Research on factors influencing these components of well-being has mainly focused on environmental characteristics. The aim of this study was to investigate the relative contribution of genes and environment to individual differences in life satisfaction in a large sample of Dutch twins and their singleton siblings.

Method. Life satisfaction of 5668 subjects registered with The Netherlands Twin Registry (NTR) was measured with a Dutch version of the self-reported Satisfaction with Life Scale. An extended twin design was used to obtain correlations in life satisfaction scores for monozygotic twins, dizygotic twins and sibling pairs and to estimate the contribution of genes and environment to the variation in life satisfaction.

Results. No differences between males and females were found in the mean level of life satisfaction. Broad-sense heritability was 38%. Non-additive genetic factors explained all or most of the genetic influences. The remaining 62% of the variance in life satisfaction could be attributed to unique environmental factors, both persistent and transitory, plus measurement error.

Conclusions. Individual differences in life satisfaction are determined in part by genetic factors that are largely or entirely non-additive in nature.

Background. We test the hypothesis that the neuregulin 1 (NRG1) gene at chromosome 8p22-p12, which has been implicated as a susceptibility gene to schizophrenia, is associated with variations in schizotypal personality in non-clinical populations.

Method. A randomly selected sample of 905 adolescents were assessed for their personality features using the Perceptual Aberration Scale (PAS) and the Schizotypal Personality Questionnaire (SPQ) and genotyped for three single nucleotide polymorphisms (SNP8NRG221533, rs3924999, and rs2954041) at the NRG1 gene. Relations between the three genetic variants and continuous schizotypal personality scores were evaluated using ANOVA for single-locus analyses and haplotype trend regression test for multi-locus analyses.

Results. Single locus analysis showed that the A allele of rs3924999, a functional polymorphism in exon 2, had the largest effect size and exhibited a prominent allele–dose trend effect for the PAS score. Haplotype analyses using the haplotype trend regression test indicated that the A allele of rs3924999 was mainly responsible for the association with the PAS but not with the SPQ or its three factors, and the magnitude of significance was not strengthened by the combination of this allele with adjacent locus.

Conclusions. Our study provides the first evidence for the association of NRG1 with schizotypal personality and indicates a possible role of NRG1 in the genetic etiology of schizophrenia through perceptual aberrations.

Background. Individual genome-wide linkage scans and meta-analyses support that one or more susceptibility genes for schizophrenia are located in chromosome 8p. A gene from this region, neuregulin 1 (NRG1), known to be involved with glutamatergic function, has been found to be associated in some studied samples.

Method. We have examined a new combined schizophrenia sample with 136 schizophrenia families largely of European ancestry (EA) and 646 subjects with DNA. We genotyped 14 single nucleotide polymorphisms (SNPs) in NRG1 including those reported to comprise schizophrenia-associated haplotypes in Icelandic, Scottish, Irish, and Chinese Han populations.

Results. We found no evidence of association at a single-marker or a haplotypic level. We review methodological aspects of previous studies to enable us to put our findings into context.

Conclusions. Our failure to find an association between NRG1 and schizophrenia might reflect different linkage disequilibrium (LD) patterns found in different populations, disease allelic heterogeneity, clinical heterogeneity of schizophrenia, or inadequate statistical power deriving from moderate sample size. NRG1, if a true gene for schizophrenia, accounts for a small fraction of the disease in most populations. The confirmation of NRG1 as a schizophrenia susceptibility gene will require studies with a comprehensive set of markers and in larger samples. The possibility remains that reports of NRG1 association might reflect false positives.

Background. This and the companion paper present two sequential family studies of obsessive–compulsive disorder (OCD) conducted by the same research group, but with different sampling and best-estimate procedures. In addition to providing further data on familial transmission of OCD, we used comparison of disparate findings (moderate, specific familial aggregation in this first study versus a stronger effect for other anxiety disorders than for OCD alone in the second) to examine possible effects of proband characteristics and informant data on outcome.

Method. In this initial study we interviewed 179 first-degree relatives of 72 OCD probands and 112 relatives of 32 never mentally ill (NMI) controls. Informant data were obtained on an additional 126 relatives (total ‘combined’ samples of 263 and 154 respectively). Analyses used best-estimate diagnoses made by consensus of two ‘blinded’ senior clinicians who reviewed all diagnostic materials including proband informant data about relatives.

Results. Significantly higher risk for OCD but not other anxiety disorders was found in relatives of OCD probands compared to relatives of controls in both the directly interviewed and combined samples. There was no relationship between proband age at onset of OCD and strength of familial aggregation.

Conclusions. These data indicate moderate familial aggregation of OCD, but do not support increased transmission by early onset probands, or a familial relationship between OCD and other anxiety disorders with the possible exception of generalized anxiety disorder.

Background. Overall findings of our first direct interview family study of obsessive–compulsive disorder (OCD) indicated that OCD is familial. In this replication study, we carefully examined the role of informant data in ascertaining OCD in relatives.

Method. We interviewed 112 relatives of 57 OCD patients and 115 relatives of 41 not ill controls predominantly by telephone. Additional analyses included a combined sample of relatives about whom any diagnostic information was available (228 OCD and 239 controls). To examine the contribution of proband information about relatives, we considered two sets of best-estimate diagnoses. First, we ascertained best-estimate diagnoses for relatives using information from direct interviews and from all informants except the proband. Then, we re-diagnosed relatives based on all available information, including reports from the proband about their relatives.

Results. When relative diagnoses were derived without the benefit of proband informant reports, no evidence of familial OCD transmission was found. When diagnoses were made including information from the proband about the relative, evidence of familial OCD was found, but only when the diagnostic threshold was lowered to include cases with probable OCD or OCD symptoms. Other diagnoses (generalized anxiety disorder, social phobia, drug use disorder) were also higher among OCD relatives.

Conclusions. This second study provides less robust support for familial transmission of OCD. Evidence for familial transmission of OCD was found only when diagnoses were made using information from the affected proband about their relatives. Taken in context of past findings, our own inconsistent results suggest that OCD may be heterogeneous with regard to familial transmission. Also, more careful attention should be paid to the contribution of informant reports, especially from relatives affected by the same disorder.

Background. There are significant barriers to accessing effective psychological therapy in primary care resulting from a lack of suitably trained therapists to meet current demand. More efficient service delivery using minimal interventions (such as bibliotherapy) provided by paraprofessional therapists may be one method of overcoming these problems, and is the subject of attention in the UK and elsewhere. A randomized trial was conducted to test the clinical effectiveness of this model. Assistant psychologists delivered a guided self-help intervention to patients with anxiety and depression who were currently waiting for psychological therapy.

Method. A total of 114 patients were randomized either to guided self-help or a waiting-list control group. All patients were followed up 3 months later, prior to starting conventional psychological therapy. Measures included self-reported adherence to the intervention, anxiety and depressive symptoms, social functioning and patient satisfaction.

Results. Adherence to the guided self-help intervention was acceptable and patients reported satisfaction with the intervention. However, there were no statistically significant differences between groups in anxiety and depression symptoms at 3 months.

Conclusions. The results demonstrate that this model of guided self-help did not provide additional benefit to patients on a waiting list for psychological therapy. The results are considered in the context of possible internal and external validity threats, and compared with previous trials of minimal interventions. The implications of the results for the design of future minimal interventions are considered.

Background.Randomized clinical trials indicate a benefit from combining medications with cognitive behavioral therapy (CBT) relative to medication alone for panic disorder. Using an as-treated analysis, we evaluated whether the addition of CBT enhanced outcomes for panic disorder relative to medications alone in the primary-care setting.

Method.Primary-care patients with panic disorder reported on their receipt of CBT and medications over the 3 months following baseline assessment. The degree to which outcomes for those who used anti-panic medications were enhanced by the receipt of at least one component of CBT was analyzed using a propensity score model that took into account observable baseline patient characteristics influencing both treatment selection and outcomes.

Results.The addition of CBT resulted in statistically and clinically significant improvements at 3 months on anxiety sensitivity, social avoidance, and disability. Also, patients receiving CBT in the first 3 months of the study were more improved at 12 months than patients who took medications only during the first 3 months of the study.

Conclusions.The clinical utility of the findings are discussed in terms of the importance of primary-care physicians encouraging their panic disorder patients to receive CBT as well as medications.

Background. Subjective quality of life has gained a crucial role as a global measure of outcome in mental health care. This study aimed to investigate the impact of meeting needs for care, as assessed by both patients and mental health professionals, to improve the subjective quality of life in a sample of patients receiving community-based psychiatric care.

Method. The study was conducted using a 4-year prospective longitudinal design. A cohort of patients from the South-Verona Community-based Mental Health Service (CMHS) was assessed at baseline and follow-up using, among other social and clinical measures, the Camberwell Assessment of Need (both staff and patient versions) and the Lancashire Quality of Life Profile. Predictors of changes of subjective quality of life were explored using block-stratified multiple regression procedures.

Results. Improvement in patients' clinical conditions as well as the reduction in patient-rated unmet needs in the social domain predicted an increase in subjective quality of life over 4 years; changes in staff-rated needs did not show any association with changes in subjective quality of life.

Conclusions. Meeting self-perceived social needs, beyond symptoms reduction, seems to be of particular importance for ensuring a better quality of life for people with mental disorders. If the main goal of mental health care is to improve the quality of life of users, a policy of actively addressing patient-rated needs should be implemented.

Background. There are no modern cross-cultural comparative studies of the frequency and clinical presentation of catatonia in a Western country and India using standardized rating instruments and diagnostic criteria.

Method. A total of 104 consecutively admitted patients in Wales and in India were screened for catatonic features using the same standardized rating instrument by the same psychiatrist to generate DSM-IV and other diagnostic criteria for catatonia, and a profile of signs in catatonia. Inter-rater reliability for the ratings made by the research psychiatrist was established with local psychiatrists at each unit.

Results. The frequency of DSM-IV criteria catatonia was 13·5% in India versus 9·6% in Wales (N.S.). The severity of catatonia did not differ between the two units. However, retarded catatonia was more common in India (12·5%) versus Wales (p<0·05) whereas the frequency of excited catatonia was equally common in both units. Catatonia was found in many different mental disorders not just schizophrenia and affective disorder.

Conclusions. Catatonia is commonly found among psychiatric in-patients with a similar frequency and severity but differing clinical presentations in Wales and India. Some classic signs of catatonia like posturing, catalepsy, staring and stupor were more frequent among psychiatric admissions in India than Wales. The differing clinical presentations may be due to differences in demographic features rather than cultural or aetiological factors.