Prediction of lithium response using clinical data.

OBJECTIVE: Promptly establishing maintenance therapy could reduce morbidity and mortality in patients with bipolar disorder. Using a machine learning approach, we sought to evaluate whether lithium responsiveness (LR) is predictable using clinical markers. METHOD: Our data are the largest existing sample of direct interview-based clinical data from lithium-treated patients (n = 1266, 34.7% responders), collected across seven sites, internationally. We trained a random forest model to classify LR-as defined by the previously validated Alda scale-against 180 clinical predictors. RESULTS: Under appropriate cross-validation procedures, LR was predictable in the pooled sample with an area under the receiver operating characteristic curve of 0.80 (95% CI 0.78-0.82) and a Cohen kappa of 0.46 (0.4-0.51). The model demonstrated a particularly low false-positive rate (specificity 0.91 [0.88-0.92]). Features related to clinical course and the absence of rapid cycling appeared consistently informative. CONCLUSION: Clinical data can inform out-of-sample LR prediction to a potentially clinically relevant degree. Despite the relevance of clinical course and the absence of rapid cycling, there was substantial between-site heterogeneity with respect to feature importance. Future work must focus on improving classification of true positives, better characterizing between- and within-site heterogeneity, and further testing such models on new external datasets.

Mapping susceptibility genes for bipolar disorder: a pharmacogenetic approach based on excellent response to lithium.

Genetic mapping studies in bipolar disorder (BD) have been hampered by the unclear boundaries of the phenotypic spectrum, and possibly, by the complexity of the underlying genetic mechanisms, and heterogeneity. Among the suggested approaches to circumvent these problems, a pharmacogenetic strategy has been increasingly proposed. Several studies have indicated that patients with BD who respond well to lithium prophylaxis constitute a biologically distinct subgroup. In this study we have conducted a complete genome scan using 378 markers spaced at an average distance of 10 cM in 31 families ascertained through excellent lithium responders. Response to lithium was evaluated prospectively with an average follow-up of 12 years. Evidence for linkage was found with a locus on chromosome 15q14 (ACTC, lod score = 3.46, locus-specific P-value = 0.000014) and suggestive results were observed for another marker on chromosome 7q11.2 (D7S1816, lod score = 2.68, locus-specific P-value = 0.00011). Other interesting findings were obtained with markers on chromosomes 6 and 22, namely D6S1050 (lod score = 2.0, locus-specific P-value = 0.00004) and D22S420 (lod score = 1.91). Nonparametric linkage analysis provided additional support for the role of these loci. Further analyses of these results suggested that the locus on chromosome 15q14 may be implicated in the etiology of BD, whereas the 7q11.2 locus may be relevant for lithium response. In conclusion, our results provide original evidence suggesting that loci on 15q14 and 7q11.2 may be implicated in the pathogenesis of BD responsive to lithium.

Is lithium response related to G(s)alpha levels in transformed lymphoblasts from subjects with bipolar disorder?

BACKGROUND: A series of studies from independent laboratories have found increased levels of G(s)alpha in bipolar disorder in postmortem brain and peripheral blood cells. Long-term lithium administration blunts G-protein coupled cAMP signaling and may regulate G(s)alpha levels. METHODS: We measured G(s)alpha in transformed lymphoblasts obtained from subjects with bipolar disorder and compared the findings with 23 age- and sex-matched controls. To reduce patient heterogeneity, we included only patients with an excellent response to lithium prophylaxis. RESULTS: We found no differences in G(s)alpha protein levels measured with immunoblotting. G(s)alpha levels did not correlate with age, age of onset or duration of lithium therapy. LIMITATIONS: This study made use of transformed lymphoblasts, which may not fully represent changes that occur in regionalized brain tissue. Furthermore, the transformed lymphoblasts used in this study were acquired from a select group of bipolar disorder subjects that responded to lithium treatment. Lastly, consideration has to be given to the small sample size of the study. CONCLUSIONS: These results are consistent with recent observations suggesting that mood state and treatment effects may account at least in part for increased G(s)alpha levels in bipolar disorder. CLINICAL RELEVANCE: This study suggests a need to further characterize biological phenotypes in subjects with mood disorders to enhance genetic studies.

Anticipation in bipolar affective disorder: is age at onset a valid criterion?

Anticipation has been suggested among the genetic mechanisms of bipolar disorder (BD), prompting the search for unstable DNA sequences. Past studies of anticipation in BD have generally relied on observed shift in the age at onset between parental and offspring generations. Such a shift, however, may be caused by a number of other factors difficult to correct for. We investigated age at onset distributions in a sample of 161 related subjects and in a sample of "pseudofamilies" consisting of 320 unrelated subjects selected from a large epidemiological cohort using Monte-Carlo simulation to mimic the family sample. Comparison of age at onset distributions in both samples shows a difference between the generations, but of a similar magnitude in each sample. This suggests that age at onset alone may not be a sufficient criterion of anticipation. Am. J. Med. Genet. (Neuropsychiatr. Genet.) 96:804-807, 2000.

Association and linkage studies of candidate genes involved in GABAergic neurotransmission in lithium-responsive bipolar disorder.

OBJECTIVE: To test for genetic linkage and association with GABAergic candidate genes in lithium-responsive bipolar disorder. DESIGN: Polymorphisms located in genes that code for GABRA3, GABRA5 and GABRB3 subunits of the GABAA receptor were investigated using association and linkage strategies. PARTICIPANTS: A total of 138 patients with bipolar 1 disorder with a clear response to lithium prophylaxis, selected from specialized lithium clinics in Canada and Europe that are part of the International Group for the Study of Lithium-Treated Patients, and 108 psychiatrically healthy controls. Families of 24 probands were suitable for linkage analysis. OUTCOME MEASURES: The association between the candidate genes and patients with bipolar disorder versus that of controls and genetic linkage within families. RESULTS: There was no significant association or linkage found between lithium-responsive bipolar disorder and the GABAergic candidate genes investigated. CONCLUSIONS: This study does not support a major role for the GABAergic candidate genes tested in lithium-responsive bipolar disorder.

Association and linkage studies of CRH and PENK genes in bipolar disorder: a collaborative IGSLI study.

Corticotropin-releasing hormone (CRH) and proenkephalin (PENK) are hypothalamic peptides involved in the stress response and hypothalamic-pituitary axis regulation. Previous research has implicated these peptides in the pathogenesis of affective disorders. In this study we investigated two polymorphisms located in the genes that code for CRH and PENK by means of association and linkage analyses. A total of 138 bipolar patients and 108 controls were included in the association study. In addition, 24 families were available for linkage analysis, including six families of probands with documented periodic positivity of dexamethasone suppression tests (DST) during remission. We found no association of bipolar disorder with either gene. Similarly, we did not find any evidence of linkage (P = 0.56 for CRH and 0.52 for PENK) in the entire sample or in the subsample of families of DST positive probands. In conclusion, our study does not support the hypothesis that genes coding for CRH or PENK contribute to the genetic susceptibility to bipolar disorder. Am. J. Med. Genet. (Neuropsychiatr. Genet.) 96:178-181, 2000.

Polyglutamine coding genes in bipolar disorder: lack of association with selected candidate loci.

BACKGROUND: Several studies have suggested that expanded trinucleotide repeats, particularly CAG, may have a role in the etiology of BD. Results obtained with the repeat expansion detection technique (RED) have indicated that bipolar patients have an excess of expanded CAG repeats. However, it is not clear which loci account for this difference. METHODS: Using lithium-responsive bipolar patients in order to reduce heterogeneity, we investigated five loci that are expressed in the brain and contain translated CAG repeats. A sample of 138 cases and 108 controls was studied. Genotypes were coded quantitatively or qualitatively and repeat distributions were compared. RESULTS: No difference was found in allele distribution between cases and controls for any of the loci studied. In one locus - L10378 - patients had a tendency to present shorter alleles (28.1 versus 27.9 repeats; t=2.55, df=205, P=0.011), however, this difference disappeared after correction for multiple testing. LIMITATIONS: The study has limitations common to most candidate gene association studies, that is, limited number of loci investigated and limited power to detect loci that account for a small proportion of the total genetic variability. CONCLUSIONS: Our results suggest that the loci investigated have no major role in the genetic predisposition to bipolar disorder.

Patterns of DST positivity in remitted affective disorders.

BACKGROUND: While the Dexamethasone Suppression Test (DST) has been extensively used in cross-sectional observations of patients with major affective disorders, studies have tended to ignore the longitudinal application of the DST in patients stabilized on long-term prophylactic medication. METHODS: Monthly DST's were performed on 19 patients, 16 with bipolar disorder and 3 with recurrent major depression. All cases had an excellent response to lithium treatment, and family history positive for bipolar disorder. The average duration of observation was 4 years. RESULTS: All patients remained clinically stable throughout the period of observation. Eleven patients showed intermittent DST positivity ranging from 10% to 60% of tests, and 2 patients exhibited no positivity. Six patients had fewer than 10% positive DST's. Females showed significantly higher positivity than males. The frequency of positivity did not correlate with current age, age of illness onset, duration of illness, duration of lithium treatment, or season. The risk of primary affective disorders in first-degree relatives was also unrelated to the frequency of positivity. CONCLUSIONS: While the highly selected and small sample population limits generalizability, our observations suggest that clinically sufficient lithium prophylaxis does not automatically prevent intermittent HPA dysregulation. We hope that a better understanding of this phenomenon will offer new approaches to the long-term management of mood disorders.

Lithium responsive bipolar disorder, unilineality, and chromosome 18: A linkage study.

Over the last three years several studies have investigated the hypothesis of linkage between bipolar disorder and markers on chromosome 18. Although independent groups have reported positive results, it is still not clear how these should be interpreted, as linkage spans a considerably large segment of the chromosome. In this study we have investigated linkage with chromosome 18 markers in 19 families of lithium-responsive bipolar patients, as a way to select a more homogeneous population. In addition, we have investigated whether there is evidence of a parent-of-origin effect as suggested by previous studies. Eleven markers spanning the whole chromosome were typed and linkage analysis was carried out using parametric and nonparametric methods. Analysis of the whole sample provided nonsignificant linkage results. However, when the sample included only unilineal families, and was further stratified according to parental origin, two chromosomal regions provided modestly positive lod scores. Maximum lod scores of 1.04 (P = 0.001) at D18S53 and 0.87 (P = 0.045) at D18S61 were observed for maternal and paternal pedigrees, respectively. Nonparametric analysis yielded similar results. In conclusion, our results are congruent with previous reports that suggest an advantage of unilineal pedigrees in linkage analysis of bipolar disorder and cannot rule out a parent-of-origin effect in this genomic region.

MAOA: association and linkage studies with lithium responsive bipolar disorder.

A number of association studies have investigated the role of the monoamine oxidase A (MAOA) gene in the susceptibility to bipolar disorder. Although some studies have reported positive findings, there remains some controversy, because results from different studies have not been consistent. A common explanation for inconsistencies between studies is genetic heterogeneity. We have focused on lithium responsive bipolar disorder as a way to reduce heterogeneity. In this study, we investigated the role of MAOA in lithium responsive bipolar patients using association and linkage study designs. The investigation used 138 patients and 108 normal controls. In addition, 25 families were also studied. Our results were not supportive of a major role of MAOA in the predisposition to bipolar disorder.

Evidence for a role of phospholipase C-gamma1 in the pathogenesis of bipolar disorder.

Several studies have indicated that patients with bipolar disorder (BD) who respond well to lithium prophylaxis constitute a biologically distinct subgroup. Lithium is thought to stabilize mood by acting at the phosphoinositide cycle. We have investigated a polymorphism located in the gene (PLCG1) that codes for a gamma-1 isozyme of phospholipase (PLC), an enzyme that plays an important role in the phosphoinositide second messenger system. A population-based association study and a family-based linkage study were carried out on patients who were considered excellent responders to lithium prophylaxis. Response to lithium was evaluated prospectively with an average follow-up of 14.4 +/- 6.8 years. The PLCG1 polymorphism was investigated in 136 excellent lithium responders and 163 controls. In addition, the segregation of this marker was studied in 32 families ascertained through lithium-responsive bipolar probands. The allele distributions between lithium-responsive bipolar patients and controls were different, with a higher frequency of one of the PLCG1 polymorphisms in patients (chi2 = 8.09; empirical P = 0.033). This polymorphism, however, confers only a small risk (OR = 1.88, CI 1.19-3.00). Linkage studies with the same marker yielded modest support for the involvement of this gene in the pathogenesis of BD when unilineal families were considered (Max LOD = 1.45; empirical P = 0.004), but not in the whole sample. Our results provide preliminary evidence that a PLC isozyme may confer susceptibility to bipolar disorder, probably accounting for a fraction of the total genetic variance. Whether this polymorphism is implicated in the pathogenesis of BD or in the mechanism of lithium response remains to be determined.

Evidence supporting the independent inheritance of primary affective disorders and primary alcoholism in the families of bipolar patients.

BACKGROUND: This study explored the nature of the association between bipolar disorder and alcoholism. METHODS: The authors studied 814 first-degree relatives of 121 bipolar patients, divided on the basis of response to lithium prophylaxis. Logistic regression analysis was used to analyze the contribution of demographic, familial and clinical variables to the risk of primary alcoholism in the relatives. RESULTS: The risk of primary alcoholism in relatives was not related to the degree of affective loading in the family or to the proband's lithium response. CONCLUSION: This study does not support a shared genetic liability between bipolar disorder and alcoholism. LIMITATIONS: This study lacked a control group, but the analysis accounted for this. CLINICAL RELEVANCE: These disorders are not alternative forms of the same illness.

MN blood groups and bipolar disorder: evidence of genotypic association and Hardy-Weinberg disequilibrium.

BACKGROUND: MN blood groups have been studied in the past as a genetic marker of biopolar disorder (BD). Several previous studies reported an association of the illness with lower frequency of blood group NN. METHODS: We analyzed distribution of MN blood groups in a sample of 174 patients with BD, 176 with unipolar depression, 98 with schizophrenia, and 331 healthy controls. In addition, we tested whether the inferred genotypes. conform to Hardy-Weinberg equilibrium (HWE). RESULTS: The frequency of NN phenotype was significantly lower among the bipolar patients than in any of the other three groups (p < .001). The genotype frequencies in the BD group deviated significantly from those expected under HWE (p < .01). CONCLUSIONS: These results suggest a possible locus on chromosome 4 (4q28-q31.1) associated with genetic susceptibility to bipolar illness.

Autosomal recessive inheritance of affective disorders in families of responders to lithium prophylaxis?

In this paper we report the results of a study of the mode of inheritance in affective disorders responsive to lithium. Earlier we described a series of 71 families in which the genetic transmission was compatible with a single-gene model. We have now carried out an independent replication study on 25 newly recruited families in a different geographical location. The autosomal recessive model from our original study could not be rejected with the new data. In a subsequent analysis of the pooled sample of 96 families, a recessive model with a common predisposing allele (q = 0.16) and sex-specific penetrance (0.35 in males, 0.66 in females) fitted the data best. On the other hand, X-chromosome and polygenic models could be rejected. The finding of a major-gene effect represents a specific hypothesis that can be tested by molecular genetic techniques.

Mortality of patients who dropped out from regular lithium prophylaxis: a collaborative study by the International Group for the Study of Lithium-treated patients (IGSLI).

Excess mortality has repeatedly been found in patients with recurrent affective disorders. In previous studies our group has shown that during long-term lithium treatment the mortality of such patients is not significantly higher than that of the general population. In the present study, we extended our investigation to 273 patients from the earlier IGSLI cohort who subsequently dropped out from regular lithium prophylaxis. The standardized mortality ratio (SMR) for the whole group was 2.5, significantly higher (P < 0.01) than 1.0, which is the SMR of the general population. Furthermore, the SMR of the patients from each of the participating countries, namely Denmark, Germany and Austria, was significantly higher than 1.0. These findings strengthen the evidence accumulated in previous investigations that regular long-term lithium treatment does in fact markedly reduce the excess mortality of patients with recurrent affective disorders.

Lithium-responsive affective disorders: no association with the tyrosine hydroxylase gene.

Family, adoption, and twin studies have demonstrated the involvement of genetic factors in the etiology of major affective disorders. In an attempt to identify the involved genes, several linkage and association studies have focused on the gene coding for tyrosine hydroxylase, the rate-limiting enzyme in catecholamine synthesis. The discrepant results to date could be explained by etiological heterogeneity, which may be substantially reduced by selecting patients according to lithium response. Therefore, we investigated 54 patients who had shown definite long-term response to lithium monotherapy in spite of a high risk of recurrence as indicated by the previous clinical course. All the subjects suffered from major affective disorder by Research Diagnostic Criteria (48 bipolar, 6 recurrent unipolar). They were compared to 94 population controls of similar ethnic background to test for association with a penta-allelic microsatellite marker found within the tyrosine hydroxylase gene. No significant differences in allele and genotype frequencies were observed between the two groups, providing further evidence against a major role for the tyrosine hydroxylase gene in the etiology of major affective disorders.

How to interpret findings on mortality of long-term lithium treated manic-depressive patients?! Critique of different methodological approaches.

Many studies have shown that patients with affective disorders have a mortality markedly higher than that of the general population. Studies of manic-depressive patients given prophylactic lithium treatment have yielded varying results. Some authors have reported mortalities which were more than four times that of the general population. Others have found mortalities which did not differ significantly from that of the general population. In order to examine these discrepancies we re-analysed the data of a previous study by IGSLI, using three different methods to calculate the Standardised Mortality Ratio (SMR). The data base was enlarged by adding mortality data from two additional centres. The results indicate that the discrepancies may result from the common use of a 'cumulative' approach which produces a distortion of the data if the treatment duration is not taken into consideration properly. An analysis which eliminates this artefact and takes the treatment duration into account ('year-by-year' approach) provides the strongest evidence that the mortality of lithium treated patients is not significantly different from that of the general population.

Excess cardiovascular and suicide mortality of affective disorders may be reduced by lithium prophylaxis.

The mortality of patients suffering from affective disorders is much higher than that of the general population; this excess is due to both suicides and cardiovascular disease. During long-term lithium treatment, the overall mortality has not been found to differ significantly from that of the general population but the question remains whether this lowering, if it is in fact caused by lithium, is due to a reduction in suicide frequency or cardiovascular mortality, or both. We analysed data from 827 previously studied patients and used a procedure that estimated both overall mortality and cause-specific mortalities by single-case analysis. For overall mortality, the ratio of observed deaths (among the patients) to expected deaths (in the general population) was 1.14, which is not significantly different from 1.0; this was also found in our previous analysis. In the whole patient group, comprising 5600 patient years under lithium treatment, seven suicides were observed and 1.3 expected, resulting in a standard mortality ratio of 5.22; this is significantly > 1.0, but markedly lower than that found in patients with affective disorders not given lithium. Cardiovascular mortality was not found to be higher in our patients than in the general population. In view of the fact that a placebo-controlled mortality study under long-term conditions is neither ethically nor practically feasible, our findings cannot prove definitively that long-term lithium treatment counteracts factors responsible for the excess suicide and cardiovascular mortality of affective disorders. However, our observations are compatible with such a notion.

Mortality during initial and during later lithium treatment. A collaborative study by the International Group for the Study of Lithium-treated Patients.

We have previously shown that the mortality of patients with recurrent affective disorders in long-term lithium treatment is not higher than that of the general population. In the present study on 471 patients from Denmark and Germany, we examined mortality during the initial year of lithium treatment and during later lithium treatment. During initial lithium treatment, the total mortality was twice as high as in the general population (difference not significant) and the mortality due to suicide 16 times higher. During later lithium treatment, the mortality rates did not differ from those in the general population. Our results indicate that patients with frequent, often severe recurrences, those chosen for prophylactic lithium treatment, are at risk of high mortality, which then diminishes as the prophylactic action of the treatment takes effect.