Pharmacogenetic studies in depression - Nature

The Pharmacogenomics Journal (2008) 8, 90–100 & 2008 Nature Publishing Group All rights reserved 1470-269X/08 $30.00 www.nature.com/tpj

REVIEW

Pharmacogenetic studies in depression: a proposal for methodologic guidelines A Serretti1, M Kato1,2 and JL Kennedy3 1 Institute of Psychiatry, University of Bologna, Bologna, Italy; 2Department of Neuropsychiatry, Kansai Medical University, Osaka, Japan and 3 Neurogenetics Section, Centre for Addiction and Mental Health (CAMH), University of Toronto, Toronto, Ontario, Canada

Correspondence: Dr A Serretti, Department of Psychiatry, University of Bologna, Viale Carlo Pepoli 5, Bologna 40123, Italy. E-mail: [email protected]

Pharmacogenetic studies in mood disorders are rapidly proliferating after the initial reports linking gene variants to treatment outcomes. However, a considerable range of methodologies has been used, making it difficult to compare results across studies and limiting the representativeness of findings. Specification of sampling source (inpatients vs outpatients, primary vs tertiary settings), standardization of diagnostic systems and treatments, adequate monitoring of compliance through plasma levels, sufficient length of observation (at least 6 weeks for acute antidepressant treatments, though 3–6 months are preferable), the use of a range of response criteria and the inclusion of possible environmental confounding variables (life events, social support, temperament) are all potentially important issues when planning pharmacogenetic studies. We reviewed the state-of-the-art methodology and suggested possible guideline for future studies. The Pharmacogenomics Journal (2008) 8, 90–100; doi:10.1038/sj.tpj.6500477; published online 7 August 2007 Keywords: pharmacogenetics; depression; genetics; methodology; antidepressants

State-of-the-art and unmet needs

Received 1 January 2007; revised 20 June 2007; accepted 9 July 2007; published online 7 August 2007

In the last few years, a large amount of effort has been directed to the search of genetic predictors of drug efficacy in mood disorders. In this emergent discipline a number of papers have reported positive associations between gene variants and response to antidepressants; at present the whole impact of those results is however quite difficult to understand because of the heterogeneity of samples and methods. The present paper is an attempt to describe potential biases and suggest a more homogeneous study strategy based on available evidence. Earlier papers on pharmacogenetics usually originated from clinical trials, in which subsequent genotype analyses were performed; however, pharmacologic studies have different purposes from genetic ones, and that sample selected for those specific purposes should be recruited. At present, researchers in the field of pharmacogenetics of mood disorders do not have a reference framework when planning their studies. Genetic studies are much more prone to possible stratification biases, the comparison between two genotypes may lead to spurious results when genotype samples are not balanced for relevant features, this is particularly important given the much smaller effect size of gene variants compared to psychopharmacology studies.1 Further, the investigation of gene variants influencing response is prone to biases due to possible circular interactions between genes and clinical conditions. Gene variants may influence not only one but multiple conditions and these conditions are intertwined with each other and with response. As an

Pharmacogenetics in depression A Serretti et al

91

example, the variation in the promoter region of the 5-HTT gene (5-HTTLPR) has been associated with poor response2 and it also modulates depressogenic and anxiety responses to life stressors, which are both associated with poor antidepressant outcome.3 A large number of reviews of pharmacogenetic results on mood disorders have been previously published;4,5 the aim of present paper is to describe some of the most problematic issues to deal with; following, we compared the suggested methods with the state-of-the-art methodology in a pool of recent pharmacogenetic studies on 5-HTTLPR. Diagnostic and clinical issues Depression is a phenotype complex to define: each single episode of illness is characterized by parameters such as severity and duration, while the whole disorder presents a largely variable number of episodes, with a range of severities and different remission patterns. It is basically a continuum more than a discrete entity.6 Drug response could be a simpler trait compared to depression, but still, it is difficult to define its boundaries: apart from the possibility of spontaneous remission, partial, complete manic switch responses are possible. Moreover drug response is not completely stable across episodes. The possible relationship between such phenotypes (depression and drug response) and a genotype is evidently not straightforward; for this reason the study of drug response in depression must consider such aspects. Clinical features of depression In psychiatry, the clinical approach typical for other medical fields, such as to narrow the ‘affected’ phenotype with an objective measure associated with the disorder (that is, blood pressure or glycemia), cannot be applied: there is no pathognomonical sign or symptom of disease and diagnosis is made in accord to presence of a number of symptomatologic criteria. A structured diagnostic interview plus clinical information (from siblings, caregivers, prescriptions, and so on) is considered the ‘gold standard’,7 consequently the complexity of the diagnostic effort seems to be overcome: a subject affected by a major depressive episode has been clearly identified and included in distinction from episodes with different etiopathological origins. However, the clinical experience is much more multifaceted and the most conservative option would be to exclude all the cases that do not satisfy completely and unanimously the criteria, but this could determine the exclusion of a large number of subjects; in a similar way the over-inclusion of any depression probably will never lead to strong genetic results, given that many other factors intervene in the outcome of broadly considered depression.8 Moreover, a longitudinal observation has revealed that about 50% of young people hospitalized for a major depressive episode will develop manic or hypomanic episodes in the following 10–15 years,9 therefore some

actually depressive future bipolar subject could be involuntarily included in pharmacogenetic study. It has been also reported that the stability of diagnosis seems to improve as severity increases10 and that bipolar disorder has probably a stronger genetic load compared to ‘major depressive single episode’;11 from this point of view only subjects with severe episodes, long history of illness, higher genetic predisposition or bipolar diagnosis should be included, but we would loose the representativeness of the studied sample compared to the subject population.12 At the moment, we have no a priori basis to determine if the mechanism of antidepressant response could be the same in delusional and non-delusional depression, in mild or severe, in typical or atypical depression and conversely no evidence is actually supporting the hypothesis that a drug could act differently in various situations. Nevertheless, there are some variables that have been repeatedly associated to a worse probability of treatment response, and that have always to be considered in pharmacogenetic studies, like the symptomatology severity.13–15 Also the mean duration of index episode should be considered, because a proportion of patients might already be on the road to recovery at the beginning of the study, following the natural course of an untreated depression. At the opposite side, chronic depression (with episode length 42 years) is thought to respond poorly to treatments.16,17 Other factors may influence response such as baseline Hamilton Depression Rating Scale (HAM-D), duration of illness, number of episodes, h(UP/BP), presence of side effects, current medical condition, delusional features, lithium augmentation, the level of previous treatments and personality disorders could be some predictors.18,19 However, no definite data is available on this issue.20 The lack of possibility for rational guidelines on this issue, make reasonable to suggest that a careful description of clinical features of the sample, in terms of severity of symptomatology, number of episodes, presence of delusional symptomatology, duration of illness, duration of the episode and the level of previous treatment is suggested (Table 1). Clinical features of comorbid disorders: axis I Comorbidity is also a very important response modifier21 that should be controlled for. The presence of a severe comorbid medical illness, although very frequent in clinical practice,22 together with psychiatric comorbidity with substance abuse are unequivocally reported to have a detrimental effect on outcome23–25 and are usually excluded from pharmacogenetic studies. However, substance abuse is common in depressives and it could be also as a consequence of depression. Therefore, we should check for collateral information about this issue and exclude only primary substance abuse and not completely exclude secondary substance abuse, consequence of depression, but consideration for this bias as possible confounder is needed. A particular issue is the comorbidity with mild cognitive impairment in middle-aged and elderly patients, because assessment during the acute phase is difficult given the

The Pharmacogenomics Journal


92

Table 1 Summary of the suggested issues for pharmacogenetic studies in depression Diagnostic and clinical issues 1. Structured diagnostic interview plus clinical information with DSM IV 2. Clinical features of the sample (subgroup of depressive disorder and with or without BP) 3. Severity of baseline symptomatology 4. Number of episodes 5. Presence of delusional symptomatology 6. Duration of the illness 7. Duration of the episode 8. When not drug naı¨ve, the level of previous treatments 9. Exclusion of intentional recruiting bias 10. Exclusion of severe comorbid medical illness 11. Exclusion of psychiatric comorbidity with substance abuse 12. Exclusion of severe psychiatric comorbidities (severe OCD, severe eating disorder, and so on) 13. When inclusion of mild-to-minor comorbidities (including personality disorder), a detailed assessment reported Socio-demographical and psychological and social issues 14. Older people should be analyzed separately or considered as covariate 15. Sex 16. Ethnicity 17. Inclusion of psychological and social modulators Treatment issues 18. Exclusion of different mechanisms of action (SSRI, TCA, ECT) or covariate analyses 19. Assessment of compliance through plasma levels 20. Washout period of at least more than 1 week 21. Control of concomitant drugs Assessment issues 22. Use of more than one outcome definition with at least one standard assessment tool also in second or later treatment 23. Frequent assessment during the first 2 weeks 24. Assessment at least biweekly 25. Duration of the study of at least 6 weeks 26. Side effects assessment Statistical issues 27. To avoid analysis made in the same population repeatedly 28. Presence of power analysis 29. Hardy–Weinberg equilibrium 30. To present basic univariate analyses combined with more sophisticated approaches (ANOVA, GLM, RRM) that allow the inclusion of covariates Genetic issues 31. The use of relatively small haplotypes and the use of empirical P-values Abbreviations: ANOVA, analysis of variance; BP, bipolar disorder; ECT, electroconvulsive therapy; GLM, General Linear Model; RRM, Random Regression Model; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic anti-depressant.

substantial overlap with depressive symptomatology, it becomes evident only when a deteriorating course is observed.26,27


Comorbid psychiatric disorders (that is, dysthymia or obsessive–compulsive disorder (OCD)) could also reduce the overall response to treatment, or in some case could modify its timing.28–30 Moreover, from a genetic perspective, the comorbidity with anxiety disorders, in particular the presence of panic-agoraphobic spectrum symptoms has been reported to be related with a different genetic background.31 However, comorbidity is present to a large extent in depression,32 and the exclusion of all comorbid patients will eventually again reduce representativeness. Based on the abovementioned evidence, we suggest to exclude only those patients with severe psychiatric comorbidities (severe OCD, severe eating disorder, and so on) that could heavily bias results, and to include mild to minor comorbidities but with a detailed assessment that allows to control for this confounder in the statistical analysis and in replication studies. Clinical features of comorbid disorders: axis II The even more intricate issue of comorbidity with personality disorders lacks of unequivocal conclusion. It is a very common phenomenon33 and a recent meta-analysis reported that it has been considered a possible clinical negative predictor of response to treatment,34 consistently with early studies35–38 whereas others reported a similar outcome for people with and without personality disorders comorbidity using only well-designed study39 consistent with some recent studies.40–42 Apart from specific personality disorders, also temperament traits have been associated with response, generally in the direction of a poorer response in subjects with high harm avoidance.43,44 The influence of genes on personality is debated as well, though some studies reported the associations of personality disorders with genetic polymorphism, particularly DRD4 and 5-HTTLPR,45 also in depressed patients46 these remain still controversial. Overall we cannot at present time disentangle the relation between genetic background, personality, depression and response to treatment. Genes could predispose to temperamental and personality features that increase liability to depression and modulate response, but other causal relationships are possible as well, such as depression inducing personality changes or depression and personality being a common manifestation of liability genes.47,48 We should not therefore exclude patients with personality disorder but a careful assessment should be performed in order to control for this possible bias. Socio-demographical and psychological issues From a clinical point of view distribution of many sociodemographical characteristics have failed to produce findings robust enough to be clinically relevant as response predictors,49,50 nevertheless they have a mild influence depending form the sample recruitment strategy; in particular, old age is often considered an exclusion criteria in clinical trials, because older people are considered more prone to concomitant medical conditions and to cognitive


93

impairment, more sensitive to adverse effects and with lower response rate to antidepressants;51 however, in literature elderly condition has been related to a better and a worse outcome in affective disorders.52,53 In the same way sex and ethnicity has never been clearly related to a different response rate; while sexes may differ in the physiological pathway of response, ethnicity may have implications for drug metabolism and tolerance, and from a genetic perspective different population could have different allelic frequencies. Positive results deriving from one population could not be subsequently replicated in another one: as an example, pharmacogenetic results for the serotonin transporter gene promoter polymorphism (5-HTTLPR) and antidepressant response seem confirmed on western populations but are conflicting in eastern populations.54 Also younger age at onset55 and duration of illness53 have been suggested as predictors of poor response. Social support and marital status have been repeatedly associated with increased probability to recover from a depressive episode, as well as good level of functioning and higher educational level.8,23 At the opposite, other studies failed to detect any difference in response rate according to marital status and social class.50 Religious belief is usually ignored; however, it has been reported to be associated with earlier remission.56,57 In general, psychosocial issues have been scarcely kept in consideration in pharmacogenetic studies; the role of nonpharmacologic treatments and environment factors is also important as there is evidence that application of nonsomatic interventions can have a significant impact. A very basic aspect is the difference in response observed in inpatients vs outpatients,58 apart from compliance aspects, inpatients are subject to a much more intensive non-pharmacologic treatments. A much larger number of psychological features may influence outcome but they are not sufficiently validated to be suggested for routine inclusion. They include intelligence level,59 premorbid functioning,60 defense mechanism style,61 and stressful life events,62 just to cite the most important ones. Finally, circular or nonlinear influences are possible also with psychosocial issues and antidepressant outcome as reported for 5-HTTLPR and depression course.3 In conclusion, we suggest to include at least age, sex and ethnicity and additionally as much as possible sociodemographical and psychological and modulators of antidepressant response; however, a cost/benefit evaluation should be performed in each study (Table 1).

are a quite homogenous class in terms of pharmacogenetic effects,5 the inclusion of very different mechanisms of action (SSRI, tricyclic antidepressant (TCA), electro-convulsive therapy (ECT)) should be avoided in order to reduce variability.63 The low antidepressant dosages could be some variables that have been associated to a worse probability of treatment response.13 The use of concomitant antipsychotics, sometimes needed in the clinical practice for depression, could strongly modulate the genetic influence to antidepressant response and should be avoided or at least separate subjects treated with concomitant antipsychotics from one treated with only antidepressant to evaluate genetic efficacy of antidepressant or antidepressant plus antipsychotics. Use of benzoziazepines also could influence outcome in particular ones or in case of change of dose during the study. Moreover, pharmacokinetic factors, that could influence bioavailability, could also influence the response of the individual to the drug and they should be analyzed if possible; for example the assessment of compliance thorough plasma levels could help to exclude a common cause of nonresponse. However, a direct relationship between serum levels and clinical response, as well as administered dose and clinical response, has not been observed for SSRIs but this is controversial.64–68 To develop the investigation of antidepressant response in a clinical trial, a washout period from all previous treatment is necessary, but this could be a serious challenge in acutely ill patients. In fact from one side, if symptoms exacerbation occurs, this could complicate baseline setup, as well as it add ethical concerns and management issues. On the other hand, the absence of washout, means that true baseline is not achieved and antidepressant response could be over or underestimated. Obviously an inpatient situation is far easier to manage for the clinician, and for outpatients is advisable to arrange frequent short meetings to assess increasing severity to respond to buildup of suicidal intention at once. There is no simple suggestion that could help to resolve this point: in principle, prescribed drugs should be at a sufficient dosage with homogenous mechanism of action or, if they have different mechanisms of action, this should be considered in the analysis. Avoid the use of antipsychotics and unstable use of other concomitant psychotropics, evaluate compliance and pharmacokinetic factors through plasma level, and, if no suicidal intention emerges in washout period, patients may be included in the trial (Table 1). Whatever strategy is chosen, it should be clearly reported in the paper.

Treatment issues

Assessment issues

In general, in published papers, the drug administration protocols are specified well enough to permit replications; in the context of pharmacogenetics, the choice of treatment molecule is one of the variables to investigate; however, studies using all together different antidepressants, at times with antipsychotics, have been published. Although it seems that selective serotonin reuptake inhibitors (SSRIs)

In psychiatry, we do not have objective measures and the subjectivity of many symptoms creates special challenges. Although many rating scales have been developed (The Bech–Rafaelsen Melancholia Scale,69 The Inventory of Depressive Symptomatology (IDS),70 Quick Inventory of Depressive Symptomatology, Clinician Rating (QIDS-C)71) the need of a rating scale sensitive to clinically important



94

differences lead the authors to routinely use HAM-D,72 or less frequently the Montgomery Asberg Depression Rating Scale (MADRS),73 usually weekly. Inter-rater reliability is rarely reported and a description of raters experience and procedure of blindness for rating is often lacking. Subtle study design issues as the choice of whole score on HAMD17 or HAM-D21, score reduction of HAM-D, response rate defined as at least 50% improvement or remission rate defined as 7 or less in HAM-D score would lead to different conclusions.74 For example, in a sample of severe or delusional depressives, a cut-off of 50% could include in the ‘responder’ group subjects for whom HAM-D dropped from 40 to 20, still satisfying criteria for a full episode and conversely a cut-off of 7 points could lead to exclusion from remission group for a subject for whom the score dropped from 40 to 9 points. What is more, unlike virtually all other fields of medicine, drug response in psychiatry is based on behavioral observation rather than on direct objective measures and the definition of drug response is based on changes in behavioral ratings; thus an anxious subject could have a score much more high (only anxiety items 9, 10, 11, 12, 13 of HAM-D sums to 24) than a slowed, hypersomnic, hyperphagic but more severe depressive one. One suggestion could be to use more than one single analysis, along with at least one standard tool (total score of HAM-D or MADRS); for example analyzing all or two of HAM-D score reductions, response rate and remission rate, using Hamilton together with self-administered scales for depression (such as the Visual Analog Scale75), or the ‘core symptoms’ of HAM-D to exclude possible bias due to anxiety features.76 However, trying them all would compromise our type I error rate significantly, therefore, a choice should be performed when planning the study according also to the expertise of the administering team. Another main issue of assessment is associated with how to evaluate and assess subjects with rapid onset of effect within 1 week after treatment.77,78 Delayed and persistent (‘true drug’) improvement was reported to characterize the response to antidepressant medication; while early or non-persistent (‘placebo’) benefit is typical of a placebo response.79 Particularly, in inpatient setting it is possible to observe early recovery due to the environment modification, but it usually persists only a few days after discharge; therefore, subjects with a rapid onset of effect (that is, less than 1 week) should be included cautiously, as possible placebo responders77 and the possibility of a placebo response should be investigated through different strategies such as a run-on period and long-term observation. Until now there are some genetic studies of placebo like response;80,81 it could be possible that some gene are effectively predisposing to a better placebo response and this hypothesis should also be investigated. On the other hand, one may postulate that maximum improvement occurs during the first 2 weeks, with some improvement within the first 3 days82,83 and some studies reported associations between gene variants and rapid onset of effect.84,85 We may not rule out the possibility that gene


variants may be indirectly associated with rapid onset through a mechanism of action independent from drugs, for example, predisposing to shorter episodes. This would then imply that gene variants are also associated with placebo response. Side effect as well as severe adverse event are also what we should monitor and assess during the study. Up to now, few studies assessed both treatment response and side effects of antidepressants; however, it is important because the evaluation of side effects are more clear cut than that of clinical response with possible use of objective measures. Therefore, information of association between side effects and polymorphisms could be useful for clinicians immediately. Assessments tools for pharmacognetic studies in depression are not still well established but some scales were suggested in the study of treatment algorithms, they are Frequency, Intensity, and Burden of Side Effects Rating, Global Rating of Side Effect and Burden and Patient-Rated Inventory of Side Effects described in the Sequenced treatment alternatives to relieve depression (STAR*D).86 Suggestion could be to assess symptomatology frequently during the first 2 weeks (say every 3 days) and weekly or at least biweekly after 2 weeks from medication. The length of the trial in published studies is variable from 2 to 6 weeks, infrequently from 12 weeks to 3 year. A reasonable duration of the study could be of at least 6 weeks.87 An even longer observation would be preferable in inpatient studies, to control the stability of remission after hospitalization period and in populations with a slower response as elderly or delusional patients. However, the experimenter should be aware that some gene variants may influence time to response more than overall response rate, and that for longer observation more and more factors influence response. Side effects should be monitored and assessed.

Sequential treatment In recent years, the study of treatment algorithms, guideline-based procedures and collaborative care systems have been developed and they have been recognized as effective methods to optimize the delivery of treatment and to avoid treatment resistant depression.86,88–92 However, physicians have to choose the second or following treatment to the patients who was resistant to initial treatment from a number of suggested treatment and, at this time, there are no evidence to chose the best therapeutic tool for those patients even in these established treatment algorithm. Pharmacogenetic perspective could help to solve this issue and pharmacogenetic study on second or later treatment response of the patients who was resistant to first therapy also could be suggested using these treatment algorithms. STAR*D is the one of the largest and most sophisticated example that is noteworthy for its huge sample, the use of trained raters, well-organized treatment and assessment protocols; however, it also implies biases from multi-ethnicity, multi-drug use and multiple sites of recruitment.86


95

Each treatment algorithm has different assessment tools for depressive symptoms; for example, STAR*D86–QIDS-C16 (p5: remission), The German Algorithm Project (GAP)89– HAM-D (p9: remission), Texas Medication Algorithm Project (TMAP)91–IDS-C30 and genome-based therapeutic drugs for depression (GENDEP)92–HAM-D or MADRS. Depressive symptoms on pharmacogenetic studies on antidepressant response have been usually evaluated by HAM-D or MADRS and many important data were accumulated by these assessment tools. We suggest that at least one of these standard assessment tools should be used with other one if needed that might help to establish general validity of the new other assessment tools. Even if complex treatment algorithms may complicate interpretation of results, yet they should be applied for a more ‘real world’ investigation. The suggestion about this issue is to use more than one outcome definition with at least one standard assessment tool.

Statistical issues The most common approach in pharmacogenetic depression studies is the traditional analysis of variance (ANOVA), modified to evaluate time course (repeated measures ANOVA) and to include concomitant factors or covariates (multivariate analysis of variance (MANOVA) or General Linear Models (GLM)). However, it is well known that traditional parametric approaches rely on distribution assumptions that, if violated, may strongly bias results.93 The dependent variable should be normally distributed and equal variances across different genotypes should be observed, this is not always the case as variances unequally increase during treatment. A departure from normal baseline distribution should lead to non-parametric methods as well as the observation of unequal variances across genotypes, a fact commonly observed particularly in the end point of the trial when patients spread for magnitude of response. In order to avoid the loss of power of nonparametric analyses, it has been suggested to use Random Regression Models94 which overcome many of those biases. We applied such methods in some studies but the problem of replicability from other groups limits this method. Recently, survival analysis has been applied and it may overcome some biases as it is robust to variance biases; however, only a categorical outcome can be considered.86 The large amount of explanatory variables with reciprocal nonlinear interactions in pharmacogenetic studies suggested also the use of neural networks.95 However, the exploratory use of such methods should be cautioned in absence of strong underlying known associations. In any case univariate analyses should be reported together with multivariate approaches to improve readability and replicability of results. The methods of analysis should be therefore planned a priori, including subgroups analysis helping to generate hypothesis, but for which all positive results must be tested subsequently in newly

recruited patients. In the same way, analysis made in the same population repeatedly, using different phenotypes or different genes should be avoided. For these reasons, correction for multiple test and for possible confounding variables (that is, baseline level of severity among groups) should be always performed to rule out false positives. Conversely, there is a risk of missing genuine associations because of multiple test correction; for this reason, a power analysis is compulsory. The sample size has to be sufficient to detect clinically significant differences, but pharmacological studies are often not designed for pharmacogenetic analysis purposes. To analyze a gene-by-drug interaction a case control study requires a minimum of 200 subjects to detect odd ratios for a main effect gene of around 4; for an odd ratio between 2 and 4, 400 cases and 800 controls are necessary, and following these indications almost all the existing studies are underpowered.96 Therefore, the frequency of the genotype under study will determine the number of case and controls that need to be recruited to achieve sufficient statistical power to detect a difference. Particular attention must be paid to sample sizes when more variables are investigated simultaneously, such as multiple traits, interaction gene–environment or between different genes.97 In fact, the emerging issue of gene interaction is not yet satisfactory addressed by present statistical tools: GLM,98 linear99,100 and nonlinear95 models have been proposed but scarcely used. Further, it is possible that positive association results derive not only from direct effect of polymorphism under study, but also from a polymorphism in linkage disequilibrium, or from a spurious association due to ethnic differences, has to be performed, and, if possible, an analysis of multiple markers could be useful to exclude stratification bias.101 Statistical issues may greatly influence results, our suggestion is to avoid analysis made in the same population repeatedly and to present basic univariate analyses combined with more sophisticated approaches (MANOVA, GLM, Random Regression Model (RRM)) that allow the inclusion of covariates, power analysis that help the reader to evaluate the study and control of Hardy–Weinberg equilibrium of the sample (Table 1).

Genetic issues Genetic issues are not the core of the present paper and they will be just summarized.96 With about 20 000 genes, many of which expressed in the brain, and 3 000 000 of SNPs the choice of the right candidate for pharmacogenetic analysis is quite difficult. The issues to be addressed are whether to use a genome-wide, hypotheses-free approach or to follow a candidate strategy. For candidate genes, the functional significance should be investigated both in vitro and in vivo, with proteomics and metabolomics approaches; however, this may not always be feasible.102 Animal models could be of great help but this strategy has not been extensively applied yet.



96

The impact of a gene could be at pharmacokinetic or at the pharmacodynamic level; it could modify the process of absorption, distribution or metabolism of the drug, as for example the cytochrome P-450 system, or else influence response at one of the several steps from the receptor-drug binding, through the cascade of following signals; finally it could act through long-term modulations and other different epigenetic mechanisms such as methylation and posttranscriptional modifications. The exact mechanism of action of antidepressants is only partially known; the theories of antidepressant action are focused on serotonergic pathway, in particular, on the 5-HT transporter, 5-HT receptors and enzymes involved in monoamine biosynthesis and degradation. However, many other systems are possibly involved.103,104 We could estimate that the variance explained by the most confirmed gene variant for antidepressant response (5-HTTLPR) is about 2–3%. This is in accordance with the range of effect of single gene variants in complex disorders1 and it means that a large number of genes remain to be identified. Moreover, the system is much more complex that previously hypothesized and genes may interact with the environment in determining the final phenotype3 that in turn interact with the drug. This makes selection of candidate quite a difficult step. Haplotype analysis has been suggested as more informative compared to single SNPs, moreover, much larger areas of the genome may be investigated. However, this technique is not free of criticisms, haplotypes can only be inferred in association studies such as pharmacogenetic ones, therefore liability haplotype identification could be biased.105 Moreover the risk of false positive when high numbers of haplotypes are analyzed is possible. The knowledge of those

limitations, the use of relatively small haplotypes and the use of empirical P-values are suggested (Table 1). Ethical issues Ethical issues have been widely reported in separate reviews106 and consortium statements (http://www.royalsoc. ac.uk/displaypagedoc.asp?id ¼ 17570). Recent editorial policies by a number of journals require that ethical committee permission should be presented in the process of paper submission to avoid misuse of genetic information. Analysis of published papers Following the methodology dissection of the previous sections, we present here the result of the analysis of the published papers about association of antidepressant response with 5-HTTLPR, the most reported polymorphism, in the light of the suggested summary of the present paper, some issues we could not evaluate from the papers were excluded in analysis (Figure 1 and Table 2). As selection criteria, we searched Medline for all publications available up to October 2006 studying the association between antidepressant treatments and 5-HTTLPR in depressive patients with the keywords: affective, depression, mood, treatment response, SERTPR and 5-HTTLPR. Obviously all published papers are lacking in some of the aspects we just discussed, this is not intended as a criticism but it is due to the fact that our analysis is based on existing papers and contains suggestion for future ones. In the table we reported all published papers and the

Figure 1 Analysis of pharmacogenetic papers on 5-HTTLPR. The total and subcategory scores are presented. The Arabic numeral is based on the summary of issues (Table 1). I. Diagnostic and clinical issues, II. Socio-demographical and psychological and social issues, III. Treatment issues, IV. Assessment issues, V. Statistical issues, VI. Genetic issues.



97

Table 2

Analysis of the published pharmacogenetic papers including 5-HTTLPR

Author

I Total:12

II Total:1

Arias et al.107 Durham et al.85 Hong et al.108 Joyce et al.109 Kato et al.110 Kato et al.111 Kim et al.112 Kim et al.113 Kirchheiner et al.114 Kraft et al.115 Lee et al.116 Minov et al.117 Murphy et al.63 Pollock et al.84 Raush et al.81 Serretti et al.118 Smeraldi et al.119 Yoshida et al.120 Yoshida et al.121 Yu et al.122 Zanardi et al.123 Zanardi et al.124

8 8 6 5 7 6 6 6 10 7 7 7 5 8 9 10 8 5 6 6 9 10

2 3 3 3 2 2 3 3 3 2 3 0 3 3 3 3 1 2 3 2 2 2

(66.67) (66.67) (50.00) (41.67) (58.33) (50.00) (50.00) (50.00) (83.33) (58.33) (58.33) (58.33) (41.67) (66.67) (75.00) (83.33) (66.67) (41.67) (50.00) (50.00) (75.00) (83.33)

(50.00) (75.00) (75.00) (75.00) (50.00) (50.00) (75.00) (75.00) (75.00) (50.00) (75.00) (0.00) (75.00) (75.00) (75.00) (75.00) (25.00) (50.00) (75.00) (50.00) (50.00) (50.00)

III Total:4 4 2 3 3 4 4 3 4 2 3 0 0 3 3 4 3 4 4 4 3 4 4

(100.00) (50.00) (75.00) (75.00) (100.00) (100.00) (75.00) (100.00) (50.00) (75.00) (0.00) (0.00) (75.00) (75.00) (100.00) (75.00) (100.00) (100.00) (100.00) (75.00) (100.00) (100.00)

IV Total:5 2 (40.00) 3 (60.00) 1 (20.00) 2 (40.00) 4 (80.00) 4(80.00) 2 (40.00) 4 (80.00) 1 (20.00) 2 (40.00) 1 (20.00) 1 (20.00) 4 (80.00) 2 (40.00) 2 (40.00) 3 (60.00) 2 (40.00) 3 (60.00) 3 (60.00) 2 (40.00) 1 (20.00) 3 (60.00)

V Total:3 2 2 2 2 1 1 0 2 3 2 1 1 3 0 1 3 1 0 1 2 2 2

(66.67) (66.67) (66.67) (66.67) (33.33) (33.33) (0.00) (66.67) (100.00) (66.67) (33.33) (33.33) (100.00) (0.00) (33.33) (100.00) (33.33) (0.00) (33.33) (66.67) (66.67) (66.67)

VI Total:1 0 0 0 0 0 0 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (100.00) (100.00) (0.00) (100.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

The score (and percentage) for each study related to the criteria presented in Table 1. Some issues (numbers 9 and 27) were not evaluated because of lack of information from papers. I. Diagnostic and clinical issues, II. Socio-demographical and psychological and social issues, III. Treatment issues, IV. Assessment issues, V. Statistical issues, VI. Genetic issues.

correspondence with the criteria we proposed. In detail, no paper included psychological and social modulations nor assessed symptom frequently during the first 2 weeks. Only three papers used haplotypes. Few studied side effects. There are large heterogeneities of the ratio of each issue among the papers. Conclusion We described some of the most problematic issues to deal with pharmacogenetic studies in mood disorders and also summarize the suggested issues at Table 1. We hope that this article could be helpful as guideline for both clinicians and researchers of the field to detect the robust predictor of antidepressant response in treatment of depressive patient in the future. Duality of interest No conflict of interest or financial support is present for the authors. References 1 Kendler KS. ‘A gene fory’: the nature of gene action in psychiatric disorders. Am J Psychiatry 2005; 162: 1243–1252. 2 Serretti A, Masaki K, De Ronchi D, Kinoshita T. Meta-analysis of serotonin transporter gene promoter polymorphism (5-HTTLPR) association with selective serotonin reuptake inhibitor efficacy in depressed patients. Mol Psychiatry 2007; 12: 247–257.

3 Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science 2003; 301: 386–389. 4 Malhotra AK, Murphy Jr GM, Kennedy JL. Pharmacogenetics of psychotropic drug response. Am J Psychiatry 2004; 161: 780–796. 5 Serretti A, Artioli P, Quartesan R. Pharmacogenetics in the treatment of depression: pharmacodynamic studies. Pharmacogenet Genomics 2005; 15: 61–67. 6 Kendler KS, Gardner Jr CO. Boundaries of major depression: an evaluation of DSM-IV criteria. Am J Psychiatry 1998; 155: 172–177. 7 Leckman JF, Sholomskas D, Thompson WD, Belanger A, Weissman MM. Best estimate of lifetime psychiatric diagnosis: a methodological study. Arch Gen Psychiatry 1982; 39: 879–883. 8 Hirschfeld RM, Russell JM, Delgado PL, Fawcett J, Friedman RA, Harrison WM et al. Predictors of response to acute treatment of chronic and double depression with sertraline or imipramine. J Clin Psychiatry 1998; 59: 669–675. 9 Goldberg JF, Harrow M, Whiteside JE. Risk for bipolar illness in patients initially hospitalized for unipolar depression. Am J Psychiatry 2001; 158: 1265–1270. 10 Rice JP, Rochberg N, Endicott J, Lavori PW, Miller C. Stability of psychiatric diagnoses. An application to the affective disorders. Arch Gen Psychiatry 1992; 49: 824–830. 11 Tsuang M, Faraone S. The Genetics of Mood Disorders. The Johns Hopkins University Press: Baltimore, 1990. 12 Zimmerman M, Mattia JI, Posternak MA. Are subjects in pharmacological treatment trials of depression representative of patients in routine clinical practice? Am J Psychiatry 2002; 159: 469–473. 13 Meyers BS, Sirey JA, Bruce M, Hamilton M, Raue P, Friedman SJ et al. Predictors of early recovery from major depression among persons admitted to community-based clinics: an observational study. Arch Gen Psychiatry 2002; 59: 729–735. 14 Paykel ES, Ramana R, Cooper Z, Hayhurst H, Kerr J, Barocka A. Residual symptoms after partial remission: an important outcome in depression. Psychol Med 1995; 25: 1171–1180. 15 Keller MB, Lavori PW, Mueller TI, Endicott J, Coryell W, Hirschfeld RM et al. Time to recovery, chronicity, and levels of psychopathology in



98

16 17 18

19

20

21 22 23

24

25

26 27 28 29

30

31

32

33

34

35

36

37 38

39

major depression. A 5-year prospective follow-up of 431 subjects. Arch Gen Psychiatry 1992; 49: 809–816. Keller MB. Long-term treatment of recurrent and chronic depression. J Clin Psychiatry 2001; 62(Suppl 24): 3–5. Fava M. Diagnosis and definition of treatment-resistant depression. Biol Psychiatry 2003; 53: 649–659. Serretti A, Zanardi R, Mandelli L, Smeraldi E, Colombo C. A neural network model for combining clinical predictors of antidepressant response in mood disorders. J Affect Disord 2007; 98: 239–245. Serretti A, Olgiati P, Liebman MN, Hu H, Zhang Y, Zanardi R et al. Clinical prediction of antidepressant response in mood disorders: linear multivariate vs neural network models. Psychiatry Res, (In press). Winterer G, Ziller M, Linden M. Classification of observational data with artificial neural networks versus discriminant analysis in pharmacoepidemiological studies—can outcome of fluoxetine treatment be predicted? Pharmacopsychiatry 1998; 31: 225–231. Krishnan KR. Comorbidity and depression treatment. Biol Psychiatry 2003; 53: 701–706. Murrell SA, Himmelfarb S, Wright K. Prevalence of depression and its correlates in older adults. Am J Epidemiol 1983; 117: 173–185. Bagby RM, Ryder AG, Cristi C. Psychosocial and clinical predictors of response to pharmacotherapy for depression. J Psychiatry Neurosci 2002; 27: 250–257. Sullivan LE, Fiellin DA, O’Connor PG. The prevalence and impact of alcohol problems in major depression: a systematic review. Am J Med 2005; 118: 330–341. Small GW, Birkett M, Meyers BS, Koran LM, Bystritsky A, Nemeroff CB. Impact of physical illness on quality of life and antidepressant response in geriatric major depression. Fluoxetine collaborative study group. J Am Geriatr Soc 1996; 44: 1220–1225. Wells CE. Pseudodementia. Am J Psychiatry 1979; 136: 895–900. Bulbena A, Berrios GE. Pseudodementia: facts and figures. Br J Psychiatry 1986; 148: 87–94. Keller MB, Lavori PW, Endicott J, Coryell W, Klerman GL. ‘Double depression’: two-year follow-up. Am J Psychiatry 1983; 140: 689–694. Frank E, Cyranowski JM, Rucci P, Shear MK, Fagiolini A, Thase ME et al. Clinical significance of lifetime panic spectrum symptoms in the treatment of patients with bipolar I disorder. Arch Gen Psychiatry 2002; 59: 905–911. Zimmerman M, Chelminski I, Posternak MA. Generalizability of antidepressant efficacy trials: differences between depressed psychiatric outpatients who would or would not qualify for an efficacy trial. Am J Psychiatry 2005; 162: 1370–1372. Rotondo A, Mazzanti C, Dell’Osso L, Rucci P, Sullivan P, Bouanani S et al. Catechol o-methyltransferase, serotonin transporter, and tryptophan hydroxylase gene polymorphisms in bipolar disorder patients with and without comorbid panic disorder. Am J Psychiatry 2002; 159: 23–29. Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE. Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the national comorbidity survey replication. Arch Gen Psychiatry 2005; 62: 593–602. Corruble E, Ginestet D, Guelfi JD. Comorbidity of personality disorders and unipolar major depression: a review. J Affect Disord 1996; 37: 157–170. Newton-Howes G, Tyrer P, Johnson T. Personality disorder and the outcome of depression: meta-analysis of published studies. Br J Psychiatry 2006; 188: 13–20. Pilkonis PA, Frank E. Personality pathology in recurrent depression: nature, prevalence, and relationship to treatment response. Am J Psychiatry 1988; 145: 435–441. Shea MT, Widiger TA, Klein MH. Comorbidity of personality disorders and depression: implications for treatment. J Consult Clin Psychol 1992; 60: 857–868. Reich JH, Green AI. Effect of personality disorders on outcome of treatment. J Nerv Ment Dis 1991; 179: 74–82. Sato T, Sakado K, Sato S. Is there any specific personality disorder or personality disorder cluster that worsens the short-term treatment outcome of major depression? Acta Psychiatr Scand 1993; 88: 342–349. Kool S, Schoevers R, de Maat S, Van R, Molenaar P, Vink A et al. Efficacy of pharmacotherapy in depressed patients with and without


40

41 42

43

44

45 46

47 48

49

50 51

52

53

54

55

56

57

58 59

60

61 62

personality disorders: a systematic review and meta-analysis. J Affect Disord 2005; 88: 269–278. Joyce PR, Mulder RT, Luty SE, McKenzie JM, Sullivan PF, Cloninger RC. Borderline personality disorder in major depression: symptomatology, temperament, character, differential drug response, and 6-month outcome. Compr Psychiatry 2003; 44: 35–43. Mulder RT. Personality pathology and treatment outcome in major depression: a review. Am J Psychiatry 2002; 159: 359–371. Russell JM, Kornstein SG, Shea MT, McCullough JP, Harrison WM, Hirschfeld RM et al. Chronic depression and comorbid personality disorders: response to sertraline versus imipramine. J Clin Psychiatry 2003; 64: 554–561. Joyce PR, Mulder RT, Cloninger CR. Temperament predicts clomipramine and desipramine response in major depression. J Affect Disord 1994; 30: 35–46. Joffe RT, Bagby RM, Levitt AJ, Regan JJ, Parker JD. The tridimensional personality questionnaire in major depression. Am J Psychiatry 1993; 150: 959–960. Ebstein RP. The molecular genetic architecture of human personality: beyond self-report questionnaires. Mol Psychiatry 2006; 11: 427–445. Serretti A, Mandelli L, Lorenzi C, Landoni S, Calati R, Insacco C et al. Temperament and character in mood disorders: influence of DRD4, SERTPR, TPH and MAO-A polymorphisms. Neuropsychobiology 2006; 53: 9–16. Millon T, Davis RD. Disorders of Personality: DSM-IV and Beyond, 2nd edn. John Wiley: Toronto, CA, USA, 1995. Battaglia M, Przybeck TR, Bellodi L, Cloninger CR. Temperament dimensions explain the comorbidity of psychiatric disorders. Compr Psychiatry 1996; 37: 292–298. Nierenberg AA. Predictors of response to antidepressants general principals and clinical implications. Psychiatr Clin North Am 2003; 26: 345–352. Esposito K, Goodnick P. Predictors of response in depression. Psychiatr Clin North Am 2003; 26: 353–365. Zanardi R, Cusin C, Rossini D, De Ronchi D, Serretti A. Comparison of response to fluvoxamine in nondemented elderly compared to younger patients affected by major depression. J Clin Psychopharmacol 2003; 23: 535–539. Ezquiaga E, Garcia A, Bravo F, Pallares T. Factors associated with outcome in major depression: a 6-month prospective study. Soc Psychiatry Psychiatr Epidemiol 1998; 33: 552–557. O’Leary D, Costello F, Gormley N, Webb M. Remission onset and relapse in depression. An 18-month prospective study of course for 100 first admission patients. J Affect Disord 2000; 57: 159–171. Serretti A, Kato M, De Ronchi D, Kinoshita T. Meta-analysis of serotonin transporter gene promoter polymorphism (5-HTTLPR) association with selective serotonin reuptake inhibitor efficacy in depressed patients. Mol Psychiatry 2007; 12: 247–257. Klein DN, Schatzberg AF, McCullough JP, Dowling F, Goodman D, Howland RH et al. Age of onset in chronic major depression: relation to demographic and clinical variables, family history, and treatment response. J Affect Disord 1999; 55: 149–157. Koenig HG, George LK, Peterson BL. Religiosity and remission of depression in medically ill older patients. Am J Psychiatry 1998; 155: 536–542. Koenig HG, Cohen HJ, Blazer DG, Pieper C, Meador KG, Shelp F et al. Religious coping and depression among elderly, hospitalized medically ill men. Am J Psychiatry 1992; 149: 1693–1700. Moncrieff J. A comparison of antidepressant trials using active and inert placebos. Int J Methods Psychiatr Res 2003; 12: 117–127. Mandelli L, Serretti A, Colombo C, Florita M, Santoro A, Rossini D et al. Improvement of cognitive functioning in mood disorder patients with depressive symptomatic recovery during treatment: an exploratory analysis. Psychiatry Clin Neurosci 2006; 60: 598–604. Papakostas GI, Petersen T, Mischoulon D, Hughes ME, Spector AR, Alpert JE et al. Functioning and interpersonal relationships as predictors of response in treatment-resistant depression. Compr Psychiatry 2003; 44: 44–50. Hoglend P, Perry JC. Defensive functioning predicts improvement in major depressive episodes. J Nerv Ment Dis 1998; 186: 238–243. Dew MA, Reynolds CF, Houck PR, Hall M, Buysse DJ, Frank E et al. Temporal profiles of the course of depression during treatment—


99

63

64

65

66

67

68

69 70

71

72 73 74

75

76

77

78

79

80

81

82 83 84

predictors of pathways toward recovery in the elderly. Arch Gen Psychiatry 1997; 54: 1016–1024. Murphy Jr GM, Hollander SB, Rodrigues HE, Kremer C, Schatzberg AF. Effects of the serotonin transporter gene promoter polymorphism on mirtazapine and paroxetine efficacy and adverse events in geriatric major depression. Arch Gen Psychiatry 2004; 61: 1163–1169. American Psychiatric Association. Practice guideline for the treatment of patients with major depressive disorder (revision). American psychiatric association. Am J Psychiatry 2000; 157: 1–45. Bauer M, Whybrow PC, Angst J, Versiani M, Moller HJ. World federation of societies of biological psychiatry (WFSBP) guidelines for biological treatment of unipolar depressive disorders, part 1: acute and continuation treatment of major depressive disorder. World J Biol Psychiatry 2002; 3: 5–43. Lundmark J, Bengtsson F, Nordin C, Reis M, Walinder J. Therapeutic drug monitoring of selective serotonin reuptake inhibitors influences clinical dosing strategies and reduces drug costs in depressed elderly patients. Acta Psychiatr Scand 2000; 101: 354–359. Adli M, Baethge C, Heinz A, Langlitz N, Bauer M. Is dose escalation of antidepressants a rational strategy after a medium-dose treatment has failed? A systematic review. Eur Arch Psychiatry Clin Neurosci 2005; 255: 387–400. Lotrich FE, Bies RR, Smith GS, Pollock BG. Relevance of assessing drug concentration exposure in pharmacogenetic and imaging studies. J Psychopharmacol 2006; 20: 33–40. Bech P, Gram LF, Dein E, Jacobsen O, Vitger J, Bolwig TG. Quantitative rating of depressive states. Acta Psychiatr Scand 1975; 51: 161–170. Rush AJ, Gullion CM, Basco MR, Jarrett RB, Trivedi MH. The inventory of depressive symptomatology (IDS): psychometric properties. Psychol Med 1996; 26: 477–486. Trivedi MH, Rush AJ, Ibrahim HM, Carmody TJ, Biggs MM, Suppes T et al. The inventory of depressive symptomatology, clinician rating (IDS-C) and self-report (IDS-SR), and the quick inventory of depressive symptomatology, clinician rating (QIDS-C) and self-report (QIDS-SR) in public sector patients with mood disorders: a psychometric evaluation. Psychol Med 2004; 34: 73–82. Hamilton M. Development of a rating scale for primary depressive illness. Br J Soc Clin Psychol 1967; 6: 278–296. Montgomery SA, Asberg M. A new depression scale designed to be sensitive to change. Br J Psychiatry 1979; 134: 382–389. Mulder RT, Joyce PR, Frampton C. Relationships among measures of treatment outcome in depressed patients. J Affect Disord 2003; 76: 127–135. Bech P, Malt U, Dencker S, Ahlfors U, Elgen K, Lewander T et al. Scales for assessment of diagnosis and severity of mental disorders. Acta Psychiatr Scand 1993; 87(Suppl 372): 37–40. Serretti A, Lattuada E, Zanardi R, Franchini L, Smeraldi E. Patterns of symptom improvement during antidepressant treatment of delusional depression. Psychiatry Res 2000; 94: 185–190. Quitkin FM, Rabkin JD, Markowitz JM, Stewart JW, McGrath PJ, Harrison W. Use of pattern analysis to identify true drug response. A replication. Arch Gen Psychiatry 1987; 44: 259–264. Quitkin FM, McGrath PJ, Rabkin JG, Stewart JW, Harrison W, Ross DC et al. Different types of placebo response in patients receiving antidepressants. Am J Psychiatry 1991; 148: 197–203. Stewart JW, Quitkin FM, McGrath PJ, Amsterdam J, Fava M, Fawcett J et al. Use of pattern analysis to predict differential relapse of remitted patients with major depression during 1 year of treatment with fluoxetine or placebo. Arch Gen Psychiatry 1998; 55: 334–343. Peters EJ, Slager SL, McGrath PJ, Knowles JA, Hamilton SP. Investigation of serotonin-related genes in antidepressant response. Mol Psychiatry 2004; 9: 879–889. Rausch JL, Johnson ME, Fei Y-J, Li JQ, Shendarkar N, Mac Hobby H et al. Initial conditions of serotonin transporter kinetics and genotype: influence on ssri treatment trial outcome. Biol Psychiatry 2002; 51: 723–732. Mitchell AJ. Two-week delay in onset of action of antidepressants: new evidence. Br J Psychiatry 2006; 188: 105–106. Posternak MA, Zimmerman M. Is there a delay in the antidepressant effect? A meta-analysis. J Clin Psychiatry 2005; 66: 148–158. Pollock BG, Ferrell RE, Mulsant BH, Mazumdar S, Miller M, Sweet RA et al. Allelic variation in the serotonin transporter promoter affects

85

86

87

88

89

90 91

92

93

94

95 96 97 98 99

100

101

102

103

104

105

106

onset of paroxetine treatment response in late-life depression. Neuropsychopharmacology 2000; 23: 587–590. Durham LK, Webb SM, Milos PM, Clary CM, Seymour AB. The serotonin transporter polymorphism, 5-HTTLPR, is associated with a faster response time to sertraline in an elderly population with major depressive disorder. Psychopharmacology (Berl) 2004; 174: 525–529. Rush AJ, Fava M, Wisniewski SR, Lavori PW, Trivedi MH, Sackeim HA et al. Sequenced treatment alternatives to relieve depression (STAR*D): rationale and design. Control Clin Trials 2004; 25: 119–142. Donovan SJ, Quitkin FM, Stewart JW, Ocepek-Welikson K, Harrison W, McGrath PJ et al. Duration of antidepressant trials: clinical and research implications. J Clin Psychopharmacol 1994; 14: 64–66. Adli M, Bauer M, Rush AJ. Algorithms and collaborative-care systems for depression: are they effective and why? A systematic review. Biol Psychiatry 2006; 59: 1029–1038. Adli M, Rush AJ, Moller HJ, Bauer M. Algorithms for optimizing the treatment of depression: making the right decision at the right time. Pharmacopsychiatry 2003; 36(Suppl 3): S222–S229. Mann JJ. The medical management of depression. N Engl J Med 2005; 353: 1819–1834. Trivedi MH, Rush AJ, Crismon ML, Kashner TM, Toprac MG, Carmody TJ et al. Clinical results for patients with major depressive disorder in the Texas medication algorithm project. Arch Gen Psychiatry 2004; 61: 669–680. Aitchison K, Huezo-Diaz P, Williamson R, Hoda F, Nash M, Rietschel M et al. Genetic Association Data from Gendep, a Multicentre European Study. XIVth World Congress of Psychiatric Genetics. American Journal of Medical Genetics, Neuropsychiatry Genetics: Cagliari, Itlay, 2006, p 690. Gibbons RD, Hedeker D, Elkin I, Waternaux C, Kraemer HC, Greenhouse JB et al. Some conceptual and statistical issues in analysis of longitudinal psychiatric data. application to the NIMH treatment of depression collaborative research program dataset. Arch Gen Psychiatry 1993; 50: 739–750. Hedeker D, Gibbons RD. MIXREG: a computer program for mixedeffects regression analysis with autocorrelated errors. Comput Methods Programs Biomed 1996; 49: 229–252. Serretti A, Smeraldi E. Neural network analysis in pharmacogenetics of mood disorders. BMC Med Genet 2004; 5: 27. Licinio J, Wong ML. Pharmacogenomics in psychiatry: genomic considerations. Mol Psychiatry 2005; 10: 713. Cardon LR, Bell JI. Association study designs for complex diseases. Nat Rev Genet 2001; 2: 91–99. Kelly PJ, Stallard N, Whittaker JC. Statistical design and analysis of pharmacogenetic trials. Stat Med 2005; 24: 1495–1508. Serretti A, Macciardi F, Cusin C, Lattuada E, Lilli R, Di Bella D et al. No interaction of GABA(A) alpha-1 subunit and dopamine receptor D4 exon 3 genes in symptomatology of major psychoses. Am J Med Genet 1999; 88: 44–49. Lerer B, Macciardi F, Segman RH, Adolfsson R, Blackwood D, Blairy S et al. Variability of 5-HT2C receptor cys23ser polymorphism among European populations and vulnerability to affective disorder. Mol Psychiatry 2001; 6: 579–585. Pritchard JK, Rosenberg NA. Use of unlinked genetic markers to detect population stratification in association studies. Am J Hum Genet 1999; 65: 220–228. Tsapakis E, Basu A, Aitchison K. Transcriptomics and proteomics: advancing the understanding of psychiatric pharmacogenomics. Clin Neuropsychiatry 2004; 1: 117–124. van Rossum EFC, Binder EB, Majer M, Koper JW, Ising M, Modell S et al. Polymorphisms of the glucocorticoid receptor gene and major depression. Biol Psychiatry 2006; 59: 681–688. Binder EB, Salyakina D, Lichtner P, Wochnik GM, Ising M, Putz B et al. Polymorphisms in FKBP5 are associated with increased recurrence of depressive episodes and rapid response to antidepressant treatment. Nat Genet 2004; 36: 1319–1325. Curtis D, Sham PC. Estimated haplotype counts from case–control samples cannot be treated as observed counts. Am J Hum Genet 2006; 78: 729–730. Serretti A, Artioli P. Ethical problems in pharmacogenetic studies of psychiatric disorders. Pharmacogenom J 2006; 6: 289–295.



100

107

108

109

110

111

112

113

114

115

Arias B, Catalan R, Gasto C, Gutierrez B, Fananas L. 5-HTTLPR polymorphism of the serotonin transporter gene predicts nonremission in major depression patients treated with citalopram in a 12-weeks follow up study. J Clin Psychopharmacol 2003; 23: 563–567. Hong CJ, Chen TJ, Yu YW, Tsai SJ. Response to fluoxetine and serotonin 1A receptor (C-1019G) polymorphism in Taiwan Chinese major depressive disorder. Pharmacogenom J 2006; 6: 27–33. Joyce PR, Mulder RT, Luty SE, McKenzie JM, Miller AL, Rogers GR et al. Age-dependent antidepressant pharmacogenomics: polymorphisms of the serotonin transporter and G protein beta3 subunit as predictors of response to fluoxetine and nortriptyline. Int J Neuropsychopharmacol 2003; 6: 339–346. Kato M, Ikenaga Y, Wakeno M, Okugawa G, Nobuhara K, Fukuda T et al. Controlled clinical comparison of paroxetine and fluvoxamine considering the serotonin transporter promoter polymorphism. Int Clin Psychopharmacol 2005; 20: 151–156. Kato M, Fukuda T, Wakeno M, Fukuda K, Okugawa G, Ikenaga Y et al. Effects of the serotonin type 2A, 3A and 3B receptor and the serotonin transporter genes on paroxetine and fluvoxamine efficacy and adverse drug reactions in depressed Japanese patients. Neuropsychobiology 2006; 53: 186–195. Kim DK, Lim SW, Lee S, Sohn SE, Kim S, Hahn CG et al. Serotonin transporter gene polymorphism and antidepressant response. Neuroreport 2000; 11: 215–219. Kim H, Lim SW, Kim S, Kim JW, Chang YH, Carroll BJ et al. Monoamine transporter gene polymorphisms and antidepressant response in koreans with late-life depression. JAMA 2006; 296: 1609–1618. Kirchheiner J, Nickchen K, Sasse J, Bauer M, Roots I, Brockmoller J. A 40-basepair VNTR polymorphism in the dopamine transporter (DAT1) gene and the rapid response to antidepressant treatment. Pharmacogenom J 2006. Kraft JB, Slager SL, McGrath PJ, Hamilton SP. Sequence analysis of the serotonin transporter and associations with antidepressant response. Biol Psychiatry 2005; 58: 374–381.


116

117

118

119

120

121

122

123

124

Lee MS, Lee HY, Lee HJ, Ryu SH. Serotonin transporter promoter gene polymorphism and long-term outcome of antidepressant treatment. Psychiatr Genet 2004; 14: 111–115. Minov C, Baghai TC, Schule C, Zwanzger P, Schwarz MJ, Zill P et al. Serotonin-2A-receptor and -transporter polymorphisms: lack of association in patients with major depression. Neurosci Lett 2001; 303: 119–122. Serretti A, Cusin C, Rossini D, Artioli P, Dotoli D, Zanardi R. Further evidence of a combined effect of SERTPR and TPH on SSRIs response in mood disorders. Am J Med Genet 2004; 129B: 36–40. Smeraldi E, Zanardi R, Benedetti F, Dibella D, Perez J, Catalano M. Polymorphism within the promoter of the serotonin transporter gene and antidepressant efficacy of fluvoxamine. Mol Psychiatry 1998; 3: 508–511. Yoshida K, Ito K, Sato K, Takahashi H, Kamata M, Higuchi H et al. Influence of the serotonin transporter gene-linked polymorphic region on the antidepressant response to fluvoxamine in Japanese depressed patients. Prog Neuropsychopharmacol Biol Psychiatry 2002; 26: 383–386. Yoshida K, Takahashi H, Higuchi H, Kamata M, Ito K, Sato K et al. Prediction of antidepressant response to milnacipran by norepinephrine transporter gene polymorphisms. Am J Psychiatry 2004; 161: 1575–1580. Yu YW, Tsai SJ, Chen TJ, Lin CH, Hong CJ. Association study of the serotonin transporter promoter polymorphism and symptomatology and antidepressant response in major depressive disorders. Mol Psychiatry 2002; 7: 1115–1119. Zanardi R, Benedetti F, DiBella D, Catalano M, Smeraldi E. Efficacy of paroxetine in depression is influenced by a functional polymorphism within the promoter of serotonin transporter gene. J Clin Psychopharmacol 2000; 20: 105–107. Zanardi R, Serretti A, Rossini D, Franchini L, Cusin C, Lattuada E et al. Factors affecting fluvoxamine antidepressant activity: influence of pindolol and 5-HTTLPR in delusional and nondelusional depression. Biol Psychiatry 2001; 50: 323–330.