Mood disorders and consequences of pharmacological treatment

Abstract

Mental health issues are medical problems getting increased attention throughout the world, and depression, only one among several types of mental illnesses, are currently reported as the second leading cause of disability world wide by the World Health Organization. Along this development, increased number of psychiatric diagnoses and increased utilization of medication can be observed. However, whether this increase is because of an actual increase in number of individuals suffering from these disorders, or due to improvements in diagnostics and understanding, or both, is not fully known. What is known, however, is that mental health related suffering is nothing new and can be traced back several thousands of years in human history. Throughout this history, our understanding and view of these disorders that affect our mood and behavior have changed substantially, from being assumed caused by gods or magic, to our modern view of combined environmental and genetic causes. But psychiatry is still a comparatively recent field of medicine and it was not long ago effective drugs targeting these disorders were first introduced. Our understanding of the underlying mechanisms of this suffering is still limited, but recent developments in psychiatric genetics have provided some of the first stable biological underpinnings to mental disorders. In light of these findings, a complicated picture is beginning to take form, in which the etiology appears caused by thousands of genetic factors, but also that a lot of the underlying genetics is shared among the disorders previously thought of as separate.

It is within this setting, where many of the current disorder definitions are starting being questioned and the appreciation of genetics behind these disorder is increasing, the studies within thesis have been conducted.

In study I, the potential side effect of manic switching due to antidepressant treatment was investigated in a population sample of 3,240 individuals with bipolar disorder. A within-individual design was used to adjust for otherwise unmeasured genetic confounding, and the results indicated that manic switching was confined to bipolar disorder patients treated with an antidepressant monotherapy, whereas patients treated with an antidepressant in combination with a mood stabilizer rather displayed a reduced risk.

Study II instead focused on depression around the time of pregnancy, perinatal depression, and to what extent genetics explain the variance in this disorder, and to what extent perinatal depression genetically overlap with depression at other time of life. This was studied with a twin design in a sample of twin mothers (N=3,427) that had answered the Edinburgh Postnatal Depression Scale, and with a sibling design in a sample of sisters (N=580,006) using register data of healthcare contacts. The genetic contribution to, or heritability of, perinatal depression was estimated at 54 and 44% in the twin and sibling designs respectively. Using a bivariate model, a third of the genetic contribution to perinatal depression was found unique for the disorder and not shared with depression at other times of life.

Study III was also related to perinatal depression, but focused on potential side effects of antidepressants. In a sample of 392,029 pregnancies, associations between prenatal SSRI exposure and offspring birth size and gestational age was observed. This was followed by within-family analyses (N=1,007) that adjusted for genetic and familial environmental confounding, where the associations between SSRIs and offspring birth size was attenuated, indicating that these associations were likely due to familial confounding. An association between prenatal SSRI exposure and reduced gestational age was observed in both analyses, and could be either due to a causal effect of the medication, or due to confounding factors that the withinfamily design could not adjust for.

Perinatal depression was further explored in study IV, where patterns of healthcare utilization were studied among both mothers and fathers based on register data. This included 3.6 million parents and 3.5 million pregnancies, and the occurrence of diagnoses for depressive illness, anxiety disorders, and mental illness in general around the time of pregnancy was contrasted to the occurrence of these diagnoses at other times of life. Overall, a reduction in healthcare utilization for all studied disorder types were observed around this time of life, which may indicate barriers to getting a diagnosis during this time of life.

In conclusion, the studies within this thesis demonstrate that genetically informed designs are very useful in epidemiological research. And through the application of these designs with large-scale register data, the studies of this thesis provide enhanced understanding of mental illness in general, and bipolar disorder and perinatal depression in particular.