As a mom of young children, I was very much affected by the Andrea Yates case. If you recall she experienced successive bouts of psychosis and pursued multiple attempts at suicide following the delivery of each of her children. After her fifth child, she drowned her children, killing them all. The case was a heartbreaking, and I believe, a totally preventable tragedy had her psychosis been taken seriously by medical professionals, family and others in the community. In spite of her psychosis and suicide attempts she was sent home to care for her children, as if a woman with postpartum psychosis is somehow less ill than a man or than a woman whose psychosis develops unrelated to her pregnancies. Raising young children is difficult, even under the best of circumstances, but sending a severely depressed and psychotic woman home to care for young children is just downright negligent. Although there was blame to go around among the doctors, her family and community, I couldn’t but shake the feeling that this tragedy could have been averted if her illness had been taken more seriously.
Identifying the Biological Underpinnings of Maternal Psychiatric Disturbances
The Andrea Yates tragedy inspired me to research and identify the clinical and biological components of perinatal mental illness. My goal was to identify early warning markers; biological tests, that would give women, their physicians and family members a way to predict the possibility of illness and confirm that illness once it had arrived. I thought that if we could predict and identify the risk for this illness, then the families could prepare and maybe even lessen the severity the disease process. At the very least, tragedies like the Andrea Yates case could be prevented.
I knew hormones would be key to the onset and maintenance of perinatal psychiatric symptoms. What I didn’t know is which hormones, when, and related to which symptoms. It seems that no one else did either. Despite years of research and a clear temporal association between the onset of psychiatric distress and childbirth, only tenuous connections between maternal hormone concentrations and varying degrees of postpartum depression had ever been established. This was primarily because the research was focused so narrowly upon the relationships among what are often referred to as the female hormones, progesterone and estradiol, and depressive symptoms. Very little research had examined associations between a broader range of steroid hormones and the full spectrum of potential psychiatric symptoms. This didn’t make sense to me. Certainly, other hormones affected by pregnancy, might also impact brain chemistry; certainly, the range of clinical symptoms that women might experience would go beyond the blues and depression. Even when psychosis appeared, I wasn’t convinced that the psychosis of pregnancy and postpartum was clinically similar to the psychoses that developed irrespective of the vast biochemical changes that took place across pregnancy, parturition and in the weeks and months that followed. If the biochemistry was different, as it most necessarily had to be, wouldn’t everything else about maternal psychiatric disturbances be different as well?
Looking Beyond the Boundaries
And so began my research. For the first study: Beyond Progesterone and Estrogen: Maternal Psychiatric Disturbances Linked to Adrenal Androgens, I recruited healthy, medication free, first time moms, with no previous history of mental illness. This was no easy feat. I soon realized that many women, even pregnant women, were using antidepressants and anxiolytics and many other medications. It seems the old adage that pregnant women should not take medications lest it cross the placental barrier and affect the developing fetus, had fallen by the wayside.
To assess the psychiatric distress, I abandoned the singular blues, depression and anxiety scales used so often in this research and found a broad-based, standardized assessment of psychiatric distress called the Symptom Check List 90R (SCL90R). SCL-90R is a 90-item psychiatric self-report inventory designed to measure the severity and intensity of psychiatric symptoms in both inpatient and outpatient populations. Participants rate the severity of distress experienced during the prior seven-day period using a 0-4 Likert-type scale (0=no distress-“not at all” to 4=extreme). Symptoms measured included: anxiety, hostility (aggression, irritability, etc.) phobic anxiety, paranoid ideation, psychoticism, somatization (perceptions of pain or other physical disturbances), obsessive-compulsive behavior, interpersonal sensitivity (feelings of personal inadequacy), depression and the global severity index (GSI), which reflects the overall symptom severity.
Along with the clinical symptoms, I measured five hormones, progesterone, DHEAS, testosterone, estrone, estradiol and estradiol, using saliva based testing. Symptoms and hormones were assessed twice, first in late pregnancy at 37 weeks (n =32) and again within 10 days following the delivery of their children (n=28, four were lost to attrition). I also conducted a year long follow up of those same participants and will report those data soon.
It’s Not Just Depression and It’s Not Just Postpartum
As I suspected, symptoms were present in late pregnancy and in some cases, increased in severity postpartum, but in other cases, decreased in severity. For some women, pregnancy was more problematic than postpartum, especially those with obsessive compulsive symptoms.
Fully 50% of the women tested experienced symptoms during pregnancy and 57% postpartum. This means maternal psychiatric distress is far more common than generally ascribed. As a group the anxiety related symptom scales, particularly the anxiety and obsessive compulsive scales, had the highest individual scores at each test time and when combined with hostility, phobia and psychoticism contributed the largest increase in symptom severity from pregnancy to postpartum. So it wasn’t the blues and depressive type symptoms that were most troubling, but the agitated, anxiety and even psychotic type symptoms that were the most severe.
Current research suggests that for only 1-2 per 1000 pregnancies psychosis will develop. What I found with this research and from another study, is that psychotic symptoms were far more prevalent than recognized and may be the symptoms that drive the depression. In this study, we found sub-threshold, but clinically relevant, psychotic symptoms present in several of the women postpartum. Their symptoms were absent concurrent elevations in paranoia (paranoia and psychosis often go hand in hand). The most frequently ascribed to symptoms within this cluster included fears of serious illness (n=8), loss of mind (n=7) and isolation (n=12). Surprisingly, three women showed mild to moderate distress about thought insertion and thought broadcasting, two were concerned about thought control and one woman indicated distress about auditory hallucinations. Interestingly, it was these very same women who had the most dysregulated hormone profiles.
In speaking with the women who indicated these symptoms, the visual hallucinations, involved their children suffering; usually graphic intrusive thoughts, seeing images of their children being burned, thrown out windows, cut with a butcher knife, strangled with the breast pump tubing and the like. When auditory hallucinations were present they berated the women for their weakness, bad mothering etc., inducing guilt and one can only assume, depression. We confirmed the prevalence of these types of symptoms in two subsequent studies, the first published here: Dimensions of postpartum psychiatric distress: preliminary evidence for broadening clinical scope, the second unpublished as of yet.
Aberrant Androgen Metabolism may be to Blame for Maternal Psychiatric Symptoms
As I suspected and as much research had shown, no symptom clusters were correlated with progesterone, estrone or estriol either pre- or postpartum. While expected to be a close correlate of postpartum psychiatric symptoms, estradiol was associated with very few symptom clusters in the present study. Instead, it was the androgens that were linked to the symptoms at both time periods and not in a way that might be expected.
Low late pregnancy testosterone was not only related to late pregnancy psychiatric symptoms, but significantly predicted postpartum symptom severity. In conjunction, and this is where the endocrinology gets interesting, elevated late pregnancy DHEAS and supra-elevated postpartum DHEAS were associated with pre – and postpartum symptoms, respectively. This was exciting, because in theory these two hormones should not be aligned. That is, high DHEAS should correlate with high testosterone and it didn’t. So somewhere between DHEAS>DHEA>androstenedione> testosterone there was a problem and I had pretty good idea where.
For now though, we had a pilot study that ripped open the notions that maternal psychiatric distress occurred only during postpartum, was depressive in nature, was rare and was related to the normal or expected hormone changes of pregnancy. It was none of these things. The psychiatric distress was present at both test points, was more agitated, included a spectrum of symptoms, and most importantly, was related to aberrant changes in hormones that were likely exacerbated by the normal or expected hormone disruptions of pregnancy. I was very excited. If we could identify the problem, then we could fix it right?
Not so fast. I could never get the research published and though I carried on with research I could do without funding, including a long term follow-up of the same participants (to be self-published soon) and an online study of the symptoms of psychiatric distress, the hormone work was routinely and summarily rejected. I learned very quickly how controversial studying hormones in women’s health was. So there it stands, the work was good, it pointed to a biomarker that could be used to identify and then treat a group of women who suffer horribly, but the study needs to be replicated with a much larger and more diverse population of women. It is likely that this is but one of many potential markers along this hormone pathway that could be used to predict and prevent perinatal psychiatric distress. It is also likely that this pattern of metabolism is linked to a host of other mental health and physical health issues. It was because of this research that I began Hormones Matter and have worked so arduously to increase awareness about the need for more research in women’s health. Hormones ought to be measured consistently across a woman’s life span, they aren’t and we need to change that.
Here are the full study details and the article, now officially self-published: Beyond Progesterone and Estrogen: Maternal Psychiatric Disturbances Linked to Adrenal Androgens.
Another portion of this study included assessing cognitive changes: Mommy Brain: Pregnancy and Postpartum Memory Deficits.
We Need Your Help
More people than ever are reading Hormones Matter, a testament to the need for independent voices in health and medicine. We are not funded and accept limited advertising. Unlike many health sites, we don’t force you to purchase a subscription. We believe health information should be open to all. If you read Hormones Matter, like it, please help support it. Contribute now.
Yes, I would like to support Hormones Matter.
Just came across your work while checking pregnancy effects on memory and psychiatric symptoms for a report I’m doing – fascinating stuff – thanks for persevering against the odds with this ground-breaking stuff. All the best with your future research. Have you considered peer-reviewed publications based in Europe to showcase your work?
Thank you. I am so glad you found it useful. Regarding where to submit – I submitted everywhere and it was rejected numerous times both in US and international journals. Because you have to submit the article to each journal independently, wait for the review, this process ate up a few years. I finally gave up and when I began HM, decided to publish it myself.
Thanks for the info. I am in a WHNP program, we just covered this topic and androgens were not mentioned at all! Looking forward to future research.