Estrogen Patch a Boon for Schizophrenia in Women?

Pauline Anderson

August 14, 2019

Adjunctive estrogen patches significantly improve positive and negative symptoms of schizophrenia in premenopausal women, new research suggests.

The study, a randomized controlled trial, replicates previous research but also extends it. The beneficial effect of the patch was especially evident in women older than 38 years.

The new findings suggest that estrogen patches should be considered for premenopausal women with schizophrenia, lead investigator Mark Weiser, MD, Sheba Medical Center and Tel Avi University, Israel, and associate director for treatment trials at Stanley Medical Research Institute, a nonprofit charitable organization in Kensington, Maryland, told Medscape Medical News.

"Pharmaceutical companies have put a lot of money into compounds which look very promising early on but later didn't pan out. Psychiatry is full of these promising discoveries and new hope for patients, which make a lot of noise and then disappear," said Weiser.

These new study findings "are for real," he added.

However, at least one expert had concerns about the study results, including the magnitude of improvement.

"Things can be statistically different but clinically meaningless," Dolores Malaspina, MD, professor of psychiatry, neuroscience, genetics, and genomic sciences at Icahn School of Medicine at Mount Sinai, New York City, told Medscape Medical News.

The study was published online July 31 in JAMA Psychiatry.

No Commercial Involvement

In previous studies, use of the estradiol patch led to improvement in measures of positive symptoms and in scores on the general psychopathology subscale and total Positive and Negative Syndrome Scale (PANSS) in women. In the largest of these studies, a higher dose (200 μg) had greater effects than a lower dose (100 μg).

The researchers aimed to replicate the results for the higher dose of estradiol using the same study design but in a different country and "without any commercial involvement," which, Weiser said, "is extremely rare."

The study included 200 premenopausal women (mean age, 38 years) who lived in the Republic of Moldova, had schizophrenia or schizoaffective disorder, and were receiving antipsychotic medication.

Weiser noted that he is based in Israel, so it is easier to carry out studies in Eastern Europe than in the United States and that he knows researchers in Moldavia to be "trustworthy and experienced."

Participants had to have a total PANSS score of 60 or greater. The mean baseline PANSS score was 82.8, and the mean PANSS positive subscale score at baseline was 19.6, reflecting severe symptomatology.

The primary outcome was change in the positive PANSS subscale score. Secondary outcomes included total PANSS score and changes in negative symptoms.

Positive symptoms of schizophrenia refer to such things as delusions and hallucinations, whereas negative symptoms include lack of initiative and poor social skills that may result in difficulty holding down a job or maintaining relationships. Symptoms on the PANSS are rated from 1 to 7, with 1 being normal.

Participants were randomly assigned to receive either an estradiol patch or a placebo patch. Twelve participants did not complete the 8-week study.

Investigators found that participants in both study groups experienced considerable improvement. At week 8, the mean PANSS positive subscale score was 14.4 in the placebo group and 13.4 in the estradiol group.

The estradiol group had a statistically significant improvement vs placebo in the PANSS positive subscale scores (–0.94; 95% confidence interval [CI], 1.64 to –0.24; P = .008; effect size, .38.)

The treatment group also had better outcomes on the PANSS total score (difference, –4.10; 95% CI, –6.73 to –1.47; P = .002; effect size, 0.45) and on the PANSS subscales of negative symptoms and general psychopathology.

"Really Exciting"

"What's really exciting about our data is that we found improvements in the positive but also the negative symptoms above and beyond the improvement with antipsychotics," Weiser said.

The effect sizes "might be considered somewhat significant from a clinical perspective," the investigators write.

There were also significantly greater improvements in the treatment group vs the placebo group in Clinical Global Impression–Severity and Montgomery Depression Scale scores.

During post hoc analyses, the researchers unexpectedly found that response to treatment was much better in women older than 38 years. In this group, the effect size was 0.79 on the positive PANSS subscale score. Estradiol had no significant effect on younger women.

The investigators were somewhat surprised by this finding.

"We looked at the data to better understand the profile of people who responded and didn't respond, and this jumped out at us. We then went ahead and analyzed this more fully," said Weiser.

Results for women aged 38 to 45 years were "certainly clinically significant," he added. The investigators are currently conducting a randomized trial of the estrogen patch vs placebo in women in this age group.

In the current study, the baseline dose or type of antipsychotic drug did not affect the response to estradiol, nor did baseline symptom severity.

Weiser noted that "everything got a little bit better" rather than only certain symptoms improving.

As expected, more women in the treatment group than in the placebo group reported breast discomfort and weight gain.

The investigators also carried out a meta-analysis of four studies, including the current one.

This meta-analysis showed an effect size of –0.318 (95% CI, –0.504 to –0.132; P = .001) on the PANSS total score favoring estradiol vs placebo. Results were similar across subscales.

Unclear Mechanism

The mechanism by which estrogen patches might alleviate schizophrenia symptoms is unclear. This is also the case with antipsychotic drugs, said Weiser.

"We know there are estrogen receptors in various areas of the brain that are certainly relevant for schizophrenia, but we don't know exactly what they're doing there or what the estrogen is doing," he added.

That this new study replicated previous findings is novel, Weiser noted. "In my career, I have done many replication studies of a great new finding from somebody, and they have all failed," he said.

Weiser is keen to investigate whether boosting hormone levels is effective in older women with schizophrenia. He has applied for a grant from the Stanley Medical Research Institute to carry out a study involving women aged 50 to 60 years who are in early menopause.

This strategy may also be effective in men with schizophrenia, because men also have estrogen receptors in the brain. However, such a study would be "problematic," because the men would develop secondary female characteristics, Weiser said.

Whether the estrogen patch is effective as monotherapy is an issue for further research, he said.

Getting the estradiol patch approved for the treatment of schizophrenia would be a difficult, because the US Food and Drug Administration's approval process typically requires "very stringent and expensive testing," Weiser noted.

"Estrogen is not patented, and nobody is going to make a billion dollars by doing these studies," he said.

However, he added, "there are many off-label medications used in psychiatry."

A limitation of the study was that it did not determine the durability of the treatment effect after treatment was stopped. In addition, the dose used in the trial was relatively low, and inclusion criteria did not differentiate between patients with acute, chronic, and/or treatment-resistant schizophrenia.

Retroactive Trial Registration

An accompanying editorial by Dost Ongür, MD, PhD, McLean Hospital, Belmont, Massachusetts, who is editor of JAMA Psychiatry, outlined the challenges of conducting complex international trials.

The original plan was to conduct the study in Romania as well as Moldavia, but the trial's approval in Romania was delayed. Moldavia is not a member of the European Union, so there was a problem registering the study. The study had to be registered retroactively.

"In reaching a decision to publish their article, we considered the investigators' transparency concerning the problems they experienced as well as the scientific merits of the work," writes Ongür. "The investigators have described the relevant events in the article and supplementary materials online, and the retroactive trial registration is freely available on"

Malaspina added that she had several serious concerns with the study. These included the retroactive trial registration and the post hoc analyses.

Ongür acknowledged to Medscape Medical News that such a move is "unusual" and that this was the reason for his editorial comment.

He pointed out that the study protocol had been submitted to the European registry but was never posted. However, it is now registered on

The fact that the analysis, which showed a better response in women aged 38 to 45 years, was not included in the protocol is not unusual, said Ongür.

"Investigators often carry out post hoc analyses that were not prespecified in the protocol, and in that case, it is important to acknowledge these analyses were not in the protocol by labeling them as post hoc," he said.

Weiser reiterated that post hoc analyses are a "scientifically essential part of the data analysis as long as this is clearly stated in the manuscript," as was done in this case.

It wasn't clear to Malaspina how the post hoc analysis was carried out.

"I don't know how they looked at the data; they could have just divided it at any point," she said.

Weiser explained that the study used a median split, "which we and others have used previously in many papers." The investigators also performed an analysis using age as a continuous variable, which also showed a significant effect of age on response, he added.

Questions Asked and Answered

Malaspina also noted that the study did not assess pretreatment estrogen levels and that "some women with normal or even elevated levels may have received the intervention, which is a concern."

Weiser pointed out that the investigators analyzed the issue concerning plasma estradiol levels "in depth" in the Web supplement to the article. They provided information on estradiol measurements both at baseline and at the end of the study and included an exploratory analysis of the role of plasma estradiol on the effect of the estradiol patch on PANSS, he said.

Estrogen treatment is not without risks, said Malaspina. She noted that a meta-analysis published last year in JAMA Psychiatry showed that women with schizophrenia are at increased risk for breast cancer.

Although the authors did discuss the possible risk for cancer, "it's not considered to be a limitation of the study," said Malaspina.

She also took issue with the magnitude of improvement in the study. Although the authors highlighted significant improvement in positive symptoms, this may not mean there was a change in functioning, she said.

How large an improvement in PANSS scores has to be in order to be clinically meaningful "is in the eye of the beholder," Weiser responded.

He pointed out that the effect sizes for PANSS total and positive symptoms uncovered in this study "are similar, and in some cases higher" than those found in previous studies comparing antipsychotics with placebo.

This new study also showed there was improvement in negative symptoms, "which are highly correlated with social abilities and functioning," Weiser said.

Although the authors maintain their "independence," Malaspina noted that Weiser is not without conflicts. "He has a very big conflict. He directs the use of all the money in the Stanley Foundation, so he would want to show an important positive outcome," she said.

In response, Weiser said that over several years, he has performed many replication studies funded by the Stanley Medical Research Institute that showed negative findings, which have been published. These included studies of allopurinol, minocycline, raloxifene, valnoctamide, and d-serine.

The study was funded by the Stanley Medical Research institute. Weiser and Malaspina have disclosed no relevant financial relationships.

JAMA Psychiatry. Published online July 31, 2019. Abstract, Editorial

For more Medscape Psychiatry news, join us on Facebook and Twitter


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.