Guest Commentary by Rajnish Mago, MD
- Testosterone improves antidepressant-emergent loss of libido in women: findings from a randomized, double-blind, placebo-controlled trial
- Effects of antidepressants and soybean association in depressive menopausal women
- Meta-analysis of selective serotonin reuptake inhibitor-associated QTc prolongation
- EEG neurofeedback treatments in children with ADHD: an updated meta-analysis of randomized controlled trials
- The McGill Geriatric Lithium-Induced Diabetes Insipidus Clinical Study (McGLIDICS)
Fooladi et al.
J Sex Med. 2014;11(3):831-9.
Why is this paper important?
Serotonergic antidepressants (SSRIs or SNRIs) are associated with a high incidence of sexual dysfunction. While there are many strategies for managing this problem, each of them has some limitations, so there is no doubt that we need more options. Previously, testosterone has been shown in several studies to be efficacious for decreased libido in both menopausal and premenopausal women who were not depressed or on antidepressants. So, it is a good question as to whether testosterone would work for women with decreased libido that emerged or worsened after an antidepressant was started for the treatment of a depressive disorder.
- To assess whether transdermal testosterone can treat loss of libido due to an SSRI or SNRI
- Study design: Randomized, double-blind, placebo-controlled clinical trial
- Subjects: Forty-four women, aged 35-55 years, on a stable dose of SSRI or SNRI, who had loss of libido that started during treatment with the antidepressant
- Interventions: 12-week treatment with either transdermal testosterone patch (300 mcg of testosterone/day) or a placebo patch
- At baseline, there were no differences between the treatment groups. (Good that they checked for this.)
- The authors looked at improvement in sexual functioning in three different ways:
- First, they looked at their “primary outcome measure” — how much the score on the Sabbatsberg Sexual Self-rating Scale had changed from before treatment to after treatment. On this, there was no difference between patients on transdermal testosterone or on placebo. (Disappointing!)
- Next, they looked at how many Satisfactory Sexual Events occurred over 4 weeks. The number of such Satisfactory Sexual Events increased more in patients on transdermal testosterone (average increase of 2.3 events) than in those on placebo (average increase of 0.1 events). This difference was statistically significant, which means that it is unlikely that this difference was just due to chance.
- Thirdly, they looked at the change in scores on the Female Sexual Distress Scale-Revised before and after treatment. Patients on transdermal testosterone had greater improvement than patients on placebo, but the difference was not statistically significant. What this means is that we cannot rule out the possibility that the difference in improvement could have occurred by chance. But wait a minute…the p value was .06. That means that they estimated that there was a 6% chance that the difference could have occurred simply due to chance. We can say, as the authors did, that the finding “approached statistical significance.” As you know, we have collectively decided to only consider as “statistically significant” findings that have less than a 5% probability (p < .05) of being due to chance alone. (6% vs. less than 5% -- so close! But sorry guys, if we decided that less than 5% was the finishing line, it would be cheating to change it later.)
- What did the transdermal testosterone do to the serum testosterone level? The mean total serum testosterone level at 12 weeks in patients on transdermal testosterone was 2.1 nmol/L.
- Any side effects? No androgenic adverse effects.
- Well, at least transdermal testosterone increased the number of Satisfactory Sexual Events in these patients.
- The authors wondered if the lack of improvement in the Sabbatsberg Sexual Self-rating Scale total score may reflect lack of sensitivity of this scale for measuring change in sexual function.
- That testosterone was superior to placebo on some measures but not on the primary outcome measure may be because the relatively small sample size. When you study something in a smaller number of people, it is naturally hard to be sure that any differences observed are not due to chance alone. That is what “statistical significance” means.
- Now, here’s another tip: whenever this happens, look at whether the treatment at least appeared to be better than placebo in terms of numbers, even if you cannot be sure that this apparent superiority was not just due to chance. In this study, testosterone was numerically superior to placebo on the Sabbatsberg scale total score as well as subscales.
- It is important to note that patients in this study were not selected for having low testosterone levels. For total testosterone, the normal range for premenopausal women is 0.33–1.7 nmol/L. In this study, the mean level was 0.52 prior to treatment (both groups combined) and 2.1 after treatment (testosterone group). Unfortunately, they were not systematic in when in the day they measured testosterone and they did not measure free testosterone. Therefore, we don’t know whether baseline testosterone level or change in testosterone level can predict which patients are more likely to benefit. One would guess that testosterone would be most helpful in those with low levels before treatment that improve with the medication, but we must wait for this to be tested in a future study.
Estrella et al.
Acta Pol Pharm. 2014. PubMed PMID: 25272653.
Why is this paper important?
Patients frequently ask us about the use of complementary treatments. In addition, many patients use complimentary treatments without telling their physicians. Therefore, in my experience, clinicians are very interested in learning more about the potential use of complimentary treatments in mental health. One of these potential treatments is soy isoflavones that were evaluated in this study for augmenting an antidepressant.
- Depressive disorders in menopausal women are well known and may be related in part to decline in estrogen levels after menopause.
- This was a pilot study that assessed whether adding soy isoflavone extract to an SSRI may be more efficacious than an SSRI alone in menopausal women.
- 40 menopausal women with clinical depression (diagnosis unspecified) were recruited
- The subjects were treated with one of the following for three months:
- fluoxetine (10 mg/day),
- soy isoflavone extract (100 mg/day),
- sertraline (50 mg/day), or
- sertraline (50 mg/day) plus soy isoflavone extract (100 mg/day).
- Soy isoflavone concentrate supplied by GNC was used. It comes as 50 mg capsules.
- Change in Hamilton Depression Rating Scale scores from baseline to end of three months of treatment was statistically significantly greater in patients who received the combination of sertraline and soy isoflavone compared to sertraline alone or to soy isoflavone alone.
- The combination was not shown to be superior to fluoxetine alone.
- On the Zung patient-rated depression scale, the combination of sertraline and soy isoflavone was statistically significantly superior to each of the other treatment groups.
- This study suggests that in menopausal women with depression, adding 100 mg/day of soy isoflavone to an SSRI may be more efficacious than the SSRI alone.
- Soy isoflavones are phytoestrogens for which there is a lot of literature. A number of benefits have been claimed: relief of menopausal symptoms, and prevention of heart disease, breast cancer, prostate cancer, and osteoporosis. However, they may also be associated with some risks like increased risk of breast cancer, male hormonal and fertility problems, and hypothyroidism.
- You must have noticed that the dose of fluoxetine was less than what we usually prescribe and the dose of sertraline, while within the therapeutic range, was on the low side. However, if this supplement does work (and that is not yet convincingly shown), maybe one could get away with using lower doses of the antidepressant?
- My main concern with this study is that it is not clear from the paper that there was adequate blinding of the subjects and the treating clinicians. Thus, a placebo effect cannot be ruled out.
- Because there was no placebo group, I cannot agree with the Authors’ conclusion that soy isoflavone alone was also efficacious. The finding that patients on soy isoflavone alone improved does not mean that it was efficacious. The improvement may have been spontaneous or due to a placebo effect.
- At GNC (a chain of stores that sell health- and nutrition-related products), sixty capsules of soy isoflavone concentrate 50 mg each (a month’s supply at two capsules per day) costs about $12, so at least this supplement is not particularly expensive.
Beach et al.
J Clin Psychiatry. 2014;75(5)
Why is this paper important?
- Ever since the FDA first issued a Drug Safety Communication in 2011 about the risk of QTc prolongation with citalopram, there has been a lot of interest in this topic.
- Is it really true? If studies disagree, what should we conclude if we put all the prospective studies together? A well-done meta-analysis can move the literature forward by combining a large number of studies into one result
- Also, does citalopram have a greater risk of prolonging QTc or is this something that all the SSRIs do equally?
- To use meta-analysis of all available prospective studies to evaluate whether use of selective serotonin reuptake inhibitors (SSRIs) is associated with prolongation of the corrected QT interval (QTc)
- Search of PubMed/MEDLINE database (January 1, 1975-August 15, 2012)
- Also, hand searching of references of the articles
- Included articles in English, Spanish, and German (Good!)
- Two reviewers independently identified the studies and three reviewers independently extracted the data (Very good!)
- Sixteen articles involving 4,292 patients were included.
- SSRIs were associated with a statistically significant, increase in QTc interval compared to placebo (average of +6.10 milliseconds)
- Tricyclic antidepressants (TCAs) were associated with a statistically significantly greater QTc increase than SSRIs (average 7.05 milliseconds greater than SSRIs)
- Citalopram was associated with statistically significantly greater QTc prolongation than sertraline, paroxetine, and fluvoxamine.
- SSRIs were associated with a small but statistically significant increase in the QTc interval
- The increase in QTc interval with SSRIs was less than with TCAs.
- Citalopram was associated with a greater QTc prolongation than three of the other SSRIs.
- By the way, I have reviewed this topic elsewhere (Mago et al. Cardiovascular adverse effects of newer antidepressants. Expert Rev Neurother. 2014;14(5):539-51. PubMed PMID: 24738823). Without going into more detail here, to me there is little doubt that SSRIs can prolong QTc. In my opinion, it would not be correct to say that this association has not been convincingly shown.
- Thankfully, this is not a problem in most patients. But in a high-risk patient it can become an issue and may be associated with increased risk of serious ventricular arrythmias. In such patients, caution and monitoring are advisable.
Micoulaud-Franchi et al.
Front Hum Neurosci. 2014;8:906
Why is this paper important?
- ADHD is diagnosed in an increasingly larger number of children
- Two problems: sometimes the parents are reluctant to put the child on medication and medications are not usually 100% effective
- There are many different non-medication treatments for children with ADHD that are talked about. One of these is biofeedback using EEG. But does it work? We hear contradictory things about it.
- Well, this paper puts together the studies on this topic and helps us to come to a state-of-the-art conclusion
- To conduct a meta-analysis of published randomized controlled trials that evaluated whether EEG neurofeedback works for children with Attention Deficit Hyperactivity Disorder (ADHD).
- Only randomized controlled trials that used either a semi-active control (a control treatment that has known efficacy to some extent, e.g., EMG biofeedback) and sham control (a control “treatment” that does not have any efficacy other than a placebo effect) were included
- The paper evaluates efficacy of EEG neurofeedback for overall symptoms, for inattention, and for hyperactivity/impulsivity
- The studies included looked at
a) parent assessment that was probably unblinded assessment, and
b) teacher assessment that was probably blinded assessment
- Five studies were included
- These included a total of 263 children with ADHD
- On parent assessment (probably unblinded assessment), EEG neurofeedback was statistically significantly better than the control interventions on the overall ADHD score, the inattention score, and the hyperactivity/impulsivity score.
- On teacher assessment (probably blinded assessment), EEG neurofeedback was superior to control on only the inattention score, and not on the overall ADHD score or on the hyperactivity/impulsivity score.
- EEG neurofeedback seems to be efficacious at least for inattention in children with ADHD
- A limited number of patients have been studied, significant bias is introduced by how the assessment was conducted, and I have many technical questions about how to actually implement this intervention. Therefore, at this time I don’t think that we should be routinely recommending EEG neurofeedback for patients with ADHD.
- Previous literature has shown that if you consider assessments that are probably unblinded, you will get a much greater apparent benefit from EEG neurofeedback. It is very important to realize that when parents are helping the child do the EEG neurofeedback, they are not blinded. Therefore, their assessment of how the child is doing is should not be the main way of evaluating the efficacy of a treatment. If we read a study of any intervention for ADHD in children, we should immediately look to see if the assessment was truly blinded.
Rej et al.
Can J Psychiatry. 2014;59(6):327-34
Why is this paper important?
- While all prescribers are well aware of the importance of checking serum creatinine in patients being treated with lithium, in my experience, polyuria and nephrogenic diabetes insipidus are often not monitored for and are frequently missed even when present
- Can’t we just monitor for nephrogenic diabetes insipidus by asking patients about their symptoms like how much they have been urinating?
- We usually have serum sodium available as part of the Basic Metabolic Panel (a group of laboratory tests commonly done together, at least in the USA). Is that a good indicator for diabetes insipidus because serum sodium tends to increase in patients who have a decrease in the kidney’s ability to concentrate urine?
- What about urine specific gravity? We get that result as part of a simple urinanalysis, but getting urine osmolality will require ordering it separately
- Nephrogenic diabetes insipidus is a common and potentially serious condition associated with lithium treatment
- This study aimed to compare nephrogenic diabetes insipidus symptoms and urine osmolality between elderly and adult patients.
- It also assessed whether symptoms, serum sodium, and urine specific gravity can be used as surrogate measures of decreased urine osmolality
- Lastly, it assessed whether certain clinical factors that may be associated with nephrogenic diabetes insipidus are independently correlated with decreased urine osmolality.
- Cross-sectional study (i.e., no follow up)
- 100 consecutive outpatients treated with lithium from 6 tertiary care clinics
- 45 of these patients were geriatric (aged 65 years and older) and 55 were adults (aged 18 to 64 years).
- Assessments were a patient-rated symptom questionnaire and laboratory tests (including urine osmolality, serum sodium, and urine specific gravity)
- The proportion of patients on lithium who had decreased urine osmolality (defined as < 300 mOsm/kg) was not statistically significantly different between geriatric (12.5%) and adult (17.9%) adults
- Geriatric patients reported statistically significantly less symptoms than adult patients
- Urine osmolality did not correlate with symptoms or current serum sodium
- However, low urine osmolality (i.e., < 300 mOsm/kg) was suggested by a urine specific gravity of < 1.010
- Age, duration of lithium treatment, and serum lithium level were independently associated with urine osmolality.
- Older patients are less likely to report urinary and thirst symptoms.
- Subjective symptoms and serum sodium do not correlate with urine osmolality
- However, urine specific gravity may be a cost-efficient surrogate measure for urine osmolality.
- Clinicians should be carefully looking for nephrogenic diabetes insipidus, especially in patients who are older, have taken lithium for a longer time, and have higher lithium levels.
- These patients are at risk of lithium intoxication, falls, hypernatremia, and renal dysfunction.
- The key thing to learn from this study is that just asking for symptoms and looking at serum sodium are not adequate.
- When we suspect nephrogenic diabetes insipidus in a particular patient, then what we need to do is to get a combination of urine and serum osmolality, done simultaneously