GME Logo

User login

User menu

GME Research Review

GME Research Review is a monthly newsletter where internationally recognized experts select, summarize, and provide a clinical commentary on the latest published research in psychiatry. Each summary has been derived from the relevant article’s abstract and the clinical commentary has been provided by our expert.



Issue 28, July 2014: Guest Commentator—Chittaranjan Andrade, MD, National Institute of Mental Health and Neurosciences, Bangalore, India



Negative Outcome of Charles Bonnet Syndrome

Cox TM, Ffytche DH.
Br J Ophthalmol. 2014 [Epub ahead of print]


Objective: The Charles Bonnet syndrome (CBS) describes the occurrence of visual hallucinations in persons with visual (typically, macular) impairment. It is generally believed that these hallucinations are non-distressing, and that the syndrome is uncommon, transient, and self-limiting. Most studies on CBS are based on small samples. This study described CBS in what is probably the largest study to date.

: The researchers sent a questionnaire to 4,000 members of the Macular Society in the UK. There were 1,254 (31%) respondents.

Results: Visual hallucinations were reported by 492 (39%) of the 1,254 respondents, and are outlined in the Table. 


Percentage (%)


 -    Patterns

 -    Faces

 -    Objects

 -    Figures

 -    Animals







Length of time

-       Seconds

-       Minutes




-       Constant

-       Daily

-       Weekly

-       Monthly







>1 daily activity


Almost all (93%) of participants with hallucinations reported their experiences to one or more persons. Most participants were curious, intrigued, or startled at the time of onset of the hallucinations. Only 15% were afraid, and <5% were terrified. Distress diminished substantially with time, and indifference (39%) supervened as the commonest late outcome. Survival analysis suggested that 88% of the sample had hallucinations for >2 years, and 75% had at least occasional hallucinations for >5 years. Overall, one third of the sample considered the syndrome to have a negative effect on their lives. Less than 10% considered the symptoms to be pleasant. The rest of the respondents were neutral.

Conclusion: The Charles Bonnet syndrome may be more common, less benign, and longer lasting than is commonly believed. 

Clinical Commentary
Given the low (31%) response rate, it is likely that persons with CBS, especially those with more severe and long-lasting CBS, self-selected themselves into the sample. The percentages reported in this study could therefore be suspect. Furthermore, given that the data were collected through a questionnaire (or a phone call, in some cases) and not through face-to-face clinical assessments, the data could also be suspect. Curiously, the authors did not discuss either limitation.



Rate of Chiari I Malformation in Children of Mothers with Depression With and Without Prenatal SSRI Exposure

Knickmeyer RC, Meltzer-Brody S, Woolson S, et al
Neuropsychopharmacology 2014 May 19. doi: 10.1038/npp.2014.114. [Epub ahead of print]


Objective: Selective serotonin reuptake inhibitors (SSRIs) have previously been associated with rare but specific congenital malformations after first trimester exposure during pregnancy. This magnetic resonance imaging (MRI) study suggests that SSRI exposure may also increase the risk of the Chiari I malformation (CIM), a condition in which the cerebellar tonsils extend significantly below the foramen magnum. 

  The researchers conducted a two-cohort study: 

1. The first cohort included 33 children (20 male) whose mothers (mean age, 30 years) were confirmed to have been treated with an SSRI during pregnancy for a diagnosis of depression. Sertraline (25–200 mg/day) was the most common treatment (n=25); other children received fluoxetine (20–60 mg/day), citalopram (20–40 mg/day), or paroxetine (20 mg/day) (n=2 to 4, each). The duration of exposure ranged from 3 weeks to whole pregnancy. These children were matched with 66 children whose mothers did not experience depression during pregnancy and who had no SSRI exposure during pregnancy. 

2. The second cohort included 30 children (18 male) whose mothers (mean age, 29 years) had a history of depression but who had not received antidepressants during pregnancy. These children were matched with 60 children whose mothers did not have a history of depression and who had no antidepressant exposure during pregnancy. 

Matching was based on propensity scores using variables such as date of birth, gender, twin status, maternal ethnicity, maternal age, maternal education, household income, number of MRI scans (1 or 2), and age at scan. Both cohorts were formed from volunteer samples drawn from neuroimaging clinics in the USA. Women entered the study during pregnancy or shortly after delivery. Women with major medical illness, substance abuse, or exposure to (most groups of) non-SSRI psychotropics were excluded. Other exclusion criteria were premature delivery, major postnatal complications, and presence of major congenital malformations. The infants underwent MRI screening for CIM at age 1 and 2 years. The scans were read by two independent neuroradiologists who were blinded to the groupings. 


1. In the first cohort, CIM was significantly more common in SSRI-exposed children than in controls (18% vs 2%, respectively; OR, 10.32; 95% CI, 2.04-102.46). The finding remained significant with different cut-offs for the diagnosis of CIM. The odds of CIM were increased if there was a maternal family history of depression, if there was SSRI exposure at the time of conception, and if duration of SSRI exposure was greater. 

2. In the second cohort, there was no significant difference in the prevalence of CIM in children whose mothers were depressed during pregnancy relative to children whose mothers were not depressed (7% vs 5%, respectively; OR, 1.44; 95% CI, 0.23-7.85). 

No child with CIM had an additional MRI anomaly; 6 control children in the 2 cohorts had non-CIM anomalies detected on MRI. SSRI-exposed children differed from controls in having a shorter gestational age, smaller birth length, longer duration of stay in neonatal intensive care, and association with maternal hypertension during pregnancy. However, in secondary analyses in children unexposed to either SSRIs or maternal depression, none of these variables differentiated children with CIM (n=8) from those without CIM (n=427). No children in the two cohorts had CIM in association with exposure to non-SSRI (permitted) psychotropics during pregnancy. 

The authors note several important limitations to this study (Table). They note that replication of this study is required, especially in larger samples and with better control over potentially confounding variables. The findings do not as yet recommend changes in prescription practices.

Study Limitations

• Small cohort sample sizes, with 17 comparison children overlapping between the cohorts

• Maternal depression was likely more severe in first cohort

-       Explains the greater duration of antidepressant exposure

-       Likely also explains the increased risk of CIM

-       Severity of depression and factors r/t it (rather than SSRI exposure) may be responsible for CIM risk

• History of maternal depression sufficed for inclusion criterion in the second cohort.

-       Unclear how many women were depressed during pregnancy and truly exposing fetus to developmental stress

-       Findings from the second cohort were biased toward the null hypothesis.

• Sources of confound that were not controlled for included risky maternal behavior during pregnancy

-       Poor diet, smoking, alcohol intake, substance abuse, maternal obesity, gestational diabetes

-       Other, unmeasured, confounds may also have differentiated the SSRI-exposed and control children. 

Conclusion: Exposure to SSRIs during pregnancy is associated with a marked increase in the risk of MRI-diagnosed Chiari I malformation at 1 and 2 years of age. This risk is increased if SSRI exposure occurs at the time of conception, and if the duration of exposure is greater. This risk does not appear to be mediated by maternal depression. 

Clinical Commentary
It is interesting that a maternal family history of depression increased the SSRI-associated risk of CIM. Knickmeyer et al suggested that this may represent a double-hit situation where genetic (family history) and environmental (SSRI exposure) factors may interact to trigger an adverse outcome (CIM). In addition, underdevelopment of the posterior cranial fossa can crowd the developing hindbrain; CIM may be a consequence; the condition develops after birth. Whereas CIM is often asymptomatic, it may be associated with headache, disturbances of hearing or vision, lower cranial nerve disturbances, hydrocephalus, spinal cord syrinx, or other neurological conditions. A follow up of the CIM children in this study would educate readers about the clinical importance of the findings. Knickmeyer et al provided a competent explanation for and a responsible interpretation of their findings.



Preventive Effects of Ramelteon on Delirium: A Randomized Placebo-controlled Trial

Hatta K, Kishi Y, Wada K, et al
JAMA Psychiatry. 2014;71:397-403.


Objective: Medically ill inpatients, especially those who are elderly and those in intensive care units, are at increased risk of delirium. Medications such as antipsychotics1 but not cholinesterase inhibitors2 have been found effective in preventing delirium in elderly patients at risk. Encouraged by reports of benefit with melatonin 0.5 mg/night.3 This study examined whether the MT1 and MT2 melatonin receptor agonist ramelteon, an approved treatment for sleep-onset insomnia, offers benefits in this regard. 


  • The authors conducted a 5-center, randomized, placebo-controlled, inpatient trial conducted in Japan. Newly admitted nondelirious patients 65–89 years of age (n=67) were recruited from intensive care units and acute wards. All patients had serious medical problems but were sufficiently well to be able to take medications orally. 
  • Patients with an expected life expectancy (or hospital stay) of <48 hours were excluded. Patients were also excluded if they had conditions (e.g. liver disease, Lewy body disease) characterized by fluctuation in cognition, where differentiation from delirium could be difficult. Other exclusions included patients in alcohol or drug withdrawal, those receiving fluvoxamine (which interacts with ramelteon), and those receiving antipsychotics (which may by themselves prevent delirium). 
  • The patients were randomized to receive ramelteon (8 mg/night; n=33) or placebo (n=34) every night for 1 week. The mean age of the sample was 78 years. The sample was 40% male. The mean body-mass index was 22. Dementia was present in 19% of the sample. Stroke, infection, and fracture (n=47, in all) were the most common reasons for admission.
  • Raters, nursing staff, and treating staff but not patients and their families were blinded to the treatment assignment. Hydroxyzine (25 mg) was allowed as an emergency treatment for insomnia. Many patients left the study early (exact numbers not clearly stated in the paper) but no patient was lost to follow up and all patients were included in an intent-to-treat analysis. 

Delirium (DSM-IV) was significantly less common with ramelteon than with placebo (Table). The advantage for ramelteon remained even after adjustment for risk factors such as age, presence of dementia, and diagnosis of infection. Hydroxyzine was used by twice as many ramelteon patients; however, the prophylactic effect of ramelteon against delirium remained statistically significant even after patients receiving hydroxyzine were excluded from analysis. Survival analysis also showed a significantly later onset of delirium in ramelteon- vs placebo-treated patients. There were no differences between groups in any of the sleep outcomes assessed; these included initial, middle, and terminal insomnia; sleep quality; disturbances of the sleep-wake cycle; and sleep duration. There were no adverse events potentially attributable to the study medications. 





Presence of delirium



RR, 0.09; 95% CI, 0.01-0.69

Onset of delirium

6.9 days

5.7 days

95% CI, 6.8-7.1 vs 5.1-6.4

Use of hydroxyzine




Conclusions: Ramelteon (8 mg/night for 1 week) is associated with a substantial reduction in the risk of delirium in seriously medically ill inpatients. Onset of delirium is also approximately 1 day later in ramelteon-treated patients. 

Clinical Commentary
Ramelteon did not improve any sleep outcome; in fact, numerically (but not statistically) more ramelteon patients required emergency medication for insomnia. Therefore, improved sleep or mechanisms related thereto seems an unlikely explanation for the benefits observed with ramelteon. Perhaps modulation of melatonergic neurotransmission, or of hypothalamic activity, reduces the risk of delirium in elderly patients at risk. Melatonin and ramelteon are far better tolerated than antipsychotic drugs and are therefore more important as delirium preventive agents in elderly patients at risk. One wonders whether these drugs would also be effective in delirium prophylaxis in younger patients, post-surgical patients, and patients with alcohol or drug withdrawal states. 

1. Teslyar P, Stock VM, Wilk CM, et al. Prophylaxis with antipsychotic medication reduces the risk of post-operative delirium in elderly patients: a meta-analysis. Psychosomatics. 2013 [Epub ahead of print]
2. van Eijk MM, Roes KC, Honing ML, et al. Effect of rivastigmine as an adjunct to usual care with haloperidol on duration of delirium and mortality in critically ill patients: a multicentre, double-blind, placebo-controlled randomised trial. Lancet. 2010; 376: 1829-1837.
3. Al-Aama T, Brymer C, Gutmanis I, et al. Melatonin decreases delirium in elderly patients: a randomized, placebo-controlled trial. Int J Geriatr Psychiatry. 2011; 26: 687-694.



Increased Risk of Acute Cardiovascular Events After Partner Bereavement: A Matched Cohort Study

Carey IM, Shah SM, DeWilde S, et al.
JAMA Intern Med. 2014; 174: 598-605.

Objective: Death of a spouse is given the highest ranking (100 out of 100) on the Social Readjustment Rating Scale.1 Can the high emotional turmoil associated with partner bereavement result in adverse cardiovascular and other important health outcomes? If yes, what is the magnitude of the risk? This subject was addressed in this population-based matched cohort study conducted in the UK.


  • The sample was drawn from a primary care database containing information for 401 general practices. The sample identified persons aged 60 years and older, who were married and cohabiting, and whose spouse was aged not greater than 10 years older or younger. 
  • There were 30,477 subjects, 60–89 years of age, who had experienced partner bereavement. The mean age of the sample was 76 years. The sample was 64% female. These subjects were age-, sex-, date-, and practice-matched with 83,588 nonbereaved controls (maximum, 3 controls per case).
  • Important outcomes included myocardial infarction, stroke, non-myocardial infarction acute coronary syndrome, and pulmonary embolism. Incidence rate ratios (IRRs) for these outcomes were calculated after adjustment for age, smoking, a deprivation index, and history of chronic medical disease (diabetes mellitus, hypertension, coronary heart disease, cerebrovascular disease).

 The 30-day risk of fatal or non-fatal myocardial infarction or stroke was significantly higher in the bereaved group than in the nonbereaved group (0.16% vs 0.08%, respectively; IRR, 2.20; 95% CI, 1.52-3.15). This risk remained statistically significant but was substantially attenuated at 90 (0.38% vs 0.27%; IRR, 1.59) and 365 (1.28% vs 1.11%; IRR, 1.14) days. The risk of myocardial infarction and stroke was not significantly modified by age, sex, or preexisting cardiovascular disease at any time point (30, 90, and 365 days). The risk of myocardial infarction and stroke were each significantly elevated at 30 days (IRR, 2.14 and 2.40, respectively) and 90 days (IRR, 1.47 and 1.52, respectively), but not at 365 days (IRR, 1.20 and 1.02, respectively). The risks of non-myocardial infarction acute coronary syndrome and pulmonary embolism could not be assessed at 30 days because there were too few events. However, these risks were significantly increased at 90 but not at 365 days.

Clinical Commentary
The excess risk of fatal or non-fatal myocardial infarction or stroke was 0.08% in the first 30 days after partner bereavement; that is, 1 in 1,250. This statistic may be important at the population level but is perhaps not high at the individual level. The risk of acute myocardial infarction is elevated as early as a day after the experience of bereavement.2 A meta-analysis suggested that mortality after partner bereavement is higher in men than in women; however, this difference decreased with increasing age.3 Predictors of the risk of adverse outcomes after partner bereavement need to be identified. Whereas elegant biological mechanisms involving stress, cortisol, autonomic nervous system disturbances, changes in heart rate variability, and others may be validly proposed, it should also be remembered that bereaved persons, especially elderly bereaved persons, may neglect their own health care needs in matters such as compliance to current health instructions and attention to emerging medical symptoms.

1. Holmes TH, Rahe RH. The social readjustment rating scale. J Psychosom Res 1967;11:213-218.
2. Mostofsky E, Maclure M, Sherwood JB, et al. Risk of acute myocardial infarction after the death of a significant person in one's life: the Determinants of Myocardial Infarction Onset Study. Circulation. 2012;125:491-496.
3. Shor E, Roelfs DJ, Curreli M, et al. Widowhood and mortality: a meta-analysis and meta-regression. Demography. 2012;49:575-606.



Preterm Birth and Antidepressant Medication Use During Pregnancy: A Systematic Review and Meta-analysis

Huybrechts KF, Sanghani RS, Avorn J, et al
PLoS One. 2014; 9: e92778.  

Objective: Antidepressant drugs have been associated with an increased risk of preterm birth. However, the risk is small.1 For example, a recent meta-analysis of 15 studies reported that gestational age was 3.2 (95% CI, 1.8-4.5) days shorter in women who had received an antidepressant during pregnancy relative to controls who had not; there was little difference in analyses that were restricted to different sets of controls.2 Given that preterm birth is associated with increased neonatal morbidity and mortality,3 as well as with adverse long-term outcomes,4,5 these authors subjected the research in the field to a systematic review and meta-analysis.

Methods: The authors searched electronic databases, reference lists, and other sources, and identified 41 relevant observational cohort studies of women who had used antidepressant drugs during pregnancy. Of these, 21 were prospective, 16 were retrospective, and the rest recruited women both prospectively and retrospectively. Sample size varied very widely from 44 to 1,618,255 participants. In all but one study, preterm birth was defined as birth occurring before week 37 of gestation. There was medium to high heterogeneity in the different analyses, and so random effects models were applied.

Results: Both unadjusted and adjusted results were reported. This article presents only the adjusted outcomes.  Antidepressant use at any time during pregnancy was associated with an increased risk of preterm birth (17 studies; OR, 1.53; 95% CI, 1.40-1.66). First trimester antidepressant exposure did not increase the risk of preterm birth (8 studies; OR, 1.16; 95% CI, 0.92-1.45). Third trimester antidepressant exposure was associated with an increased risk of preterm birth (12 studies; OR, 1.96; 95% CI, 1.62-2.38). Preterm birth risk associated with antidepressant exposure was not eliminated after controlling for a diagnosis of maternal psychiatric illness (12 studies; OR, 1.61; 95% CI, 1.26-2.05). Sensitivity analysis suggested that unmeasured confounding would have to be strong to explain the findings of the meta-analysis. 

Conclusions: Third but not first trimester antidepressant exposure is associated with an increased risk of preterm birth.

Clinical Commentary
Most of the studies in this meta-analysis adjusted risks for various confounding variables. However, no amount of adjustment, or even propensity matching, can alter the fact that illness characteristics are likely to be more severe in depressed women who take an antidepressant during pregnancy relative to depressed women who do not. Therefore, illness behavior, or genetic concomitants thereof (and not necessarily antidepressant use), may explain adverse pregnancy outcomes. This may be why antidepressant use was no longer associated with preterm birth in one study after depression severity was controlled for.6 Examples of illness behavior that is often not controlled for in such studies include nutrition, smoking, and alcohol and substance use or abuse. If illness behaviors increase adverse pregnancy outcomes, then antidepressant use might actually reduce the adverse pregnancy outcomes to the extent that antidepressants are able to reduce the adverse illness behaviors. This study provided a good discussion on the limitations of the literature in the field. For example, women may fill a prescription for an antidepressant drug but may not take a drug; or, they may take the drug irregularly; or, they may take the drug during part of the pregnancy but not during the whole pregnancy. For these and other reasons, it is hard to correctly classify antidepressant use during pregnancy, and hence interpretation of data becomes difficult.

1. Andrade C. Antenatal exposure to selective serotonin reuptake inhibitors and duration of gestation. J Clin Psychiatry. 2013;74:e633-635.
2. Ross LE, Grigoriadis S, Mamisashvili L et al. Selected pregnancy and delivery outcomes after exposure to antidepressant medication: a systematic review and meta-analysis. JAMA Psychiatry. 2013;70:436-443.
3. Belizan JM, Hofmeyr J, Buekens P, Salaria N. Preterm birth, an unresolved issue. Reprod Health. 2013;10:58.
4. Dong Y, Chen SJ, Yu JL. A systematic review and meta-analysis of long-term development of early term infants. Neonatology. 2012;102:212-221.
5. Rai E. Long-term outcomes of prematurity in adolescence. Adolesc Med State Art Rev. 2013;24:330-340.
6. Oberlander TF1, Warburton W, Misri S, Aghajanian J, Hertzman C. Neonatal outcomes after prenatal exposure to selective serotonin reuptake inhibitor antidepressants and maternal depression using population-based linked health data. Arch Gen Psychiatry. 2006;63:898-906.



To contact GME, email us at [email protected]

GME does not provide medical advice. The website and articles are intended for informational purposes only. They are not a substitute for professional medical advice, diagnosis or treatment. Never ignore professional medical advice in seeking treatment because of something you have read on the GME Website. If you think you may have a medical emergency, immediately call your doctor or dial 911.

Log In

Join GME For Free

Become a GME subscriber and gain full access to our extensive library of 700+ psychiatric medical education videos, free CME webcasts, latest research updates, and more. To sample our content, watch the featured videos below.
Join Today!