Letters COMMENT & RESPONSE
Ramelteon and the Prevention of Delirium To the Editor In their article, Hatta et al1 demonstrated that ramelteon may prevent the development of delirium. They suggested that this prophylactic effect might be mediated by the action of ramelteon on melatonin receptors 1 and 2 (MT1 and MT2) directly regulating activity of the suprachiasmatic nucleus. Two other potential mechanisms should be considered. One hypothesis of delirium pathophysiology is that pathologically sustained high levels of cortisol occurring with acute stress can precipitate and/or sustain delirium.2 The pars tuberalis is the rostral part of the anterior pituitary.3 Most pituitary hormones, including adrenocorticotropic hormone (ACTH) and follicle-stimulating hormone, are produced in the pars distalis of the anterior pituitary. Endocannabinoid signaling molecules have been detected in the human pars tuberalis.3 Cannabinoid receptor 1 has been detected in the human pars distalis, localized mainly on ACTH and folliclestimulating hormone cells.3 Mammalian pars tuberalis contains high levels of MT1 receptors.3 Endocannabinoid signaling is involved in the regulation of the hypothalamic-pituitaryadrenal axis, and low endocannabinoid levels have been associated with an increased risk for delirium.4 It is conceivable that ramelteon MT1 receptor binding might regulate pituitary endocannabinoid signaling and cannabinoid receptor 1–mediated ACTH release, thereby attenuating an aberrant hypothalamic-pituitary-adrenal axis stress response that could trigger delirium in a group of high-risk patients.2 Vasopressin stimulates the secretion of ACTH and facilitates the effect of corticotropin-releasing hormone on ACTH secretion. Therefore, vasopressin modulates the release of corticosteroids from the adrenal gland in response to stress. Delirium is associated with elevated cerebrospinal fluid levels of serotonin metabolites. Central serotonin 5-HT2B receptors regulate serotonin release via the serotonin transporter, and released vasopressin has been shown to activate 5-HT2B receptors indirectly.5 The major active metabolite of ramelteon is M-II, which has approximately one-tenth and one-fifth the binding affinity of ramelteon for human MT1 and MT2 receptors, respectively. Unlike ramelteon, M-II has weak affinity for the 5-HT2B receptor. Although the MT1 and MT2 receptor potency of M-II is lower than ramelteon, M-II circulates at substantially higher concentrations than ramelteon (producing a 20-fold to 100-fold greater mean systemic exposure compared with ramelteon). It is possible that M-II might interact with 5-HT2B receptor signaling, influencing central serotonergic transmission and/or the effects of vasopressin in such a way that reduces the risk for developing delirium in highrisk patients. jamapsychiatry.com
If the effectiveness of ramelteon for preventing delirium is confirmed, these potential mechanisms should be investigated in subsequent research. Robert H. Howland, MD Author Affiliation: Western Psychiatric Institute and Clinic, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania. Corresponding Author: Robert H. Howland, MD, Western Psychiatric Institute and Clinic, University of Pittsburgh Medical Center, 3811 O’Hara St, Pittsburgh PA 15213 ([email protected]). Conflict of Interest Disclosures: None reported. 1. Hatta K, Kishi Y, Wada K, et al; DELIRIA-J Group. Preventive effects of ramelteon on delirium: a randomized placebo-controlled trial. JAMA Psychiatry. 2014;71(4):397-403. 2. Maclullich AMJ, Ferguson KJ, Miller T, de Rooij SE, Cunningham C. Unravelling the pathophysiology of delirium: a focus on the role of aberrant stress responses. J Psychosom Res. 2008;65(3):229-238. 3. Yasuo S, Korf HW. The hypophysial pars tuberalis transduces photoperiodic signals via multiple pathways and messenger molecules. Gen Comp Endocrinol. 2011;172(1):15-22. 4. Hauer D, Weis F, Campolongo P, et al. Glucocorticoid-endocannabinoid interaction in cardiac surgical patients: relationship to early cognitive dysfunction and late depression. Rev Neurosci. 2012;23(5-6):681-690. 5. Knowles ID, Ramage AG. Evidence for a role for central 5-HT2B as well as 5-HT2A receptors in cardiovascular regulation in anaesthetized rats. Br J Pharmacol. 1999;128(3):530-542.
Family-Based Replication Study of Schizophrenia Genes To the Editor In response to our family-based replication study of schizophrenia genes,1 Lazzeroni2 expressed concerns regarding the interpretation of the replication results and argued that simpler studies may ultimately yield faster progress in psychiatric genetics. We presented 4 arguments that speak to the replication status of our results: (1) the highly significant enrichment of singlenucleotide polymorphisms (SNPs) with small P values in the replication; (2) an 89% to 93% agreement between the direction of effects in our case-control genome-wide association study meta-analysis used to select SNPs and the SNPs that replicated; (3) several findings that replicated across ancestral groups; and (4) replicating SNPs were organized in meaningful biological patterns. Lazzeroni2 focused mainly on the second argument. Based on a simulation study involving 800 unrelated cases and control individuals, she suggested that odds ratios greater than 1 may be enriched among the top results (even if none of the SNPs have an effect) and that this artifact could potentially have caused the high agreement in our study. On behalf of the Family-Based Replication Consortium, we have 2 responses. 1. The simulations performed have no direct relation to our study. We performed logistic regression analyses in a sample JAMA Psychiatry October 2014 Volume 71, Number 10
Copyright 2014 American Medical Association. All rights reserved.
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of St. Andrews Library User on 06/04/2015
1195
Letters
of 11 185 schizophrenia cases plus 10 768 control individuals to select more than 9000 SNPs that were replicated in 6298 individuals from 1811 nuclear families. The agreement measure we reported involved the concordance between the directions of effects in the genomewide association study meta-analysis and the markers that replicated in our family-based replication study. Because of the obvious differences (eg, test performed, design, types of samples, and sample sizes) between our investigation and the simulation performed by Lazzeroni,2 we think that the extrapolation of findings from her simulation study is speculative. 2. The simulation study assumed a disease prevalence of 5% in the population, a minor allele frequency of 0.05, and that study participants were randomly selected from the population. However, the median lifetime prevalence of schizophrenia is less than 1%,3 a minor allele frequency of 0.05 is often the lower bound after quality control, and studies are more typically designed to have an approximately equal number of cases and control individuals (as we did for our study). When we replicated the simulations assuming an equal number of cases and control individuals, we did not observe the reported bias anymore and even under a random sampling scheme, it became very small, assuming minor allele frequency of 0.5. Thus, it is unclear how relevant the author’s basic observation is. The use of a familial data set for the replication of findings from genome-wide association study meta-analyses did add complexity. However, we disagree with Lazzeroni that using statistical tools, such as a sign test, was overly complex. In genetics, most studies involve a case-control design
1196
that may introduce artificial differences. Although most biases can be controlled for, subtle artifacts can potentially remain (eg, population and ascertainment differences). When the sample sizes become increasingly larger, these artifacts can result in false discoveries. Because family-based studies are not susceptible to the same biases, they provide an opportunity to validate case-control findings, which, we believe, outweighs the disadvantage of the added complexity. Karolina A. Aberg, PhD Edwin J. C. G. van den Oord, PhD Author Affiliations: Center for Biomarker Research and Personalized Medicine, Virginia Commonwealth University, Richmond. Corresponding Author: Edwin J. C. G. van den Oord, PhD, Center for Biomarker Research and Personalized Medicine, Virginia Commonwealth University, 1112 E Clay St, McGuire Hall, Room 209B, Richmond, VA 23298-0581 (ejvandenoord @vcu.edu). Conflict of Interest Disclosures: None reported. Additional Information: The Family-Based Replication Consortium consists of Karolina A. Aberg, PhD; Youfang Liu, PhD; Jozsef Bukszar, PhD; Joseph L. McClay, PhD; Amit N. Khachane, PhD; Ole A. Andreassen, PhD; Douglas Blackwood, PhD; Aiden Corvin, PhD; Srdjan Djurovic, PhD; Hugh Gurling, PhD; Roel Ophoff, PhD; Carlos N. Pato, MD; Michele T. Pato, MD; Brien Riley, PhD; Todd Webb, PhD; Kenneth Kendler, MD; Mick O’Donovan, PhD; Nick Craddock, PhD; George Kirov, PhD; Mike Owen, PhD; Dan Rujescu, PhD; David St Clair, PhD; Thomas Werge, PhD; Christina M. Hultman, PhD; Lynn E. Delisi, MD; Patrick Sullivan, MD; and Edwin J. van den Oord, PhD. 1. Aberg KA, Liu Y, Bukszár J, et al. A comprehensive family-based replication study of schizophrenia genes. JAMA Psychiatry. 2013;70(6):573-581. 2. Lazzeroni LC. Evaluating the evidence of replication for genetic associations with schizophrenia. JAMA Psychiatry. 2014;71(1):94-95. 3. Murray CJL, Lopez AD. The Global Burden of Disease. Geneva, Switzerland: World Health Organization; 1996.
JAMA Psychiatry October 2014 Volume 71, Number 10
Copyright 2014 American Medical Association. All rights reserved.
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of St. Andrews Library User on 06/04/2015
jamapsychiatry.com
Ramelteon and the Prevention of Delirium To the Editor In their article, Hatta et al1 demonstrated that ramelteon may prevent the development of delirium. They suggested that this prophylactic effect might be mediated by the action of ramelteon on melatonin receptors 1 and 2 (MT1 and MT2) directly regulating activity of the suprachiasmatic nucleus. Two other potential mechanisms should be considered. One hypothesis of delirium pathophysiology is that pathologically sustained high levels of cortisol occurring with acute stress can precipitate and/or sustain delirium.2 The pars tuberalis is the rostral part of the anterior pituitary.3 Most pituitary hormones, including adrenocorticotropic hormone (ACTH) and follicle-stimulating hormone, are produced in the pars distalis of the anterior pituitary. Endocannabinoid signaling molecules have been detected in the human pars tuberalis.3 Cannabinoid receptor 1 has been detected in the human pars distalis, localized mainly on ACTH and folliclestimulating hormone cells.3 Mammalian pars tuberalis contains high levels of MT1 receptors.3 Endocannabinoid signaling is involved in the regulation of the hypothalamic-pituitaryadrenal axis, and low endocannabinoid levels have been associated with an increased risk for delirium.4 It is conceivable that ramelteon MT1 receptor binding might regulate pituitary endocannabinoid signaling and cannabinoid receptor 1–mediated ACTH release, thereby attenuating an aberrant hypothalamic-pituitary-adrenal axis stress response that could trigger delirium in a group of high-risk patients.2 Vasopressin stimulates the secretion of ACTH and facilitates the effect of corticotropin-releasing hormone on ACTH secretion. Therefore, vasopressin modulates the release of corticosteroids from the adrenal gland in response to stress. Delirium is associated with elevated cerebrospinal fluid levels of serotonin metabolites. Central serotonin 5-HT2B receptors regulate serotonin release via the serotonin transporter, and released vasopressin has been shown to activate 5-HT2B receptors indirectly.5 The major active metabolite of ramelteon is M-II, which has approximately one-tenth and one-fifth the binding affinity of ramelteon for human MT1 and MT2 receptors, respectively. Unlike ramelteon, M-II has weak affinity for the 5-HT2B receptor. Although the MT1 and MT2 receptor potency of M-II is lower than ramelteon, M-II circulates at substantially higher concentrations than ramelteon (producing a 20-fold to 100-fold greater mean systemic exposure compared with ramelteon). It is possible that M-II might interact with 5-HT2B receptor signaling, influencing central serotonergic transmission and/or the effects of vasopressin in such a way that reduces the risk for developing delirium in highrisk patients. jamapsychiatry.com
If the effectiveness of ramelteon for preventing delirium is confirmed, these potential mechanisms should be investigated in subsequent research. Robert H. Howland, MD Author Affiliation: Western Psychiatric Institute and Clinic, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania. Corresponding Author: Robert H. Howland, MD, Western Psychiatric Institute and Clinic, University of Pittsburgh Medical Center, 3811 O’Hara St, Pittsburgh PA 15213 ([email protected]). Conflict of Interest Disclosures: None reported. 1. Hatta K, Kishi Y, Wada K, et al; DELIRIA-J Group. Preventive effects of ramelteon on delirium: a randomized placebo-controlled trial. JAMA Psychiatry. 2014;71(4):397-403. 2. Maclullich AMJ, Ferguson KJ, Miller T, de Rooij SE, Cunningham C. Unravelling the pathophysiology of delirium: a focus on the role of aberrant stress responses. J Psychosom Res. 2008;65(3):229-238. 3. Yasuo S, Korf HW. The hypophysial pars tuberalis transduces photoperiodic signals via multiple pathways and messenger molecules. Gen Comp Endocrinol. 2011;172(1):15-22. 4. Hauer D, Weis F, Campolongo P, et al. Glucocorticoid-endocannabinoid interaction in cardiac surgical patients: relationship to early cognitive dysfunction and late depression. Rev Neurosci. 2012;23(5-6):681-690. 5. Knowles ID, Ramage AG. Evidence for a role for central 5-HT2B as well as 5-HT2A receptors in cardiovascular regulation in anaesthetized rats. Br J Pharmacol. 1999;128(3):530-542.
Family-Based Replication Study of Schizophrenia Genes To the Editor In response to our family-based replication study of schizophrenia genes,1 Lazzeroni2 expressed concerns regarding the interpretation of the replication results and argued that simpler studies may ultimately yield faster progress in psychiatric genetics. We presented 4 arguments that speak to the replication status of our results: (1) the highly significant enrichment of singlenucleotide polymorphisms (SNPs) with small P values in the replication; (2) an 89% to 93% agreement between the direction of effects in our case-control genome-wide association study meta-analysis used to select SNPs and the SNPs that replicated; (3) several findings that replicated across ancestral groups; and (4) replicating SNPs were organized in meaningful biological patterns. Lazzeroni2 focused mainly on the second argument. Based on a simulation study involving 800 unrelated cases and control individuals, she suggested that odds ratios greater than 1 may be enriched among the top results (even if none of the SNPs have an effect) and that this artifact could potentially have caused the high agreement in our study. On behalf of the Family-Based Replication Consortium, we have 2 responses. 1. The simulations performed have no direct relation to our study. We performed logistic regression analyses in a sample JAMA Psychiatry October 2014 Volume 71, Number 10
Copyright 2014 American Medical Association. All rights reserved.
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of St. Andrews Library User on 06/04/2015
1195
Letters
of 11 185 schizophrenia cases plus 10 768 control individuals to select more than 9000 SNPs that were replicated in 6298 individuals from 1811 nuclear families. The agreement measure we reported involved the concordance between the directions of effects in the genomewide association study meta-analysis and the markers that replicated in our family-based replication study. Because of the obvious differences (eg, test performed, design, types of samples, and sample sizes) between our investigation and the simulation performed by Lazzeroni,2 we think that the extrapolation of findings from her simulation study is speculative. 2. The simulation study assumed a disease prevalence of 5% in the population, a minor allele frequency of 0.05, and that study participants were randomly selected from the population. However, the median lifetime prevalence of schizophrenia is less than 1%,3 a minor allele frequency of 0.05 is often the lower bound after quality control, and studies are more typically designed to have an approximately equal number of cases and control individuals (as we did for our study). When we replicated the simulations assuming an equal number of cases and control individuals, we did not observe the reported bias anymore and even under a random sampling scheme, it became very small, assuming minor allele frequency of 0.5. Thus, it is unclear how relevant the author’s basic observation is. The use of a familial data set for the replication of findings from genome-wide association study meta-analyses did add complexity. However, we disagree with Lazzeroni that using statistical tools, such as a sign test, was overly complex. In genetics, most studies involve a case-control design
1196
that may introduce artificial differences. Although most biases can be controlled for, subtle artifacts can potentially remain (eg, population and ascertainment differences). When the sample sizes become increasingly larger, these artifacts can result in false discoveries. Because family-based studies are not susceptible to the same biases, they provide an opportunity to validate case-control findings, which, we believe, outweighs the disadvantage of the added complexity. Karolina A. Aberg, PhD Edwin J. C. G. van den Oord, PhD Author Affiliations: Center for Biomarker Research and Personalized Medicine, Virginia Commonwealth University, Richmond. Corresponding Author: Edwin J. C. G. van den Oord, PhD, Center for Biomarker Research and Personalized Medicine, Virginia Commonwealth University, 1112 E Clay St, McGuire Hall, Room 209B, Richmond, VA 23298-0581 (ejvandenoord @vcu.edu). Conflict of Interest Disclosures: None reported. Additional Information: The Family-Based Replication Consortium consists of Karolina A. Aberg, PhD; Youfang Liu, PhD; Jozsef Bukszar, PhD; Joseph L. McClay, PhD; Amit N. Khachane, PhD; Ole A. Andreassen, PhD; Douglas Blackwood, PhD; Aiden Corvin, PhD; Srdjan Djurovic, PhD; Hugh Gurling, PhD; Roel Ophoff, PhD; Carlos N. Pato, MD; Michele T. Pato, MD; Brien Riley, PhD; Todd Webb, PhD; Kenneth Kendler, MD; Mick O’Donovan, PhD; Nick Craddock, PhD; George Kirov, PhD; Mike Owen, PhD; Dan Rujescu, PhD; David St Clair, PhD; Thomas Werge, PhD; Christina M. Hultman, PhD; Lynn E. Delisi, MD; Patrick Sullivan, MD; and Edwin J. van den Oord, PhD. 1. Aberg KA, Liu Y, Bukszár J, et al. A comprehensive family-based replication study of schizophrenia genes. JAMA Psychiatry. 2013;70(6):573-581. 2. Lazzeroni LC. Evaluating the evidence of replication for genetic associations with schizophrenia. JAMA Psychiatry. 2014;71(1):94-95. 3. Murray CJL, Lopez AD. The Global Burden of Disease. Geneva, Switzerland: World Health Organization; 1996.
JAMA Psychiatry October 2014 Volume 71, Number 10
Copyright 2014 American Medical Association. All rights reserved.
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of St. Andrews Library User on 06/04/2015
jamapsychiatry.com
