The Journal: May 2009

 SSRI Use and Behavioral Disruption Among Children and Adolescents at Austin State Hospital

Texas Medicine Magazine Logo

The Journal - May 2009  

  Tex Med. 2009;105(5):e1-e5.  

By Emilie A. Becker, MD; M. Lynn Crismon, PharmD; Alan Shafer, PhD; and Jabeen Hayat, MD  

Dr Becker, psychiatrist, Texas Department of State Health Services, Austin, Texas; Dr Crismon, dean and James T. Doluiso Chair and Behrens Centennial Professor in Pharmacy, The University of Texas College of Pharmacy, Austin, Texas; Dr Shafer, research specialist, Texas Department of State Health Services, Austin, Texas; and Dr Hayat, child psychiatrist, Department of Children and Family Services, Fresno, Calif. Send correspondence to Emilie A. Becker, MD, Department of State Health Services, PO Box 149347, Austin, TX 78714.


This study examined the effect of selective serotonin reuptake inhibitors (SSRIs) on children and adolescents' disruptive behavior. Administrative data were obtained for 1 year of child and adolescent admissions at Austin State Hospital. Two measures of disruptive behavior were operationally defined as the use of mechanical restraints and emergency medication. Between-subjects analysis compared patients who did versus those who did not receive SSRIs. Within-subject analysis was conducted only for patients receiving SSRIs to examine their disruptive behavior during medication initiation. No differences were found between patients who did versus those who did not receive SSRIs. No differences were found among patients who received SSRIs during initial versus later phases of drug use. No significant relationship between SSRI use and increased agitation, hostility, or self-harm as manifested by the number of mechanical restraints or emergency medications at any time during hospitalization was found. 


Increasingly, physicians are prescribing selective serotonin reuptake inhibitors (SSRIs) for children and adolescents. Questions regarding the safety of these drugs have resulted in new investigations worldwide. The US Food and Drug Administration (FDA) now requires a black box warning on the label of all antidepressants to alert prescribers about the potential risk of suicidal ideation and self-injurious behavior, especially during the first few months of receiving treatment.

A recent meta-analysis 1 of antidepressant clinical trials in children and adolescents indicated an increased risk of suicidal ideation and suicide attempts for drugs versus placebos. However, the meta-analysis also showed that, relative to placebos, antidepressants were effective treatments for depression and anxiety among children. A recent task force 2 reviewed a wide group of current studies for efficacy and safety and found that only fluoxetine showed evidence for efficacy in treatment of depression but that all SSRIs showed some small increase in risk of suicidal ideation or attempts.

Several epidemiological studies have shown either no increased risk or a decreased risk of completed suicides with SSRIs. 3-6 An evidence-based, expert consensus process evaluated both the efficacy and safety of antidepressants and concluded that antidepressants are a reasonable treatment option but also recommended thorough diagnostic assessment and careful follow-up monitoring of treated youth with depression. 7 Thus, practicing physicians who need to treat depressed children and adolescents are faced with conflicting information but with a limited number of remaining therapeutic options.

That children can have psychiatric adverse reactions to medicines commonly given to adults is well established. For example, antihistamines typically cause sedation in adults, but activation and agitation are more frequent in children and adolescents. If SSRIs affect children differently than adults, then we must determine the mechanism by which this occurs. 8

One hypothesis is that drug-induced behavior disinhibition (DIBD) is expressed by irritability, aggression, angry outbursts, and agitation. A study by Carlson and Mick 9 measured DIBD as exhibited by an increase in time-outs over the previous week. Among several hundred hospitalized children, about 7.5% were found to be potentially disinhibited during their stay. After pervasive developmental disorder and attention deficit disorder, SSRIs were found to be the third largest predictor of disinhibition.

The use of restraints and sedation among hospitalized children could represent an alternative measure of potential behavioral disinhibition. Restraints are typically used as a last resort to prevent aggressive or self-destructive behavior. 10,11  Emergency medications are often used to treat acutely agitated patients who show potential for self-harm or aggression. 12 Therefore, incidents of restraints or emergency medication administration or both may represent severe loss of control.

This study retrospectively examined records of the Child Adolescent Psychiatric Services (CAPS) inpatient units at Austin State Hospital (ASH) with the hypothesis that SSRIs would increase agitation and hostility, particularly in the initial phase of SSRI administration, and would result in greater restraint frequency and use of emergency medication than in children not given SSRIs. 


Subject Inclusion Criteria  

The subject inclusion criteria were all admissions aged 4 to 8 years to ASH-CAPS from April 2006 to April 2007. Further, subject exclusion criteria for clients in this study included the diagnosis of mental retardation or borderline IQ, as these patients may respond differently to psychotropic medication than those without IQ limitation.

Sources of Data  

Patient demographic, diagnostic, and hospitalization information was gathered from the Texas Mental Health Mental Retardation Client Assignment and Registration database, prescription data from the ASH pharmacy, restraint information from Health Information Management at ASH, and the remainder from hospital medical records. This study was approved by the Texas Department of State Health Services Institutional Review Board for Protection of Human Subjects.


The primary treatment medications evaluated were SSRIs, including citalopram, escitalopram, fluoxetine, and sertraline. For the purpose of completeness, all other treatment medications prescribed were also obtained. Treatment medications were defined as multiday prescriptions and were obtained from hospital pharmacy records.

Emergency medications and mechanical restraints formed the two outcome variables. These variables were obtained from chart orders. Emergency medications were defined as any drugs ordered to be given stat or as soon as possible for the purpose of calming a patient. Emergency medications are not multiday prescriptions but usually are for single episodes. Data collected included the date of the first emergency medication, the total number of days on which any emergency medications were given, the number of emergency medications given during weeklong intervals, and the total number of emergency medications given during a hospitalization.

The single nonmedication variable used for clients in the study was mechanical restraints. The study did not distinguish between the types of restraints used. Collected information included the date or day of the first restraint, the total number of restraints, the number of days on which restraints were used, and the number of restraints used during weeklong intervals.


Three major analyses were conducted: an overall comparison for demographic information, diagnosis, and rates of mechanical restraints and emergency medications; of the patients who received only SSRIs, the number of mechanical restraints or emergency medications during weeks 1 through 5; and a comparison between the group that received SSRIs and the group that did not receive SSRIs during weeks 1 through 5. To make the comparison between the latter two groups roughly equivalent, only patients (about 95%) who received SSRIs in the first week of their hospital stay were retained for these analyses. The remainder (5% of patients) were excluded because they had medications that started much later during their length of stay. All analyses in this study treated each individual admission as an independent observation. 


The final sample consisted of 352 child and adolescent (aged 4 to 17 years) admissions to ASH between April 1, 2006 and April 30, 2007. Patient demographics are outlined in Table 1. Patients on SSRIs were significantly older at admission than the non-SSRI group and included significantly more females than the non-SSRI group. Traditionally under-represented minorities were well represented in the patient sample. No significant differences occurred between treatment groups in length of stay. Bipolar disorder was the most frequent primary diagnosis in the patient sample. The population was highly comorbid, with 40% having two diagnoses; 14%, three diagnoses; and 3.5%, four or more diagnoses. The SSRI group had significantly more patients diagnosed with major depression and posttraumatic stress disorder than did the non-SSRI group. Diagnoses are listed in Table 1  [ PDF ].

Psychotropic Medications Prescribed  

A total of 127 (36%) patients had an SSRI prescribed during their hospitalization. One hundred twenty (95%) patients in the study received only a single SSRI. Only 7 patients in the SSRI group received two SSRIs, concomitantly or sequentially. The percentage totals are greater than 100% because of the 7 patients receiving two SSRIs.

No one prescribed paroxetine, and fluvoxamine was given only once; thus, these SSRIs were excluded from the analyses. The average SSRI medication order lasted for approximately 2 weeks (X = 14 days ± 15 days). These SSRI days covered slightly more than half of these patients' hospital stays, with specific SSRI dates ranging from 51% to 57% of patients' complete hospital stays.

Second-generation antipsychotics (SGAs) were the most commonly prescribed medications across both SSRI and non-SSRI groups, with 76% of all patients receiving SGAs. Among the specific SGAs prescribed, a significant difference was observed between the SSRI group and the non-SSRI group: the SSRI group received less risperidone than did the non-SSRI group. No other significant differences occurred between the SSRI and non-SSRI groups in the frequency of SGAs prescribed. In addition to SGAs and SSRIs, the most frequently prescribed drugs to patients in this study were divalproex, trazodone, and oxcarbazepine. Three significant differences emerged among other drug orders: the SSRI group had significantly more orders for trazodone than did the non-SSRI group; the SSRI group had significantly fewer orders for divalproex than did the non-SSRI group; and the SSRI group had significantly fewer orders for bupropion than did the non-SSRI group. The mean number of medications ordered (for all the drugs described above including SGAs and SSRIs) for the 328 patients receiving these drugs was 2.67 (SD = 1.30). Eleven medications accounted for about 80% of all the emergency medications. Table 2 [ PDF ] details the prescribing information.

Mechanical Restraints and Emergency Medications  

Table 3 [ PDF ] summarizes the data for mechanical restraints and emergency medications for the entire sample of 352 clients, with additional columns for the SSRI group and the non-SSRI group.

A total of 245 mechanical restraints (representing individual incidents or episodes of mechanical restraint) occurred for all patients in the study. No significant differences were observed between the two groups for any of the mechanical restraint variables. Overall, 66 (19%) clients were restrained during their stay. The mechanical restraints occurred, on average, 3.71 times (SD = 4.62).

In sum, 116 (33%) patients received some form of emergency medications during their stay, resulting in 369 individual emergency medications prescribed (representing individual episodes of emergency medical intervention with an emergency medication). For the 116 patients who received some form of emergency medication, the average given was 3.18 times (SD = 3.04). No significant differences appeared between the two groups for any of the emergency medications.

The total number of patients (either of SSRI patients as measured by their first day of medication or of all patients as measured from admission) who had mechanical restraints, emergency medications, or both, as measured from their starting point, show a distinctly higher number of restraints or medications per day in the first week compared with the remainder of their stay. However, this increased number is due in part to the larger number of patients in the hospital during the first week compared with the lesser number of patients at later times. The cumulative percentage of patients discharged during the first 5 weeks of their stay increases rapidly. By the end of week 1, 25% of the patients had been discharged; by the end of week 2, 50% had been discharged; week 3 resulted in the discharge of 66% of the patients; the end of week 4 showed 75% had been discharged; and, by the end of week 5, 80% had been discharged. Because of the relatively large numbers of patients discharged as the length of stay increases, two additional analyses examined the number of mechanical restraints and emergency medications while holding the number of patients in each observation period equal. Patients retained for these analyses had to have either a minimum of 1 week on SSRIs or a minimum length of stay of 1 week, respectively.

Table 4 [ PDF ] summarizes the results of mechanical restraints and emergency medications for the group prescribed SSRIs for a 5-week period. No significant difference was found in the total number of restraints, emergency medications, or combination of the two between the initial SSRI period and any subsequent time period during the study. For example, no differences occurred in the use of restraints or emergency medications or both during the first week of the study versus the second week of the study for those patients who were present for the entire 2 weeks.

Table 5 [ PDF ] summarizes the analysis comparing the group prescribed SSRIs and those who were not prescribed them for the total number of mechanical restraints, emergency medications, and the combination of the two. No significant differences were apparent in the total number of restraints, emergency medications, or the combination of the two at the first week or at any of the subsequent weeks during the study. 


Overall, the study results showed no significant relationship between SSRI use and increased agitation, hostility, or self-harm as manifested by the number of mechanical restraints or emergency medications at any time during hospitalization.

The most important limitation to this study is that it is not a prospective randomized study, and differences between the two groups not detected by the data collected may have affected the results. For example, one of the groups could have been more severely ill than the other, which would have created potential bias with regard to the results. Other potential limitations were present in the study: patient compliance, drug dosage, and age were not factored into outcome. For example, whether a medication was ordered did not mean that a patient agreed to take it. Further, low-dose SSRIs may have less effect overall on patients than those given a higher dose. The study did not address any dose changes that might have been ordered as a result of a restraint or administration of an emergency medication. Differential effects of age were not calculated - possibly the effects of SSRIs might vary with developmental age. The mean length of SSRI treatment that most patients received may not have been long enough to observe a therapeutic effect.

The rate of restraint and emergency medication use may have masked any differential effects of medications. In other words, if 1 in 5 patients were restrained, the medicines administered may have been irrelevant. Further, if 1 in 3 received emergency medicines, additional medicines given routinely may not have had substantial effect.

The use of SGAs in children and adolescents has increased markedly in recent years. Three-quarters of the patients received treatment with SGAs, either as mood stabilizers or medications for agitation and aggression. The use of SGAs may have dampened any potential agitating effects of the SSRIs. The prescribing practices in this study reflect those practices elsewhere, whereby complicated patients with multiple diagnoses are often treated with multiple medications. Sorting out any potential behavioral toxicity with SSRIs is best observed in monotherapeutic trials, not in the midst of polypharmacy.

This study cannot answer solely the pertinent question others have investigated: are SSRIs safe for use in children and adolescents? However, this study reveals that children and adolescents for whom these drugs were prescribed showed no more agitation, aggression, or hostility as measured by restraint and emergency medication orders than did children and adolescents on the same unit during the same time frame who were given other medications.


  1. Bridge JA, Iyengar S, Salary CB, et al. Clinical response and risk for reported suicidal ideation and suicide attempts in pediatric antidepressant treatment: a meta-analysis of randomized controlled trials. JAMA . 2007;297(15):1683-1696.
  2. Mann JJ, Emslie G, Baldessarini RJ, et al. ACNP Task Force report on SSRIs and suicidal behavior in youth. Neuropsychopharmacology. 2006;31(3):473-492.
  3. Gibbons RD, Hur K, Bhaumik DK, Mann JJ. The relationship between antidepressant prescription rates and rate of early adolescent suicide. Am J Psychiatry . 2006;163(11):1898-1904.
  4. Simon GE, Savarino J, Operskalski B, Wang PS. Suicide risk during antidepressant treatment. Am J Psychiatry . 2006;163(1):41-47.
  5. Tiionen J, Lönnqvist J, Wahlbeck K, et al. Antidepressants and the risk of suicide, attempted suicide, and overall mortality in a nationwide cohort. Arch Gen Psychiatry. 2006;63(12):1358-1367.
  6. Simon GE, Savarino J. Suicide attempts among patients starting depression treatment with medications or psychotherapy. Am J Psychiatry. 2007;164(7):1029-1034.
  7. Hughes CW, Emslie GJ, Crismon ML, et al. The Texas Children's Medication Algorithm Project: update from the Texas Consensus Conference Panel on Medication Treatment of Childhood Major Depressive Disorder. J Am Acad Child Adolesc Psychiatry. 2007;46(16):667-686.
  8. Carlson G.  Medication-induced activation in children and adolescents . Psychiatric Times. 2005;22(10).
  9. Carlson GA, Mick E, Drug-induced disinhibition in psychiatrically hospitalized children. J Child Adolesc Psychopharmacol.  2003;13(12):153-163.
  10. Delaney KR. Evidence base for practice: reduction of restraint and seclusion use during child and adolescent psychiatric inpatient treatment. Worldviews Evid Based Nurs . 2006;3(1):19-30.
  11. Nelstrop L, Chandler-Oatts J, Bingley W, et al. A systematic review of the safety and effectiveness of restraint and seclusion as interventions for the short-term management of violence in adult psychiatric inpatient settings and emergency departments. Worldviews Evid Based Nurs . 2006;3(1):8-18.
  12. Marder SR. A review of agitation in mental illness: treatment guidelines and current therapies. J Clin Psychiatry . 2006;67(suppl 10):13-21.

May 2009 Texas Medicine Contents
Texas Medicine Main Page