|
 |
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2007 American Public Health Association
DOI: 10.2105/AJPH.2006.087262
RESEARCH AND PRACTICE |
Paul A. Kurdyak is with the Centre for Addiction and Mental Health, Toronto, Ontario, and the Insitute for Clinical Evaluative Sciences, Sunnybrook Health Science Centre, Toronto. David N. Juurlink and Muhammad M. Mamdani are with the Institute for Clinical Evaluative Sciences, Sunnybrook Health Science Centre, Toronto.
Correspondence: Requests for reprints should be made to Dr. Paul A. Kurdyak, Centre for Addiction and Mental Health, 33 Russell Street T311, Toronto, Ontario, Canada, M5S 2S1 (e-mail: paul_kurdyak{at}camh.net).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Objectives. We studied whether 5 regulatory agency advisories concerning the possible increased risk of suicidal behavior during antidepressant therapy had an effect on antidepressant prescription trends in Ontario.
Methods. We conducted a population-based, time-series analysis of new monthly antidepressant prescriptions dispensed by the Ontario Drug Benefits program in 3 age groups (younger than 20 years, 20–65 years old, and older than 65 years) over a 7-year period (April 1998 to March 2005). The impact of five advisories about the possible risk of suicide during antidepressant therapy was also analyzed.
Results. The number of new prescriptions for selective serotonin reuptake inhibitors as a group did not change after any antidepressant warning in any age group. However, the rate of new paroxetine prescriptions in patients younger than 20 years declined by 54% immediately after the first warning for paroxetine was issued in the United Kingdom in June 2003. That same warning had no effect on new paroxetine prescriptions in the other age categories.
Conclusion. The warning about paroxetine use in depressed patients younger than 18 years that was issued in the United Kingdom led to a significant decrease in new paroxetine prescriptions for young patients in this country. By contrast, warnings in North America did not influence new antidepressant prescription rates in any patient group.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
The possibility that selective serotonin re-uptake inhibitors (SSRIs) might occasionally increase the risk of suicide has been debated for more than a decade.1 Studies have provided conflicting evidence about the risk of suicide related to SSRI use.2–7 On June 10, 2003, the United Kingdom (UK) Committee on Safety of Medicines published a report advising physicians and patients not to use paroxetine for depression in patients younger than 18 years.8 On October 27, 2003, the US Food and Drug Administration (FDA) issued a public health advisory that emphasized all newer antidepressants should be used with caution in pediatric patients.9 A year later (October 15, 2004), the FDA issued another public health advisory that included a "black box" warning and an expanded statement about a possible increased risk of suicidal ideation and behavior in children and adolescents who were being treated with all antidepressants.10 These warnings were issued in response to reports of increased suicidal ideation in children who had been prescribed paroxetine.11
These warnings were directed toward a pediatric population. On March 22, 2004, the FDA issued a public health advisory about the need to closely monitor patients of all ages for worsening depression or suicidality after the initiation of antidepressant therapy or changes in dose.12 On June 3, 2004, Health Canada issued a similar advisory.13
The timing and specificity of the 5 warnings differed. The paroxetine warning issued in the UK on June 10, 2003, was the first and most specific warning. We studied the effect of these 5 warnings on new antidepressant prescription trends in Ontario. We particularly looked at new antidepressant prescriptions in patients younger than 20 years old. We hypothesized that the warnings would result in an immediate reduction in the number of people who were prescribed antidepressant therapy.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
This study was a population-based, time-series analysis of new monthly antidepressant prescriptions in Ontario, Canada, over a 7-year period (April 1998 to March 2005).
We examined the computerized prescription records of the Ontario Drug Benefit (ODB) program, which records prescription information for all Ontario residents aged 65 years and older. The ODB also provides prescription coverage for eligible recipients younger than 65 years old. Eligibility for recipients younger than 65 years old is on the basis of the presence of disability, financial need, or both. Information on prescriptions for individuals younger than 65 years old is also captured in the computerized prescription records of the ODB. During the study period, there was a 40% decline in enrollment for ODB beneficiaries younger than 20 years old and an 11% decline in enrollment for ODB beneficiaries between 10 and 65 years old. Both declines reflect stricter eligibility criteria for financial aid.
For the analysis, we used prescription rates aggregated monthly and stratified the sample on the basis of the following age categories: younger than 20 years, 20 to 65 years old, and 66 years and older. We speculated that antidepressant warnings would most likely affect new prescriptions at the time the warning was published. To qualify as a new antidepressant prescription, each prescription had to be preceded by a 12-month period during which no other antidepressant prescription was issued to that patient.
We used time-series analysis to examine patterns of new antidepressant prescription rates on a monthly basis during the study period. Time-series analysis is a collection of techniques used for modeling autocorrelation in temporally sequenced data, as we have previously described.14,15 We examined trends in antidepressant prescription rates, defined as the rate of new prescriptions per 10 000 Ontario residents every month, and effects related to the publication of both the UK paroxetine warning (June 10, 2003), the 3 FDA Public Health Advisories (October 27, 2003; March 22, 2004; and October 15, 2004) and the Health Canada Advisory (June 3, 2004). The effect of the advisories was assessed by using interventional seasonal autoregressive integrated moving-average models, a particular type of time series analysis, and the use of a ramp function, which describes a point in time in the data series at which the linear trend suddenly changes slope.16 We assessed the effect of all 6 advisories on all prescriptions of SSRIs (fluoxetine, fluvoxamine, paroxetine, sertraline, and citalopram) aggregated as a group of medications and by SSRIs individually. Escitalopram, a SSRI included in the warnings, was not included in the analysis because it was not in the ODB formulary during the study period.
The autocorrelation, partial autocorrelation, and inverse autocorrelation functions were assessed for model parameter appropriateness. Stationarity (the degree to which there is homogeneity for the mean and variance over the observation period) was assessed using autocorrelation functions. The presence of white noise (an uncorrelated sequence of random variables with constant variance) was assessed by examining the autocorrelations at various lags using the Ljung–Box 
2 statistic.17 All P values were 2-sided, and analyses were conducted using SAS, version 8.2 (SAS Institute Inc., Cary, NC).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
During the study period, the ODB enrollment for individuals aged younger than 20 years declined from 526675 in 1998 to 315887 in 2005. For individuals aged 20 to 65 years, the enrollment declined from 812452 in 1998 to 726099 in 2005. For individuals aged older than 65 years, ODB enrollment increased from 1 483 388 to 1590719. The mean (SD) number of new prescriptions for any SSRI per 10000 individuals in the group aged younger than 20 years was 5.5 (1.1). For the group aged 20–65 years, the mean (SD) new prescriptions per 10000 individuals was 29.7 (1.9) and for individuals aged 66 years and older, the mean (SD) new prescriptions per 10000 individuals was 16.4 (1.8).
We tested the effect of the 5 antidepressant advisory warning dates on total new prescriptions for SSRIs as a group and for SSRIs individually. None of the 5 antidepressant warnings had an affect on new prescriptions for SSRIs as a group in any age category (Table 1
; Figure 1). The June 10, 2003, warning in the UK about the use of paroxetine resulted in a statistically significant 54% decrease (P = .03) in new prescriptions of paroxetine issued to patients aged younger than 20 years (Table 1; Figure 2). The North American antidepressant warnings in 2004 had no effect on the rate of new prescriptions for paroxetine in the age category younger than 20 years (Table 1; Figure 2). None of the 5 antidepressant warnings were associated with a change in new prescription rates for paroxetine in the 2 age categories for older patients (Table 1; Figure 3). Finally, the 5 antidepressant warnings had no affect on new antidepressant prescription rates for antidepressants other than paroxetine in any age category (data not shown).
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
We found that the June 2003 UK Committee on Safety of Medicines advisory against the use of paroxetine for depression in patients aged younger than 20 years resulted in a persistent and significant reduction in the rate of new paroxetine prescriptions. This antidepressant warning did not influence new paroxetine prescriptions in the other 2 age groups. Total new SSRI prescription rates did not change in association with any antidepressant advisory dates.
Physicians frequently do not heed adverse drug event warnings,18 particularly when the warnings are for newer medications.19 We believe there are 3 main reasons why the UK warning resulted in a reduction in the number of paroxetine prescriptions for adolescents, whereas the North American warnings had no obvious affect. First, the UK warning was specific. Unlike the other warnings issued, the UK warning advocated for avoiding a specific agent in a defined group of patients.7 By contrast, the later North American warnings were related to emergent suicidal behavior in connection with several antidepressants regardless of indication and, in the case of 2 of the North American warnings in 2004, of age. Other authors have advocated for greater specificity to improve adherence to drug warnings.18 Second, after the UK warning, physicians may have prescribed an alternate antidepressant for patients who otherwise might have received paroxetine. Conversely, more general warnings about multiple antidepressants may have left physicians with no acceptable therapeutic options, which caused physicians to disregard the warning. Finally, the UK advisory and the FDA counterpart were the first of their kind. Physicians who were most likely to respond to antidepressant warnings by changing prescribing behaviors may have responded to the June 2003 warning, which would have reduced the likelihood of an effect from the subsequent warnings.
Some limitations of this study merit emphasis. The ODB data on patients aged younger than 65 years represent an economically disadvantaged population and are not representative of the general population. However, these economically disadvantaged individuals have drug costs covered by ODB, thereby removing the most significant financial barrier—the cost of drugs—to receiving and filling prescriptions. Also, for our definition of new antidepressants, an individual could not have a prescription for any antidepressant in the 12 months preceding the index prescription. Although unlikely, it is possible that patients had been receiving antidepressant prescriptions outside of ODB coverage, in which case they would be falsely classified as new. This could bias against finding an effect, because such patients are less likely to have their treatment discontinued in response to warnings than are true new cases. Nonetheless, a substantial and statistically significant reduction was observed for new paroxetine prescriptions in patients aged younger than 20 years, which suggests that such a bias did not have an effect on new prescription rates. Finally, the 2004 North American antidepressant warnings were temporally close, which makes interpretation of the individual effects difficult. However, with multiple warnings so temporally close, we would expect to see a reduction in new prescription rates on the basis of the multitude of warnings, which we did not.
We found that a specific warning issued in the UK influenced the prescribing of paroxetine in Ontarians aged younger than 20 years, whereas subsequent, more generalized warnings issued in the United States and Canada did not. We speculate that the UK warning had an effect because it was the first of its kind, because alternatives to paroxetine were available to physicians and patients, and because the message contained in the warning was very specific. The subsequent North American warnings, which did not target a specific antidepressant or provide specific instructions regarding the use of antidepressants, did not have an effect on antidepressant prescriptions. We hypothesize that the lack of alternate treatment options and timing explain, at least in part, the stationary antidepressant prescription trends in response to the North American warnings. The reduction in paroxetine prescription rates in response to the UK warning has implications for agencies responsible for issuing drug safety warnings.
|
Acknowledgments |
|---|
P. A. Kurdyak was supported by a fellowship award from the Canadian Institute of Health Research, Rx&D–AstraZeneca–Canadian Psychiatric Research Foundation (CIHR DFE-64660), and by a Research Training Award from the Ontario Mental Health Foundation. D.N. Juurlink was supported by a New Investigator Award from the Canadian Institute of Health Research (CIHR MSH-69119) and the University of Toronto Drug Safety Group. M. M. Mamdani was supported by a Canadian Institute of Health Research New Investigator Award and by a New Investigator Award from the New Emerging Teams of the Canadian Institute of Health Research. The New Emerging Team program receives joint sponsorship from the Canadian Diabetes Association, the Kidney Foundation of Canada, Heart and Stroke of Canada, and the Institutes of Nutrition, Metabolism and Diabetes, and Circulatory and Respiratory Health.
Human Participant Protection
The ethics review board at Sunnybrook and Women’s College Health Sciences Centre approved this research project.
|
Footnotes |
|---|
Peer Reviewed
Contributors
All authors contributed to the origination of the study, as well as to analyzing data and writing and revising the article.
Accepted for publication July 3, 2006.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
1. Teicher MH, Glod C, Cole JO. Emergence of intense suicidal preoccupation during fluoxetine treatment. Am J Psychiatry. 1990;147:207–210.
2. Khan A, Khan S, Kolts R, Brown WA. Suicide rates in clinical trials of SSRIs, other antidepressants, and placebo: analysis of FDA reports. Am J Psychiatry. 2003;160:790–792.
3. Jick H, Kaye JA, Jick SS. Antidepressants and the risk of suicidal behaviors. JAMA. 2004;292: 338–343.
4. Fergusson D, Doucette S, Glass KC, et al. Association between suicide attempts and selective serotonin reuptake inhibitors: systematic review of randomised controlled trials. BMJ. 2005;330:396–399.
5. Gibbons RD, Hur K, Bhaumik DK, Mann JJ. The relationship between antidepressant medication use and rate of suicide. Arch Gen Psychiatry. 2005;62: 165–172.
6. Simon GE, Savarino J, Operskalski B, Wang PS. Suicide risk during antidepressant treatment. Am J Psychiatry. 2006;163:41–47.
7. Juurlink DN, Mamdani MM, Kopp A, Redelmeier DA. The risk of suicide with selective serotonin reuptake inhibitors in the elderly. Am J Psychiatry. 2006; 163:813–821.
8. Duff G. Safety of Seroxat (Paroxetine) in children and adolescents under 18 years–contraindication in the treatment of depressive illness. London, England: Committee on Safety of Medicines; 2003. Available from: http://www.mhra.gov.uk/home/idcplg?IdcService=GET_FILE&dID=2491&noSaveAs=0&Rendition=WEB. Accessed: November 3, 2004.
9. US Food and Drug Administration. FDA issues public health advisory entitled: Reports of suicidality in pediatric patients being treated with antidepressant medications for major depressive disorder (MDD). Rockville, Md: US Food and Drug Administration, Department of Health and Human Services; 2003.
10. US Food and Drug Administration. Suicidality in children and adolescents being treated with antidepressant medications. Rockville, Md: US Food and Drug Administration, Center for Drug Evaluation and Research; 2004.
11. Waechter F. Paroxetine must not be given to patients under 18. BMJ. 2003;326:1282.
12. US Food and Drug Administration. Worsening depression and suicidality in patients being treated with antidepressant medications. Rockville, Md: US Food and Drug Administration, Center for Drug Evaluation and Research; 2004.
13. Health Canada. Health Canada advises Canadians of stronger warnings for SSRIs and other newer antidepressants. Ottawa, Ontario: Health Canada; 2004. Available from: http://www.hc-sc.gc.ca/ahc-asc/media/advisories-avis/2004/2004_31_e.html. Accessed November 3, 2004.
14. Juurlink DN, Mamdani MM, Lee DS, et al. Rates of hyperkalemia after publication of the Randomized Aldactone Evaluation Study. N Engl J Med. 2004; 351:543–551.
15. Tu K, Mamdani MM, Jacka RM, Forde NJ, Rothwell DM, Tu JV. The striking effect of the Heart Outcomes Prevention Evaluation (HOPE) on ramipril prescribing in Ontario. CMAJ. 2003;168:553–557.
16. Helfenstein U. Box-Jenkins modelling in medical research. Stat Methods Med Res. 1996;5:3–22.
17. Ljung G, Box G. On a measure of lack of fit in time series models. Biometrika. 1978; 67:297–303.
18. Lasser KE, Seger DL, Yu DT, et al. Adherence to black box warnings for prescription medications in outpatients. Arch Intern Med. 2006;166:338–344.
19. Lasser KE, Allen PD, Woolhandler SJ, Himmelstein DU, Wolfe SM, Bor DH. Timing of new black box warnings and withdrawals for prescription medications. JAMA. 2002;287:2215–2220.
This article has been cited by other articles:
|
|
 |
|
  L. Y. Katz MD, A. L. Kozyrskyj PhD, H. J. Prior MSc, M. W. Enns MD, B. J. Cox PhD, and J. Sareen MD Effect of regulatory warnings on antidepressant prescription rates, use of health services and outcomes among children, adolescents and young adults Can. Med. Assoc. J., April 8, 2008; 178(8): 1005 - 1011. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. M. Mamdani PharmD MPH Health advisories: when good intentions go bad Can. Med. Assoc. J., April 8, 2008; 178(8): 1025 - 1026. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Olfson, S. C. Marcus, and B. G. Druss Effects of Food and Drug Administration Warnings on Antidepressant Use in a National Sample Arch Gen Psychiatry, January 1, 2008; 65(1): 94 - 101. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |