Concomitant Use of Atypical Antipsychotics With Other Psychotropic Medication Classes and the Risk of Type 2 Diabetes Mellitus

Accepted Manuscript Concomitant Use of Atypical Antipsychotics With Other Psychotropic Medication Classes and the Risk of Type 2 Diabetes Mellitus Mehmet Burcu, PhD, MS, Julie M. Zito, PhD, Daniel J. Safer, MD, Laurence S. Magder, PhD, Susan dosReis, PhD, Fadia T. Shaya, PhD, MPH, Geoffrey L. Rosenthal, MD, PhD PII:

S0890-8567(17)30201-0

DOI:

10.1016/j.jaac.2017.04.004

Reference:

JAAC 1749

To appear in:

Journal of the American Academy of Child & Adolescent Psychiatry

Received Date: 28 January 2017 Revised Date:

15 April 2017

Accepted Date: 26 April 2017

Please cite this article as: Burcu M, Zito JM, Safer DJ, Magder LS, dosReis S, Shaya FT, Rosenthal GL, Concomitant Use of Atypical Antipsychotics With Other Psychotropic Medication Classes and the Risk of Type 2 Diabetes Mellitus, Journal of the American Academy of Child & Adolescent Psychiatry (2017), doi: 10.1016/j.jaac.2017.04.004. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Concomitant Use of Atypical Antipsychotics With Other Psychotropic Medication Classes and the Risk of Type 2 Diabetes Mellitus RH: Concomitant AAP Use and Diabetes Risk

RI PT

Mehmet Burcu, PhD, MS; Julie M. Zito, PhD; Daniel J. Safer, MD; Laurence S. Magder, PhD; Susan dosReis, PhD; Fadia T. Shaya, PhD, MPH; Geoffrey L. Rosenthal, MD, PhD This article is discussed in an editorial by Dr. Christoph U. Correll on p. xx. Supplemental material cited in this article is available online.

SC

Accepted May 1, 2017

M AN U

Drs. Burcu, Zito, Magder, dosReis, Shaya, and Rosenthal are with University of Maryland, Baltimore. Dr. Safer is with Johns Hopkins Medical Institutions, Baltimore.

The study was supported, in part, by the University of Maryland, Baltimore and, in part, by a grant from the US Food and Drug Administration (1U01FD004320). The supporting organizations had no role in the design, conduct, or reporting of the study. The views and opinions expressed in this article are those

TE D

of the authors and do not necessarily reflect the views and opinions of the supporting organizations. The preliminary findings from this study were presented at the 2016 Mid-Year Meeting of the International Society for Pharmacoepidemiology, Baltimore, MD, April 10-12, 2016. This study is also a

EP

part of Dr. Burcu’s doctoral dissertation.

AC C

Drs. Burcu and Magder served as the statistical experts for this research. Disclosure: Drs. Burcu, Zito, Safer, Magder, dosReis, Shaya, and Rosenthal report no biomedical financial interests or potential conflicts of interest. Correspondence to Mehmet Burcu, PhD, MS, Department of Pharmaceutical Health Services Research, University of Maryland, Baltimore, 220 Arch Street, Baltimore, MD 21201; email: [email protected].

ACCEPTED MANUSCRIPT ABSTRACT Objective: More than half of youth treated with atypical antipsychotic (AAP) medications are also treated with concomitant antidepressants or stimulants. This study assessed the association between

RI PT

antidepressant or stimulant use concomitant with AAPs and the risk of incident type 2 diabetes mellitus (T2DM).

Method: Medicaid Analytic eXtract data were used to conduct a retrospective cohort study of youth (520 years) who initiated AAP treatment. Concomitant antidepressant (selective serotonin reuptake

SC

inhibitors [SSRI]/serotonin-norepinephrine reuptake inhibitors [SNRIs], tricyclic/other cyclic

M AN U

antidepressants [TCAs], and other antidepressants) or stimulant use was assessed. The risk of incident T2DM was estimated using discrete time failure models, adjusting for disease risk score estimated using >125 baseline and time-dependent covariates.

Results: Among 73,224 AAP initiators, 43.0% had concomitant antidepressant use (76.4% were SSRI/SNRIs) and 43.8% had concomitant stimulant use. The study cohort had an average follow-up of

TE D

24.8 months (median=22.0 months, interquartile range [IQR]=10.0-38.0 months). In current AAPtreated youth, concomitant SSRI/SNRI (relative risk [RR]=1.84, 95% CI=1.30-2.59) or TCA use (RR=2.75, 95% CI=1.28-5.87) was associated with an increased risk of T2DM. By contrast,

EP

concomitant use of other antidepressants or stimulants was not associated with an increased risk of

AC C

T2DM. In concomitant users of AAPs and SSRI/SNRIs, the risk of T2DM increased with the duration of SSRI/SNRI use (RR=2.35, 95% CI=1.15-4.83 for ≥180 days vs. 1-180 days) as well as with the cumulative SSRI/SNRI dose (RR=1.99, 95% CI=1.08-3.67 for >2,700 mg vs. 1-2,700 mg fluoxetine dose equivalents)—after adjusting for the duration and cumulative dose of AAP use. By contrast, in concomitant users of AAPs and stimulants, neither duration nor cumulative dose of stimulants was associated with an increased risk of T2DM. Conclusion: In AAP-treated Medicaid-insured youth, concomitant SSRI/SNRI use was associated with

1

ACCEPTED MANUSCRIPT a heightened risk of T2DM, which intensified with increasing duration and dose. Key words: type 2 diabetes mellitus; atypical antipsychotics; antidepressants; stimulants; Medicaid INTRODUCTION

RI PT

Over the last two decades, second-generation antipsychotics, i.e., atypical antipsychotics (AAPs) have largely replaced first-generation antipsychotics.1 The increasing use of atypical antipsychotic medications in US youth has been profound,2,3 outpacing its growth in adults,2 particularly for off-label use in the treatment of externalizing (behavioral) disorders.2-4 Such expanded AAP use was particularly

SC

notable among Medicaid-insured US youth, whose prevalence of antipsychotic use was nearly 6-fold

M AN U

greater than their privately insured counterparts.5 The growth in pediatric use of AAPs did not just occur in the US but also in other western countries such as France,6 Germany,7 the Netherlands,8 and UK.9 This increased pediatric AAP use occurred amidst safety warnings. In 2003, in response to emerging cardiometabolic safety concerns, the US Food and Drug Administration (FDA) issued class warnings for AAPs related to their treatment-emergent risks, such as metabolic syndrome and diabetes.10

TE D

Even in many short-term clinical studies of AAPs in youth, treatment-emergent cardiometabolic abnormalities were prominent.11,12 More recently, population-based studies reported an increased risk of incident type 2 diabetes mellitus (T2DM) associated with antipsychotic exposure in community-treated

EP

youth populations.13-16 For example, in a retrospective cohort study of Medicaid-insured youth,13 Bobo

AC C

et al. showed that youth initiating treatment with antipsychotics had a 3-fold increased risk of developing T2DM compared with youth who initiated treatment with other psychotropic medication classes.

More than half of antipsychotic-treated, Medicaid-insured youth have concomitant exposure with other psychotropic medications, most commonly with stimulants or antidepressants,17,18 raising concerns regarding the metabolic effects of such combinations in youth. However, little attention has been given to the possibility of an increased risk of diabetes in youth populations associated with concomitant use

2

ACCEPTED MANUSCRIPT of AAPs with other major psychotropic medication classes. While there is no known biological plausibility for stimulant treatment-emergent risk of diabetes, there is growing evidence—mostly in the adult literature—that antidepressants are independently associated with an increased risk of T2DM.19-21

RI PT

The evidence in youth is limited to a secondary analysis in a recent study wherein antidepressant use during three months prior to AAP initiation was associated with an increased risk of T2DM.14 In another population-based study in youth, the absence of a difference in the risk of T2DM between antidepressant and AAP treatment (a risk factor for T2DM) also suggested that antidepressants may carry a T2DM

SC

risk.22 Thus, there is a need for a comprehensive assessment of the T2DM risk for AAPs in combination

of use, and cumulative dose exposure.

M AN U

with major psychotropic medication classes that accounts for time-dependent exposure status, duration

Consequently, the primary objective of the present study was to assess the risk of incident T2DM associated with concomitant use of antidepressant subclasses with AAPs in Medicaid-insured youth. We hypothesized that among AAP-treated youth, concomitant exposure to antidepressants would be

TE D

associated with an increased risk of T2DM. Also, we assessed the risk for T2DM associated with concomitant stimulant use with AAPs, hypothesizing that this combination would not be associated with an increased risk. Finally, we evaluated whether the risk of T2DM is affected by the duration of use and

AC C

METHOD

EP

the cumulative dose exposure to AAPs as well as to antidepressants and stimulants.

Data source

This study analyzed computerized administrative claims from the Medicaid Analytic eXtract database for 4 large, geographically diverse states (California, Florida, Illinois, New Jersey) for calendar years 2004 through 2009.23 The data included enrollment files and claim files for inpatient, outpatient, and physician services, and prescription drug dispensings. The enrollment files were used to derive information on monthly enrollment and eligibility status and sociodemographic characteristics of the

3

ACCEPTED MANUSCRIPT enrollees. An encrypted identification number was assigned to each Medicaid-insured youth to link enrollment files to service claims. Study Design and Population

RI PT

In a retrospective cohort design, the study population was comprised of Medicaid-insured youth who were 5-20 years of age. Since antipsychotic use in preschool-age children is not common,24 children <5 years of age were not included. A new-user design25 was applied by restricting the study cohort to those who initiated treatment with an oral atypical antipsychotic medication. The treatment

SC

initiation date served as the index date for cohort entry. There were 73,224 youth who initiated AAP

online). Incident Type 2 Diabetes Mellitus

M AN U

treatment and met the other study eligibility criteria (Figure 1) (Supplement 1, Table S1, available

The main study outcome was incident diagnosis of T2DM, which was expressed as the number of new cases of T2DM per 10,000 person-months of follow-up. Incident T2DM was ascertained

TE D

adapting a computerized database algorithm that was previously validated in a cohort of Medicaidinsured youth and had a positive predictive value of 83.9% (Supplement 2, available online).26 Youth in the study were followed until the incident T2DM. Youth were censored if they had a

EP

diabetes-related medical care encounter that did not meet the case definition of T2DM (i.e., polycystic

AC C

ovarian syndrome diagnosis, type 1 diabetes, or antidiabetic medication use in the absence of diabetes diagnosis) (Supplement 2, available online). For those without such events, the study cohort was followed until their 21st birthday, the end of continuous enrollment in the state Medicaid systems, or the end of study (December 31, 2009), whichever came first. Main Study Exposures In this cohort of AAP initiators, we also identified youth who had an exposure to antidepressant subclasses (selective serotonin reuptake inhibitors/serotonin-norepinephrine reuptake inhibitors

4

ACCEPTED MANUSCRIPT [SSRI/SNRIs], tricyclic and other related cyclic antidepressants [TCAs], and other antidepressants) or stimulants (Table S2, available online). During follow-up, the usage of study medications was operationalized using three time-

RI PT

dependent measures that accounted for exposure status (current use, former use, or nonuse), duration of use (in days), and cumulative dose exposure. The exposure status was considered current use unless medications were discontinued for more than 90 days. The 90-day time window aimed to account for carryover effects27 of recently discontinued medications and also to allow for a biologically plausible

SC

time for the development and detection of treatment-emergent T2DM. Former use was defined as having

M AN U

previous use that did not occur within the past 90 days. Otherwise, the exposure status during follow-up was categorized as nonuse. Of note, because the study was nested in a cohort of AAP initiators, AAP exposure status was never classified as nonuse throughout follow-up. For antidepressants and stimulants, the exposure status was categorized as nonuse if there was no dispensing during follow-up as well as during 365 days prior to the cohort entry date.

TE D

The duration of exposure was assessed using the dispensing date and the days of supply information available from the prescription drug claims, and it was calculated as the sum of total days of supply that were available starting from the cohort entry date. Similarly, the cumulative dose exposure

EP

was assessed using the dispensing date, strength, and quantity supplied information available in the

AC C

prescription drug claims, and it was calculated as the sum of total dosage that were available starting from the cohort entry date. The cumulative dose was calculated in risperidone dosage equivalents for AAPs, in fluoxetine dosage equivalents for antidepressants, and in methylphenidate dosage equivalents for stimulants (Supplement 3, available online). Analysis All analyses were conducted using SAS version 9.3 (SAS Institute, Inc., Cary, NC). To estimate the adjusted incidence of T2DM, we conducted several discrete time failure models wherein the unit of

5

ACCEPTED MANUSCRIPT analysis was person-months. Discrete time analyses have been previously used14,28,29 and produce estimates approximately identical to those from Cox proportional hazard regression models while providing computational efficiency and having practical advantages in the presence of time-dependent

RI PT

exposures.30,31 To account for confounding, we constructed a disease risk score to be included within the discrete time failure models. Disease risk score, analogous to propensity score, is a summary confounder score that allows for parsimonious models. It has advantages over the propensity score in the presence of

SC

multiple time-dependent exposure groups.32-34 In the current study, to estimate the disease risk score,

M AN U

which is the probability of incident T2DM conditional on study covariates, we used the Miettinen fullcohort approach.35,36 The study covariates (>125 baseline and time-dependent covariates) included a range of sociodemographic, administrative, clinical, and other health care utilization characteristics (Table S3, available online).

The final models that assessed the risk of T2DM according to main study exposures included

TE D

only the variables for the main study exposures, disease risk score (expressed as percentile ranks, Figure S2, available online) and time from cohort entry. In these models, we first compared the risk of T2DM according to current vs. former AAP use. The subsequent analyses were restricted to youth with current

EP

AAP use (i.e., excluding former use) to assess the risk according to AAP use concomitant with

AC C

antidepressant subclasses or stimulants.

Duration-response analyses. In youth with current AAP use, we assessed the risk of incident T2DM according to duration of AAP exposure. Subsequently, we evaluated whether the risk associated with concomitant AAP use with SSRI/SNRIs or stimulants differed with long-term (>180 days) vs. short-term (1-180 days) AAP exposure. In a subset of youth who were using AAPs concomitantly with SSRI/SNRIs, we also assessed the risk of T2DM according to duration of SSRI/SNRI use. Likewise, we assessed the risk according to duration of stimulant use in youth who were using AAPs concomitantly

6

ACCEPTED MANUSCRIPT with stimulants. Dose-response analyses. In youth with current AAP use, we assessed the risk of incident T2DM according to cumulative AAP dose (in risperidone dose equivalents). Also, we evaluated whether the

RI PT

risk associated with concomitant AAP use with SSRI/SNRIs or stimulants differed in youth with high (>500 mg) vs. low (1-500 mg) cumulative AAP dose. In a subset of youth who were using AAPs concomitantly with SSRI/SNRIs, we further assessed the risk of T2DM according to cumulative SSRI/SNRI dose (in fluoxetine dose equivalents). Likewise, in youth who were using AAPs

SC

concomitantly with stimulants, we assessed the risk according to cumulative stimulant dose (in

M AN U

methylphenidate dose equivalents).

Of note, the duration-response and dose-response analyses were not conducted for TCAs or other antidepressants due to their limited use.

Subgroup and sensitivity analyses. We assessed the risk of T2DM according to SSRI/SNRI and stimulant use concomitant with AAPs in youth only enrolled in fee-for-service programs, in older youth

TE D

(10-20 years) whose antipsychotic use was most prevalent,24 and in youth without a clinician-reported diagnosis of an FDA-labelled/evidence-based indication for AAPs.37 Also, we assessed the T2DM risk within the first year of follow-up (to assure that baseline covariates remained unchanged). In another

EP

analysis, we did not require the 90-day time window that could account for carry-over effects.

AC C

Additionally, we conducted analyses excluding youth who had exposures to antidepressant or stimulant medications prior to the index date (cohort entry) so as to attenuate prevalent user bias.25 Because depressive symptoms may be independently associated with an increased risk of T2DM,38 we assessed the risk of T2DM according to concomitant SSRI/SNRI use by restricting the cohort to youth diagnosed with depressive disorders (to attenuate confounding by indication). Finally, we conducted an analysis in youth diagnosed with attention-deficit/hyperactivity disorder (ADHD)—the leading condition for stimulant use in youth—to estimate the risk of T2DM according to stimulant use concomitant with

7

ACCEPTED MANUSCRIPT AAPs. RESULTS Characteristics of the Study Cohort

RI PT

The study cohort largely comprised youth between ages 5 and 14 (68.6%), males (64.6%), and non-Caucasian youth (55.5%) (Table 1). More than half of the study youth were eligible for Medicaid based on low family income (55.7%). Likewise, most study youth were enrolled in fee-for-service programs (57.6%). The vast majority (92.2%) had a clinician-reported psychiatric diagnosis. The leading

SC

clinician-reported psychiatric diagnoses, in rank order, were ADHD (42.9%), depressive disorders

M AN U

(34.4%), and disruptive behavioral disorders (29.7%). In the year prior to AAP treatment initiation, more than half (56.8%) of the study youth had exposure to at least one other psychiatric medication class. Only one-quarter (28.5%) of youth had a baseline metabolic monitoring-related procedure. The study cohort had an average follow-up of 24.8 months (median=22.0 months, interquartile range [IQR]=10.0-38.0 months) (Table S4, available online). The average duration of AAP use was

TE D

294.2 days (median=154.0 days, IQR=60.0-403.0 days). During follow-up, 43.8% of the study cohort had concomitant use of stimulants, and 34.4% had concomitant use of SSRI/SNRIs. Far fewer had concomitant use of TCAs (2.6%) and other antidepressant medications (16.4%).

EP

The Risk of Incident Type 2 Diabetes Mellitus

AC C

There were 182 incident cases of T2DM, most of which occurred during current AAP use (Table 2). Compared to youth without T2DM, youth who developed incident T2DM during follow-up were more likely to be older (mean age=13.7 vs. 12.5 years), female (55.0% vs. 35.3%), non-white (67.6% vs. 55.5%), and enrolled in Medicaid through the Supplemental Security Income (SSI) program (29.7% vs. 19.6%) (data not shown). By contrast, there was no statistically significant difference between youth with and without incident T2DM with respect to state of residence, fee-for-service vs. managed care enrollment, or calendar year of AAP treatment initiation (data not shown). Following treatment initiation

8

ACCEPTED MANUSCRIPT with AAPs, there was a 3.07-fold greater risk (95% CI=2.02-4.65) of incident T2DM during current than former AAP use (absolute risk=1.35 vs. 0.44/10,000 person-months) (Table 2). In current AAP-treated youth, concomitant exposure to SSRI/SNRIs (relative risk [RR]=1.84,

RI PT

95% CI=1.30-2.59) or to TCAs (RR=2.75, 95% CI=1.28-5.87) was associated with increased risk of T2DM (Table 2). By contrast, concomitant use of other antidepressants (RR=1.42, 95% CI=0.92-2.18) or stimulants (RR=0.71, 95% CI=0.46-1.08) were not associated with an increased risk.

Among current AAP users, the risk of T2DM increased with the duration of AAP use and

SC

cumulative AAP dose (in risperidone dose equivalents) (Tables S5 and S6, available online). The risk of

M AN U

incident T2DM associated with SSRI/SNRI use concomitant with AAPs varied according to the duration of AAP use and cumulative AAP dose. Concomitant SSRI/SNRI use was associated with an increased risk of T2DM in youth with >180 days of AAP use (RR=2.21, 95% CI=1.42-3.44) and youth who had >500 mg cumulative AAP dose (RR=1.95, 95% CI=1.24-3.05), but not in youth with 1-180 days of AAP use (RR=1.34. 95% CI=0.77-2.35) and youth with 1-500 mg cumulative AAP dose (RR=1.58,

TE D

95% CI=0.92-2.71). By contrast, concomitant use of stimulants with AAP use was not associated with an increased risk of T2DM, regardless of duration of AAP use and cumulative AAP dose. Subsequently, we assessed whether the increased risk of T2DM associated with concomitant

EP

SSRI/SNRI use was dependent on the duration of SSRI/SNRI use and cumulative SSRI/SNRI dosage (in

AC C

fluoxetine dose equivalents). In youth with concomitant AAP and SSRI/SNRI use, the risk of incident T2DM increased significantly with the duration of SSRI/SNRI use and cumulative SSRI/SNRI dose (Table 3). Youth with ≥180 days of SSRI/SNRI use had a 2.35-fold increased risk of T2DM (95% CI=1.15-4.83) compared to those with 1-180 days of SSRI/SNRI use. Similarly, youth who had >2,700 mg (median) cumulative SSRI/SNRI dose had a 1.99-fold increased risk of T2DM (95% CI=1.08-3.67) compared to youth who had ≤2,700 mg cumulative SSRI/SNRI dose. Conversely, in youth with concomitant AAP and stimulant use, the duration of stimulant use or the cumulative stimulant dose were

9

ACCEPTED MANUSCRIPT not associated with an increased risk of incident T2DM (Table 3). Sensitivity and Subgroup Analyses Among current AAP users (Figure 2a and Table S7, available online), the risk of incident T2DM

RI PT

remained elevated in youth with concomitant SSRI/SNRI use in the following circumstances: a) when the study cohort was restricted to fee-for-service enrollees (RR=1.68, 95% CI=1.10-2.57); b) among older youth, i.e., 10–20-year-olds (RR=1.85, 95% CI=1.30-2.63); c) in youth without a clinicianreported diagnosis of an FDA-labelled/evidence-based indication for AAPs (RR=2.40, 95% CI=1.51-

SC

3.84); d) during the first year of follow-up (RR=1.92, 95% CI=1.18-3.12); e) when there was no time

M AN U

window allowance for carry-over effects (RR=2.13, 95% CI=1.46-3.11); f) when youth with SSRI/SNRI use prior to cohort entry were excluded in a cohort of AAP initiators (new use of both AAPs and SSRI/SNRIs) (RR=1.58, 95% CI=1.04-2.41); g) when youth with SSRI/SNRI use prior to cohort entry were excluded in a cohort that also included prevalent AAP users (RR=2.02, 95% CI=1.45-2.81); and when the study cohort was restricted to youth diagnosed with depressive disorders (RR=2.65, 95%

TE D

CI=1.50-4.69). By contrast, in youth with current AAP use, concomitant stimulant use was not associated with an increased risk of incident T2DM across similar sensitivity and subgroup analyses

Post Hoc Analyses

EP

(Figure 2b and Table S7, available online).

AC C

Because clozapine (an atypical antipsychotic agent) may be used in the treatment of complicated (severe) cases, we also conducted a post hoc analysis excluding youth treated with clozapine (N=93). Similar results were obtained in that post hoc analysis: SSRI/SNRI use concomitant with AAPs was significantly associated with an increased risk of T2DM (RR=1.83, 95% CI=1.29-2.58), whereas stimulant use concomitant with AAPs was not (RR=0.70, 95% CI=0.46-1.07) (data not shown). In another post hoc analysis, we assessed the effect of the combined use of stimulants and SSRI/SNRIs plus AAPs on the risk of incident T2DM. The combined use of stimulants and SSRI/SNRIs

10

ACCEPTED MANUSCRIPT in AAP-treated youth was not significantly associated with an increased risk of T2DM compared with AAP-treated youth with no concomitant use of stimulants and SSRI/SNRIs (RR=1.30, 95% CI=0.732.31) (data not shown). Finally, to reduce potential surveillance bias, we conducted a post hoc analysis

RI PT

in youth who did not receive a baseline metabolic monitoring. Youth who received metabolic monitoring services may be more likely to have complex medication use patterns and/or more clinically complex disorders. Nonetheless, in AAP-treated youth who did not receive a baseline metabolic screening procedure, concomitant SSRI/SNRI use was still associated with an increased risk of T2DM

SC

(RR=2.37, 95% CI=1.48-3.82) while concomitant stimulant use was not associated with an increased

M AN U

risk of T2DM (RR=0.59, 95% CI=0.33-1.06) (data not shown). DISCUSSION

In this cohort of Medicaid-insured youth who initiated treatment with atypical antipsychotics (AAPs), a major finding on concomitant AAP use with SSRI/SNRIs—the most commonly used antidepressant subclass—was the 1.84-fold increased risk of incident type 2 diabetes mellitus.

TE D

Moreover, the increased risk intensified markedly with increasing duration of SSRI/SNRI use and cumulative SSRI/SNRI dose. There was a 2.35-fold increased risk of T2DM for youth with >180 days of exposure compared youth with 1-180 days of exposure and a nearly 2-fold increased risk of T2DM

EP

for youth with >2,700 mg cumulative SSRI/SNRI dose (in fluoxetine dose equivalents) compared with

AC C

youth 1-2,700 mg cumulative SSRI/SNRI dose. By contrast, as hypothesized, stimulant use concomitant with AAP use was not associated with an increased risk of T2DM. Given the recent growth in concomitant AAP use with antidepressants among Medicaid-insured youth,18,39 these findings merit attention.

To our knowledge, this is the first comprehensive study that assesses the risk of T2DM associated with concomitant use of AAPs with the other leading psychotropic medication classes. As a recently published systematic review revealed,16 the risk of T2DM associated with antidepressant use in

11

ACCEPTED MANUSCRIPT AAP-treated youth has been thus far reported in a single study.14 In that large retrospective cohort study of Medicaid-insured youth,14 Rubin et al. reported that antidepressant use in the month of AAP initiation or in the 3 months prior was associated with 54% increased risk of incident T2DM compared to youth

RI PT

who did not have antidepressant exposure in that time window. Our study corroborates and strengthens this finding by providing a comprehensive analysis examining the T2DM risk according to

antidepressant subclass use, duration of use, and cumulative dose exposure in a cohort of AAP-treated youth.

SC

While pediatric safety data on AAP and antidepressant combinations are scarce, there is a

M AN U

growing body of research reporting antidepressant treatment-emergent risk of T2DM in adult populations.19-21 For example, in a case-control study nested in a large cohort of adults who were new users of antidepressants,20 long-term use of SSRIs and TCAs were, respectively, associated with 1.77and 2.06-fold increased risk of incident T2DM. Several reasons have been postulated to explain the association between antidepressant exposure and the risk of T2DM. These include significant weight

TE D

gain associated with long-term antidepressant use19,40; altered glucose metabolism through insulin resistance and inhibition of insulin secretion21,41; and hyperglycemia, particularly with antidepressants that have high affinity for norepinephrine reuptake transporter, 5-HT2c (serotonin) receptor, and H1

EP

(histamine) receptor.42 Additionally, several SSRIs are known to inhibit certain cytochrome P450

AC C

enzymes,43 thereby potentially slowing AAP metabolism and prolonging their treatment-emergent metabolic effects. On the other hand, there is also evidence that in short-term treatment, serotonin reuptake inhibitors may improve glucose homeostasis in non-diabetic patients.44 Further research is warranted to better elucidate the underlying biological mechanism for the incidence of T2DM associated with concomitant AAP and antidepressant exposure in youth populations. At the service delivery level, there are potentially serious implications from the study findings. In response to public health and government concerns,37,45-47 several state Medicaid agencies have

12

ACCEPTED MANUSCRIPT implemented oversight policies targeting pediatric use of AAP medications.48 However, many of these programs do not have cardiometabolic monitoring criteria for complex AAP use and do not apply to older youth who exhibit high rates of AAP use with other psychotropic medication classes.24 Additional

RI PT

efforts are needed to implement systematic drug prescribing oversight to assure appropriate use and metabolic monitoring of atypical antipsychotic therapy, particularly in complex multiple drug regimens and for use in disorders that lack sufficient evidence of benefit-risk.

Rigorous statistical methods were employed to control for a wide range of potential confounders

SC

at baseline as well as over the course of follow-up, and the drug exposures were operationalized using

M AN U

multiple measures (i.e., current/former/nonuse, duration of use, and cumulative dose) to assess the risk of T2DM. In addition, we conducted several sensitivity/subgroup analyses to examine the robustness of the study design/results. Nevertheless, several limitations should be noted. First, unmeasured confounding cannot be completely eliminated in observational studies. For example, many common risk factors for T2DM such as unhealthy diet, physical inactivity, overweight and obesity, family history of

TE D

diabetes, and untreated hypertension are not available or sparsely available in administrative Medicaid claims data. However, in the present study, we used a “new user” design25 and nested our study in a cohort of AAP initiators. This design has strengths in mitigating potential bias due to unmeasured

EP

confounding (e.g., disease severity, unmeasured clinical characteristics, differential metabolic

AC C

monitoring/surveillance between AAP users and non-users).49,50 There is also evidence that depressive disorders may be independently associated with T2DM risk,38 and depression and obesity/metabolic syndrome may have a reciprocal association.51,52 Therefore, AAP-treated youth who initiated concomitant SSRI/SNRI treatment may differ from youth who did not initiate concomitant SSRI/SNRI treatment. Consequently, to further attenuate potential confounding by indication, when the study cohort was restricted to youth diagnosed with depressive disorders, the risk of incident T2DM remained elevated in youth with concomitant SSRI/SNRI use (Figure 2).

13

ACCEPTED MANUSCRIPT Second, medication dispensings from administrative claims do not always equate with actual consumption. However, the focus on long-term use and cumulative dose exposures increases the likelihood of actual consumption. Third, the data cannot rule out missing diagnoses, but our study

RI PT

outcome (T2DM) was ascertained using a computerized algorithm that was previously validated against medical records of Medicaid-insured youth.26 Fourth, duration- and dose-response analyses were not conducted for non-SSRI/SNRI antidepressants due to inadequate exposure, but our study provided a comprehensive analysis by featuring the major antidepressant subclass, duration of use, and cumulative

SC

dose. Fifth, the study did not examine the effect of individual AAP, antidepressant, or stimulant

M AN U

medications on the risk of T2DM. However, the main focus of the study was to assess the risk of T2DM in association with antidepressant subclass or stimulant use concomitant with AAP drug class. The findings provide motivation for future research to identify distinct patterns of individual antipsychotic combination regimens that are more likely to be associated with an increased risk of T2DM. Overall, the absolute risk for incident T2DM was low and should be interpreted in light of the

TE D

general issue of potentially undetected T2DM in the study population. In addition, the relative risk estimates should be interpreted with caution by considering the available evidence for both benefits and risks of complex AAP combinations in youth. However, in a sensitivity analysis, the risk remained

EP

elevated for concomitant AAP and SSRI/SNRI exposure in youth without a clinician-reported diagnosis

AC C

of an FDA-labelled/evidence-based indication for AAPs.37 More than half of AAP-treated youth in the study did not have a clinician-reported diagnosis of an FDA-labelled indication, and only 28.5% of AAP-treated youth had received baseline metabolic monitoring—a finding consistent with previous studies in youth.53-55 These findings highlight public health concerns regarding the use of AAPs in youth populations. Also, metabolic adverse effects may not be limited to T2DM. Finally, the findings may not necessarily apply to privately insured or uninsured US youth, but published utilization studies5,17 indicate that the relatively greater prevalence of AAP use in the Medicaid population warrants an

14

ACCEPTED MANUSCRIPT emphasis on this group. Of note, Medicaid-insured youth from the 4 large geographically diverse study states represent more than one-quarter of the total Medicaid-insured youth in the US.56 In a large cohort of Medicaid-insured youth between the ages of 5 and 20, concomitant use of

RI PT

SSRI/SNRIs with AAPs was associated with a greater risk of type 2 diabetes mellitus, which further intensified significantly with the increasing duration of SSRI/SNRI use and cumulative SSRI/SNRI dose. In view of the growing complexity of atypical antipsychotic regimens in treating Medicaid-insured youth18,39 and low rates of baseline metabolic monitoring in youth initiating AAP treatment,53-55 the

SC

study findings suggest that complex AAP regimens should be used judiciously with appropriate

M AN U

cardiometabolic monitoring. Continued efforts are warranted to support Medicaid oversight policies48 so as to assure safe and effective use of complex AAP regimens in youth populations. REFERENCES 1.

Alexander GC, Gallagher SA, Mascola A, Moloney RM, Stafford RS. Increasing off-label use of

2011;20:177-184. 2.

TE D

antipsychotic medications in the United States, 1995-2008. Pharmacoepidemiol Drug Saf.

Olfson M, Blanco C, Liu SM, Wang S, Correll CU. National trends in the office-based treatment of children, adolescents, and adults with antipsychotics. Arch Gen Psychiatry. 2012;69:1247-56. Zito JM, Burcu M, Ibe A, Safer DJ, Magder LS. Antipsychotic use by Medicaid-insured youths:

EP

3.

4.

AC C

impact of eligibility and psychiatric diagnosis across a decade. Psychiatr Serv. 2013;64:223-229. Burcu M, Zito JM, Ibe A, Safer DJ. Atypical antipsychotic use among Medicaid-insured children and adolescents: duration, safety, and monitoring implications. J Child Adolesc Psychopharmacol. 2014;24:112-119. 5.

Crystal S, Olfson M, Huang C, Pincus H, Gerhard T. Broadened use of atypical antipsychotics: safety, effectiveness, and policy challenges. Health Aff (Millwood). 2009;28:w770-781.

6.

Verdoux H, Pambrun E, Cortaredona S, Tournier M, Verger P. Antipsychotic prescribing in

15

ACCEPTED MANUSCRIPT youths: a French community-based study from 2006 to 2013. Eur Child Adolesc Psychiatry. 2015;24:1181-1191. 7.

Bachmann CJ, Lempp T, Glaeske G, Hoffmann F. Antipsychotic prescription in children and

2012. Dtsch Arztebl Int. 2014;111:25-34. 8.

Kalverdijk LJ, Tobi H, van den Berg PB, et al. Use of antipsychotic drugs among Dutch youths between 1997 and 2005. Psychiatr Serv. 2008;59:554-560.

Rani F, Murray ML, Byrne PJ, Wong IC. Epidemiologic features of antipsychotic prescribing to

SC

9.

RI PT

adolescents: an analysis of data from a German statutory health insurance company from 2005 to

10.

M AN U

children and adolescents in primary care in the United Kingdom. Pediatrics. 2008;121:1002-9. FDA Patient Safety News, Show #28, June 2004. Warning about hyperglycemia and atypical antipsychotic drugs. http://www.fda.gov/downloads/Safety/FDAPatientSafetyNews/UCM 417797.pdf. Accessed January 2, 2017. 11.

Cohen D, Bonnot O, Bodeau N, Consoli A, Laurent C. Adverse effects of second-generation

2012;32:309-316. 12.

TE D

antipsychotics in children and adolescents: a Bayesian meta-analysis. J Clin Psychopharmacol.

Correll CU, Manu P, Olshanskiy V, Napolitano B, Kane JM, Malhotra AK. Cardiometabolic risk

EP

of second-generation antipsychotic medications during first-time use in children and adolescents.

13.

AC C

JAMA. 2009;302:1765-1773.

Bobo WV, Cooper WO, Stein CM, et al. Antipsychotics and the risk of type 2 diabetes mellitus in children and youth. JAMA Psychiatry. 2013;70:1067-1075.

14.

Rubin DM, Kreider AR, Matone M, et al. Risk for incident diabetes mellitus following initiation of second-generation antipsychotics among Medicaid-enrolled youths. JAMA Pediatr. 2015;169:e150285.

15.

Sohn M, Talbert J, Blumenschein K, Moga DC. Atypical antipsychotic initiation and the risk of

16

ACCEPTED MANUSCRIPT type II diabetes in children and adolescents. Pharmacoepidemiol Drug Saf. 2015;24:583-591. 16.

Galling B, Roldan A, Nielsen RE, et al. Type 2 diabetes mellitus in youth exposed to antipsychotics: a systematic review and meta-analysis. JAMA Psychiatry. 2016;73:247-259. Burcu M, Safer DJ, Zito JM. Antipsychotic prescribing for behavioral disorders in US youth:

RI PT

17.

physician specialty, insurance coverage, and complex regimens. Pharmacoepidemiol Drug Saf. 2016;25:26-34. 18.

Kreider AR, Matone M, Bellonci C, et al. Growth in the concurrent use of antipsychotics with

SC

other psychotropic medications in Medicaid-enrolled children. J Am Acad Child Adolesc

19.

M AN U

Psychiatry. 2014;53:960-970 e962.

Kivimaki M, Hamer M, Batty GD, et al. Antidepressant medication use, weight gain, and risk of type 2 diabetes: a population-based study. Diabetes Care. 2010;33:2611-2616.

20.

Andersohn F, Schade R, Suissa S, Garbe E. Long-term use of antidepressants for depressive disorders and the risk of diabetes mellitus. Am J Psychiatry. 2009;166:591-598. Barnard K, Peveler RC, Holt RI. Antidepressant medication as a risk factor for type 2 diabetes

TE D

21.

and impaired glucose regulation: systematic review. Diabetes Care. 2013;36:3337-3345. 22.

Andrade SE, Lo JC, Roblin D, et al. Antipsychotic medication use among children and risk of

Medicaid. Medicaid Analytic eXtract (MAX) general information.

AC C

23.

EP

diabetes mellitus. Pediatrics. 2011;128:1135-1141.

https://www.medicaid.gov/medicaid-chip-program-information/by-topics/data-andsystems/max/max-general-information.html. Accessed Janurary 2, 2017. 24.

Olfson M, King M, Schoenbaum M. Treatment of young people with antipsychotic medications in the United States. JAMA Psychiatry. 2015;72:867-874.

25.

Ray WA. Evaluating medication effects outside of clinical trials: new-user designs. Am J Epidemiol. 2003;158:915-920.

17

ACCEPTED MANUSCRIPT 26.

Bobo WV, Cooper WO, Stein CM, et al. Positive predictive value of a case definition for diabetes mellitus using automated administrative health data in children and youth exposed to antipsychotic drugs or control medications: a Tennessee Medicaid study. BMC Med Res

RI PT

Methodol. 2012;12:128. 27.

Cleophas TJ. Carryover bias in clinical investigations. J Clin Pharmacol. 1993;33:799-804.

28.

Magder LS, Petri M. Incidence of and risk factors for adverse cardiovascular events among patients with systemic lupus erythematosus. Am J Epidemiol. 2012;176:708-719.

Winterstein AG, Gerhard T, Kubilis P, et al. Cardiovascular safety of central nervous system

SC

29.

30.

M AN U

stimulants in children and adolescents: population based cohort study. BMJ. 2012;345:e4627. Allison PD. Analysis of tied or discrete data with proc logistic. In: Allison PD, ed. Survival Analysis Using SAS: A Practical Guide. 2 ed. Cary, NC: SAS Institute, Inc.; 2010:235-256. 31.

D'Agostino RB, Lee ML, Belanger AJ, Cupples LA, Anderson K, Kannel WB. Relation of pooled logistic regression to time dependent Cox regression analysis: the Framingham Heart

32.

TE D

Study. Stat Med. 1990;9:1501-1515.

Arbogast PG, Ray WA. Use of disease risk scores in pharmacoepidemiologic studies. Stat Methods Med Res. 2009;18:67-80.

Hansen BB. The prognostic analogue of the propensity score. Biometrika. 2008;95:481-488.

34.

Arbogast PG, Ray WA. Performance of disease risk scores, propensity scores, and traditional

AC C

EP

33.

multivariable outcome regression in the presence of multiple confounders. Am J Epidemiol. 2011;174:613-620. 35.

Glynn RJ, Gagne JJ, Schneeweiss S. Role of disease risk scores in comparative effectiveness research with emerging therapies. Pharmacoepidemiol Drug Saf. 2012;21 Suppl 2:138-147.

36.

Miettinen OS. Stratification by a multivariate confounder score. Am J Epidemiol. 1976;104:609-20.

37.

ABIM Foundation’s Choosing Wisely. American Psychiatric Association: five things physicians

18

ACCEPTED MANUSCRIPT and patients should question. http://www.choosingwisely.org/doctor-patient-lists/americanpsychiatric-association/. Accessed January 2, 2017. 38.

Rotella F, Mannucci E. Depression as a risk factor for diabetes: a meta-analysis of longitudinal

39.

RI PT

studies. J Clin Psychiatry. 2013;74:31-37. Fontanella CA, Warner LA, Phillips GS, Bridge JA, Campo JV. Trends in psychotropic polypharmacy among youths enrolled in Ohio Medicaid, 2002-2008. Psychiatr Serv. 2014;65:1332-40. 40.

Blumenthal SR, Castro VM, Clements CC, et al. An electronic health records study of long-term

Isaac R, Boura-Halfon S, Gurevitch D, Shainskaya A, Levkovitz Y, Zick Y. Selective serotonin

M AN U

41.

SC

weight gain following antidepressant use. JAMA Psychiatry. 2014;71:889-896.

reuptake inhibitors (SSRIs) inhibit insulin secretion and action in pancreatic beta cells. J Biol Chem. 2013;288:5682-5693. 42.

Derijks HJ, Meyboom RH, Heerdink ER, et al. The association between antidepressant use and

2008;64:531-538. 43.

TE D

disturbances in glucose homeostasis: evidence from spontaneous reports. Eur J Clin Pharmacol.

Nemeroff CB, DeVane CL, Pollock BG. Newer antidepressants and the cytochrome P450 system. Am J Psychiatry. 1996;153:311-320. Deuschle M. Effects of antidepressants on glucose metabolism and diabetes mellitus type 2 in

EP

44.

45.

AC C

adults. Curr Opin Psychiatry. 2013;26:60-65. United States Government Accountability Office (GAO). Foster children: HHS guidance could help states improve oversight of psychotropic prescriptions. GAO-12-201. Washington, DC: December 14, 2011. 46.

United States Government Accountability Office (GAO). Children's Mental Health: Concerns about appropriate services for children in Medicaid and foster care. GAO-13-15. Washington, DC: December 10, 2012.

19

ACCEPTED MANUSCRIPT 47.

United States Government Accountability Office (GAO). Foster children: HHS could provide additional guidance to states regarding psychotropic medications. GAO-14-651T. Washington, DC: May 29, 2014. Schmid I, Burcu M, Zito JM. Medicaid prior authorization policies for pediatric use of

RI PT

48.

antipsychotic medications. JAMA. 2015;313:966-968. 49.

Schneeweiss S, Patrick AR, Sturmer T, et al. Increasing levels of restriction in

results. Med Care. 2007;45(10 Supl 2):S131-142.

Schneeweiss S, Suissa S. Advanced approaches to controlling confounding in

M AN U

50.

SC

pharmacoepidemiologic database studies of elderly and comparison with randomized trial

pharmacoepidemiologic studies. In: Strom BL, Kimmel SE, Hennessy S, eds. Textbook of Pharmacoepidemiology. 2 ed. West Sussex, UK: Wiley Blackwell; 2013:324-336. 51.

Luppino FS, de Wit LM, Bouvy PF, et al. Overweight, obesity, and depression: a systematic review and meta-analysis of longitudinal studies. Arch Gen Psychiatry. 2010;67:220-229. Pan A, Keum N, Okereke OI, et al. Bidirectional association between depression and metabolic

TE D

52.

syndrome: a systematic review and meta-analysis of epidemiological studies. Diabetes Care. 2012;35:1171-1180.

Morrato EH, Nicol GE, Maahs D, et al. Metabolic screening in children receiving antipsychotic

EP

53.

54.

AC C

drug treatment. Arch Pediatr Adolesc Med. 2010;164:344-351. Raebel MA, Penfold R, McMahon AW, et al. Adherence to guidelines for glucose assessment in starting second-generation antipsychotics. Pediatrics. 2014;134:e1308-1314. 55.

Rodday AM, Parsons SK, Mankiw C, et al. Child and adolescent psychiatrists' reported monitoring behaviors for second-generation antipsychotics. J Child Adolesc Psychopharmacol. 2015;25:351-361.

56.

The Henry J. Kaiser Family Foundation. Distribution of Medicaid enrollees by enrollment group,

20

ACCEPTED MANUSCRIPT FY2011. http://kff.org/medicaid/state-indicator/distribution-by-enrollment-group/. Accessed January 2, 2017.

With Atypical Antipsychotic Medications, 2005-2009, N=73,224 %

22,505 27,722 17,670 5,327

30.7 37.9 24.1 7.3

47,315 25,909

64.6 35.4

32,590 17,844 15,943 6,847

44.5 24.4 21.8 9.4

18,080 14,336 40,808

24.7 19.6 55.7

38,109 6,589 20,482 8,044

52.0 9.0 28.0 11.0

42,168 31,056

57.6 42.4

67,512 5,508 5,490 421 9,877 21,752 31,418 25,156 11,707 9,530 6,398 861 896 268 3,210

92.2 7.5 7.5 0.6 13.5 29.7 42.9 34.4 16.0 13.0 8.7 1.2 1.2 0.4 4.4

EP

AC C

M AN U

SC

n

TE D

Characteristic Age group 5-9 years 10-14 years 15-17 years 18-20 years Gender Male Female Race/ethnicity Caucasian African American Hispanic Othera Medicaid eligibility group Foster care SSI TANF/CHIP State medicaid program California Florida Illinois New Jersey Payment system Fee-for-service Managed care Psychiatric diagnostic group Any psychiatric diagnosis Schizophrenia/other psychoses PDD/ID Tic disorders Bipolar disorder Disruptive behavior disorders ADHD Depressive disorders Anxiety disorders Adjustment disorder Communication and learning disorder Personality disorder Somatoform spectrum disorders Sleep disorders of nonorganic origin Alcohol and other substance abuse

RI PT

Table 1. Baseline Characteristics of Medicaid-Insured Youth (5-20 Years) Who Initiated Treatment

21

14,652

20.0

41,555 29,361 5,069 8,478 25,536 6,880 18,015 12,954 1,436 6,493 2,590 557 20,840 3,258 7.884 18,646

56.8 40.1 6.9 11.6 34.9 9.4 24.6 17.7 2.0 8.9 3.5 0.8 28.5 4.5 10.8 25.5

SC

Other psychiatric disorders Psychiatric medications Any psychiatric medication use ADHD drugs Atomoxetine Central alpha agonists Stimulants Anticonvulsant-mood stabilizers Antidepressants SSRI/SNRI TCA Other antidepressants Anxiolytic/hypnotics Lithium Metabolic screening procedures Diabetes screening Hyperlipidemia screening Metabolic panelb

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

Note: ADHD = attention-deficit/hyperactivity disorder; PDD/ID = pervasive developmental disorder/intellectual disability; SSI = supplemental security income; SSRI/SNRI = selective serotonin reuptake inhibitor/serotonin-norepinephrine reuptake inhibitor; TANF/CHIP = Temporary Assistance for Needy Families/Children’s Health Insurance Program (youth eligible for Medicaid based on low family income); TCA = tricyclic and other related cyclic antidepressant. a Other (race/ethnicity) includes youth of Asian, Native Hawaiian or other Pacific Islander race/ethnicity, and youth with more than one race or unknown race/ethnicity b Metabolic panel includes testing for blood glucose level

22

ACCEPTED MANUSCRIPT

Table 2. The Risk of Incident Type 2 Diabetes and Concomitant Use of Atypical Antipsychotics With

Cases

Absolute risk per 10,000 person-months

Adjusted risk ratioa

95% CI

687,565 1,125,602

30 152

0.44 1.35

1 3.07

Reference 2.02-4.65

745,566 125,967 254,069

66 17 69

0.89 1.35 2.72

1 0.88 1.84

Reference 0.51-1.52 1.30-2.59

1.32 0.75 3.91

1 0.50 2.75

Reference 0.12-2.03 1.28-5.87

107 19 26

1.16 2.11 2.27

1 1.26 1.42

Reference 0.76-2.09 0.92-2.18

102 20 30

2.01 1.44 0.63

1 1.03 0.71

Reference 0.63-1.70 0.46-1.08

b b

921,288 89,885 114,429

M AN U

1,081,165

508,127 138,670 478,805

SC

Personmonths

143

b b

TE D

Exposure status Atypical antipsychotics Former use Current use Among current users of atypical antipsychotics SSRI/SNRIs Nonuse Former use Current use TCAs Nonuse Former use Current use Other antidepressantsc Nonuse Former use Current use Stimulants Nonuse Former use Current use

RI PT

Antidepressant and With Stimulant Medications

AC C

EP

Note: SSRI/SNRI = selective serotonin reuptake inhibitor/serotonin-norepinephrine reuptake inhibitor; TCA = tricyclic and other related cyclic antidepressant a Adjusted for disease risk score (expressed as percentile ranks) and time from cohort entry (i.e., follow-up month). b One of the cells may have <11 cases. To protect patient confidentiality, the current policy by the United States Centers for Medicare and Medicaid Services (CMS) stipulates that no cell (e.g., admissions, discharges, patients, services) with counts of 10 or less may be displayed. Also, no use of other data or other mathematical formulas may be used if they result in the display of a cell 10 or less. c Other antidepressants included bupropion, mirtazapine, nefazodone, and trazodone. Among youth treated with these “other” antidepressants, 48.0% used bupropion, 43.8% used trazodone, 21.0% used mirtazapine, and <1.0% used nefazodone.

23

ACCEPTED MANUSCRIPT

Table 3. The Risk of Incident Type 2 Diabetes According to Duration and Cumulative Dose of Selective Serotonin Reuptake Inhibitor (SSRI)/Serotonin-Norepinephrine Reuptake Inhibitor (SNRI) or Stimulant

Stimulants

Cases

152,890 101,179

27 42

127,040 127,029

213,529 265,276

Adjusted risk ratioa

95 %CI

1.77 4.15

1 2.35

Reference 1.15-4.83

19 50

1.50 3.94

1 1.99

Reference 1.08-3.67

13 17

0.61 0.64

1 0.98

Reference 0.34-2.78

11 19

0.46 0.79

1 1.62

Reference 0.66-4.00

SC

Absolute risk per 10,000 person-months

M AN U

TE D

Among concomitant users of SSRI/SNRIs and AAPs Duration of SSRI/SNRI use ≤180 days of SSRI/SNRI use >180 days of SSRI/SNRI use Cumulative SSRI/SNRI doseb ≤Median >Median Among concomitant users of stimulants and AAPs Duration of stimulant use ≤180 days of stimulant use >180 days of stimulant use Cumulative stimulant dosec ≤Median >Median

Personmonths

RI PT

Exposure in Concomitant Users of Atypical Antipsychotic (AAP) Medications With SSRI/SNRIs or

239,425 239,380

AC C

EP

Note: a Models were adjusted for disease risk score (expressed as percentile ranks), time from cohort entry (i.e., follow-up month), duration of AAP use, cumulative AAP dosage, and exposures to tricyclic and other related cyclic antidepressants and other antidepressant medications. In concurrent users of SSRI/SNRIs and AAPs, exposure to stimulants was also adjusted. Likewise, in concurrent users of stimulants and AAPs, exposure to SSRI/SNRIs was also adjusted. b Cumulative SSRI/SNRI dose was calculated in fluoxetine equivalents (median=2,700 mg, interquartile range [IQR]= 9007,436 mg) c Cumulative stimulant dose was calculated in methylphenidate equivalents (median=8,450 mg, interquartile range [IQR]=2,700-21,708 mg)

24

ACCEPTED MANUSCRIPT

FIGURES Figure 1. Flow chart for the study cohort of Medicaid-insured youth (5-20 years) who were new users of

RI PT

atypical antipsychotic medications, 2005-2009. Note: aLife-threatening or serious somatic conditions included sickle cell disease, cystic fibrosis, cerebral palsy, cancer, human immunodeficiency virus (HIV) infection, organ transplant, dialysis/end stage renal disease, respiratory failure, aplastic anemia,

SC

congenital immune deficiencies, Down syndrome, other lethal chromosomal anomalies, fatal metabolic

M AN U

diseases, serious neuromuscular disease. bYouth were also excluded if they had any inpatient or outpatient claim with a diagnosis of diabetes (type 1 or type 2) and if they had a dispensing of an antidiabetic medication during 365 days prior to the index date.

Figure 2. Subgroup and sensitivity analyses assessing the risk of incident type 2 diabetes according to concomitant use of selective serotonin reuptake inhibitor (SSRI)/serotonin–norepinephrine reuptake

TE D

inhibitors (SNRIs) or stimulants with atypical antipsychotic (AAP) medications. Note: Panel A shows relative risk for concomitant SSRI/SNRI use; Panel B shows relative risk for concomitant stimulant use.

AC C

EP

FDA = Food and Drug Administration.

25

ACCEPTED MANUSCRIPT

RI PT

Youth (5-20 years) with a dispensing of an oral atypical antipsychotic medication between January 2005 and December 2009 n=183,263

AC C

EP

TE D

M AN U

SC

Youth excluded (in the following order): • Dispensing for an oral or injectable conventional or atypical antipsychotic during 365 days preceding the index date, n=49,387 • Partial enrollment over 12-month period preceding the index date, n=31,013 • Enrollment in a comprehensive or behavioral managed care plan (Florida Medicaid only), n=16,246 • <2 outpatient visits on separate days (regardless of diagnosis) during 365 days preceding the index date, n=1,680 • An admission to a long-term care facility during 365 days preceding the index date, n=6,231 • Hospice care or diagnoses/medical services for life-threatening serious somatic illnessa during 365 days preceding the index date, n=3,256 • Pregnancy-related diagnoses or medical services during 365 days preceding the index date, n=1,640 • Polycystic ovarian syndrome diagnosis during 365 days preceding the index date, n=54 • Diagnosis of diabetes, or dispensing for insulin or oral antidiabetic medications during 365 days preceding the index date, n=532b

Final study population of Medicaid-insured youth (5-20 years) who were new users of atypical antipsychotic medications N=73,224

A

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

B