How Much Is Too Much? Escalating Psychotropic Class Polypharmacy,Diabetes Risk,and Clinical Complexity in Stimulant New Users

Polypharmacy refers to the administration or use of multiple medications at once. However, it is often taken to mean excessive prescribing, the importance of which rests on the observation that the probability of drug-related adverse events grows as the number of medications taken by a given patient increases (Lavan and Gallagher, 2016). Estimates of the frequency, drivers, and adverse consequences of polypharmacy have historically focused on elderly patients, with good reason. The elderly are especially prone to the use of multiple medications due to multimorbidity from chronic diseases (Nicholson et al., 2024). At the same time, they are at greater risk than their younger counterparts for adverse medication effects due to age-related changes in body composition, physiological functioning (e.g., decreased renal and hepatic function), and functional capacity (e.g., changes in strength, mobility, cognition, and perception/sensation) (Wastesson et al., 2018). With the aging of the general population, polypharmacy in the setting of multimorbidity from chronic conditions has become a worldwide public health priority (Davies et al., 2020).

Polypharmacy in pediatric populations is also a growing concern (Bolger et al., 2019). The subject of this short discussion, pediatric psychotropic polypharmacy, refers to the concomitant administration of multiple psychotropic medications to patients 21 years of age and younger, thus spanning childhood and adolescence (Zito et al., 2021). Children and adolescents typically lack the extensive lists of co-occurring general medical conditions that characterize many elderly patients; however, the mental health conditions for which psychotropic medications are often prescribed to young people have overlapping characteristics, high rates of diagnostic comorbidity, and other therapeutic challenges that may prompt multiple prescriptions (Baker et al., 2017). In addition, the chronic or recurring nature of psychiatric conditions and comorbidities in children and youth may result in greater cumulative exposure to multiple concomitant medications and medication-associated health risks, especially if initiated at a younger age.

As is the case with adults, polypharmacy in children and adolescents is simple in concept but difficult to operationally define. A threshold of two or more concomitant medications is often used to define polypharmacy in pediatric patients (Bakaki et al., 2018); however, this simple definition may overestimate polypharmacy rates if the true goal is to identify potentially harmful prescribing patterns. On the contrary, up to 300,000 U.S. children and youth take three or more concomitantly prescribed psychotropic medications from different pharmacological classes—a threshold that is more clearly associated with higher risk of adverse medication-related events (Zito et al., 2021). Regardless of definitional thresholds, several reports document an increasing prevalence of pediatric psychotropic polypharmacy, both overall and within vulnerable patient subgroups (reviewed in Jureidini et al., 2013; Zito et al., 2021). These findings highlight growing concerns about potential overprescribing of psychotropic medications in pediatric patients and underscore the importance of quantifying potential harms associated with specific polypharmacy patterns, independent of underlying indication, indication severity, or clinical complexity—factors that may increase both the risk of polypharmacy and adverse health outcomes.

In this issue of the Journal of Child and Adolescent Psychopharmacology, Zhu and colleagues present the results of a retrospective cohort study examining the association of newly diagnosed diabetes with increasing levels of psychotropic polypharmacy in children and adolescents treated with stimulants. The data source for this study was linked regional Medicaid data (2007–2014), used to identify a cohort of stimulant new users aged 2–17 years, with an average follow-up time of 6.4 years after the initiation of stimulant treatment. Concomitantly prescribed psychotropic classes included antipsychotics, antidepressants, anticonvulsant mood stabilizers, anxiolytics/hypnotics, alpha-2 receptor agonists, and lithium. These medication classes were used to create exposure groups defined by stimulant monotherapy and two- to five-drug class polypharmacy. Statistical analyses focused on the risk of newly diagnosed diabetes, adjusted for disease risk scores, time of follow-up, and other factors. Of the 30,112 stimulant new users, 43 were classified as having newly diagnosed diabetes. The relative risk of newly diagnosed diabetes was significantly increased with three- to five-class polypharmacy (vs. stimulant monotherapy), with longer cumulative duration of polypharmacy exposure (vs. exposure duration <120 days), and with higher disease risk score (vs. the lowest quartile). There was a significant linear trend for increased risk of newly diagnosed diabetes with ascending levels of polypharmacy.

The study by Zhu and colleagues is important because it goes beyond merely describing the frequency of and risk markers for psychotropic polypharmacy. The growth in pediatric psychotropic polypharmacy has paralleled rapid increases in childhood obesity and diagnosed diabetes (Misra et al., 2023; Wagenknecht et al., 2023). Adding to these concerns, youth-onset type 2 diabetes mellitus (T2DM) is considered more aggressive than adult-onset T2DM, characterized by faster decline in pancreatic function, more rapid accumulation of diabetes-associated complications, and higher mortality (TODAY Study Group, 2021). Stimulants, themselves, are not readily associated with adverse changes in glycemic profile or diabetes (Dong et al., 2024). However, drug-associated increases in body weight and new-onset diabetes are associated with medications that are often co-prescribed with stimulants, including orexigenic antipsychotic drugs, mood stabilizers, and antidepressants (Alruwaili et al., 2023; Barnard et al., 2013). Moreover, children and adolescents may be especially prone to their dysmetabolic side-effects (Correll, 2008; Safer, 2004; Verrotti et al., 2011; Wirrell, 2003). By rigorously linking psychotropic polypharmacy in stimulant-treated youth with newly diagnosed diabetes, the authors raise the hypothesis that multi-class psychotropic polypharmacy may be a preventable risk factor for new-onset diabetes in children and youth.

The methodological approach taken by Zhu and colleagues has several advantages. The use of claims data in this study enabled the assembly of a large cohort of stimulant new users, thus providing a high level of statistical power to examine the association of newly diagnosed diabetes with increasing levels of polypharmacy in a manner that was relatively free of recall bias or known difficulties with ascertaining clinical information from children and youth with severe mental health or neurodevelopmental conditions (Crystal et al., 2007). Analyses focused on the duration of polypharmacy appropriately emphasized daily chronic polypharmacy, reducing the odds of polypharmacy misclassification from as-needed psychotropic medications. Although the use of medication counts to define polypharmacy does not consider the appropriateness of co-prescribed medications (Cadogan et al., 2016; Masnoon et al., 2017), an analysis of risk by increasing levels of polypharmacy makes sense given the correlation of adverse event risk with the number of concurrent medications mentioned earlier. The authors’ focus on polypharmacy from different medication classes avoided the mistake of conflating the augmentation of long-acting stimulants with short-acting (immediate-release) stimulants, which is often needed to optimize therapeutic effects and should be considered appropriate use (Wolraich et al., 2019). Additionally, the validated computer case definition for newly diagnosed diabetes in this study was previously developed using state-level Medicaid data to identify incident diabetes in psychotropic-exposed children and youth (Bobo et al., 2013), making a strong case for its application to the present cohort.

The use of disease risk scores represents an additional methodological strength of the study by Zhu and colleagues. The ways in which clinical indications influence treatment decisions and adverse outcome risk are complex, presenting formidable methodological challenges outside of randomized studies (Sendor and Sturmer, 2022). In observational studies of medication risk, confounding by indication is a common threat to validity. The most important indications for stimulants, attention-deficit/hyperactivity disorder (ADHD) and binge-eating disorder, are each associated with increased T2DM risk (Garcia-Argibay et al., 2023; Raevuori et al., 2015). Common ADHD comorbidities, such as mood and anxiety disorders, are also associated with heightened risk of diabetes (Liu et al., 2022). As discussed further below, if more severe or complex cases are disproportionately represented in the polypharmacy groups, higher diabetes detection rates may be observed for reasons other than medication effects, such as poorer general health, higher medical service use, and more intensive medical surveillance. Many of these factors were accounted for in the disease risk score, which included structured information on psychiatric and general medical diagnoses, metabolic testing procedures, prescribed medications, and medical service use. However, residual bias from unmeasured factors such as family history and other genetic, dietary, and lifestyle factors may remain.

These limitations notwithstanding, the current study contributes to a broader discussion on how the term “polypharmacy” should be framed (Bakaki et al., 2018). While there is little doubt that concomitant medication prescribing introduces risk, the use of multiple medications is often necessary and appropriate for managing complex psychiatric conditions. For instance, despite their effectiveness in reducing ADHD symptoms and improving important aspects of functioning, many patients respond poorly or incompletely to stimulants (Chen et al., 2019; Childress and Sallee, 2014). Under such circumstances, clinicians may prescribe additional medications to improve overall efficacy and prevent known complications of under-treated ADHD (Thapar and Cooper, 2016). Furthermore, over half of patients with ADHD have additional psychiatric disorders or symptoms (such as aggression) requiring active management (Biederman et al., 1991; Cumyn et al., 2009; Saylor and Amann, 2016). These cases are often more severe and less likely to respond to stimulants than “pure” ADHD cases (Njardvik et al., 2025). Therefore, in some cases, polypharmacy may be less about excess and more about treatment resistance, clinical complexity, or both. On the contrary, even if increasing class polypharmacy represents greater severity or complexity of underlying indications, the observed linear trend in diabetes risk in this study supports a causal argument, and the adjusted effect sizes were convincingly large. This was especially so for four- and five-drug class polypharmacy, which may be difficult to justify in stimulant-treated children and youth in the absence of a compelling clinical reason for such aggressive regimens.

In sum, the study by Zhu and colleagues is a model for future research linking polypharmacy patterns with adverse health outcomes that significantly impact children and youth and have well-triangulated associations with the medications of interest. If further established, an association of psychotropic polypharmacy with newly diagnosed diabetes in young people should strongly inform polypharmacy prevention and deprescribing efforts. Within individual drug classes, however, the risk of adiposity, insulin resistance, and diabetes is known to vary (McIntyre et al., 2024). To inform prescribing decisions, more studies are needed to estimate the risk of diabetes associated with specific medication combinations. Such studies should prioritize polypharmacy patterns that include multiple orexigenic drugs. Future studies should also include additional measures or steps to validly judge the appropriateness of observed polypharmacy patterns. Though limited by the quality of clinical documentation and the need for expert case review (Bobo et al., 2019), longitudinal patient-level data from linked electronic health records may be used in future studies to integrate rigorous drug-outcome risk associations with clinically based estimations of the appropriateness of drug exposures.