Outcome measures in pediatric polypharmacy research: a scoping review

Abstract

Introduction

Various methods have been used to interpret the reports of pediatric polypharmacy across the literature. This is the first scoping review that explores outcome measures in pediatric polypharmacy research.

Objectives

The aim of our study was to describe outcome measures assessed in pediatric polypharmacy research.

Methods

A search of electronic databases was conducted in July 2017, including Ovid Medline, PubMed, Elsevier Embase, Wiley Cochrane Central Register of Controlled Trials (CENTRAL), EBSCO CINAHL, Ovid PsyclNFO, Web of Science Core Collection, ProQuest Dissertations and Thesis A&I. Data were extracted about study characteristics and outcome measures, and also synthesized by harms or benefits mentioned.

Results

The search strategy initially identified 8169 titles and screened 4398 using the inclusion criteria after de-duplicating. After the primary screening, a total of 363 studies were extracted for the data analysis. Polypharmacy (prevalence) was identified as an outcome in 31.4% of the studies, prognosis-related outcomes in 25.6%, and adverse drug reactions in 16.5%. A total of 265 articles (73.0%) mentioned harms, including adverse drug reactions (26.4%), side effects (24.2%), and drug-drug interactions (20.9%). A total of 83 studies (22.9%) mentioned any benefit, 48.2% of which identified combination for efficacy, 24.1% combination for treatment of complex diseases, and 19.3% combination for treatment augmentation. Thirty-eight studies reported adverse drug reaction as an outcome, where polypharmacy was a predictor, with various designs.

Conclusions

Most studies of pediatric polypharmacy evaluate prevalence, prognosis, or adverse drug reaction-related out-comes, and underscore harms related to polypharmacy. Clinicians should carefully weigh benefits and harms when introducing medications to treatment regimens.

Introduction

Medications play an essential role in effectively treating patients at any age [1]. The use of more than one medication is an inevitable consequence of pharmacotherapy for patients with multiple chronic conditions. The term ‘polypharmacy’ is often used to describe the use of multiple medications to treat one or multiple diseases. Polypharmacy is commonly used to describe the use of five or more prescribed medications for an adult, or two or more for a child [2–8]. Pediatric polypharmacy varies based on disease condition, medication class, duration of use, and healthcare setting [9]; however, a recent publication defined pediatric polypharmacy as the consumption of two or more distinct medications for at least 1 day [10]. Polypharmacy is associated with both benefits [11–13] and harms to patients; however, systematic or scoping reviews evaluating these outcomes are limited [9, 14–16]. Although polypharmacy is known to be prevalent among adult and elderly patients, there are also many reports of polypharmacy in pediatric populations [1, 6, 9]. For instance, the prevalence of antipsychotic polypharmacy in children has been reported to range from 2.9 to 27%. Often times, antipsychotic polypharmacy coincides with undesirable therapy outcomes including increased length of hospitalizations when polypharmacy is administered in inpatient settings [17, 18]. Feinstein and colleagues studied data from Colorado Medicaid beneficiaries and revealed that ≈ 35% of children were exposed to two or more concurrent medications for at least 1 day [9].

Polypharmacy in adults is sometimes necessary, but it may also be associated with an increased risk of adverse drug reactions or events, drug-drug interactions, hospitalizations, poor medication adherence, and mortality [19–21]. These increased risks may be extrapolated to the pediatric population. Most medications receive drug approval through the US Food and Drug Administration after solitary evaluation, and not in combination with other medications. Therefore, we are vastly unaware of possible drug-drug interactions, rare adverse events, or other health consequences that may occur with polypharmacy regimens. Many pediatric patients receive medications for chronic conditions for far beyond the length of time approved for drug trials, which limits our ability to predict possible health consequences from sustained use as the child receiving these medications matures and grows [22, 23]. The lack of safety and efficacy studies involving children is even more concerning, for it further limits our clinical decisions when managing complex disease states. A fine balance of preventing overtreatment while avoiding undertreatment is imperative to reduce harms and increase the benefits of polypharmacy in this patient population.

Compared with other children, those with chronic conditions are often exposed to more medications over longer periods of time, such as psychotropic agents, anticonvulsants, cardiovascular agents, and opioids [17, 18]. Yet, there are few literature evaluations describing outcomes associated with pediatric polypharmacy. This information is essential for advancing and guiding research in clinical decision-making processes, and for the formulation of clinical practice guidelines. Currently, we are unsure whether the prescribing patterns of physicians who employ the use of polypharmacy to manage complex conditions is problematic in the pediatric population.

Due to the varied nature of pediatric polypharmacy reports across the literature, there is a need to examine the differences through a comprehensive scoping review. Like a systematic review, a scoping review uses transparent and reproducible processes to define a research question, search for studies, and synthesize findings. However, compared with a systematic review, a scoping review tends to address a broader research question, may develop inclusion/exclusion criteria a posteriori or in an iterative fashion as the group reviews the existing literature, may not assess quality of studies, and the literature synthesis is more qualitative or descriptive than it is quantitative [24, 25].

To the best of our knowledge, this is the first scoping review that explores outcome measures related to pediatric polypharmacy. The aims of our study were to describe outcome measures assessed in pediatric polypharmacy research and to conduct a detailed evaluation of adverse outcomes associated with polypharmacy.

Methods

Search strategy

We conducted a large scoping review of pediatric polypharmacy and limited our search to children aged 18 years old or younger to explore or map the literature. We used the methodological framework proposed by Arksey and O’Malley [26] and enhanced and adopted by Levac et al. [25]. Our transdisciplinary team utilized the methods established by previous scoping review methodology papers [25–27]. A search of electronic databases was conducted in October 2016, which included Ovid Medline, PubMed, Elsevier Embase, Wiley Cochrane Central Register of Controlled Trials (CENTRAL), EBSCO CINAHL, Ovid PsycINFO, Web of Science Core Collection, ProQuest Dissertations, and Theses A&I. An updated search of electronic databases was conducted in July 2017. Search keywords included polypharmacy, multiple drug(s), hyperpolypharmacy, polytherapy, poly-medication(s), multiple medication(s), multiple prescription(s), and combination pharmacotherapy. To define the pediatric population, we used related vocabularies including child, infant, teen, pediatric, paediatric, school, pre-school, boy, girl, baby(ies), newborn, adolescent, juvenile, and minors (see Appendix 1 in the electronic supplementary material).

Eligibility criteria

Only original, primary studies on pediatric polypharmacy that defined or assessed polypharmacy as an aim, outcome, predictor, or covariate were included. We excluded non-English studies, studies pertaining solely to perinatal polypharmacy, case series, case reports, clinical trials, conference abstracts, letters, editorials, opinion pieces, review studies, and book chapters. We excluded experimental studies because we aimed to understand practice patterns and out-comes, which cannot be addressed by experimental studies due to their methodological uniqueness. Consistent with scoping review methodology, we did not assess the quality of the included studies.

Study selection

Screening studies for inclusion and exclusion began with a pilot screen of 100 abstracts, with the goal of reaching 80% interrater agreement at each level of screening. Screening was conducted in two stages. First, articles were independently screened based on title and abstract for inclusion or exclusion by two reviewers. Second, two independent reviewers screened all accessible full-text articles. Any disagreement between co-reviewers was resolved through a discussion or after reaching consensus from the full research team.

Data items and data collection process

Detailed information for the included studies was coded using a standardized data extraction tool that incorporated basic study information on characteristics listed in Table 1. Polypharmacy was identified as an outcome, aim, covariate, or predictor. Data about disease conditions, medication classes and categories, and definitions of polypharmacy (including number of medications and concurrence of medications) were also collected. Polypharmacy was also extracted as a primary or secondary aim, because it is often related to outcome measures. Any mention of harm and/or benefit in the included studies was also collected. Extracted harms varied and encompassed adverse drug reactions, side effects, drug interactions, adverse events, metabolic disorders, poor adherence, abnormal labs, toxicity, or death [28, 29]. Extracted benefits included drug combination for efficacy, treating complex diseases or augmentation (inherently referred to as beneficial polypharmacy), and rational polypharmacy. Other outcomes were compiled and labeled as ‘prognostics’, ‘pharmacokinetics’, ‘medication use evaluation (MUE)’, and ‘diagnostics’. Any clinical outcomes or those exploring behavioral change, measurement, exam, and action were reclassified under prognostics. Pharmacokinetic outcomes were those reports that studied dose and metabolism of medications or activity of blood markers. MUE included those non-polypharmacy outcomes reviewing the effect of the medication’s treatment duration, therapy adherence, and prescribing pattern. Diagnostic outcomes applied to any laboratory reports, imaging, and other measuring tools used in diagnosing disease (see Appendix 2 in the electronic supplementary material). Given the importance of adverse outcomes of polypharmacy, we further reviewed, mapped, and classified the characteristics of studies that reported adverse drug reaction(s) as their primary outcome measure and polypharmacy as a predictor.

Table 1.

Characteristics of 363 included studies

Characteristic	No. of studies (%)	Characteristic	No. of studies (%)
Country		Data sources
USA	126 (34.7)	Primary	172 (47.4)
India	23 (6.3)	Chart review	90 (24.8)
Japan	17 (4.7)	Claims	44 (12.1)
Italy	15 (4.1)	Electronic health records	17 (4.7)
Brazil	13 (3.6)	Healthcare setting
Canada	11 (3.0)	Outpatient	197 (54.3)
Turkey	11 (3.0)	Inpatient	77 (21.2)
UK	10 (2.8)	Combination	27 (7.4)
Egypt:	8 (2.2)	Others	51 (14.0)
Germany	8 (2.2)	Not reported	11 (3.0)
Nigeria	8 (2.2)	Study sample
Poland	8 (2.2)	1–49	41 (11.3)
Sweden	7 (1.9)	50–99	64 (17.6)
China	6 (1.7)	100–499	145 (39.9)
South Korea	6 (1.7)	500–999	19 (5.2)
Multiple	6 (1.7)	1000+	85 (23.4)
Netherlands	5 (1.4)	Not reported	9 (2.5)
Spain	5 (1.4)	Age
Others	70 (19.3)	Mean reported	165 (45.5)
Year		Median reported	32 (8.8)
1980–1985	5 (1.4)	Minimum reported	310 (85.4)
1986–1990	9 (2.5)	Maximum reported	320 (88.2)
1991–1995	20 (5.5)	Children-only study
1996–2000	29 (8.0)	No	44 (12.1)
1996–2005	58 (16.0)	Yes	319 (87.9)
2006–2010	74 (20.4)	Sex reported
2011–2015	117 (32.2)	No	75 (20.7)
2016–2017	51 (14.0)	Yes	288 (79.3)
Design		Health insurance
Case-control	4 (1.1)	Public	46 (12.7)
Cross sectional	235 (64.7)	Private	8 (2.2)
Prospective cohort	56 (15.4)	Combination	19 (5.2)
Retrospective cohort	68 (18.7)	Not Reported	290 (79.9)
		Polypharmacy identified as:
		Primary aim	238 (65.6)
		Primary outcome	114 (31.4)
		Main predictor	124 (34.2)
		Covariate	99 (27.3)

Data Synthesis

EPPI-Reviewer 4 software was used for screening, data extraction, mapping, and synthesis [27]. Using this software, we were able to determine the frequency distribution of coded information and qualitative synthesis of text information at the study level [30].

Results

The search strategy identified 8169 titles (Fig. 1). After de-duplicating records, a total of 4398 abstracts and titles remained for the first screening step. During the first screening step, 1082 studies qualified for screening on full text, resulting in a total of 363 studies, which were extracted for the primary data analysis. Most studies originated in the USA (n = 126, 34.7%), India (n = 23, 6.3%), Japan (n = 17, 4.7%) or Italy (n = 15, 4.1%). Polypharmacy was identified as an aim, outcome, main predictor, or covariate in 238 (65.6%), 114 (31.4%), 124 (34.2%), and 99 (27.3%) of the studies, respectively (Table 1). Polypharmacy was identified as the most frequent outcome measure (n = 114, 31.4%), followed by prognosis (n = 93, 25.6%), adverse drug reaction (n = 60, 16.5%), medication use (n = 34, 9.4%), pharmacokinetics (n = 29, 8.0%), and diagnostics (n = 17, 4.7%) [Fig. 2].

Fig. 1.

Flow diagram of studies identified, screened, and extracted [Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2009]

Fig. 2.

Primary outcome measures identified in 363 studies

Some of the included studies reported more than one harm. A total of 265 articles (73.0%) mentioned harms. Adverse drug reaction was mentioned the most (n = 96, 26.4%), followed by side effect (n = 88, 24.2%), drug-drug interaction (n = 76, 20.9%), adverse event (n = 75, 20.7%), metabolic disorder (n = 30, 8.3 %), poor adherence (n = 55, 15.2%), abnormal lab result (n = 28, 7.2%), behavioral toxicity (n = 18, 5.0%), and death (n = 16, 4.4%) [Table 2]. Benefits of pediatric polypharmacy were less frequently reported in the included studies. Out of all pediatric polypharmacy articles, only 83 studies (22.9%) stated any benefit, 40 of which (48.2%) identified combination for efficacy, 20 (5.5%) combination for treatment of complex diseases, 16 (4.4%) combination for augmentation, and 14 (3.9%) combination for rational drug use.

Table 2.

Benefits and harms mentioned in any section of 363 included studies

Harms and benefits	No. of studies (%)
Harms
Mentioned any harm	265 (73.0)
Adverse drug reactions (injury happens by taking medication at the normal doses; could be predictable or unpredictable [74])	96 (26.4)
Side effects (an undesired effect that occurs when the medication is administered regardless of the dose [29])	88 (24.2)
Drug-drug interaction	76 (20.9)
Adverse event (an undesired occurrence that results from taking medication correctly that includes adverse drug reaction, medication error, dose reduction and discontinuation of therapy [74])	75 (20.7)
Poor adherence (limited patient compliance with agreed recommendation from a health care provider [75])	55 (15.2)
Metabolic disorder	30 (8.3)
Abnormal laboratory results	28 (7.7)
Behavioral toxicity (common among medicinal drug users. Drugs frequently produce adverse effects and cause the users to not be adherent to the accurate therapy regimen [76])	18 (5.0)
Death	16 (4.4)
Benefits
Mentioned any benefit	83 (22.9)
Combinations for efficacy (any disease treatment with two or more drugs aiming to reach efficacy with lower doses or lower toxicity drugs; gain additive or synergistic effects; or combat expected acquired resistance or minimize the possibility for development of drug resistance [77])	40 (11.0)
Treat complex diseases (when disease is chronic or refractory and the monotherapy is not effective. In this situation, add-on therapy is frequently practiced)	20 (5.5)
Augmentation (the addition of a second agent to an existing medication regimen with the aim of achieving improved clinical response [78])	16 (4.4)
Rational drug use (when a patient receives appropriate medications based on their clinical needs, with individually adjusted dosage, for an adequate period of time, and at the lowest cost to them, the health system and the community [79])	14 (3.9)

Thirty-eight studies reported adverse drug reactions as an outcome, where polypharmacy was a predictor [31–68]. As summarized in Table 3, these studies were designed variably, including retrospective (n = 8) [33, 36, 37, 50, 54, 58, 62, 65], prospective (n = 8) [34, 35, 41, 45, 51, 56, 60, 67] and cross-sectional designs (n = 22) [31, 32, 38–40, 42–44, 46–49, 52, 53, 55, 57, 59, 61, 63, 64, 66, 68]. Most of the studies that reported adverse outcome variables (n = 28) were conducted in the USA, in an outpatient setting (n = 23), in neurological disorders (n = 36), and on central nervous system (CNS) medication(s) [n = 36]. Only four studies examined outcomes in cohorts of more than 1000 patients. As cited in Table 3, Bali et al. studied the cardiovascular safety of concomitant use of atypical antipsychotics and long-acting stimulants using a large cohort through claims data on children aged 6–16 years [36]. They showed polypharmacy was associated with increased risk of adverse drug reactions by 19%, but it was neither clinically nor statistically significant. McIntyre and Jerrell utilized South Carolina Medicaid claims data of children and examined the association between multiple antipsychotic treatment and cardiovascular and metabolic adverse events and showed an increased risk in multiple therapy compared with monotherapy [54]. Hilt et al. also reported a higher risk of adverse effects with multiple therapies, in a large cohort of children aged 3–17 years [48]. Patel et al., in their large retrospective cohort study, utilized an electronic medical record database and compared treated children and adolescents with bipolar disease with untreated pediatric patients. They concluded that antipsychotics and mood stabilizers are significantly associated with increased body mass index, especially when used in combination with other medication(s) [58].

Table 3.

Included studies that reported adverse drug reactions as an outcome variable

Study (y of publication)	Country	No. of par- ticipants (age < 21 y)	Aim	Study design	Care setting	Global pharmaco- logical category	Diagnostic group
Al-Qudah et al. (1991) [31]	Canada	88	To study further the role of carbamazepine-10–11-epoxide in producing neurotoxicity	Cross-sectional	NR	CNS agents	Neurology
Allarakhia et al. (1996) [32]	USA	259	[To] retrospectively review [clinical] experience with VPA and thrombocytopenia in 167 children evaluated at [an] institution over a 3-y period	Cross-sectional	Outpatient	CNS agents	Neurology
Amitai et al. (2015) [33]	Israel	104	To investigate the long-term effects of VPA treatment on thyroid-stimulating hormone levels, white blood cell and platelet count, liver function tests, and metabolic parameters in adolescents	Retrospective cohort	Inpatient	CNS agents	Combination
Anderson et al. (2015) [34]	UK	180	To prospectively determine the nature and rate of ADRs in children on AEDs	Prospective cohort	Outpatient	CNS agents	Neurology
Anghelescu et al. (2008) [35]	USA	117	To examine the safety of concurrent use of epidural and intravenous opioids in a consecutive series of 117 epidural infusions in pediatric pts and compare findings to those reported by other investigators	Prospective cohort	Inpatient	Analgesic antiinflammatory	Somatic
Bali et al. (2019) [36]	USA	37,903	To examine cardiovascular safety of concomitant use of long acting stimulant and atypical antipsychotic in children and adolescents with ADHD	Retrospective cohort	Outpatient	Psychotropic agents	Psychiatric
Becker et al. (2005) [37]	USA	228	[To study] the effects of psychotropic medications on the weight of adolescent pts	Retrospective cohort	Inpatient	Psychotropic agents	Psychiatric
Burch (2011) [38]	USA	59	To discover whether children in a pediatric rehabilitation setting are at higher risk for AEs because of polypharmacy, or at a lower risk because of relative overall clinical stability	Cross-sectional	Inpatient	Combination	Not specific
Cramer et al. (2011) [39]	Multiple	255	To compare AEs in patients with epilepsy taking different AEDs using standardized physician-completed questionnaires	Cross-sectional	Outpatient	CNS agents	Neurology
De Las Salas et al. (2016) [40]	Colombia	772	To describe the ADRs in inpatient children aged < 6 y in two general pediatrics wards located in Barranquilla, Colombia	Cross-sectional	Inpatient	Combination	Somatic
Dos Santos and Coelho (2006) [41]	Brazil	272	To investigate the occurrence of ADRs and associated risk factors in a pediatric hospital in northeast Brazil, from August to December 2001	Prospective cohort	Inpatient	CNS + psychotropic agents + others	Combination
El-Khayat et al. (2003) [42]	Egypt	130	To investigate the effect of epilepsy and AEDs on both the physical and hormonal aspects of the sexual development of male pts with epilepsy	Cross-sectional	Outpatient	CNS agents	Neurology
El-Rashidy et al. (2015) [43]	Egypt	60	To evaluate cardiac autonomic status in children with epilepsy on AEDs	Cross-sectional	Outpatient	CNS agents	Neurology
Farhat et al. (2002) [44]	Lebanon	29	[To study] the prevalence of vitamin D deficiency and insufficiency and bone mineral density in ambulatory pts on AEDs, [and to further explore] the impact of chronicity of intake of AEDs on bone density	Cross-sectional	Outpatient	CNS agents	Neurology
Galas-Zgorzalewicz et al. (1996H45]	Poland	63	To investigate the periodontal condition of children and adolescents using various AEDs such as carbamazepine, VPA or a combination of these drugs	Prospective cohort	NR	CNS agents	Neurology
Ghose and Taylor (1983) [46]	UK	84	To assess the serum copper and zinc levels in epileptics receiving long term medication	Cross-sectional	Schools	CNS agents	Neurology
Hallioglu et al. (2008) [47]	Turkey	92	To investigate the effects of epilepsy and various AEDs on heart rate variability	Cross-sectional	Outpatient	CNS agents	Neurology
Hilt et al. (2014) [48]	USA	1347	To investigate the side effect risks from using one or more psychiatric medications (including antipsychotics, antidepressants, α2-agonists, benzodiazepines, mood stabilizers, and stimulants) among a national cohort of children and adolescents	Cross-sectional	Parents	CNS + psychotropic agents	Psychiatric
Junger et al. (2015) [49]	USA	NR	To establish the MCIDs for the PESQ (Pediatric Epilepsy Side Effects Questionnaire) using a distribution-based method	Cross-sectional	Outpatient	CNS agents	Neurology
Kaplan et al. (2016) [50]	USA	108	To determine which anticonvulsants provided optimal seizure control and which resulted in the fewest side effects	Retrospective cohort	Outpatient	CNS agents	Neurology
Kim et al. (2010) [51]	South Korea	151	To verify the actual incidence and define the characteristics of topiramate-induced hypohidrosis	Prospective cohort	Outpatient	CNS agents	Neurology
Ko et al. (2001) [52]	Hong Kong	144	To investigate the relationship of thrombocytopenia, defined as a platelet count of <150 × 109 /L (normal range, 150–450 × 109 /L), with VPA therapy, age, duration of VPA therapy, and polytherapy	Cross-sectional	Outpatient, Inpatient	CNS agents	Neurology
Kwon et al. (2004) [53]	South Korea	152	[Evaluate] for drug-induced QT prolongation to see if the use of AEDs contributes to or controls sudden unexpected death in epileptic pts	Cross-sectional	Outpatient	CNS agents	Neurology
McIntyre and Jerrell (2008) [54]	USA	4140	To compare the incidence/prevalence of metabolic, cardiovascular, and cerebrovascular events in an antipsychotic - treated cohort	Retrospective cohort	Outpatient	Psychotropic agents	Psychiatric
Mikati et al. (2007) [55]	USA	143	To identify risk factors for subclinical hypothyroidism (thyroid-stimulating hormone levels >5 mlU/mL) in pts receiving VPA therapy	Cross-sectional	Outpatient	CNS agents	Neurology
Ohtahara and Yamatogi (2004) [56]	Japan	928	To examine the safety of zonisamide, including AEs and teratogenicity in a post marketing prospective follow-up survey of Japanese pts treated for 1–3 ys with zonisamide, either as monotherapy or as a component of polytherapy	Prospective cohort	Outpatient	CNS agents	Neurology
Otsuka et al. (1994) [57]	Japan	79	To evaluate the significance of measuring urinary W-acetyl-beta-glucosaminidase and guanidinoacetic acid in the chronic treatment of childhood epilepsies	Cross-sectional	Outpatient	CNS agents	Neurology
Patel et al. (2017) [58]	USA	2299	To assess the long-term effect of all treatment options for pediatric bipolar disorders on body mass index	Retrospective cohort	parents	Psychotropic agents	Psychiatric
Rashed et al. (2012) [59]	China	329	To determine the epidemiology of and identify risk factors for drug-related problems in hospitalized children in Hong Kong	Cross-sectional	Inpatient	Combination	Not specific
Rufo-Campos et al. (2006) [60]	Spain	62	To document the usefulness, safety, ease of dosing, and acceptance of oxcarbazepine oral suspension in monotherapy or polytherapy in a large pediatric population of pts	Prospective cohort	Outpatient	CNS agents	Neurology
Sarajlija et al. (2013) [61]	Serbia	35	To investigate the prevalence of vitamin D deficiency in females with Rett syndrome	Cross-sectional	Outpatient	CNS agents	Neurology
Shores (2005) [62]	USA	16	To monitor metabolic changes, including hyperprolactinemia, in adolescents medicated with atypical antipsychotics, especially when polypharmacy is involved	Retrospective cohort	Inpatient	Psychotropic agents	Psychiatric
Sobaniec et al. (2006) [63]	Poland	90	To estimate, first, the activity of the antioxidant enzymes superoxide dismutase, glutathione peroxidase, and glutathione reductase and, second, the malondialdehyde concentration in the erythrocytes in children and adolescents newly diagnosed with epilepsy and receiving either carbamazepine or VPA monotherapy or polytherapy with various AEDs	Cross-sectional	Outpatient	CNS agents	Neurology
Tekgul et al. (2005) [64]	Turkey	56	To investigate bone mass density values of ambulatory pediatric pts receiving long term AED treatment + a standard vitamin D3 supplement (400 U/day)	Cross-sectional	Outpatient	CNS agents	Neurology
Thomé-Souza et al. (2003) [65]	Brazil	28	To assess the risks and benefits of the co-administration of lamotrigine and VPA in a pediatric population with refractory epilepsy	Retrospective cohort	Outpatient	CNS agents	Neurology
Unay et al. (2006) [66]	Turkey	114	To assess the effects of AEDs on renal tubular function	Cross-sectional	Outpatient	CNS agents	Neurology
Verrotti et al. (1999) [67]	Italy	84	To evaluate plasma ammonia levels association with clinical signs and symptoms, plasma carnitine concentrations, and the eventual normalization after VPA withdrawal in a group of children taking VPA	Prospective cohort	Outpatient	CNS agents	Neurology
Wonodi et al. (2007) [68]	USA	118	To examine motor side effects associated with the use of antipsychotic drugs in children	Cross-sectional	Outpatient, Inpatient	Psychotropic agents	Psychiatric

ADHD attention deficit hyperactivity disorder, ADR adverse drug reaction, AE adverse event, AED antiepileptic drug, CNS central nervous system, MCID minimally clinically important difference, NR not reported, pts patients, VPA valproic acid/valporate, y year

Discussion

This is the first scoping review which identified outcome measures in pediatric polypharmacy research over the last 27 years. Out of the 363 studies included in this scoping review, 31.4% identified pediatric polypharmacy as an out-come, 73% reported harms, and 22.9% reported benefits. In articles where adverse effects were the outcome measures and polypharmacy the predictor variable, anticonvulsants [31–34, 39, 41–53, 55–57, 60, 61, 63–67] and antipsychotics [36, 37, 41, 48, 54, 58, 62, 68] were among the most common studied groups of CNS medications.

As the number of new medications released into the market continues to increase and the number of published pediatric polypharmacy studies dwindles, there becomes a growing risk for evidence-based practice diversions despite the recognized importance in identifying undesirable outcomes in pediatric polypharmacy [31–68]. Furthermore, because different phrases are often used to define harms throughout the pediatric polypharmacy literature, this may lead to misinterpretation of therapeutic outcomes. The terms ‘adverse event’, ‘adverse drug reaction’, and ‘side effect’ appeared interchangeably in various studies of this scoping review, even though they are not identical terms and may affect the accuracy of clinical judgments [29].

This review revealed that most studies identified polypharmacy as the primary outcome. Typically, these studies evaluated prevalence of pediatric polypharmacy. Although most studies reported on harms associated with polypharmacy, it is important to know that polypharmacy is sometimes necessary and beneficial in treating specific patient cohorts to increase the efficacy of drug therapy or decrease the chance of refractory diseases [69]. As indicated in Table 3, most adverse outcome-related reports examined the effect of polypharmacy in treating neurological diseases such as epilepsy and mood disorder [31–34, 37–39, 41–58, 60–68]. Anticonvulsants, psychotropics, and other CNS medications were among the most common drug categories found in scoping pediatric polypharmacy literature evaluating adverse outcomes. However, only a few reports are currently available that evaluate the effects of combination drug therapy when treating other common conditions such as cardiovascular disease, allergies, infectious diseases, and respiratory disorders in children [70–73].

This scoping review showed that there have not been many publications examining the association between polypharmacy and adverse outcomes, especially outside of psychiatric research. Further, identifying polypharmacy adverse out-comes is important for the determination of pharmacotherapy options, but these outcomes have not been studied extensively in the pediatric polypharmacy literature. Future large longitudinal cohort studies should determine the specific harms and benefits associated with pediatric polypharmacy, especially in vulnerable groups (e.g. patients with newly initiated medications and patients with multiple conditions), and aim to examine which patients benefit from polypharmacy in terms of quality of life, morbidity, and mortality.

Our search strategy consisted of a comprehensive literature search of eight bibliographic databases from their inception. Our transdisciplinary team followed a strict process to decrease inconsistency in data screening and extraction. A few limitations should also be noted. Health insurance status was not reported in the majority of included studies, which may influence the generalizability of these results for all children. Moreover, the lack of standardized medical subject headings, terms, and definitions for polypharmacy is a limiting factor. In accordance with scoping review methodological frameworks, no assessment of quality was performed for included studies, which inherently limits our knowledge about credibility. Of note, many disease states not listed in our findings may use polypharmacy; thus such use was not reported in our research. In addition, we did not synthesize specific information about interventions studied and reviewed within the literature as it is not the intended purpose of a traditional scoping review. Finally, although having a transdisciplinary team from different backgrounds offers various perspectives throughout a scoping review process of this nature and helps to strengthen the approach overall, variability among reviewers is still possible despite a predefined process, tool, and classification plan.

Conclusion

Most studies on pediatric polypharmacy evaluate prevalence, prognosis, or adverse drug reaction-related outcomes, and underscore harms related to polypharmacy. Neurological and psychiatric complex conditions were frequent in those studies reporting adverse drug reactions as an outcome variable. Clinicians should carefully weigh the benefits and harms when deciding to add medications to treatment regimens, especially when these are chronic medications adding to polypharmacy.

Further research is needed to characterize the outcomes related to benefits and harms around polypharmacy for children with comorbidity from various complex chronic conditions both individually and combined. Information gathered from such research will help in the creation of thorough guidelines and pediatric-focused initiatives. Using scoping review methodology allowed us to include a broader range of literature, characterize aims and outcome measures in pediatric polypharmacy studies, and identify existing limitations in research compared with traditional synthesis methodology.