Metabolic and Endocrine ADRs of Atypical Antipsychotics (AAPs) in Paediatric Patients with Autism Spectrum Disorder (ASD): A Review of Prevalence, Risk Factors, and Implications for Clinical Monitoring

Abstract

Atypical antipsychotics (AAPs) remain the most effective treatment to control irritability associated with autism spectrum disorder (ASD). Although there is no pharmaceutical treatment to target the core symptoms of ASD, AAPs reduce their severity. However, AAPs have been reported to be associated with severe adverse drug reactions (ADRs) that may lead to long-term conditions such as diabetes mellitus and heart disease. Their prevalence varies depending on the type of AAP prescribed, age, ethnicity, gender, healthcare systems, and the severity of the ASD. Current ADR monitoring guidelines exist, but they are broad in scope and do not fully account for these factors. Therefore, the need to develop ADR monitoring guidelines considering these factors has increased with the expanded use of AAPs in paediatrics with ASD. This gap in knowledge and clinical practice highlights the ongoing need for research to explore these factors and how they can inform the creation of tailored guidelines for monitoring ADRs in this population.

1. Introduction

Autism spectrum disorder (ASD) is a neurological condition characterised by challenges in social interaction and repetitive interests [1]. Recently, its prevalence has been reported to be 1% of the population in the UK, 2.5% in Saudi Arabia, and 2.8% in the USA [2,3,4]. The UK and the USA are presented as well-studied populations, whereas Saudi Arabia as an underrepresented population [5]. To date, there is no pharmacological treatment targeting core symptoms (difficulties in social interaction and restricted interests) [6]. However, atypical antipsychotics (AAPs) are more effective at reducing core symptoms than other treatment options, such as antidepressants, stimulants, and mood stabilisers [7]. Additionally, AAPs have been shown to be most effective in managing irritability and comorbidities associated with ASD [8,9]. Furthermore, two AAPs, risperidone and aripiprazole, have been approved by the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) for children aged 5 and older with ASD [10,11]. Risperidone and aripiprazole are therefore the first-line treatments for irritability in children and adolescents with ASD [11]. While other AAPs, including olanzapine, quetiapine, lurasidone, and ziprasidone, are still prescribed for off-label use. These differences in FDA approval status may explain why risperidone and aripiprazole are more commonly prescribed and prioritised in clinical guidelines compared to other AAPs [12]. However, administration of AAPs is associated with neurological, cardiac, metabolic, and endocrine adverse drug reactions (ADRs). Recent studies have focused on metabolic and endocrine ADRs than other ADRs for several reasons. Firstly, they may contribute to the development of long-term conditions, such as diabetes and heart disease [13]. Secondly, they are associated with the risk of osteoporosis and reproductive system issues (e.g., hypogonadism and menstrual irregularities), particularly in children and adolescents [14]. Thirdly, they are easily detected and managed in the early stages, whereas preventing or slowing the progression of their complications can be challenging [15]. These ADRs and their clinical complications are influenced by various risk factors and also impose a burden on healthcare systems [16]. Therefore, investigating the factors affecting ADR prevalence is essential for developing monitoring guidelines for paediatric patients with ASD.

The aim of this review is to explore the factors related to medications, patients, healthcare systems, and disease that influence the prevalence of ADRs, specifically metabolic and endocrine ADRs, and their clinical implications.

2. Adverse Drug Reactions (ADRs)

The World Health Organization (WHO) plays a crucial role in ensuring safety for those who are treated with medications and other therapeutic agents by developing clinical guidelines and initiatives. Health organisations worldwide have recently renewed their focus on pharmacovigilance due to the complexity of drug evaluation and the rapid evolution of medications and diagnostic tools [17,18]. WHO defines pharmacovigilance as “The science and activities related to the detection, assessment, understanding, and prevention of adverse drug effects or any other possible drug-related problems” [19,20].

The drug event terminology includes ADRs and medication errors (MEs) [21]. According to WHO, MEs is “any preventable event that may cause or lead to inappropriate medication use or patient harm while the medication is in the control of the healthcare professional, patient, or consumer” [17]. By contrast, WHO defines ADR as “Any response to a drug which is noxious and unintended which occurs at doses normally used in man for prophylaxis, diagnosis, or therapy of disease, or for the modifications of physiological function” [18,22]. In parallel, EMA and the Medicines and Healthcare Products Regulatory Agency (MHRA) in the UK, define ADR as “a response to a medicinal product which is noxious and unintended” [18]. From these definitions, the key difference lies in preventability. To illustrate this explicitly, MEs can be prevented through correct use, while ADRs may occur even when medication is used properly [23].

National and international health organisations encourage health professionals and patients to report any ADRs through national spontaneous reporting systems such as the FDA, the MHRA and the Saudi Food and Drug Authority (SFDA), which in turn analyse the data and report to the Uppsala Monitoring Centre (UMC) [24]. In the UK, ADRs comprised 16% of all hospital admissions, with an average length of stay of approximately one week. The direct cost of ADR admissions for one month was around GBP 468,993 (USD 634,409) in the UK [25]. Similarly, ADRs account for 11.7% of all hospital admissions in the USA, with the average length of stay reaching 11 days. Conversely, in Saudi Arabia, ADRs account for 1.6% of hospital admissions with no prolongation of hospital stay [26].

ADRs are an avoidable strain on the finite resources of healthcare systems. Thus, investigation of these ADRs is essential, particularly in medications potentially associated with severe ADRs or in vulnerable populations such as children or patients with mental health disorders [27]. Based on the classification of Rawlins and Thompson, ADRs can be categorised into six types: A, B, C, D, E, and F [18,28]. Type A (augmented reaction) is dose-dependent and results from the pharmacological activity of the drug, while Type B (bizarre reaction) is idiosyncratic and unrelated to the pharmacological activity of the drug. Type C (chronic reaction) is a time-dependent reaction related to the accumulation of the dose over time, whereas Type D (delayed reaction) appears long after drug use. Type E (end-of-use) occurs upon withdrawal of the drug, and Type F (failure of treatment) refers to unpredictable treatment failure [18,28].

Another scale for ADR classification is the Hartwig Scale (

Table 1. The classification of atypical antipsychotics (AAPs) ADRs based on the Hartwig scale.

Level	Severity	Definition	Example
1	Mild	ADR is detected without requiring discontinuation of medication or intervention.	Nausea, vomiting, constipation, dizziness, and headache
2	Mild	ADR detected, leading to discontinuation of medication but without need for intervention.	Sedation, extrapyramidal side effects
3	Moderate	ADR detected, requiring medication discontinuation or modification but no need for prolonged hospitalisation.	Weight gain
4	Moderate	ADR is detected and results in prolonged hospitalisation.	Hyperglycaemia
5	Severe	ADR is detected and leads to life-threatening situations or temporary disability.	Hyperlipidaemia
6	Severe	ADR contributes to permanent disability.	Hyperprolactinemia
7	Severe	ADR results in death.	Myocarditis

). It classifies ADRs into seven levels, ranging from mild (levels 1–2) and moderate (levels 3–4) to severe (levels 5–7) [27]. This scale is preferred over other scales for several reasons. Firstly, the Hartwig scale better guides clinical decision-making and patient management by prioritising ADR severity. Secondly, it is more applicable to pharmacovigilance and hospital settings [29]. Thirdly, it is directly relevant to patient safety because it highlights ADRs requiring intervention [30]. This review emphasises levels 3–6. However, level 7 is associated with clozapine administration, which is not used in ASD management [31].

2.1. ADRs Associated with AAPs

2.1.1. Metabolic ADRs

Metabolic ADRs encompass metabolic abnormalities that include hyperglycaemia (fasting blood glucose ≥ 110 mg/dL) and hyperlipidaemia (triglycerides ≥ 100 mg/dL or total cholesterol ≥ 170 mg/dL), which contribute to type 2 diabetes and cardiovascular diseases [13]. Children and adolescents receiving AAPs seem to be more prone to metabolic ADRs compared to adults [32]. In particular, children with ASD are more likely to develop metabolic ADR than adults or children with other mental disorders [33,34]. For example, the prevalence of metabolic ADRs reaches up to 55% with AAPs in ASD patients [35].

A recent umbrella review was conducted by Carnovale et al. (2024) [36], evaluating 23 studies (six systematic reviews, 13 meta-analysis, and four networking meta-analysis) to compare the prevalence of metabolic ADRs among AAPs. It found that no AAP showed a statistically significant risk of hyperglycaemia compared to the placebo, as all 95% confidence intervals (CI) crossed zero, indicating a lack of statical significance However, among AAPs, olanzapine (median difference (medianD) = 4.51, 95% CI = −1.70, 10.72) and aripiprazole (medianD = 3.53, 95% CI = −1.66, 8.84) showed comparatively higher effect size on glucose levels, while ziprasidone (medianD = −5.93, 95% CI = −13.23, 1.37) and lurasidone (medianD = 1.67, 95% CI = −3.50, 6.88) showed comparatively smaller effect [36,37]. However, there was no statistically significant difference among these AAPs. Regarding hyperlipidaemia, olanzapine (standardised mean difference (SMD) = 0.40, 95% CI = −0.01, 0.80) showed the highest increase in triglyceride levels, although this was not statistically significant, followed by quetiapine (SMD = 0.37, 95% CI = 0.06, 0.68), which was statistically significant. In contrast, ziprasidone, followed by aripiprazole and risperidone, showed comparatively smaller effect size for hypertriglyceridemia (SMD = 0.09, 95% CI = −0.25, 0.43), (SMD = 0.06, 95% CI = −0.29, 0.42), (SMD = 0.12, 95% CI = −0.07, 0.32), respectively. However, none were statistically significant, as all 95% CIs crossed zero. With respect to weight gain, olanzapine (odds ratio (OR) = 17.34, 95% CI = 3.97, 75.65) posed the highest estimated risk among AAPs, followed by risperidone (OR = 9.01, 95% CI = 2.10, 38.69) and quetiapine (OR = 8.4, 95% CI = 1.58, 44.82), while ziprasidone (OR = 0.45, 95% CI = 0.04, 5.49), followed by aripiprazole (OR = 2.34, 95% CI = 0.47, 11.64), were associated with the lowest estimated risk [36,38]. However, the wide 95% CIs indicate considerable uncertainty that may be due to a small sample size. A study by DelBello et al. (2022) indicated that olanzapine may be associated with an increased risk of weight gain (OR = 44.81, 95% CI = (11.19, 147.70)), while lurasidone appeared to have a lower risk of weight gain (OR = 0.82, 95% CI = 0.22, 2.13) [36,39]. However, the wide 95% CIs indicate considerable uncertainty. The umbrella review demonstrated a wide range and lack of precision in 95% CI findings, which may be due to heterogeneity resulting from the inclusion of studies that use different effect sizes, such as SMD and OR, making direct comparisons and evidence synthesis more challenging. Additionally, the review does not focus specifically on ASD populations, which contributes to inconsistencies in results. Ultimately, the lack of consideration for key demographic factors such as age, gender, and ethnicity limits the generalisability and applicability of the findings.

2.1.2. Endocrine ADRs

Endocrine ADRs are characterised as unintended and harmful effects on the endocrine system that result in abnormalities in hormone secretion, such as hyperprolactinaemia (for <12 years: >25.4 μg/L; for ≥12 years: >18.4 μg/L for males and >24.1 μg/L for females) and thyroid dysfunction (for 2 to ≤17 years; thyroid-stimulating hormone (TSH) is ≥10 mIU/L and thyroxine (T4) is <0.80–1.8 ng/dL) [40,41].

Hyperprolactinaemia ADRs

Hyperprolactinaemia is studied more extensively compared to thyroid dysfunction. Although children and adolescents have a higher risk for hyperprolactinaemia, its contributions are primarily examined in adults [42,43]. Hyperprolactinaemia may cause serious complications such as galactorrhoea, menstrual irregularities, infertility, and changes in bone density [14]. The prevalence of hyperprolactinaemia ADR with AAP use was 38.5% in children and adolescents with neurological disorders, including those with ASD [44].

AAPs were linked to significantly increased prolactin levels in children and adolescents compared to placebo (mean difference (MD) = 10.10 μg/L, 95%CI  =  6.97,13.23) [42]. AAPs differ in their risk of inducing hyperprolactinaemia. For example, the prolactin level increases by 28.24 μg/L (p < 0.05) and 11.34 μg/L (p < 0.01) in risperidone and olanzapine, respectively [42]. By contrast, aripiprazole significantly decreased prolactin to 4.91 μg/L (p < 0.01). Other AAPs were not associated with significant elevation, included quetiapine (p < 0.06) and lurasidone (p = 0.22) [42]. Although ziprasidone is associated with hyperprolactinaemia, the limitations of the studies make it challenging to provide robust conclusions [42]. The main limitations include a small sample for ziprasidone (16 patients), a predominantly male adolescent population (up to 66%), and although the risk is low and unclear, bias could remain, which may limit generalizability.

Thyroid Dysfunction ADRs

Quetiapine is the most extensively studied AAP and has been confirmed to cause significant hypothyroidism (p < 0.01) with low prevalence [45,46]. Other AAPs lack a significant impact on thyroid function, such as risperidone (p > 0.05) and olanzapine (p > 0.05) [45]. Due to a lack of controlled evidence supporting the positive correlation between AAP and hypothyroidism, several studies did not demonstrate the need for monitoring serum thyroid hormones in patients with normal thyroid levels hormones [46,47,48]. However, the Canadian Alliance for Monitoring Effectiveness and Safety of Antipsychotics (CAMESA) recommended monitoring thyroid hormones for children and adolescents receiving quetiapine [46,49]. For example, patients with normal TSH levels should undergo repeated testing, while those with abnormal TSH levels require ongoing evaluation monitoring [46]. Additionally, the Maudsley prescribing guidelines recommend annual monitoring of thyroid function in patients taking quetiapine, even though the incidence of thyroid dysfunction is low [50]. Contradictory results have been reported regarding the administration of quetiapine and thyroid dysfunction [46]. For instance, studies by Alvarez-Herrera et al. (2020) [51] and Khoodoruth, Abdo, and Ouanes, (2022) [50] reported elevated TSH levels and a decline in T4 with quetiapine. Meanwhile, the study by Findling et al. (2014) [52] observed no change in T4 and elevation in TSH with quetiapine compared to placebo (p = 0.25).

3. Risk Factors Influencing the Prevalence of Metabolic and Endocrine ADRs

3.1. Medication-Related Factors

3.1.1. Mechanism of Metabolic ADRs

AAPs induce metabolic ADRs by affecting neurotransmitter signalling (H₁, D₂, 5-HT₂A) or by impairing the function of molecules that regulate appetite, glucose, and lipid metabolism in the central nervous system (CNS) and peripheral organs, including the liver, pancreas, and adipose tissues (Figure 1) [13,15].

Weight Gain and Hyperglycaemia

Dopamine receptor D₂:

The blockage of D₂ receptors may indirectly induce metabolic ADRs. For instance, D₂ inhibits orexin, a neuropeptide that regulates appetite and energy homeostasis [53]. When AAPs block the D₂ receptor, it causes stimulation of orexin expression, resulting in increased hyperphagia and reduced physical activity [15]. Additionally, the prolactin hormone is regulated by D₂ in the hypothalamus, which affects biological pathways in pancreatic β-cells and adipose tissues [13]. Thus, blocking the D₂ receptor leads to elevated prolactin levels, decreased insulin sensitivity, and increased food consumption by dysregulating glucose metabolism [54]. Taken together, these mechanisms contribute to weight gain and hyperglycaemia [13,15].

5-HT_2C/2A receptors

In the hypothalamus, 5-HT₂C activates Pro-opiomelanocortin (POMC) neurons to express α-melanocyte stimulating hormone (α-MSH), which is a key peptide in the regulation of appetite as well as glucose and lipid metabolism. AAPs block 5-HT₂C, causing inhibition of the POMC/α-MSH (anorexigenic) pathway, leading to hyperphagia and decreased fatty acid oxidation and glucose uptake [13,55]. In contrast, AAPs stimulate the neuropeptide Y (NPY)/Agouti-related protein (AgRP) (orexigenic) pathway to induce hyperphagia and weight gain [13]. Additionally, AAPs block 5-HT₂C in pancreatic β-cells, resulting in a decrease in insulin secretion and hyperglycaemia [13]. Moreover, the blockade of 5-HT₂C activity disrupts leptin regulation and increases food intake, which leads to weight gain [55,56].

In contrast, AAPs block 5-HT₂A in skeletal muscles and indirectly impair glucose transporter type 4 (GLUT4), leading to a reduction in glucose uptake by cells [57,58]. This mechanism causes an increase in insulin resistance and hyperglycaemia [13].

H₁ receptors

One proposed mechanism is that AAPs induce metabolic ADRs through H₁ antagonism, which indirectly interferes with the AMP-activated protein kinase (AMPK) pathway, which regulates appetite in the hypothalamus. AAPs block H₁ receptors in the hypothalamus, leading to the activation of AMPK activity, which increases hyperphagia and weight gain [15,59]. In skeletal muscles, AAPs block H₁ receptors, inducing insulin resistance and impairing insulin signalling by inhibiting phosphoinositide 3-kinase (PI3K), insulin receptor substrate (IRS) proteins, and extracellular signal-regulated kinase (ERK) cascades [15,60]. Furthermore, the activation of the H₁ receptors in pancreatic β-cells increases insulin secretion [61]. Hence, interfering with insulin signalling or secretion by AAPs contributes to hyperglycaemia and elevation of risk of type 2 diabetes [62].

α2-adrenergic and M₁ acetylcholine receptors

AAPs modulate α2-adrenergic receptors activity, subsequently altering sympathetic activity. This contributes to increased glucose production and lipolysis, resulting in short-term hyperglycaemia [58,63]. They also act as M₁ antagonists in the hypothalamus and impair acetylcholine activity in glucose metabolism by inhibiting insulin signalling [64,65].

Hyperlipidaemia

All the underlying pharmacological mechanisms of AAPs induce hyperlipidaemia through an indirect effect via the inhibition of neurotransmitter receptors, leading to disruption of glucose and lipid metabolism and insulin resistance [13,15]. However, AAPs also directly affect biological molecules that regulate lipid metabolism [13]. For example, AAPs directly inhibit the AMPK pathway in the liver or do so indirectly through adiponectin signalling in adipose tissue [66]. This leads to the stimulation of lipogenesis by reducing fatty acid oxidation, resulting in the accumulation of lipids in hepatic tissues and weight gain [15,67]. In contrast, AAPs activate the glucagon receptor (GCGR), which stimulates gluconeogenesis by increasing the expression of phosphoenolpyruvate carboxykinase (PEPCK) and G6Pase (glucose-6-phosphat(ase)) while reducing glucose uptake. The activation of GCGR contributes to hyperglycaemia and insulin resistance [15]. Moreover, in adipose tissue, AAPs stimulate the mammalian target of rapamycin (mTOR), which activates lipid metabolism by enhancing the expression of peroxisome proliferator-activated receptor-γ (PPARγ), leading to increased lipid storage [15]. Additionally, AAPs upregulate sterol regulatory element-binding protein 1/2 (SREBP1/2) in the liver, resulting in increased lipogenesis [68]. AAPs also enhance the expression of IL-2, IL-6, and TNF-α in adipose tissue that activates lipolysis and free fatty acids and causes hyperlipidaemia [69,70].

Olanzapine carries the highest risk of metabolic ADRs, followed by risperidone and quetiapine, and then aripiprazole, ziprasidone, and lurasidone [71,72]. Olanzapine carries the highest risk of metabolic ADRs compared to other AAPs, possibly due to its greatest affinity for H₁ and M₁ receptors. In contrast, lurasidone presents the lowest risk of metabolic ADRs because it lacks activity on the H₁ receptor. Consequently, other biological pathways regulated by H₁, which impact body weight and glucose and lipid metabolism, such as AMPK, PI3K, IRS proteins, and ERK, are not dysregulated by lurasidone [15].

3.1.2. Mechanism of Endocrine ADRs

Hyperprolactinaemia

Hyperprolactinaemia induces menstrual disorders and sexual dysfunction and inhibits bone formation in paediatric patients [41]. This occurs through blockade of D₂ receptors, leading to elevated prolactin levels. Elevated prolactin levels directly inhibit gonadotropin-releasing hormone (GnRH), which in turn decreases the secretion of follicle-stimulating hormone (FSH) and luteinizing hormone (LH) (Figure 2) [42]. Both FSH and LH play crucial roles in menstrual regulation and fertility. Thus, a decline in these hormones contributes to infertility and sexual dysfunction [73]. Additionally, reduced levels of sex hormones (estradiol and testosterone) interfere with bone mineralization and decrease bone density, particularly in adolescents during peak bone development [74]. The risk of hyperprolactinaemia tends to be increased with risperidone due to its highest affinity for D2 (Ki = 2.29 nM) [75].

Thyroid Dysfunction

The underlying mechanism of quetiapine induced thyroid dysfunction remains unclear. However, multiple hypothesised mechanisms have been proposed (Figure 3) [50]. The blockade of D₂ receptors by AAPs may reduce the release of thyrotropin-releasing hormone (TRH) from the hypothalamus, resulting in decreased TSH secretion from the anterior pituitary gland. This leads to decreased triiodothyronine (T3) and T4, causing hypothyroidism [46,50,76]. A major criticism of this hypothesised mechanism, linking D₂ receptor blockade to reduced TRH and subsequent hypothyroidism, is that quetiapine has a low affinity for D₂ receptors (Ki = 567 nM) compared to other AAPs [77,78]. If this mechanism were the primary driver of thyroid dysfunction, one would expect higher frequency for this ADR with risperidone, which has the highest D₂ receptor affinity (Ki = 2 nM) among AAPs [75,77].

Quetiapine has a high affinity for the H₁ receptor, which could disrupt the hypothalamic-pituitary-thyroid (HPT) axis through alternative neuroendocrine pathways, such as the leptin pathway [79]. However, this proposed mechanism has limitations, olanzapine exhibits higher H1 affinity (2.3 nM) than quetiapine (8.7 nM), which would suggest a higher risk of hypothyroidism. Yet, clinical outcomes differ that may be due to pharmacokinetic factors (e.g., higher c_max for quetiapine [1291.4 nM] and differences in tissue distribution compared to olanzapine) and downstream signalling effects rather than receptor affinity alone. Another mechanism associated with quetiapine that induces thyroid dysfunction is the competition between thyroid hormone and quetiapine, and quetiapine being metabolised by uridine diphosphate glucuronosyltransferase (UGT), leading to reduced thyroid hormone levels [76]. This mechanism aligns more closely with olanzapine, for which UGT is one of the primary enzymes involved in metabolism, whereas it plays a secondary role in quetiapine. One further hypothesis considered is that quetiapine stimulates autoimmune thyroiditis, although this has been reported in an inconsistent manner in adults [50,80].

3.1.3. Pharmacokinetic Factors

Risperidone and olanzapine pharmacokinetics (PK) have been the most extensively researched AAP’s in children and adolescents (

Table 2. Summary of pharmacokinetics of AAPs in paediatric populations. Abbreviation: cmax: maximum plasma concentration, F: female, ↑: increase, ↓: decrease, X: no effect, V_d; volume distribution, PPB: plasma protein binding, t_1/2: half-life, UGT: uridine diphosphate glucuronosyltransferase, ^a: not effected by low-fat meal, ^b: no active metabolites, PM: Poor metabolizer, EM: extensive metabolizers.

Drug/PK	Absorption			Distribution		Metabolism	Excretion
	cmax (nM)	Bioavailability (%)	Food Effect	Vd (L/kg)	PPB (%)	Enzyme(s)	t1/2 (Hour)	Route
Risperidone	36.5	70	X	1–2	90	CYP2D6 -With CYP2D6-inhibitors: ↓ by 55%	Risperidone and 9-OH-risperidone: 20 -PM: 20 -EM: 3	Urine: 70% Faeces: 14%
Aripiprazole	240.8 -F: ↑ by 11%	87	X	~5	99	CYP2D6	Aripiprazole: 75 Dehydroaripiprazole: 94 -PM: 150	Urine: 30% Faeces: 55%
Olanzapine	48	≥65	X	21	93	CYP1A2 CYP2D6 UGT1 -F: ↓ by 30% -With CYP1A2 inducer (smoker): ↑ by 40%	30–52 b	Urine: 57% Faeces:30%
Quetiapine	1291.4	10	X a	6–14	83	CYP3A4 -Hepatic impairment: dose ↓ by 0.1–0.5-fold	Quetiapine: 6 -N-desalkyl quetiapine: 12 -CYP3A4 inhibitor: ↑ by 92%	Urine: 73% Faeces: 20%
Lurasidone	60.9 -CYP3A4 inhibitor: ↑ by 7 times	19 September	↑ by 2–3 times	2.4–20	99	CYP3A4 -CYP3A4 inhibitor: contraindication	lurasidone: 29 ID-14283: ↓ by 2.5 times	Urine: 9% Faeces: 80%
Ziprasidone	121 -F: ↑ by 25% -CYP3A4 inhibitor: ↑ by 34%	60	Increased	1.5	99	glutathione and aldehyde oxidase CYP3A4 (lesser extent) -CYP3A4 inhibitor: monitoring therapy	5	Urine: 20% Faeces:66%

) [81].

Risperidone

A single oral dose (2 mg) of risperidone is rapidly absorbed and exhibits 70% bioavailability (%F), regardless of food intake. Risperidone reaches maximum plasma concentration (c_max = 36.5 nM) within one hour and three hours for the active metabolite (9-OH-risperidone) [82,83,84]. It is highly protein bound (90%), with a volume distribution (V_d) ranging between 1 and 2 L/kg. It exhibits linear pharmacokinetics within its therapeutic dose [56]. Risperidone is primarily metabolised by CYP2D6, converting it to its active metabolite (9-OH-risperidone). The half-life (t_1/2) for risperidone and 9-OH-risperidone is 20 h. The kidneys primarily excrete risperidone, and its clearance significantly declines in patients with renal impairment, causing drug accumulation and an increased risk of ADRs [85].

Generally, in children aged 4 to 15 years receiving risperidone, the pharmacokinetics are similar to those observed in adults [86]. However, the concentration-to-dosage (C/D) ratio of risperidone and 9-OH-risperidone in children is lower than in adolescents and adults [81]. Compared to adults, protein binding is 16% lower in children and adolescents. As a result, the risk of ADRs may be higher in the paediatric population due to an increased free fraction (26%) of risperidone [87].

The pharmacokinetics of risperidone are influenced by several factors, including concomitant medications and CYP2D6 phenotype. The co-administration of risperidone with CYP2D6 inhibitors (e.g., fluoxetine, valproate) reduces the metabolism of risperidone and increases the level of exposure by up to 55%, resulting in an increased risk of ADRs [81]. Moreover, the metabolism of risperidone is influenced by the CYP2D6 phenotype, which may lead to variability in the prevalence of ADRs due to differences in drug metabolism activity across populations and ethnicities [81,88,89]. In the study by Biswas, Vanwong, and Sukasem (2022) [31], weight gain and hyperglycaemia ADRs are reported among poor metabolizers (PMs). Therefore, dose adjustment of risperidone should be considered for patients who are PMs or those who concomitantly take CYP2D6 inhibitors [88].

Aripiprazole

Aripiprazole is well absorbed with or without food, showing an 87% bioavailability [90]. It exhibits linear pharmacokinetics at therapeutic doses, reaching c_max, which is approximately 240.8 nM, within 3–5 h after 2–15 mg of oral aripiprazole [91,92]. It has a V_d of around 5 L/kg with 99% plasma protein binding, predominantly to albumin [56]. It is mainly metabolised by CYP2D6, producing the active metabolite (dehydroaripiprazole) and inactive metabolite (aripiprazole metabolites-1451) [81,93]. The aripiprazole t_1/2 is 75 h and becomes longer by 20% in dehydroaripiprazole. Aripiprazole is excreted mainly via faeces (55%) and urine (30%) [56].

Aripiprazole pharmacokinetics are comparable between children aged 10–17 years old and adults [94,95]. Nonetheless, factors such as gender and enzyme phenotype play a crucial role in the variability of aripiprazole pharmacokinetics [96,97]. Regarding gender differences, at an equivalent therapeutic dose of aripiprazole, the c_max was 11% higher in females than in males (p < 0.05) [97].

There is no significant correlation between the concentration of aripiprazole in plasma and the co-administration of CYP2D6 inhibitors. However, this concentration is influenced by the CYP2D6 genotype [98,99]. For example, the t_1/2 is doubled in PM, reaching 150 h compared to extensive metabolizers (EMs) [93,96].

Olanzapine

Olanzapine is well absorbed after oral administration, and its absorption is not affected by food. It exhibits linear pharmacokinetics, with a bioavailability of ≥65% [32,100,101]. Its c_max is 48 nM after oral administration [102]. It binds approximately 93% of albumin and α1−acid glycoprotein and has a V_d of 21 L/kg [103]. Regarding its metabolism, CYP2D6, CYP1A2, and UGT1 convert olanzapine into three inactive metabolites, including 2-hydroxymethylolanzapine, N-desmethylolanzapine, and 10′/4′-N-glucuronides, respectively [100,103,104]. The t_1/2 of olanzapine ranges from 30 to 52 h, with an average clearance rate of 25 L/h. Up to 57% of olanzapine is renally excreted, while the remaining amount is excreted through faeces [100,103].

Factors such as age, gender, and concomitant medications that may affect the pharmacokinetics of olanzapine have shown conflicting results. Studies indicated that the plasma concentrations of olanzapine differ between adolescents and adults [32,81,101,105]. In contrast, several studies reported that a negative association was identified between age and plasma concentration of olanzapine [81,106]. These divergent outcomes may result from overlooking the factors influencing the pharmacokinetic profile. For instance, the t_1/2 and c_max in adults are comparable to those observed in children aged 10–18 years old [107]. By contrast, clearance is reduced by 48% in younger children (those under 12 years old), leading to a 43% increase in olanzapine exposure compared to adults [101,108]. Additionally, females tend to have lower CYP1A2 activity than males, contributing to a 30% reduction in clearance. Consequently, both children and females may require a lower dose of olanzapine than adult males [32,81]. On the other hand, the co-administration of CYP1A2 inducers, such as cigarette smoking, increases olanzapine clearance by 40% in patients who smoke tobacco or e-cigarette (vaping) [109,110].

Quetiapine

Oral quetiapine is rapidly absorbed and is not affected by a low-fat meal, showing approximately 10% bioavailability and demonstrating linear pharmacokinetics [111,112]. Following 300 mg of quetiapine, the c_max was observed to be 1291.4 nM [113]. Its V_d varies from 6 to 14 L/kg, and albumin binding reaches 83% [111,114]. It is extensively metabolised by CYP3A4, which converts it to the active metabolite. N-desalkyl quetiapine [115]. The t_1/2 of quetiapine is approximately 6 h, with an elevation of 12 h for N-desalkyl quetiapine [114]. Quetiapine is excreted through urine and faeces, accounting for 73% and 20%, respectively [111].

The influence of age on quetiapine pharmacokinetics has been reported in several studies [81,116,117]. For example, the plasma concentration of quetiapine is 41% higher in adults compared to children aged 10–17 years old [117]. In contrast, in the study by McConville et al. (2000) [118], the t_1/2 and c_max in adults are similar to those observed in children aged 12–17 years. Specifically, this discrepancy between these findings may be due to variations in study design, sample size, and age range of participants. The strength of this study is due to the systematic, population-based pharmacokinetics data across a wide paediatric age range. Meanwhile, the open-label trial conducted by McConville et al. (2000) [118] had smaller sample sizes that followed DSM-IV criteria and provided limited pharmacokinetic detail, making them less suitable for modelling or dose optimisation.

The pharmacokinetics of quetiapine are influenced by concurrent medications, CYP3A4-dependent factors, and hepatic impairment [119,120,121]. The administration of 400 mg/day of quetiapine with CYP3A4 inhibitors (e.g., erythromycin) increases its t_1/2 by 92%. Therefore, a six-fold dose reduction is essential for the co-administration of quetiapine with CYP3A4 inhibitors [120,121,122]. The quetiapine dose is reduced by 10% to 50% due to decreased clearance in patients with hepatic impairment [121]. There is a lack of evidence to support a gender difference with quetiapine [123].

Lurasidone

The absorption of lurasidone increases by 2 to 3 times when taken with food [124]. Its c_max is 60.9 nM after a 20 mg dose with bioavailability varying between 9 and 19% due to first-pass metabolism [124,125,126]. Lurasidone binds approximately 99% to plasma proteins, with V_d varying between 2.4 and 20 L/kg [127]. It is primarily metabolised by CYP3A4, producing two active metabolites, namely exo-hydroxy metabolite ID-14283 and endo-hydroxy metabolite ID-14326 [115]. The average t_1/2 of lurasidone is 29 h as the parent drug and becomes 2.5-fold shorter in ID-14283. Approximately 80% of lurasidone is excreted in faeces, while 9% is excreted in urine [128].

Numerous studies show that pharmacokinetics (e.g., plasma concentration) negatively correlates with age [129,130,131]. However, both gender and concurrent medication influence its pharmacokinetics. In the study by Yang et al. (2022b) [132], it was reported that the area under the curve (AUC) of lurasidone is approximately 15% higher in females than in males. In relation to concurrent medications, prescribing lurasidone (40 mg/day) with CYP3A4 inhibitors (e.g., ketoconazole 400 mg/day) increases AUC by approximately ten times and C_max by seven times compared to lurasidone alone [72,133]. Consequently, it is advisable to avoid co-administration of ketoconazole with lurasidone and switch to an alternative antifungal [72].

Ziprasidone

Ziprasidone is well absorbed, and its c_max reaches 121 nM within 6 h after a 20 mg dose [134,135]. The bioavailability of ziprasidone is approximately 60% and increases to 100% when consumed with food [136]. It is 99% bound to plasma albumin and α1 acid glycoprotein. It has a V_d of ~1.5 L/kg. The majority of ziprasidone’s metabolism is mediated by glutathione and aldehyde oxidase, while CYP3A4 plays a lesser role [72,115]. Ziprasidone’s t_1/2 is approximately 7 h, with the main route of elimination being faecal (66%), while the remainder is excreted in the urine [137].

The pharmacokinetics of ziprasidone in children are comparable to those in adults [81,138]. Several factors, including gender and polypharmacy, play a critical role in ziprasidone pharmacokinetics. The c_max of ziprasidone is approximately 25% higher in females than in males, which is statistically significant (p  <  0.05) [123]. Despite CYP3A4 having a minor effect on ziprasidone metabolism, the co-administration of 400 mg/day of ketoconazole significantly increases the AUC and c_max by approximately 34% for a 40 mg/day dose of ziprasidone (p < 0.05) [72,139].

3.2. Patient-Specific Factors

3.2.1. Age

Children and adolescents, particularly those with ASD who are administered AAPs, have a greater prevalence of metabolic and endocrine ADRs compared to adults [32,34]. There are multiple reasons that may explain why paediatric populations are more likely to develop these ADRs. As noted, their physiological systems for metabolising and eliminating medications are not as fully developed as those in adults [81,140].

In children, gastric pH is initially higher and gradually decreases to reach adult levels in the second year of childhood due to an immature digestive system. This can affect the absorption of AAPs [141]. For distribution, body composition changes by decreasing extracellular water and increasing fat storage throughout childhood. This causes a continuous change in V_d of drugs and t_1/2, necessitating adjustments in dosage for AAPs [101,142]. In addition, paediatric populations have lower plasma protein levels compared to adults. For example, AAPs are highly protein-bound, leading to increased free AAP concentrations and a higher risk of ADRs in children [141].

In addition, historically, children have been insufficiently included in clinical trials, which results in limited paediatric studies and creates challenges in determining safe and effective doses for children [141,143]. Furthermore, during puberty, these differences become more pronounced as hormonal fluctuations influence the concentration and clearance of the drug [81,141].

3.2.2. Gender

With respect to gender variation, female adolescents who are receiving risperidone or aripiprazole are more prone to metabolic ADRs, particularly weight gain, than males [144]. Furthermore, a statistical difference in the prevalence of weight gain (i.e., the increase in body mass index (BMI) was 1.3–1.5 kg/m²) was observed between males and females receiving AAPs (i.e., aripiprazole, olanzapine) (p < 0.001) [97,145]

Generally, the prevalence of endocrine ADRs, particularly hyperprolactinaemia, is 20% higher in females than in males who are receiving AAPs (i.e., risperidone, olanzapine, quetiapine, ziprasidone, and lurasidone), which may be due to the estrogenic effect on prolactin levels [42]. Another study showed that females receiving AAPs (i.e., aripiprazole, olanzapine, and quetiapine) experienced a higher rate of thyroid dysfunction ADRs compared to males, although the difference was not statistically significant (p  = 0.134) [146].

There are several hypotheses for the variation in the prevalence of ADRs between males and females [147]. In the past, females were not included in clinical trials due to concerns about the teratogenic risk to those who could become pregnant [148]. Additionally, females typically have lower body mass and a higher proportion of adipose tissue, which influences the absorption and distribution of medications [149]. Furthermore, females exhibit a lower glomerular filtration rate (GFR) and P-glycoprotein (P-gp) renal efflux of medication compared to males, resulting in a prolonged t_1/2 [150,151]. Moreover, females report ADRs more often than males [152]. Therefore, gender differences should be considered in guidelines to reduce the risk of ADRs.

3.2.3. Ethnicity

Ethnicity also plays a crucial role in determining the prevalence of metabolic and endocrine ADRs among AAPs. The Caucasian group is overrepresented compared to other ethnicities, such as Middle Eastern populations, which limits the generalizability of these study findings [153].

This lack of diversity is particularly significant when considering pharmacogenetic factors that influence drug metabolism. AAPs are primarily metabolised by three key enzymes: CYP2D6, CYP3A4, and CYP1A2 [154]. The activity of these enzymes varies across ethnic groups due to differences in the frequency of genetic variants, which can, in turn, affect the efficacy and safety profiles of AAPs [129].

Regarding CYP2D6, the metabolites OH-risperidone and dehydroaripiprazole decrease approximately two times in PMs and one and a half times in intermediate metabolizers [155,156]. A cohort study showed that CYP2D6 PMs of risperidone are most prevalent in Oceania (21%). In contrast, in the Middle East, CYP2D6 PMs rate is around 12.7%, followed by Europe (9.3%) and America (8.4%). While CYP2D6 PMs of aripiprazole are most prevalent among Europeans at 6.1%, the rate in Americans is 3.7%, and in Middle Easterns, it is 1.2% [157]. This indicates that the risk of risperidone ADRs is most frequently observed in Oceania, followed by the Middle East, Europe, and America. In contrast, the risk of ADRs associated with aripiprazole appears to be higher in Europe than in the Middle East populations. According to the Dutch Pharmacogenetics Working Group (DPWG) guideline, in CYP2D6 PMs, the dose of risperidone and aripiprazole should be prescribed at 50–67% of the therapeutic dose to reduce the risk of ADRs. In contrast, in ultrarapid metabolizers, risperidone should be switched to other medications to prevent treatment failure [157]. Interestingly, the rate of CYP2D6 PMs in the East Asian population is between 0.9% and 1.2% for aripiprazole and risperidone, respectively. This may explain why the therapeutic dose of risperidone in East Asians is reduced by 26% compared to Europeans (around 50%) [157,158,159].

With reference to CYP3A4 PM, quetiapine plasma concentrations were observed to be more than three times higher. Therefore, they need a dose reduction to less than a third or a switch to an alternative medication with lower CYP3A4 dependence based on clinical use [119,154]. N-Desalkyl quetiapine exhibits antidepressant effects and is present in reduced concentrations in PM. Therefore, switching to an alternative drug is highly recommended for treating depression in PMs [154].

In terms of CYP1A2, there is no significant correlation between olanzapine exposure and its phenotype. Numerous clinical studies did not observe any difference in the prevalence of ADRs across different populations [100,160,161]. Therefore, the DPWG recommended that no further action be taken [154].

3.3. Healthcare System Factors

The differences in strategies and policies for reporting and managing ADRs are among the key factors that contribute to variations in ADR prevalence between countries [162]. For instance, the MHRA partnered with patient organisations to develop pharmacovigilance guidelines and enhance patient awareness [163]. Similarly, the SFDA collaborated with the WHO to enhance ADR reporting among patients [164]. In contrast, in some countries, the absence of guidance for patients on reporting ADRs has led to a decline in the number of reported cases [165]. Moreover, some countries, such as the UK, standardise ADR reporting for both medications and herbal medicines, while Taiwan uses different forms that may delay responses from healthcare professionals (HCPs) [163]. Furthermore, the use of information technology in ADR reporting has increased to improve the validity and reliability of ADR data and to speed up response times for curating, managing, and analysing ADRs [166]. However, this facility remains limited in low-income countries, resulting in a misrepresentation of ADR prevalence globally [167].

3.4. Disease-Related Factors

As described, ASD is a heterogeneous disorder that produces varying severity of symptoms and sensitivity to medications, resulting in increased susceptibility to ADRs [168,169]. In addition, the increased risk of ADRs in ASD patients may occur due to several reasons. Children and adolescents with ASD are typically prescribed more than one medication to manage ASD along with comorbidities, which increases the risk of drug–drug interactions (DDIs) and the likelihood of ADRs [170]. For example, in Ritter, Hewitt and McMorris (2021) [171] study indicated that the rate of concurrent medications in this population reached up to 87%. Another reason is due to gut microbiota imbalance in ASD patients, which affects the absorption of medications and alters their pharmacokinetics [172]. Additionally, most ASD patients follow restricted diets (e.g., gluten-free), which contributes to a deficiency in co-factors (e.g., zinc, selenium) that negatively affects the metabolism of medications and alters their safety profiles [173]. Notably, ASD is linked to thyroid and adrenal dysfunction, which may potentiate endocrine-related ADR [174]. Similarly, a relationship exists between ASD and developed metabolic ADRs [33]. HCPs and caregivers also face difficulties differentiating between ADRs and ASD symptoms, such as weight gain due to physical inactivity [175,176,177]. However, there is currently no standardised framework for assessing and controlling this factor, which may contribute to variability and inconsistent ADR reporting across studies. This lack of standardisation highlights the need for an ASD-specific approach to evaluate ADR susceptibility rather than relying solely on data from the general population of mental disorders and overlooking the potential impact of ASD on ADR prevalence [178].

4. Research Gap and Future Directions

With the increasing global prevalence of heterogeneous diseases such as ASD and the use of AAPs as first-line treatments, the need for guidelines that consider population variation has become essential [157]. Until now, the DPWG, Clinical Pharmacogenetics Implementation Consortium (CPIC), and CAMESA have been the only organisations developing monitoring guidelines that considered individual variation within the population [49,56,154].

Over the last five years, only four studies Al-Huseini et al. (2022), Alsabhan et al. (2024), Shilbayeh and Adeen (2023), and Makary et al. (2023)—[35,179,180,181]—from the Middle Eastern regions have examined the prevalence of metabolic or endocrine ADRs associated with AAPs. Furthermore, thyroid dysfunction remains unexplored in the Middle East and is rarely investigated in other regions [50]. However, current evidence remains limited, particularly regarding population-specific ADR prevalence in paediatric patients with ASD, especially within the Middle Eastern population. This underscores the urgent need for further research and the development of tailored monitoring guidelines [157].

5. Clinical Implications

The study of the prevalence of metabolic and endocrine ADRs associated with AAPs in paediatric populations with ASD across different countries improves medication safety. Risk assessment helps identify high-risk populations and highlights variability in ADRs due to ethnicity, age, gender, disease, and healthcare-related factors [182,183]. As a result, this study encourages personalised treatment approaches and supports the development of tailored monitoring guidelines based on regional and individual risk profiles, including age and specific disease [184]. Additionally, it helps HCPs select safer AAPs based on the prevalence of ADRs in susceptible individuals based on their unique characteristics, thereby improving prescribing practices [185]. This study also contributes to the early detection and optimised monitoring of ADRs to prevent long-term complications, such as type 2 diabetes and cardiovascular diseases [186]. Moreover, it underscores the impact of healthcare infrastructure (e.g., spontaneous reporting systems) on the reporting and management of ADRs [163].

6. Conclusions

AAPs have remained the first line of treatment for paediatric patients’ irritability with ASD due to their efficacy. However, AAPs use in ASD is associated with metabolic and endocrine ADRs that contribute to public health concerns. Although data on the prevalence of these ADRs inform current monitoring guidelines, such guidelines often rely on findings from specific populations (e.g., UK, USA) without accounting for inter-individual diversity factors such as ethnicity and age.

In the future, tailored monitoring guidelines for paediatric patients with heterogeneous disorders like ASD, particularly in underrepresented populations such as those in the Middle East, should be considered in clinical practice and research. This would support developing more personalised therapies, aiming to avoid both over-monitoring of rare or non-serious ADRs and under-monitoring serious ADRs not currently included in guidelines. Taken together this would promote the safer use of AAPs in paediatric populations with ASD.