The Impact of Caregiving for Children with Chronic Conditions on the HPA Axis: A Scoping Review

Abstract

Caregiving has been robustly linked to caregiver health through the dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis in the context of caregiving for an adult with a chronic illness. However, little research examines the physiological impact of caregiving for a child with a chronic illness despite high burden and unique stressors. In this review, we explore the links of caregiving for a child with a congenital, chromosomal, or genetic disorder to the regulation or dysregulation of the HPA axis. A search was conducted in PubMed, Embase, and the Web of Science and 15 studies met inclusion criteria. Overall, there were inconsistent links of caregiving to HPA axis functioning, perhaps due to the heterogeneity across disease contexts, study designs, and biomarker measurement. Future research should standardize measurement and study designs, increase participant diversity, and examine moderators of the links of caregiving to the HPA axis.

A nationally representative survey conducted in 2020 in the U.S. suggested more than 1 in 5 adults report providing informal care to an adult or child with health or functional limitations (National Alliance for Caregiving & AARP, 2020). The term “informal caregivers” refers to those individuals who provide regular, unpaid care or assistance to a family member or friend with a disability or illness, as opposed to formal caregivers who are paid to take on this role (CDC, 2020; Toledano-Toledano et al., 2019). In 2020, 53 million adults identified as informal caregivers, an increase from 2015 when an estimated 43.5 million adults reported providing informal care to another (National Alliance for Caregiving & AARP, 2020). Given these increasing prevalence rates, it is critical to understand the impact of the caregiving process on caregivers’ health and well-being.

There are a number of models that outline how caregiving-related behaviors can accumulate to result in high stress and burden for caregivers. The stress process model is one such conceptualization that states caregiving-related stressors may impact health and well-being (Pearlin et al., 1990). This model posits that poor health outcomes result from the stressors associated with the caregiving experience, including both primary stressors that derive directly from providing care (e.g., perceived caregiver burden) as well as secondary stressors that arise from these initial primary stressors (e.g., financial strains, family conflict) (Manalel et al., 2022; Pearlin et al., 1990). Initially developed in the context of Alzheimer’s disease and related dementias (ADRD; e.g., Pearlin et al., 1981; Zarit et al., 1980; Zarit et al., 1986), studies of caregiver stress and burden often explore the psychological impacts experienced by informal caregivers. For example, past research consistently shows that caregivers of individuals with ADRD report higher rates of depressive symptoms (Schulz & Williamson, 1991), anxiety (Parks & Pilisuk, 1991) and lower health-related quality of life (Markowitz et al., 2003) compared to non-caregivers. These findings are not limited to informal caregiving in an ADRD context either, as caregivers of adults with other chronic illnesses also self-report elevated psychological distress (see del Pino-Casado et al., 2019 for a review). While we note that the majority of past research suggests caregiving for an adult with a chronic illness such as ADRD is linked to negative health and well-being outcomes, there are some studies that find caregiving can be linked to positive outcomes—primarily psychological processes—such as benefit-finding (e.g., Cohen et al., 2002; Tarlow et al., 2004).

In addition to the psychological toll of caregiving, the stress associated with caregiving can detrimentally impact caregivers’ physical health as well. Caregivers often report a range of physical symptoms, including lack of appetite, headaches, dizziness, and cold-like symptoms (see Pelentsov et al., 2015 for a review), as well as backaches, insomnia, arthritis, gastrointestinal issues, hearing problems, and memory problems (Aldwin et al., 2007). Caregivers also report poor diet, alcohol consumption, nicotine use, lack of exercise, and missed personal health appointments (Aldwin et al., 2007; Beach et al., 2000; Schulz & Sherwood, 2008). Over time, these behaviors can increase risk for more distal, chronic health conditions such as cardiovascular, metabolic, and immunological problems (Aldwin et al., 2007). Further, caregivers of children with a life-limiting condition have higher incidence rates of psychological and physical health conditions, as well as higher rates of mortality compared to caregivers of typically developing children (Fraser et al., 2021).

Caregiving and the HPA Axis

One way the psychological stress of caregiving can impact downstream physical health outcomes is through the body’s physiological stress response. One of the main components of the physiological stress response is the hypothalamic-pituitary-adrenal (HPA) axis, which connects the hypothalamus, pituitary, and adrenal glands via hormonal signals. The HPA axis contains a network of neurohormones, neurons, and steroids and is regulated by glucocorticoids and cytokines that manage physiological mechanisms of stress (Spencer & Deak, 2017). The primary biomarkers of the HPA axis include corticotropin releasing hormone (CRH), adrenocorticotropic hormone (ACTH), arginine vasopressin (AVP), cortisol, and dehydroepiandrosterone (DHEA) (see Figure 1). First, CRH—a neuropeptide secreted by the hypothalamus—and AVP—a hormone produced by the posterior pituitary gland—signal the anterior pituitary gland to release ACTH (Lovell & Wetherell, 2012). ACTH then travels into the bloodstream to the adrenal glands which drive the production of cortisol (Spencer & Deak, 2017). Cortisol is a common glucocorticoid hormone that is responsible for a number of physiological effects such as redirection of cellular processes including increasing blood pressure and cardiac output as well as diverting the blood supply away from processes not critical during periods of high stress (e.g., immune responses, digestive processes) (Kamin & Kertes, 2017; Kudielka & Kirshbaum, 2005; Stephens & Wand, 2012). Finally, DHEA—a steroid secreted by the adrenal cortex—functions as a buffer to the potential harm caused by increased levels of cortisol circulating throughout the body (Kamin & Kertes, 2017). In sum, each of these biomarkers both regulates and provides critical information regarding the overall function of the HPA axis, one system through which caregiving stress can “get under the skin” to impact physical health.

Figure 1.

Caregiver Stress and the HPA Axis

While activation of the HPA axis is adaptive during instances of acute stress, chronic stress and subsequent dysregulation of the HPA axis can lead to negative distal health conditions. Consistently elevated cortisol levels may lead to a suppressed immune response and symptoms that mimic illness, which can lead to further susceptibility to chronic health conditions (Holsboer, 2000; Morey et al., 2015; Reagan et al., 2008). For example, repeated HPA axis activation has been linked to chronic conditions such as type 2 diabetes, obesity, and cardiovascular disease (Adam et al., 2017; Matthews et al., 2006). Continual activation of the HPA axis can also lead to glucocorticoid resistance (GCR) during which prolonged periods of elevated cortisol limit the body’s ability to recognize the appropriate time to deactivate the inflammatory process (Cohen et al., 2012). Furthermore, over-activation of the HPA axis can result in patterns of either hypercortisolism or hypocortisolism, which can impact multiple bodily systems. Such patterns of HPA axis dysregulation may repress the expression of pro-inflammatory cytokines by immune cells (Chrousos, 2009; Cruz-Topete & Cidlowski, 2015) and result in visceral adipose deposition, insulin resistance, hyperlipidemia, cardiovascular disease, and hypertension (Adam et al., 2017; Chrousos, 2009). Thus, proper functioning of the HPA axis is crucial in responding to stressors in an individual’s environment, but chronic stress has the potential to disrupt the HPA axis in detrimental ways.

Given the role of the HPA axis in responding to stress, it is critical to understand how caregiving specifically impacts caregivers’ HPA axis given the high prevalence rates and chronic stress associated with caregiving. Past research suggests informal caregiving is indeed disruptive to the HPA axis—at least in the context of individuals caregiving for adults with chronic conditions. For example, multiple reviews of informal caregivers caring for a family member with ADRD found clear links of caregiving stress to endocrine function dysregulation, such as elevated cortisol (see Allen et al., 2017; Lavretsky, 2005 for reviews). Research also suggests that even though differential caregiving roles may be associated with differential levels of physical burden (e.g., helping a care recipient in and out of bed vs. providing medication reminders), informal caregivers of spouses with ADRD experienced elevated cortisol regardless of the type of caregiving provided or physical burden experienced (Davis et al., 2004).

Caregiving in a Pediatric Context

While it is clear the HPA axis is critical in maintaining caregiver health, most of the research investigating the impact of caregiving stress on the HPA axis has been conducted in caregivers of adults with chronic health conditions. Additionally, the overwhelming majority of existing literature that does focus on caregivers within a pediatric context examines the psychosocial aspects of stress and burden—often using survey methodology (e.g., Cohn et al., 2020; Dijkstra-de Neijs et al., 2020)—and does not measure biological stress (e.g., biomarkers of HPA axis). Thus, there is a clear need to examine the impact of caregiving stress within a pediatric context because this type of stress may differ from adult caregiving in ways that may have important implications for traditional stress and coping frameworks.

First, from a life-course perspective, caregiving in the pediatric context may present a particularly potent stressor because a childhood disease is less normative than chronic illness in older adulthood (Berg et al., 2021). Second, the cumulative cost and burden of caregiving in pediatric conditions is high, as caregivers of children with chronic conditions often continue to be the primary caregiver of these individuals well into adulthood, as caregivers themselves reach old age; in this context, care is a lifetime commitment that includes changes in income and domestic responsibilities, as well as chronic stress (Gilligan et al., 2017; Pelentsov et al., 2015). Additionally, research indicates that less normative stressors may require different coping behaviors than more normative stressors, which may impact the ways in which stress influences health (Wrosch & Freund, 2001).

In addition to the high stress associated with caring for a child with a chronic illness, there is an increasing need to focus research on this population due to high prevalence rates. In the United States, about 40% of school-aged children suffer from a chronic health condition (CDC, 2021) – totaling about 29 million children who are cared for by informal caregivers. Additionally, it is estimated that 15 million children with chronic health conditions are affected by rare diseases, about 80% of which are estimated to have a genetic etiology (Government Accountability Office, [GAO], 2011). These disorders require additional responsibilities and add stress to caregivers, as caregivers must support activities of daily living, perform complex medical tasks, and advocate for healthcare and educational needs among other services (National Alliance for Caregiving & AARP, 2020); caregivers of children with such conditions therefore become experts in both the condition itself and specialized care tasks.

Current Study

In this paper, we seek to examine the impact of chronic stress on the HPA axis—specifically the stress associated with caring for a child with an inborn condition or illness. We focus on the cumulative and chronic stress associated with caregiving given the robust literature that demonstrates potentially deleterious health effects of caregiving. We specifically examine conditions that are congenital, chromosomal, or genetic in nature because of the high burden and unique stress placed on these caregivers. Conditions that are congenital, chromosomal, or genetic in nature present early in life—often identified at birth—and are associated with high caregiving burden with numerous responsibilities (Moretti et al., 2021), high healthcare costs and longer hospitalizations (Miller et al., 2020), and a lifetime of care (Gilligan et al., 2017). Additionally, given the genetic etiology and presentation early in life, it may be the case that caregivers may have the same condition themselves or may have other children with the disorders. Given these unique stressors associated with the experience of caring for a child with a congenital, chromosomal, or genetic condition we seek to review studies that examine the impact on a caregiver’s HPA axis.

To accomplish our goal of exploring the impact of caring for a child with a congenital, chromosomal, or genetic condition on the HPA axis, we utilize a scoping review format in order to chart the current state of the literature and identify gaps to inform future research. Whereas a systematic review is designed to answer clearly defined questions or assess the quality of a given body of literature, a scoping review is indicated when the question is broad or seeks to summarize findings from a body of literature that is heterogenous and diverse (e.g., Tricco et al., 2018). It can also be used as a preliminary step if it is not clear if a systematic review is warranted. Therefore, given the variability in inborn conditions, we first review the congenital, chromosomal, and genetic conditions of the care recipients and caregiving contexts that have served as the focus of past research. It is important to understand which conditions have served as the focus in past research to understand the caregiving context and how various illness management requirements may impact the HPA axis. We also review relevant contextual factors that may impact the caregiving process such as whether studies examine gender (e.g., only mothers included), the geographic location where the study was conducted, and other relevant demographic variables (e.g., additional children with the same inborn condition) that impact caregiving.

Next, we seek to summarize the methodological designs of current studies that assess the impact of caring for a child with a congenital, chromosomal, or genetic condition given the variety of potential methods to assess HPA axis function. Specifically, we review the aspects of the HPA axis that have been examined in this context, and the study designs utilized by past research—which includes the collection type, biomarker calculation, and more general assessment design. Given that the HPA axis regulates physiological mechanisms of stress, it is important to understand the potential biomarker pathways associated with chronic caregiver stress and the various roles these biomarkers play in the stress process. For example, some biomarkers (e.g., AVP) stimulate the HPA axis whereas others reflect regulation (or dysregulation) of the HPA axis (e.g., cortisol; Spencer & Deak, 2017). We also examine various study characteristics including study design and biomarker measurement assessment. Because there are a variety of data collection methods (e.g., blood vs. salivary markers) we chart the collection types and methods, again in line with the goal of a scoping review (Arksey & O’Malley, 2005; Munn et al., 2018). Specifically, we examine both the method of measurement (e.g., cortisol awakening response (CAR) vs. cortisol/DHEA) and number of assessments over time (e.g., a single assessment within a cross-sectional study design vs. multiple assessments through a longitudinal study design). We examine specific biomarker calculation because even within type of biomarker, certain calculations have different meanings—for example, there are numerous calculations of cortisol used in the literature that may provide a variety of information regarding an individual’s health (see Supplemental Table 1 for more details). In doing so, we seek to summarize and examine patterns across currently available information while identifying consistencies/inconsistencies across design, measurement type, and caregiving context.

In sum, caregiving has been identified as a potent stressor that impacts a caregivers’ physical health through the dysregulation of the HPA axis. Past studies have robustly linked caregiving stress to HPA axis function, but this work is primarily conducted in caregivers of adults. Caring for a child with a congenital, chromosomal, or genetic condition is a particularly stressful experience given the disease context, high healthcare costs, unique requirements for care, and presentation early in life. Thus, this scoping review seeks to assess the current state of the literature related to caregivers of children with congenital, chromosomal, or genetic diseases to establish both what is currently known in this domain as well as guide future research.

Methods

Protocol and Registration

A protocol was developed a priori and can be viewed in Supplemental Materials Appendix A. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) was used for reporting this scoping review (2018). A full table detailing the checklist and where we report information on each criterion in the manuscript can be viewed in Supplemental Table 2. We also followed the five-stage scoping review methodology framework outlined by Arksey and O’Malley (2005).

Eligibility Criteria

The inclusion criteria for studies were as follows: 1) focused on caregivers and parents of children with chromosomal, congenital, or genetic disorders; 2) measured HPA axis function through naturally occurring ACTH, AVP, cortisol, CRH and/or DHEA/DHEA-S because they are the main biomarkers of the HPA axis; 3) published in English within the 20-year window; 4) included original data; and 5) were published in peer-reviewed journals. We include studies that both directly assess caregiving stress as well as those that simply compare caregivers to non-caregiving controls in order to garner a more comprehensive set of manuscripts. Articles were excluded if they did not fit the inclusion parameters, if full text could not be accessed, or if the manuscript was written in a language other than English.

Information Sources and Search

A search was conducted by the National Institutes of Health (NIH) Librarian (author AL) in three databases: PubMed (US National Library of Medicine), Embase (Elsevier), and the Web of Science: Core Collection (Clarivate Analytics). Articles published between the 20-year window of January 2002 and January 2022 (month determined by date of search) were considered based on the search terms used in Table 1. A combination of keywords and controlled vocabulary terms (e.g., Medical Subject Headings [MeSH], EMTREE) were used for each concept of interest (e.g., caregiving/caregivers, HPA axis, pediatrics, chronic disease/genetic disorders/chromosome disorders). Searching for chromosomal, congenital, and genetic disorders is complicated, as there are over 7,000 known rare diseases, 80% of which have a genetic etiology (Elliott, 2020; Ferreira, 2019; GAO, 2011). Further, as around two-thirds of these rare conditions impact children (Elliott, 2020; Nguengang Wakap et al., 2020)—resulting in over 4,000 genetically characterized rare diseases that impact children—we chose to use broad phrases to describe the highest level of a condition or disease (e.g., chronic illness, genetic disorder, rare disease, chromosome abnormality, etc.) instead of including any specific diseases or conditions (e.g., Down Syndrome, Autism Spectrum Disorder, etc.). As such, the searches may miss studies where the authors did not use these keywords in the titles, abstracts, or author keywords. To try to compensate for this limitation, we used the “Genetic Diseases, Inborn” and “Chromosome Disorders” MeSH terms which would automatically extend to include more specific diseases of condition of interest under these MeSH terms. We also included the MeSH term “Rare Disease” because 80% of these disorders are genetic in nature as mentioned above. After full text screening was completed, we reviewed the reference lists of all included articles to identify any additional articles of possible relevance. Any articles identified in this manner were screened using the same steps listed below.

Table 1.

Final PubMed Search Strategy

Component	Terms
Pediatric	(pediatric[tiab] OR pediatrics[tiab] OR children[tiab] OR child[tiab] OR toddler*[tiab] OR infant*[tiab] OR newborn*[tiab] OR baby[tiab] OR babies[tiab] OR neonate*[tiab] OR juvenile[tiab] OR juveniles[tiab] OR adolescent*[tiab] OR adolescence[tiab] OR childhood[tiab] OR teen[tiab] OR teens[tiab] OR teenager*[tiab] OR perinat*[tiab] OR schoolchild*[tiab] OR paediatric*[tiab] OR boy[tiab] OR boys[tiab] OR girl[tiab] OR girls[tiab] OR “Child”[Mesh] OR “Child, Preschool”[Mesh] OR “Infant”[Mesh] OR “Infant, Newborn”[Mesh] OR “Adolescent”[Mesh])
Condition	(“chronic illness”[tiab] OR “chronic illnesses”[tiab] OR “chronic disease”[tiab] OR “chronic diseases”[tiab] OR “chronic disorder”[tiab] OR “chronic disorders”[tiab] OR “chronic condition*”[tiab] OR “rare disease”[tiab] OR “rare diseases”[tiab] OR “rare illness*”[tiab] OR “rare disorder*”[tiab] OR “genetic disease*”[tiab] OR “genetic disorder*”[tiab] OR “genetically linked disease*”[tiab] OR “genetically linked disorder*”[tiab] OR “chromosomal abnormalit*”[tiab] OR “chromosomal disorder*”[tiab] OR “chromosome disorder*”[tiab] OR “chromosome disease*”[tiab] OR “chromosomal disease*”[tiab] OR “inborn genetic disease*”[tiab] OR “inborn genetic disorder*”[tiab] OR “inborn error”[tiab] OR “inborn errors”[tiab] OR “hereditary disorder*”[tiab] OR “hereditary disease*”[tiab] OR “hereditary illness*”[tiab] OR “hereditary condition*”[tiab] OR “genetic defect*”[tiab] OR “pediatric illness*”[tiab] OR “pediatric disorder*”[tiab] OR “pediatric disease*”[tiab] OR “childhood illness*”[tiab] OR “childhood disease*”[tiab] OR “childhood disorder*”[tiab] OR “genetic abnormalit*”[tiab] OR “gene abnormalit*”[tiab] OR “autosomal recessive”[tiab] OR “special health care need*”[tiab] OR “special healthcare need*”[tiab] OR “Genetic Diseases, Inborn”[Mesh] OR “Chromosome Disorders”[Mesh] OR “Rare Diseases”[Mesh])
Caregivers	(caregiver[tiab] OR caregivers[tiab] OR caregiving[tiab] OR “care support”[tiab] OR carer[tiab] OR carers[tiab] OR “care giver*”[tiab] OR “care giving” [tiab] OR “care burden”[tiab] OR “Caregivers”[Mesh] OR “Caregiver Burden”[Mesh]
HPA Axis	(“HPA axis*”[tiab] OR “Hypothalamo pituitary adrenal axis*”[tiab] OR “hypothalamic pituitary adrenal axis*”[tiab] OR “biological stress*”[tiab] OR “adrenal axis*”[tiab] OR hypothalamus[tiab] OR “hypothalamic hormone*”[tiab] OR “adrenal gland*”[tiab] OR Glucocorticoid*[tiab] OR endocrine[tiab] OR hormone*[tiab] OR hormonal[tiab] OR “stress system*”[tiab] OR “stress response*”[tiab] OR stressor*[tiab] OR “chronic stress*”[tiab] OR “physiologic stress*”[tiab] OR “physiological stress*”[tiab] OR adrenal[tiab] OR biomarker*[tiab] OR “biological marker*”[tiab] OR neuroendocrine[tiab] OR “biological outcome*”[tiab] OR cortisol[tiab] OR hydrocortisone[tiab] OR DHEA[tiab] OR Dehydroepiandrosterone[tiab] OR hypercortisolism[tiab] OR hypercortisolemia[tiab] OR hyperhydrocortisonism[tiab] OR hypocortisolism[tiab] OR “adrenocorticotropic hormone*”[tiab] OR ACTH[tiab] OR Corticotropin[tiab] OR CRH[tiab] OR “corticotropin-releasing hormone*”[tiab] OR “arginine vasopressin”[tiab] OR argipressin[tiab] OR “hypothalamic pituitary unit*”[tiab] OR “Hypothalamo-Hypophyseal System*”[tiab] OR “Hypothalamo-Hypophyseal System”[Mesh] OR “Pituitary-Adrenal System”[Mesh] OR “pituitary adrenal system*”[tiab] OR “hypothalamus hypophysis adrenal axis”[tiab] OR “hypophysis adrenal system*”[tiab] OR “adrenal hypophyseal axis”[tiab] OR hypoadrenocorticism[tiab] OR “Adrenocorticotropic Hormone”[Mesh] OR “Dehydroepiandrosterone”[Mesh] OR “Corticotropin-Releasing Hormone”[Mesh] OR “Arginine Vasopressin”[Mesh] OR “Biomarkers”[Mesh] OR “Hydrocortisone”[Mesh] OR “Glucocorticoids”[mesh] OR “Pituitary Gland, Anterior”[mesh])

Selection of Sources of Evidence

To streamline the review process, we used the web-based software Covidence (Veritas Health Innovation, Ltd.), which facilitates tasks such as detecting and deleting duplicate study entries, keeping decisions blinded from other reviewers, tracking excluded studies, and customizing data extraction templates (Kellermeyer et al., 2018). A pilot of the screening was conducted using similar search terms and inclusion criteria in Covidence to practice and refine reviewing techniques and familiarize the review team with the software.

Articles were first screened for inclusion by title and abstract only by authors LM and TR, and reviewer disagreements were resolved by author MZ. Articles that passed this first round of review were then screened for inclusion via full text using the same eligibility criteria by authors LM and TR; author MZ again resolved any discrepancies.

To find additional relevant articles, the reference lists of all articles included after full-text screening were reviewed by LM and TR. All articles identified as potentially relevant by reference list scanning proceeded through the same process of title/abstract and full-text screening described above with the same inclusion criteria. Before starting data collection, all included articles were screened by authors LM and MZ as a quality control check to ensure all included studies were appropriate to present in the review (e.g., whether manuscripts report on the same data).

Data Charting Process & Data Items

Using Covidence, authors LM, TR, KK, BK, and MZ extracted data from included studies, and LM and MZ resolved any discrepancies. Specifically, the country in which the study was conducted, congenital, chromosomal, or genetic condition of the child, number of children, number of children with condition, caregiver sex, caregiver health status, study design, biomarker collection details (biomarker collected, collection type and method, time of collection, place of collection, quality and compliance checks), biomarker calculation, links of caregiving to the HPA axis, role of sex in the links of caregiving to the HPA axis, caregiver stress measures, and links of caregiver stress to the HPA axis were collected.

We categorized each study based upon study design (case-control, cross-sectional, longitudinal, intervention) in Tables 2–4. Table 2 reports relevant demographic and contextual background details of each study, Table 3 details the biomarker collected and information regarding biomarker collection, and Table 4 reports the links of caregiving and caregiving stress to the HPA axis. Below, we report the findings classified by the main aims of the study: 1) reporting the caregiving contexts that have been studied in the current literature, 2) examining the ways in which the HPA axis has been studied in the literature, and 3) linking caregiving and caregiving stress to the HPA axis.

Table 2.

Study Context and Demographics

	First author (Year)	Country	Diagnosis	Number of children	Number of affected children	CG sex	CG health status: Covariates	CG health status: Genetic
Case control
	Ljubičić (2020)	Croatia	DS, ASD, CP, T1D	Collected data; not included in analysis	Exclusion criteria > than 1 child with condition	61% female	Specific exclusion criteria (e.g., chronic diseases)	Not discussed
	Nozoe (2014)	Brazil	DMD	Not discussed	Not discussed	100% female	Menstrual status (e.g., hormonal levels), illness history	Not discussed
	Wong (2014)	United States	ASD, FXS	Two groups did not differ; M 2.55	Not discussed	100% female	Medications	All carriers of the FXS premutation gene
	Ruiz-Robledillo (2013)	Spain	ASD	Not discussed	Not discussed	61% female	Medications, Selfreported health	Not discussed
	Lovell (2012)	United Kingdom	ASD and ADHD	CG: 2.3 +/− 1.3 (1–6) Control: 1.7 +/− 0.6 (1–3) P 0.10 No sig diff	At least one child with diagnosis	86% female	Specific exclusion criteria (e.g., no chronic illness, not pregnant)	Not discussed
	Bella (2011)	Brazil	CP	CG: 2.0 +/− 1.1, (1–6) Control: 2.2 +/− 1.3, (1–7) P 0.29	Not discussed	100% female	Specific exclusion criteria (e.g., minimal chronic illness), SF-36, medications	Not discussed
	Miller (2002)	United States	Cancer	Did not differ	Not discussed	76% female	Specific exclusion criteria (e.g., no chronic illness,	Not discussed
Cross-sectional
	Kyritsi (2017)	Greece	CAH, HT, TS	Not discussed	Not discussed	55% female	Specific exclusion criteria (e.g., no family history of mental illness)	Carriers of 21-OHD
	Chen (2015)	Chile	Mixed	Not discussed	Only one affected child per family included	88% female	Weight, height, BMI	Not discussed
	Foody (2015)	Ireland	ASD	Not discussed	Not discussed	50% female	Report of chronic illnesses, BMI, mood measures	Not discussed
	Ruiz-Robledillo (2014)	Spain	ASD	Not discussed	Not discussed	60% female	GHQ-28	Not discussed
	Dykens (2013)	United States	DS, PWS, ASD, and WS	M = 2.45	Not discussed	100% female	Medical history, height, weight, blood pressure	Not discussed
	Stoppelbein (2012)	United States	Cancer and T1D	Not discussed	Not discussed	100% female	Specific exclusion criteria (e.g., endocrine disorders)	Not discussed
Longitudinal
	Hartley (2012)	United States	FXS	M = 2.49, SD =1.04, Range 1–6	One designated “target child”	100% female	Medications	Premutation; excluded mothers with full mutation or normal FMR1 gene
Intervention
	Tsiouli (2014)	Greece	T1D	Not discussed	Not discussed	79.5% female	Specific exclusion criteria (e.g., no psychiatric diagnosis), selfreported health, sleep quality, health habits	Not discussed

Note: DS refers to down syndrome, ASD refers to autism spectrum disorder, CP refers to cerebral palsy, T1D refers to type 1 diabetes, DMD refers to Duchenne muscular dystrophy, FXS refers to Fragile X syndrome, ADHD refers to attention deficit hyperactivity disorder, CAH refers to congenital adrenal hyperplasia, HT refers to hypothyroidism, TS refers to Turner syndrome, PWS refers to Prader-Willi syndrome, and WS refers to Williams syndrome, CG refers to caregiver

Table 4.

Links of Caregiving to the HPA Axis

	First author (Year)	Diagnosis	Measurements	Links of CG to HPA axis	CG gender comparison	Stress measure	Links stress to the HPA axis
Case control
	Ljubičić (2020)	DS, ASD, CP, T1D	Slope of CAR, CAR salience, mean cortisol increase, CAR increase, AUC, aucG, aucI	All groups low CAR; DS and T1D reduced AUCI	No associations between sex and cortisol	PSS, PRSS- 18, CCBS	No
	Nozoe (2014)	DMD	Concentrations	CGs higher cortisol levels than NCs	N/A (Only mothers)	No	N/A
	Wong (2014)	ASD, FXS	CAR	Nonstressor days: CGs lower CAR than NCs	N/A (Only mothers)	DISE	Yes
	Ruiz-Robledillo (2013)	ASD	CAR, AUC	No difference CGs vs. NCs	Controlled for sex	CBI, EAE-G, Barthel Index	No
	Lovell (2012)	ASD and ADHD	Diurnal rhythm, diurnal slope, evening levels	No difference CGs vs. NCs	No significant difference, trend (p > .05) for higher cortisol for women	PSS	No
	Bella (2011)	CP	CAR, AUC, concentrations	CGs lower CAR and AUC than NCs but preserved cortisol rhythm	N/A (Only mothers)	PSS	No
	Miller (2002)	Cancer	AUC, diurnal slopes	CGs flatter diurnal slope than NCs	Did not examine	PSS	No
Cross-sectional
	Kyritsi (2017)	CAH, HT, TS	AUC, concentrations (peak and mean)	No difference CAH CGs vs. CGs of children with endocrine disorders	Did not examine	No	N/A
	Chen (2015)	Mixed	Concentrations	Perceived stress not correlated with cortisol	Did not examine	PSS	Yes
	Foody (2015)	ASD	AUCG, AUCI	Lower than average postawakening cortisol	No differences between groups in cortisol	PSI-SF, PRS	No
	Ruiz-Robledillo (2014)	ASD	CAR, AUC, concentration levels	Negative correlation between resilient coping and CAR and AUCG	Controlled for sex	EAE-G, CBI	No
	Dykens (2013)	DS, PWS, ASD, and WS	CAR, evening levels	One group of CGs lower mean cortisol and blunted trajectories of change vs. one group higher cortisol and steeper trajectories of change	N/A (Only mothers)	PSI, CBCL	Yes
	Stoppelbein (2012)	Cancer and T1D	Concentrations	CGs with PTSD higher cortisol levels vs. those without at Time 1	N/A (Only mothers)	SCID-PTSD, stressful life events checklist	No
Longitudinal
	Hartley (2012)	FXS	CAR	Greater genetic vulnerability in CGs and lower CAR	N/A (Only mothers)	SIB-R	Yes
Intervention
	Tsiouli (2014)	T1D	Diurnal rhythm	No differences; low-normal range	Did not examine	PSS, PSI-SF	No

Note: CGs = Caregivers; NCs = Non-caregivers; PSS = Perceived Stress Scale; PRSS-18 = Parental Stress Scale; CCBS = Child’s Challenging Behavior Scale; DISE = Daily Inventory of Stressful Events; CBI = Caregiver Burden Inventory; EAE-G = Escala General de Apreciación al Estrés; SF-36 = Short Form Health Survey of the Medical Outcomes Study; PSI-SF = Parenting Stress Index-Short Form; SIB-R = Scales of Independent Behavior-Revised; PRS = Parental Responsibility Scale; GQH-28 = General Health Questionnaire; CBCL = Child Behavior Checklist; SCID-PTSD = Structured Clinical Interview for the DSM-5

Table 3.

Biomarker Information

	First author (Year)	Biomarker(s) measured	Collection type and method	Time of collection	Setting	Measurements	Quality and compliance checks
Case control
	Ljubicic (2020)	Cortisol	Saliva; Salivettes	Before bed, waking next morning, 15-, 30-, and 60- minutes post-waking	Home	Slope of CAR, CAR salience, mean cortisol increase, CAR increase, AUC, AUCG, AUCi	Not discussed
	Nozoe (2014)	Cortisol, ACTH	Blood	~7am	Lab	Concentrations	Not discussed
	Wong (2014)	Cortisol	Saliva; Salivettes	Waking, 30 minutes postwaking, before lunch, before bed (2 days)	Home	CAR	Dropped if noncompliant with complete saliva collection
	Ruiz- Robledillo (2013)	Cortisol	Saliva; Salivettes	Waking, 30-, 45-, and 60- minutes post-waking	Home	CAR, AUC	Assay sensitivity .5 ng/dl; intraassay CV 2.8%; inter-assay CV 5.3%
	Lovell (2012)	Cortisol	Saliva; Salivettes	Waking, 30 minutes postwaking, noon, 10pm (2 consecutive days)	Home	Diurnal rhythm, diurnal slope, evening levels	Dropped if noncompliant with complete saliva collection
	Bella (2011)	Cortisol	Saliva; Salivettes	Waking, 30 minutes postwaking, before lunch, before dinner	Home	CAR, AUC, concentrations	Intra-assay CV 6.9%; inter-assay CV 6.2%; All cortisol samples considered in analysis
	Miller (2002)	Cortisol	Saliva; Salivettes	1-, 4-, 9-, 11-, and 13- hours post-waking (2 days)	Home	AUC, diurnal slopes	Intra-assay CV <10%; inter-assay CV <12%
Cross- sectional
	Kyritsi (2017)	Cortisol, ACTH, DHEA-s	Saliva, blood, and urine	Saliva cortisol 08:00h, 12:00h, 16:00h, 20:00h and 24:00h (2 days), 24- hour urine specimen Blood (ACTH, Cortisol) at −5, 0, 15, 30, 45, 60, 90 and 120 min after stimulation of oCRH (given at 8:00h) Blood (DHEA-S) drawn at 8:00h prior to oCR administration	Lab	AUC, concentrations (peak and mean)	Intra-assay CVs 6.2%, 5.6% and 5.1% at 1.30 and 7.12 and 22.88 nmol/L, respectively. The inter-assay CVs were 9.2% and 5.2 % at 2.33 and 5.97 nmol/L, respectively
	Chen (2015)	Cortisol	Hair	Not specified	Clinic	Concentrations	Dropped if noncompliant with hair collection
	Foody (2015)	Cortisol	Saliva; Salimetrics Oral Swab	Waking, 15-, 30-, and 45- minutes post-waking, noon	Home	AUCG, AUCI	Intra-assay CV <4%; inter-assay CV <7%; Dyad dropped if noncompliant with complete saliva collection
	Ruiz-Robledillo (2014)	Cortisol	Saliva; Salivettes	Waking, 30-, 45-, and 60- minutes post-waking (2 days)	Home	CAR, AUC, concentration levels	Assay sensitivity .5 ng/dl; intra-assay CV 2.8%; inter-assay CV 6.53%
	Dykens (2013)	Cortisol	Saliva; Salivettes	Waking, 30 minutes postwaking, before lunch, ~4pm, ~6pm, before bed	Home	CAR, evening levels	Dropped if noncompliant with complete saliva collection
	Stoppelbein (2012)	Cortisol	Saliva; Salivettes	Between 3:00 and 5:00 (2 days)	Home and lab	Concentrations	Intra-assay CV 3.41%
Longitudinal
	Hartley (2012)	Cortisol	Saliva; Sarstedt salivette collection devices	Waking, 30 minutes post getting out of bed	Home	CAR	Dropped if noncompliant with complete saliva collection
Intervention
	Tsiouli (2013)	Cortisol	Saliva; Salivettes	08:00, 12:00, 15:00, 18:00, 21:00, bedtime	Home or work	Diurnal rhythm	Compliance not considered a factor differentiating intervention group stress levels

Note. CAR = Cortisol awakening response; AUC = Area under the curve; AUC_G = Area under the curve with respect to ground; AUC_I = Area under the curve with respect to increase; CV = Coefficient of variation

Results

Included Studies

A total of 1,348 records were retrieved of which 344 were duplicates thereby yielding 1004 unique records. See Figure 2 for a flow diagram of the review process. During title and abstract screening, a total of 954 articles did not meet inclusion criteria. The remaining 50 articles moved to full-text screening. During this stage, 40 articles were excluded largely because physiological outcomes were not measured (n = 16), were reviews (n = 10), full text could not be accessed (e.g., poster presentations; n = 4); the study collected other biomarkers (n = 4); or the study did not focus on our target population (e.g., adult care recipients; n = 6). While we note that not all cancers are chromosomal, congenital or genetic in nature, we opted to retain the studies focused on pediatric cancer that met all inclusion criteria (n = 2) because 1) these particular studies were returned by the search terms focused on chromosomal, congenital or genetic conditions and 2) recent studies suggest a genetic component in some pediatric cancers (see Dean & Farmer, 2017 for a review). Thus, 10 studies were found to fit all inclusion criteria at this stage. To ensure the search did not miss relevant articles, the reference lists of the 10 included articles were screened and revealed nine additional articles to include. Finally, these 19 studies underwent a final round of review before data extraction. During this stage, four studies were found to be unfit for the review due to insufficient number of pediatric participants with a genetic condition (e.g., one out of 31 hospitalized children with a genetic condition; n = 1) and the repeated use of the same dataset by a team of researchers (n = 3). We included one paper by this group (Hartley et al., 2012) that focused on caregivers of children with Fragile X syndrome that most clearly reflected our topic of interest: caregiving stress and the HPA axis, as well as an additional paper that utilizes the sample of Fragile X caregivers but also compared across additional datasets (e.g., caregivers of children with autism; Wong et al., 2014). Thus, there were 15 total studies that met the inclusion criteria for this review.

Figure 2:

PRISMA Flow diagram

Caregiving Contexts & Study Characteristics

Results can be seen in Table 2. These studies were conducted in a variety of countries, including the United States (n = 5), Brazil (n = 2), Greece (n = 2), Spain (n = 2), Chile (n = 1), Croatia (n = 1), Ireland (n = 1), and the United Kingdom (n = 1). A total of 1034 caregivers were captured across studies, with caregiver sample sizes ranging from 19 (Nozoe et al., 2014) to 124 (Wong et al., 2014). Studies focused on caring for children with autism spectrum disorder (n = 7), type 1 diabetes (n = 3), cerebral palsy (n = 2), Down syndrome (n = 2), Fragile X syndrome (n = 2), cancer (n = 2), ADHD (n = 1), congenital adrenal hyperplasia (n = 1), Duchenne muscular dystrophy (n = 1), Prader-Willi syndrome (n = 1) and Williams syndrome (n = 1); some studies (n = 6) focused on multiple illness contexts. Additionally, another study (Chen et al., 2015) examined multiple contexts, but aggregated diagnoses within broader categories of illness, rather than specific diagnoses. Categories investigated in Chen et al., (2015) include: developmental conditions (e.g., autism spectrum disorder), diseases of the musculoskeletal system and connective tissue (e.g., Duchenne muscular dystrophy), conditions of the nervous system (e.g., cerebral palsy), and congenital malformation and chromosomal abnormalities (e.g., Down syndrome). Seven of the studies were case-control designs (caregiver vs. non-caregiver parents), six of the studies were cross-sectional designs, one was a longitudinal design, and one was a randomized control study focused on a stress reduction intervention for caregivers.

All caregivers identified as a primary caregiver to the care recipient, with one study (Chen et al., 2015) that included non-parent primary caregivers as well, such as grandmothers, older sisters, older cousins, and stepfathers. Six of the studies focused solely on mothers of affected children, eight of the studies focused on both mothers and fathers of affected children, and one study included caregivers outside of the parental dyad. Because this scoping review focuses on the biological impact of caregiving stress, we report biological sex of the caregiver. Of the studies (n = 9) that did not solely focus on mothers, many were still predominantly female samples. Seven studies reported more than 55% female caregivers (e.g., mothers, grandmothers, etc.). Only two studies included roughly equal numbers of male and female caregivers (Foody et al., 2015 [50% female]; Kyritsi et al., 2017 [55% female]. Results can be seen in Table 2.

Three studies simply controlled for the effect of biological sex (Chen et al., 2015; Ruiz-Robledillo & Moya-Albiol, 2013; Ruiz-Robledillo et al., 2014) and three studies did not examine the role of biological sex, but two of these studies were sex-matched (Kyritsi et al., 2017; Miller et al., 2002) and one found no differences in sex across groups (e.g., caregiver vs. controls; Tsiouli et al., 2014). Finally, three studies examined how biological sex was linked to the HPA axis and overall found no difference in cortisol between male and female caregivers. Specifically, Ljubičić et al., (2020) found no links of biological sex to cortisol, Foody et al., (2015) found no differences in cortisol among parental caregiving dyads, and Lovell et al., (2012) saw a trend for female caregivers to have higher cortisol than male caregivers, but the sample was only 13% male caregivers, and the test did not reach standard statistical significance (p > .05). See Table 4 for more details.

Across studies, caregivers were approximately 42 years of age, on average. Only five studies reported racial demographics, all of which were conducted in the U.S. Of these, four (Dykens & Lambert, 2013; Hartley et al., 2012; Wong et al., 2014; Miller et al., 2002) examined primarily White caregivers (77%−97%) while the other (Stoppelbein et al., 2012) was approximately evenly split between Black Americans (57%) and White Americans (43%). No studies examined caregivers that had the same condition as their child, though in some studies caregivers were identified as carriers of genetic mutations or pre-mutations (e.g., Hartley et al., 2012; Kyritsi et al., 2017; Wong et al., 2014). Further, studies largely did not report whether the examined families had more than one child with the affected condition, as one study excluded families with more than one child with the condition (Ljubičić et al., 2020), an additional study states at least one child had an inborn condition (Lovell et al., 2012), while two studies designated one child to be included in the study without explicitly listing how many children in the family were diagnosed with a given condition (Chen et al., 2015; Hartley et al., 2012). Caregiver health comorbidities and covariates can be seen in Table 2.

Biomarker Measurement

Results can be seen in Table 3. All 15 studies examined cortisol as a marker of HPA axis function, while two studies studied ACTH in addition to cortisol (Nozoe et al., 2014; Kyritsi et al., 2017), and one study also examined DHEA/DHEA-S (Kyritsi et al., 2017). The majority of studies assessed cortisol through salivary measures (n = 12), one employed hair measures (Chen et al., 2015), one blood collection (Nozoe et al., 2014), and one through a combination of saliva and blood measures (Kyritsi et al., 2017). Most studies (n = 11) collected biomarkers at home, three collected measures in the lab (Nozoe et al., 2014; Chen et al., 2015; Kyritsi et al., 2017), and one study collected markers both in the lab and at home (Stoppelbein et al., 2012). Studies used a range of measurement types including CAR (n = 7), CAR increase (n = 1), slope of CAR (n = 1), CAR salience (n = 1), mean cortisol increase (n = 1), area under the curve (AUC) (n = 6), AUC with respect to ground (AUC_G; n = 2), AUC with respect to increase (AUC_I; n = 2), concentration levels (n = 5), diurnal cortisol rhythm (n = 2), diurnal cortisol slope (n = 2), and evening cortisol levels (n = 1). Additional discussion on cortisol measurement types can be seen in Supplemental Table 1. Additional details regarding biomarker measurement (e.g., time of biomarker collection, collection setting) are in Table 3.

Links of Caregiving to HPA Axis Biomarkers

The links of caregiving stress to the HPA axis revealed mixed results and a variety of analytic strategies. Below we discuss these results in terms of study design: case control (n = 7), cross-sectional (n = 6), and other designs (i.e., longitudinal (n = 1) and intervention (n = 1)). Results can be seen in Table 4.

Case Control Study Design

Seven of the studies compared caregivers to non-caregivers using a case control study design; one of these seven studies also employed the use of daily diary methodology (Wong et al., 2014). Of these studies, there were a variety of findings. Three of the studies found caregivers had lower CAR than non-caregivers (Ljubičić et al., 2020; Bella et al., 2011; Wong et al., 2014) and two found caregivers had lower cortisol assessed via AUC compared to non-caregivers (Bella et al., 2011; Miller et al., 2002). In contrast, one study found caregivers had higher cortisol compared to non-caregivers (Nozoe et al., 2014). Further, two found no differences in cortisol across groups (Lovell et al., 2012; Ruiz-Robledillo & Moya-Albiol, 2013). However, Ruiz-Robledillo and Moya-Albiol (2013) found that caregivers experience increased CAR compared to non-caregivers after controlling for negative affect.

Six out of the seven case control studies explicitly measured stress in caregivers (e.g., perceived stress, stressful life events). However, only one (Wong et al., 2014) empirically examined the links of stress to HPA biomarkers. Results found stress interacted with caregiving status such that when caregivers reported a stressor on a given day, they had higher CAR the following morning compared to days without a stressor. This pattern was not seen for non-caregivers.

Cross-Sectional Study Design

Of the six cross-sectional studies, there were similar mixed findings regarding the HPA axis. One study found that caregivers had lower post-awakening cortisol levels compared to average values reported in past literature (Foody et al., 2015). Another study identified two abnormal patterns of cortisol experienced by maternal caregivers: one pattern exhibiting lower mean cortisol and blunted trajectories of change, and another pattern exhibiting higher cortisol and steeper trajectories of change (Dykens & Lambert, 2013). Similarly, another study found that caregivers of children with congenital adrenal hyperplasia did not differ significantly from caregivers of children with endocrine disorders with respect to peak, mean, or AUC calculations of cortisol regardless of collection type (plasma, urine, and salivary cortisol); this study also examined ACTH and found no differences across groups in ACTH (Kyritsi et al., 2017).

Five out of six cross-sectional studies directly measured stress in caregivers. Of these five studies, two empirically examined the links of stress to the HPA axis. Dykens and Lambert (2013) found that mothers that report high perceived stress had blunted cortisol trajectories while mothers with low perceived stress had steeper cortisol trajectories. Chen et al., (2015) found that caregiver perceived stress was not related to hair cortisol concentrations.

Finally, two of the cross-sectional studies focused on other aspects of the caregiving process and links to the HPA axis. One study found that greater resilient coping was linked to lower CAR and AUC_G but did not directly examine the relationship between stressful life events and cortisol (Ruiz-Robledillo et al., 2014). The other found that caregivers who met diagnostic criteria for PTSD had higher cortisol levels compared to those caregivers that do not meet diagnostic criteria for PTSD (Stoppelbein et al., 2012).

Other Study Designs

There were two additional study designs: one longitudinal study and one intervention study. The sole longitudinal study measured caregiving stress and HPA axis function within caregivers of a child with Fragile X syndrome, but largely focused on other aspects of this context. Specifically, they focused on whether caregivers’ own genetic status—specifically premutation in the FMR1 gene—served as a moderator of the link between caregiving stress and cortisol. They found a relationship between premutation gene status in caregivers and lower CAR (Hartley et al., 2012). The single intervention study (Tsiouli et al., 2014) examined caregivers of children with type 1 diabetes and randomly assigned caregivers to either a stress management intervention focused on relaxation techniques plus diabetes education or a control group that received only diabetes education. The two groups did not differ at baseline or at 2-month follow-up in mean salivary cortisol. Notably, the paper notes that cortisol ranges in both groups were in the low-normal range.

Both Hartley (2012) and Tsiouli (2014) assessed caregiver stress, but only Hartley linked caregiver stress to HPA axis and found different patterns of cortisol based on the level of caregiver genetic vulnerability. The morning after exposure to caregiving stressors, caregivers with greater genetic vulnerability had lower levels of cortisol while caregivers with less vulnerability had higher levels of cortisol.

Discussion

Caregivers of children with a chronic illness—especially congenital, chromosomal, and genetic conditions —experience a great deal of stress and high caregiver burden. While numerous studies examine psychological aspects of caregiving such as support gaps (Pelentsov et al., 2015), the prevalence of psychological distress (Cohn et al., 2020; Dijkstra-de Neijs et al., 2020), and the impact on familial processes (Pelentsov et al., 2016), there are very few studies that examine the physical impact of caregiving stress for parents of children with a congenital, chromosomal, or genetic condition. Studies in the adult caregiving literature demonstrate a clear physical toll of caregiving, but there are fewer studies conducted in the pediatric caregiving domain. This study therefore sought to provide a comprehensive review mapping the current state of literature linking pediatric caregiving stress to the HPA axis—a primary driver of the physiological stress response—and outline future research directions. After review, 15 studies met inclusion criteria spanning multiple diseases, countries, and links of caregiving stress to health. Contrary to findings in the adult caregiving literature that identify consistent links of caregiving stress to physical health outcomes, there were inconsistent links of caregiving stress in a congenital, chromosomal, or genetic context to the HPA axis across studies. Below we review the main findings, discuss how the various study designs may contribute to the inconsistent links, and consider potential future avenues for research.

One goal of this review was to examine the ways in which the HPA axis has been assessed to gain greater clarity as to the current state of the literature and identify future directions. Across studies, the most common measure of HPA axis function was cortisol and other markers of HPA axis function (e.g., AVP, DHEA) were not often assessed. One reason for this pattern may be the low burden of collection—cortisol can be captured at home through salivary measures, is non-invasive compared to blood draws that measure AVP or DHEA, and is relatively stable when stored at room temperature (Kristenson et al., 2012). However, while all studies assessed cortisol in some form, there was a great deal of variability in both the methods and calculation of cortisol, as collection types included hair, salivary, urinary, and blood samples. Further, there was a great deal of variability in collection across studies, as location of collection (e.g., home vs. in the lab), time of day, and reporting of procedures varied, resulting in a literature with heterogenous study procedures and findings.

Calculations of cortisol were even more varied and included CAR, AUC, and diurnal cortisol slope, among others. This variability in measurement may be due to the complex nature of analyzing data across multiple timepoints and decisions regarding how to translate these data points into meaningful calculations; oftentimes researchers choose to reduce the complexity and utilize a single measure (e.g., CAR, AUC) rather than model trajectories of cortisol. Moreover, the variability in biomarker measurements may make it difficult to interpret links of caregiving to the HPA axis across studies, as patterns of healthy vs. unhealthy trajectories can differ based on metric. For example, while higher circulating cortisol levels are generally considered unhealthy, a higher or steeper cortisol CAR often represents a healthy pattern (see Supplemental Table 1 for more details). One advantage to leveraging multiple types of metrics though is that they may provide differing insights into HPA axis function—as some measures may represent more transient stress levels whereas others represent more cumulative measures of stress.

While it is interesting that these various metrics may provide differential information regarding HPA axis function and caregiver stress, perhaps as a result, there were inconclusive links of caregiving to cortisol outcomes in this context. For example, while some studies found no evidence of HPA axis dysregulation for caregivers (Lovell et al., 2012; Ruiz-Robledillo & Moya-Albiol, 2013), others found caregivers did indeed have less healthy patterns of cortisol (Ljubičić et al., 2020; Bella et al., 2011; Wong et al., 2014), while at least one found caregivers had healthier patterns of cortisol after controlling for additional factors such as negative affect (Ruiz-Robledillo & Moya-Albiol, 2013). We may have seen these varied and inconclusive links of caregiving to HPA axis function because cortisol measurement is subject to variation across data collection methods. Inconsistencies in study collection and instructions may result in additional noise that obscures links between caregiving and cortisol levels; thus, future research should describe and standardize biomarker collection procedures to validly examine these links. We also note that in this scoping review we focused on the primary markers of HPA axis function (e.g., cortisol), but future reviews may want to examine other physiological systems that interact with the HPA axis and biomarkers such as interleukin-6 and C-reactive protein. Additionally, enhanced standardization of best practices for biomarker collection as well as reporting of those biomarker collections may lead to greater clarity in the role of caregiving stress on the HPA axis.

The various calculations of cortisol may also provide challenges in capturing stress in the caregiving population specifically. Within the examined studies, by far the most common measurement of cortisol was CAR—a measurement that examines the increase in cortisol within the first hour of waking. However, caregivers of young children or those with a complex medical condition may wake multiple times in the night to check on the care recipient—thereby impacting the CAR measurements, as poorer sleep quality has been linked to disrupted CAR and awakening cortisol (e.g., Vargas & Lopez-Duran, 2014). In this review, multiple studies did find caregivers had lower CAR than non-caregivers (e.g., Bella et al., 2011; Wong et al., 2014), while others found no differences (e.g., Ruiz-Robledillo & Moya-Albiol, 2013). In contrast, some studies use metrics of cortisol that may be more cumulative in nature, such as hair collection or calculation of AUC, but these studies again suggested mixed findings. It may be the case that caregivers in some situations have normalized or adapted to the caregiving stress process, which may result in a lack of differences across caregiving for a child with a chronic illness vs. a child that is typically developing.

Additionally, the studies in this review focused on a variety of caregiving contexts. For instance, diagnoses ranged from developmental disorders such as autism spectrum disorder, to endocrine disorders such as type 1 diabetes, to illnesses that are more complex and impact multiple domains such as cerebral palsy. While there are many facets of the caregiving experiences that are common across disorders—such as the need for familial support (Pelentsov et al., 2015)—specific aspects of the caregiving stress process may vary based on age of diagnosis, longitudinal prognosis and developmental trajectories, and caregiver-required activities of daily living. The variability inherent across illnesses may be another contributing factor in the inconsistent links of caregiving stress to HPA axis function.

Many of the reviewed studies were also cross-sectional in design, making it impossible to understand how chronic caregiving stress may impact caregivers’ HPA axis regulation over an extended period given the singular measurement timepoints. Due to both normative developmental milestones and disease progression, stress may fluctuate over time and impact the caregivers’ physiological system differently throughout the child’s life. For example, stress may be critically high initially after diagnosis as daily patterns shift, new routines are adopted, and uncertainty is elevated, but may decrease over time as caregivers adapt to these new responsibilities. Further, stress may mirror disease trajectories, and if there are periods of worse health or new difficulties in development, stress may be heightened again. Future research is necessary to understand the stress process over time and the longitudinal impact on caregivers’ HPA axis; examining the length of diagnosis and disease course may also provide additional insights into these patterns.

The differential links of caregiving to HPA axis function may also depend on the conceptualization of stress across studies. There are a variety of ways to capture stress within caregiving—through activities of daily living as a measurement of objective caregiving burden, through perceived stress in general, through parenting-specific stress, and numerous additional measurement types. These various metrics may capture different aspects of the caregiving process and may also contribute to inconsistent findings. Future studies may seek to examine the relations among stress measures and whether the links to the HPA axis are consistent across measures within the same study. Additionally, many reviewed studies did not measure stress explicitly, but rather compared caregivers to non-caregivers as a comparison group; however, this type of analysis presupposes that caregivers experience more stress than parents of typically developing children instead of directly modeling the links of stress to health. Such a design also limits the number of contextual factors (e.g., length of caregiving, number of children, parental health comorbidities) that can be examined that may provide further nuance in understanding the links of caregiving to the HPA axis. Therefore, future studies may want to consider directly assessing the role of contextual variables and model the amount of stress caregivers report to facilitate comparisons across studies and provide additional richness to our understanding of how the caregiving process can impact physical health.

Finally, while there were only 15 studies included in this review, they represented 8 different countries. Although this factor contributed to the heterogeneity in the studies, it is important to note the global diversity of the studies as a strength. This finding exemplifies the ubiquity of the burden: caregivers of children with congenital, chromosomal, and genetic illnesses face a great deal of stress and more research is necessary worldwide. However, this research does exhibit a common pattern in psychological research wherein western, educated, industrialized, rich, and democratic (WEIRD) samples, especially from the United States, are overrepresented (Cheon et al., 2020); there is a lack of representation in caregiving communities in Africa, Asia, and Australia/Oceania. However, multiple aspects of the caregiving experience may vary based on geographic location—as structural differences in healthcare policies, access to care, or societal influences may impact caregivers’ stress and health. An intriguing pattern observed was that a majority of studies (10/15) did not report racial or ethnic demographics, and the only studies that did include such information were conducted in the United States. Therefore, there is a need for future caregiving studies to measure and report demographic variables and to represent the diverse perspectives of caregiver experiences and health implications. Further, it is important for future research to examine whether these demographic and contextual factors may play a role in the links between caregiving stress and the HPA axis. For example, the biological impact of caregiving stress may differ based on collectivistic vs. individualistic cultural backgrounds or even by caregiver gender or biological sex, as women traditionally assume primary caregiver roles.

In conclusion, this study sought to review the current research examining the impact of caregiving for children with congenital, chromosomal, or genetic conditions on the HPA axis. The review highlights a number of future directions and next steps that can guide this research. First, standardization across measures is critical in order to draw conclusions across studies. For example, hair cortisol represents a more global measure of cortisol that captures a multi-month period, whereas salivary cortisol is able to capture diurnal rhythms and changes at a more transient level. However, dysregulated cortisol in one domain should theoretically capture a dysregulated HPA axis more broadly and indicate poorer caregiver health. Given the variability in both stress measures and biomarker calculations, it is hard to generalize across studies and identify the impact of stress on the HPA axis in caregivers of children with chronic illnesses. A review of parents caring for a child with autism spectrum disorders found similarly inconsistent links between stress and psychophysiological outcomes (e.g., cardiovascular reactivity, cortisol; Padden et al., 2019). It may be the case that caregivers of children with a chronic condition experience dysregulation of the HPA axis similar to that seen in the adult caregiving literature. However, it may also be the case that caregivers of children with a chronic illness somehow normalize the unique chronic stress they experience and adapt to their daily caregiving routines. Because these caregivers are also likely younger in age than ADRD caregivers, they may be more physically healthy and more biologically resilient. In addition, more diversity in studies is also necessary to elucidate the caregiving stress process.

Future research should also take into account additional contextual factors that may shape the caregiving process. For example, when caregiving for a child with a congenital, chromosomal, or genetic illness, it is likely that other family members may also be diagnosed with the same condition, or the parents themselves may be genetic carriers. Such contextual factors may add to the overall burden of caregiving and result in a uniquely stressful experience that could impact physical health in important ways. It may also be the case that pediatric caregiving is uniquely stressful itself, as it is less normative than caregiving for a family member that is older (e.g., Berg et al., 2021). In addition, the length of caregiving and accumulation of burden may differ across pediatric and adult caregiving contexts, as caregiving for a child with a chronic condition is often a lifetime commitment and requires specialized caregiving skills (Pelentsov et al., 2015). Thus, future research is needed that examines the links of caregiving stress to HPA axis function in pediatric caregiving contexts that both standardize biomarker measurement and directly examine these potential contextual factors over time.

This review also highlighted a number of gaps that exist in the pediatric caregiving literature. First, it is possible caregiver burden may vary based on any number of characteristics such as healthcare systems within countries, collectivistic vs. individualistic cultures of caregiving, or access to mental health resources. Thus, increasing the amount and diversity of research on caregivers of children with chronic conditions and the HPA axis is critical in ensuring equity in future representations of the global caregiver burden. Additionally, more research needs to be done to examine moderators that may explain caregiving’s effect on the physiological response to stress. Future research should include more varied cultures (e.g., non-WEIRD populations and familial-focused cultures), caregiver demographics (e.g., fathers), length and severity of diagnosis, and care recipient activities of daily living, all of which may serve to impact the links between caregiving and the HPA axis.

Before concluding, we note that this review was not without limitations, and some of these limitations may also inform future research. First, we did not include specific diagnoses within our search criteria in order avoid biasing the search towards chronic conditions that may be more well-known and in order to be as broad as possible when reviewing the literature. We sought to truly understand the scope of the current literature in terms of caregiving context and assess which conditions have been studied. However, it is possible that diagnoses not specifically categorized as congenital, chromosomal, or genetic in nature may have been missed. In addition, our knowledge of conditions that are known to have a genetic component changes over time with advancement in genomic discoveries. Thus, studies that are older may not be indexed as chromosomal, congenital, or genetic in nature as our knowledge of genetics and genomics is constantly evolving.

In conclusion, the caregiving stress process model (Pearlin et al., 1990) outlines a number of pathways through which stress from caregiving can impact physical health outcomes—a pattern observed in caregivers of adults with chronic illnesses. This review sought to summarize the current state of the literature to understand what is currently known about the impact of chronic stress on the HPA axis in the context of caregiving for a child with a congenital, chromosomal, or genetic condition. Overall, we found that more work is necessary to make claims about the impact of caregiving on the physiological stress response—as measured by the HPA axis—in caregivers of children with a chromosomal, congenital, or genetic disorder given the variability in study design and disease context. However, from these studies we identify several potential future directions and urge researchers to continue to examine the physiological burden of caregiving for a child with a chronic condition.

• Few studies examine the physical impact of pediatric caregiving

• Inconsistent links between stress and the HPA axis in pediatric caregivers

• Future standardization across both stress and physiological measures is critical