1 INTRODUCTION

Current literature suggests the greatest impact of food allergy on affected individuals and their caregivers is a reduced health-related quality of life (HRQL).^1-3 The need for rigorous dietary restrictions to avoid food allergens, lack of treatment options, and fear of accidental exposure have been implicated as the major contributors to poor HRQL.^3-5 Previous reviews have reported reduced emotional well-being and significant lifestyle impacts in individuals with food allergy and their families compared to non-allergic populations.^1-3 However, these reviews were restricted by limited considerations of patient demographics, especially gender, in the existing literature at the time they were completed. Understanding factors that impact HRQL is essential to inform personalised allergy management and identification of those most in need of intervention. No recent collation of gender-related evidence has been conducted and the modifying role of gender on the association between food allergy and HRQL is not well understood.

Sex and gender differences in HRQL have been observed in other allergic conditions, though are conflated and limited to investigations of the gender binary. An investigation of venom allergy found women had poorer anxiety-related HRQL, especially after experiencing anaphylactic reactions, than men.⁶ Similarly, poorer asthma-related HRQL was observed in women, despite having similar disease statuses and management practices as men.⁷ Studies in food allergy have reported conflicting findings on whether gender has a similar influence on HRQL and have largely only captured the perspectives of children with peanut allergy.^8-10 A previous food allergy review did suggest girls reported lower scores in most areas of HRQL than boys, though was restricted by limited existing literature and HRQL was not considered in the presence of intervention.²

As food allergy commonly arises during childhood, many of the burdens related to allergy management and treatment are delegated to caregivers.^{1, 18-23} Poorer HRQL has been shown in caregivers, mostly mothers, where the experience of food allergy-related anxiety was suggested to be the largest contributing factor.²⁴ An overrepresentation of mothers in the food allergy literature has been noted and can likely be attributed to gender roles in parenting.^{1, 2, 25} Mothers are still most often responsible for managing elements of family life that can be affected by food allergy, such as meal preparation, social outings and healthcare.^{2, 25} Little is yet known about the food allergy-related experience of mothers in relation to other caregivers. Exploration of HRQL across caregiver genders is needed to fully understand the scope of the food allergy burden.

A holistic understanding of the complex interactions between gender and HRQL in food allergy is needed to improve personalised allergy management and understanding of treatment impact. Here, we present a systematic exploration of the existing literature regarding gender influences on food allergy-related HRQL in the setting of standard care (allergen avoidance) and following food allergen intervention, considering both patient and caregiver perspectives.

2 METHODS

This review was conducted and reported consistent with PRISMA guidelines.²⁶

3 RESULTS

3.1 Included studies

The search strategy identified a total of 4799 records between Medline and Embase. After removal of duplicates, 3927 publications were screened through titles and abstracts, 426 were reviewed in full and 34 were included in this systematic review (Figure 1).

Of the included studies, one investigated HRQL across all age groups,³⁰ five considered only cross-sectional HRQL of adults with food allergy,^31-35 and 28 focused on paediatric populations with food allergy^{10, 36-62} (Table 1). Caregiver HRQL was investigated in seven of the studies of children with food allergy.^{10, 36, 40, 44, 49, 53, 62} Ten interventional studies of children with food allergy were identified by our search, nine of which investigated the effects of immunotherapy on HRQL,^{53-60, 62} and one of which considered the impact of exclusion diets.⁶¹ No studies of intervention in adult populations were identified. A variety of HRQL instruments were used in studies, though age-specific versions of the FAQLQ were the most common.

TABLE 1. Study characteristics by the age category of participants and interventional versus non-interventional study types.

Author, year	Study design	Age, range or mean (SD), years	Gender/sex definition	Intervention/Exposure	Comparator	HRQL instruments	Outcome
Non-interventional studies of individuals of all ages with food allergy
Stensgaard, 2017	Cross-sectional	All ages	Male/Female or Mother/Father	Food allergy	None	FAIM-AF FAIM-TF FAIM-CF FAIM-PF FAQLQ-AF FAQLQ-TF FAQLQ-CF FAQLQ-PF	Adult, adolescent and child self-reported and proxy-reported HRQL
Non-interventional studies of adults with food allergy
Antolin-Amerigo, 2015	Validation (cross-sectional)	31.4 (13.6)	Men/Women	Food allergy	None	FAQLQ-AF	Adult self-reported HRQL
Flokstra-de Blok, 2009	Validation (cross-sectional)	37.2 (14.3)	Male/Female	Food allergy	None	FAQLQ-AF	Adult self-reported HRQL
Jansson, 2013	Cross-sectional	19–78	Male/Female	Food allergy	None	FAQLQ-AF	Adult self-reported HRQL
Saleh-Langenberg, 2015	Cross-sectional	30–40a	Male/Female	Food allergy	None	FAQLQ-AF	Adult self-reported HRQL
Warren, 2021	Cross-sectional	18+	Male/Female	Food allergy	None	FAIM-AF	Adult self-reported HRQL
Non-interventional studies of children with food allergy
Acaster, 2020b	Cross-sectional	4–15	Male/Female	Peanut allergy	None	FAQLQ-PF FAIM-PF EQ-5D-Y	Child HRQL by parent-proxy
Acaster, 2020b	Cross-sectional	4–15	Male/Female	Peanut allergy	None	FAQL-PB EQ-5D-5L	Caregiver self-reported HRQL
Arik Yilmaz, 2018	Validation (cohort)	0–12	Male/Female	Food allergy	None	FAQLQ-PF	Child HRQL by parent-proxy
Cummings, 2010	Cross-sectional	6–16	Boy/Girl	Food allergy	Normative data	PedsQL 4.0	Child self-reported HRQL
Fathi, 2016	Validation (cross-sectional)	0.5–12	Male/Female	Food allergy	None	FAQL-PB	Caregiver self-reported HRQL
Flokstra-de Blok, 2008	Validation (cross-sectional)	13–17	Male/Female or Boy/Girl	Food allergy	None	FAQLQ-TF	Adolescent self-reported HRQL
Flokstra-de Blok, 2009	Validation (cross-sectional)	8–12	Male/Female or Boy/Girl	Food allergy	None	FAQLQ-CF	Child self-reported HRQL
King, 2009	Cross-sectional	8–12	Male/Female or Boy/Girl	Peanut allergy	Normative data	PedsQL 4.0 WHOQOL-BREF	Child self-reported HRQL and Caregiver self-reported HRQL
Knibb, 2013b	Parallel validation (cross-sectional)	6–16 6–16	Male/Female or Boy/Girl Male/Female or Boy/Girl	Food allergy Food allergy	None None	PFA-QL PFA-QL-PF	Child self-reported HRQL Child HRQL by parent-proxy
Knibb, 2013b	Validation (cross-sectional)	8.9 (NR)	Male/Female	Food allergy	None	FAQL-PB Modified FAQL-PBc	Caregiver self-reported HRQL
Manso, 2017	Validation (cross-sectional)	0–12	Male/Female	Food allergy	None	FAQLQ-PF	Child HRQL by parent-proxy
Mizuno, 2017	Validation (cross-sectional)	0–12	Male/Female	Food allergy	No food allergy	FAQLQ-PF	Child HRQL by parent-proxy
Morou, 2021	Cross-sectional	7.5–12.3	Male/Female	Food allergy	None	FAQLQ-CF FAIM-CF PedsQL 4.0	Child self-reported HRQL
Ogg, 2017	Cross-sectional	1–17	Mother/Father Son/Daughter	Food allergy	Fracture or bony injury	FAIM-PF	Child HRQL by parent-proxy
Pappalardo, 2022	Cohort	0–12	Male/Female	Race (Black)	Race (White)	FAIM-PF FAQL-PB	Child HRQL by parent-proxy 3 months post-baseline and Caregiver self-reported HRQL 9 months post-baseline
Soller, 2020	Cross-sectional	0–12	Male/Female	Food allergy	None	FAQLQ-PF10	Child HRQL by parent-proxy
Thörnqvist, 2019	Cross-sectional	0–12	Boy/Girl	Food allergy	None	FAQLQ-PF	Child HRQL by parent-proxy
Wassenberg, 2012	Validation (cross-sectional)	0–12	Male/Female or Boy/Girl	Food allergy	None	FAQLQ-PF FAQLQ-CF	Child HRQL by self-report and by parent-proxy
Interventional studies of children with food allergy
Arasi, 2014	Single-arm trial	4–18	Male/Female	IT	None	FAQL-PB	Change in caregiver HRQL after 24 months of child receiving intervention
Dantzer, 2020	RCT	7–12	Male/Female	Active SLIT/Placebo OIT	Active OIT/Placebo SLIT	FAQLQ-PF FAQLQ-CF	Change in child HRQL by self-report and by parent-proxy up to 8 years post-trial
Dunn-Galvin, 2021	RCT	4–11	Male/Female	EPIT	Placebo	FAQLQ-CF	Change in child self-reported HRQL after 24 months of intervention
Epstein-Rigbi, 2016b	Validation (case–control)	0–12	Male/Female	OIT	OFC for suspected food allergy	FAQLQ-PF	Child HRQL by parent-proxy 1 week after OIT induction.
Epstein-Rigbi, 2017b	Clinical trial	4–12	Male/Female	OIT	No OIT	FAQLQ-PF	Change in child HRQL by parent-proxy after 4 months of intervention
Epstein-Rigbi, 2019b	Clinical trial	4–12	Male/Female	OIT	No OIT	FAQLQ-PF	Change in child HRQL by parent-proxy after treatment maintenance reached
Epstein-Rigbi, 2020b	Clinical trial	8–12	Male/Female	OIT	No OIT	FAQLQ-CF	Change in child self-reported HRQL after treatment up-dosing
Epstein-Rigbi, 2021b	Cross-sectional	4–12	Male/Female	Medical clown entertainment during OIT induction	No medical clown entertainment during OIT induction	FAQLQ-PF FAQLQ-CF	Child HRQL by self-report and by parent-proxy
Indinnimeo, 2013	Validation (cross-sectional)	0–12	Male/Female	Exclusion diet for one or more allergenic foods	None	FAQLQ-PF	Child HRQL by parent-proxy
Reier-Nilsen, 2019	RCT	5–15	Male/Female	OIT	No OIT	PedsQL 4.0 FAQL-PB	Change in child HRQL by self-report and by parent-proxy, and change in caregiver self-reported HRQL, after 2 years of intervention

• Abbreviations: EPIT, epicutaneous immunotherapy; EQ-5D-5L, EuroQol five-dimension scale questionnaire–five level version; EQ-5D-Y, EuroQol five-dimension scale questionnaire–Youth version; FAIM-AF, Food Allergy Independent Measure–Adult Form; FAIM-CF, Food Allergy Independent Measure–Child Form; FAIM-PF, Food Allergy Independent Measure–Parent Form; FAIM-TF, Food Allergy Independent Measure–Teen Form; FAQL-PB, Food Allergy Quality of Life–Parental Burden; FAQLQ-AF, Food Allergy Quality of Life Questionnaire–Adult Form; FAQLQ-CF, Food Allergy Quality of Life Questionnaire–Child Form; FAQLQ-PF, Food Allergy Quality of Life Questionnaire–Parent Form; FAQLQ-PF10, short version of the Food Allergy Quality of Life Questionnaire–Parent Form; FAQLQ-TF, Food Allergy Quality of Life Questionnaire–Teen Form; HRQL, health-related quality of life; IT, immunotherapy; NR, not reported; OFC, oral food challenge; OIT, oral immunotherapy; PedsQL 4.0, Paediatric Quality of Life inventory version 4.0 generic core scale; PFA-QL, Paediatric Food Allergy Quality of Life Questionnaire; PFA-QL-PF, Paediatric Food Allergy Quality of Life Questionnaire–Parent Form; RCT, randomised control trial; SLIT, sublingual immunotherapy; UK, United Kingdom; US, United States; WHOQOL-BREF, abbreviated World Health Organisation quality of life questionnaire.
• ^a Range of mean ages of participants across included countries.
• ^b Multiple reports with the same primary author compiled from the same study.
• ^c Modified version of the FAQL-PB with no constraints on recall time.

While studies varied in whether they claimed to investigate the relationship between gender or sex and HRQL, all studies considered these concepts in the same way: according to a biologically determined binary categorisation. We hereby refer to all investigations of these concepts using the broader term ‘gender’ to reflect the intersectionality between sex and gender in included literature, and as they pertain to relationships with a psychosocial construct: HRQL.

Investigations of gender on HRQL were performed as secondary, subgroup or exploratory analyses in all studies and many did not investigate or report on relevant HRQL subscores or total scores. Consequently, selective reporting of partial results limited the data available for extraction.

3.2 Risk of bias assessment

All studies, except that by Dantzer et al.,⁵⁴ received serious or critical risk of bias ratings according to the ROBINS-I (Figure 2). The most common sources of bias between studies were in the measurement of the HRQL outcome due to participants and study investigators being aware of intervention or exposure statuses (D6), and via inadequate or inconsistent consideration of confounding factors (D1). Sensitivity analysis with the ROBINS-E found minor differences in the overall risk of bias classifications of four non-interventional studies.^34-37 All four studies received a slightly poorer quality assessment with the ROBINS-E for inadequate consideration of major sources of bias. However, ROBINS-E results for all other non-interventional studies were consistent with those obtained using the ROBINS-I.

3.3 Gender-HRQL relationships identified

Collation of included studies revealed five key gender-HRQL relationships: the association between the participant's gender and (1) their self-reported HRQL, (2) their caregiver proxy-reported HRQL, and (3) their caregivers' self-reported HRQL, and also the association between caregiver gender and (4) caregiver HRQL and (5) the proxy-reported HRQL of the participant with food allergy (Figure 3).

3.4 Gender in non-interventional studies

Strong evidence for poorer baseline self-reported HRQL in female than male participants with food allergy was present, regardless of age (Figure 4A; Table S3). Female adults self-reported poorer baseline total or subscore HRQL than males across five of the six cross-sectional studies that investigated adult HRQL. Poorer baseline HRQL scores were also observed in female than male children with food allergy when HRQL was self-reported in five of the eight studies to investigate child gender on child self-reported HRQL. However, there were differing findings about affected HRQL subscores and between HRQL instruments across studies of child self-reported HRQL. Both Cummings et al.³⁹ and King et al.¹⁰ observed poorer physical subdomain HRQL as measured with the PedsQL 4.0, while Morou et al.⁴⁷ observed no gender difference in any PedsQL 4.0 subdomains. Amongst other studies employing the use of FAQLQ instruments, gender differences were commonly observed in subdomains related to dietary risk and emotional impact. One study observed clinically significantly poorer HRQL in females than males across all subscores of the FAQLQ.⁵²

There was limited cumulative evidence for the associations between child gender or caregiver gender and caregiver proxy-reported HRQL (Figure 4; Tables S4 and S5). Only two of the nine studies to report data for proxy-reported child HRQL by child gender observed potentially poorer total HRQL in female than male children, with the rest observing no difference.

3.5 Gender in interventional studies

The change in child HRQL between baseline and follow-up in interventional studies appeared to differ between genders, though the direction and presence of association is unclear (Figure 5). Data provided by three immunotherapy studies suggested a potential gender difference in the change in HRQL between baseline and follow-up. Dantzer et al.⁵⁴ observed a strong association between gender and HRQL across all scores and measures, reporting male children experienced greater HRQL improvement during follow-up than their female counterparts. Dunn-Galvin et al.⁵⁵ and Reier-Nilsen et al.⁶² also produced results that suggest slightly poorer HRQL in one gender after follow-up. However, neither of these studies defined a reference gender, such that the direction of association could not be ascertained for these results. In contrast, the results of Epstein-Rigbi et al.^56-60 differed from other included studies, with no gender differences in total HRQL scores being observed at any point during treatment, regardless of whether HRQL was proxy- or self-reported. However, poorer self-reported emotional impact subdomain HRQL, and poorer proxy-reported food anxiety subdomain HRQL were observed during oral immunotherapy induction.⁶⁰ The one study to investigate exclusion diets as a form of intervention observed no gender differences in child HRQL⁶¹ (Table S6).

4 DISCUSSION

This review suggests there may be gender differences in self-reported baseline HRQL of adults and children with food allergy, in child HRQL change over the course of immunotherapy treatment, and potentially in caregiver subdomain HRQL. Poorer HRQL was most often observed in females, whether adult, child or caregiver, as compared to their male counterparts. The only study without a critical or serious risk of bias observed a strong association between gender and HRQL, where male children experienced a greater HRQL improvement over the course of treatment than females.⁵⁴ Additionally, differences in HRQL between mothers and fathers arose in subscores and components of HRQL, but not in overall HRQL. The heterogeneity in results obtained from included studies emphasises the relationship between gender and HRQL in populations with a food allergy is complex and likely limited by variation in HRQL instruments applied and narrow considerations of gender.

Credibility is provided to our finding that females with food allergy self-reported poorer baseline HRQL total scores than males by reports of a similar direction of association occurring in studies of sex and gender in other allergic conditions.^{6, 7, 63} We identified similar gender differences across HRQL subdomains, with females with food allergy, especially children, reporting poorer outcomes across a range of physical, psychological and social HRQL subscores. The female predominance in anxiety, avoidant eating disorders and distorted body image symptoms may help explain some of the gender differences observed in these subscores.^64-66 These conditions often occur in those with food allergy,^{67, 68} and their influences may subsequently have been captured by generic HRQL measures used in the included studies. Multiple authors have argued generic HRQL instruments lack the sensitivity needed to assess food allergy-specific life experiences.^{5, 69} As such, the use of both food allergy-specific and general HRQL measures contributes to the heterogeneity in identified gender-affected subdomains. Limitations in the availability of validated food allergy-specific HRQL instruments (i.e. the FAQLQ) at the time included studies were conducted, or because food allergy-specific measures are not yet validated in as many languages as general measures (such as the PedsQL),^{70, 71} means inclusion of studies using a variety of HRQL instruments is required to capture all relevant literature in reviews of food allergy-related HRQL.^{2, 72-74} Taking this limitation into account, our findings therefore primarily act to convey that gender could interact with multiple aspects of well-being in food allergy populations. Purposeful future measurement of the impact of gender on subdomain HRQL, not only total HRQL scores, is recommended.

Collation of interventional studies revealed the direction and presence of association between gender and HRQL was somewhat unclear. Multiple authors have speculated the HRQL of individuals with food allergy is dependent on the allergen involved.^{73, 75, 76} This may explain the differences observed in our data, as all interventional studies to observe a difference between genders that was of, or approached, significance were peanut immunotherapy studies.^{54, 55, 62} In contrast, the trial of oral immunotherapies for a diversity of allergens observed no gender difference in Total HRQL.^56-60 As such, our findings extend the observations of previous authors, suggesting a potential interaction between gender and allergen type could affect HRQL outcomes during treatment. Despite this potential finding, we are unable to speculate whether the male or female gender is likely to have a greater influence on treatment outcomes due to the poor definition of a reference gender in two of the three identified interventional studies that observed a potential difference between genders.^{55, 62} Future investigations that specifically focus on the role of gender in HRQL are needed to reduce reliance on subgroup and sensitivity analyses that are often limited by selective reporting, as was experienced in this review.⁷⁷

The observation that caregiver HRQL was not likely affected by child gender across studies is consistent with previous findings of general parental well-being not being associated with child gender.⁷⁸ However, observations of a difference in subdomain HRQL between mothers and fathers of participants with food allergy suggested an association between caregiver HRQL and caregiver gender may have been present.^{10, 40} Authors have observed parents who were more involved in their child's allergy management consequently reported more greatly impaired HRQL.^{23, 25} The influence of gender roles in parenting should be acknowledged here, with primary caregiving roles that pose the most risk of food allergy burden, such as food preparation, school and social activities, most often being undertaken by mothers.²⁵ As mothers tend to be more involved in allergy-related care,^{23, 25, 76} it is reasonable that psychosocial-domain associations would be observed more so in mothers than fathers, providing support to our review's findings. However, due to the limited number of studies to observe this, future investigation of the lived experience of gender and gender roles in parenting may be beneficial to fully understand the relationship between caregiver gender and caregiver HRQL. This association is also an important consideration for study designers, as the gender of caregivers completing questionnaires throughout a study has the potential to influence results.

Strengths of this study include the extensive search strategy, large number of full-text manuscripts screened, and dual investigator screening review, which reduced risk of biased selection of studies for inclusion.⁷⁹ The initial observation that all eligible interventional studies investigated populations of children with food allergy was unexpected. A possible reason for this is that allergy in adults is often understudied compared to allergy in paediatric populations.^{80, 81} Therefore, adults would be underrepresented in studies of the role of gender in food allergy-specific HRQL interventions. Nonetheless, this does indicate results of this review collated from interventional studies are largely only generalisable to children with food allergy. Further research into adult populations is still required.

The high risk of bias amongst included studies should also be addressed. As the ROBINS tools are the preferred tools for assessing non-randomised studies, the poor quality ratings of included non-randomised studies are likely to be reliable.⁸² While this indicates our results are impacted by the biases present in the included studies, the severity of bias across these studies is fairly consistent and their findings are still comparable.⁸³ To prevent any studies that presented extreme quantitative results from severely influencing our findings, we extracted and synthesised data in terms of the general direction of association and the presence of both clinical and statistical significance (see the Supplementary Material). It is also possible that included randomised control studies may have received suboptimal quality assessment results given the ROBINS tools are best-suited to non-randomised studies.⁸² As such, additional quality assessment with tools specifically suited to assessment of RCTs should be considered in future studies of a diversity of study types.^{82, 84} Subgroup analysis by risk of bias strata could also be considered to observe whether bias risk influences the direction of results, though this would require a greater number of included studies with variation in their risk of bias assessments than was achieved in this review.⁸³

A key limitation to the generalisability of our findings is that they only address a narrow definition of gender, with all studies defining sex or gender constructs according to the same biologically determined binary. Little has yet been done to investigate genders outside of binary classification, and as such it is unclear whether biological determinants or psychosocial experience are the main drivers of HRQL outcomes in allergy.^{2, 10} To completely understand the role of gender in food allergy-specific HRQL, clear definitions of sex and gender and analysis across a diversity of genders (beyond the male–female dichotomy) are required.⁸⁵ Furthermore, the extensive reliance on p-values below a .05 significance threshold to determine a gender difference amongst the studies included in this review does indicate pressure to communicate only results that would be considered statistically significant.^{86, 87} This again emphasises the need for more studies to consider the role of gender in HRQL to determine if our results are consistent and not a consequence of publication bias.

5 CONCLUSION

This review indicates gender is a likely modifier of HRQL in IgE-mediated food allergy. Observation of poorer baseline total HRQL in female adults and children with food allergy, poorer subdomain HRQL in mothers, and potential gender differences in HRQL post-intervention, highlight that a tailoring of allergy management approaches to account for gender is needed. Our results are of particular relevance to study design and indicate the influences of both participant and caregiver genders need to be considered to advance understanding of allergy-specific HRQL. Stratification of HRQL results by gender may be especially important in emerging allergen immunotherapy trials, where it remains unclear as to whether one gender experiences a greater magnitude of the effect of treatment than others. This study highlights the need for purposeful consideration of a diversity of gender identities, ages and allergens in future investigations of HRQL to establish whether psychosocial experience or biological factors drive differences in food allergy outcomes.

AUTHOR CONTRIBUTIONS

SA Rosser, M Lloyd and MLK Tang conceived this systematic review and designed the methodology. SA Rosser, M Lloyd and A Hu contributed to dual-reviewer screening of titles, abstracts and full texts. SA Rosser conducted the data extraction and analysis and drafted the preliminary version of the manuscript. Critical appraisal of included studies using the ROBINS-I was performed by SA Rosser and A Hu before sensitivity analysis with the ROBINS-E was performed by SA Rosser and M Lloyd. All authors assisted with interpretation of the results and reviewed and critically appraised the manuscript before submission.

CONFLICT OF INTEREST STATEMENT

MLK Tang declares consultant fees from Pfizer; was a past employee (ended July 2022) of and share interest/options in Prota Therapeutics; is a member of the Medical Advisory Board of Anaphylaxis & Anaphylaxis Australia; is a member of the Board of Directors of Asia Pacific Association of Allergy Asthma and Clinical Immunology and was a past member of the Board of Directors of the WAO (ended 2019); is a member of expert committees of the American Academy of Allergy, Asthma & Immunology, Asia Pacific Association of Allergy Asthma and Clinical Immunology, Australasian Society of Clinical Immunology and Allergy, WAO; and was a past member of the International Union of Immunological Societies (ended 2019). All other authors have no competing interests to declare.