The burden of mitochondrial disease with associated seizures: systematic literature reviews of health-related quality of life, utilities, costs and healthcare resource use data

Background Mitochondrial disease is a degenerative, progressive, heterogeneous group of genetic disorders affecting children and adults. Mitochondrial disease is associated with morbidity and mortality, with predominantly neurological and neuromuscular symptoms including dystonia, weakness, encephalopathy, developmental delay and seizures. Seizures are one of the most common and severe manifestations of mitochondrial disease. These seizures are typically refractory to common anti-seizure therapies. There are no approved disease-modifying treatments for mitochondrial disease. Our objective was to conduct two systematic literature reviews to identify health-related quality of life (HRQoL), utilities, costs and healthcare resource use data in mitochondrial disease with associated seizures. Methods A range of databases and information sources were searched up to July 2022 to identify eligible studies. Search strategies included a range of variant terms for mitochondrial disease and HRQoL, utilities, cost and healthcare resource use outcomes. Two reviewers independently assessed articles against the eligibility criteria; studies were extracted by one reviewer and checked by a second. Risk of bias was assessed for studies reporting HRQoL data. Results were narratively assessed. Results Seven studies were eligible for the HRQoL and utilities review. The studies used different tools to report data, and despite the variability in methods, HRQoL scores across the studies showed moderate/severe disease in patients with mitochondrial disease with associated seizures. Parents of patients with mitochondrial disease with associated seizures were characterised by high total parenting stress. No studies reported utilities data. Two case reports and one retrospective review of medical records of children who died in hospital were eligible for the costs and resource use review. These provided limited information on the duration of hospital stay, in an intensive care unit (ICU), on mechanical ventilation. No studies reported costs data. Conclusion These reviews highlight the limited HRQoL, utilities, costs and resource use data and the variability of instruments used in mitochondrial disease with associated seizures. However, the data available indicate that mitochondrial disease with associated seizures affects patients’ and caregivers’ HRQoL alike. No robust conclusion can be drawn on the impact of mitochondrial disease with associated seizures on hospital or ICU length of stay. Trial registration PROSPERO: CRD42022345005. Supplementary Information The online version contains supplementary material available at 10.1186/s13023-023-02945-6.


Background
Mitochondrial disease is a diverse group of rare disorders, resulting from mutations in both mitochondrial and nuclear DNA, causing deficiencies in mitochondrial function [1].These serious and progressive diseases have an unpredictable disease course and occur in approximately 1 in 5,000 individuals [2,3].Mitochondrial diseases affect all organs in the body but especially those with high energy requirements including the brain and heart.Symptoms include muscular and neurological impairments alongside other complications, such as fatigue, behaviour and speech disturbances, impaired vision, and diabetes [4].Seizures are a common neurological manifestation of mitochondrial disease, affecting up to 40% of adults and 60% of paediatric cases [5,6].
The pathophysiology of seizures associated with mitochondrial disease is not fully understood, but this is mostly due to the disruption of the mitochondrial respiratory chain that results in cellular energy deficiency and promotes oxidative stress [5].This can cause an accumulation of lipid peroxides and depletion of reduced glutathione, the cell's natural antioxidant.These conditions trigger a feed-forward cycle promoting proinflammatory signalling, additional increases in oxidative stress, and, ultimately, cell death.The resultant cellular dysfunction and cell death drives the downstream pathologies observed with mitochondrial disease.In the brain, these events can create an excitatory/inhibitory imbalance that disrupts normal neuronal circuits and can cause seizures.Such seizures can increase morbidity and result in a poor prognosis or epileptic encephalopathy [5].Children with mitochondrial disease that experience seizures are characterised by more severe disease than those without [7].Seizures associated with mitochondrial disease are typically refractory to common anti-seizure therapies, as many of the approved seizure therapies are toxic to the mitochondria and exacerbate the underlying pathology of mitochondrial disease.There are no approved disease-modifying treatments for mitochondrial disease or the associated seizures.Currently there are very limited treatments for these patients.Standard of care includes multidisciplinary management and treatment of symptoms, including nutritional supplements and exercise.Furthermore, the diagnosis and management of mitochondrial disease with associated seizures is challenging because the diseases are biochemically and genetically heterogeneous and require multidisciplinary healthcare.
To the authors' knowledge, there are no published reviews assessing the health-related quality of life (HRQoL) and economic impact of mitochondrial disease with associated seizures, which may present a burden for patients, caregivers and healthcare systems.To address this evidence gap, two systematic literature reviews (SLRs) were undertaken to identify HRQoL, utilities, costs and healthcare resource use data in these patients.Together, the SLRs aimed to provide a holistic understanding of the HRQoL, utility, cost and healthcare resource use burden of mitochondrial disease with associated seizures; therefore, they are both presented within this publication.The SLRs may be used to inform future health technology assessment (HTA) submissions of possible treatments.

Methods
The SLRs were undertaken following the principles of systematic reviewing embodied in the Cochrane handbook [8] and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [9].The questions and methods of both SLRs were predefined in one protocol that was registered on the PROS-PERO database (CRD42022345005) [10].The completed PRISMA checklist is presented in Additional file 1.

Eligibility criteria
Full details of the eligibility criteria for both SLRs are presented in Table 1.The eligibility criteria were consistent between the SLRs and only the eligible outcomes differed.Studies of patients with mitochondrial disease with associated seizures or studies reporting data for caregivers of these patients were eligible for inclusion.The specific eligible mitochondrial diseases (Table 1) were selected because of their increased likelihood of phenotypes of patients with seizures [5].Studies reporting data from HRQoL and health state utility tools were eligible for the first SLR, and those reporting data on direct monetary costs, indirect monetary costs, and non-monetary healthcare resource use were eligible for the second SLR (Table 1).Eligible studies were limited to those in English language.

Searches
The original searches were conducted in August 2021 and then updated in July 2022.Two separate searches were conducted to inform the two SLRs.Table 2 shows the resources searched for each SLR.Date restrictions were not applied to searches; where appropriate, language restrictions reflecting the eligibility criteria (English language only) were applied.For both SLRs, reference lists of included studies and retrieved relevant SLRs published from 2016 were checked for additional eligible studies.Search methods for the original and update searches (including full search strategies for each database and information source) are presented in Additional file 1.

Study selection, data extraction and quality assessment
A single researcher assessed the search results according to their relevance in providing information for the reviews and removed the obviously irrelevant records such as those about animals.Two reviewers independently assessed the titles and abstracts then the full texts for relevance against the eligibility criteria, with any disagreements adjudicated by the third reviewer.The number of records included and removed at each selection stage was recorded in a PRISMA flow diagram.
Data extraction (into an Excel template) was conducted by one reviewer, and the second reviewer checked all the data points.Disagreements were adjudicated by the third reviewer.The Excel template was piloted on three studies before progressing to full data extraction.When different studies seemed to match recruitment dates and centres, and patient characteristics, we tried to reach the authors to identify whether the same patients were reported.We aimed to include the data from the same patient only once.
A quality assessment was performed for studies reporting HRQoL data [11].Full details of the quality assessment are presented in Supplementary Table 5 (in Additional file 1).One reviewer assessed the quality of each study, and the second reviewer checked the assessment.Any disagreements were resolved through discussion or by consulting the third reviewer.No quality assessment was conducted for studies reporting resource use data because the transferability of these data for potential use in future HTAs would not be impacted by an assessment of the elicitation study's conduct.
The included studies were summarised in tables and through a qualitative synthesis providing data on their methods and results.

HRQoL studies identified and selected
The original searches identified 1,704 records.Four studies (reported in four documents) were eligible for the SLR.The update searches identified 1,967 records.Three new studies (reported in three documents) were eligible for the SLR.The PRISMA diagram is presented in Fig. 1.
Lists of included and excluded studies (with reasons for exclusion) are presented in Additional file 1.

HRQoL study characteristics
Overall, seven studies (in seven publications) reporting HRQoL data were eligible for the SLR [12][13][14][15][16][17][18].The characteristics of the included studies are presented in  It must be noted that the patients included in van Kempen et al. [17] are likely the same as those included in Hendrix et al. [18] given the similarities in patient characteristics, study dates and recruitment.We contacted the study authors to clarify if this was the case, but we did not receive a response before completion of this SLR.Therefore, we have reported the results from both studies as separate sources of data because different eligible outcomes were reported in both studies.
Four studies enrolled paediatric patients [12,13,15,16], and the other three studies recruited adult patients [14,17,18]; one study also recruited mothers of the paediatric patients [12].None of the studies estimated utility data.

HRQoL outcome data
Data for the HRQoL outcomes are shown in Table 4.

Newcastle Mitochondrial Disease Adult Scale (NMDAS)
The most reported score was the NMDAS (maximum score 145; higher scores reflected more severe disease) which was used in three studies [14,17,18].
In the first study, NMDAS scores ranged from 108 to 141 in three eligible patients (aged 18-22 years) [14].

NewcastlePaediatric Mitochondrial Disease Scale (NPMDS)
The NPMDS (maximum score 107) was used in two studies [13,15].Higher scores reflected more severe disease, with total scores > 25 indicating severe disease.In the first study, NPMDS scores ranged from 36.2 to 48.8 at peak phase (before medication) and from 38.7 to 53.1 at last assessment in three eligible patients (age at diagnosis: 1 year 2 months to 5 years 8 months) [13].
In the second study, NPMDS total scores were 6, 20 and 32 for three eligible patients with Leigh syndrome and epilepsy (aged 2-12 years) [15].In this study, total scores of 0-14 represented mild disease, 15-25 moderate disease, and > 25 severe disease, thus the three patients had different disease severity levels.

Other HRQoL scores
The Japanese Mitochondrial Disease Rating Scale (JMDRS; maximum score 80; higher scores reflected more severe disease) was adopted in one study, and scores ranged from 65 to 70 in three eligible patients (aged 18-22 years) [14].
The Korean Child Behaviour Check List (K-CBCL; maximum score 100; clinical cut-off score 63) was used in one study [12].In this study, the mean K-CBCL score for total behavioural problems was 66.8 in children with mitochondrial diseases and intractable epilepsy (age not reported).The same study also reported data derived from mothers of children diagnosed with mitochondrial diseases and intractable epilepsy using the Korean version of the Parenting Stress Index (K-PSI; maximum score 100; clinical cut-off score 84) and the Beck Depression Inventory (BDI; maximum score 63; clinical cut-off score 11).For both tools, higher scores reflected more severe disease.In this study, the mean K-PSI score for total parenting stress was 93.9, and the mean BDI score for maternal depression was 14.7 [12].

Costs and healthcare resource use studies identified and selected
The original searches identified 4,633 records.Two studies (reported in two documents) were eligible for the SLR.The update searches identified 5497 records.One study (reported in one document) was eligible for inclusion in the SLR.The PRISMA diagram is presented in Fig. 2. Lists of included and excluded studies (with reasons for exclusion) are presented in Additional file 1.

Costs and healthcare resource use study characteristics
Three studies reporting healthcare resource use data were eligible for the SLR [19][20][21].No studies reported costs data.The characteristics of the included studies are presented in Table 3.Two studies were carried out in Asia (Japan and Korea) [19,20], and one was conducted in Turkey [21].The studies included two case reports [20,21] and one retrospective review of medical records [19]; all three were conducted in hospitals.Overall, limited information on healthcare resource use data (methods and results) was provided.Therefore, data from the two case reports are only presented in the narrative summary.

Healthcare resource use outcome data
One study [19] analysed the clinical characteristics and cause of death in paediatric patients (aged < 15 years) with mitochondrial diseases to establish potential risk factors associated with mortality.For the 31 eligible patients (age: 1.8 ± 2.0 years) with mitochondrial diseases and death in hospital, the number of intensive care unit (ICU) admissions per year was 0.62 (SD: 0.86), and the mean duration of ICU stay was 14.59 days (SD: 23.68) [19].Results by age subgroups are presented in Table 5.The mean number of causes of death per patient was 2.03 (SD: 0.91), with sepsis (17 patients), pneumonia (13 patients), disseminated intravascular coagulation (9 patients), and sudden unexpected death (9 patients) the most common causes.In addition, early death (children who died aged six years or younger) was associated with lesions in the thalamus, the number of organs involved, and Leigh syndrome [19].
Another study [20] presented a case report of a Japanese woman (aged 24 years) who was diagnosed with MELAS and admitted to hospital.The patient was transferred to ICU after 19 days of hospitalisation, and mechanical ventilation in ICU was required for 25 days.
A second case report [21] presented a 12-year-old male patient who was referred to a paediatric intensive care unit (PICU) in Turkey.The patient was diagnosed with MELAS on the eighth day of PICU hospitalisation and was intubated until three days after the start of MELASspecific treatment, then non-invasive mechanical ventilation was initiated.The patient was discharged on the 36th day of admission.The total length of stay in PICU was not reported, but this was until at least day eight when MELAS was diagnosed.

Discussion
The SLRs identified data on the HRQoL, utilities, costs and healthcare resource use burden in patients with mitochondrial disease with associated seizures.Overall, seven studies reporting HRQoL data and three studies reporting healthcare resource use data were identified.No studies reporting utilities or monetary costs data for this specific patient population were found in the literature.The studies included both children and adult patients, and one study also reported HRQoL data derived from mothers of children diagnosed with mitochondrial disease [12].
Mitochondrial disease is a progressive, heterogeneous group of genetic disorders, and seizures are one of the most common and severe manifestations of the disease.The presence of seizures is associated with increased morbidity and mortality and can result in a poor prognosis or epileptic encephalopathy [5] and severe disease in children [7], negatively impacting QoL and placing a burden on both patients and caregivers.The management of symptoms in these patients is challenging because seizures associated with mitochondrial disease are typically refractory to common anti-seizure therapies, and there are no approved disease-modifying treatments.
Across the included studies, HRQoL data were reported using different tools, thus limiting the comparability of results.However, despite this variability in study methods, a degree of consistency among the results was observed, with patients reporting HRQoL scores that indicate moderate to severe disease and parents scoring highly for total parenting stress.
Of the HRQoL tools used in the included studies, the NMDAS was the most commonly reported in adult patients [14,17,18].In children, the NPMDS was reported in two studies [13,15].Data from other diseasespecific tools were also reported, including the JMDRS in one study of adult patients [14] and the IPMDS in another study of paediatric patients [16].In the final HRQoL study, Eom and Lee [12] used two other generic tools to capture behavioural data in children (K-CBCL) and its impact on mothers' mental health (K-PSI).
Although the included studies provide insight into the HRQoL burden of patients with mitochondrial disease with associated seizures, they are limited in number and are generally based on small samples and retrospective or cross-sectional study designs.Further robust studies, with larger sample sizes and prospective study designs, are needed to fully understand the implications on HRQoL for patients and caregivers.Future work should use a common HRQoL tool to allow for consistent reporting and to enable further comparisons between studies.The NMDAS or NPMDS are the most commonly used currently in the literature.
In terms of the economic burden, only very limited non-monetary healthcare resource use data were identified, and the data available were in small and specific populations.Therefore, transferability of these data to other contexts and their usefulness for populating future economic models should be carefully assessed.For example, Eom et al. [19] investigated the reasons of early death in children with mitochondrial diseases, with only children who died included in the evaluation.This implies that the number of ICU admissions and the ICU length of stay data reported by Eom et al. [19] are likely to reflect the inherent disease severity of these patients and are unlikely to be representative of the average ICU admission and length of stay for children with mitochondrial disease with associated seizures.
Although no monetary costs data were identified for mitochondrial disease with associated seizures, previous studies have demonstrated that mitochondrial diseases more broadly (i.e.without specifying seizures) present an economic burden for both patients and healthcare systems [3,22,23].Patients with mitochondrial diseases (and their families) may also experience substantial outof-pocket expenses, as with other rare diseases, related to caregiver responsibilities, work productivity impairment, healthcare visits, and non-prescription medications [22].The economic burden, including out-of-pocket expenses for patients and their caregivers, is potentially increased in mitochondrial disease with associated seizures.However, this is unknown given that no monetary costs data for this population were identified in the literature, and more work is needed to establish if this is the case.

Strengths and limitations of the included studies
The strengths of the studies included the clear reporting of inclusion/exclusion criteria for patient enrolment and the explicit discussion of critical areas of the research.However, eight of the included studies were based on small samples, often as subgroups of already small cohorts with mitochondrial diseases, affecting the representativeness of the patient population and the transferability of the results to other settings.Furthermore, the epidemiological and clinical settings of some studies may not be directly applicable to populations in other areas, with potential differences in the management and treatment of these patients between countries, limiting the external validity.Further research is needed to explore the HRQoL and economic burden of mitochondrial disease with associated seizures across different countries.
The SLRs were focused on patients with mitochondrial disease with associated seizures; therefore, studies evaluating patients with mitochondrial disease, but without seizures, were not eligible and were excluded.With respect to study design, most studies were retrospective or crosssectional, and two were single case reports.These designs are methodologically weak and lack rigour when compared with randomised controlled trials (RCTs).

Strengths and limitations of the SLRs
Two SLRs were undertaken following systematic review guidance [8,9], with extensive searches in several databases and information sources.The SLRs were originally conducted in 2021 and were updated in 2022.
Limitations of the SLRs include the restriction to English-language studies only and the exclusion of conference abstracts, both of which mean there is a risk of relevant research-either published in non-English language or as a conference abstract, including ongoing research-not being identified.Furthermore, although studies reporting on any non-monetary resource use data were eligible for the costs and healthcare resource use SLR, the search strategy was only designed to retrieve records that referred to a selection of non-specific healthcare resource use related terms or the following specific non-monetary resource use outcomes: hospitalisation, visits/appointments or length of stay.This approach was prospectively discussed and agreed within the research team at protocol stage, but did potentially increase the risk of not retrieving relevant data on additional specific non-monetary resource outcomes.

Conclusions
Overall, the SLRs demonstrated the paucity of studies reporting HRQoL, utilities, costs, and healthcare resource use data in patients with mitochondrial disease with associated seizures.The HRQoL of patients and caregivers is affected with patients showing moderate to severe symptoms.No robust conclusions can be reached on resource use because of the limited data available in these patients.The limitations of the analysis are mainly related to the rarity of disease, but also to the lack of instruments that can capture the specific clinical condition that diminishes social support and family function, and presents cognitive challenges, medical and psychiatric comorbidities, and physical limitations in daily activities.Further prospective studies are needed to fully understand the HRQoL and economic burden of mitochondrial disease with associated seizures.Alongside seizures, future studies should also consider the indicators of increased severity and prognostic factors in these patients, including early age of disease onset; magnetic resonance imaging and clinical signs of brainstem involvement; and respiratory drive involvement [24].

Table 1
Summary of the eligibility criteria for both SLRs

Table 1
(continued) 17D 17-dimensional, adPEO Autosomal dominant PEO, AHS Alpers-Huttenlocher syndrome, AHUM Adolescent Health Utility Measure, arPEO Autosomal recessive PEO, AQoL Assessment of Quality of Life, CHEQOL-25 Health-Related Quality of Life Measure for Children with Epilepsy, CHU9D Child Health Utility instrument, ECQ Epilepsy and Children Questionnaire, EFA Epilepsy Foundations of America Concerns Index, ELDQOL Epilepsy and Learning Disability Quality of Life, EQ-5D EuroQol 5 dimensions, EQ-5D-Y EuroQol 5 dimensions (youth), GEOS-YP Glasgow Epilepsy Outcome Scale, G-QOLCE Global Quality of Life in Childhood Epilepsy, HARCES The Hague Restrictions in Childhood Epilepsy Scale, HSCS-PS Health Status Classification System-Preschool, HUI Health Utilities Index, ICER Incremental cost-effectiveness ratio, ICI Impact of Childhood Illness Scale, ICND Impact of Childhood Neurologic Disability, IPES Impact of Paediatric Epilepsy Scale, MCHS Childhood myocerebrohepatopathy spectrum, MELAS Mitochondrial encephalopathy, lactic acidosis, and stroke-like episode, MEMSA Myoclonic epilepsy myopathy sensory ataxia, MERRF Myoclonus epilepsy with ragged-red fibers, MIRAS Mitochondrial recessive ataxia syndrome, NeuroQol Neurology Quality of Life Measurement System, PDHC Pyruvate dehydrogenase complex, PedsQL Paediatric Quality of Life Inventory, PEO Progressive external ophthalmoplegia, PESQ Perceptual Evaluation of Speech Quality, POLG Polymerase gamma, PROMIS Patient-Reported Outcomes Measurement Information System, QALY Quality adjusted life year, QOLCE Quality of Life in Childhood Epilepsy, QOLIE-AD-48 Quality of Life in Epilepsy Inventory for Adolescents, QOLIE Quality of Life in Epilepsy Inventory, QOLPES Quality of Life in Paediatric Epilepsy, QWB Quality of Well-Being Scale, RARS2 Arginyl-tRNA synthetase 2, SANDO Sensory ataxia neuropathy dysarthria and ophthalmoplegia, SCAE Spinocerebellar ataxia with epilepsy, SF-36 36-Item Short Form Survey, SF-6D Short Form-6 dimensions, SLR Systematic literature review

Table 3
Study and participant characteristics

Table 4
HRQoL studies outcomes

Table 4 (continued) Study Population in which health effects were measured Source of perspective of the values and population characteristics Brief description of the tool and tool outcome grading Total outcome data (SD)
BDI Beck Depression Inventory, CM Cardiomyopathy, DD Developmental delay, HIBCH 3-hydroxyisobutryrly-CoA hydrolase, HRQoL Health-related quality of life, IQ Intelligence quotient, IQR Interquartile range, JMDRS Japanese Mitochondrial Disease Rating Scale, K-CBCL Korean Child Behaviour Check List, K-PSI Korean version of the Parenting Stress Index, MELAS Mitochondrial encephalopathy, lactic acidosis, and stroke-like episode, NMDAS Newcastle Mitochondrial Disease Adult Scale, NPMDS Newcastle Paediatric Mitochondrial Disease Scale, NR Not reported, PMD Primary mitochondrial disorder, QoL Quality of life, SD Standard deviation, SF-12v2 Short Form 12-item version 2