Establishing core outcome sets for phenylketonuria (PKU) and medium-chain Acyl-CoA dehydrogenase (MCAD) deficiency in children: study protocol for systematic reviews and Delphi surveys

Inherited metabolic diseases (IMD) are a large group of single-gene diseases, most often diagnosed early in life, that are individually rare but collectively have an estimated birth prevalence of at least 1 in 2500 [1, 2]. These conditions exemplify the challenges of delivering high-quality care for children with rare chronic diseases [3]. Healthcare needs are disproportionately high and often complex [4, 5], caregiver burden is known to be substantial [6], and evidence on the effectiveness of care is sparse and challenging to generate [7, 8]. With promising new interventions for IMD (including drug, dietary, stem cell, and surgical therapies) rapidly emerging [9], it is vital that their efficacy be evaluated in explanatory studies within controlled settings and that their effectiveness comparative to current standard treatment be assessed in pragmatic studies within real-world settings [8, 10, 11]. Such evaluations should be based upon outcomes of greatest clinical importance and of importance to patients and their family members and should capture measures of cost and health system impact [11, 12]. The careful consideration of these different perspectives is particularly critical when studying treatments for rare diseases, including IMD, since the relative benefits and harms of alternative therapies may differ depending on which outcomes are studied from across Berwick and colleagues’ ‘triple aim’ (i.e., medically-defined outcomes, patient experiences and quality of life, and health system impacts) [12].

While randomized controlled trials (RCTs) represent the evaluative study design with the greatest ability to minimize sources of bias and maximize internal validity, several challenges have been documented previously with the implementation of this design when assessing interventions for rare conditions. These include poor feasibility to attain adequate sample size to achieve planned statistical power [13‐15], reliance on surrogate endpoints of limited importance to patients and their caregivers [14, 16, 17], and, associated with inadequate sample size, the challenge of sufficiently evaluating heterogeneity of benefits across relevant subgroups of patients [11, 13, 18]. To address these challenges, experimental and analytic methods tailored to rare disease settings have been developed [19‐22]. One such method that has been described as an important innovation is the registry-based randomized trial [23]. This approach involves the embedding of intervention trials within observational cohort studies or patient registries and adopting a pragmatic rather than explanatory focus to support real-world decision-making [23‐29]. In the context of evaluating interventions for rare paediatric conditions, registry-based trials may offer numerous advantages including efficiencies in patient recruitment and data collection (given reliance on existing cohorts and routinely collected data), better access to follow-up data to understand long-term outcomes, suitability for addressing comparative effectiveness questions (given that registries recruit patients from routine clinical settings), and a high degree of external validity (given that registries are frequently population-based).

An important design element in the development of registry-based RCTs and other robust evaluative studies is the high-quality standardized collection of data on important outcomes, i.e., a core outcome set (COS) [28]. The Core Outcome Measures in Effectiveness Trials (COMET) initiative (www.​comet-initiative.​org) promotes the performance of literature reviews and multi-stakeholder consensus approaches for the establishment of standardized COS to be used in subsequent evaluative studies [30‐33]. These and similar methods have been used to develop outcome sets for paediatric conditions (e.g., traumatic brain injury, acute and chronic pain, fever and neutropenia in cancer, otitis media, inflammatory bowel disease) [34‐40] and rare conditions [41‐43]. Currently, there is no standardized COS for studies evaluating interventions for paediatric IMD. The Canadian Inherited Metabolic Diseases Research Network (CIMDRN) is ideally suited to address this research gap. CIMDRN is a pan-Canadian multidisciplinary network of investigators funded by the Canadian Institutes of Health Research, with an established cohort of children born during the years 2006–2015, diagnosed with one of 31 targeted IMD, and treated at one of 13 participating Canadian centres [8]. CIMDRN collects observational data for participating children from a range of sources, including retrospective review of patient chart data. Developing and implementing standardized COS and data collection tools will enable the transition of the CIMDRN data platform to a registry format that can support registry-based randomized comparative effectiveness trials.

This protocol outlines a two-part study that will establish a COS for each of two of the most common IMD in children, phenylketonuria (PKU) and medium-chain acyl-CoA dehydrogenase (MCAD) deficiency (our study is registered in the COMET database, http://​comet-initiative.​org/​studies/​details/​995). The first part of the study is a knowledge synthesis project that includes a rapid review and development of an evidence map to identify a comprehensive listing of outcomes reported or suggested in past studies of PKU and MCAD deficiency. The review findings will be used to inform the second part of the study, a Delphi consensus process involving patients or their family members, healthcare providers, and health system decision-makers, to establish priority outcomes. The final COS will be used in future studies to evaluate therapies for children diagnosed with PKU or MCAD deficiency.

PKU is an inborn error of the phenylalanine hydroxylase enzyme resulting in elevated phenylalanine (phe) and reduced tyrosine concentrations in the blood and tissues, which if untreated is associated with both behavioural and intellectual disabilities. Current treatment involves dietary restriction of phe beginning in the newborn period [44]. While this strategy is largely successful in preventing the most negative outcomes, its implementation is challenging. Currently available phe-restricted formulae and foods are unpleasant with regard to taste and require daily calculation and meal planning of phe/protein intake, resulting in challenges to diet adherence for patients and their caregivers [45, 46]. There remains a relatively high prevalence of neuropsychiatric (features of attention deficit hyperactivity disorder (ADHD), anxiety, depression) and executive functioning problems in this population [47, 48]. Recent therapeutic developments include pharmacological treatments such as sapropterin dihydrochloride, and supplements such as large neutral amino acids; both are add-ons to standard diet therapy and have the potential to further reduce blood phe levels. Sapropterin dihydrochloride therapy has been found to improve blood phe levels, dietary phe tolerance, ADHD symptoms and executive function [49, 50] in a subset of patients who respond to this therapy, but has not been evaluated with respect to other key patient-oriented outcomes. Sapropterin has a high annual cost (range C$12 K to C$170 K in Canada from infancy to adulthood and dependent on body weight [51]) with variable reimbursement coverage within Canada and internationally. Large neutral amino acids have not to date been evaluated using robust trial methodology with patient-oriented outcomes [52‐54].

MCAD deficiency is an inborn error of the MCAD enzyme leading to impaired oxidation of medium-chain fatty acids. Affected patients are at risk of metabolic decompensation precipitated by catabolic stress during periods of prolonged fasting and fever. Clinical manifestations include hypoglycaemia, encephalopathy, cardiomyopathy, and cardiac arrest. Potentially lethal decompensation can be avoided by providing sufficient rapidly accessible caloric intake (e.g., carbohydrates) in at-risk situations. However, in addition to avoidance of prolonged fasting, it is unclear which, if any, interventions (e.g., routine dietary fat restriction, preventive use of cornstarch) have true preventive effectiveness. While daily oral L-carnitine supplementation is used for preventive management in some patients with MCAD deficiency, there is considerable practice variation [55] and concerns about potential adverse effects [56, 57]. To our knowledge there are no prospective trials (or other robust sources of evidence) on the use of L-carnitine for MCAD deficiency, nor studies that have focused on meaningful patient-oriented outcomes [58‐65].

PKU and MCAD deficiency represent two of the most common IMD and are associated with important health effects for both children and caregivers. There are sizable gaps in the knowledge related to the evaluation of new and existing therapies for both diseases. Given these important limitations, further research into both conditions is of great importance. Thus, we have prioritized these two IMD for the initial establishment of COS for implementation in future registry-based trials.

The rapid review and evidence map will address the following primary research question: “What outcomes are reported in key publications related to children diagnosed with phenylketonuria and medium-chain acyl-CoA dehydrogenase deficiency?” A secondary research question is “Among key publications as described above, what outcome measurement instruments are described or used to collect data for these outcomes?” The review will build on CIMDRN’s previous work, including a scoping review of patient/parent-reported outcomes in chronic paediatric illness [66] and qualitative studies of patient/family priorities [6, 66].

The rapid review protocol was prepared in consultation with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)-P checklist (Additional file 1.1) [67] and registered with PROSPERO (CRD42017073524). The review will be performed according to the methods outlined below, and any post-hoc protocol modifications will be described in the final report.

To establish consensus on COS for PKU and MCAD deficiency, we will use separate (disease-specific) Delphi surveys with patients/family members, healthcare providers, and health system decision-makers as survey participants, following methods recommended for COS development [32] and consensus approaches used by members of our research team [71, 72].