The Limb Girdle Muscular Dystrophy Health Index (LGMD-HI)

Abstract

Limb girdle muscular dystrophy (LGMD) is a debilitating disease and the fourth most common muscular dystrophy. This study describes the development of the LGMD-Health Index (LGMD-HI). Participants were aged >18 years and recruited from three LGMD registries and GRASP-LGMD consortium. The initial instrument, comprised of 16 thematic subscales and 161 items, underwent expert review resulting in item removal as well as confirmatory factor analysis followed by inter-rater reliability and internal consistency of the subscales. Following expert review, one subscale and 59 items were eliminated. Inter-rater reliability was assessed and five items were removed due to Cohen’s kappa <0.5. The final subscales had high internal consistencies with an average Cronbach alpha of 0.92. Test re-test reliability of the final instrument was high (intraclass correlation coefficient=0.97). Known groups validity testing showed a statistically significant difference in LGMD-HI scores amongst subjects based on ambulation status (28.7 vs 50.0, p<0.0001), but not sex, employment status, or genetic subtype. The final instrument is comprised of 15 subscales and 97 items. The LGMD-HI is a disease-specific, patient-reported outcome measure designed in compliance with published FDA guidelines. This instrument is capable of measuring burden of disease with no significant variation based on LGMD subtype.

1. Introduction

Limb girdle muscular dystrophy (LGMD) is a genetically diverse group of rare, autosomal muscular dystrophies caused by mutations in genes encoding proteins that are needed for muscle maintenance, repair and normal function.[1,2] Using next-generation sequencing (NGS), more than 30 subtypes of LGMD at different autosomal loci have been identified in LGMD.[2] Since 2018 these have been categorized with the nomenclature “LGMD D” for autosomal dominant and “LGMD R” for autosomal recessive mutations.[3] For example, LGMD R1 is caused by autosomal recessive mutations in the CAPN3 gene and LGMD R9 is caused by autosomal recessive mutations in the FKRP gene.[4] The 30+ subtypes of LGMD are characterized by a common phenotype: progressive weakness and wasting of the hip and/or shoulder muscles, onset after age two, and creatinine kinase elevation.[1] LGMD is the fourth most common muscular dystrophy with a prevalence of 1.63 per 100,000 individuals.[1] The variability in presentation and lack of diagnostic specificity create challenges when conducting and designing therapeutic trials in LGMD.

The Food and Drug Administration (FDA) has issued guidelines for creating patient-reported outcome measures designed to measure the burden of disease.[5] These instruments are structured questionnaires meant to give “reproducible, meaningful, quantitative assessments of how patients feel and how they function.”[6] There is currently no widely-used patient-reported outcome measure specific to LGMD. A 2017 study examined quality of life in patients with LGMD but did so using nonspecific instruments, the Short-Form-36 and the Individualized Neuromuscular Quality of Life (INQoL) questionnaire. [7,8] It is unlikely that these instruments were able to comprehensively evaluate the myriad of symptoms experienced by patients with LGMD.

This research group previously evaluated symptom frequency and impact on quality of life in patients with LGMD via a cross-sectional survey.[9] The survey results were used to create the Limb Girdle Muscular Dystrophy Health Index (LGMD-HI), a disease-specific patient-reported outcome measure. The purpose of LGMD-HI is to assess the overall disease burden associated with LGMD. The LGMD-HI was designed in compliance with FDA guidelines for use in therapeutic trials.[10] Here we describe the development of the LGMD-HI, including its content validation, test retest reliability, and ability to discriminate between groups thought to have different disease burden.

2. Patients and Methods

2.1. Eligibility Criteria

Participants were aged >18 years and had a diagnosis of genetically confirmed or clinically diagnosed LGMD.[1] Participants were recruited from three LGMD registries: the Coalition to Cure Calpain-3 registry ((https://www.curecalpain3.org/registry/), the Jain Foundation registry (https://dysferlinregistry.jain-foundation.org/), and the Global FKRP Registry (https://www.fkrp-registry.org/).[9] An additional 29–32 participants were recruited from ongoing Genetic Resolution and Assessments Solving Phenotypes in LGMD (GRASP-LGMD) consortium natural history studies (NCT 03981289, 04202627). In total, 163 individuals completed the initial survey. All research activities were approved by the institutional review board at Virginia Commonwealth University. All participants gave informed consent.

2.2. LGMD-HI Creation

Question Selection and Content Validity

We previously used structured interviews to develop a survey that identified the major symptoms experienced by patients with LGMD. The survey was administered to 134 individuals and the results were used to identify each symptom’s frequency and impact on quality of life.[9] This methodology has previously been used to assess major symptoms in different neurologic populations (Patient Reported Impact of Symptoms, or PRISM, studies).[11–16] In this project, the same survey with 16 sections and 161 questions was administered to a new cohort of 163 participants. The participants were asked to rate the impact of each symptom on a 6-point Likert scale, as previously described.[9] For each symptom, participants were asked “How much does the following impact your life now?” Participants responded on a scale of 1–6 as follows: (1) I don’t experience this; (2) I experience this but it does not affect my life; (3) It affects my life a little; (4) It affects my life moderately; (5) It affects my life very much; and (6) It affects my life severely.

Each symptom was assigned a population impact score based on its prevalence amongst survey participants as well as the relative impact of each symptom. Using the Likert scale above, both items 1 and 2 are scored a 0 and the remainder of the items are scored a 1, 2, 3, or 4 respectively. Therefore, while the Likert scale has six items, the population impact scale ranges from 0–4 with 4 representing the highest level of burden. Only questions with a population impact score >0.5 were included in the initial instrument. Subsequently questions were eliminated if the investigators agreed by full consensus that these were redundant, vague, not amenable to therapeutic intervention, or non-generalizable.

2.3. Confirmatory Factor Analysis

Questions were grouped based on content into 15 subscales representing LGMD themes and these subscales subsequently underwent confirmatory factor analysis (CFA). This process was followed by evaluation of the internal consistency of the subscales.

2.4. Test-Retest Reliability

Test-retest reliability of the LGMD-HI was assessed in 32 participants with LGMD. Participants were provided the survey in paper format. Participants completed the initial survey and received the exact same survey 14 days later but were not provided their original survey responses. Test-retest reliability was assessed using the total score across the instrument and each subscale score. The reliability of the instrument was quantified with intraclass correlation coefficients (ICC).

2.5. Scoring of the LGMD-HI

Each question in the LGMD-HI is weighted to reflect the corresponding symptom’s importance to survey respondents. Each question was weighted in proportion to its population impact score as determined in Kovalchik et al.[9] The weighted sum of responses was converted to a 0–100 scale by expressing the sum as a percentage of the total possible value with a score of 100 representing the most severe disease burden and a score of zero representing no burden. The weighted overall score is referred to as the LGMD-HI total score.

2.6. Known Groups Validity

The mean final LGMD-HI score was determined for predefined subgroups that may have different disease severities. Groups were defined by sex (male versus female), genetic subtype, ambulation status (defined as those who could walk independently versus those who could not walk or required assistive devices (e.g., walkers)), and employment (employed versus nonemployed). A two-sample t-test was used for the analysis of all groups with the exception of genetic subtypes, which were analyzed using a non-parametric one-way ANOVA due to small sample sizes in some groups.

3. Results

3.1. Patients

The demographic and clinical characteristics for patients who participated in the cross-sectional survey are provided in Table 1. The majority of patients (57%) possess the FKRP mutation.

Table 1.

Demographic and Clinical Characteristics of Participants

Variable	Cohort (n=163)
Age, median	37
Women, n (%)	95 (58.3%)
Race, n (%)
White	150 (92.0%)
Black	5 (3.1%)
Asian	2 (1.2%)
Other/Not Specified	6 (3.7%)
Ethnicity, n (%)
Hispanic	8 (4.9%)
Non-Hispanic	152 (93.3%)
Not Specified	3 (1.8%)
Age of Symptom Onset, median	14.5
Age of Diagnosis, median	25
Duration of disease, average years	10.3
LGMD Mutation, n
FKRP	93
CAPN3	29
DNAJB6	10
ANO5	9
DYSF	4
SGCA	4
SGCG	3
SGCB	1
DVSF	1
c.826C>A	1
Not Specified	8
Homozygosity, n (%)
Homozygous	79 (48.5%)
Non-homozygous	74 (45.4%)
Not Specified	10 (6.1%)
Employment, n (%)
Working	85 (52.1%)
Not Working	53 (32.5%)
Not Specified	25 (15.3%)
Ambulation, n (%)
Ambulatory	98 (60.1%)
Non-Ambulatory	64 (39.3%)
Not Specified	1 (0.6%)

3.2. Initial Question Selection

The initial survey included 161 items representing 16 themes. We eliminated 59 questions and one subscale based off lack of perceived responsiveness to intervention (24), lack of generalizability (22), vague wording (7), unclear language (4), and redundancy (2).

3.3. Confirmatory Factor Analysis

The initial survey results from all participants were used for confirmatory factor analysis. Sixteen factors were identified. Inter-rater reliability was assessed using Cohen’s kappa, and five items were removed due to kappa <0.5 (see Fig 1). The five questions were: “Weak Grasp” (kappa=0.45), “Muscle Aching with Activity” (kappa=0.45), “Stress” (kappa=0.47), “Shortness of Breath with Activity” (kappa=0.47) and “Shoulder Weakness” (kappa=0.495). The final subscales had high internal consistencies with an average Cronbach alpha of 0.92. The internal consistency of each subscale is shown in Table 2.

Figure 1.

Instrument Development

Table 2:

Internal Consistency of Final Instrument Subscales (15 Themes, 97 Questions)

Section	Questions, n	Internal Consistency, Cronbach Alpha
2: Ambulation	14	0.97
3: Lower Extremity Function	8	0.92
4: Hand Function	4	0.83
5: Arm and Neck Strength	3	0.8
6: Emotional Health	16	0.96
7: Embarrassment of Gait	1	N/A
8: Social Performance	8	0.9
9: Social Satisfaction	8	0.96
10: Daytime Sleepiness	1	N/A
11: Activity Participation	16	0.96
12: Fatigue	3	0.87
13: Pain	12	0.95
14: Swallowing	1	N/A
15: Breathing	1	N/A
16: Heart Palpitations	1	N/A

3.4. Test-Retest Reliability

32 participants completed the LGMD-HI twice over a 14 day time period. Participants were not allowed access to their prior answers during the second administration of the LGMD-HI. The test re-test reliability for the overall instrument was high (ICC=0.97)(Figure 2). The ICC scores ranged from 0.56 to 0.96 among the individual items. Following the removal of the five questions that had a Cohen’s kappa below 0.5, the ICC scores ranged from 0.61 to 0.96 (see Table 2). The final version of the LGMD-HI consists of 15 subscales and 97 questions that measure symptoms of importance to individuals with LGMD. Survey participants utilized a 6-point Likert scale when responding to each question to demonstrate the current impact of symptoms on their quality of life. The Likert scale and the survey format have been previously validated through qualitative assessment in multinational populations with neuromuscular disorders.[11–21] The measured subscales and the specific questions associated with each are provided in Table 2.

Figure 2.

Test-Retest Reliability of the LGMD-HI

3.5. Known Groups Validity and Concurrent Validity

There was no statistically significant difference in the final LGMD-HI scores amongst subjects separated by sex, employment status or genetic subtype. There was a statistically significant difference in the final LGMD-HI scores amongst ambulatory versus non-ambulatory subjects (28.7 vs 50.0, p<0.0001) (Figure 3, Table 3). The median LGMD-HI score varied amongst groups with different genetic subtypes but this variation was not statistically significant (Figure 4, Table 4). There was a weak, positive association between disease duration and final score (r = 0.28, p-value = 0.0003) demonstrating that as disease duration increases, the LGMD-HI score also increases. A subsequent simple linear regression indicates that approximately 7% of the variability in the LGMD-HI score is explained by disease duration.

Figure 3. LGMD-HI by Known Groups: Sex, Ambulation and Employment.

*indicates p<0.0001

Table 3:

LGMD-HI by Known Groups: Sex, Ambulation and Employment

	LGMD-HI score, Mean (SE)
Sex
Male (n=67)	36.08 (2.50)
Female (n=95)	37.83 (2.10)
Difference	1.75 (3.26)
Ambulation status
Ambulatory (n=98)	28.65 (1.77)
Non-ambulatory (n=64)	50.02 (2.20)
Difference	21.37 (2.82)*
Employment status
Employed (n=85)	35.84 (2.30)
Non-employed (n=53)	38.94 (2.91)
Difference	3.10 (3.71)

Figure 4.

LGMD-HI by Known Groups: Genetic Subtype

Table 4:

LGMD-HI by Known Groups: Genetic Subtypes

	LGMD-HI score, Median (Min, Max)
Genetic subtype
ANO5 (n=9)	20.97 (2.87, 64.42)
CAPN3 (n=29)	36.94 (1.85, 85.65)
DNAJB6 (n=10)	36.50 (1.34, 55.43)
DYSF (n=5)	46.84 (24.00, 86.83)
FKRP (n=93)	39.52 (0.10, 68.94)
SGC (n=8)	49.36 (6.05, 67.14)

4. Discussion

The Limb Girdle Muscular Dystrophy Health Index (LGMD-HI) is a disease-specific, patient-reported outcome measure designed for this diverse category of inherited muscular dystrophies. The creation of the LGMD-HI complied with FDA guidelines, producing an instrument designed to capture the disease burden of LGMD in the context of a clinical trial. This paper demonstrates that the LGMD-HI is able to capture disease burden in LGMD.

The known groups assessment of the LGMD-HI indicates a statistically significant difference in total LGMD-HI score for ambulating versus non-ambulating patients, as would be expected. There is no significant difference amongst males versus females. The data also demonstrate a weakly positive correlation between disease burden and LGMD-HI score, indicating that patients with more severe disease score higher on the instrument.

This paper also demonstrates that the LGMD-HI is capable of being used to measure disease burden in a variety of LGMD subtypes. The most represented subtype in this study is the FKRP mutation. While other subtypes were less robustly represented, we found no significant differences in LGMD-HI score among the six major LGMD subtypes. The LGMD-HI has some features that provide advantages over generic patient reported outcome measures. This disease-specific outcome measure focuses on questions that are most relevant to the LGMD population. Disease-specific instruments have been shown to improve sensitivity to change in clinical trials.[17–19] We anticipate that this will also be true for this instrument; however additional longitudinal studies are needed to confirm this.

There are some limitations to this study. Although the instrument was developed based on patients with LGMD from multiple different consortiums, we cannot state definitively that the population sampled represents the greater LGMD community. For example, in this study the FKRP mutation is overrepresented; the majority of patients (57%) in the study possess this mutation, although it is not the most common type of LGMD mutation worldwide. It is also possible that the group participating in the reliability study was not the same as that in the cross-sectional component. Additionally, LGMD has many subtypes and the less common subtypes are not as well represented in this health index. It is also important to note that 92% of the survey population identifies as White. While this number is consistent with a previous study from Moore et al showing that the majority of patients with LGMD in the United States are White (87%), LGMD is also diagnosed amongst other racial/ethnic groups; consider studies from Brazil, Mexico and South China.[22–25] Therefore the high proportion of White participants in this group may limit the applicability of the survey to general LGMD population. In addition, the responsiveness of this instrument is unknown. The known groups validity assessment provided a preliminary analysis, but the true responsiveness of disease progression may only be confirmed with a longitudinal study. Finally, the index takes about fifteen minutes to complete, which may potentially limit its use in some time-sensitive environments. Despite their length, instruments developed for other diseases using identical methodology have been found by patients to be more meaningful and relevant, have fewer irrelevant questions, be easier to use understand and complete, correlate better with strength and function, and be preferred overall compared to the shorter SF-36 and INQoL assessments.[26] In situations where a faster assessment is needed, the 16 item short form version of the LGMD-HI can be considered.

5. Conclusion

The Limb Girdle Muscular Dystrophy Health Index (LGMD-HI) is a disease-specific, patient-reported outcome measure designed in compliance with FDA guidelines. This instrument accurately captures the burden of disease amongst all survey respondents, the majority of which possess the FKRP mutation. The LGMD-HI is a reliable and validated tool designed for use in clinical trials.

Highlights.

• Patient reported outcome measure for all limb girdle muscular dystrophies (LGMDs)

• This outcome measure is specific to LGMD and measures 15 areas of disease burden

• This measure is designed for use in clinical trials to capture patient experience

Abbreviations:

CFA

Confirmatory factor analysis

GRASP-LGMD

Genetic Resolution and Assessments Solving Phenotypes in LGMD

ICC

intraclass correlation coefficient

LGMD

limb girdle muscular dystrophy

LGMD-HI

Limb Girdle Muscular Dystrophy-Health Index

PRISM

Patient Reported Impact of Symptoms