Health-Related Quality of Life in FKRP-Related Limb-Girdle Muscular Dystrophy R9

Participants

The present study is a part of “The Norwegian LGMDR9 cohort study” at the National Neuromuscular Centre Norway (NMK), University Hospital of North Norway (UNN). A total of 153 subjects (135 adults, i.e., 16+ years) in Norway with clinically and genetically confirmed LGMDR9 were invited to the cohort study. Of these, 101 subjects gave consent and participated [2]. The participants completed a study-specific questionnaire, and patient notes were collected from the specialist health care. All adults (16+ years) in the cohort (90/101) were invited to the present study of HRQoL.

Instruments

Procedure

The HRQoL instruments were distributed in paper versions by regular mail to adult participants of «The Norwegian LGMDR9 cohort study» at three consecutive time points: 2020-June, 2021-February, and 2021-August, respectively. One reminder was sent per administration. In addition, FSS and PSQI were administered one month after the second HRQoL collection. The second HRQoL collection was used for correlation analyses between HRQoL, FSS, and PSQI. Baseline HRQoL was used for correlation analyses between INQoL and SF-36 and for comparison with reference data and of subgroups except genotype subgroups (c.826 C > A homozygotes vs c.826 C > A compound heterozygotes). To enable inclusion of all compound heterozygotes, who were few, the first responses of each unique participant was used rather than the baseline HRQoL. Repeated measurements of HRQoL were applied to evaluate the level of fluctuations in cross-sectional measurements over time and to measure longitudinal changes at the individual level and the rank-order stability of the subscales (Sect. “Statistical analyses”). Social and clinical background data were collected from the questionnaire of the main study and included marital status, living situation, work status, educational achievement, activities of daily living (ADL), ambulation, use of Positive Airway Pressure (PAP) therapy (i.e., mask-based therapy for e.g., sleep apnea or hypoventilation), cardiac involvement, Percutaneous Endoscopic Gastrostomy (PEG), and consumption of pain medication for myalgia. Higher education was defined as at least one year at university or college. Age at disease onset, for calculation of disease duration, was primarily collected from patient notes. Disease onset was defined as the first clinical sign or symptom of muscle disease, such as muscle weakness, myalgia, or myoglobinuria and did not include incidental finding of elevated muscle enzymes. Norwegian normative SF-36 data, used as reference, were obtained from published material [28, 29], and social population statistics from Statistics Norway [30].

Statistical analyses

Data were analyzed using IBM SPSS Statistics for Windows (Version 27.0. Armonk, NY: IBM Corp.). Distribution of continuous variables are described using mean and standard deviation (±SD), or median and interquartile range (IQR), as appropriate. For simple group comparisons, independent t-test with bootstrapping or Kruskal-Wallis (multiple subgroups) was applied. Bootstrapping was always performed with 5000 resamples. In comparisons with normative SF-36 data, independent t-test was found acceptable given the sample sizes. In genotype comparison, multiple linear regression (MLR) with bootstrapping was applied to adjust for sex and age. To assess the change in INQoL domains from baseline to 14 months, paired t-test with bootstrapping (homozygotes), or paired Wilcoxon-signed rank test (compound heterozygotes due to a very limited sample size) was used. Categorical variables are presented as frequencies. Pearson chi-square was used for variable cell comparisons, and effect size is presented as odds ratio (OR) with confidence interval (CI). Bivariate correlations were examined with Pearson correlation (r) on normal distributions, otherwise with Spearman rank correlation (r_s). Within the context of the current study, we considered correlation coefficients  < 0.40 as weak, 0.40–0.69 as moderate,≥0.70 as strong, and≥0.90 as very strong.

For repeated measurements, we used generalized linear mixed models (GLMM) as it is a flexible regression analysis with regarding handling of missing data and data dependencies. The rank-order stability of the HRQoL domain scores was estimated as an Intraclass Correlation Coefficient (ICC) as the ratio between the intersubject variance (random intercept) and the total variance in a model without any fixed effects. ICC values < 0.50, 0.50 –0.75, 0.75 –0.90, > 0.90 may be considered as poor, moderate, good, or a high degree of rank-order stability, respectively. A high ICC thus means that if changes in HRQoL occur, the participants show the same change pattern with low variability in change across time. In order to regress the independent variables on the repeatedly measured dependent continuous variables of interest (perceived muscle weakness and the INQoL index), GLMM regression was applied. Choice of covariance structure for the residual matrix was decided based upon which structure that yielded the lowest Bayesian Information Criterion value. All GLMM analyses were performed with robust estimation of the standard errors. To identify predictors of dependent continuous variables measured once (FSS and PSQI), MLR with backward elimination was applied. The selection of independent variables in GLMM, or MLR before backward elimination, was based on the p-values of simple regression analyses of the variable of interest (p < 0.20 considered relevant). Age and sex were included as control variables. Assumptions of normality of the MLR residuals were checked using Q-Q plots, skewness, and kurtosis, homoscedasticity by scatter plotting predicted versus residual values, and influential cases by Cook's distance. The alpha level was set to p < 0.05. Considering the exploratory nature of the study and small subgroups (low power), correction for multiple testing was not used in this study [31].

Approvals and patient consent

All participants provided a written informed consent for the collection and use of clinical data. The study was approved by the Regional Committee for medical and health research ethics of Northern Norway (2018/1968/REK nord), and by the Data Protection Officer at UNN. The Norwegian version of SF-36 v. 1 was freely distributed by Knowledge Centre for the Health Services, Norwegian Institute of Public Health [32]. Permission to use the INQoL [20], the PSQI [33], and the FSS [34], respectively, was obtained.

Participants

The response rates are shown in Fig. 1. The overall HRQoL response rate was 84/90 (93%). Of all HRQoL respondents, 59 (70%) participated all three times, 16 (19%) twice, and 9 (11%) participated only once. Baseline characteristics of the sample comprising 75 c.826 C > A homozygotes and nine c.826 C > A compound heterozygotes are provided in Table 1. The compound heterozygotes had comparable disease duration to homozygotes but were younger (median 31 vs 49 years, p = 0.033). The status regarding work, educational achievement, ADL, ambulation, PAP therapy, and PEG, was relatively poorer among compound heterozygotes than homozygotes, whereas the frequency of cardiac involvement was comparable and the consumption of pain medication relatively higher among homozygotes. The subgroup of homozygotes was gender balanced, and males and females had comparable age and disease duration. Among homozygotes, females showed relatively poorer ADL and ambulation, a higher rate of PAP therapy, and a greater consumption of pain medication compared to males, whereas a higher proportion of males reported cardiac involvement. Comparison of compound heterozygous males and females was not applicable due to low sample size, heterogeneous genotypes, and unequal age distributions (Table 1 and 2).

Fig. 1

Flow chart.

Internal validity

Global scores for the PSQI, the FSS, and the HRQoL could be calculated for 89%, 97%, and 96–100% of the responses, respectively. SF-36 showed considerable floor effects on physical functioning, ceiling effects on social relationships, and both floor- and ceiling effects on role physical and role emotions (Table 3). The INQoL scale was reverse scored and had negligible ceiling effects but considerable floor effects on myalgia, fatigue, and dysphagia (Table 4). ICCs were moderate to high (0.62–0.85) on all SF-36 subscales except role emotions and change (Table 3). INQoL showed overall relatively higher ICCs (0.73–0.91) (Table 4) indicating higher rank-order stability across time. There were moderate to strong correlations between SF-36 and INQoL subscales sharing conceptual similarity (Table 5).

HRQoL over time

The measures of HRQoL at the three time points showed comparable mean scores (Table 3 and 4). The greatest difference between the highest and the lowest mean score was found on the subscales of bodily pain (4.9 points) and vitality (6.5 points) of the SF-36 (Table 3) and, correspondingly, myalgia (5.9 points) and fatigue (6.8 points) of the INQoL (Table 4). The change in the INQoL subscale scores from baseline to 14 months showed, in homozygotes, a worsening in muscle weakness and the INQoL index and a borderline significant worsening in activities (Table 4). Correspondingly, GLMM regression, including all three measurements of the homozygotes, showed an increase in muscle weakness (Table 6A), and an incremental tendency in the INQoL index (p = 0.08) (Table 6B). Five compound heterozygotes, all wheelchair dependent, completed the INQoL at baseline and 14 months. They showed a tendency towards worsening in muscle weakness but, more strongly, an improvement of the INQoL index (Table 4). On the particular SF-36 item concerning change in general health over the past year, worsening was reported by 42–57% of the homozygotes at the three respective measurements, compared to 17–29% of the compound heterozygotes. Among the homozygotes, 2–9% reported severe worsening, and 4–12% reported improvement. None of the compound heterozygotes reported severe worsening or any improvement.

HRQoL compared to normative data, and gender differences

Compared to the SF-36 normative data, the present study population showed significant impairments on all subscales except bodily pain, and the impairment of mental health in males was borderline significant (Table 3). Females showed relatively poorer scores both in the reference- and the present study population (not shown), but the female impairments were more pronounced among our participants compared to reference population (Table 3). Overall, homozygous females reported lower levels of HRQoL than homozygous males (Fig. 2A) and significantly on bodily pain, vitality, and social relationships (Table 3). Moreover, they showed a higher burden on several aspects of INQoL (Fig. 2B) and significantly on activities and independence (Table 4). GLMM analyses, including all three measurements of the homozygotes, showed that females tended to have more severe muscle weakness (Table 6A) and a significantly worse INQoL index than males (Table 6B).

HRQoL and genotypes

SF-36 showed no significant differences in HRQoL between the two genotype groups but a tendency towards poorer physical functioning, yet better social and mental HRQoL among the compound heterozygotes (Table 3, Fig. 2C). On INQoL, the compound heterozygotes reported significantly higher burden on the domains of dysphagia and independence (Table 4, Fig. 2D). The prevalence of dysphagia was 6/9 (67%) in compound heterozygotes compared to 14/80 (18%) in homozygotes.

Fig. 2

Box plot showing (A) SF-36 scores in FKRP c.826 C > A homozygous females vs males. (B) INQoL scores in FKRP c.826 C > A homozygous females vs males. (C) SF-36 scores in FKRP c.826 C > A homozygotes vs c.826 C > A compound heterozygotes. (D) INQoL scores in FKRP c.826 C > A homozygotes vs c.826 C > A compound heterozygotes. Notes: SF-36 scales range from lowest (0) to highest (100) HRQoL. The INQoL subscales range from no (0) to maximal (100) burden. FKRP c.826 C > A compound heterozygotes are significantly younger than c.826 C > A homozygotes (Table 1).

HRQoL and age/ disease duration

In homozygotes, both age and disease duration correlated positively with the following INQoL subscales: muscle weakness (p = 0.017 and p = 0.004), fatigue (p = 0.038 and p = 0.004), dysphagia (p = 0.030 and p < 0.001), activities (p = 0.036 and p = 0.011), independence (p = 0.002 and p < 0.001), and the INQoL index (p = 0.039 and p = 0.039) (r_s = 0.24–0.44). Duration was also positively correlated with the subscale of relationships (r_s = 0.31, p = 0.008). Myalgia, emotions, and body image did not correlate with age or disease duration in homozygotes. In compound heterozygotes, age and disease duration did not correlate with any of the INQoL subscales.

HRQoL and LGMDR9-related symptoms/ involvements

The frequency of any level of muscle weakness was 93–97%, fatigue 67–82%, myalgia 63–68%, and dysphagia 17–28% at the three respective measurements. Multivariate regression showed that muscle weakness and fatigue were independent predictors of the INQoL index (Table 6C). Comparing of the INQoL index of the independent walkers and the wheelchair and walking aid subgroups showed that the INQoL index was associated with wheelchair dependency (p = 0.002) and more strongly with walking aids (p < 0.0001) but did not differ significantly between the wheelchair and walking aid subgroups (p = 0.10).

Fatigue and subjective sleep quality

The FSS cut-off score gave a fatigue prevalence of 41% (14/34) in females and 38% (15/39) in males. FSS correlated positively with age, PSQI, and the INQoL domains of emotions, myalgia, and muscle weakness but was unrelated to sex. Only emotions (p = 0.007) and myalgia (p = 0.009) were identified as independent correlates (Supplementary Table 1). Median FSS was relatively higher in those using walking aids (5.3) compared to wheelchair (4.2) and independent walkers (4.3), but the subgroup differences were not significant. FSS correlated strongly with the fatigue scale of INQoL (r_s = 0.76, p < 0.0001), moderately with the INQoL index (r = 0.46, p < 0.0001) and, as expected, negatively with the vitality scale of SF-36 (r_s = –0.68, p < 0.0001).

Applying the PSQI cut-off score, 19/34 (56 %) of females and 21/37 (43 %) males were classified as poor sleepers. PSQI was related to myalgia (p = 0.003) but not age, sex, PAP therapy, wheelchair dependency, or the INQoL domains of muscle weakness or emotions (Supplementary Table 2).

HRQoL and socideomographic variables

The proportion of participants 30–69 years old living alone was 18/62 (29%), which is comparable to corresponding age groups in the general population (21–26%) [30]. Among participants age 25–39 years, 9/14 (64%) females and 5/14 (36%) males had higher education, which are relatively similar to the 60% of females and 40% of males reported in the general population [30].

The participants living alone did not differ on any aspect of HRQoL from those living with a partner or family. Nevertheless, there was a negative trend between living alone and most aspects of HRQoL except independence. In participants 25+ years, those with higher education showed poorer HRQoL on mental health (p = 0.047) and marginally for social function (p = 0.068) as well. The INQoL index was unrelated to educational achievement (Table 6C). Median age of participants younger than 67 years with long-term or permanent 100% incapacity for work or studies was 54.5 years (range 24–66). In the age group 25–67 years, 100% study/work incapacity was associated with lower level of education (OR 3.72, 95% CI: 1.31–10.53, p = 0.011) (Fig. 3), more severe bodily pain (p = 0.021) and myalgia (p = 0.007), and poorer general health (p = 0.035), role emotions (p = 0.018), activities (p = 0.006), and independence (p < 0.001). However, the INQoL index was unrelated to study/work incapacity.

Fig. 3

Bar chart showing the distribution of participants 25–67 years old with 100% long-term sickness absence or incapacity for work or studies, according to the highest educational level achieved (n = 69).

Supportive care

In total, 17/40 (43%) females and 21/44 (45%) males received physiotherapy treatment regularly. Reasons provided for not having physiotherapy included absent treatment effect (15/46), absent needs (11/46), costs (8/46), school/work (6/46), lack of available services (4/46), transportation/ travel difficulties (3/46), and 13/46 reported other various factors. Mental health treatment was received by 3/40 females and 3/43 males while 9/40 females and 4/43 males did not receive although they reported needs for mental health treatment.

Strengths and limitations

We believe the findings are representative of the Norwegian LGMDR9 population 16+ years given the nation-wide recruitment, high response rates, and gender balance. Measurement at three time points provided insight into the longitudinal stability and change at cohort and individual level. INQoL provided information about expression of disease-related symptoms, burden, and relationships with burden. Our analyses indicated good internal validity. The use of SF-36 enabled comparison with a reference population, which was useful in the interpretations of the INQoL with regard to impaired areas of HRQoL.

This study also has some important limitations. Due to the exploratory design we accepted a higher risk of conducting type I error (false positive) and urge for caution in interpreting p-values particularly in the zone of 0.01 to 0.05 as significant. Findings that are flagged as significant should be replicated in future studies. The sample size of c.826 C > A compound heterozygotes was low, which clearly limits the statistical power. The study was conducted during the COVID-19 pandemic, which may have introduced period-specific results generalizing less well to post-COVID life. The increase in perceived muscle weakness observed in homozygotes could potentially result from inactivity due to restrictions [50]. Nevertheless, since all data collections were performed in comparable and relatively COVID-19 free periods, the cross-sectional measurements were comparable, and no longitudinal change was detected in the aspects of relationships, myalgia, fatigue, or emotions at the individual level, we believe the pandemic had minor effects on the results. Lastly, interpretations of the prospective study were also hampered by the fact that INQoL has not been validated on sensitivity in LGMD. Neither a 6-year follow-up study of patients with Dystrophia myotonica type 1 [51] nor a 20-month follow-up study of patients with Oculopharyngeal muscular dystrophy [52] detected deterioration in the INQoL measurements despite objective progression of muscle weakness.

Conclusions

This study provided insight into the impacts of LGMDR9 on HRQoL, subjective changes over a 14-months period, and relationships with disease burden. Both physical, emotional, and social aspects of HRQoL were impaired, which advocates for a multi-disciplinary care. In c.826 C > A homozygotes, perceived muscle weakness and disease burden deteriorated during the study, which is relevant information for clinical trials. However, validation of sensitivity of the INQoL measure in LGMD remains and potential COVID-19 effects should be considered. Among homozygotes, females reported a higher burden, which indicates the need for a gender-specific perspective in treatment and care. Muscle weakness and fatigue proved particularly important where fatigue is amenable to treatment. Associations suggest that myalgia and mental distress are potential targets for the treatment of fatigue, but further studies are needed to establish evidence for effective interventions. Poor sleep was frequent and will be investigated further. The combined use of a generic and a disease-specific HRQoL is particularly useful for research purposes. However, a generic HRQoL that is brief to minimize patient burden and optimize response rate, constructed for utility analyses, and with accessible norms may be preferable. Adding an item of overall QoL or happiness would have provided insight into the significance of the disease in the patient' lives as a whole and can be considered in future research.