Real‐Life Effectiveness After Switching to Avalglucosidase Alfa in Late‐Onset Pompe Disease Patients Worsening on Alglucosidase Alfa Therapy: A French Cohort Study

ABSTRACT

Late‐onset Pompe disease (LOPD) is a progressive myopathy. Enzyme replacement therapy is effective, but long‐term outcomes vary. Avalglucosidase alfa, shown to be non‐inferior to alglucosidase alfa in a phase 3 trial, became available in France through compassionate use for patients with insufficient response to alglucosidase alfa.

Methods

Data from the French Pompe registry were analyzed for patients who switched to avalglucosidase alfa with at least 1 year of follow‐up. Respiratory function (forced vital capacity, FVC) and motor function evaluated with gait performance (Six‐Minute Walk Test, 6MWT) were assessed before the switch, and one and 2 years after. Individual changes were compared using paired‐sample tests.

Results

Forty‐seven adult patients were included. A stabilization of motor decline was observed: prior to switching, the 6MWT decreased by −27 m/year, whereas an improvement of +17 m/year was seen during the first year after the switch (p = 0.001), followed by overall stability in the second year (−10 m/year, p = 0.280). Respiratory changes were not statistically significant: a decline of 60 mL/year before the switch versus 10 mL/year after 1 year (p = 0.161), and 20 mL/year during the second year (p = 0.346). Three patients died during follow‐up, with causes unrelated to the disease or treatment.

Discussion

Gait deterioration halted during the first year after transitioning to avalglucosidase, with sustained stabilization thereafter, while respiratory parameters showed minimal change. For patients experiencing significant walking decline under alglucosidase alfa therapy, switching to avalglucosidase alfa resulted in disease stabilization, beginning with mild improvement in the first year and a return to pre‐switch baseline thereafter.

Illustrating motor function changes before the switch, during the first year, and during the second year after switching treatment.

1. Introduction

Late‐onset Pompe disease (LOPD) is a recessively inherited glycogen storage disorder (GSDII) (MIM#232300) caused by a deficiency of the lysosomal enzyme acid alpha‐glucosidase (GAA). This deficiency leads to glycogen accumulation and disrupted autophagy in skeletal muscle, resulting in progressive proximal and axial muscle weakness, as well as diaphragmatic involvement that may cause respiratory insufficiency [1]. Since its introduction in 2006, enzyme replacement therapy (ERT) with recombinant human GAA (rhGAA), alglucosidase alfa, has shown the ability to improve or stabilize disease progression in LOPD patients. However, many patients experience clinical decline after 3–5 years of treatment [1, 2, 3, 4, 5].

Avalglucosidase alfa is a next‐generation rhGAA ERT developed for improved uptake via the mannose‐6‐phosphate (M6P) receptor [6, 7, 8]. The phase 3 COMET trial compared the efficacy of both ERTs in treatment‐naïve patients and demonstrated that avalglucosidase alfa was non‐inferior to alglucosidase alfa. Moreover, after 49 weeks of treatment, patients on avalglucosidase alfa showed greater improvements in forced vital capacity (FVC, percent predicted) and Six‐Minute Walk Test (6MWT) distance than patients on alglucosidase alfa [9], globally stable in the extension [10]. Long‐term extension of the previous ERT clinical trials (17 patients) has shown a global stability of respiratory parameters (−0.64% of the predicted forced vital capacity per year up to 6.5 years in the switched group) and gait parameters over time (−0.85% of the predicted 6‐min walking test per year, up to 6.5 years in the switched group) [11]. Concerning the patients who switched from alglucosidase alfa to avalglucosidase alfa after 49 weeks in the COMET trial, they improved at the 97th week (i.e., about 1 year after the switch) their 6MWT of 5 (±11) m (385 ± 140 at the switch moment) [10], maintained at week 145 (about 1.5 year after the switch) [12]. Concerning respiratory data, for the 24 patients having a deterioration of the FVC (−3.4 (1.0)% the first year under alglucosidase alfa), they stabilized it after the switch (+0.3 (0.7)% the year after) [13].

Based on these findings, avalglucosidase alfa received regulatory approval in several countries for the treatment of LOPD. In France, as of June 11, 2020, avalglucosidase alfa became available through compassionate use for patients previously treated with alglucosidase alfa who showed inadequate response. This access was later extended through an early access program starting on September 22, 2022 (“Haute Autorité de Santé”, Decision 2022.0330/DC/SEM). Criteria for treatment failure were defined during monthly meetings of the French neuromuscular network (FILNEMUS) and included significant respiratory and/or motor decline after at least 1 year of alglucosidase alfa treatment, not explained by other factors such as severe COVID‐19, intercurrent diseases (severe depression, cancer, fractures, etc.), significant weight changes, or physiotherapy interruption. The initial response to alglucosidase alfa was also considered using data from the French Pompe registry—for instance, a good early response was defined as an improvement of ≥ 1.4% per year in the 6MWT during the first 2 years of treatment [1, 14].

This study aimed to provide real‐world evidence on the effectiveness of switching therapy in a selected adult population with LOPD who showed an inadequate response to alglucosidase alfa, using data from the French Pompe registry. We previously reported the 29 first patients switched in France with a global stabilization of motor worsening 1 year after the switch [15]. Reporting these findings and evolution over time is crucial, as all participants in the COMET trial were treatment‐naïve [9]. In this paper, we explore the benefits of avalglucosidase alfa in patients who experienced secondary loss of response or a waning response over time to alglucosidase alfa in our registry, with a longer analysis window of 2 years.

2. Methods

2.1. The French Pompe Registry

Data were extracted from the French Pompe Registry (about 150 active patients in 2019–2020). This registry falls within the scope of Chapter IX of the French Data Protection Act and follows the Recommendations of Ethics and Good Practices in Epidemiology (version France—2007). The protocol, patient information note/consent, and all other relevant documents have been submitted to the competent authorities for approval in accordance with national legislation. In particular, the authorizations of the National Commission for Information Technology and Freedoms (CNIL No. 909139) and the Consultative Committee for the Treatment of Information in the field of Health (CCTIRS No. 08.461) were obtained.

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

2.2. Switch Decision

For each patient considered for a treatment switch, individual trajectories of motor and respiratory parameters were carefully reviewed in a multidisciplinary meeting to determine whether the switch was appropriate (see Data S1 for examples). At least 1 year of previous ERT therapy was necessary to decide if the patient was an alglucosidase alfa responder.

If there was a clear confounding factor (typically a hip fracture), an early reassessment was requested at 3 and 6 months in order to assess whether the patient was improving again with rehabilitative therapies and taking care of the confounding factor, or whether the worsening continued after its treatment/management. For recurrent pneumonitis (or a respiratory comorbidity such as chronic obstructive pulmonary disease), the same procedure was followed with several assessments at a distance from the acute episodes to have a measurement linked as much as possible only to the specific evolution of Pompe disease.

2.3. Endpoints

Respiratory function was assessed by measuring the FVC (in a sitting position, expressed as volume), whereas motor function was assessed with the 6MWT (m) [10]. In order to better capture the individual variations, we also expressed respective variations of the FVC and the 6MWT for each patient for 1 year (deltaFVC and delta6MWT, respectively, in liters and meters), the year before the switch (pre‐), and the year just after the switch (post1‐), and the further year after (post2‐). For instance, the post1‐delta6MWT for one patient was “6MWT 1 year after‐ 6MWT at the switch moment (m)”. A positive difference indicated an improvement, both for FVC and 6MWT, whereas a negative one showed a worsening over time. Finally, to compare the characteristics of good and bad responders, we considered a significant change in the slope of 10% after the switch [16, 17]. It was determined a “(post1‐pre1)/T6MWT at the switch” (see Data S2 for examples).

We also explored the Motor Function Measure (MFM), which is a functional scale designed to evaluate motor function in individuals with neuromuscular disorders. It assesses three functional domains: standing and transfers (D1), axial and proximal motor function (D2), and distal motor function (D3). It includes 32 items scored on a scale from 0 to 3, where zero indicates that the task cannot be performed and three indicates that the task is performed with no difficulty and no compensation. The total score ranges from 0 to 100%, with higher scores indicating better motor function [18].

The Rasch‐Built Pompe‐Specific Activity scale (R‐PAct) [19] was also used to carry out daily life activities and the social participation of patients. R‐PAct consists of 18 questions scored on a scale from 0 to 2, with 0 = no, 1 = yes, but with difficulty, and 2 = yes, without difficulty. The total R‐PAct score is based on the summed‐up score across 18 items, which ranges from 0 to 36, with higher scores indicating less impact of the disease.

2.4. Statistical Analysis

The results are presented as the mean and standard deviation for continuous variables with a normal distribution (according to the Shapiro–Wilk test) and as the median and range for non‐normally distributed continuous variables. FVC and 6MWT, as the deltaFVC and delta6MWT, were compared before the switch and 1 year later with a paired sample t‐test (Student's or Wilcoxon according to normality), and then between 1 and 2 years.

Categorical variables are expressed as the number and percentage. The significance threshold was set to p = 0.05 for all analyses. Statistical analyses were performed using JASP version 0.18.1 (JASP Team, 2020).

To explore the parameters associated with the switch response, we compared parameters between bad and good responders with a t‐test. We also used bivariate Pearson correlation between clinical parameters and the clinical motor response “(post1‐pre1)/T6MWT”. Collinearity between pairs of independent variables was used (correlation coefficients, p < 0.1). To select potentially predictive factors, backward elimination was then used to obtain the best explanatory model.

3. Results

A total of 47 patients were included in this study, 29 of whom had 2 years of follow‐up after the switch. These patients were aged 43.0 (±13) years at the time of diagnosis, with a mean evolution of 10 (±14) years between symptoms and diagnosis, 10.0 (±4.5) years of prior treatment. Their weight was stable over time (p ≥ 0.05): 71.3 (±17.9) kg, BMI 25.17 (±6.07) kg/m². Main characteristics are presented in Table 1.

TABLE 1.

Main disease characteristics of patients who switched treatment.

	1 year pre‐switch	At the switch	1 year post‐switch	2 years post‐switch
Motor parameters
6MWT (m)	255 [165; 407]	237 [133; 360]	245 [147; 426]	163 [104; 348]
Total MFM score (%)	72 (±15)	70 (±16)	71 (±15)	67 (±14)
D1 subscore (gait), in %	44 [16; 67]	39 [18; 69]	36 [21; 64]	23 [15; 51]
D2 subscore (proximal), in %	90 [79; 97]	92 [81; 97]	89 [78; 97]	86 [76; 94]
D3 (distal), in %	100 [95; 100]	100 [95; 100]	100 [95; 100]	100 [95; 100]
Delta‐MWT (in m, negative value: decrease)*	−27 (±49)		17 (±67)	−10 (±44)
Respiratory parameters
FVC (in liters)	2.19 (±1.24)	2.19 (±1.24)	2.15 (±1.18)	1.95 (±0.85)
FVC (in % of predicted values)	60.77 (±26.81)	61 (±27)	60 (±25)	60 (±24)
Functional parameters
R‐PACT36	12.0 [7.8; 21.8]	13.5 [9.3; 20.8]	11.0 [9.0; 19.0]	10.0 [8.8; 17.5]

*p < 0.05.

For previous data, 1 year before the switch, the 6MWT was 255 [165; 407] meters, the R‐PACT36 was 12.0 [7.8; 21.8], the total MFM score 72 (±15) % with respectively D1 subscore (gait) 44 [16; 67] %, D2 subscore (proximal) 90 [79; 97] % and D3 (distal) 100 [95; 100] %. At this time, the FVC was 1.88 [1.24; 2.91] liters corresponding to 60.77 (±26.81) percent predicted values (mean 2193 (±1240) mL). The MIP was 32 [26; 54] and MEP 46 [38; 66] cm H₂O.

The longitudinal 6MWT was not different (p = 0.66): 272 (±169) m at the time of the switch versus 269 (±170) m 1 year after the switch and 2 years after the switch: 219 (±178) m (p = 0.27). The FVC at the time of the switch was 2.19 (±1.24) L and was stable (p = 0.08) 1 year after the switch at 2.15 (±1.18) L and 2 years after (p = 0.69): 1.95 (±0.85) L, corresponding successively to 61 (±27), 60 (±25) and 60 (±24) percent predicted values (p ≥ 0.05).

Concerning the differences at the individual levels, the delta6MWT showed a decrease of 27 (±49) m during pre1 versus an increase of 17 (±67) m for post1 (p = 0.001), and a further stabilization: −10 (±44) m for post2 (p = 0.28) (Figure 1a). On the contrary, the deltaFVC was stable, showing a decrease of 62 (±247) mL during the pre1 period versus a decrease of 13 (±395) mL during post1 (p = 0.16) and 22 (±332) mL during post (p = 0.35) (Figure 1b).

FIGURE 1.

Differences in and Six‐Minute Walk Test (6MWT) (a) the forced vital capacity (FVC) (b) the year before the switch (left, Pre1), the year after the switch (middle, Post1), and 2 years after (right, Post2). A negative value indicates a worsening, whereas a positive value indicates an improvement. The graphs show a point‐per‐point representation. Note the significant change in 6MWT 1 year after the switch with a stabilization of the gait function, sustained at 2 years.

Three patients died along the follow‐up, not related to the disease or treatment.

Concerning the motor evaluation with the MFM (available for 36 patients), the D1 subscore in percentage slightly worsened before switch: 42 (±2) vs. 41 (±27) (p = 0.03) and then stayed stable at 41 (±26) the year after and 34 (±26) the second year (p ≥ 0.05) reflecting on the total MFM score in percentage: 72 (±15) the year before the switch with a slight worsening at the moment of the switch 70 (±16) (p = 0.02), and a further stabilization 71 (±15) the year after and 67 (±14) 2 years after (p ≥ 0.05). No changes were observed along time for the D2 dimension. No differences were observed concerning the delta pre and post MFM (D1, D2 nor total in percentage).

The R‐PACT36 (available for 27 patients at 1 year, 17 at 2 years) was also stable according time: 14.8 (±14.1) 1 year before the switch, 15.2 (±6.9) at the time of switch, 12.9 (±7.8) the year after and 13.0 (±7.5) 2 years after, representing a deltaPre1 of −0.7 (±3.9), deltaPost1 of −0.1 (±1.9) and deltaPost2 of 0.1 (± 3.1) (p ≥ 0.05). In particular, 10 patients had a deltaPre1 null, 11 had a worsening (from −10 points to −1 points) and 7 patients had an improvement (from 2 to 11 points) whereas 13 had a deltaPost1 null, 7 a worsening (from −4 to −1) and 7 an improvement (from 1 to 4 points).

Concerning the individual rate of response on the 6MWT, 20 patients were good responders with an increase from 10.58% up to 209.09% increase (mean 48.46%) and 3 bad responders with a decrease from −43.81% up to −84.39% (mean −64.62%) (other patients stable, i.e., variations <|10|%). The significant differences between good responders and others were (i) the disease duration at the time of switch, shorter in good responders: 20.2 (±7.7) years vs. 26.7 (±9.5) years (t = 2.17, p = 0.038) (ii) the CK levels, lower in good responders 316 (±197) UI/mL vs. 545 (±356) UI/mL (t = 2.20, p = 0.037) and the 6MWT performance at the time of switch, lower in good responders 225 (±130) m vs. 328 (±156) m (t = 2.05, p = 0.049). On the other hand, the significant differences between poor responders and others were also related to the age at the switch (t = −3.27, p = 0.003), disease duration (t = −2.94, p = 0.006), and the duration of alglucosidase alfa prior exposure (t = −2.94, p = 0.006), as well as the delta MFM (D1 and total) the year before the switch (respectively t = 2.66, p = 0.013; t = 2.34, p = 0.027): 76 (±10) vs. 52 (±12) years at the time of switch, with 36 (±10) years of disease duration (compared to 21 (±8) years) and 15 (±0.8) years of prior enzyme therapy (vs. 8 (±4) years). The MFMD1 score of bad responders the year before the switch was already severely affected, and more than for others patients: 15 (±10)% versus 54 (±24) and MFMtotal scores 56 (±11)% vs. 77 (±13)%.

The delta6MWT (post1‐pre1) was correlated with the pre1T6M (rho = −0.518, p = 0.003), i.e., the response post‐switch was greater the more significant the worsening the year before the switch was. It also correlated with the BMI (rho = 0.401, p = 0.021), i.e., the response was better for patients with higher BMI.

A three‐step backward regression analysis yielded a model with three predictors explaining 55.2% of variance (AIC = 321.8; BIC = 329.0; F = 11.11; p < 0.001): BMI (p < 0.001), age at the switch (p = 0.011), pre1‐6MWT (p < 0.001). Younger age, more severe pre‐switch 6MWT worsening, and a higher BMI were associated with a better response.

4. Discussion

This study highlights the effect of avalgucosidase alfa in LOPD patients previously treated who showed a waning response over time to alglucosidase alfa, based on an analysis of clinical data prospectively collected in the French Pompe registry. We previously reported a halt in gait worsening [15], and in this study, we confirm these data with more patients (47). We also proposed a longer follow‐up of 2 years for previously switched patients with an observed stabilization. Concerning respiratory data, no significant modifications were observed, but patients were most often switched for motor rather than respiratory worsening.

These results provide additional efficacy information on this “next‐generation” ERT in real life for patients suffering from advanced LOPD and declining despite alglucosidase alfa therapy, which could guide the treatment‐switching decision. However, we observed various responses within our cohort of switched patients, with 20 patients (43%) showing a significant improvement (more than 10% improvement on the delta6MWT) and three patients (6%) showing a considerable worsening (more than 10% worsening in their delta6MWT), emphasizing the need to better understand the prognostic factors of the therapeutic response.

Concerning the MFM subscore, the previous worsening before switch was observed on D1 subscore, which observes the gait and stand‐up position, and this was also slightly but significantly improved. Other dimensions (proximal muscles and distal ones) were not modified, neither before the switch nor after.

Our findings showed no meaningful change in FVC, with about 1% of variation in percent predicted FVC. It's difficult to determine a significant threshold of response. Some pulmonary studies sometimes use stringent 10% [17] or 6% [20] thresholds, while research in Pompe disease may rely solely on the direction of change—positive or negative— [21] or even consider variations as small as 3% [13]. Three percent of percent predicted is the within‐participant clinically meaningful threshold found in the COMET trial [22]. Recently, the minimum clinically important difference indeed has been determined at 3.4% in the lung fibrosis [23] and this threshold could also be investigated in further studies.

The variations of the 6MWT was sometimes important (up to 200%). However, this test was conducted using a standardized protocol. Participants were instructed to “walk as far as possible” along a 25‐m corridor for six minutes. In line with the American Thoracic Society guidelines [24], they were permitted to slow down or stop and rest if needed, but were encouraged to resume walking as soon as possible. When necessary, participants could use their usual assistive devices. Throughout the test, verbal encouragement (e.g., “You're doing well!”) was provided every 60 s. All patients included in this study were already familiar with the test, having performed it for several years, making any learning effect negligible [25]. However, such variability may still raise the question of fatigue or a potential placebo or motivational effect. Despite its limitations, this outcome measure remains the most widely validated in Pompe disease.

Surprisingly, the R‐Pact scale scores didn't change significantly, either before or after the switch. The gait speed worsening was not associated with a further deterioration of functional activities, reflecting the clinical experiences shared during the collegial discussions. This is evidenced by the fact that patients rarely ask for switching treatment because they don't experience a worsening (maybe because very insidious and progressive) of their daily activities that are already limited. Once again, it raises the question of our evaluation criteria, maybe more objective but not enough clinically relevant. The R‐PAct was already severely impacted (mean 15 out of 36), corresponding to walking difficulties with the use of devices and even intermittent wheelchair use. It seems that the score is not sensitive enough to very light tendencies (not significant worsening before the switch of about 1 point and no more after); moreover, when the patients already have lost several motor abilities. Other patient‐related outcomes to observe the response to treatment have to be developed. The minimal clinically significant difference of R‐PACT is not yet determined [19, 26], however, in the evaluation of cipaglucosidase (another ERT for Pompe disease), the cut‐off of > 0 has been used with a difference of 0.2 points between the two enzymotherapies [27]. Another explanation, nevertheless, could be related to the missing data (only 27 patients of all the cohort had repeated R‐PAct). Changes in R‐PAct were inconsistent, with some patients showing improvements prior to the treatment switch. Consequently, the interpretation of these limited results remains challenging and may be influenced by compensatory strategies, rehabilitation, or other confounding factors. Concerning the factors associated with a good motor response, a shorter disease duration with lower CK levels could reflect a less aggressive disease. However, the 6MWT performance was also more impaired in the good responders, possibly due to a statistical artifact related to a greater improvement threshold. These subdata have to be cautiously considered. Nevertheless, the three patients who continued to significantly worsen were very aged with a very severe disease (maybe outdated stage for enzymotherapy). The three poor responders had a longer disease duration (by comparisons with others), were older, and had a more severe motor impairment (on MFMD1 evaluating gait and transfers) and probably were too advanced in their disease for ERT alone to halt muscle loss. We hoped that muscle MRI could be a valuable tool in the future for differentiating the various pathophysiological processes involved (energy deficit linked to glycolysis abnormalities, autophagy [28], secondary muscular dystrophy, glycogen overload, etc.) and determining the most appropriate treatments. Regarding the regression analysis, younger age and higher BMI were associated with a better response, supporting the idea of a narrow switching window before the disease progresses to advanced stages with significant muscle wasting. Concerning the impact of pre‐switch decline in the 6MWT, patients who experienced recent and marked deterioration (similar to evolutionary “spurts”) may be more likely to show recovery. However, defining response based on the comparison between pre‐ and post‐switch slopes introduces a methodological bias when including this parameter as a predictor of response.

In the Phase 3 COMET trial, switching to avalglucosidase was protocol‐driven after 1 year of alglucosidase alfa exposure. In contrast, in our real‐world cohort, switching occurred in the context of clinical decline, representing a markedly different setting. This distinction may inform the optimal timing and patient selection for treatment switching. Regarding respiratory outcomes, switching may be more beneficial earlier in the disease course [13] than in patients who have already declined (as suggested by our data). With regard to motor outcomes, the only conclusion that can be drawn is that patients treated with avalglucosidase during the first year of therapy did not exhibit any worsening after 3 years of follow‐up [12]. In contrast, some patients receiving alglucosidase alfa—both in the present cohort and in previously reported cohorts [2, 14, 29]—showed a decline in motor function, suggesting that switching therapy may be beneficial for these patients, as described recently by other ERT switches [30].

The main limitation of this study is related to the number of patients. Only 29 patients were followed for 2 years after changing their enzymotherapy. However, other studies or real‐world experience are even more limited [12, 31]. Moreover, patients were generally switched due to motor decline, which could partly explain the lack of observable changes in respiratory outcomes, with effects mainly seen in motor parameters. Furthermore, MEP and MIP (as well as supine FVC) are more sensitive indicators of respiratory variation than FVC; however, these measures were generally unavailable in real‐world data.

In conclusion, we have demonstrated in this real‐life study that patients suffering from LOPD on prior enzymotherapy with alglucosidase alfa, having been switched to avalglucosidase alfa because of respiratory and/or motor worsening, present a significant stabilization of the motor decline, sustained at 2 years. On the respiratory level, no significant change was observed after 1 or 2 years since the switch. Longer follow‐up assessments over several years, ideally including data from other countries, will be needed in order to draw conclusions in larger cohorts.

Author Contributions

Maxime Fournier: investigation, validation, visualization, writing – review and editing. Françoise Bouhour: investigation, visualization, writing – review and editing, validation. Nadjib Taouagh: funding acquisition, data curation, supervision, resources, writing – review and editing. Florence Demurger: investigation, validation, visualization, writing – review and editing. Emmeline Lagrange: investigation, validation, visualization, writing – review and editing. Stephane Beltran: investigation, validation, visualization, writing – review and editing. Céline Tard: conceptualization, investigation, writing – original draft, methodology, validation, visualization, formal analysis. Jean‐Baptiste Noury: investigation, validation, visualization, writing – review and editing. Maud Michaud: investigation, validation, visualization, writing – review and editing. Sabrina Sacconi: investigation, validation, visualization, writing – review and editing. Pascal Cintas: investigation, validation, writing – review and editing, visualization. Dimitri Renard: investigation, validation, visualization, writing – review and editing. Claire Lefeuvre: investigation, validation, visualization, writing – review and editing. Azzeddine Arrassi: investigation, funding acquisition, data curation, supervision, resources. Armelle Magot: validation, visualization, writing – review and editing, investigation. Emmanuelle Salort‐Campana: investigation, validation, writing – review and editing, visualization. Shahram Attarian: investigation, validation, visualization, writing – review and editing. Jean‐Baptiste Davion: investigation, validation, visualization, writing – review and editing. Joëlle Deibener‐Kaminsky: investigation, validation, visualization. Diana Chitimus: investigation, validation, writing – review and editing, visualization. Marco Spinazzi: investigation, validation, visualization, writing – review and editing. Dalil Hamroun: investigation, funding acquisition, data curation, supervision, project administration, resources, software. Pascal Laforêt: investigation, conceptualization, funding acquisition, writing – review and editing, validation, project administration, supervision, resources. Sylvain Nollet: validation, visualization, writing – review and editing, investigation.

Funding

This work was supported by Sanofi, Amicus Therapeutics and Astellas Pharma Europe.

Conflicts of Interest

C. Tard declares personal fees and non‐financial support from Sanofi‐Genzyme, Amicus, non‐financial support from Astellas. J.B. Noury declares consulting fees from Sanofi‐Genzyme and Amicus, support for attending meetings from Sanofi‐Genzyme. F. Bouhour declares personal fees and non‐financial support from Sanofi‐Genzyme, Amicus. M. Michaud declares non‐financial support from Sanofi. P. Cintas declares personal fees and non‐financial support from Sanofi‐Genzyme. M. Spinazzi reports personal stock ownership in Sanofi. M. Fournier declares personal fees from Sanofi. C. Lefeuvre declares personal fees and non‐financial support from Sanofi‐Genzyme, and fees from Astellas. E. Salort‐Campana declares personal fees and non financial support from Sanofi‐Genzyme, Amicus. P. Laforet declares personal fees and non‐financial support from Amicus Therepeutics, Astellas, Sanofi and Shionogi. A. Magot reports support for conference attendance from Sanofi‐Genzyme, and serves as an investigator for Astellas. S. Beltran, F. Demurger, D. Renard: no disclosures.

Supporting information

Data S1: Examples of individual disease trajectories of four switched patients.

Data S2: Examples of individual disease trajectories of four switched patients after the switch.

Appendix A.

Pompe Study Group: D. Orlikowski⁴, M. Masingue²¹, T. Stojkovic²¹, A. Béhin²¹, F. Lallement²⁸, L. Kouton²⁰, A. Verschueren²⁰, Al. Kaminsky⁶, G. Sole¹³, F. Duval¹³, F. Bouibede²⁶, M. C. Minot¹³, Y. Pereon¹⁵, B. Acket¹⁶, A. Nadaj‐Pakleza², R. J. Morales²³, F. Esselin¹³, R. Jaussaud²⁹, K. Ghorab¹³, L. Magy¹³, E. Krim²⁷, C. Dahan²⁷, F. Taithe⁶, A. L. Bedat‐Millet², S. Toquet², E. Diab², M. Maillet‐Vioud³⁰, J. Y. Hogrel²¹, A. Echaniz‐Laguna³¹, R. Thomas¹, G. Nicolas⁴

¹Inserm, CHU Lille, U1172—LilNCog—Lille Neuroscience and Cognition, University of Lille, Lille, France; ²Centre de Référence des Maladies NeuroMusculaires Nord—Est—Ile‐de‐France, Filnemus, Garches, France; ⁴Neurology Department, APHP, Raymond Poincaré University Hospital, FHU PHENIX, Garches, France; ⁶Centre de Référence des Maladies NeuroMusculaires PACA Réunion Rhône Alpes, Filnemus, Marseille, France; ¹³Centre de Référence des Maladies NeuroMusculaires AOC, Filnemus, Bordeaux, France; ¹⁵Euro‐NMD, CHU de Nantes, Nantes, France; ¹⁶CHU Toulouse, Toulouse, France; ²⁰Service de Neurologie du Professeur Attarian, ERN Neuro‐NMD, La Timone, Marseille, France; ²¹Institut de Myologie, Hôpital La Pitié‐Salpétrière, FHU PHENIX, AP‐HP, Paris, France; ²³Centre Hospitalo‐Universitaire de Montpellier, Direction de la Recherche et de L'innovation, Montpellier, France; ²⁶Service médecine interne, pôle médecine forte consultation, CHR d’Orléans, Orléans, France; ²⁷Service de Neurologie, Hôpital F. Mitterrand, 4 bd Hauterive, 64046 Pau, France; ²⁸Service Neurologie, hopital de Saint Brieuc, Saint Brieuc, France; ²⁹Département de médecine interne et d'immunologie clinique, CHU de Nancy, Université de Lorraine, Vandoeuvre‐lès‐Nancy, France; ³⁰Centre Hospitalier Général de Montluçon, 03113 Montluçon, France; ³¹CHU de Bicêtre, Assistance Publique‐Hôpitaux de Paris, Université Paris‐Saclay, Le Kremlin‐Bicêtre, France.

Tard C., Sacconi S., Taouagh N., et al., “Real‐Life Effectiveness After Switching to Avalglucosidase Alfa in Late‐Onset Pompe Disease Patients Worsening on Alglucosidase Alfa Therapy: A French Cohort Study,” European Journal of Neurology 33, no. 7 (2026): e70680, 10.1111/ene.70680.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.