FormalPara Key Points

Tauro-urso-deoxycholic acid is being tested alone or in combination with sodium phenylbutyrate for the treatment of amyotrophic lateral sclerosis.

A comparative analysis of designs and methodologies shows similarities and differences between studies and nearly identical results on disease modification.

Tauro-urso-deoxycholic acid slows down the functional decline of patients with amyotrophic lateral sclerosis and the addition of sodium phenylbutyrate does not seem to influence this disease-modifying activity.

1 Introduction

Amyotrophic lateral sclerosis (ALS) is an aggressive and devastating neurodegenerative disease for which no effective treatment is available. Tauro-urso-deoxycholic acid (TUDCA) and urso-deoxycholic acid are highly hydrophilic bile acids synthesized in the liver, which can cross the blood–brain barrier. Preclinical studies suggested that TUDCA, the taurine conjugate of urso-deoxycholic acid, might exert a neuroprotective potential in several neurodegenerative disorders including ALS [1]. The hypothesized TUDCA mechanisms of action include the reduction of endoplasmic reticulum stress and of apoptosis, through inhibition of p53/Bax signaling, of cytochrome C release and of caspase activation [2]. Tauro-urso-deoxycholic acid has potential anti-neuroinflammatory activities, may act as a chemical chaperone and can increase neurite outgrowth in a murine model of ALS [3]. In keeping with preclinical evidence, phase II clinical trials have shown that TUDCA, alone or conjugated with sodium phenylbutyrate (NaPB), has a disease-modifying potential in ALS [4, 5]. This evidence has been reinforced by additional data on the survival of patients with ALS enrolled in an open-label follow-up study [6] and by post-hoc analyses of two phase II controlled trials [7, 8], raising further expectations on the therapeutic potential of TUDCA-containing drugs in ALS. Data obtained from the phase II study on TUDCA+NaPB have been considered of sufficient strength to approve this association for use in the USA and Canada [9, 10], while the same evidence was considered insufficient to grant approval in Europe [11].

These contrasting reactions raise increased attention on the two large phase III studies that are currently underway and expected to be completed in 2023 and early 2024. Their structure and designs are available for review. Both trials administer TUDCA to patients with ALS, as a solo compound in one case (NCT03800524) or combined with NaPB in another (NCT05021536). We therefore provide a comparative analysis of trials testing TUDCA and focus on study designs, eligibility criteria, as well as primary and secondary outcomes.

2 Methods of Literature Search

This article implemented a systematic methodology for the study search and selection [12]. A PubMed search was based on the following search strings: (((TUDCA) OR (Tauroursodeoxycholic Acid)) OR (taurursodiol)) AND (amyotrophic lateral sclerosis))). We subsequently reviewed the articles’ references to help identify studies not populated by the search strings. Furthermore, the ClinicalTrials.gov repository was searched using the terms TUDCA, tauroursodeoxycholic acid, or taurursodiol, combined with ALS or Amyotrophic Lateral Sclerosis. Listed trials were associated with their corresponding scientific publications and their contents were compared; current ongoing studies with no published results were identified. Furthermore, the main ALS databases and pharmaceutical sites were searched for the original datasets of the studies retrieved. The search collected data available until 18 October, 2023.

Studies were included if they were randomized, double-blind, placebo-controlled clinical trials assessing the efficacy of TUDCA-containing compounds in patients with ALS. Protocol descriptions, open-label extensions and post-hoc analyses of these trials were also included. Publications were excluded if they were reviews, letters, or preclinical. Trials were excluded if TUDCA-containing compounds were over-the-counter formulations, concomitant medications or prescribed for compassionate use. Open-label studies unrelated to an included randomized trial were also excluded. Information on the study design was collected for all the selected studies, including details on the number and location of centers, the medicinal products (doses, route of administration), trial duration, eligibility criteria, and outcomes. For trials already completed, published data were compared to information listed in the ClinicalTrials.gov repository and inconsistencies were annotated. Safety and efficacy results were collected from studies with published results. Data collected from published studies were comparatively analyzed. To make data comparable, the phase II TUDCA data were reassessed to calculate the monthly functional decline using an analysis of variance model with the baseline score as a covariate.

3 Comparative Analysis

Our initial search identified 45 articles in PubMed and 13 trials in ClinicalTrials.gov. After excluding duplicates, titles and abstracts were screened and nine publications were included (Fig. 1), which referred to three completed trials (two phase II randomized studies and one open-label extension), and one ongoing phase III randomized study. Of the studies retrieved from the trial repository, two published phase II studies, one published phase II open-label extension, and four ongoing unpublished trials (two phase III randomized trials and two open-label follow-up extensions) were included. These trials were conducted in North America and Europe. The criteria used for diagnosis and eligibility were comparable across studies, whereas the main differences regarded study duration, sample size, and concomitant treatments (Tables 1 and 2, Electronic Supplementary Material [ESM]). One set of studies tested the safety and efficacy of a proprietary formulation of TUDCA, another set of clinical trials tested a proprietary co-formulation of TUDCA and NaPB called AMX0035. No other combinations or co-formulations of TUDCA were detected.

Fig. 1
figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram outlining the identification, screening, evaluation, inclusion, and exclusion of articles. TUDCA tauro-urso-deoxycholic acid, NaPB sodium phenylbutyrate

Full size image
Table 1 Characteristics of phase II studies and of open-label follow-up testing TUDCA in ALS. Data are retrieved from published results
Full size table
Table 2 Characteristics of phase III studies and of open-label follow-up testing TUDCA in ALS. Data are retrieved from the ClinicalTrials.gov repository and from scientific publications
Full size table

3.1 Published Trials

3.1.1 TUDCA Trial

An independent phase II trial investigating the therapeutic potential of TUDCA in ALS was conducted in Italy (NCT00877604) [4]. This study tested the safety and efficacy of a branded TUDCA formulation (Bruschettini srl, Genova, Italy), in addition to a conventional therapy with riluzole. The study design and the main inclusion and exclusion criteria are listed in Table 1 and Table 1 of the ESM. Briefly, patients had clinically probable or definite ALS disease, as defined by the revised El Escorial diagnostic criteria, a phenotype with spinal onset and a disease duration < 18 months. Eligible participants were required to have forced vital capacity (FVC) ≥ 75% of normal and a stable riluzole regimen for at least 3 months. The included patients were observed for 12 weeks (run-in period) with riluzole only (50 mg twice daily). Subsequently, they were randomized 1:1 to receive TUDCA (1 g twice daily) or placebo for 54 weeks while continuing riluzole treatment. As reported in the published report [4], the primary outcome measure was the percentage of responding patients in the two treatment groups, defined as improving by at least 15% in the Amyotrophic Lateral Sclerosis Functional Rating Scale Revised (ALSFRS-R) at the end of the treatment period compared with the run-in period. Secondary outcomes included between-group comparisons of: (1) ALSFRS-R scores at study end; (2) ALSFRS-R mean linear regression slopes during the treatment period; (3) survival time; (4) FVC variation at study end; (5) physical component summary and mental component summary scores of SF-36; and (6) Medical Research Council scores for each side muscle groups. Additional secondary outcomes were listed in the ClinicalTrials.gov repository, but not in the publication: time to tracheostomy, and the incidence and severity of adverse events. By contrast, the ALSFRS-R mean linear regression slopes during the treatment period were listed in the publication but not in the online repository.

In this study, 17 patients were randomized to TUDCA and as many to placebo. Tauro-urso-deoxycholic acid was well tolerated, with no between-group differences for adverse events. Overall, two treatment-related adverse events were listed: mild diarrhea and anorexia. The trial fulfilled its primary objective. The responder rate was significantly higher in the TUDCA group than in the placebo group (87% vs 43%; p = 0.021). The calculated ALSFRS-R monthly functional decline was on average slower in the active group by – 1.13 points, compared with − 1.60 points in the placebo group (difference: 0.48 points/month; 95% confidence interval [CI] 0.13–0.83; p = 0.009; Table 3). At study end, the baseline-adjusted ALSFRS-R was significantly higher in the TUDCA group compared with placebo by a 7-point difference (mean value: 23.3 vs. 16.3; 95% CI 19.9–26.6 vs 12.9–19.7; p = 0.007). A regression slope analysis showed a milder progression in the TUDCA group than in the placebo group (p < 0.01). The remaining secondary outcome measures did not differ [4]. A post-hoc analysis of the entire dataset showed that receiving placebo treatment (p = 0.055) and having a lower ALSFRS-R total score at baseline (p = 0.128) were associated with an increased risk of death. The estimated hazard ratio (HR) for active treatment versus placebo was 0.051 (95% CI 0.002–1.059), indicating that participants in the active arm had a 95% lower risk of mortality compared with participants randomized to placebo [8] (Table 4). In a survival analysis, patients were followed for up to 66 weeks. The cumulative incidence of death during the study period was three deaths in the placebo group and one death in the TUDCA group. The results of the survival analysis were not statistically significant.

Table 3 Longitudinal pattern of ALSFRS-R functional decline (points per month) in phase II TUDCA studies
Full size table
Table 4 Overall survival time (months) in phase II TUDCA studies
Full size table

3.1.2 TUDCA+NaPB Trial

A pharma-sponsored phase II trial was conducted in North America (NCT03127514) [5] to test a proprietary co-formulation of TUDCA and NaPB (Amylyx Pharmaceuticals, Inc., Cambridge, MA, USA). The patients were free to take concomitant riluzole as well as edaravone treatments. The trial involved 137 participants with a diagnosis of definite ALS (revised El Escorial criteria), symptoms for ≤ 18 months, and slow vital capacity (SVC) >60% (Table 1 and Table 1 of the ESM). Participants were randomly assigned (2:1 ratio) to receive TUDCA (1 g) combined with NaPB (3 g) once or twice daily for 24 weeks (twice daily, if tolerated, from week 4) or alternatively placebo, in addition to riluzole, edaravone, or both. In total, 89 patients received the active treatment and 48 received placebo. At baseline, 68% of participants in the active group took riluzole and 25% took edaravone; in the placebo group, the percentages were instead 77% and 50%. In the published report, the primary outcome measure was the rate of decline in the total ALSFRS-R score over 24 weeks [5]. The ClinicalTrials.gov repository also listed additional primary outcome measures: the number of participants with adverse events and the number of subjects in each group able to remain on the study drug until planned discontinuation. Secondary outcomes were: (1) the rates of decline in isometric muscle strength (Accurate Test of Limb Isometric Strength [ATLIS] score changes); (2) the change in plasma phosphorylated axonal neurofilament H subunit; (3) the change in SVC; and (4) time to death, tracheostomy, permanent ventilation, or hospitalization.

The trial met its primary endpoint, as the ALSFRS-R score declined on average by −1.24 points per month in the active group and by −1.66 points per month in the placebo group (difference: 0.42 points/month; 95% CI 0.03–0.81; p = 0.03; Table 3). At study end, the between-group difference was 2.32 points. Secondary outcomes did not differ between groups. The most frequent adverse events were diarrhea, nausea, salivary hypersecretion, and abdominal discomfort. These gastrointestinal events were considered treatment related only during the first 3 weeks of active treatment. Nineteen percent of participants in the active group and 8% in the placebo group prematurely discontinued the trial regimen because of adverse events, among which the most common were diarrhea (6% in the TUDCA+NaPB group, none in the placebo group) and respiratory failure (6% in the placebo group, none in the TUDCA+NaPB group). Five deaths were registered in the active treatment group, two in the placebo group [5].

The open-label follow-up of this trial involved 90 patients who successfully completed the double-blind phase (56 originally randomized to TUDCA+NaPB; 34 originally randomized to placebo) and received TUDCA+NaPB for up to 30 months (NCT03488524) (Table 1 and Table 1 of the ESM). The open-label follow-up assessed the long-term safety and efficacy of TUDCA+NaPB. Primary outcomes were the quantity of adverse events and serious adverse events observed in the study; secondary outcomes were the number of hospitalizations, the rate of progression on ALSFRS-R, ATLIS strength measurement and SVC progression, the frequency of a gastric tube, and the occurrence of permanent invasive ventilation (ClinicalTrials.gov repository).

The double-blind and open-label studies were analyzed cumulatively [6, 7, 13]. Data on survival considered the cumulative post-randomization period, including an extended active treatment administration to patients originally randomized to TUDCA+NaPB during the double-blind period who received the active compound also during the follow-up, and a switch from placebo to active treatment for those patients who received placebo during the double-blind period and active compound at the follow-up. An intent-to-treat analysis of the combined datasets yielded a median overall survival of 25.8 months for participants originally randomized to TUDCA+NaPB and of 18.9 months for those originally randomized to placebo. This suggested a 43% lower risk of death in participants receiving the active compound for the entire treatment period (36 months; HR = 0.57; 95% CI 0.35–0.92; p = 0.023) [7]. A post-hoc analysis estimated an even lower median survival duration of the placebo group (15.2 months), resulting in a 10.6-month increased survival and a 61% lower risk of death for patients taking TUDCA+NaPB (HR = 0.39; 95% CI 0.17–0.88; p = 0.023) [7] [Table 4]. In keeping, the risk of any key event (all-cause death, tracheostomy, permanent assisted ventilation, or hospitalization) was 47% lower in patients originally randomized to TUDCA+NaPB compared with those who received placebo during the randomized phase (HR = 0.53; 95% CI 0.35–0.81; p = 0.003) [13]. A further analysis of cumulative survival, compared with an external control cohort, showed that the estimated median overall survival was 23.54 (14.56–39.32) months in the TUDCA+NaPB group and 13.15 (9.83–19.20) months in the external control group [14].

The following efficacy outcomes were calculated from randomization until week 24 of the open-label period (a total of 48 weeks): ALSFRS-R, upper-limb and lower-limb ATLIS scores, and SVC. Patients receiving active treatment for the total period had higher ALSFRS-total scores compared with those who received placebo during the first 24 weeks and active treatment during the last 24 weeks (4.23-point ALSFRS-R total score difference, 95% CI 0.56–7.9; p = 0.02). A similar functional benefit of early administration of TUDCA+NaPB was observed for ATLIS scores in the upper and lower limbs; between-group differences were 7.83 points in the upper limbs (95% CI 0.85–14.80; p = 0.03) and 4.74 points in the lower limbs (95% CI – 3.00 to 12.48; p = 0.23). The between-group difference for SVC was 10.66% (95% CI 0.63–20.69; p = 0.04) [15].

3.1.3 Phase II Trial Comparison

A comparison of these two independent studies showed remarkable differences. Both trials had a parallel-group double-blind design, but a run-in observation period was included only in the TUDCA solo trial, and an open-label phase was added only to the TUDCA+NaPB study. A further difference is that in the TUDCA solo trial, primary outcome variables were treated as categorical data by a predefined ALSFRS-R slope threshold to censor patients who responded to the experimental drug [4]. In the TUDCA+NaPB trial, instead, the absolute scores for all continuous efficacy outcomes were analyzed with the use of a random-slope, shared-baseline, linear mixed model adjusted for age and the pre-baseline ALSFRS-R slope [5]. Furthermore, relevant between-study differences concerned study duration and sample size. Trial duration was much longer in the TUDCA solo trial compared with TUDCA+NaPB (54 weeks vs 24 weeks). However, there were four times more patients enrolled in the TUDCA+NaPB study than in the TUDCA solo study. The randomization rate also differed, as the number of patients included in the phase II TUDCA+NaPB active arm doubled that of the placebo arm, whereas in the TUDCA solo trial, patient numbers were comparable between the arms. Finally, another between-study difference regarded the allowed concomitant medications. In the TUDCA solo trial, all patients had a standard additional treatment with riluzole, whereas in the TUDCA+NaPB trial, 71% of patients received riluzole, 34% received edaravone, and 28% received both.

A comparison of longitudinal patterns in functional decline, as measured by the ALSFRS-R and by survival time, showed no substantial between-study differences considering that the study designs had different effect sizes and durations (Tables 3 and 4, Fig. 2). The absolute difference in ALSFRS-R total scores between active and placebo arms at study end were: 7 points in the TUDCA trial (after 54 weeks), 2.32 points in the TUDCA+NaPB trial (after 24 weeks), and 4.23 points when combining double-blind and open-label observations of the TUDCA+NaPB trial (after 48 weeks; Fig. 2). Among the four subdomains of the ALSFRS-R scale, the bulbar score was prevalently influenced in the TUDCA trial, while the fine-motor subscale was prevalently influenced in the TUDCA+NaPB study. Individual patient data were unfortunately not available in public repositories and a direct statistical comparison could not be performed.

Fig. 2
figure 2

Plots of Amyotrophic Lateral Sclerosis Functional Rating Scale-Revised (ALSFRS-R) functional decline reported in phase II studies show remarkable similarities. Data on tauro-urso-deoxycholic acid (TUDCA) alone are shown in black whereas data on TUDCA + sodium phenylbutyrate (NaPB) are in red. Solid lines represent active treatment, dashed lines represent placebo. Data on TUDCA alone span 54 weeks of the double-blind phase II study. Data on TUDCA+NaPB span 48 weeks, of which 24 are double blind and the remaining 24 are open label (active compound). There is a remarkable parallelism of progression curves (either placebo or active arms) between the TUDCA and TUDCA+NaPB trials. Data are from published reports [4, 5, 15]. ALSFRS-R Amyotrophic Lateral Sclerosis Functional Rating Scale-Revised, NaPB sodium phenylbutyrate, TUDCA tauro-urso-deoxycholic acid

Full size image

3.2 Ongoing Trials

Two phase III randomized studies with open-label follow-up extensions are currently ongoing. An independent double-blind trial funded by the European framework Horizon 2020 program involves 25 centers in seven European countries and tests a proprietary TUDCA formulation (Bruschettini Srl) in combination with standard riluzole treatment (NCT03800524) [16, 17]. Patients undergo a 3-month run-in period with their standard treatment before being randomized to receive TUDCA (1 g twice daily) or placebo for 18 months. The study included patients with ALS as defined by the Revised El Escorial diagnostic criteria, having a disease duration ≤18 months, FVC or SVC ≥70% of normal, and a stable riluzole regimen for at least 3 months (Table 2 and Table 2 of the ESM). The primary outcome measure is the proportion of responding patients (defined as those showing an improvement of at least 20% in the ALSFRS-R slope during the treatment period compared with the run-in period). Secondary outcomes include: (1) survival time (as measured by death or respiratory insufficiency); (2) the rate of functional decline measured by ALSFRS-R, FVC, ALS Assessment Questionnaire-40, EuroQol 5-Dimension-5 Levels scale, and Medical Research Council scale; (3) long-term safety and tolerability of TUDCA for up to 18 months; and (4) measures of neurofilaments and of matrix metalloproteinase 9 in blood and cerebrospinal fluid. In total, 337 patients have been enrolled and the results are expected by the end of 2023. Upon completion of the double-blind phase, patients are offered to enter an open-label extension (NCT05753852) to receive the active treatment for an additional 18 months [18]. The extension study aims at collecting additional data on the long-term safety and tolerability of TUDCA, on survival, and on a change in disease progression and functional impairment as measured by ALSFRS-R (Table 2 and Table 2 of the ESM).

A pharma-sponsored, phase III, double-blind trial is ongoing in 69 centers in North America and Europe (NCT05021536) [19]. In this study, a proprietary co-formulation of TUDCA and NaPB (Amylyx Pharmaceuticals, Inc.) is administered for 48 weeks to patients with a diagnosis of definite or clinically probable ALS, disease onset < 24 months, and SVC > 55% of normal, prior to randomization. Patients are randomized 3:2 to receive either TUDCA (1 g) in combination with NaPB (3 g) or placebo, initially once daily for 3 weeks then twice daily (if tolerated) for 48 weeks. Riluzole, edaravone, or both are allowed concomitantly to TUDCA+NaPB (Table 2 and Table 2 of the ESM). Three primary outcome measures are listed: the rate of decline in the total ALSFRS-R score and survival, the number of participants with adverse events, the number of participants in each group able to remain on the study drug until planned discontinuation. Secondary outcomes are: (1) the rate of functional decline measured by SVC; (2) participant quality of life; (3) decline in King’s and MiToS stages; (4) ventilation-free survival; (5) EQ-5D descriptive system and the EQ visual analog scale; and (6) long-term survival. The estimated enrollment is 600 patients. In addition, patients completing the randomized phase can access an open-label extension (NCT05619783) to receive the active compound for an additional 108 weeks [20]. The aim of the open-label trial is to evaluate the safety and tolerability of TUDCA+NaPB by assessing the incidence of treatment-emergent adverse events and the impact of long-term treatment on survival (Table 2 and Table 2 of the ESM).

Phase III study designs are comparable to those of parent phase II studies, including the open-label follow-up periods. Compared to the phase II counterparts, few minor changes on eligibility criteria have been implemented in phase III studies. An endpoint on biomarker analysis has been included in the TUDCA trial, while it is not listed in the TUDCA+NaPB study. In the TUDCA solo trial, primary outcome variables will be treated as categorical data by a predefined ALSFRS-R slope threshold to censor patients who will respond to the experimental drugs. Compared to the phase II study, the threshold was raised from 15% to 20% [17]. Criteria for statistical analysis of the TUDCA trial have been recently published in a preprint article [21]. Criteria for the data analysis of the TUDCA+NaPB trial have not been yet declared, but the ClinicalTrials.gov repository lists for the phase III trial the same three endpoints reported in the phase II study. It can be expected that, similarly to the parent phase II study, the first of these will be chosen for a final analysis.

4 Discussion

Outcomes of two phase II studies suggested that TUDCA and TUDCA+NaPB might represent a valuable add-on therapeutic option for patients with ALS, and prompted the initiation of larger phase III studies. The present review reports the similarities and differences of these trials and highlights the opportunities and challenges derived from their comparative assessment.

The TUDCA and TUDCA+NaPB trials share similar eligibility criteria, endpoints, and outcome measures; but display remarkable differences in the overall design, duration, and sample size. Notwithstanding, both phase II studies showed that active treatments slowed the rate of ALS functional decline by comparable degrees. It has been observed that disease progression trajectories during the double-blind periods were remarkably parallel, either for the active or placebo arms [1]. This review prolongs the progression trajectories in TUDCA+NaPB studies to 48 weeks and confirms that parallelism between trials continues during the open-label observation period (Fig. 2). The TUDCA and TUDCA+NaPB data are both in keeping with an earlier report on urso-deoxycholic acid, which calculated a similar dimension in decline of the Appel ALS rating scale (by 1.17 points/month in participants who received an urso-deoxycholic acid dose of 3.5 g/day) [22]. It is quite notable that none of the secondary outcomes was achieved by the studies under comparison, suggesting that their designs and sample size may have been inadequate or that secondary outcomes were not influenced by the investigational medicinal products, at least for the duration of the double-blind periods.

The similarity of efficacy results reported by this comparative analysis strengthen the question of whether addition of NaPB to TUDCA is advantageous, neutral, or detrimental [1, 8]. This direct comparison suggests that the observed disease-modifying effect of TUDCA is not influenced by the addition of NaPB. An earlier trial reported that NaPB alone (from 9 to 21 g daily) was well tolerated in patients with ALS, with approximately 57% of them tolerating the highest dose [23]. In addition, this study also measured functional outcomes of disease progression, such as ALSFRS-R, vital capacity, and isometric strength testing. These measures were not different from the progression rates reported for the placebo cohort of a trial testing celecoxib in ALS [23], suggesting that NaPB alone does not influence disease progression. There are no planned studies on NaPB in ALS; therefore, further evidence of its efficacy may be obtained by comparing the results of the ongoing phase III studies, one of which is based on TUDCA solo, the other on TUDCA+NaPB.

This review highlights discrepancies in reporting outcomes between the ClinicalTrials.gov repository and the corresponding scientific publications. The most common occurrence refers to outcomes listed in the repository but missing in publications; occasionally the opposite was found. Remarkably, these discrepancies regarded primary as well as secondary outcomes. Primary outcomes are key variables for planning the statistical power of a study once it is designed [24]. It is unlikely that a study powered to test three primary outcomes is then analyzed for one only. It is more likely that the original power size calculations considered one outcome measure only, to be chosen from a list of different possibilities. Primary outcomes are key elements in the study design and usually remain unchanged. There have been previous reports on discrepancies between the ClinicalTrials.gov repository and a corresponding scientific publication [25], which usually refer to secondary outcomes that are subject to reconsideration during the course of the study. Guidance from European and American regulatory agencies on multiple endpoints in clinical trials highlights methodological issues concerning multiple comparisons, and the control of type I errors [26, 27]. An excessive number of outcomes and treatment effects may limit the possibility to compare studies and requires the collection of larger samples to attain significance of trial data.

This review indicates that some remarkable differences between trials concerned the design of both phase II and phase III trials. The TUDCA+NaPB trials performed an ordinal analysis, comparing the means of the two experimental treatments based on the hypothesis that the two experimental arms differed only for the given medication. The TUDCA solo trial, instead, considered that within each treatment group there may be heterogeneous populations of patients with ALS, some of whom may respond to treatment while others may not. The TUDCA solo trial, therefore, included a run-in phase to perform a categorical data analysis based on the classification of individual patient response related to a prespecified threshold. Differences in design among ALS trials have been the object of a dedicated review [28]. Including a run-in period may introduce some bias or delay the experimental treatment by some months, and may potentially result in lower recruitment rates and an increased dropout of patients with more rapid disease progression [28]. However, this design type has the advantage of considering the heterogeneity of patient response within each treatment arm.

An additional difference between phase III trials is the assessment of biomarkers. The TUDCA solo study included two biomarker measurements in the cerebrospinal fluid and blood: neurofilaments to monitor disease progression, and matrix-metalloproteinase-9 to monitor drug efficacy. The TUDCA+NaPB study, instead, is not planned to measure biomarkers.

Analysis of the open-label extension period is particularly delicate. While open-label extensions facilitate the assessment of long-term survival and allow a reduction in the length of the double-blind period, the switch of the placebo arm to active treatment increases the risk of type I errors and requires the incorporation of statistical methods to adjust survival estimates [29]. The open-label TUDCA+NaPB phase II extension study was subject to repeated post-hoc reassessments [7, 14] that were variably considered during the approval processes. An unplanned post-hoc reassessment of data incorporating both the double-blind and the open-label extension periods can be useful to generate scientific hypotheses or explore unanticipated gaps, but cannot be reliably used to strengthen the results of primary outcome measures [30]. Each post-hoc reassessment can add significant bias to the interpretation of data; differently from the prospective design, a post hoc analysis must not assume that if an event or intervention precedes another, it is necessarily associated with or has a causal relationship to the endpoint chosen [31].

5 Conclusions

The present review analyzes comparatively all data collected by the phase II, double-blind TUDCA studies, including the follow-up open-label extensions. The design of the phase III studies was also compared based on information stored in the ClinicalTrials.gov repository. The TUDCA+NaPB trials are shorter than TUDCA trials, 44% in phase II, and 62% in phase III. The short duration of the phase II TUDCA+NaPB study is compensated by a long open-label extension period, which was subject to repeated post-hoc analyses. While in the TUDCA solo study data on progression and survival are derived from the double-blind phase only [4, 8], for the TUDCA+NaPB study, data on progression derive from the double-blind study [5], whereas survival data derive from combined observations of the double-blind and open-label phases [6, 7]. These combined data were instrumental for obtaining approval in the USA [9] and Canada [10]. By contrast, marketing authorization was recently denied by the European Medicines Agency as it was considered that the evidence of efficacy was not compelling and combined survival data could not be relied upon [11]. It may be considered that the Accelerating Access to Critical Therapies for ALS Act approved in December 2021 facilitated approval in the USA [32].

Combining all the data reviewed, it emerges that TUDCA is very well tolerated and likely efficacious in ALS, in line with evidence recently collected by an independent real-world observational study [33]. All trials used medicinal products of pharmaceutic quality, supplied by pharma companies. There is no reason to expect differences in the TUDCA components used in these studies. It should be considered, however, that TUDCA can also be obtained as a generic or non-pharmaceutic grade product. Therefore, a caveat must be maintained on the generalizability of the observations reviewed here to non-pharmaceutical quality preparations. No evidence emerged that adding NaPB improves the efficacy of TUDCA. To draw a final conclusion, it would be important to investigate the efficacy of NaPB alone in patients with ALS. It would also be important to better ascertain the interactions between TUDCA and NaPB. In the wake of these studies, the ongoing phase III trials and their open-label follow-up extensions will provide additional information on the efficacy of these medicinal products in ALS.