Huoling Shengji granule in amyotrophic lateral sclerosis: A multicenter, randomized, double-blind, riluzole-controlled trial

Abstract

Currently, no curative therapies exist for Amyotrophic Lateral Sclerosis (ALS). This study aimed to evaluate the efficacy and safety of Huoling Shengji granules (HLSJ), a traditional Chinese medicine (TCM), compared to the standard treatment riluzole. A multicenter, randomized, double-blind, double-dummy, active-controlled Phase II clinical trial was conducted across 11 centers in China. A total of 140 ALS patients were randomly assigned (1:1) to receive either HLSJ or riluzole for 48 weeks. The primary endpoint, analyzed in the full analysis set (FAS), was the change in the ALS Functional Rating Scale-Revised (ALSFRS-R) score from baseline to Week 48. Safety profiles were comparable between groups, with no significant difference in adverse events (80.28% vs. 80.56%, P = 0.9671). Analysis using the pre-specified Last Observation Carried Forward (LOCF) method showed a numerical advantage for HLSJ (1.07 points) but lacked statistical superiority. However, a more scientific analysis using the Mixed Model for Repeated Measures (MMRM), recommended for progressive diseases, indicated that HLSJ was significantly superior to riluzole in slowing ALSFRS-R score decline (Least Squares Mean Difference [LSMD]: 2.29 points; 95% confidence interval [CI]: 0.52 to 4.06; P = 0.0114). While the LOCF model showed no statistical difference, the MMRM analysis confirmed that HLSJ demonstrated significant efficacy superior to riluzole with a favorable safety profile. These findings provide a critical basis for conducting pivotal Phase III confirmatory trials of HLSJ for ALS treatment.

Introduction

ALS is a fatal progressive neurodegenerative disease characterized by selective loss of motor neurons, leading to progressive muscle weakness, paralysis, and typically death from respiratory failure within 3–5 years of symptom onset [1]. The pathogenesis of ALS is complex and multifactorial, involving interplays between genetic susceptibility, glutamate excitotoxicity, oxidative stress, neuroinflammation, RNA metabolism dysregulation, and protein homeostasis disruption [[2], [3], [4], [5], [6], [7]]. While global incidence varies, approximately 90% of cases are sporadic, with a slight male predominance. In China, epidemiological data indicate that about 54.87% of ALS patients are aged 45–64 years—a demographic that constitutes the core of family and societal productivity—highlighting the profound personal, familial, and public health burden imposed by this disease [8].

To date, no cure exists for ALS. Since the approval of riluzole in 1995, only a few therapies have received regulatory authorization: edaravone (approved in 2017) has been shown to slow functional decline in a subset of patients, and the antisense oligonucleotide tofersen (approved in 2022) benefits individuals with SOD1 mutations. However, these treatments offer limited efficacy and apply to narrow patient populations [[9], [10], [11], [12]]. Most current therapies target single pathways, failing to address the marked heterogeneity and multi-pathway nature of ALS pathology. Thus, there remains a critical unmet need for novel therapeutic strategies that can modulate multiple biological targets, delay disease progression, and improve quality of life.

Traditional Chinese medicine (TCM) offers a holistic, multi-component approach well-suited to complex chronic diseases. HLSJ is a TCM formula composed of Epimedium folium, Astragalus membranaceus, Cornus officinalis, Atractylodes macrocephala, Rehmannia glutinosa, and Poria cocos. Its major bioactive constituents—including icariin, astragaloside IV, loganin, atractylenolide, catalpol, and verbascoside—have demonstrated anti-inflammatory, antioxidant, mitochondrial protective, and neuroprotective properties in preclinical studies [[13], [14], [15], [16], [17], [18]]. Notably, several of these compounds have shown potential to improve neurological function in animal models of Alzheimer's disease, Parkinson's disease, and ALS.

In the SOD1^G93A transgenic mouse model, HLSJ significantly extended survival and reduced motor neuron loss in the spinal cord anterior horn, suggesting robust neuroprotective effects [19]. Years of clinical observation further indicate that HLSJ alleviates symptoms such as dyspnea, spontaneous sweating, and lassitude of the waist and knees, improving overall well-being in ALS patients [20]. These translational insights—from bench to bedside—support the hypothesis that HLSJ may slow ALS progression through synergistic, multi-target mechanisms.

Despite the widespread use of TCM in ALS management, no herbal formula has ever been evaluated against standard therapy in a rigorously designed head-to-head RCT. To address this gap, we conducted a multicenter, randomized, double-blind, double-dummy, active-controlled trial across 11 medical centers in China to assess the efficacy and safety of HLSJ compared with riluzole in slowing functional decline in ALS, with the primary endpoint being change in the ALSFRS-R score at week 48. This study represents the first head-to-head RCT of a TCM formula versus standard care in a rare neurological disorder, providing pivotal evidence for the modernization and integration of traditional medicine into contemporary neurodegenerative disease therapeutics.

Results

Fingerprint chromatography study of HLSJ

HLSJ was produced and provided by Zhengda Qingchunbao Pharmaceutical Co., LTD. The established fingerprint analysis method was used for analysis, and the results showed that the main characteristic peaks of HLSJ were basically displayed (Fig. 1). Meanwhile, the similarity of multiple batches of HLSJ was above 95% (Table 1), indicating that the components of this botanical drug were stable and controllable.

Fig. 1.

HPLC fingerprint of three different batches of HLSJ.

Table 1.

Fingerprint similarity analysis of three different batches of HLSJ.

	2005001	2007001	2007002	Reference fingerprint
2005001	1	0.984	0.991	0.988
2007001	0.984	1	0.987	0.990
2007002	0.991	0.987	1	0.997
Reference fingerprint	0.988	0.990	0.997	1

Study participants

Between July 22, 2021, and May 26, 2022, a total of 144 patients with ALS were enrolled across 11 centers in China. Following randomization, one subject was excluded due to not meeting the inclusion criteria. Ultimately, 143 patients (71 in the HLSJ group and 72 in the Riluzole group) received at least one dose of the study medication and were included in the Safety Set (SS). The FAS was constructed based on the intention-to-treat (ITT) principle, formed after excluding subjects who clearly did not meet the study requirements. In the Riluzole group, three patients were excluded from the FAS due to trial discontinuation and lack of post-medication efficacy data: one withdrew informed consent, one experienced a serious adverse event, and one was lost to follow-up. Therefore, the FAS comprised a total of 140 patients (71 in the HLSJ group and 69 in the Riluzole group) (Fig. 2). Given that the trial was conducted during the COVID-19 pandemic, when China implemented stringent containment measures leading to a higher dropout rate, an additional 21 patients withdrew due to major protocol deviations: 10 in the HLSJ group (6 withdrew informed consent, 4 experienced adverse events) and 11 in the Riluzole group (8 withdrew informed consent, 1 was lost to follow-up, 1 experienced an adverse event, and 1 died). Ultimately, a total of 119 patients (61 in the HLSJ group and 58 in the Riluzole group) completed the entire trial process. Baseline demographic and clinical characteristics were well balanced between the two groups: the mean age was 54.8 years (SD 6.90) in the HLSJ group and 56.4 years (SD 6.29) in the riluzole group; baseline ALSFRS-R scores were 42.6 (SD 2.43) and 42.2 (SD 2.45), respectively (Table 2).

Fig. 2.

Trial profile. Abbreviations: FVC=Forced Vital Capacity.

Table 2.

Baseline demographic and clinical characteristics of the study population (full analysis set).

	Riluzole (n = 69)	Huoling shengji (n = 71)	Totala (n = 140)
Age, years b	56.4 (6.29)	54.8 (6.90)	55.6 (6.63)
45–65 y	55.6 (5.63)	53.4 (5.61)	54.5 (5.70)
66–70 y	67.8 (0.84)	68.0 (1.15)	67.9 (1.00)
Sex
Male	40 (57.97)	45 (63.38)	85 (60.71)
Female	29 (42.03)	26 (36.62)	55 (39.29)
Body mass index, kg/m2	23.32 (2.692)	23.01 (3.063)	23.16 (2.880)
Disease type
Non-bulbar	63 (91.30)	60 (84.51)	123 (87.86)
Bulbar	6 (8.70)	11 (15.49)	17 (12.14)
ALS diagnosis of updated WFN criteria c
Definite	18 (26.09)	24 (33.80)	42 (30.00)
Probable	24 (34.78)	27 (38.03)	51 (36.43)
Probable laboratory-supported	27 (39.13)	20 (28.17)	47 (33.57)
History of ALS treatment
Negative	22 (31.88)	13 (18.31)	35 (25.00)
Positive	47 (68.12)	58 (81.69)	105 (75.00)
Time of diagnosis, mod	18.2 (8.63)	18.2 (8.98)	18.2 (8.78)
Rate of disease progression, points/moe	0.417 (0.324)	0.401 (0.384)	0.409 (0.355)
ALSFRS-R score at baseline	42.2 (2.45)	42.6 (2.43)	42.4 (2.44)
ROADS score at baseline	92.9 (9.82)	94.5 (14.16)	93.7 (12.20)
ALSAQ-40 score at baseline	20.4 (12.60)	18.8 (11.31)	19.6 (11.95)
FVC at baseline, %	96.938 (16.238)	98.121 (14.524)	97.538 (15.349)
TCM syndrome scale scoref	12.6 (3.30)	12.7 (3.58)	12.7 (3.43)

Abbreviations: ALS, amyotrophic lateral sclerosis; ALSFRS-R, amyotrophic lateral sclerosis functional rating scale-revised; ROADS, rasch-built overall amyotrophic lateral sclerosis disability scale; ALSAQ-40, amyotrophic lateral sclerosis assessment questionnaire-40; FVC, forced vital capacity; TCM, traditional Chinese Medicine; WFN, World federation of neurology.

^aData are n (%), mean (SD), unless stated otherwise. A total column has been added to facilitate the description of the overall study population.

^bThe definition of “age 65″ as the boundary between adulthood and older age is based largely on the 1956 United Nations report on Population Aging and its Socio-economic Implications.

^cALS diagnosis includes four categories: Definite, Probable, Probable laboratory-supported, and Possible, as outlined in the El Escorial revisited: revised criteria for the diagnosis of amyotrophic lateral sclerosis published by the WFN study group in 2000.

^dTime of diagnosis is calculated as the date of randomization minus the date of first symptom onset.

^eRate of disease progression is calculated by subtracting the baseline ALSFRS-R total score from 48 points and dividing by the duration of disease (mo).

^fTCM syndrome scale score is formulated according to the TCM Diagnosis and Treatment Plan of Wei disease (Motor neuron disease) issued by the Chinese Association of TCM in 2018.

Prespecified efficacy outcomes

In the primary efficacy analysis, the least squares mean (LSM) change in ALSFRS-R score from baseline to week 48 was −10.46 (95% CI: −12.10 to −8.82) in the HLSJ group and −11.53 (95% CI: −13.19 to −9.87) in the riluzole group (Fig. S1 in Supplementary File S2). The between-group difference was 1.07 in favor of HLSJ (95% CI: −1.27 to 3.40; P = 0.3674), which did not meet the prespecified threshold for statistical superiority (Table S1 in Supplementary File S2). In a post hoc exploratory subgroup analysis of participants aged 45–65 years (n = 128), HLSJ showed a trend toward greater benefit compared with riluzole, with a LSMD of 2.20 (95% CI: −0.06 to 4.46; P = 0.0564). However, this result also did not reach statistical significance for superiority. No secondary efficacy endpoints achieved statistical significance for superiority (Table S1 in Supplementary File S2).

Post hoc efficacy analysis

In a MMRM analysis of the FAS, HLSJ was associated with a significantly slower decline in ALSFRS-R scores compared to riluzole, with a LSMD of 2.29 (95% CI: 0.52 to 4.06; P = 0.0114) (Fig. 3 and Table 3), indicating a statistically significant treatment effect over time. For secondary efficacy endpoints, HLSJ demonstrated statistically significant advantages over riluzole at multiple timepoints: ALSFRS-R at week 36: difference of 1.93 (95% CI: 0.30 to 3.56; P = 0.0208); ROADS scale at week 24: difference of 3.99 (95% CI: 0.27 to 7.70; P = 0.0356); ROADS at week 36: difference of 4.39 (95% CI: 0.55 to 8.23; P = 0.0252); ROADS at week 48: difference of 4.80 (95% CI: 0.66 to 8.94; P = 0.0233) (Table 3). The treatment effect was further strengthened in the modified intention-to-treat population (mITT) (n = 128), which included participants aged ≤65 years. After confirming baseline balance between groups (Table S2 in Supplementary File S2), the LSMD in ALSFRS-R decline favored HLSJ by 2.88 (95% CI: 1.02 to 4.73; P = 0.0025) (Table S3 and Fig. S2 in Supplementary File S2). Sensitivity analyses consistently supported the robustness of this treatment benefit (Table S4 in Supplementary File S2).

Fig. 3.

Subgroup analysis for the primary end point in post hoc analysis (full analysis set). Abbreviations: ALS, Amyotrophic Lateral Sclerosis; ALSFRS-R, Amyotrophic Lateral Sclerosis Functional Rating Scale-Revised; FVC, Forced vital capacity; MD, Mean difference.

Table 3.

Primary and secondary efficacy outcomes in post hoc analysis (full analysis set).

	Mean (SD)		Difference (95% CI)	P value
	Riluzole (n = 69)	Huoling shengji (n = 71)
Primary outcome
Change in ALSFRS-R score from baseline to week 48	−11.8 (7.61)	−10.5 (6.53)	2.29 (0.52–4.06)	0.0114
Secondary outcomes
Change in ALSFRS-R score from baseline to week 12, 24, and 36
12	−3.3 (2.84)	−3.1 (3.12)	1.21 (−0.40 to 2.82)	0.1390
24	−6.8 (5.44)	−5.8 (4.09)	1.57 (−0.01 to 3.15)	0.0509
36	−9.6 (6.60)	−8.0 (5.02)	1.93 (0.30–3.56)	0.0208
Change in ROADS score from baseline to week 12, 24, 36, and 48
12	−8.0 (8.08)	−7.4 (12.50)	3.58 (−0.20 to 7.37)	0.0633
24	−14.6 (10.14)	−11.9 (13.45)	3.99 (0.27–7.70)	0.0356
36	−19.8 (13.36)	−16.8 (14.32)	4.39 (0.55–8.23)	0.0252
48	−22.9 (14.62)	−20.9 (15.93)	4.80 (0.66–8.94)	0.0233
Change in FVC% from baseline to week 12, 24, 36, and 48
12	−5.298 (10.518)	−6.254 (10.418)	1.36 (−4.60 to 7.33)	0.6524
24	−9.284 (13.884)	−11.688 (14.354)	1.17 (−4.79 to 7.12)	0.6989
36	−13.856 (16.105)	−13.119 (12.276)	0.97 (−5.19 to 7.14)	0.7558
48	−12.892 (15.189)	−15.473 (14.472)	0.78 (−5.80 to 7.35)	0.8158
Change in ALSAQ-40 score from baseline to week 12, 24, 36, and 48
12	9.1 (10.76)	10.0 (12.80)	0.31 (−2.47 to 3.10)	0.8255
24	15.6 (12.98)	14.9 (13.80)	−0.83 (−4.03 to 2.38)	0.6116
36	23.7 (18.31)	20.7 (17.05)	−2.64 (−6.32 to 1.04)	0.1589
48	28.6 (18.48)	27.4 (19.70)	−2.15 (−6.16 to 1.86)	0.2931
Change in TCM syndrome scale score from baseline to week 12, 24, 36, and 48
12	1.6 (3.73)	0.9 (4.14)	−0.46 (−1.72 to 0.80)	0.4717
24	1.9 (4.62)	1.9 (4.76)	−0.62 (−1.86 to 0.61)	0.3205
36	3.2 (5.10)	2.7 (4.35)	−0.78 (−2.07 to 0.50)	0.2306
48	4.4 (5.63)	3.2 (4.59)	−0.95 (−2.36 to 0.47)	0.1878
Rate of post-treatment end point events	3 (4.35)	4 (5.63)	NA	>0.9999

Abbreviations: ALSFRS-R, amyotrophic lateral sclerosis functional rating scale-revised; ROADS, rasch-built overall amyotrophic lateral sclerosis disability scale; ALSAQ-40, amyotrophic lateral sclerosis assessment questionnaire-40; FVC, forced vital capacity; TCM, traditional Chinese medicine.

Given prior evidence suggesting age as a potential modifier of treatment response—and because the majority of FAS participants were aged ≤65 years—we conducted an exploratory subgroup analysis in this population. In mITT participants aged ≤65 years, HLSJ consistently showed a favorable trend over riluzole across multiple subgroups, including those with disease duration >1 year, forced vital capacity (FVC) ≥80%, baseline ALSFRS-R score of 41–47, definite or laboratory-supported probable ALS diagnosis, non-bulbar onset, prior use (or non-use) of riluzole or edaravone, female sex, and faster disease progression rate (Fig. S2 in Supplementary File S2).

Safety outcomes

Both HLSJ and riluzole were well tolerated by all participants. A total of 457 treatment-emergent adverse events (TEAEs) were reported in 115 participants (80.42%) during the study period (Table 4). Of these, 232 TEAEs were reported in 57 participants (80.28%) in the HLSJ group, and 225 TEAEs in 58 participants (80.56%) in the riluzole group. There was no statistically significant difference in TEAE incidence between the two groups (P = 0.9671). Thirty-seven serious adverse events (SAEs) occurred in 27 participants (18.88%): 15 SAEs in 12 participants (16.90%) in the HLSJ group and 22 SAEs in 15 participants (20.83%) in the riluzole group (Table 4). No significant differences were observed between groups in the number, frequency, severity, or drug-relatedness of TEAEs or SAEs. Six deaths occurred during the study. Five were attributed to disease progression or complications of ALS. One death was due to coronavirus disease 2019 (COVID-19) and was determined to be unrelated to the investigational product. The HLSJ group reported no grade ≥3 TEAEs, no SAEs leading to study drug discontinuation, and no drug-related deaths. Laboratory assessments revealed no hematological abnormalities in the HLSJ group, and no confirmed cases of moderate or severe hepatic or renal dysfunction. Transient, mild fluctuations in liver or kidney function tests were observed in a few participants but were not associated with clinical symptoms and did not require treatment interruption. One participant with a history of coronary artery atherosclerosis developed an isolated T-wave abnormality on electrocardiogram (incidence: 1.41%) during HLSJ treatment, which was classified as a cardiac-related AE. The event did not lead to study drug discontinuation and resolved spontaneously during follow-up. In summary, HLSJ demonstrated a favorable safety profile. The TEAE spectrum was comparable to that of the riluzole group, and no new safety signals were identified. Detailed safety data are presented in Table 4.

Table 4.

Safety analysis in huoling shengji and riluzole groups during the randomized controlled period (safety analysis set).

Variablea	Riluzole (n = 72)		Huoling Shengji (n = 71)		P value
	Total reported events, No.	Participants with adverse event, No. (%)b	Total reported events, No.	Participants with adverse event, No. (%)b
Total TEAE	225	58 (80.56)	232	57 (80.28)	0.9671
Severity level c
Level 1	111	14 (19.44)	131	12 (16.90)	NA
Level 2	92	29 (40.28)	86	33 (46.48)	NA
Level 3	19	12 (16.67)	11	8 (11.27)	NA
Level 4	0	0 (0.00)	1	1 (1.41)	NA
Level 5	3	3 (4.17)	3	3 (4.23)	NA
Correlation with trial drugs
Definite	0	0 (0.00)	1	1 (1.41)	NA
Probable related	3	1 (1.39)	1	1 (1.41)	NA
Possible related	54	27 (37.50)	59	25 (35.21)	NA
Possible unrelated	117	21 (29.17)	130	24 (33.80)	NA
Irrelevant	51	9 (12.50)	41	6 (8.45)	NA
Serious adverse event	22	15 (20.83)	15	12 (16.90)	0.5480
Significant adverse event	115	45 (62.50)	111	49 (69.01)	0.5948
TEAE leading to premature withdrawal	5	3 (4.17)	4	4 (5.63)	0.9849
TEAE leading to death	3	3 (4.17)	3	3 (4.23)	1.0000
Total adverse reactions	57	28 (38.89)	61	27 (38.03)	0.9158
Serious adverse reactions	2	1 (1.39)	0	0 (0.00)	1.0000
Significant adverse reactions	25	16 (22.22)	25	13 (18.31)	0.6782

Abbreviations: TEAE, treatment-emergent adverse event.

^aIf multiple identical adverse events occurred in the same subject, the severity level or related category of the adverse event in that subject was counted only once, but the number of events was counted as the actual number of occurrences. If these same adverse events differed in severity or causality, counted once in the most serious or most relevant category.

^bCalculated based on the number of subjects in each group.

^cRecorded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0 and graded from 1 to 5.

Discussion

ALS is a highly heterogeneous progressive neurodegenerative disorder, with only a limited number of disease-modifying therapies proven to slow disease progression or extend survival. Riluzole, the first approved treatment for ALS, has demonstrated efficacy through multiple RCTs. This study represents the first multicenter, double-blind, randomized, active-comparator, head-to-head trial evaluating the novel TCM formula HLSJ against riluzole in patients with ALS. In the FAS, HLSJ did not meet the prespecified threshold for statistical superiority in slowing the decline of the ALSFRS-R score, but suggests that its efficacy may be comparable to riluzole, with a favorable safety profile. These findings establish the initial clinical feasibility of HLSJ as a potential therapeutic option for ALS.

A key finding emerged from post hoc exploratory subgroup analyses: in patients aged ≤65 years, HLSJ showed a trend toward superiority that approached, but did not reach, statistical significance. When further restricted to participants who received treatment for at least 12 weeks (to account for drug exposure), the difference in ALSFRS-R change favored HLSJ with statistical significance (mean difference: 2.61; 95% CI: 0.32–4.89; P = 0.0259; n = 123). This effect size is not only clinically meaningful—exceeding the previously reported benefit of edaravone versus placebo (2.49 points)—but also biologically plausible. Multi-component herbal medicines often exert cumulative effects, and their pharmacokinetic and pharmacodynamic profiles suggest that sustained exposure may be necessary for optimal therapeutic activity. This aligns with findings from Chen et al., which showed that long-term riluzole use (cumulative defined daily dose ≥16,800 mg) significantly improves prognosis [21].

This age-dependent treatment response is consistent with the epidemiology and biology of ALS. Population data from urban China indicate that approximately 77% of ALS patients are aged ≤65 years, with peak incidence between 45 and 65 years [8]. Importantly, recent studies have shown that neutrophil-mediated neuroinflammation is more pronounced in younger ALS patients and closely associated with lower motor neuron axonal dysfunction [22,23]. With advancing age, neutrophil phagocytic and degranulation capacity declines [24,25], potentially attenuating inflammation-driven pathology and reducing responsiveness to immunomodulatory interventions. Thus, age may not only be a prognostic factor but also a key effect modifier in treatment response.

Further support for this hypothesis comes from an analysis of “poor responders"—defined as individuals with a >50% decline in ALSFRS-R score from baseline, indicating rapid disease progression. In the FAS, significantly fewer poor responders were observed in the HLSJ group compared to the riluzole group (5 vs. 14 patients). Stratification by median age (57.5 years) revealed a striking pattern: among patients aged 45–57.5 years, no individual in the HLSJ group was a poor responder, whereas 8 were identified in the riluzole group; in contrast, differences were minimal in the 57.5–70-year age group. These results suggest that HLSJ may confer stronger neuroprotective effects during earlier stages of ALS, particularly in middle-aged and younger adults, potentially reducing the risk of rapid progression. This has important clinical implications, as evidence suggests that riluzole's benefits are more pronounced in later-stage disease, with limited efficacy in early-phase ALS [26]. The advantage of HLSJ in early or rapidly progressing phenotypes implies a mechanism of action that may complement or differ from current standard therapies, offering a potential solution to a key unmet need and highlighting its promise as a targeted intervention for high-risk subgroups.

Although no significant difference in composite endpoint events was observed between groups, indirect modeling based on external data estimates that HLSJ reduces the rate of ALSFRS-R decline by 11.7%, corresponding to a potential extension in median survival of 2.8–3.5 months [21,27,28]. While this inference requires validation through long-term follow-up, it provides a compelling rationale for future studies.

This study has several limitations. First, although the primary analysis of this study adopted the protocol-pre-specified LOCF method, we fully recognize its limitations in progressive neurodegenerative diseases (such as potential bias in estimating efficacy). To provide a more robust assessment of efficacy, we concurrently conducted a MMRM analysis. The discrepancy between the results of the two methods highlights the importance of selecting appropriate statistical methods in ALS studies. Second, the absence of a long-term extension phase limits direct assessment of survival benefit and long-term safety. Third, the study population was restricted to Chinese participants, and generalizability to other racial and geographic populations remains to be established. Finally, the study did not systematically collect or analyze key biomarkers such as neurofilament light chain (NfL), which limits the in-depth exploration of the drug's biological effects and the objective quantification of treatment efficacy.

In conclusion, this head-to-head Phase II randomized controlled trial provides the first robust clinical evidence supporting the potential benefit of the traditional Chinese medicine formula HLSJ in the treatment of ALS. Both the subgroup analysis of patients aged ≤65 years who received at least 12 weeks of treatment and the MMRM analysis of the FAS consistently suggest that HLSJ may be superior to riluzole in slowing disease progression. Given the high heterogeneity of ALS, precise clinical trial design is critical. Therefore, we recommend that the subsequent Phase III confirmatory trial adopt an enrichment strategy, focusing on younger and middle-aged patients (≤65 years) in the early-to-mid stages of ALS, use MMRM as the primary statistical method, and incorporate biomarker assessments—such as NfL—to more accurately evaluate the clinical value of HLSJ.

Material and methods

Fingerprint analysis of HLSJ

According to the established HPLC fingerprint analysis methodology [19], the similarity of each batch of HLSJ used in this clinical trial was compared to confirm the quality consistency of this botanical drug between batches. The HLSJ batch numbers used for comparison were 2005001, 2007001 and 2007002, respectively.

Study design and oversight

This study was a multicenter, randomized, double-blind, double-dummy, active-controlled phase II clinical trial conducted at 11 centers across China, designed to evaluate the efficacy and safety of HLSJ in slowing disease progression in ALS. The total study duration was approximately 49 weeks, comprising a baseline period of up to 1 week and a 48-week treatment period. The Clinical Study Protocol (version 2.0) and Statistical Analysis Plan (SAP, version 1.0) were approved by the institutional ethics committees at all participating centers, prospectively registered on a public clinical trials platform (ChiCTR2100044085) in July 2021. The study design has been previously published [29]. This study was approved by the Ethics Committee of Peking University Third Hospital (Approval No.: (2020) Drug Ethics Review No. 095-02). All study procedures were conducted in strict accordance with the ethical principles of the Declaration of Helsinki and relevant national and institutional ethical guidelines. The ethics committees and approval numbers for each participating center are as follows: Peking University Third Hospital (Ethics No: Clinical Trial [Medicine], number 095-05, 2020), the Second Hospital of Hebei Medical University (Ethics No: 2021EC07-01-1), Henan Provincial People's Hospital (Ethics No: 2020-130-02), the First Affiliated Hospital of Sun Yat-sen University (Ethics No: 2021-025-01), Guangdong Provincial Hospital of Traditional Chinese Medicine (Ethics No: AF2020-287-01), the Second Affiliated Hospital of Zhejiang University School of Medicine (Ethics No: Clinical Trial [Medicine], number 866, 2020), West China Hospital of Sichuan University (Ethics No: Clinical Trial [Chinese Medicine], number 1, 2021), the First Affiliated Hospital of Xi ‘an Jiaotong University (Ethics No: XJTU1AF2021LSY-31), Fujian Medical University Union Hospital (Ethics No: 2021YW009-02), the First Affiliated Hospital of Harbin Medical University (Ethics No: Clinical Trial [Medicine], number 202115), Renmin Hospital of Wuhan University (Ethics No: WDRY2022-Y028). The relevant data were monitored by an Independent Data Monitoring Committee.

Participants

Inclusion criteria were as follows: age 45–70 years; diagnosis of ALS classified as “definite,” “probable,” or “probable–laboratory-supported” according to the World Federation of Neurology El Escorial criteria; ALSFRS-R score ≥2 on all individual items; normal respiratory function, defined as FVC ≥70% of predicted; symptom duration of ≤3 years from onset to enrollment; TCM pattern diagnosis of qi deficiency with kidney yang deficiency; and provision of voluntary, written informed consent with commitment to complete all study visits and follow-up. Specifically, the corresponding syndrome differentiation were determined by TCM physicians according to the TCM Diagnosis and Treatment Regimen of Motor Neuron Disease issued by the China Association of Chinese Medicine in 2018 (https://www.cacm.org.cn/2018/11/30/2946/), and the methods of tongue- and pulse-diagnosis of TCM. Exclusion criteria were reported previously [29]. All participants were fully informed by the investigator about the study objectives, procedures, and potential risks prior to randomization. Written informed consent was obtained from each participant or their legally authorized representative, with concurrent signature by the investigator to confirm the process.

Randomization, masking, and procedures

Participants were randomized in a 1:1 ratio to receive either HLSJ or riluzole using a blocked randomization method. An independent statistician generated the randomization sequence with concealed block sizes using SAS 9.4 software. The allocation schedule was implemented via an interactive web-based response system (IWRS) to ensure dynamic and secure assignment. The study employed a double-blind, double-dummy design to maintain blinding integrity: the HLSJ group received active HLSJ granules (20 g twice daily) plus placebo tablets matching the appearance of riluzole (50 mg); the riluzole group received active riluzole tablets (50 mg twice daily) plus placebo granules matching the appearance of HLSJ (20 g). All study medications were identically packaged with uniform labeling and coded identifiers. The details of the medicinal product are provided in Supplementary File S1. A total of four scheduled visits were conducted during the treatment period, occurring at weeks 12, 24, 36, and 48.

Outcomes

The primary efficacy endpoint was the change in the ALSFRS-R score from baseline to week 48. Secondary efficacy endpoints included: changes in ALSFRS-R scores at weeks 12, 24, and 36; changes at weeks 12, 24, 36, and 48 in the Revised Urinary and Bowel Function Scale (RODS), the ALS Assessment Questionnaire–40 (ALSAQ-40) for health-related quality of life, TCM syndrome scores, and percent predicted forced vital capacity (FVC%); and time to first occurrence of a composite endpoint event—defined as death, tracheostomy, or permanent ventilator dependence (>22 h/day for ≥14 consecutive days). Safety assessments included all adverse events (AEs), SAEs, laboratory abnormalities, and changes in vital signs.

Sample size

Based on previously reported data [21], we assumed a mean difference of 2.0 points in the change in ALSFRS-R score at week 48 between the HLSJ and riluzole groups (μT − μC = 2.0), with a pooled standard deviation of 4.0 points. With a two-sided α level of 0.05 and a target statistical power (1 − β) of at least 80%, a sample size of 64 evaluable participants per group was required. Accounting for an anticipated dropout rate of approximately 10%, we planned to enroll 144 participants (72 per group) to ensure sufficient statistical power to detect the prespecified clinically meaningful difference.

Statistical analysis

Analysis populations

FAS: Defined as all randomized participants who received at least one dose of study medication and had at least one post-baseline efficacy assessment. The FAS was analyzed according to the ITT principle and served as the primary efficacy population. mITT: Prespecified for exploratory analyses, this population included the subset of participants in the FAS who were aged ≤65 years. SS: Included all participants who received at least one dose of study medication and had any safety data recorded.

Primary outcome analysis

The change in ALSFRS-R score from baseline to week 48 was compared between groups using analysis of covariance (ANCOVA), with baseline ALSFRS-R score as a covariate and treatment group as a fixed effect. HLSJ was considered superior to riluzole if the lower limit of the two-sided 95% CI for the LSMD exceeded zero and the corresponding P value was <0.049.

Secondary outcome analyses

Continuous variables were analyzed using the same ANCOVA model as the primary outcome. Ordinal categorical variables were assessed using the Wilcoxon rank-sum test. Categorical variables were analyzed using the Cochran–Mantel–Haenszel test stratified by study center, with risk difference (RD) and its 95% CI reported. Time-to-event data for the composite endpoint (death, tracheostomy, or permanent ventilator dependence) were estimated using the product-limit method (Kaplan–Meier), with median time and 95% CI reported.

Handling of missing data

Missing values for the primary outcome were imputed using the LOCF method. No imputation was performed for secondary efficacy or safety variables.

Post hoc efficacy analysis

To provide a more robust assessment of treatment effects over time, a MMRM was applied in a post hoc analysis. The model included fixed effects for time, treatment group, and the time-by-treatment interaction, with adjustment for age (≤65 vs > 65 years) and ALS diagnostic certainty as stratification factors. For the mITT analysis, the age covariate was excluded. All observed data across time points were included, and no additional imputation for missing data was performed.

Sensitivity analysis

A sensitivity analysis was conducted by applying LOCF imputation to the primary outcome and re-running the MMRM model to evaluate the robustness of results to missing data assumptions.

Statistical software and significance level

All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc.). All hypothesis tests were two-sided, and a P value of <0.049 was considered statistically significant.

Author contributions

XLL, LS, JY, YJZ, and DSF contributed to the concept or study design. XLL, XYC, LS, JY, SCZ, LZH, PY, DSF were involved in analyzing or interpreting the data. XLL, WCF, LS, JY, YJZ, and SCZ contributed to the drafting of manuscript. XLL, WCF, LS, YF, BW, XX, XLY, YLL, MMM, BXD, JXD, ZYZ, HLF, YZ, ZNL, YZ, ZYW, HFS, JSZ, PY, and DSF contributed to critical review for important content. XYC, SCZ, and PY contributed to the statistical analysis. LS, XX, BW, and YF contributed to financial support. JQZ contributed to the supply of the trial drugs. LS, XX, BW, YF, and DSF contributed to supervision. All authors had full access to all study data and were responsible for the decision to submit for publication.

Ethical approval and consent to participate

This study was approved by the ethics committees of 11 hospitals, led by Peking University Third Hospital, and all participants gave written informed consent.

Consent for publication

Not applicable.

Data availability

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Role of the funding source

The funder of the study was involved in the study design, data gathering, data analyzing, data interpretation, report and manuscript writing, and the decision to submit for publication. All of the data analyses and interpretation were confirmed by Wuhan Zhizhi Medical Technology Co. LTD.

Declaration of generative AI and AI-assisted technologies in the manuscript preparation process

During the preparation of this work, the authors used [Qwen3-Max] in order to [assist with language editing]. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

Funding

The study was funded by Shanghai Pharma Rare Disease Medicine Co., LTD.

Declaration of competing interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Appendix A. Supplementary data

The following are the Supplementary data to this article.