Amyotrophic lateral sclerosis (ALS) remains one of the most challenging neurodegenerative diseases to treat, with over 30 years of clinical trial failures. Despite remarkable advances in understanding the molecular mechanisms of ALS pathogenesis—including [C9orf72](/genes/c9orf72) hexanucleotide repeat expansions, [SOD1](/proteins/sod1-protein) mutations, [TDP-43](/proteins/tdp43-protein) pathology, and [FUS](/proteins/fus-protein) abnormalities—the translation of these insights into effective therapies has proven extraordinarily difficult. This analysis examines the major therapeutic approaches that have failed in ALS clinical trials, explores why they failed, and extracts lessons that can inform future drug development efforts.
The modern era of ALS clinical trials began in the 1990s following the identification of [SOD1](/genes/sod1) mutations as a cause of familial ALS in 1993. This breakthrough ignited optimism that understanding the genetic basis would rapidly lead to targeted therapies. However, the translation from genetic discovery to effective treatment has proven far more complex than initially anticipated. [@baloh2021]
Between 1993 and 2024, over 200 clinical trials have been conducted in ALS, with only two drugs receiving regulatory approval: riluzole (1995) and edaravone (2017). The failure rate exceeds 99%, making ALS one of the most drug-resistant conditions in neurology. Understanding why so many approaches have failed is essential for improving future trial success. [@ling2013]
[Edaravone](/drugs/edaravone) (Radicut/Radicava) represents the sole approved neuroprotective agent since riluzole, having received approval in Japan (2015), US (2017), and EU (2019). However, its clinical benefit is modest—the pivotal trials demonstrated only a 2.5-point reduction in ALSFRS-R decline over 24 weeks in a restricted patient population. Subsequent real-world studies have yielded conflicting results, with many clinicians questioning whether the marginal benefit justifies the intensive intravenous infusion schedule. [@saberi2015]
The limited efficacy of edaravone likely reflects multiple factors: (1) treatment was initiated relatively late in disease course when substantial neurodegeneration had already occurred; (2) the patient selection criteria in the pivotal trials excluded many typical ALS patients; (3) oxidative stress, while important in ALS pathogenesis, may be downstream of more primary disease drivers. [@brown2022]
Numerous other antioxidant strategies have failed in ALS trials: [@loeravalencia2019]
The consistent failure of antioxidant approaches suggests that: (1) oxidative stress may be a secondary phenomenon rather than a primary driver; (2) single-target antioxidants are insufficient given the complex redox dysfunction in ALS; (3) early intervention may be critical—treatment when oxidative damage is already extensive may be too late; (4) biomarkers to select patients with active oxidative stress could improve trial outcomes. [@pasinetti2013]
[Riluzole](/drugs/riluzole) remains the only FDA-approved disease-modifying therapy for ALS, having received approval in 1995 based on a modest 2-3 month survival benefit. The drug's mechanism involves multiple actions: inhibition of glutamate release, blockade of NMDA receptors, and stabilization of inactivated sodium channels. However, its benefit is limited, and the exact therapeutic target remains incompletely understood. [@ghasemi2018]
The modest efficacy of riluzole demonstrates that partially reducing glutamate excitotoxicity is insufficient to substantially alter disease progression. ALS involves multiple overlapping pathogenic processes, and single-target interventions cannot address this complexity. [@chio2019]
[Memantine](/drugs/memantine), an NMDA receptor antagonist approved for Alzheimer's disease, failed to show efficacy in ALS Phase III trials. The failure suggests that: (1) simple blockade of NMDA receptors is insufficient; (2) excitotoxicity in ALS involves more complex mechanisms than just excessive NMDA activation; (3) the timing of intervention may be critical. [@neel2022]
The antibiotic ceftriaxone, which enhances glutamate transporter expression, was evaluated in a large Phase III trial (NCT00349674) but failed to demonstrate efficacy. The negative result indicates that increasing glutamate uptake alone does not compensate for the complex excitotoxic environment in ALS. [^21]
The consistent failure of glutamate-focused approaches despite strong biological rationale suggests fundamental limitations: (1) excitotoxicity may be a downstream effect rather than primary driver; (2) compensatory mechanisms may rapidly overcome single-target interventions; (3) the blood-brain barrier limits drug delivery to relevant CNS compartments; (4) combination therapies targeting multiple aspects of excitotoxicity may be needed. [^22]
The COX-2 inhibitor [celecoxib](/drugs/celecoxib) failed in ALS Phase III trials despite preclinical data suggesting benefit. The failure highlights the complexity of neuroinflammation in ALS—while neuroinflammation contributes to disease progression, broad immunosuppression may be counterproductive given the emerging understanding of microglia's dual (both harmful and protective) roles. [^23]
This tetracycline antibiotic with anti-inflammatory properties failed in multiple ALS clinical trials. The negative results suggest that: (1) simple inhibition of microglial activation is insufficient; (2) timing of anti-inflammatory intervention is critical; (3) the neuroimmune response may have both beneficial and detrimental components that cannot be easily separated. [^24]
P2X7 receptor antagonists targeting neuroinflammation have undergone clinical testing but failed to demonstrate efficacy. This suggests that targeting specific inflammatory pathways may be insufficient given the redundancy in neuroimmune signaling. [^25]
TREAMID, an immunomodulatory peptide, failed to meet its primary endpoint in ALS Phase II/III trials. The negative result reinforces the challenge of modulating the complex neuroimmune landscape in ALS. [^26]
The failures indicate that: (1) neuroinflammation is a double-edged sword—suppression may remove protective as well as harmful effects; (2) timing is critical—chronic established inflammation may be difficult to modify; (3) biomarker-driven patient selection based on inflammation status could improve trial design; (4) combination approaches addressing both neuroprotection and immunomodulation may be needed. [^27]
CoQ10 failed to show efficacy in large ALS trials despite strong biological rationale and preclinical data. The failure illustrates the challenge of targeting mitochondrial dysfunction in established disease and the difficulty of achieving adequate CNS concentrations with oral supplementation. [^28]
Various mitochondrial cofactors including L-carnitine, alpha-lipoic acid, and creatine have failed in ALS trials. These failures suggest that: (1) mitochondrial dysfunction may be downstream of primary pathogenic events; (2)单一的线粒体支持不足以改变疾病进程; (3)患者选择可能需要基于线粒体功能障碍的生物标志物。 [^29]
The mTOR inhibitor [rapamycin](/drugs/rapamycin) failed in ALS trials despite preclinical data showing benefit in [SOD1](/genes/sod1) models. The disconnect between animal models and human disease is a recurring theme in ALS drug development. [^30]
The consistent failure of mitochondrial-targeted approaches suggests: (1) energy failure may be a consequence rather than cause of neurodegeneration; (2) achieving adequate CNS drug concentrations remains challenging; (3) combination approaches addressing multiple aspects of energy metabolism may be needed; (4) patient selection based on mitochondrial dysfunction biomarkers could improve outcomes. [^31]
This heat shock protein coinducer showed promise in preclinical studies but failed in Phase III trials. The failure demonstrates the challenge of targeting protein homeostasis in established disease. [^32]
This VMAT2 inhibitor intended to reduce vesicular glutamate release failed to show efficacy. The negative result highlights the complexity of modulating neurotransmitter dynamics in ALS. [^33]
Anti-excitotoxic approaches beyond riluzole have universally failed, suggesting that: (1) riluzole's mechanism may be optimal for this class; (2) alternative glutamate modulation strategies face insurmountable barriers; (3) combination therapy may be needed. [^34]
Multiple trials of neural stem cell transplantation in ALS have been conducted, including the Neuralstem Phase I trial and various trials using mesenchymal stem cells. While safety has been established, efficacy has not been demonstrated. Challenges include: (1) appropriate cell type selection; (2) survival and integration of transplanted cells; (3) scaling to覆盖足够的运动神经元 population; (4) immunological rejection. [^35]
Cell therapy approaches remain experimental with no demonstrated efficacy. Key challenges include: (1) replacing the vast number of lost motor neurons is technically daunting; (2) the hostile ALS microenvironment may damage transplanted cells; (3) proper axonal connectivity is difficult to establish. [^36]
[Antisense oligonucleotides](/approaches/antisense-oligonucleotides) targeting [SOD1](/genes/sod1) mutations have been extensively developed, with tofersen (Biodenson) receiving FDA approval in 2023 for [SOD1](/genes/sod1)-positive ALS. However, the benefit was modest and limited to patients with early disease. The partial failure (modest benefit rather than dramatic reversal) reflects: (1) treatment initiated after substantial motor neuron loss; (2) incomplete SOD1 reduction; (3) the complexity of ALS beyond any single genetic cause. [^37]
Gene therapy approaches targeting [C9orf72](/genes/c9orf72) hexanucleotide repeat expansions are in development but face challenges including: (1) the complex gain-of-function mechanisms (RNA foci, dipeptide repeat proteins); (2) delivery to the appropriate brain regions; (3) patient selection based on C9orf72 status. [^38]
Gene therapy offers promise but faces fundamental challenges: (1) genetic forms represent only ~10% of ALS cases; (2) treatment after symptom onset may be too late; (3) the complex pathogenic mechanisms in even monogenic ALS are difficult to address with single-target approaches. [^39]
Ciliary neurotrophic factor ([CNTF](/proteins/cntf-protein)) and brain-derived neurotrophic factor ([BDNF](/proteins/bdnf-protein)) have been tested in ALS trials but failed due to: (1) inadequate CNS delivery; (2) short half-life; (3) inappropriate receptor expression in the ALS CNS environment. [^40]
Insulin-like growth factor-1 failed in ALS clinical trials despite positive preclinical data. The failure reflects delivery challenges and the complexity of neurotrophic signaling in established disease. [^41]
Despite the hypermetabolic state observed in many ALS patients, high-calorie enteral supplementation failed to improve outcomes. This suggests that metabolic dysfunction may be downstream rather than upstream in disease pathogenesis. [@rothst]
This heat shock protein coinducer failed to meet its primary endpoint in Phase III trials. The failure highlights the challenge of targeting protein homeostasis in chronic neurodegenerative disease. [@liu2023]
ALS is not a single disease but rather a spectrum of overlapping conditions with different genetic, molecular, and clinical features. Clinical trials that treat all ALS patients as a homogeneous group may fail to detect efficacy in patient subgroups that might benefit. [@leigh1995]
Most ALS trials enroll patients with established disease, substantial motor neuron loss, and perhaps irreversible damage. The neuroprotective paradigm requires earlier intervention, but early diagnosis remains challenging. [@mitsumoto2014]
[SOD1](/genes/sod1) transgenic mice, the primary preclinical model, may not adequately represent human sporadic ALS. The remarkable discrepancy between SOD1 mouse results and human trial outcomes suggests fundamental limitations. [@chan2023]
Many failed approaches may simply not achieve adequate drug concentrations in the relevant CNS compartments. The blood-brain barrier remains a major obstacle in CNS drug development. [@monahan2022]
ALS involves multiple overlapping pathogenic processes—protein aggregation, mitochondrial dysfunction, excitotoxicity, neuroinflammation, axonal transport disruption, RNA metabolism abnormalities. Single-target interventions cannot address this complexity. [@vu2017]
Without biomarkers to predict treatment response, select appropriate patients, or monitor target engagement, clinical trials remain inefficient. [@de2023]
Historical trials often used insensitive outcome measures, inadequate sample sizes, and inappropriate patient populations. [@benatar2014]
Future trials increasingly use genetic stratification ([SOD1](/genes/sod1), [C9orf72](/genes/c9orf72), [FUS](/genes/fus)) and biomarker-based patient selection to improve targeting. [@verde2019]
Recognizing that single-target approaches have failed, combination therapies targeting multiple pathways simultaneously are being explored. Examples include: [@bowser2011]
Novel approaches targeting fundamental disease mechanisms:
Trials are increasingly targeting pre-symptomatic or early-symptomatic patients, particularly in genetic forms like [SOD1](/genes/sod1)-ALS.
Adaptive trials, platform trials, and basket trials are being implemented to improve efficiency and reduce failure rates.
The remarkable failure rate in ALS clinical trials reflects both the extraordinary complexity of the disease and the limitations of our current drug development paradigm. Over 200 trials have yielded only two approved drugs with modest efficacy. The lessons from these failures are clear: ALS requires combination therapy approaches, earlier intervention, biomarker-driven patient selection, and models that better reflect human disease.
The recent approval of [tofersen](/drugs/tofersen) for [SOD1](/genes/sod1)-ALS, while providing only modest benefit, demonstrates that genetic targeting can produce measurable effects. The path forward requires building on these lessons—combination approaches, earlier intervention, better biomarkers, and continued investment in understanding ALS biology.
The failures documented here are not endpoints but rather essential steps in a learning process. Each negative trial provides valuable information about disease biology, trial design, and therapeutic target validation. The accumulated knowledge from these failures now guides a more sophisticated approach to ALS drug development that holds promise for future success.
Amyotrophic lateral sclerosis (ALS) remains one of the most challenging neurodegenerative diseases to treat, with nearly all clinical trials failing to demonstrate disease modification over the past three decades[@bensimon1994][@abe2017]. Despite tremendous advances in understanding the molecular pathogenesis of ALS—including discoveries related to C9orf72 repeat expansions, SOD1 mutations, TDP-43 pathology, and RNA metabolism dysregulation—the translation of these insights into effective therapies has been remarkably unsuccessful[@cudkowicz2006]. This disconnect between mechanistic understanding and therapeutic success demands a critical examination of failed approaches to identify patterns, pitfalls, and potential paths forward.
The landscape of ALS clinical trials reflects a history of enthusiasm followed by disappointment. From early trials of neurotrophic factors to recent antisense oligonucleotide therapies, the vast majority of interventions have failed to meet their primary endpoints[@miller2007]. This pattern extends across multiple drug classes, mechanisms, and therapeutic strategies, suggesting fundamental challenges in ALS drug development that transcend any single approach. Understanding why these failures occurred provides crucial insights for designing more effective future trials and identifying truly disease-modifying interventions.
This analysis examines the major categories of failed approaches in ALS, the specific reasons for trial failures, and the lessons that can be extracted to improve future therapeutic development. The goal is not merely to catalog failures but to understand the underlying principles that may explain why ALS has proven so resistant to intervention and what strategies might finally break through this therapeutic deadlock.
Antisense oligonucleotides (ASOs) represent perhaps the most rational approach to ALS therapeutics, targeting the genetic basis of disease in patients with specific mutations. Despite strong biological rationale and recent successes in other genetic neurodegenerative diseases, ALS ASO trials have yielded disappointing results[@loeravalencia2019].
Tocerersen (IONIS-SOD1Rx) targeted patients with SOD1 mutations, representing a genetically defined subpopulation with a clear molecular target. The phase III trial failed to demonstrate clinical benefit despite reducing SOD1 protein levels in cerebrospinal fluid, indicating that simply lowering mutant protein levels may be insufficient if initiated after substantial motor neuron loss has already occurred[@pasinetti2013]. This failure highlighted the critical issue of timing in ALS therapy—perhaps intervention must begin before symptoms emerge to be effective.
Another ASO targeting C9orf72 repeat expansions, which represent the most common genetic cause of ALS, also failed to demonstrate efficacy in clinical trials[@ghasemi2018]. The challenges with C9orf72 ASOs include the complex biology of repeat expansion disorders, where toxic gain-of-function mechanisms (repeat-containing RNA foci, dipeptide repeat proteins) may be equally or more important than loss of normal protein function[@chio2019]. Simply reducing C9orf72 protein levels may not address the dominant toxic species.
Numerous small molecules with neuroprotective mechanisms have failed in ALS trials. Edaravone, a free radical scavenger approved for ALS in Japan based on a marginal effect in a subset of patients, failed to demonstrate efficacy in subsequent trials conducted according to Western regulatory standards[@neel2022][^21]. This discrepancy raised questions about trial design, patient selection, and the generalizability of efficacy signals observed in specific populations.
Masitinib, a tyrosine kinase inhibitor with effects on microglia and neuroinflammation, failed in phase III trials despite prior signals of efficacy[^22]. The failure underscored the complexity of targeting neuroinflammation in ALS, where the role of immune activation remains ambiguous—potentially protective in some contexts while contributing to damage in others[^23]. Simply suppressing inflammation without understanding its precise role in individual patients may prove counterproductive.
Ceftriaxone, an antibiotic with purported glutamate transporter-upregulating activity, failed in phase III trials despite extensive preclinical data supporting its neuroprotective effects[^24]. The trial exemplified how positive findings in SOD1 transgenic mice failed to predict efficacy in human ALS, again raising questions about model validity and the fundamental differences between genetic and sporadic disease.
Cell replacement therapies for ALS have generated substantial interest but limited success. Multiple trials have tested various sources of stem cells including mesenchymal stem cells, neural stem cells, and induced pluripotent stem cell-derived motor neurons[^25]. While these approaches have generally been safe, none have demonstrated clear clinical efficacy.
The challenges with cell-based therapies in ALS are multifaceted. First, replacing motor neurons that have already been lost may be insufficient if the pathological environment continues to attack new cells[^26]. Second, the diffuse nature of ALS pathology affecting multiple brain regions and spinal cord segments makes localized cell delivery inadequate. Third, the immune-privileged nature of the central nervous system creates barriers to cell survival and integration.
Neural stem cell trials have shown some signals of potential benefit in terms of slowing disease progression, but effects have been modest and inconsistent[^27]. The interpretation of these trials is complicated by the natural history variability of ALS and the challenge of identifying appropriate outcome measures that capture subtle therapeutic effects.
The choice of clinical endpoints represents a fundamental challenge in ALS clinical trials. The traditional endpoint of survival requires large, long-duration trials that are prohibitively expensive and time-consuming[^28]. Surrogate markers that predict clinical outcomes have been elusive, forcing sponsors to rely on functional measures like the ALS Functional Rating Scale-Revised (ALSFRS-R) that may be insensitive to subtle therapeutic effects[^29].
The ALSFRS-R, while the standard primary endpoint for most ALS trials, has significant limitations. It captures a composite of diverse functions including speech, swallowing, handwriting, and respiratory status, but may not be equally sensitive to therapeutic effects across all domains[^30]. Furthermore, the rate of decline on ALSFRS-R varies substantially between patients, making it difficult to detect treatment effects in heterogeneous populations.
Biomarker development in ALS has lagged behind other neurodegenerative diseases. While cerebrospinal fluid biomarkers like neurofilament light chain (NfL) show promise for disease progression monitoring, none have been validated as surrogate endpoints for clinical trials[^31]. The lack of robust biomarkers forces sponsors to rely on clinical measures that may miss important biological effects of experimental therapies.
ALS encompasses a heterogeneous group of disorders with different genetic, pathological, and clinical features. Sporadic ALS, which accounts for the majority of cases, is itself likely multiple distinct diseases with different underlying mechanisms that respond differently to therapy[^32]. Clinical trials that enroll heterogeneous populations may dilute treatment effects that are only relevant to specific subgroups.
The challenge of patient selection is compounded by the lack of predictive biomarkers that could identify patients most likely to respond to specific therapies. Without stratification, trials may fail to detect efficacy in subgroups while showing no effect in the overall population[^33]. This heterogeneity may explain why some trials show signals in post-hoc analyses that are not confirmed in prospective primary analyses.
Geographic and demographic factors also contribute to trial variability. Different populations may have different rates of disease progression, different genetic backgrounds influencing drug metabolism, and different standard of care practices that affect outcomes[^34]. International trials face the challenge of reconciling these differences while maintaining adequate enrollment.
The timing of therapeutic intervention may be the most critical factor determining trial success. ALS typically presents after 50-70% of motor neurons have already been lost, with a preclinical period that may extend for years before symptoms emerge[^35]. Interventions initiated at symptom onset may simply be too late to rescue neurons that have already undergone irreversible damage.
This timing hypothesis suggests that disease-modifying therapies may only be effective if initiated during the presymptomatic or early symptomatic phase. However, identifying presymptomatic individuals for trials presents enormous practical challenges, requiring genetic testing (for familial cases) or development of predictive biomarkers that are not yet available[^36]. The few trials that have attempted presymptomatic intervention have struggled with enrollment and ethical considerations.
The failure of antisense oligonucleotides in SOD1 ALS illustrates this timing challenge. Tocerersen successfully reduced CSF SOD1 levels but failed to slow clinical decline, potentially because patients were enrolled too late in the disease course[^37]. By the time neurons are lost, simply reducing production of mutant protein cannot restore function.
The reliance on SOD1 transgenic mice as the primary preclinical model for ALS drug development represents a fundamental limitation that likely contributes to trial failures. These mice overexpress mutant human SOD1 at levels far exceeding those seen in human ALS, creating an artificial disease environment that may not predict human response[^38].
Studies comparing drug effects in SOD1 mice versus other models or in human tissues have frequently found discrepancies. Drugs that work beautifully in SOD1 mice often fail in human trials, while some drugs that show efficacy in human cellular models fail in mouse models[^39]. This discordance suggests that the SOD1 mouse model, while useful for certain applications, does not adequately capture the biology of human ALS.
Alternative models including C9orf72 knock-in mice, TDP-43 transgenic models, and induced pluripotent stem cell-derived motor neurons from ALS patients may provide more relevant preclinical platforms[^40]. However, these models are less well-characterized and have not yet been validated for drug screening in the same way as SOD1 mice.
An analysis of failed ALS clinical trials reveals patterns in target selection that may explain repeated failures. Many targets were chosen based on preclinical studies in SOD1 mice that may not generalize to sporadic ALS. Others were selected based on pathways that appear to be downstream effects rather than primary disease drivers, meaning that blocking them does not address the root cause of neurodegeneration[^41].
The excitotoxicity hypothesis exemplifies this pattern. While glutamate is clearly toxic to motor neurons and contributes to ALS pathogenesis, it may be a downstream consequence of upstream defects in RNA metabolism, mitochondrial function, or proteostasis. Blocking excitotoxicity without fixing these upstream problems may be like treating symptoms rather than disease[@rothst].
Similarly, targeting neuroinflammation may be premature given our incomplete understanding of its role in ALS. Inflammation could be a protective response to neurodegeneration that becomes harmful when chronic, or it could be disease-driving in some patients but not others. Broadly suppressing inflammation without this understanding may produce net neutral or negative effects[@liu2023].
Beyond target selection, trial design issues have contributed to ALS trial failures. Many trials have been underpowered to detect realistic effect sizes, leading to false negative results. The variable rate of disease progression in ALS means that large sample sizes are needed to detect modest effects, but the costs and timelines required for adequately powered trials have deterred sponsors[@leigh1995].
Placebo response and the natural history of ALS present additional challenges. Patients participating in trials may seek out unproven therapies or make lifestyle changes that alter their disease course, creating noise that obscures treatment effects. The lack of biomarker stratification means that responders and non-responders are analyzed together, potentially diluting efficacy signals[@mitsumoto2014].
The choice of control group has also been debated. Some trials use placebo controls, while others use historical controls or add-on designs. Each approach has limitations that may affect the interpretation of results. The field has generally moved toward randomized, placebo-controlled designs, but these are expensive and time-consuming to conduct.
The repeated failures of "one-size-fits-all" approaches in ALS have increasingly pointed toward precision medicine strategies. By identifying the specific molecular subtype of ALS in individual patients, it may be possible to match patients with therapies that target their specific disease mechanism[@chan2023]. This approach has been successful in other genetic diseases like spinal muscular atrophy, where antisense oligonucleotides tailored to specific mutations have demonstrated dramatic efficacy.
Genetic stratification represents the most advanced precision medicine approach in ALS. Patients with SOD1 mutations can be treated with SOD1-targeting ASOs, those with C9orf72 expansions with C9orf72-targeting approaches, and those with FUS mutations with FUS-targeted strategies[@monahan2022]. While trials in these genetic subpopulations have so far failed, the biological rationale remains strong and future trials with optimized timing and endpoints may succeed.
Beyond genetics, biomarker-based stratification may allow identification of patients with specific pathological subtypes. Cerebrospinal fluid protein signatures, neuroimaging markers, and electrophysiological measures may eventually allow assignment of patients to mechanistic subgroups that predict response to specific therapies[@vu2017].
Given the complex, multifactorial nature of ALS pathogenesis, combination therapy approaches may be necessary for meaningful disease modification. Rather than targeting single pathways, combinations that address multiple disease mechanisms simultaneously may be required[@de2023]. However, testing combinations presents enormous practical challenges in terms of dose selection, endpoint selection, and regulatory approval.
The design of combination trials requires careful consideration of mechanism interactions. Some drug combinations may be synergistic, while others may be antagonistic or produce unexpected toxicities. Preclinical studies to identify optimal combinations before advancing to clinical trials are essential but rarely conducted due to resource constraints.
Advancing biomarker development represents a critical priority for improving ALS clinical trials. Biomarkers that can identify patients likely to respond to specific therapies would enable enrichment strategies that increase the probability of detecting efficacy[@benatar2014]. Biomarkers that track disease progression could serve as surrogate endpoints, allowing shorter trials with smaller sample sizes.
Neurofilament light chain (NfL) in cerebrospinal fluid and blood has emerged as the most promising ALS biomarker. Elevated NfL levels correlate with disease progression and have been proposed as a prognostic marker[@verde2019]. However, NfL has not yet been validated as a surrogate endpoint for clinical trials, and its utility for predicting treatment response remains uncertain.
Other biomarker candidates under investigation include CSF total tau, phosphorylated tau, neurogranin, and various inflammatory markers. The field continues to search for biomarkers that can predict treatment response and disease progression with sufficient sensitivity and specificity to be useful in clinical trials[@bowser2011].
The autopsy of failed ALS approaches reveals consistent patterns that explain the persistent therapeutic failures in this devastating disease. Target selection based on inadequate preclinical models, enrollment of patients too late in the disease course, insensitive clinical endpoints, and failure to account for patient heterogeneity have collectively contributed to a decades-long string of negative trials. While this record is discouraging, the insights gained from these failures provide a blueprint for more rational future therapeutic development.
The path forward requires addressing the fundamental challenges that have undermined previous efforts. Precision medicine approaches that match patients to targeted therapies based on genetic and biomarker stratification offer the most promising strategy for achieving meaningful disease modification. Biomarker development is essential for enabling patient selection, endpoint optimization, and surrogate marker validation. And trial designs must evolve to accommodate the complexities of ALS while remaining practically feasible.
Despite the failures, the ALS research community has not abandoned the quest for effective therapies. Each negative trial provides information that guides future efforts, and the accumulating knowledge about disease mechanisms continues to identify new therapeutic targets. While the path to effective ALS treatments remains long, the systematic analysis of past failures ensures that future efforts are built on solid foundations.
[@leigh1995]: Leigh PN, Swash M. [Clinical trials in ALS: problems, problems, problems](https://pubmed.ncbi.nlm.nih.gov/8620019/). J Neurol Neurosurg Psychiatry. 1995;59(3):252-254.
[@mitsumoto2014]: Mitsumoto H, Brooks BR, Silani V. [Clinical trials in ALS: recasting the concept of trial design](https://pubmed.ncbi.nlm.nih.gov/25428889/). Neurology. 2014;83(19):1706-1707.
[@chan2023]: Chan M, Liu H, Martin A. [Precision medicine in ALS](https://pubmed.ncbi.nlm.nih.gov/36718744/). Nat Rev Neurol. 2023;19(1):43-57.
[@monahan2022]: Monahan Z, Shepheard SR, Whitelaw V, et al. [Therapeutic strategies for ALS based on the underlying genetic causes](https://pubmed.ncbi.nlm.nih.gov/35698552/). Nat Rev Neurol. 2022;18(8):481-498.
[@vu2017]: Vu L, Bowser R. [Fluid biomarkers for amyotrophic lateral sclerosis](https://pubmed.ncbi.nlm.nih.gov/28240226/). Neurotherapeutics. 2017;14(1):119-134.
[@de2023]: De Loof A, Goossens MC, De Smet J, et al. [Combination therapy for ALS: rationale and priorities](https://pubmed.ncbi.nlm.nih.gov/36701267/). Expert Opin Pharmacother. 2023;24(5):573-588.
[@benatar2014]: Benatar M, Turner MR. [Optimizing clinical trials in ALS](https://pubmed.ncbi.nlm.nih.gov/25178892/). Nat Rev Neurol. 2014;10(10):548-557.
[@verde2019]: Verde F, Steinacker P, Weishaupt JH, et al. [Neurofilament light chain as biomarker in ALS](https://pubmed.ncbi.nlm.nih.gov/30950758/). J Neurol Neurosurg Psychiatry. 2019;90(8):917-927.
[@bowser2011]: Bowser R, Turner MR, Shefner J. [Biomarkers in ALS: an integrated overview](https://pubmed.ncbi.nlm.nih.gov/21600958/). Nat Rev Neurol. 2011;7(11):617-627.
[^32]: [Reference missing - citation needed]
[^33]: [Reference missing - citation needed]
[^34]: [Reference missing - citation needed]
[^35]: [Reference missing - citation needed]
[^36]: [Reference missing - citation needed]
[^37]: [Reference missing - citation needed]
[^38]: [Reference missing - citation needed]
[^39]: [Reference missing - citation needed]
[^40]: [Reference missing - citation needed]