Efficacy of Corticosteroid Therapy for HTLV-1-Associated Myelopathy: A Randomized Controlled Trial (HAMLET-P)

Corticosteroids are most commonly used to treat HTLV-1-associated myelopathy (HAM); however, their clinical efficacy has not been tested in randomized clinical trials. This randomized controlled trial included 8 and 30 HAM patients with rapidly and slowly progressing walking disabilities, respectively. Rapid progressors were assigned (1:1) to receive or not receive a 3-day course of intravenous methylprednisolone in addition to oral prednisolone therapy. Meanwhile, slow progressors were assigned (1:1) to receive oral prednisolone or placebo. The primary outcomes were a composite of ≥1-grade improvement in the Osame Motor Disability Score or ≥30% improvement in the 10 m walking time (10 mWT) at week 2 for rapid progressors and changes from baseline in 10 mWT at week 24 for slow progressors. In the rapid progressor trial, all four patients with but only one of four without intravenous methylprednisolone achieved the primary outcome (p = 0.14). In the slow progressor trial, the median changes in 10 mWT were −13.8% (95% CI: −20.1–−7.1; p < 0.001) and −6.0% (95% CI: −12.8–1.3; p = 0.10) with prednisolone and placebo, respectively (p for between-group difference = 0.12). Whereas statistical significance was not reached for the primary endpoints, the overall data indicated the benefit of corticosteroid therapy. (Registration number: UMIN000023798, UMIN000024085)


Introduction
Human T-lymphotropic virus type 1 (HTLV-1) infects at least 5-10 million people globally. Moreover, it causes rare but devastating diseases, including HTLV-1-associated myelopathy (HAM) and adult T-cell leukemia/lymphoma (ATL) in a small proportion of infected individuals [1][2][3]. HAM is characterized by chronic spinal cord inflammation, particularly at the thoracic level, resulting in neurological disorders such as spastic paraparesis, sensory disturbance in the legs, and bladder and bowel dysfunction [4]. To date, there is no treatment for HAM. Interferon-α is the only drug that demonstrated clinical efficacy in a randomized controlled trial (RCT) [5]. However, this medication is seldom used because it is not highly efficient [6]. Although data are limited, corticosteroids are most commonly used to maintain motor function by suppressing inflammation [4, 6,7].
Typically, neurological deterioration progresses slowly in HAM. However, the rate of progression or the disease activity varies among patients, ranging from minor walking abnormalities for more than a decade to the inability to walk within years [8][9][10][11]. The international HAM guidelines (2019) stated that patients should be classified according to the three types of progression (rapid, slow, and very slow) [12]. Low-dose steroid therapy with or without high-dose induction may be recommended based on the category. However, these recommendations are based on retrospective observational studies and clinical experiences [13][14][15]. Therefore, to our knowledge, this multicenter RCT is the first to assess the efficacy of intravenous methylprednisolone induction therapy for patients with rapidly progressing HAM (rapid progressors) and oral prednisolone treatment for those with slowly progressing HAM (slow progressors). before enrollment. The study protocol, statistical analysis plan, and CONSORT checklist are available in Files S1-S4.

Participants
Patients aged ≥18 years and diagnosed with definite HAM based on the Belem criteria were eligible for this study [16]. Upon enrollment, to assess motor ability, the participants had to walk for ≥10 m with or without walking aids. The exclusion criteria included those with complications such as neurological diseases other than HAM, comorbidities affecting motor function (e.g., osteoarthritis, rheumatoid arthritis), severe organ dysfunction, cancer, and contraindications to corticosteroids; those receiving corticosteroids or drugs targeting HAM (e.g., interferon-alpha, immunosuppressive agents, antiretroviral agents, and valproic acid) within 12 or 48 weeks prior to providing consent in rapid and slow progressors, respectively; and those treated with drugs strongly affecting CYP3A4 metabolism.

Study Design
This study comprised two trial arms: the prospective, randomized, open, blinded endpoint trial to assess the efficacy of intravenous methylprednisolone induction therapy for rapid progressors; and the prospective, randomized, double-blind, placebo-controlled trial to evaluate the efficacy of oral prednisolone therapy for slow progressors ( Figure S1). Study visits were performed for progression rate assessment at −12, −8, −4, and 0 weeks (last assessment) (progression assessment period) as well as at day 1 (baseline) and 2, 4, 8,12,24,28,32,36, and 48 weeks (treatment period). After eligibility screening, patients were registered for the progression rate assessment based on changes in motor function. Then, they were classified as rapid progressors at the time of registration if the clinical history within 12 weeks prior to registration met one of the following rapid progressor criteria: ≥30% worsening in the time taken to walk 10 m (10 m walking time [10 mWT]) or ≥1-grade deterioration in the Osame Motor Disability Score (OMDS), which is a specific scale for HAM, ranging from 0 to 13, with higher scores indicating greater disability [17]. For other patients, motor function was assessed during the 12-week progression assessment period and was classified as follows: rapid progressors, those who fulfilled the aforementioned rapid progressor criteria; slow progressors, those who experienced 10-<30% worsening in 10 mWT at the last assessment; and non-progressors, those who experienced <10% worsening. After the progression assessment period, non-progressors were followed-up without the trial drug for 48 weeks and were reclassified as rapid or slow progressors if their 10 mWT worsened by ≥30% or 10-<30% compared with the best time recorded during the progression assessment period. After the participants were classified into the progression groups, they were assigned to a specific treatment and followed up for 48 weeks.

Trial Intervention
Rapid progressors were randomly assigned (1:1) to receive a 3-day course of intravenous methylprednisolone at a dose of 1 g/day (Pfizer Japan Inc., Tokyo, Japan) along with oral prednisolone once per day (Nipro Pharma Corporation, Osaka, Japan) (pulse group), or oral prednisolone alone (non-pulse group; Figure S2). The non-pulse group did not receive placebo. In both groups, oral prednisolone was administered at a dose of 0.5 mg/kg body weight (BW) until week 2 and was then tapered to 20 mg by week 4 and to 5 mg by week 24. After alternate-day administration of 5 mg prednisolone for 14 days (at week 26), all patients discontinued prednisolone, except those who required a prespecified additional treatment for deterioration. Then, they were followed up until week 48. If patients experienced motor function deterioration after week 4, additional treatment (oral prednisolone with or without intravenous methylprednisolone) was administered according to the protocol ( Figure S2).
Slow progressors were randomly assigned (1:1) to receive oral prednisolone or matching placebo once daily ( Figure S3). In the prednisolone group, prednisolone was administered at a dose of 0.5 mg/kg BW until day 7, tapered to 5 mg/day by week 24, and then Viruses 2022, 14, 136 4 of 17 continued at 5 mg/day until week 48. In the placebo group, 5 mg/day prednisolone was initiated after the 24 week placebo period and was continued until week 48. If patients experienced ≥10% worsening in 10 mWT twice or ≥1-grade deterioration in OMDS after week 24, prednisolone was increased according to the protocol.
Medications for urinary symptoms and spasticity that had been started prior to enrollment were continued without changes throughout the trial.

Randomization and Blinding
We used web-based interactive response technology for randomization. Rapid progressors were stratified according to the use of walking aids (none or unilateral/bilateral) and randomized with permuted blocks of two patients. Slow progressors were randomized using the minimization method, adjusting for sex (male/female), use of walking aids (none or unilateral/bilateral), and trial site. Only physicians who performed clinical evaluations were blinded to the treatment allocation of rapid progressors. Therefore, patients and treating physicians knew their treatment. For slow progressors, all patients and study staff were blinded, and study drugs in numbered containers were supplied by an external vendor to ensure blinding.

Disease Evaluation
In clinical evaluations, we assessed mobility, dysuria, sensory dysfunction of the lower extremities, and the patient's self-assessment of HAM-related symptoms. Mobility was evaluated using the OMDS, 10 mWT, 2 and 6 min walk tests (distance walked in 2 and 6 min, 2 MWD, and 6 MWD), and the timed up-and-go test (time required to stand up from a chair, walk 3 m away, turn, walk back, and sit down again). Spasticity of the knee extensor and flexor muscles was evaluated using the Modified Ashworth Scale (MAS; grades: 0, 1, 1+, 2, 3, and 4), with higher grades indicating more severe spasticity [18]. Dysuria was assessed using the International Prostate Symptom Score (IPSS), with higher scores (range: 0-35) indicating more difficulty in urinating [19]; the Overactive Bladder Symptom Score (OABSS), with higher scores (range: 0-15) indicating more severe urinary urgency [20]; the International Consultation on Incontinence Questionnaire-Short Form (ICIQ-SF), with higher scores (range: 0-21) indicating more severe incontinence [21]; and the Nocturia-Quality of Life Questionnaire (N-QOL), with lower scores (range: 0-100) indicating a lesser quality of life [22]. The patient's subjective assessment of the global condition of HAM, mobility, and sensory dysfunction was performed using a 100 mm visual analog scale (VAS), in which higher values indicate more severe conditions. The Insituto de Pesquisa Clinica Evandro Chagas disability score (IPEC) 1 was used to comprehensively evaluate HAM-related symptoms [23].
Regarding laboratory analyses, we performed complete blood count, blood chemistry test, and cerebrospinal fluid (CSF) analysis. We also evaluated the CSF concentrations of neopterin (a marker of immune system activation) and CXCL10 (a chemokine mainly induced by interferon gamma), as well as HTLV-1 proviral loads in the peripheral blood mononuclear cells (PBMCs) and CSF cells, as described in the Supplementary Materials [24]. For safety evaluation, adverse events (AEs) were assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.0. Serious AEs were defined as life-threatening conditions, congenital abnormalities or birth defects, diseases considered fatal by the investigator, and conditions resulting in hospitalization or prolonged hospital stay, persistent or clinically serious disability or incapacity, and death. Whether AEs were attributed to the treatment regimen was confirmed by the investigators.

Outcomes
In rapid progressors, the primary outcome was a composite of ≥1-grade improvement in OMDS or ≥30% improvement in 10 mWT at week 2 compared with baseline. The secondary outcomes included: each component of the primary outcome; changes in 10 mWT, 2 MWD, and 6 MWD, and CSF neopterin and CXCL10 concentrations; motor In slow progressors, the primary outcome was a change in 10 mWT at week 24 from baseline. The secondary outcomes included: changes in 2 MWD, 6 MWD, and CSF markers; differences in changes in 10 mWT, 2 MWD, 6 MWD, and CSF markers between the placebo and active drug periods (from baseline to week 24 and from week 24 to week 48) in the placebo group; and safety.

Statistical Analysis
We analyzed the treatment efficacy using the intention-to-treat population (full analysis set, FAS), which comprised all eligible patients who received at least one dose of treatment. Moreover, the per-protocol set (PPS), in which patients with serious protocol violations were excluded from the FAS, was evaluated. All patients who received at least one dose of the trial drug underwent safety analysis.
The proportions of rapid progressors who met the primary outcome, with 95% confidence intervals (CIs), were analyzed. Next, between-group differences were evaluated using Fisher's exact test. Because this study arm was underpowered for detecting even large differences due to its small sample size, we hypothesized that intravenous methylprednisolone therapy is effective if the proportion of patients who meet the primary endpoint in the pulse group is higher than that in the non-pulse group.
In slow progressors, the primary endpoint was analyzed using the mixed-effect model with the repeated-measures (MMRM) approach. The model included the fixed effects of treatment (prednisolone or placebo), 10 mWT at baseline, time points (week 4, 12, and 24), and treatment-by-time-point interactions. We calculated the least-squares (LS) means with 95% CIs at each time point and assessed the between-group differences at week 24. Of the key secondary outcomes, 2 MWD and 6 MWD were analyzed using MMRM. The CSF neopterin and CXCL10 concentrations at week 24 were analyzed by calculating the LS means with 95% CIs using analysis of covariance (ANCOVA). A natural logarithmic transformation was applied to produce a normal distribution in 10 mWT and CSF marker values. To interpret data, the results were expressed as median percent change with 95% CI, which were calculated from the exponential of the LS mean estimates. Further details are provided in the Supplementary Materials.
Statistical analysis was performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). p-values were two-sided, and a significance threshold of 0.05 was applied for all tests.

Sample Size Calculation
It was challenging to recruit rapid progressors because the annual incidence of rapidly progressing walking disability is low in Japan. Therefore, we set a minimum sample size of only four patients per group. One more patient was recruited if five of eight were assigned to the non-pulse group. Assuming that gait function will improve in ≥50% of patients receiving intravenous methylprednisolone and none among patients receiving oral prednisolone alone, the probability that more patients in the pulse group can achieve the primary endpoint compared with those in the non-pulse group was >80%.
Based on our clinical experience, we assumed that the 10 mWT of slow progressors could improve by at least 15% (0.165 in log) with treatment and could worsen by at least 6% without treatment (0.058 in log). Therefore, the sample size was set at 20 patients per group based on the estimation that this could provide 90% power for detecting a between-group difference of 21% (0.223 in log) using ANCOVA with a 5% significance threshold, assuming that the standard deviation was 0.21 (in log). Under the same assumption, 15 patients per group were estimated to provide a power of 80%.

Study Population
From 31 August 2016 to 24 June 2019, 46 Japanese patients across 5 trial sites were enrolled, and 44 patients underwent the progression assessment ( Figure 1). After the evaluation, patients were classified as rapid (n = 7), slow (n = 14), and non-progressors (n = 22). Several of the 22 non-progressors later experienced deterioration, and they were subsequently reclassified as rapid (n = 2) and slow progressors (n = 16) accordingly. Thus, 9 rapid and 30 slow progressors were randomized. Among them, 8 and 26 completed the trial, respectively. Due to the rarity of HAM, the target number of slow progressors (n = 40) was not achieved. group difference of 21% (0.223 in log) using ANCOVA with a 5% significance threshold, assuming that the standard deviation was 0.21 (in log). Under the same assumption, 15 patients per group were estimated to provide a power of 80%.

Study Population
From 31 August 2016 to 24 June 2019, 46 Japanese patients across 5 trial sites were enrolled, and 44 patients underwent the progression assessment ( Figure 1). After the evaluation, patients were classified as rapid (n = 7), slow (n = 14), and non-progressors (n = 22). Several of the 22 non-progressors later experienced deterioration, and they were subsequently reclassified as rapid (n = 2) and slow progressors (n = 16) accordingly. Thus, 9 rapid and 30 slow progressors were randomized. Among them, 8 and 26 completed the trial, respectively. Due to the rarity of HAM, the target number of slow progressors (n = 40) was not achieved. Trial flow chart. * One of five rapid progressors in the non-pulse group was excluded from the full and per-protocol sets due to ineligibility associated with treatment with carbamazepine (a strong CYP3A4 inducer) throughout the trial. AE, adverse event; IV mPSL, intravenous methylprednisolone; PSL, prednisolone.
All randomized patients were included in the FAS, except for one rapid progressor who was taking the strong CYP3A4 inducer carbamazepine. All rapid progressors in the FAS were included in the PPS, whereas three slow progressors were excluded from the PPS due to lack of gait function assessment data at week 24 (n = 2) and low treatment adherence (≤80%, n = 1). The mean rates of adherence to the trial regimen (the proportion of administered doses per planned doses) were 99.8% and 99.2% in rapid and slow progressors, respectively.
Demographic characteristics were balanced in each progressor group between the treatment arms (Tables 1 and 2). The median age of rapid and slow progressors was 62.0 and 64.0 years, respectively. All rapid (n = 8) and 23 (76.7%) of 30 slow progressors were women. The two progressor groups commonly presented with grade 5 OMDS (which requires unilateral support for walking) (5 rapid [62.5%] and 12 slow [40.0%] progressors). However, slow progressors in the prednisolone arm had more severe motor disability than those in the placebo arm (median OMDS grade, 5 vs. 4; median 10 mWT, 12.5 vs. 8.1 sec, respectively). Trial flow chart. * One of five rapid progressors in the non-pulse group was excluded from the full and per-protocol sets due to ineligibility associated with treatment with carbamazepine (a strong CYP3A4 inducer) throughout the trial. AE, adverse event; IV mPSL, intravenous methylprednisolone; PSL, prednisolone.
All randomized patients were included in the FAS, except for one rapid progressor who was taking the strong CYP3A4 inducer carbamazepine. All rapid progressors in the FAS were included in the PPS, whereas three slow progressors were excluded from the PPS due to lack of gait function assessment data at week 24 (n = 2) and low treatment adherence (≤80%, n = 1). The mean rates of adherence to the trial regimen (the proportion of administered doses per planned doses) were 99.8% and 99.2% in rapid and slow progressors, respectively.
Demographic characteristics were balanced in each progressor group between the treatment arms (Tables 1 and 2). The median age of rapid and slow progressors was 62.0 and 64.0 years, respectively. All rapid (n = 8) and 23 (76.7%) of 30 slow progressors were women. The two progressor groups commonly presented with grade 5 OMDS (which requires unilateral support for walking) (5 rapid [62.5%] and 12 slow [40.0%] progressors). However, slow progressors in the prednisolone arm had more severe motor disability than those in the placebo arm (median OMDS grade, 5 vs. 4; median 10 mWT, 12.5 vs. 8.1 sec, respectively).

Efficacy Analysis of Rapid Progressors
At week 2, the OMDS improved by ≥1 grade in all 4 patients in the pulse group but none in the non-pulse group (p = 0.029; Figure 2). Meanwhile, the 10 mWT improved in 3 of 4 patients from each group after initiating steroid treatment, and 1 patient from each group experienced ≥30% improvement (p = 1.00). Thus, the primary outcome (improvement in OMDS or 10 mWT at week 2) was achieved by 4 patients with and 1 without intravenous methylprednisolone therapy (100% and 25%, respectively; difference: 75% [95% CI: −5.3-99.4; p = 0.14]; Table 3). At week 24, all patients in the pulse group maintained the OMDS improvement. In the non-pulse group, 1 patient experienced improvement in the OMDS at week 8; however, 2 patients met the deterioration criteria and required additional steroid therapy. At week 26, prednisolone treatment was discontinued in 4 and 2 patients in the pulse and non-pulse arms, respectively, according to the protocol. However, 3 and 2 patients resumed treatment due to deterioration.
A total of 3 of 4 patients in each group experienced an improvement in the 2 MWD and 6 MWD (Table 3, Figure 3 and Figure S4). Meanwhile, the CSF neopterin and CXCL10 concentrations decreased at week 2, and then they increased after the drug doses were tapered in both arms. Other disease evaluation parameters, including the timed up-and-go test, MAS, IPEC1, VAS, and urinary symptom scores, and HTLV-1 proviral loads, varied between patients ( Figure S5).

Efficacy Analysis of Slow Progressors
The median changes from baseline in 10 mWT at week 24 were −13.8% (95% CI: −20.1-−7.1; p < 0.001) and −6.0% (95% CI: −12.8-1.3; p = 0.10) in the prednisolone and placebo groups, respectively (p for between-group difference = 0.12; Figure 4 and Table 4). Although the 10 mWT at week 24 significantly improved from baseline only in the prednisolone group, the changes did not significantly differ between the prednisolone and placebo groups. The PPS analysis results were similar to those of the FAS. cerebrospinal fluid; OMDS, Osame Motor Disability Score; 2 MWD, 2 min walk distance; 6 MWD, 6 min walk distance; 10 mWT, 10 m walking time.
A total of 3 of 4 patients in each group experienced an improvement in the 2 MWD and 6 MWD ( Table 3, Figures 3 and S4). Meanwhile, the CSF neopterin and CXCL10 concentrations decreased at week 2, and then they increased after the drug doses were tapered in both arms. Other disease evaluation parameters, including the timed up-and-go test, MAS, IPEC1, VAS, and urinary symptom scores, and HTLV-1 proviral loads, varied between patients ( Figure S5).

Efficacy Analysis of Slow Progressors
The median changes from baseline in 10 mWT at week 24 were −13.8% (95% CI: −20.1-−7.1; p < 0.001) and −6.0% (95% CI: −12.8-1.3; p = 0.10) in the prednisolone and placebo groups, respectively (p for between-group difference = 0.12; Figure 4 and Table 4). Although the 10 mWT at week 24 significantly improved from baseline only in the prednisolone group, the changes did not significantly differ between the prednisolone and placebo groups. The PPS analysis results were similar to those of the FAS.
The results of other disease evaluations were analyzed via post hoc analysis (Tables S2 and S3 and Figure S7). Of 15 patients in the prednisolone group, 2 and 5 experienced ≥1-grade improvement in the OMDS from baseline at week 24 and week 48, respectively.
No patients in the placebo group showed improvement in the OMDS while on placebo, but 7 of 13 patients had improvements after receiving prednisolone. The total IPEC1 score at week 24 significantly improved with prednisolone compared with placebo (median [IQR] changes from baseline: −2.0 [−5.0-−1.0] and 0.0 [0.0-0.0]; p = 0.002). There were no remarkable differences in other parameters.

Safety Analysis
None of the participants died during the trial. All rapid progressors developed AEs, and both treatment arms presented with steroid-related AEs ( Table 5). The only serious AE, which resulted in trial discontinuation, was urinary tract infection in the non-pulse group. Table 5. Safety profile of rapid and slow progressors. During the first 24 weeks, 11 and 3 slow progressors in the prednisolone and placebo groups, respectively, developed AEs that were considered to be related to the treatment regimen; however, there were no serious AEs (Table 5). Between week 25 and week 48, during which both treatment arms received oral prednisolone, 8 and 5 patients in the prednisolone and placebo groups experienced steroid-related AEs, respectively. In addition, the prednisolone group presented with 2 serious AEs (ATL and herpes zoster). Meanwhile, the serious AE in the placebo group was facial palsy. During the 48 week treatment period, 2 patients in the prednisolone group developed herpes zoster. Moreover, urinary tract infection was observed in 4 and 2 patients in the prednisolone and placebo groups, respectively.

Discussion
We investigated the efficacy of corticosteroid therapy for HAM according to disease activity. The primary endpoint was achieved in rapid progressors. That is, the proportion of patients receiving intravenous methylprednisolone pulse therapy who achieved improvement in the OMDS or 10 mWT at week 2 was higher than that of those receiving non-pulsed steroid alone (Table 3 and Figure 2). All patients with pulse therapy but none of those without experienced improvement in the OMDS (p < 0.05). In addition, two patients in the non-pulse group required additional steroid therapy due to disease progression. Hence, oral prednisolone alone is insufficient among rapid progressors, and intravenous methylprednisolone is more effective for the rapid improvement and maintenance of motor function.
Observational studies reported that the motor function of most patients deteriorated in the long term despite corticosteroid treatment [6,25,26]. Thus, whether continuous corticosteroid therapy is beneficial for HAM has been debated. Therefore, treatment with prednisolone was discontinued at week 26 in rapid progressors. After discontinuation, five of six patients experienced deterioration and resumed prednisolone treatment, thereby indicating that the premature discontinuation of steroids could cause deterioration. Our findings support a recent retrospective study showing that continuous treatment was more effective in maintaining motor function than short-term therapy [27].
In slow progressors, we used 10 mWT as the primary outcome because it is a sensitive and objective functional measure with small intra-patient variability, and it can worsen by an average of 5.74% per 6 months in untreated patients with HAM [28]. Based on our clinical experience, there was an improvement of approximately 15% per 6 months with prednisolone (unpublished data), which is roughly equivalent to 2 years worth of gait deterioration. Because 1-grade OMDS deterioration can develop at an average of approximately 5 years, this improvement may be equivalent to 0.4 grade [6].
In this study, 10 mWT at week 24 significantly improved from baseline only in the prednisolone group, and the change was similar to our clinical experience. However, the results between the prednisolone and placebo groups did not significantly differ (Table 4 and Figure 4). Other mobility measures, including 2 MWD and 6 MWD, had similar tendencies. A statistically significant difference was not detected, presumably due to the small sample size. Nevertheless, prednisolone might be effective against HAM because the prednisolone group had better results in multiple mobility measures. Moreover, the results were consistent with significant findings in CSF markers, which are indicative of activation and inflammation [29].
This study provided important information for future clinical trials on HAM. The changes in 10 mWT, 2 MWD, and 6 MWD in the prednisolone group were consistent with the improvement in 10 mWT observed in earlier clinical observations. Therefore, 10 mWT is a simple and reliable outcome measure for HAM with a minimal burden on patients. Next, the 10 mWT of the placebo group improved, which was contrary to the results of clinical observations [28]. Thus, the placebo effect is nonnegligible in clinical trials on HAM. Since there was no data regarding the change in 10 mWT in clinical trial settings, we calculated the sample size of slow progressors using data obtained in clinical practice. If the sample size is calculated based on the current study, 100 participants (n = 50 in each group) are required to show the superiority over placebo of any therapy with similar efficacy to oral prednisolone. Recruiting such a large sample with rare diseases, including HAM, is not easy. Hence, it might be necessary to discuss whether applying general statistical thresholds to rare diseases is adequate. Moreover, developing reliable surrogate endpoints is essential to facilitate clinical trials.
CSF markers are advantageous because they can be measured objectively and do not exhibit any placebo effects. In the current research, we assessed CSF neopterin and CXCL10 concentrations. Their baseline concentrations were higher in rapid progressors than in slow progressors, well reflecting disease activity according to the classification criteria [30,31]. These markers significantly decreased in the prednisolone group, while there was minimal change in the placebo group (Figure 4). A retrospective study showed that changes in the concentrations of these markers are correlated with those in gait function in patients treated with corticosteroids [27]. Furthermore, in another retrospective study with a median follow-up of 4.1 years, patients whose CSF marker concentrations decreased during steroid therapy were at lower risk of deterioration [32]. Hence, this finding indicated a lower risk of deterioration in the future. Long-term maintenance of gait function is the gold standard endpoint for HAM. However, it is difficult to generate statistically significant outcomes if the sample size of clinical trials is limited. Nevertheless, CSF neopterin and CXCL10 can be a promising surrogate endpoint, with lower values indicating a better therapeutic efficacy.
In this study, the gait function and CSF marker concentrations of patients who received a placebo did not change. However, these parameters substantially improved after prednisolone therapy (Table S1 and Figure S6). These findings supported the aforementioned results regarding the treatment efficacy of prednisolone and the use of CSF markers.
Corticosteroid therapy was well tolerated, although steroid-related AEs were common. The overall incidence of serious AEs was similar among rapid and slow progressors between the treatment arms. Infectious disease is a common and, occasionally, serious complication of steroid therapy. Moreover, urinary tract infection is frequently observed in HAM due to neurogenic bladder issues. In this study, some patients presented with herpes zoster and urinary tract infections. Individuals infected with HTLV-1, including those with HAM, are at risk of developing ATL [33]. Those who have clonal expansion of HTLV-1-infected lymphocytes with ATL-related somatic mutations are at significantly high risk [34][35][36][37]. In this study, one slow progressor taking prednisolone developed ATL, and we retrospectively confirmed that this patient had such clones upon trial enrollment. Therefore, clonality examination should be included in eligibility screening in future trials. Whether immunosuppressive therapy can be recommended to high-risk patients with HAM and whether it affects ATL development should be investigated in future studies. When monitoring patients with HAM, special attention should be paid to infectious diseases and ATL.
The current study had several limitations. First and foremost, the sample size was extremely small. Thus, randomization might not have been significantly effective, and the baseline 10 mWT differed among slow progressors between the treatment groups. Rapid progressors did not receive placebo and were aware of their treatment, which might have affected the symptoms; moreover, some of them received additional treatment for deterioration after week 4. Meanwhile, all slow progressors received oral prednisolone after week 24. Therefore, long-term treatment efficacy was not completely validated. Lastly, all participants were Japanese, and most of them were women. Hence, the generalizability of the results might be limited.
In conclusion, to our knowledge, this is the first RCT on corticosteroid therapy for HAM. Although the primary endpoints did not significantly differ due to the small sample size, this study indicated that corticosteroid therapy was safe and beneficial for patients with HAM. Therefore, larger trials must be conducted to confirm treatment efficacy.  (Table S1: Changes in motor function and CSF marker concentrations in the placebo group during the placebo and prednisolone periods, Table S2: Summary of exploratory assessment in slow progressors, Table S3: Changes in OMDS and the Modified Ashworth Scale score at 24 and 48 weeks from baseline in slow progressors), Supplementary Figures ( Figure S1: Trial overview, Figure S2: Treatment flow chart of rapid progressors, Figure S3: Treatment flow chart of slow progressors, Figure S4: Changes in motor function and CSF marker concentrations in rapid progressors, Figure S5: Exploratory assessment in rapid progressors, Figure S6: Changes in motor function and CSF marker concentrations in slow progressors, Figure S7: Exploratory assessment in slow progressors), and Supplementary Files (File S1: Trial protocol, File S2: Statistical analysis plan for rapid progressors, File S3: Statistical analysis plan for rapid progressors, Figure S4: CONSORT Checklist).