Nusinersen treatment in adult patients with spinal muscular atrophy: a safety analysis of laboratory parameters

The condition, 5q-associated spinal muscular atrophy (5q-SMA) is a rare genetic neuromuscular disorder that leads to progressive muscle atrophy and weakness because of degeneration of anterior horn cells [28]. The “classic” clinical classification of the disease is based on the achievement of motor milestones during childhood, with a broad range of phenotypes from severely affected infants never achieving the ability to sit (SMA type 1), to individuals able to sit but never achieving the ability to walk (SMA type 2), to ambulatory patients (SMA type 3) [31].

The disorder 5q-SMA is caused by a homozygous deletion or mutation in the survival of the motor neuron 1 gene (SMN1) localized on chromosome 5q, which results in insufficient levels of SMN proteins[24, 25]. Nusinersen is an antisense oligonucleotide (ASO) that is capable of increasing SMN protein production.

In clinical trials, intrathecal treatment with nusinersen led to meaningful benefits on motor functions in infants and children[14, 29]. Based on the results of these studies, nusinersen was approved by the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA). Moreover, both the feasibility and efficacy of nusinersen therapy have also been well established in adult SMA patients [18, 26, 37, 39, 42, 43].

In addition to feasibility and efficacy, the safety of a therapy is the decisive factor for the success of the treatment. In the past, various organ toxicities have been described with the use of some ASOs, such as an increase in liver transaminases, renal impairment, coagulation abnormalities, or thrombocytopenia [3, 7, 15, 20]. Since ASOs are not able to cross the blood–brain barrier (BBB), nusinersen is administered intrathecally [16]. In terms of potential side effects, this factor could make a difference compared to other ASOs.

To date, systematic data on laboratory parameters under treatment with nusinersen are still scarce and are mainly available from pivotal studies in infants and children [11, 38]. Here, we present our data on laboratory findings in adult SMA patients under treatment with nusinersen. These data are valuable to better assess the safety of longer-term nusinersen therapy in this patient cohort.

Laboratory parameters can help to identify side effects and organ toxicities of new therapies and thus to provide information on therapy safety. In the past, a variety of side effects have been identified for different ASOs, such as platelet declines, renal impairment, increases in liver enzymes or prolongation of aPTT [3, 7, 15, 20]. In particular, thrombocytopenia, including acute severe thrombocytopenia, and nephrotoxicity are major concerns and are often considered class effects. However, these side effects seem to occur only with the use of some ASOs [8‐10, 15]. Since data on laboratory findings in adult SMA patients under therapy with nusinersen are lacking, the present study helps to assess the safety of the therapy better.

Analysis of CSF parameters can detect various pathologies, such as infectious or noninfectious CNS inflammation, intrathecal hemorrhage, changes in CSF flow, or BBB dysfunction. Changes in CSF parameters under therapy with nusinersen are of particular interest since the drug has to be administered by repeated lumbar punctures. In our cohort, the mean baseline values of the analyzed CSF parameters were all within the normal range. Three patients showed mild pleocytosis at baseline; however, an association with SMA is unlikely. In addition, the mean CSF-WBC count tended to increase during the first three injections (L1–L3). Furthermore, approximately half of the CSF samples with pleocytosis were obtained at L2 or L3. In line with our results, Müschen et al. and Wurster et al. found a transient increase in mean CSF-WBC counts during loading dosing. Also in these studies, all increases in CSF-WBC counts (four increases in three of 88 SMA patients) occurred at L2 or L3 [32, 41]. These observations are most likely a reaction to the rapid succession of lumbar punctures at the beginning of the therapy. Under therapy, we found mild sporadic or transient elevations in CSF-WBC counts in approximately one-fifth of the patients, whereas no persistent increase was seen. In contrast, Müschen et al., Wurster et al. and Kessler et al. observed only three cases of pleocytosis in 28, 60 and 10 SMA patients treated for a period of 2 to 18 months [21, 32, 41].

Total CSF protein, an indicator of BBB and CSF flow integrity, was elevated in eleven SMA type 3 patients but only in one SMA type 2 patient. However, the widely used upper limit of normal (ULN) of 45 mg/dL is now being challenged [5, 6]. Although significance was not reached in our cohort, the trend of higher CSF protein concentrations in SMA type 3 patients is consistent with the findings of Müschen et al. and Wurster et al. [32, 41]. Since total protein increases with age, the difference in patient age between the SMA type 2 and type 3 subgroups could be an explanation although it was only 5.5 years in our cohort [27]. During nusinersen therapy, two patients in our cohort developed a sustained elevation of total protein. Furthermore, mean protein levels increased significantly in the total cohort as well as in the SMA type 3 subgroup. These findings are compatible with results from other studies on nusinersen in adult patients and may indicate an alteration of CSF flow or BBB dysfunction. Whether this observation is a consequence of repeated lumbar punctures or a side effect of the drug remains unclear [21, 32, 39, 41]. This outcome may possibly be related to reports of rare cases of hydrocephalus of unknown etiology.

Recently, some cases of severe thrombocytopenia with platelet counts < 25/nL have been reported in clinical trials on ASOs. In the NEURO-TTR trial, three cases (3%) of severe thrombocytopenia, including one in association with a fatal intracranial hemorrhage, were observed during therapy with inotersen, an ASO designed to inhibit hepatic production of transthyretin in patients with hereditary transthyretin amyloidosis. In the same study, a decrease in platelet count to less than 140/nL was seen in 54% of the participants in the inotersen group vs. 13% in the placebo group [4]. In another study, severe thrombocytopenia was found in 2 of 33 patients with familial chylomicronemia treated with volanesorsen [40]. Based on experience with other ASOs, thrombocytopenia, including acute severe thrombocytopenia, is part of the warnings in the prescribing information of nusinersen. Data on platelet counts under therapy with nusinersen are mainly available from studies in infants or children. In the ENDEAR and CHERISH studies, the incidence of shifts to low platelet counts was 13 vs. 0% (nusinersen vs. placebo) and 20 vs. 26%, respectively. Platelet counts < 20/µL were seen in one infant in the ENDEAR study and in two children in the CHERISH study. However, no sustained shifts or bleeding complications occurred. Median platelet counts remained stable during therapy [11]. Another study of nusinersen therapy in presymptomatic infants also showed stable platelet counts [12]. In a cohort of 58 SMA patients aged between 1 month and 56 years, Goedeker et al. described three cases of mild thrombocytopenia, with none of them reaching a platelet count < 100/nL[17]. Szabo et al. reported singular mild and transient thrombocytopenia in a single child treated with nusinersen [38]. Remarkably, only a single case of sporadic thrombocytopenia with a platelet count of 101/nL occurred in our study. For the graduation of abnormal laboratory values, National Cancer Institute—Common Terminology Criteria of Adverse Events (NCI-CTCAE version 5.0) is often used in clinical studies, which distinguishes between four severity grades (grade 1, mild, to grade 4, severe) [1]. According to this reference, the observed thrombocytopenia would meet the criterion of grade 1 thrombocytopenia, defined as a platelet count between the lower limit of normal (LLN) and 75/nL. Similar to the data from pediatric studies, platelet levels remained stable over time. Based on our results, there is currently no evidence of an increased risk of severe thrombocytopenia in adult patients treated with nusinersen.

We observed a reduction in the mean aPTT between the first and last injections in the total cohort and the SMA type 2 subgroup. Shortened aPTTs are often considered laboratory artifacts, which are due to difficult blood sampling. In addition, there is some evidence that short aPTTs represent a hypercoagulable state and may be associated with an increased risk for thromboembolic and cardiovascular events [23, 30]. With regard to the coagulation parameters, we did not observe any persistent parameter deviations from the normal range, which is in line with the results of Goedeker et al. [17].

Nephrotoxicity is a major concern with the use of ASOs [3, 15]. A significantly increased risk of renal impairment was observed in patients with Duchenne muscular dystrophy treated with drisapersen [36]. In the abovementioned study on inotersen, three cases of glomerulonephritis and one case of tubulointerstitial nephritis were observed in 112 subjects with hereditary transthyretin amyloidosis [4]. In pivotal studies on nusinersen, shifts to high urea nitrogen or creatinine were observed only in individual cases. Additionally, proteinuria did not occur more frequently in nusinersen-treated subjects than in control subjects, so the studies did not provide evidence of renal impairment [11, 14, 29]. In the study of Goedeker et al. proteinuria was the most common laboratory abnormality. However, it was never persistent and often attributed to sample contamination [17]. A safety evaluation of 18 adult patients with SMA type 3 and type 4 treated with nusinersen over a 14-month period revealed no evidence of renal dysfunction during therapy [13]. In our cohort, mean creatinine levels at baseline were decreased in both the SMA type 2 and SMA type 3 subgroups, with SMA type 2 patients having lower values than SMA type 3 patients (0.31 vs. 0.49 mg/dL). These observations are not surprising since creatinine levels correlate with skeletal muscle mass and are consistent with a study conducted in 238 patients with SMA types 1–3. In line with our results, the authors found 1.7-fold higher creatinine levels in SMA type 3 than in SMA type 2. Furthermore, serum creatinine correlated with SMA type, SMN2 gene copy number, motor function, and severity of denervation. The authors conclude that creatinine is a candidate biomarker for SMA progression and suggest investigating whether creatinine levels respond to new SMA therapies [2, 22]. Remarkably, only nine SMA type 3 patients from our cohort had normal creatinine values at baseline, and six of them dropped below the LLN during the course of therapy, with one decrease classified as persistent. Furthermore, mean creatinine levels decreased in the SMA type 3 subgroup and the total cohort, possibly indicating further disease progression under therapy. With regard to nephrotoxicity, we found no case of a relevant increase in either serum creatinine or urea nitrogen. However, the use of serum creatinine as a marker of renal function is strongly limited in SMA patients for the reason already mentioned above. Therefore, determination of cystatin C protein might be a more appropriate method for monitoring renal function in patients with SMA [13, 22].

Hepatotoxicity or an increase in liver enzymes are other potential side effects with the use of ASOs [15]. In clinical trials, mipomersen, an ASO used to treat dyslipidemia, resulted in an increased incidence of elevated ALT above 3 × ULN and hepatic steatosis compared to placebo [33]. With regard to nusinersen, the NURTURE trial showed stable transaminases in presymptomatic children [12]. In ENDEAR and CHERISH studies, there were some shifts in ALT and AST, with no clear difference from the control groups or reports of liver failures [11, 14, 29]. Similar to the results from pediatric studies, only a few and mostly temporary elevations in liver enzymes were observed in our cohort. However, one patient with SMA type 2 with a normal baseline value developed a discrete but persistent increase in AST, while a single SMA type 2 patient showed a successive increase in GGT. Remarkably, both cases would only fulfill grade 1 according to the NCI-CTCAE, defined as enzyme levels up to 3 × ULN [1]. Regardless of the baseline value, no patient in our study developed elevations in liver enzymes above threefold the respective baseline value, and no patient with elevated liver enzymes showed increased bilirubin levels (data not shown). In conclusion, there was no evidence of a clinically relevant increase under treatment with nusinersen.

In terms of detecting potential toxicities of nusinersen therapy in adult patients, this study has some limitations. Importantly, despite the relatively long observation period, rare side effects may not have been captured because of the limited cohort size. In outpatients, coagulation parameters were not usually assessed during therapy. Thus, a large proportion of samples and results were from more severely affected patients, which may result in a bias. The other blood samples from outpatients were usually acquired before the first loading dose and each maintenance dose. Therefore, transient changes during loading dosing in these patients may have been missed. Finally, our definition of the duration of deviations from the normal range is somewhat arbitrary. In particular, "persistent" changes may have returned to normal values after the end of the observation period, and conversely, "transient" shifts may have persisted.

Our data demonstrate a favorable safety profile of nusinersen therapy in adult SMA patients under longer-term “real-world” conditions, regardless of SMA type. In particular, we found no evidence of clinically relevant platelet declines, coagulopathies, or renal or hepatic organ toxicities, which are common concerns with the use of ASOs. Whether this finding is solely attributable to the intrathecal administration route of nusinersen remains speculative.