Periodic paralysis (PP) is a rare skeletal muscle channelopathy induced by abnormal excitability of the sarcolemma, leading to episodes of flaccid paralysis in the extremities of patients. Symptoms commonly appear in the first or second decade, usually upon awakening in the middle of the night or early morning, and last for hours (occasionally days) before gradually disappearing. The calculated minimum point prevalence rates of PP have been reported at 0.38–0.69/100,000 in the UK and the Netherlands (1, 2).

Clinically, patients with PP with decreased serum potassium level (<3.5 mmol/L) during the paralytic attacks are subtyped as hypokalemic periodic paralysis (hypoPP), genetically linked to the mutations in CACNA1S (encoding α1-subunit of the skeletal muscle L-type calcium channel Cav1.1; hypoPP1) or SCN4A (encoding α1-subunit of voltage-gated sodium channel Nav1.4; hypoPP2) (3, 4). However, naming of the subtype of patients with PP with normal-range or high serum potassium levels is controversial, and in the present study, we refer to multiple recent publications and group these patients with serum potassium level ≥3.5 mmol/L as hyperkalemic periodic paralysis (hyperPP) (1, 2, 5, 6). SCN4A is the causative gene of hyperPP as well (7), and it is also responsible for SCN4A-related non-dystrophic myotonia, characterized by a heterogeneous phenotypic spectrum of myotonia (8, 9). Furthermore, mutations in the KCNJ2 gene (encoding inward-rectifier potassium channel Kir2.1), which have been linked to Andersen-Tawil syndrome (ATS), could also result in a PP phenotype, although typically accompanied by ventricular arrhythmias and dysmorphism (10).

To date, large-group genetic studies concerning both hypoPP and hyperPP are inadequate, and only a few studies have covered all three abovementioned genes, CACNA1S, SCN4A, and KCNJ2. The genetic diagnostic rate of overall patients with PP remains unclear, which is estimated to be 64.1% in the USA or 56.6% in China (11, 12). In this study, among 119 index patients with PP, referring to broad diagnostic criteria, we present a low molecular diagnostic rate in Japan and reassess multiple clinical features associated with the diagnostic yield.

In this study, based on our relaxed enrollment criteria, we collected 119 unrelated index patients with clinically suspected PP, and the genetic diagnosis was only procured from 28.6% of them. However, even using the more stringent criteria involving recurrent paralytic attacks, the diagnostic rate was 33.0% (31/94). Both rates were significantly lower than previous studies conducted in the USA and China (11, 12). In terms of hypoPP, the diagnostic rate was 26.2%, which was also much lower than in several western countries (64.3–89.2%) (12, 18–20), but higher than in a Taiwan study (12.5%) (21) (Table 2). In contrast, 18 out of 25 (72.0%) patients with FPP received a molecular diagnosis, showing a much higher yield than patients with SPP (17.0%; p < 0.0001). Otherwise, patients with early onset (≤20 years) were found to be more amenable to molecular diagnosis than that of late-onset cases (24/57 vs. 9/59).

Taken together, the aforementioned differences between present and previous studies may be contributed by but not limited to a high proportion of late-onset PP (59/116) and SPP (94/119) in our case series. This is comparable to the Taiwan study with by far the lowest diagnostic rate, where 93.8% (60/64) of their cases were SPP (21). Furthermore, we also noted that the diagnostic rates of PP were concurrently lower in Asia than that of studies in Europe or the USA, and thus, a racial difference should be taken into account as well.

When compared to multiple large-group studies that covered all CACNA1S, SCN4A, and KCNJ2 genes, the genetic proportion of our diagnosed patients had the following characteristics: (1) hypoPP1 was more common than hypoPP2 (9:5), which was comparable to a recent Japanese study (4:3) (5), but not as noticeable as the difference observed in western countries; (2) hyperPP (41.2%) was more common than either hypoPP phenotype alone (Figure 3D). Our case series included 108 men and 11 women, for a gender ratio of about 10:1. This gender disparity could be explained by females' lower penetrance, which is consistent with previous findings (18, 22, 23). Though without substantial variation, female patients (45.5%) were more likely than male patients (26.9%) to receive a molecular diagnosis.

All five CACNA1S gene variants were found at arginines of S4 voltage sensors in domains II (p.R528H/G), III (p.R900S), and IV (p.R1239H, p.R1242S) of Cav1.1. Functional assessments have been conducted for all these variants except p.R1242S. Reduced amplitude of inward Ca²⁺ currents was observed from all of the four variants, and an abnormal gating pore leak current was detected from p.R528H/G and p.R1239H (24–26). These changes would result in susceptibility to recurrent episodes of depolarization-induced loss of excitability and weakness in HypoPP (27).

Unlike CACNA1S, SCN4A variants associated with both hyperPP and hypoPP2 were found throughout the protein Nav1.4. It is of note that multiple SCN4A variants, previously reported from patients with hypoPP (p.R222W and p.R672H) or non-dystrophic myotonia (p.E452K) (19, 28, 29), developed hyperPP phenotype in our patients. These findings, together with the p.T704M variant, which is associated with both hyperPP and hypoPP2 phenotypes in the present study, highlight the phenotypic heterogeneity of sodium channelopathies. Mutations associated with hyperPP produce the gain-of-function changes for Nav1.4, commonly exhibiting defects of fast and/or slow inactivation, and occasionally showing an enhancement of activation (27). In contrast, multiple mechanisms have been elucidated from hypoPP2 mutations, consisting of loss-of-function changes of Nav1.4, such as enhanced inactivation and decoupling of voltage-sensor displacement to channel opening (29, 30), as well as the gating pore “leakage” current (31). Reduced Nav1.4 currents may also contribute to the reduced excitability of the muscle membrane, leading to paralysis. Among novel SCN4A variants (p.A37T, p.F264S, p.R1451H, and p.T1646N), p.A37T and p.T1646N locate at cytoplasmic N or C terminus of Nav1.4, the domains where multiple variants have been reported, and a p.F1705I variant was found causing fast inactivation defects (32).

Despite PP, ventricular arrhythmias, and dysmorphism being identified as the triad of ATS, patients frequently lack one or more features of the classic triad. As demonstrated in our patients, among all six patients carrying KCNJ2 variants, none of them showed any noticeable dysmorphic features or electrocardiographic abnormalities. This PP-only phenotype complicates the clinical diagnosis of ATS and emphasizes the importance of genetic screening for the KCNJ2 gene in patients with isolated PP. Our outcomes also indicate that the frequency of ATS with PP-only phenotype may be underestimated, referring to a previous report in Japan (2/57) (33). Mutations of KCNJ2 locate throughout the Kir2.1 protein, and cellular analyses revealed the loss-of-function and mostly with a dominant-negative effect on lowering the inward rectifier current, which subsequently depolarizes resting membrane potential and leads to paralysis (10, 34).

On the other hand, SPP is predominant in our case series (94:25), and the etiology requires further research. Recently, within a molecularly undiagnosed Japanese SPP cohort, disease susceptibility was confirmed for nine single-nucleotide variants (SNVs), discovered in genome-wide association studies from SPP and/or thyrotoxic PP in Asian populations (35–40). All of these SNVs are found on chromosome 17 downstream of the KCNJ2 gene, with strong linkage disequilibrium, implying a genetic basis for the undiagnosed SPP.

In this study, we adopted a relatively broad inclusion criterion and obtained a low diagnostic rate (28.6%) from a case series of patients with PP. As indicated in our subsequent statistical analyses, the diagnostic yield could be improved using a more stringent enrollment criterion, such as positive family history, early-onset, and recurrent paralysis. However, since disease-associated variants were also identified from sporadic, atypical, or first-onset cases, we decided to involve all of them in this study. We could not exclude the possibility that the part of our patients was actually not PP, particularly those patients with hyperPP. Another limitation of this study is that the pathogenicity of the VUS variants has not been functionally verified, whereas the possible existence of benign variants would make the diagnostic rate even lower.

In summary, we evaluate the low molecular diagnostic rate and specific genetic proportion of a large Japanese case series of patients suspected of PP. Our outcomes outline the racial diversity and etiological complexity of patients with PP in Japan. Future research should attempt to explore other possible causes of undiagnosed PP, the pathogenicity of detected variants in known PP disease-causing genes, particularly VUS variants, and the pathogenesis of SPP-associated SNVs.

The datasets presented in this article are not readily available because of ethical and privacy restrictions. Requests to access the datasets should be directed to the corresponding author.

The studies involving human participants were reviewed and approved by Institutional Review Board of Kagoshima University (Application ID: 490). Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

HT conceptualized the study. J-HY and AY conducted the genetic experiments and analyzed the data. J-HY, YujH, AH, MA, TN, YuH, and YS participated in the clinical data acquisition and analysis. J-HY drafted the original manuscript. All authors revised the manuscript and approved the final version.