Two hundred fifty-nine patients were identified, and 256 included in the study (Table 1). Three patients were omitted. Two of them had scarce medical records, and one did not complete the electrophysiological exercise testing. Mean age at the time of exercise tests was 43.2 years (SD 12.8; range 11–78). 122 were male (47.7%) and 134 female (52.3%). There were more female (74.1%) than male in the case group. The mean age of the case group (36.9 years) was lower than the overall mean age of the study patients (43.3 years). The most common symptoms were muscle weakness, myalgia, cramps, and muscle stiffness.

The most common clinical ICD-10 diagnoses were M79.1 Myalgia (n = 41), R29.8 Other and unspecific symptoms and signs involving the nervous and musculoskeletal system (n = 38), R25.2 Tendency for muscle cramping (n = 26), G72.9 Myopathy, unspecified (n = 24), and G72.8 Suspected metabolic myopathy (n = 22). These accounted for 59.0% of the diagnoses (Table 2). Most common after these were G71.18 Myotonia congenita (n = 10), G71.18 Paramyotonia congenita (n = 9), and G71.18 Suspected skeletal muscle channelopathy (n = 9). All patients of the case group were diagnosed with channelopathy, myotonic dystrophy type 1 or myotonic dystrophy type 2 (Tables 2 and 3). Diagnosis of suspected metabolic myopathy was based on symptoms and extensive investigations including electromyography, muscle biopsy, and magnetic resonance spectroscopy (MRS) examination. Sixteen subjects also underwent non-ischaemic forearm exercise test and cycle ergometry. Pathogenic variants could be identified in one subject diagnosed with metabolic myopathy. He was homozygous for the POLG variant c.803G > C (p.G268A).

Sixteen patients (59.3%) of the case group had an abnormal test result in SET, LET, or SET with cold provocation compatible with a pattern described by Fournier. In nine cases (33.3%) more than just one of these exercise tests was abnormal. EMG myotonia was found in 21 (77.8%) case subjects. Five patients (18.5%) had no abnormal exercise test results consistent with Fournier pattern nor EMG myotonia. These patients had either a heterozygous SCN4A variant c.3466G > A (p.A1156T) or heterozygous c.4379G > A (p.R1460Q), or were diagnosed with DM2 (Table 4).

In contrast, only two patients of the control group had a clearly abnormal exercise test result compatible with Fournier pattern. The first one was diagnosed with metabolic myopathy. He was homozygous for the POLG variant c.803G > C (p.G268A). In the long exercise test CMAP amplitude had an immediate + 69% increase which slowly decreased being − 13% at the end of the test. This meant that CMAP amplitude decrease from the peak value was -48.5%. This was compatible with Fournier pattern IV. The other one was a child diagnosed with unspecific myalgia. The patient had episodic fewer with elevated infection parameters and myalgia during the episodes. Extensive investigations including MYOcap did not reveal any specific diagnosis. Short exercise with cold provocation induced decrement which persisted and was amplified with exercise repetition, being up to − 52% after the third exercise.

Exercise test results were slightly different from normal values described by Fournier [22, 23] in 33 cases out of 256 (12.9%). Four of these were from the case group, four were suspected to have a skeletal muscle channelopathy, and 25 were from the control group. These results did not fit to patterns described by Fournier [22, 23], and the significance of these findings remained unclear. The age and sex distributions were similar to those of the whole study population. The symptoms, clinical phenotype, and acquired diagnoses were diverse. No common factor was evident among these patients. The most common slightly abnormal finding was observed in the long exercise test. Twenty-one subjects demonstrated immediate increase in CMAP amplitude after the long exercise (mean + 34.8%, SD 10.6%, range from + 17 to + 54%). The amplitude slowly returned to baseline, except in two cases. In the first one the amplitude decreased to − 26%, and in the second one increase persisted and was + 16% at the end of follow-up. The distinct late decrease described in Fournier pattern IV was absent in all these tests. The second most common finding was immediate decrease (mean − 47.8%, SD − 12.6%, range from − 32 to − 63%) in the long exercise test, found in five exercise tests. The CMAP amplitude slowly recovered, except in one test where decrement virtually disappeared by the second recording of CMAP amplitude. Other unspecific findings were even less frequent (Table 5). Slightly abnormal findings were found also in the short exercise tests in room temperature and with cold provocation.

Electrophysiological exercise testing, comprising SET, LET, SET with cold provocation, repetitive nerve stimulation, and EMG, added further evidence to confirm skeletal muscle channelopathy in five cases. One patient diagnosed with paramyotonia congenita had a novel heterozygous SCN4A c.2143G > A (p.A715T) variant. SET with cold provocation was compatible with Fournier pattern II, with CMAP amplitude decreases up to − 40%. This together with segregation data supported the diagnosis. Another patient diagnosed with myotonia congenita was compound heterozygous for the CLCN1 variants c.2680C > T (p.R894X) and c.264G > A (p.V88V). SET in room temperature was compatible with Fournier pattern II with post-exercise decreases up to − 47% which disappeared with exercise repetition. SET with cold provocation was compatible with pattern II, decreases was up to − 64%. The latter variant was later confirmed to be pathogenic [33]. It caused splicing defect to mRNA resulting in missing exon 2. One patient diagnosed with myasthenic congenital myopathy [12] had compound heterozygous SCN4A c.4379G > A (p.R1460Q) and c.3175C > T (p.R1059X), and a heterozygous CLCN1 c.2680C > T (p.R894X) variant. LET induced immediate and persisting CMAP amplitude decrease up to − 38%. Although this did not exceed 40% this was interpreted to suggest ion channel defect. This result is regarded as normal later in this paper when calculating sensitivity, specificity and odds ratio. In two subjects exercise test results were the only abnormal findings. Both were found in LET, which induced slight CMAP amplitude increase + 14% followed by late decrease which was − 54% compared to the pre-exercise value and − 48% from the peak value in the first subject. In the other subject, LET induced immediate and persisting increment of + 41% but without the late distinct decrease. These two patients were diagnosed with suspected skeletal muscle disorder.

Almost all study patients performed electrophysiological exercise tests once. Five patients performed them twice. In two such cases the results were conflicting. In the first one, short exercise test with cold provocation induced immediate decrement up to − 24% after exercises. The significance of the result was unclear. When repeated the year after, the amplitude changes in the same test ranged from − 8 to + 2% being normal and no specific diagnosis could be reached. In the other subject findings of the first tests were compatible with Fournier pattern II. Subject had up to − 84% decrease in CMAP amplitude which disappeared with exercise repetition in the short exercise test, and immediate − 45% post-exercise decrease in the long exercise test. When repeated after four years, however, test results were normal. The patient remained with an unspecific M79.1 Myalgia diagnosis.

According to our data, the sensitivity of electrophysiological exercise testing, when this includes SET, LET, and SET with cold provocation, was 59.3% and specificity 99.1% in discriminating skeletal muscle channelopathy (Table 6). Abnormal electrophysiological exercise tests compatible with Fournier patterns were associated with increased risk of skeletal muscle channelopathy or myotonic disorder (OR 164.3, 95% Cl 28.3–954.6, p < 0.001). The crude odds ratio, one not adjusted for age and sex, was almost the same (OR = 154.9). In all of these estimations exercise test was assessed to be abnormal if any of the short exercise test, long exercise test, or short exercise test with cold provocation was compatible with Fournier pattern. Normal test result refers to results not compatible with patterns described by Fournier. When examined separately, short exercise test had a sensitivity of 37.0% and specificity of 100%, and long exercise test had a sensitivity of 37.0% and specificity of 99.5%. Short exercise test with cold provocation alone had a sensitivity of 40.7% and specificity of 99.5% (Table 6). When combined, the aforementioned sensitivity of 59.3% and specificity of 99.1% was achieved.