The choice of the ulnar nerve for our study was dictated by the very large anatomical changes in the lower limbs with irreducible contractures in some SMA patients, as well as the presence of clinical features of carpal tunnel syndrome in other patients in the study group. The distal upper limbs are usually less affected than the lower limbs due to muscle atrophy and osteoarticular deformities in SMA 5q.

The most repeatable result of our study were the changes in the amplitude values of the sensory potentials. The amplitudes in the analyzed individual patient groups differed between those groups and in relation to the controls. The SNAP amplitudes were significantly higher in the study group compared to the control group, higher in the more severe forms of SMA, i.e., in type 2 compared to 3a and 3b, respectively. Amplitudes were also significantly higher in the group of patients with a more severe degree of disability as expressed by the HFSME scale. Maintaining an appropriate skin temperature, i.e. always above 32 degrees, seems to exclude the influence of this factor on the amplitude value. Assessing the impact of overweight and underweight based on BMI in SMA patients is difficult due to the heterogeneity of this group of patients; among others, some patients show marked overweight coexisting with muscular atrophy.

Large SNAP amplitudes have been previously described in spinal cord disease. Pullman et al. [22] described the occurrence of high SNAP amplitude in patients with myelopathy without identifying the possible cause. High SNAP amplitude seems to be proportional to the number of active sodium channels per unit of membrane area. In the case of spinal cord damage, axonal transport can be redirected to healthy axons of cells. This phenomenon may result in an increased density of sodium channels and other transported substances in the peripheral, healthy parts of neurons. It is probable that metabolic overactivity is an adaptive mechanism in response to damage [22, 23].

Tripton et al. [23] indicated that higher SNAP amplitudes reflect an increased number of larger diameter sensory axons with lower depolarization thresholds. Abnormalities in the conduction capacity in axons or receptors are also possible. Patients with non-specific sensory symptoms may have peripheral sensory nerve hyperactivity syndrome, correlating with SNAP high amplitude in the electrophysiological study [24]. Our patients did not complain about obvious sensory disturbances or important pain. In the electrophysiological study we found only one young man with SMA type 3b who had no sensory response.

Our results are generally contrary to the results of other authors, e.g., Sultan et al. [25] who indicated the loss of the SNAP amplitude in patients with SMA types 1 and 2. Similar data were achieved in patients with SMA type 1 by Duman et al. [11]. In their study, 26.7% patients had decreased SNAP amplitude or sensory nerve conduction velocities; in five patients SNAP could not be found. In the first study, statistically lower amplitudes were seen in the median nerve in patients with SMA type 1, while the amplitude did not differ significantly in patients with type 2. In a study by Duman et al. [11] two out of 15 patients with SMA type 1 showed a lack of sensory responses in the median and sural nerves. In the remaining patients, the amplitudes of sensory responses in the median nerve were within normal limits. Most published studies [11, 25,26,27] on this topic mainly indicate damage to the sural nerve, and therefore this is the one more likely to suffer secondary, additional damage. Pro et al. [26] stressed that SNAP amplitudes in the sural and median nerves are normal in younger patients with SMA type 1, while an axonal neuropathy appears only in older ones.

Our findings suggest sensory fiber overactivity, which seems to confirm previous studies. The predominance of lesions in patients with more advanced disease may reflect a greater intensity of metabolic overactivity as an adaptive mechanism in response to damage. However, a direct comparison of the results obtained between acute forms of SMA in children and chronic forms in adults does not seem valid.

We cannot overlook possible influences of the central nervous system, e.g., the thalamus or cerebellum, which are increasingly considered important in sensory processes, and probably involved in the pathological mechanism in SMA [8,9,10, 28]. In this context, the analysis of the results obtained in the QST study seems challenging. A QST study is based on the determination of temperature and temperature-induced pain thresholds, which allows assessment of the function of smaller myelinated and unmyelinated sensory fibers – Aδ and C [19, 29]. Thimm et al. [30] revealed significant subclinical small nerve fiber damage in the cornea in patients with SMA type 3, which correlated with motor function. In our QST study, the most significant differences were related to cold pain threshold. SMA patients, especially with milder forms, were significantly less sensitive to pain caused by cold. The threshold for low temperature was comparable in the study groups. The CP-CS difference was greater in SMA patients with less advanced disease but did not reach statistical significance, in contrast to the spread of values for high temperature. HP-WS was significantly greater in the SMA group as a whole, between the individual groups of SMA, and in SMA type 3b compared to the control group. The correlations suggest that in the early stages of the disease, the small fibers responsible for the conduction of temperature and temperature-depended pain show a moderate degree of damage, higher tolerance to more extreme temperatures (low), but with correct sensation thresholds for normal temperatures. CP thresholds appear to be more individualized, varying by body area, depending on complex psychophysical processes. Lötsch et al. [31] demonstrated the hypothesis that CP thresholds reflect the contribution of the two different cold sensors. Fewer heat receptors, greater spatial summation, and more diffuse sensation of heat may underlie the different sensations of pain caused by heat and cold [31,32,33].

Pitarch-Castellano et al. [34], Sagerer et al. [35], and Uchio et al. [36] described the occurrence of pain in SMA patients. They found a prevalence of pain in 27% to approximately 40% of patients. The pain was of moderate to low intensity, chronic in nature, and predominantly in the lower limbs. The condition of our patients with regard to pain perception was similar to the results mentioned above. Again, it seems that higher pain sensory thresholds in SMA patients may be an adaptive mechanism, active especially in more benign SMA forms. It cannot be ruled out that a deficiency of the systemic SMN protein may cause slight damage to the C and A-delta small fibers early in the course of the disease, which may be evidenced by an increase in pain thresholds.

Vibratory limits were comparable in the patient and control groups. They were significantly higher in the patients with HFSME scores above 10 points. They tended to be higher in SMA type 3b, but this was without statistical significance. These results indicate damage to large A-beta fibers in the prolonged milder form of SMA. Gregory et al. [37] indicated that in amyotrophic lateral sclerosis, vibration thresholds were elevated in comparison to controls. The authors found no reports on vibration sensation in SMA, its possible changes, and significance in the overall view of this disease entity.

In our SMA patients, we were able to show some the features of sensory nerve demyelination. They had significantly lower sensory conduction values with prolonged latencies. These parameters were worse in milder forms of SMA, mainly in type 3b (Fig. 1B) There were no significant differences in sensory conduction velocity or prolongation of latency when comparing patients with more and less disability on the HFSME scale. Therefore, it appears that the demyelination exponents shown in our patients are of relatively low severity, most likely secondary in nature, and this may be dependent on the duration of the disease.

These two findings—early damage to large A-beta fibers, and the secondary demyelination process of sensory fibers—are in agreement with our previous research results on motor fibers based on the conduction velocity distribution study. We found the presence of a coexisting, demyelinating process in the early stage of the disease [14], which was in line with previous reports, e.g., Duman et al. [11].

As a limitation of this study, we must point to the limited number of patients and the fact that the neurophysiological study was restricted to one nerve, as was explained at the beginning of the Discussion section. We must also note the difference in mean age between SMA patients and controls. However, in both groups the mean age range was in the fourth/fifth decade of life, and according to the literature, the important trend for an increase in the results of sensory latency and a decrease in the sensory conduction velocity is observed at ages ≥ 46 years [38]. We also included SMA patients with different comorbidities in the study, and this can potentially promote nerve damage. Among them were underweight and overweight patients, BMI showed differences between the study and control groups and between SMA type 2 and 3 patients. These patients were not excluded due to the low severity of the coexisting diseases, and the lack of clinical and electrophysiological signs of polyneuropathy.