COVID-19 predominantly affects the lungs but can also harm the nervous, cardiovascular, and renal systems, especially in hospitalized patients (1–3). Still, recent studies reported localized pain and numbness in some COVID-19 patients, suggesting the involvement of the peripheral nervous system (2, 4, 5).

Peripheral neuropathy (PNP) refers to conditions that involve damage to the peripheral nervous system, resulting in sensory symptoms (e.g., numbness, tingling), weakness, autonomic symptoms (e.g., early satiety, sweat abnormalities), and neuropathic pain (6). Electrodiagnostic studies are valuable for patients with suspected PNP; electrodiagnostic testing includes nerve conduction studies (NCS) that can determine the existence of PNP and the scope of lesions and evaluate damaged parts and pathological type (7). In the case of a positive diagnosis, early nerve nutrition drugs are given in combination with rehabilitation treatment to avoid more serious physical symptoms (8).

A review of 105 studies demonstrated that PNP is prevalent in patients with COVID-19, primarily due to immune mechanisms or neurotoxic side effects of medications; few patients developed PNP from prolonged bedding in the intensive care unit (ICU) and pre-existing diabetes (9). Because of the severe damage to multiple organs caused by the novel coronavirus, many patients continue to experience a range of physical and psychological symptoms beyond the acute phase of COVID-19 infection, including PNP-related symptoms like pain, limb numbness, and muscle weakness (10, 11). Previous studies also showed that 60% of the patients have persistent pain and numbness even after treatments, and these prolonged symptoms can lead to anxiety, depression, reduced productivity, and even suicide, increasing the economic burden on the family and healthcare system (12).

Despite the long-term impact of PNP, most studies on PNP in COVID-19 patients are based on case reports (13, 14). Furthermore, the risk factors of PNP in COVID-19 patients have not been explored. A better understanding of the risk factors can help healthcare professionals to identify high-risk individuals and intervene early. Therefore, this study investigated the risk factors of PNP among COVID-19 survivors at 6-month and 12-month follow-ups after diagnosis using NCS.

A total of 332 participants were enrolled and received at least one NCS: 161 completed the 6-month follow-up, and 243 completed the 12-month follow-up. Nineteen patients with unstable blood glucose levels during follow-up or incomplete medical records or tests were excluded. Finally, 313 participants [aged 49.0 (IQR 33.0–58.0) years, and 191 (61.0%) females] were included for analysis, 150 of whom completed the 6-month follow-up assessment, 230 completed the 12-month follow-up assessment (Figure 1).

During the one-year follow-up period, 232 (74%) patients developed PNP (PNP group), while 81 (26%) did not (non-PNP group). The demographics and baseline characteristics between PNP and non-PNP groups were similar (all p < 0.05), except for age [43.0 (IQR 28.0–53.0) vs. 51.0 (IQR 34.8–59.0) years, p = 0.002] (Table 1). Multivariable logistic regression analysis showed that age [odds ratio (OR) = 1.22, confidence interval (CI): 1.05–1.41, p = 0.009] was independently associated with PNP (Table 2; Figure 2). Additionally, among patients in the PNP group, 51 (16%) patients had mononeuropathy, while 181 (58%) had polyneuropathy (Table 3).

Next, multivariable logistic regression analysis showed that college or higher education (OR = 5.07, 95% CI: 1.80–13.90, p = 0.002) was independently associated with mononeuropathy, while no factors were associated with polyneuropathy.

F-wave abnormalities were detected in 22 (7%) out of 313 patients (Table 3), but no factors were independently associated with F-wave abnormality (Table 2; Figure 3).

The occurrence of polyneuropathy [37 (45%) vs. 144 (63%), p = 0.006] and tibial nerve injury [12 (15%) vs. 65 (28%), p = 0.020] were higher at 12-month follow-up compared to 6-month follow-up (Figure 2; Supplementary Table S1). Eighty-one (26%), 41 (13%), 12 (4%), and 3 (1.0%) patients had two, three, four, and five types of nerve injury that co-occurred, respectively (Table 4). According to pathological classification, the peroneal nerve type of PNP at 12-month follow-up decreased compared to 6-month follow-up [5 (6%) vs. 3 (1%), p = 0.020; Table 5]. In addition, memory loss, found in 269 patients (85.9%; 74 patients with PNP and 195 without PNP), was the most commonly reported symptom (Table 6).

In this study, we assessed the risk factors for PNP and F-wave abnormality during 1 year of COVID-19 rehabilitation in 313 participants, 232 (74%) of whom developed PNP. Our data suggested that age might be an independent risk factor for PNP, while higher education appears to be a risk factor for mononeuropathy. Also, the level of nCoV IgG at discharge might be an independent risk factor for F-wave abnormality during one-year rehabilitation from COVID-19. These findings might help clinicians formulate clinical rehabilitation directions for COVID-19 survivors.

The present study suggests that those with high education levels are at an increased risk of mononeuropathy. The literature suggests no direct, strong association between education level and the development of mononeuropathy (or polyneuropathy), but a lower education level may be associated with a higher prevalence of painful diabetic peripheral neuropathy (DPN) (21–24). Additional studies will be necessary to examine the issue. Nevertheless, we found that demyelination of the median nerve motor branch and the sensory branch is a common injury, suggesting that these patients are likely to have carpal tunnel syndrome (CTS). Causes of CTS include obesity, contraction of the wrist or repetitive flexion, wrist-related occupations, diabetes, and potential PNP leading to hypertrophy (25). We also compared the characteristics of PNP at two visits, finding that demyelination of the sensory and motor branches of the median nerve was aggravated, suggesting that COVID-19 may aggravate and lead to CTS.

The involvement of the central and peripheral nervous systems can have several pathogenic pathways, including endothelial dysfunction, direct viral lesions, thrombotic microangiopathy, hypoxia, systemic inflammation, and autoimmune reactions. In SARS-CoV-2 infection, neuroinflammatory changes of the brain and brainstem are the most common autopsy findings. Acute neurologic manifestations are more often present in the severe form of COVID-19 and are linked to poor prognosis (26).

Previous studies (20, 27, 28) assessed 1-year outcomes in hospital survivors with COVID-19, finding that most patients had at least one sequelae symptom 6 months after discharge, which is consistent with our data. However, no significant association between PNP and symptoms was observed in this study, highlighting the need to focus on peripheral nerves during follow-up of all COVID-19 survivors. Several non-respiratory manifestations were reported after COVID-19 (29, 30), and the occurrence of PNP accounts for 7.2–42.8% (31, 32). In the present study, the frequency of PNP was 74%, higher than the reported range. The higher frequency could be related to a selection bias since only patients hospitalized for COVID-19 were included, possibly shifting the disease severity distribution toward more severe disease. Different criteria for PNP among studies could also account for the differences in frequency.

Nevertheless, the specific mechanism of PNP after COVID-19 remains unclear. The current research focuses on three aspects. (1) Autoimmune response. Finsterer et al. (9) suggested that SARS-CoV-2 may have epitopes similar to peripheral nerve components (activating autoreactive B or T cells), resulting in PNP. SARS-CoV-2 spikes interact with GM1 gangliosides in peripheral nerves, which leads to cross-reactivity and the production of antibodies against antigens, inducing a peripheral demyelinating pattern of GBS (33). (2) Direct neuroinvasive effects of SARS-CoV-2. The nervous system damage of COVID-19 patients is commonly caused by acute nervous system infection (7, 34). Infection with SARS-CoV-2 increases the risk of neurological damage. In addition, SARS-CoV-2 can lead to exacerbate pre-existing baseline neurological disease severity or acute neurological damage in some patients. Another factor to consider is the possible direct neuro-muscular damage caused by the SARS-CoV-2 virus. However, direct muscle damage by the virus has not been demonstrated in autologous muscle lesions. (3) Persistent and recurring neuroinflammatory responses and damage. Immune dysregulation after infection with SARS-CoV-2 may persist in the form of persistent inflammation, immunosuppression, catabolic syndrome, and immunosuppression (35). The cytokine storm could also be involved (36). This immune state is triggered by cytokine storms during acute infection and is further promoted by the sustained release of endogenous alarmins or danger-associated molecular patterns, which can lead to chronic systemic inflammation. Chronic neuroinflammation associated with high levels of cytokines/chemokines may be involved in the pathogenesis and progression of neurological diseases (35).

In this study, the mixed type (axonal loss combined with demyelination) had the highest rate among the pathological types. The median nerve injury was the most common. In addition, sensory and motor branches were often damaged, with demyelination and distal (carpal-palm) involvement, followed by the peroneal nerve with mainly axonal loss in both proximal and distal ends. Compared to 6-month follow-up data, the sural and ulnar nerve amplitude increased at the 12-month follow-up, while the median nerve conduction velocity decreased in patients who completed two measurements. These results suggest that the demyelination of the median nerve was aggravated, and axonal loss of the sural and the ulnar nerves was relieved. Studies have found that COVID-19 patients are more at risk of having PNP in the distal limb and higher occurrence in the lower extremities, especially when undergoing long-term rehabilitation (37, 38). Thus, for COVID-19 patients, especially older adults with low IgG antibody levels, more attention should be given to the peripheral nerve condition. Early NCS screening and prophylactic administration of vitamin B6 and mecobalamin may be recommended for these patients (39, 40).

Electrodiagnostic studies are very beneficial when evaluating patients with suspected PNP (41, 42). In this study, NCS was performed using full-featured electromyography/evoked potential equipment during the follow-up. Most patients had PNP, which was consistent with previous studies. Oaklander et al. (43) evaluated the PNP of 17 patients with prolonged long COVID-19 and found that 59% of cases had ≥1 test interpretation confirming neuropathy. Moreover, another study (44) reviewed 143 studies reporting on nervous system involvement by COVID-19, detecting central and peripheral nervous system involvement in 10,723 COVID-19 patients. The study reflected the remarkable occurrence of neurological involvement in COVID-19 patients, ranging from 22.5 to 36.4%. However, previous studies on the nervous system have mainly focused on the central nervous system (45–47); also, most were case reports or focused mainly on severe COVID-19 patients (48). Finsterer et al. (9) analyzed 105 retrieved articles on SARS-CoV-2-related PNP, including 220 patients with and 41 patients without Guillain-Barre syndrome (GBS) neuropathy. A total of 168 patients reported latency between viral infection and the onset of neuropathy, ranging from −10 to 90 days. In addition, neuropathy was classified in 257 patients. Moreover, 220 patients were diagnosed with GBS, 11 with mononeuritis multiple neuropathy, 4 with severe neuropathy, 16 with plexus neuropathy, 4 with isolated sensory neuropathy, and 2 with leprosy. A review highlighted the association between GBS and COVID-19, suggesting that GBS could be caused by COVID-19 in these patients (49). COVID-19 appears to cause the senescence of dopaminergic neurons (50). However, electrophysiological testing was not performed in the studies mentioned above. PNP is relatively insidious, and the clinical manifestations are not typical (51). Many cases tend to develop PNP after acute infection (37). Also, there is a lack of large-scale studies assessing PNP in COVID-19 patients, especially those investigating neurological recovery in long-term rehabilitation of COVID-19 survivors.

NCS is important for diagnosing PNP and selecting the best therapy (52). A previous study (53) assessed the correlation between electromyography and NCS in COVID-19 patients and found that NCS is important for predicting COVID-19 prognosis and recovery. Using NCS, we found that most patients had persistent PNP, mainly polyneuropathy, 6 or 12 months after infection with SARS-CoV-2. Also, PNP was positively associated with age. The occurrence of F-wave abnormality was relatively low.

The present study has several limitations. First, some symptoms, such as fatigue, nausea, and chest tightness, were self-reported, which may lead to recall bias. Second, not all enrolled patients were measured twice, and the NCS results of patients at 6-month follow-up may change after that, which may give rise to an attrition bias. The positive patients were hospitalized at the same place, and activities were severely limited during the lockdown, but after 6–12 months and once the lockdown was lifted, the recovered patients resumed school and work or returned to their living place. Still, those patients were included in the study on the basis that they would consult if neurological symptoms occurred. A selection bias is also unavoidable since selection criteria were applied when enrolling the patients. Third, the relatively small proportion of severe or critically severe COVID-19 patients in our study limits the generalizability of the findings. The study population was limited to a single hospital in China, which may not be generalizable to populations with different demographics, healthcare access, and genetic predispositions. Fourth, this study did not describe the intrinsic link between PNP and persistent multi-organ functional injury. nCoV IgG levels were measured at discharge, while the development of F-wave abnormalities was assessed at 6 and 12 months. The IgG levels were not measured during follow-up, and no dynamic data were available after discharge. In addition, the exact moment when F-wave abnormalities developed is also unknown. Fifth, some risk factors for PNP were not accounted for, such as diabetes, lifestyle factors, and vitamin deficiencies. Sixth, the present study used the cut-offs suggested by Tankisi et al. in 2005 (18), but the selection of the cut-off can affect the results and conclusions. Finally, the use of ORs can overestimate the risk, especially when the frequency of the outcome is high. Next, we plan to analyze more clinical data of COVID-19 patients during hospitalization and discharge; blood biochemical test indicators, six-minute walk tests, lung function, anxiety/depression, etc., will be assessed during this phase. Correlation analysis with the relevant indicators of peripheral nerve electrophysiology will be carried out to fully elucidate the relationship between PNP and the function of major organs of the body and further study the pathophysiological mechanism of PNP to provide objective data for complete recovery. Nevertheless, these results suggest that older patients with COVID-19 and those with high nCoV IgG levels at discharge should be monitored for PNP during follow-up. Future studies should be performed with larger numbers of patients to confirm those factors and determine their exact predictive value for PNP and other abnormalities.

In conclusion, age might be an independent risk factor for PNP, while higher education was associated with mononeuropathy. Future studies should investigate the factors and interventions for PNP in patients with COVID-19.