Background

The diaphragm, the fundamental muscle of respiration, is innervated by the phrenic nerve which originates from the cervical ventral horn of C3–C5 roots, and it is supplied mainly by C4 [1]. Injury to the phrenic nerve and/or one of its roots may lead to diaphragmatic paralysis (DP) and dysfunction [2]. The spinal level mostly affected in cervical spondylosis is C5–C6, C6–C7, and C4–C5. Accordingly, the phrenic nerve is vulnerable to compression in cervical spondylosis [3, 4]. Phrenic nerve injury can be evaluated using multiple modalities including esophageal and gastric manometry, phrenic nerve conduction studies (NCS), diaphragmatic electromyography (EMG), or ultrasound [5]. We aimed to investigate possible subtle phrenic nerve injury in patients with cervical spondylosis.

Methods

Patient selection

This case-control study was conducted on 30 patients suffering from neck pain diagnosed as cervical spondylosis above C5 by using X-ray cervical spines. Additionally, there are 30 sex- and age-matched healthy volunteers in the control group. Both patients and volunteers were recruited from the physical medicine, rheumatology, and rehabilitation outpatient clinic.

Exclusion criteria

Patients presenting with a lesion that can possibly affect the phrenic nerve like respiratory disease, diabetes mellitus, peripheral neuropathy, multiple sclerosis, motor neuron disease, Guillain-Barre syndrome, high-level spinal cord injury, sarcoidosis, muscle dystrophy, cardiac and thoracic surgery, trauma, infections (Herpes zoster and Lyme disease), and/or malignancy were excluded from the study.

This study was submitted to and approved by the FMASU REC ethics committee, FWA 000017585. All participants signed a written consent to participate in this study.

Methodology

All participants were subjected to a full medical history and thorough clinical examination. Posteroanterior and lateral cervical X-ray films using a grading system of the cervical degenerative index by Ofiram et al. were performed [6]. Cervical spine magnetic resonance imaging (MRI) was performed on patients with grade 3 and grade 4 to exclude cervical disc prolapse. After taking the participants’ consent, electrophysiological studies were performed in a quiet room with a constant temperature of 27 °C. The participant was placed in the supine position. The operator was blinded with the status of the subjects as either patients or controls. The study was conducted using a four-channel electromyography instrument (EMG/NCV/EP system topas 230/240 V Schwarzer GmbH). Phrenic NCS were performed bilaterally on the participants with a bipolar stimulating electrode between the sternal and clavicular heads of the sternocleidomastoid muscle just above the clavicle or lateral to the clavicular head. The active electrode (G1) was fixed 5 cm above the xiphoid process, and the reference electrode (G2) was placed 16 cm from G1 on the chest margin ipsilateral to the stimulated phrenic nerve [7] (Fig. 1). Phrenic nerve CMAP onset latency (ms) and amplitude (μV) were obtained at supramaximal stimulation [7]. The values of phrenic NCS were obtained from the patient group and compared to the corresponding sides in the control group.

Fig. 1
figure 1

Application of recording electrodes for phrenic NCS

Statistical analysis

Data were calculated and tabulated, and statistical analysis was performed. IBM SPSS statistics (V. 23.0, International Business Machines Corporation, Armonk, NY) was used for data analysis. The following tests were done: comparison between two independent mean groups for parametric data using the Student t test and the Pearson correlation test to study the possible association between two variables among each group for parametric data. The probability of error P ≤ 0.05 was considered significant while P ≤ 0.001 was considered highly significant. Sensitivities and specificities of phrenic nerve DML and amplitude were analyzed using the receiver operating characteristic (ROC) curve.

Results

Our study included 30 patients with a diagnosis of cervical spondylosis with an age range of 38–57 years and were matched with 30 healthy controls with ages ranging from 34 to 55 years. In the patient group, 18 were males (M = 60%) and 12 were females (F = 40%). In the control group, 19 participants were males (M = 63.3%) and 11 were females (F = 36.7%). The patient and control groups were also matched regarding sex (P > 0.05). BMI in the patient group and control group was 27.29 ± 1.92 and 26.21 ± 1.96, respectively.

The patients with cervical spondylosis showed a statistically highly significant prolongation of phrenic nerve DML on both sides compared to the controls (P < 0.001). Evaluation of phrenic nerve amplitude revealed a non-significant difference between both groups (P > 0.05) (as shown in Table 1).

Table 1 Comparison between patients and controls regarding electrophysiological data

ROC curve analysis revealed a good ability of DML to differentiate between healthy and affected phrenic nerve (area under the curve (AUC) 76.6%) which was more than the amplitude (AUC 60%). At the individual level, values of phrenic NCS exceeding the cutoff point and normative data obtained from the controls were considered abnormal (shown in Table 2). Only 3 patients (10%) showed normal DML while 27 patients (90%) showed delayed value regarding this parameter (Fig. 2). At the same time, 14 patients (46.6%) showed lower amplitude while 16 of them (53.4%) revealed average value.

Table 2 The determined cutoff points for different electrophysiological studies
Fig. 2
figure 2

Phrenic NCS showing delayed DML (7.52) in one of our patients, phrenic n. DML cutoff > 7.2 ms and amplitude cutoff < 387 μv

Interestingly, the X-ray score has a highly significant positive correlation with right phrenic nerve DML (P < 0.001) and a significant correlation with the left (P < 0.05). There was no significant difference between the amplitudes of CMAP of Rt and Lt phrenic nerves (P < 0.05), and there was no significant correlation between the duration of neck pain and both phrenic DML and CMAP amplitude on both sides (as shown in Table 3).

Table 3 Correlation between phrenic DML and some parameters

Discussion

Though diaphragmatic dysfunction due to phrenic nerve compression by cervical spondylosis (at C3, C4, and/or C5) is not a common cause, further investigation is required to be excluded [8]. Phrenic nerve neuropathy usually has non-specific signs and symptoms, so high suspicion should be raised especially in unilateral DP when a patient is asymptomatic at rest but has dyspnea on exertion [9]. Supporting this, Sagoliocco et al. revealed a dramatic unilateral phrenic nerve involvement in the absence of clinical and laboratory evidence of diaphragmatic weakness [10].

The significant delay of phrenic nerve DML in the patient group with cervical spondylosis compared to the control group could be a sign of demyelination which could be attributed to hypoxia caused by compression affecting myelination and conduction of impulses [11]. Noticeably, we did not record a significant reduction in the amplitude of phrenic nerve CMAP of the patient group. It is thought that this could be due to the axons of the phrenic nerve being unaffected by the compression and hypoxia caused by spondylosis. Higher levels of compression are known to occur with severe disc prolapse and myelopathies [12, 13].

Our study reported a significant positive correlation between DML and X-ray score where 100% of patients with grade 4 had delayed DML. Sachin et al. supported the same concept and showed that DML is not affected in cervical radiculopathies unless there is extreme demyelination of axons [13]. The patients with high X-ray scores in our study were found to have severe phrenic demyelination.

In the current research, we reported a non-significant negative correlation between phrenic nerve CMAP amplitude and X-ray score in the patient group. This could be explained as a tendency of an axonal lesion to the phrenic nerve roots which similarly occurs with severe disc prolapse and myelopathies [12,13,14]. Additionally, a non-significant correlation between values of phrenic NCS and neck pain duration could be recorded which may indicate that phrenic nerve DML is mostly affected by the grade of cervical spondylosis and not the duration itself.

Conclusions

Cervical spondylosis is an underestimated cause of phrenic nerve delayed DML. There is a correlation between the delay of phrenic nerve DML and the severity of cervical spondylosis.

We recommend further investigation in the form of MRI of the cervical region to all patients, diaphragmatic EMG in patients with lower amplitudes of phrenic CMAP, and neuromuscular ultrasound to evaluate the phrenic nerve.

Study limitations

The limitation of the current study is the small sample size to be done on a larger scale for more accurate results. We recommend further investigations in the form of MRI of the cervical region to all patients, diaphragmatic EMG in patients with lower amplitudes of phrenic CMAP, and neuromuscular ultrasound for the phrenic nerve.