A retrospective observational cohort study of ulnar nerve compression at the elbow at Department of Hand Surgery, Plastic Surgery and Burns, Linköping University Hospital, Linköping, Sweden, where outcome of surgery was analyzed in 173 patients [29 bilateral surgery; thus, in total 202 cases (19)], was earlier performed. A majority of the original 173 patients (202 cases) had a concomitant affection of another nerve in operated arm and one third had a concomitant affection of another nerve in contralateral arm (Figure 1). Around 50% had an MRI performed preoperatively, with referral to radiology of whom 30% had signs of nerve root affection (level not specified) according to the patients' charts. For details of the patient cohort see Giöstad and Nyman 2019 (19).

In the present study, pre- or postoperative available MRI images were retrieved and re-evaluated (for details see below). The inclusion criterion of the present study was availability of MRI examination with images that could be re-evaluated as described below. Background data of the present 62, out of the earlier presented 173, patients were extracted from the file based on the patient charts, including age, sex, uni- or bilateral surgery, McGowan grading for severity of preoperative ulnar nerve problems (20), other nerve compression lesion(s) and surgery for other hand condition in the same or opposite arm, shoulder or neck problems, and presence of polyneuropathy and diabetes. Data of outcome was also extracted from the file, consisting of DASHscore and replies from two further questions: “How do you think the hand/works today compared to before surgery?” and “Are you pleased with the result from the surgery?”; both graded as completely fine – improved vs. unchanged—worse; data from 173 patients earlier published (19).

MRI images from the present 62 patients (Figure 1), investigated within the last 3 years before or after surgery, were available for re-evaluation and re-analyzed by a single experienced neuroradiologist (KAK), who was not aware of any clinical patient data, including affected side, if surgery was performed bilaterally, presence of any concomitant nerve compression disorder or outcome of surgery. The evaluation included assessment of presence of disc herniation (location, side, level), cervical spinal (central) stenosis, any spinal nerve root affection (side and level; graded as normal, probable affection and definite affection), spinal cord compression, increased T2 signal in spinal cord, disc degeneration (level) and Modic I (Modic type I endplate change) (Figures 2, 3). The Modic type endplate changes represent a classification for vertebral body endplate MRI signal. Modic type I is defined as T1: low signal; T2: high signal; represents bone marrow oedema and inflammation (Table 1). Spinal nerve root affection was defined as (a) no affection, (b) a probable involvement of the specific nerve root and (c) a definite involvement of a specific nerve root at the individual C3-Th1 nerve root levels bilaterally. Data were then further categorized into (a) no spinal nerve root affection, or (b) spinal nerve root affection C8–Th1 (origin of the ulnar nerve), and spinal nerve root affection of C3–C7, respectively, based on overall judgment, where a probable involvement was defined as a definite affection if a disc herniation, a central stenosis or a disc degeneration were present (Table 1).

Data are presented as numbers (n; %) or median (interquartile range; IQR). Any differences in categorial data were analyzed with χ²-test or Fisher’s exact test and in continuous data (not-normally distributed) with the Mann-Whitney U Test. The significance level was set at a p-value of <0.05. Statistical analyses were performed with IBM SPSS Statistics (version 26; 2019).

All methods were performed in accordance with relevant guidelines and regulations, including the Declaration of Helsinki.

Among 62 patients, 39 were women and 23 were men [latter statistical higher age (p = 0.012) and more frequently observed polyneuropathy (p = 0.016); no other significant differences were seen between sex] (Table 2). Among all patients (n = 61; information missing in one case), 15 (25%) were smokers. Smokers were more frequently observed among cases with spinal nerve root affection at C3–C7 levels at surgically treated side (p = 0.039; Table 3), but without any other significant differences in smoking habits. Unilateral surgery was performed in 45 patients (right side 23 patients; left side 22 patients) and bilateral surgery in 17 patients with a similar distribution of McGowan grading 1–3 (Table 2). Another compression neuropathy was frequently present in the same (39/62 patients; 63%) or contralateral (29/62; 47%) arm. A hand surgical co-morbidity, treated at the same time as ulnar nerve compression, was observed in 24/62 (39%) patients. Shoulder problems were less frequently reported (11/62; 18%). Neck problems were observed in around half of the patients (29/62; 47%). Data are summarized in Tables 2, 3.

In three patients, a single spinal nerve root at C3–C7 level was judged as a probable affection without any other pathology and in one patient a bilateral probable affection was observed at C7 level. These changes were also defined as spinal nerve root affection. Most changes were observed in C3–C7 spinal nerve roots. Affection of spinal nerve roots C3–C7 at surgically treated side was seen in almost half of the cases (26/62; 42%), while C8-Th1 spinal nerve root affection was rarely observed (1/62; 2%; Table 4, Supplementary Table S1). The patient with C8-Th1 spinal nerve root affection at the surgically treated side (unilateral surgery) had also C3–C7 spinal nerve root affection at the ipsi- and contralateral sides. Spinal nerve root affection at C3–C7 levels at contralateral side was seen in about the same frequency as for surgically treated side (39% and 42%, respectively) among all patients. The single case with C8-Th1 involvement had affection at the surgically treated side (Table 4). Fifty percent (31/62) of the patients had nerve root affection at any side or level of the cervical spinal nerve roots (Table 4). Disc herniation, cervical spinal stenosis, medulla compression, increased T2 signal in the spinal cord, Modic I or disc degeneration were less frequently observed (Supplementary Table S2).

There was a significant difference in age between the patients with (n = 31) and without (n = 31) spinal nerve root affection C3-Th1 at any side (p = 0.001), as well as those with (n = 26) and without (n = 36) spinal nerve root affection at C3–C7 at surgically treated side (p = 0.001), showing a higher age in those with spinal nerve root affection (Table 3). No differences regarding sex were observed between those with and without spinal nerve root affection (p > 0.05). Generally, men were significantly older than women among all patients (p = 0.012, Table 3). There were no differences concerning a concomitant neuropathy in the same or in the contralateral arm and spinal nerve root affection at surgically treated or contralateral side, respectively, at C3–C7 spinal nerve roots (p = 0.29 and p = 0.30, respectively) or C3-Th1 spinal nerve root affection at any side (p = 0.29 and p = 0.61, respectively, Table 3). There were no differences between unilateral and bilateral cases concerning presence of spinal nerve root affection at surgically treated or contralateral sides at C3–C7 spinal nerve roots (p = 0.58, and p = 0.78, respectively) or C3-Th1 spinal nerve roots at any side (p = 0.26; Table 3).

There were differences concerning C3–C7 spinal nerve root affection at surgically treated and contralateral sides in relation to McGowan grading among all cases (p = 0.017 and p = 0.022, respectively; Table 5) with McGowan grade 3 more frequently observed in those with nerve root affection. McGowan grade 3 was also related to more frequently observed ipsilateral, but not contralateral, pathology at C3–C7 spinal nerve roots among surgically treated patients at unilateral side (p = 0.045). Analyses of bilateral cases (n = 17 patients; n = 34 cases) showed no relation between spinal nerve root affection and McGowan grading (p = 0.06; Table 5).

There was a relation between presence of C3–C7 spinal nerve root affection at surgically treated and contralateral sides among patients with unilateral (p = 0.005) ulnar nerve compression at the elbow (Supplementary Table S3).

Residual problems were common [median DASH score 40 (10–51) at minimum follow up time 12 months; n = 30; no differences in sex; p = 0.80]. No significant differences in DASH scores between no affection and affection at surgically treated side of C3–C7 spinal nerve roots (p = 0.46) or no affection and affection of C3-Th1 spinal nerve roots at any side (p = 0.30) were detected. Generally, 53% of the patients (n = 30) expressed completely fine or improved hand function as result of surgery with respect to the two questions: “How do you think the hand/works today compared to before surgery?” (completely fine – improved vs. unchanged—worse) and “Are you pleased with the result from the surgery?” (completely fine – improved vs. unchanged - worse); no sex differences (p = 0.47 and p = 0.47, respectively). Concerning the two outcome questions, there were no relations between no affection and affection at surgically treated side of C3–C7 spinal nerve roots (p = 0.72 and p = 0.72, respectively) or no affection and affection of C3-Th1 spinal nerve roots at any side (p = 0.48 and p = 0.48, respectively). Data is shown in Supplementary Tables S4.