INTRODUCTION
Degenerative cervical myelopathy (DCM) arises from spinal cord compression, leading to a gradual deterioration of neurological function [1,2]. Impairment of hand motor function is one of the most common symptoms experienced by DCM patients, with reported incidences ranging from 26% to 90% [3,4]. This symptom makes daily tasks such as using utensils, sewing, writing, or buttoning up challenging, and significantly diminishes the quality of life in patients [4-6]. Given that one of the primary goals of surgical treatment is to stop the deterioration of neurological function and even try to improve it, understanding the recovery potential in DCM patients with hand dexterity impairment helps align treatment with patients’ needs and expectations [1]. However, while some previous research indicates that a majority of patients experienced improvements in dexterity [7,8], another study reported insignificant changes [9]. As preoperative neurofunction emerges as a significant contributing factor to surgical outcomes, the variability in baseline hand function may contribute to inconsistent results [10-12]. Therefore, it is essential to compare the recovery potential of patients with different severities of hand dexterity impairment.
Numerous studies have assessed the overall status of patients using scoring systems such as the modified Japanese Orthopaedic Association (mJOA) score and the JOA Cervical Myelopathy Evaluation Questionnaire (JOACMEQ) [9,13-15]. Both are multiple-item scoring systems and include domains specifically focused on the motor function of the upper limb. Questions such as how well a patient can button their shirt or feed themselves are used to evaluate hand dexterity. These subjective measurement indirectly captures patients’ perceptions of their hand functions, and exhibits an inherent limitation known as the “ceiling effect” when reaching the upper limit of its scale, making it unsuitable for classifying hand dexterity impairment [11,16]. Accordingly, it is important to employ an objective measurement that directly classifies different hand dexterity impairments.
The 10-second grip and release (10s-G&R) test, simply counting the grip and release cycles within 10 seconds, is one of the widely accepted objective assessments for hand motor function [17,18]. In our previous work, we developed an artificial intelligence (AI) video analysis system to capture and analyze the 10s-G&R test. This system demonstrated fair accuracy and inference speed, enhancing our documentation process during follow-up [19]. On this basis, we collected data and established a classification system to differentiate various levels of hand dexterity impairments, allowing for the categorization of individuals based on their actual performance [20]. In this study, we prospectively categorized patients into distinct hand dexterity impairment groups. By comparing their outcomes, our goal is to elucidate the recovery potential of hand function across different severity groups and to establish a preoperative 10s-G&R test cutoff value associated with incomplete recovery.
MATERIALS AND METHODS
1. Ethics Statements
This prospective observational study was conducted as a component of our approved cohort study focusing on hand motions in patients with DCM (Research Ethics Committee Approval No. GDREC2020291H, ClinicalTrials.gov ID: NCT06041542). Prior to the commencement of data collection, all participating patients provided informed consent. The study adhered to the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) statement [21].
2. Participants
The flow chart of the study was shown in Fig. 1. We systematically recruited 152 consecutive series of DCM patients at our hospital starting from October 2021. Inclusion criteria comprised individuals who: (1) were 18 years or older; (2) had a confirmed diagnosis of DCM, determined by the presence of at least one clinical sign or symptom of myelopathy along with evidence of spinal cord compression on imaging; (3) exhibited hand dexterity impairment, defined as a 10s-G&R test with fewer than 20 cycles.18 Notably, the use of the 10s-G&R test in this study is not for diagnosing DCM but for evaluating hand dexterity. Although the diagnostic cutoff value for DCM using the 10s-G&R test can be influenced by age and sex, a G&R cycle count of less than 20 within 10 seconds served as an indicator of abnormal hand motion by its original definition. Therefore, we utilized the threshold of 20 to distinguish between individuals with and without hand function impairment.
Exclusion criteria encompassed patients with: (1) a history of previous spine surgery; (2) concomitant traumatic spinal cord injury or other neurological disorders potentially affecting neurological function; (3) a diagnosis of rheumatoid arthritis, hand injury, or any other condition influencing hand motion. 41 individuals were excluded, and 111 participants were enrolled into follow-up. The follow-up was conducted at the outpatient department 1 year after surgery. Finally, 8 patients were lost to follow-up and 1 underwent a further cervical surgery were excluded, leaving 102 participants in the study (Fig. 1).
3. Surgical Technique
Surgical techniques for cervical decompression included anterior cervical discectomy and fusion (ACDF) and laminoplasty, performed by 3 senior spine surgeons (YC, GL, and YY). The choice of surgical approach depended on both the characteristics of the cord compression and the experience of the surgeons. ACDF was employed for patients with 1- or 2-level ventral compression caused by degenerative discs and/or osteophytes. Laminoplasty was utilized for cases of congenital and/or degenerative cervical stenosis involving 3 or more levels, with or without hypertrophy of the ligamentum flavum.
4. Measurements and Grouping
Data were collected preoperatively and 1 year postoperatively. Demographic and clinical data, including symptoms, signs, surgical approach, and radiological findings, were thoroughly recorded at the time of enrollment. Patients recorded the 10s-G&R test using our AI system mentioned above. The system is an automated process that offers higher accuracy than manual counting, particularly when the speed of the patient's hand movement is unstable. The total G&R numbers of the more affected hand were counted and selected as the primary outcome measurement for subsequent analysis. All participants were categorized into 3 groups based on the severity of hand dexterity impairment, determined by the preoperative 10s-G&R test as described earlier: mild group (17–19 cycles), moderate group (14–16 cycles), and severe group (13 cycles or less) [20]. A 10s-G&R test result of less than 20 cycles at the 1-year follow-up was considered indicative of incomplete recovery of hand dexterity [18,20]. The incomplete recovery rate was calculated as the percentage of individuals who did not achieve normal hand dexterity.
Secondary measurements that specifically for hand function included: (1) upper limb motor dysfunction scored employed form the mJOA scores (mJOA upper) [14]; (2) upper extremity function score form the JOACMEQ (JOACMEQ upper extremity function) [22]. Additional scores used to evaluate the overall status of patients included the overall mJOA score and other domains of the JOACMEQ [14,22]. Improvements of the above hand dexterity parameters were assessed based on the differences between preoperative and 1-year follow-up scores.
5. Statistical Analysis
The outcome measurements across 3 groups were compared using 1-way analysis of variance and chi-square tests. Multivariate regression analyses were conducted to identify potential predictive factors for postoperative hand function. The efficacy of the preoperative 10s-G&R test in predicting incomplete recovery of hand dexterity after surgery was assessed using receiver operating characteristic (ROC) curve analysis, with the area under the curve (AUC) calculated. The Youden index defined the cutoff value of the ROC curve, and the relative risk (RR) reflected the risk of incomplete recovery in patients with 10s-G&R test results below the cutoff value. Statistical analyses were performed using IBM SPSS Statistics ver. 23.0 (IBM Co., Armonk, NY, USA), with a significance level set at p< 0.05.
RESULTS
1. Baseline Characteristics of Different Severity Groups
The baseline demographic data were shown in Table 1. The numbers of patients in the mild, moderate, and severe groups were 30, 32, and 40. The sex, age, duration of symptom, body mass index, and smoking status were similar among the 3 groups (all p > 0.05). The severe group displayed significantly higher percentages of positive Hoffmann sign, T2-weighted imaging intramedullary high signal, and posterior approach surgeries, as well as an elevated involved segment numbers (all p< 0.05). The preoperative 10s-G&R test results demonstrated notable distinctions across the 3 groups (average cycles= 17.90± 0.71, 15.13± 0.79, 11± 2.08, respectively, p<0.001). Furthermore, all subjective measurements, encompassing mJOA upper score and JOACMEQ upper extremity function, indicated significant deterioration in the severe group (all p< 0.001).
2. Hand Functional Outcomes of Different Severity Groups
The recovery potential of the 3 groups was presented in Fig. 2. The postoperative differences among the 3 groups were significantly reduced, with the average cycles being 21.83± 5.82, 19.97± 4.45, and 18.65± 5.11, respectively (p= 0.041). The severe group exhibited a higher improvement in the 10s-G&R test (7.65± 5.02) compared to the other 2 groups (3.93± 5.70 and 4.84± 4.3, p= 0.006). Likewise, the average improvements in mJOA Upper score, JOACMEQ upper extremity function score, and overall mJOA score within the severe group surpassed those in the other 2 groups, resulting in smaller differences among these scores across the groups.
Fig. 3 illustrated the postoperative distribution of hand function across the 3 groups. The moderate group exhibited an increase in the proportions (18%) of moderate and severe impairments after surgery, with an even more pronounced (34%) rise observed in the severe group. The incomplete recovery rates were 26.67% in the mild group, 46.88% in the moderate group, and reach to 57.50% in the severe group (p< 0.05). Notably, not all patients experienced improvement in their hand function after surgery. The ratio of individuals who experienced any kind of improvement in hand function after surgery were similar among the 3 groups (73% for mild, 82% for moderate, and 83% for severe, p= 0.61). Three individuals (10%) in the mild group and 3 (9%) in the moderate group encountered hand dexterity deterioration, while 5 (17%), 3 (9%), 7 (17%) patients in the 3 groups, respectively, remained in the same hand dexterity grade after surgery.
3. ACDF and Laminoplasty Demonstrated Similar Efficacy in Improving Hand Function
The surgical approach in the severe group was not comparable to those in the other groups (Table 1). While 60% and 62% of patients in the mild and moderate groups underwent ACDF, only 22% of those in the severe group received this procedure. The preoperative 10s-G&R test score was significantly higher in the ACDF group (15.49± 2.57) compared to the laminoplasty group (13.32± 3.41, p< 0.01). Both surgical approaches yielded satisfactory outcomes at the 1-year follow-up (20.53± 5.35 for ACDF and 19.55 ± 5.17 for laminoplasty, p = 0.347). To minimize bias from preoperative severity and more accurately compare the efficacy of these 2 surgical approaches, we performed a stratified analysis within each severity group. Preoperative 10s-G&R test scores were comparable between the 2 approaches within each severity group (Fig. 4). At follow-up, all 10s-G&R test scores showed significant improvement compared to preoperative levels and remained similar between the 2 surgical approaches across all severity groups (Fig. 4). These findings suggest that, after adjusting for preoperative severity, both ACDF and laminoplasty demonstrate comparable efficacy in improving hand function.
4. Preoperative 10s-G&R Test Result Predicts Hand Functional Outcome
A multivariate linear regression analysis revealed that preoperative 10s-G&R test result, age, Hoffmann sign, Duration of symptom, and mJOA upper score serve as significant predictors for postoperative 10s-G&R test outcomes (Table 2). Among all variables, preoperative 10s-G&R test result demonstrated the highest standardized coefficient (0.416, Table 2).
5. Cutoff Value of 10s-G&R Test for Predicting Hand Function Incomplete Recovery
Subsequently, we conducted ROC curve analysis to examine the predictive efficacy and determine the cutoff value of the preoperative 10s-G&R test (Fig. 5). The ROC curve showed an AUC of 0.67 (95% confidence interval [CI], 0.558–0.769; p= 0.004), reflecting an acceptable predicting performance. We found that a preoperative 10s-G&R test of 15 cycles is the cutoff value for predicting incomplete recovery at 1 year after surgery, with a sensitivity of 64.90% and specificity of 63.00%. The RR value is 1.919 (95% CI, 1.217–3.026; p= 0.005).
DISCUSSION
This study marks the first attempt to compare hand functional outcomes in DCM patients based on different severities of hand dexterity impairment, utilizing the 10s-G&R test as a preoperative categorization tool. Since its development in 1987, the 10s-G&R test has become a widely accepted objective measure for assessing hand dexterity [17,18]. This physical performance test has proven to be accurate in classifying various severities of hand dexterity impairment [18,23]. Using this preoperative categorization tool, we found that severe functional impairment before surgery greatly reduces the chances of recovery. Specifically, if the 10-second test result is less than 15 repetitions, recovery becomes challenging.
Our study provides valuable insights for clinical practice. Patients often ask whether their hand dexterity impairment improves after surgery and to what extent it improves. However, previous studies did not consider preoperative hand function as a confounding factor, leading to inconsistent results during comparisons [7-9]. In this study, we prospectively classified patients into 3 severity groups. The average improvements of 10s-G&R test results, mJOA Upper score, JOACMEQ upper extremity function score and overall mJOA score between preoperative and 1-year follow-up remarkably underscore the recovery potential of hand dexterity impairment among patients in all 3 groups following surgical treatment. Specifically, around 70%–80% of patients experienced any kind of improvement in hand function after surgery, suggesting surgical treatment is beneficial for patients with any severity of hand dexterity impairment. However, while patients with mild impairment are more likely to achieve complete recovery, the probability decreases significantly in the moderate group and even more in the severe group. Patients with severe hand dexterity impairment may find it more challenging to recover completely compared to other groups. Our results facilitate surgeon-patient communication by providing high quality evidence for establishing appropriate expectations regarding prognosis.
Another frequently asked question is about the appropriate timing for considering surgical treatment. According to the guidelines, surgical intervention is recommended for patients with moderate and severe DCM, as well as for mild DCM patients with neurological deterioration [24]. Previously, we observed that patients with 10s-G&R test less than 14 cycles are more likely to have moderate-to-severe DCM, indicating a need for surgical intervention [20]. In this study, we moved forward and demonstrated that it may be advisable to consider surgical treatment before the 10s-G&R test result drops below 15 cycles, as it could double the risk of incomplete recovery of hand function. This surgical criterion based on 10s-G&R test serves as a crucial supplementary tool for decision-making during clinical practice.
Ossification of the posterior longitudinal ligament (OPLL) is a specific form of DCM that may influence surgical outcomes [25]. In our cohort, patients were diagnosed with either cervical spondylotic myelopathy (CSM) or OPLL based on their pathology. The proportion of OPLL cases was similar across the 3 groups (p > 0.05) (Table 1). The preoperative 10s-G&R test scores of patients with OPLL and CSM were comparable (14.00± 3.47 and 14.46 ± 3.13, respectively, p > 0.05), as were the postoperative scores (18.93± 4.81 and 20.44± 5.39, respectively, p> 0.05). Additionally, logistic regression analysis showed that diagnosis was not a predictor of postoperative hand functional outcomes (β= 0.662, p> 0.05). These findings suggested that the type of diagnosis did not affect hand dexterity outcomes after surgery.
Spinal cord compression severity may be another potential factor influencing surgical outcomes. The average preoperative narrowest anteroposterior canal diameter was only slightly smaller in the severe group (6.01± 1.22 mm, p= 0.469) compared to the mild (6.20±1.35 mm) and moderate (6.76±1.26 mm) groups, suggesting similar spinal cord compression severity across the 3 groups (Table 1). It is important to note that we categorized patients based on hand function impairment rather than overall neurological function. This means that patients in the mild group (in terms of hand function) might have had severe lower limb impairment, which explains the comparable spinal cord compression severity across the groups. The relationship between upper and lower limb function warrants further investigation.
Notably, 6 patients (5.8%) in our cohort experienced a decline in hand dexterity after surgery, while 15 patients (14.7%) maintained the same hand dexterity grade. The incidence of nonimprovement was similar across all 3 groups. Importantly, all patients in the study, including the 6 cases with deterioration, achieved a change in anteroposterior canal dimension greater than 5 mm, indicating adequate neural decompression (Table 1) [26]. Among those who experienced deterioration, 3 patients were older in age, which has been reported as a risk factor for neurofunctional deterioration after surgery [27]. However, no identifiable reason was found in the other 3 cases. As noted in previous research, patients should be aware that there is a risk that postoperative hand dexterity may not improve or may even worsen, for reasons that are not yet fully understood [27].
A limitation of this study is that the selection of the surgical approach was determined by real-world factors rather than through randomization, which inevitably introduced bias. In clinical practice, the choice of surgical approach is typically influenced by multiple factors. Therefore, maintaining the real-world approach selection in our study potentially enhanced the applicability of our findings to predict hand dexterity outcomes. Furthermore, previous research has demonstrated that overall outcomes between ACDF and laminoplasty are similar [28]. Although our analysis also indicated comparable outcomes between anterior and posterior approaches, a carefully designed randomized controlled trial is needed to confirm this finding. Additionally, our follow-up time was limited to only 1 year. A longer-term study is needed to clarify whether hand function will continue to improve.
CONCLUSION
In this prospective study, we have shown that patients exhibit potential for recovery in hand dexterity impairment regardless of their preoperative condition. However, almost half of the patients in the moderate group, and even more in the severe group, are unable to achieve complete recovery in normal hand dexterity. A preoperative 10s-G&R test < 15 cycles predicts a higher risk of incomplete recovery of hand function, suggesting that surgical treatment is recommended before the 10s-G&R test drops below 15 cycles.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
