Agarwal, Nitin

OBJECTIVE The aim of this study was to explore the preoperative patient characteristics that affect surgical decision-making when selecting an anterior or posterior operative approach in patients diagnosed with cervical spondylotic myelopathy (CSM). METHODS This was a multi-institutional, retrospective study of the prospective Quality Outcomes Database (QOD) Cervical Spondylotic Myelopathy module. Patients aged 18 years or older diagnosed with primary CSM who underwent multilevel (≥ 2-level) elective surgery were included. Demographics and baseline clinical characteristics were collected. RESULTS Of the 841 patients with CSM in the database, 492 (58.5%) underwent multilevel anterior surgery and 349 (41.5%) underwent multilevel posterior surgery. Surgeons more often performed a posterior surgical approach in older patients (mean 64.8 ± 10.6 vs 58.5 ± 11.1 years, p < 0.001) and those with a higher American Society of Anesthesiologists class (class III or IV: 52.4% vs 46.3%, p = 0.003), a higher rate of motor deficit (67.0% vs 58.7%, p = 0.014), worse myelopathy (mean modified Japanese Orthopaedic Association score 11.4 ± 3.1 vs 12.4 ± 2.6, p < 0.001), and more levels treated (4.3 ± 1.3 vs 2.4 ± 0.6, p < 0.001). On the other hand, surgeons more frequently performed an anterior surgical approach when patients were employed (47.2% vs 23.2%, p < 0.001) and had intervertebral disc herniation as an underlying pathology (30.7% vs 9.2%, p < 0.001). CONCLUSIONS The selection of approach for patients with CSM depends on patient demographics and symptomology. Posterior surgery was performed in patients who were older and had worse systemic disease, increased myelopathy, and greater levels of stenosis. Anterior surgery was more often performed in patients who were employed and had intervertebral disc herniation.

OBJECTIVE Compared with laminectomy with posterior cervical fusion (PCF), cervical laminoplasty (CL) may result in different outcomes for those operated on for cervical spondylotic myelopathy (CSM). The aim of this study was to compare 24-month patient-reported outcomes (PROs) for laminoplasty versus PCF by using the Quality Outcomes Database (QOD) CSM data set. METHODS This was a retrospective study using an augmented data set from the prospectively collected QOD Registry Cervical Module. Patients undergoing laminoplasty or PCF for CSM were included. Using the nearest-neighbor method, the authors performed 1:1 propensity matching based on age, operated levels, and baseline modified Japanese Orthopaedic Association (mJOA) and visual analog scale (VAS) neck pain scores. The 24-month PROs, i.e., mJOA, Neck Disability Index (NDI), VAS neck pain, VAS arm pain, EQ-5D, EQ-VAS, and North American Spine Society (NASS) satisfaction scores, were compared. Only cases in the subaxial cervical region were included; those that crossed the cervicothoracic junction were excluded. RESULTS From the 1141 patients included in the QOD CSM data set who underwent anterior or posterior surgery for cervical myelopathy, 946 (82.9%) had 24 months of follow-up. Of these, 43 patients who underwent laminoplasty and 191 who underwent PCF met the inclusion criteria. After matching, the groups were similar for baseline characteristics, including operative levels (CL group: 4.0 ± 0.9 vs PCF group: 4.2 ± 1.1, p = 0.337) and baseline PROs (p > 0.05), except for a higher percentage involved in activities outside the home in the CL group (95.3% vs 81.4%, p = 0.044). The 24-month follow-up for the matched cohorts was similar (CL group: 88.4% vs PCF group: 83.7%, p = 0.534). Patients undergoing laminoplasty had significantly lower estimated blood loss (99.3 ± 91.7 mL vs 186.7 ± 142.7 mL, p = 0.003), decreased length of stay (3.0 ± 1.6 days vs 4.5 ± 3.3 days, p = 0.012), and a higher rate of routine discharge (88.4% vs 62.8%, p = 0.006). The CL cohort also demonstrated a higher rate of return to activities (47.2% vs 21.2%, p = 0.023) after 3 months. Laminoplasty was associated with a larger improvement in 24-month NDI score (−19.6 ± 18.9 vs −9.1 ± 21.9, p = 0.031). Otherwise, there were no 3- or 24-month differences in mJOA, mean NDI, VAS neck pain, VAS arm pain, EQ-5D, EQ-VAS, and distribution of NASS satisfaction scores (p > 0.05) between the cohorts. CONCLUSIONS Compared with PCF, laminoplasty was associated with decreased blood loss, decreased length of hospitalization, and higher rates of home discharge. At 3 months, laminoplasty was associated with a higher rate of return to baseline activities. At 24 months, laminoplasty was associated with greater improvements in neck disability. Otherwise, laminoplasty and PCF shared similar outcomes for functional status, pain, quality of life, and satisfaction. Laminoplasty and PCF achieved similar neck pain scores, suggesting that moderate preoperative neck pain may not necessarily be a contraindication for laminoplasty.

OBJECTIVE Spondylolisthesis is a common operative disease in the United States, but robust predictive models for patient outcomes remain limited. The development of models that accurately predict postoperative outcomes would be useful to help identify patients at risk of complicated postoperative courses and determine appropriate healthcare and resource utilization for patients. As such, the purpose of this study was to develop k-nearest neighbors (KNN) classification algorithms to identify patients at increased risk for extended hospital length of stay (LOS) following neurosurgical intervention for spondylolisthesis. METHODS The Quality Outcomes Database (QOD) spondylolisthesis data set was queried for patients receiving either decompression alone or decompression plus fusion for degenerative spondylolisthesis. Preoperative and perioperative variables were queried, and Mann-Whitney U-tests were performed to identify which variables would be included in the machine learning models. Two KNN models were implemented (k = 25) with a standard training set of 60%, validation set of 20%, and testing set of 20%, one with arthrodesis status (model 1) and the other without (model 2). Feature scaling was implemented during the preprocessing stage to standardize the independent features. RESULTS Of 608 enrolled patients, 544 met prespecified inclusion criteria. The mean age of all patients was 61.9 ± 12.1 years (± SD), and 309 (56.8%) patients were female. The model 1 KNN had an overall accuracy of 98.1%, sensitivity of 100%, specificity of 84.6%, positive predictive value (PPV) of 97.9%, and negative predictive value (NPV) of 100%. Additionally, a receiver operating characteristic (ROC) curve was plotted for model 1, showing an overall area under the curve (AUC) of 0.998. Model 2 had an overall accuracy of 99.1%, sensitivity of 100%, specificity of 92.3%, PPV of 99.0%, and NPV of 100%, with the same ROC AUC of 0.998. CONCLUSIONS Overall, these findings demonstrate that nonlinear KNN machine learning models have incredibly high predictive value for LOS. Important predictor variables include diabetes, osteoporosis, socioeconomic quartile, duration of surgery, estimated blood loss during surgery, patient educational status, American Society of Anesthesiologists grade, BMI, insurance status, smoking status, sex, and age. These models may be considered for external validation by spine surgeons to aid in patient selection and management, resource utilization, and preoperative surgical planning.

OBJECTIVE The purpose of this study was to evaluate the performance of different supervised machine learning algorithms to predict achievement of minimum clinically important difference (MCID) in neck pain after surgery in patients with cervical spondylotic myelopathy (CSM). METHODS This was a retrospective analysis of the prospective Quality Outcomes Database CSM cohort. The data set was divided into an 80% training and a 20% test set. Various supervised learning algorithms (including logistic regression, support vector machine, decision tree, random forest, extra trees, gaussian naïve Bayes, k–nearest neighbors, multilayer perceptron, and extreme gradient boosted trees) were evaluated on their performance to predict achievement of MCID in neck pain at 3 and 24 months after surgery, given a set of predicting baseline features. Model performance was assessed with accuracy, F1 score, area under the receiver operating characteristic curve, precision, recall/sensitivity, and specificity. RESULTS In total, 535 patients (46.9%) achieved MCID for neck pain at 3 months and 569 patients (49.9%) achieved it at 24 months. In each follow-up cohort, 501 patients (93.6%) were satisfied at 3 months after surgery and 569 patients (100%) were satisfied at 24 months after surgery. Of the supervised machine learning algorithms tested, logistic regression demonstrated the best accuracy (3 months: 0.76 ± 0.031, 24 months: 0.773 ± 0.044), followed by F1 score (3 months: 0.759 ± 0.019, 24 months: 0.777 ± 0.039) and area under the receiver operating characteristic curve (3 months: 0.762 ± 0.027, 24 months: 0.773 ± 0.043) at predicting achievement of MCID for neck pain at both follow-up time points, with fair performance. The best precision was also demonstrated by logistic regression at 3 (0.724 ± 0.058) and 24 (0.780 ± 0.097) months. The best recall/sensitivity was demonstrated by multilayer perceptron at 3 months (0.841 ± 0.094) and by extra trees at 24 months (0.817 ± 0.115). Highest specificity was shown by support vector machine at 3 months (0.952 ± 0.013) and by logistic regression at 24 months (0.747 ± 0.18). CONCLUSIONS Appropriate selection of models for studies should be based on the strengths of each model and the aims of the studies. For maximally predicting true achievement of MCID in neck pain, of all the predictions in this balanced data set the appropriate metric for the authors’ study was precision. For both short- and long-term follow-ups, logistic regression demonstrated the highest precision of all models tested. Logistic regression performed consistently the best of all models tested and remains a powerful model for clinical classification tasks.

OBJECTIVE Depression and anxiety are associated with inferior outcomes following spine surgery. In this study, the authors examined whether patients with cervical spondylotic myelopathy (CSM) who have both self-reported depression (SRD) and self-reported anxiety (SRA) have worse postoperative patient-reported outcomes (PROs) compared with patients who have only one or none of these comorbidities. METHODS This study is a retrospective analysis of prospectively collected data from the Quality Outcomes Database CSM cohort. Comparisons were made among patients who reported the following: 1) either SRD or SRA, 2) both SRD and SRA, or 3) neither comorbidity at baseline. PROs at 3, 12, and 24 months (scores for the visual analog scale [VAS] for neck pain and arm pain, Neck Disability Index [NDI], modified Japanese Orthopaedic Association [mJOA] scale, EQ-5D, EuroQol VAS [EQ-VAS], and North American Spine Society [NASS] patient satisfaction index) and achievement of respective PRO minimal clinically important differences (MCIDs) were compared. RESULTS Of the 1141 included patients, 199 (17.4%) had either SRD or SRA alone, 132 (11.6%) had both SRD and SRA, and 810 (71.0%) had neither. Preoperatively, patients with either SRD or SRA alone had worse scores for VAS neck pain (5.6 ± 3.1 vs 5.1 ± 3.3, p = 0.03), NDI (41.0 ± 19.3 vs 36.8 ± 20.8, p = 0.007), EQ-VAS (57.0 ± 21.0 vs 60.7 ± 21.7, p = 0.03), and EQ-5D (0.53 ± 0.23 vs 0.58 ± 0.21, p = 0.008) than patients without such disorders. Postoperatively, in multivariable adjusted analyses, baseline SRD or SRA alone was associated with inferior improvement in the VAS neck pain score and a lower rate of achieving the MCID for VAS neck pain score at 3 and 12 months, but not at 24 months. At 24 months, patients with SRD or SRA alone experienced less change in EQ-5D scores and were less likely to meet the MCID for EQ-5D than patients without SRD or SRA. Furthermore, patient self-reporting of both psychological comorbidities did not impact PROs at all measured time points compared with self-reporting of only one psychological comorbidity alone. Each cohort (SRD or SRA alone, both SRD and SRA, and neither SRD nor SRA) experienced significant improvements in mean PROs at all measured time points compared with baseline (p < 0.05). CONCLUSIONS Approximately 12% of patients who underwent surgery for CSM presented with both SRD and SRA, and 29% presented with at least one symptom. The presence of either SRD or SRA was independently associated with inferior scores for 3- and 12-month neck pain following surgery, but this difference was not significant at 24 months. However, at long-term follow-up, patients with SRD or SRA experienced lower quality of life than patients without SRD or SRA. The comorbid presence of both depression and anxiety was not associated with worse patient outcomes than either diagnosis alone.

Study Design: Prospective observational study, level of evidence 1 for prognostic investigations. Objectives: To evaluate the prevalence of sleep impairment and predictors of improved sleep quality 24 months postoperatively in cervical spondylotic myelopathy (CSM) using the quality outcomes database. Summary of Background Data: Sleep disturbances are a common yet understudied symptom in CSM. Materials and Methods: The quality outcomes database was queried for patients with CSM, and sleep quality was assessed through the neck disability index sleep component at baseline and 24 months postoperatively. Multivariable logistic regressions were performed to identify risk factors of failure to improve sleep impairment and symptoms causing lingering sleep dysfunction 24 months after surgery. Results: Among 1135 patients with CSM, 904 (79.5%) had some degree of sleep dysfunction at baseline. At 24 months postoperatively, 72.8% of the patients with baseline sleep symptoms experienced improvement, with 42.5% reporting complete resolution. Patients who did not improve were more like to be smokers [adjusted odds ratio (aOR): 1.85], have osteoarthritis (aOR: 1.72), report baseline radicular paresthesia (aOR: 1.51), and have neck pain of ≥4/10 on a numeric rating scale. Patients with improved sleep noted higher satisfaction with surgery (88.8% vs 72.9%, aOR: 1.66) independent of improvement in other functional areas. In a multivariable analysis including pain scores and several myelopathy-related symptoms, lingering sleep dysfunction at 24 months was associated with neck pain (aOR: 1.47) and upper (aOR: 1.45) and lower (aOR: 1.52) extremity paresthesias. Conclusion: The majority of patients presenting with CSM have associated sleep disturbances. Most patients experience sustained improvement after surgery, with almost half reporting complete resolution. Smoking, osteoarthritis, radicular paresthesia, and neck pain ≥4/10 numeric rating scale score are baseline risk factors of failure to improve sleep dysfunction. Improvement in sleep symptoms is a major driver of patient-reported satisfaction. Incomplete resolution of sleep impairment is likely due to neck pain and extremity paresthesia.

OBJECTIVE Minimally invasive transforaminal lumbar interbody fusion (MI-TLIF) has been used to treat degenerative lumbar spondylolisthesis and is associated with expedited recovery, reduced operative blood loss, and shorter hospitalizations compared to those with traditional open TLIF. However, the impact of MI-TLIF on long-term patient-reported outcomes (PROs) is less clear. Here, the authors compare the outcomes of MI-TLIF to those of traditional open TLIF for grade I degenerative lumbar spondylolisthesis at 60 months postoperatively. METHODS The authors utilized the prospective Quality Outcomes Database registry and queried for patients with grade I degenerative lumbar spondylolisthesis who had undergone single-segment surgery via an MI or open TLIF method. PROs were compared 60 months postoperatively. The primary outcome was the Oswestry Disability Index (ODI). The secondary outcomes included the numeric rating scale (NRS) for back pain (NRS-BP), NRS for leg pain (NRS-LP), EQ-5D, North American Spine Society (NASS) satisfaction, and cumulative reoperation rate. Multivariable models were constructed to assess the impact of MI-TLIF on PROs, adjusting for variables reaching p < 0.20 on univariable analyses and respective baseline PRO values. RESULTS The study included 297 patients, 72 (24.2%) of whom had undergone MI-TLIF and 225 (75.8%) of whom had undergone open TLIF. The 60-month follow-up rates were similar for the two cohorts (86.1% vs 75.6%, respectively; p = 0.06). Patients did not differ significantly at baseline for ODI, NRS-BP, NRS-LP, or EQ-5D (p > 0.05 for all). Perioperatively, MI-TLIF was associated with less blood loss (108.8 ± 85.6 vs 299.6 ± 242.2 ml, p < 0.001) and longer operations (228.2 ± 111.5 vs 189.6 ± 66.5 minutes, p < 0.001) but had similar lengths of hospitalizations (MI-TLIF 2.9 ± 1.8 vs open TLIF 3.3 ± 1.6 days, p = 0.08). Discharge disposition to home or home health was similar (MI-TLIF 93.1% vs open TLIF 91.1%, p = 0.60). Both cohorts improved significantly from baseline for the 60-month ODI, NRS-BP, NRS-LP, and EQ-5D (p < 0.001 for all comparisons). In adjusted analyses, MI-TLIF, compared to open TLIF, was associated with similar 60-month ODI, ODI change, odds of reaching ODI minimum clinically important difference, NRS-BP, NRS-BP change, NRS-LP, NRS-LP change, EQ-5D, EQ-5D change, and NASS satisfaction (adjusted p > 0.05 for all). The 60-month reoperation rates did not differ significantly (MI-TLIF 5.6% vs open TLIF 11.6%, p = 0.14). CONCLUSIONS For symptomatic, single-level grade I degenerative lumbar spondylolisthesis, MI-TLIF was associated with decreased blood loss perioperatively, but there was no difference in 60-month outcomes for disability, back pain, leg pain, quality of life, or satisfaction between MI and open TLIF. There was no difference in cumulative reoperation rates between the two procedures. These results suggest that in appropriately selected patients, either procedure may be employed depending on patient and surgeon preferences.

OBJECTIVE Patients who undergo surgery for cervical spondylotic myelopathy (CSM) will occasionally develop postoperative neck pain that was not present preoperatively, yet the incidence of this phenomenon is unclear. The authors aimed to elucidate patient and surgical factors associated with new-onset sustained pain after CSM surgery. METHODS The authors reviewed data from the Quality Outcomes Database (QOD) CSM module. The presence of neck pain was defined using the neck pain numeric rating scale (NRS). Patients with no neck pain at baseline (neck NRS score ≤ 1) were then stratified based on the presence of new postoperative pain development (neck NRS score ≥ 2) at 12 and 24 months postoperatively. RESULTS Of 1141 patients in the CSM QOD, 224 (19.6%) reported no neck pain at baseline. Among 170 patients with no baseline neck pain and available 12-month follow-up, 46 (27.1%) reported new postoperative pain. Among 184 patients with no baseline neck pain and available 24-month follow-up, 53 (28.8%) reported new postoperative pain. The mean differences in neck NRS scores were 4.3 for those with new postoperative pain compared with those without at 12 months (4.4 ± 2.2 vs 0.1 ± 0.3, p < 0.001) and 3.9 at 24 months (4.1 ± 2.4 vs 0.2 ± 0.4, p < 0.001). The majority of patients reporting new-onset neck pain reported being satisfied with surgery, but their satisfaction was significantly lower compared with patients without pain at the 12-month (66.7% vs 94.3%, p < 0.001) and 24-month (65.4% vs 90.8%, p < 0.001) follow-ups. The baseline Neck Disability Index (NDI) was an independent predictor of new postoperative neck pain at both the 12-month and 24-month time points (adjusted OR [aOR] 1.04, 95% CI 1.01–1.06; p = 0.002; and aOR 1.03, 95% CI 1.01–1.05; p = 0.026, respectively). The total number of levels treated was associated with new-onset neck pain at 12 months (aOR 1.34, 95% CI 1.09–1.64; p = 0.005), and duration of symptoms more than 3 months was a predictor of 24-month neck pain (aOR 3.22, 95% CI 1.01–10.22; p = 0.048). CONCLUSIONS Increased NDI at baseline, number of levels treated surgically, and duration of symptoms longer than 3 months preoperatively correlate positively with the risk of new-onset neck pain following CSM surgery. The majority of patients with new-onset neck pain still report satisfaction from surgery, suggesting that the risk of new-onset neck pain should not hinder indicated operations from being performed.

OBJECTIVE Prior studies have revealed that a body mass index (BMI) ≥ 30 is associated with worse outcomes following surgical intervention in grade 1 lumbar spondylolisthesis. Using a machine learning approach, this study aimed to leverage the prospective Quality Outcomes Database (QOD) to identify a BMI threshold for patients undergoing surgical intervention for grade 1 lumbar spondylolisthesis and thus reliably identify optimal surgical candidates among obese patients. METHODS Patients with grade 1 lumbar spondylolisthesis and preoperative BMI ≥ 30 from the prospectively collected QOD lumbar spondylolisthesis module were included in this study. A 12-month composite outcome was generated by performing principal components analysis and k-means clustering on four validated measures of surgical outcomes in patients with spondylolisthesis. Random forests were generated to determine the most important preoperative patient characteristics in predicting the composite outcome. Recursive partitioning was used to extract a BMI threshold associated with optimal outcomes. RESULTS The average BMI was 35.7, with 282 (46.4%) of the 608 patients from the QOD data set having a BMI ≥ 30. Principal components analysis revealed that the first principal component accounted for 99.2% of the variance in the four outcome measures. Two clusters were identified corresponding to patients with suboptimal outcomes (severe back pain, increased disability, impaired quality of life, and low satisfaction) and to those with optimal outcomes. Recursive partitioning established a BMI threshold of 37.5 after pruning via cross-validation. CONCLUSIONS In this multicenter study, the authors found that a BMI ≤ 37.5 was associated with improved patient outcomes following surgical intervention. These findings may help augment predictive analytics to deliver precision medicine and improve prehabilitation strategies.

OBJECTIVE The aim of this study was to determine whether multilevel anterior cervical discectomy and fusion (ACDF) or posterior cervical laminectomy and fusion (PCLF) is superior for patients with cervical spondylotic myelopathy (CSM) and high preoperative neck pain. METHODS This was a retrospective study of prospectively collected data using the Quality Outcomes Database (QOD) CSM module. Patients who received a subaxial fusion of 3 or 4 segments and had a visual analog scale (VAS) neck pain score of 7 or greater at baseline were included. The 3-, 12-, and 24-month outcomes were compared for patients undergoing ACDF with those undergoing PCLF. RESULTS Overall, 1141 patients with CSM were included in the database. Of these, 495 (43.4%) presented with severe neck pain (VAS score > 6). After applying inclusion and exclusion criteria, we compared 65 patients (54.6%) undergoing 3- and 4-level ACDF and 54 patients (45.4%) undergoing 3- and 4-level PCLF. Patients undergoing ACDF had worse Neck Disability Index scores at baseline (52.5 ± 15.9 vs 45.9 ± 16.8, p = 0.03) but similar neck pain (p > 0.05). Otherwise, the groups were well matched for the remaining baseline patient-reported outcomes. The rates of 24-month follow-up for ACDF and PCLF were similar (86.2% and 83.3%, respectively). At the 24-month follow-up, both groups demonstrated mean improvements in all outcomes, including neck pain (p < 0.05). In multivariable analyses, there was no significant difference in the degree of neck pain change, rate of neck pain improvement, rate of pain-free achievement, and rate of reaching minimal clinically important difference (MCID) in neck pain between the two groups (adjusted p > 0.05). However, ACDF was associated with a higher 24-month modified Japanese Orthopaedic Association scale (mJOA) score (β = 1.5 [95% CI 0.5–2.6], adjusted p = 0.01), higher EQ-5D score (β = 0.1 [95% CI 0.01–0.2], adjusted p = 0.04), and higher likelihood for return to baseline activities (OR 1.2 [95% CI 1.1–1.4], adjusted p = 0.002). CONCLUSIONS Severe neck pain is prevalent among patients undergoing surgery for CSM, affecting more than 40% of patients. Both ACDF and PCLF achieved comparable postoperative neck pain improvement 3, 12, and 24 months following 3- or 4-segment surgery for patients with CSM and severe neck pain. However, multilevel ACDF was associated with superior functional status, quality of life, and return to baseline activities at 24 months in multivariable adjusted analyses.