Journal of Spinal Disorders and Techniques Publish Ahead of Print DOI:10.1097/BSD.0000000000000183
Perioperative outcomes after cervical laminoplasty versus posterior decompression and fusion: Analysis of 779 patients in the ACS-NSQIP database
Authors: Arya G. Varthi, M.D., Yale University School of Medicine Bryce A. Basques, B.S., Yale University School of Medicine Daniel D. Bohl, M.P.H., Yale University School of Medicine Nicholas S. Golinvaux, B.A., Yale University School of Medicine Jonathan N. Grauer, M.D., Yale University School of Medicine
Corresponding Author: Jonathan N. Grauer, MD Department of Orthopaedics and Rehabilitation, Yale University School of Medicine 800 Howard Avenue, New Haven, CT 06510 Tel: 203-737-7463, Fax: 203-785-7132, Email: [email protected]
Conflicts of Interest and Sources of Funding: No relevant conflicts of interest or sources of funding are reported.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
Study Design: Retrospective cohort study.
Objective: To compare the short-term outcomes for patients undergoing cervical laminoplasty versus posterior decompression and fusion for multilevel cervical pathology.
Summary of Background Data: There is conflicting data regarding the merits of cervical laminoplasty and posterior decompression and fusion for the treatment of multilevel cervical pathology.
Methods: Patients who underwent cervical laminoplasty or posterior decompression and fusion were identified in the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database from 2010 to 2012. Baseline patient characteristics were compared using bivariate logistic regression. Propensity-adjusted multivariate regressions were performed to assess differences in postoperative length of stay, adverse events, and 30-day readmission.
Results: A total of 779 patients were included in this study: 437 (56.1%) underwent cervical decompression and fusion and 342 (43.9%) underwent cervical laminoplasty. Decompression and fusion patients were found to be more comorbid at baseline than laminoplasty patients based
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
on increased American Society of Anesthesiologists (ASA) scores and Charlson Comorbidity Index (CCI). Propensity-adjusted multivariate analysis was used to control for differences in baseline patient characteristics, and found that compared to laminoplasty patients, decompression and fusion patients had increased length of stay (+1.2 days, p < 0.001), greater rates of any adverse event (OR = 1.7, p = 0.018), and were more likely to be readmitted (OR = 2.3, p = 0.028).
Conclusion: Posterior cervical decompression and fusion patients were found to have moderately worse short-term outcomes than laminoplasty patients. The information provided here can be used to inform patients and surgeons about the likely perioperative experience after they have made the decision to pursue one of these two procedures.
Key Words: cervical laminoplasty; cervical decompression and fusion; body mass index; NSQIP; length of stay; complications; readmission
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
Introduction
Cervical spinal stenosis is a concerning pathology that is the most common indication for multilevel posterior cervical decompression procedures.1 Several conditions can initiate a common pathway of facet arthropathy and disc degeneration, resulting in narrowing of the spinal canal. Two of the most common causes of myelopathy are cervical spondylosis and ossification of the posterior longitudinal ligament (OPLL).1 Clinically, patients with cervical myelopathy are often asymptomatic for long periods of time, followed by a stepwise decline in motor function.1 Mild myelopathy patients can be treated conservatively with physical therapy, anti-inflammatory medications and activity modification.2 However, once patients have motor weakness, difficulty with fine motor tasks and/or imaging demonstrating signal change within the spinal cord, surgical treatment is generally recommended.2 In patients with multilevel cervical stenosis and less than 10 degrees of cervical kyphosis, the two most widely accepted surgical options are currently cervical laminoplasty or posterior cervical decompression and fusion.3 There is conflicting data on which procedure is superior in terms of outcomes.4-7 When studying outcomes, there is a clear trend in spine research to turn to national databases to gather information and conduct studies with high patient numbers that are drawn from multiple institutions. The American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) is one such database that abstracts data from representative patients and multiple centers. Patient demographics, characteristics, perioperative course and 30-day follow-up are available for analysis from this database.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
The current study utilizes the ACS-NSQIP database to compare the demographics, adverse events, lengths of stay, and readmission rates for patients with multilevel cervical pathology treated with cervical laminoplasty versus decompression and fusion. Notably, as the ACS-NSQIP does not include information on mechanical instability, pathoanatomical alignment, or pain, there is potential selection bias. However, the results of this study can be used by patients and surgeons to have a more informed discussion about the likely perioperative experience for these two procedures.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
Methods
FDA Approval
In this study, posterior cervical instrumentation is discussed in general terms, however it is not approved by the Food and Drug Administration (FDA) for posterior cervical use.
Data Collection
We conducted a retrospective study using the ACS-NSQIP database. Patients who underwent cervical laminoplasty and multilevel posterior cervical decompression and fusion from 2010 to 2012 were selected using Current Procedural Terminology (CPT) codes. Laminoplasty patients were selected using CPT codes 63050 (“Laminoplasty, cervical, with decompression of the spinal cord, 2 or more vertebral segments”) and 63051 (“Laminoplasty, cervical, with decompression of the spinal cord, 2 or more vertebral segments; with reconstruction of the posterior bony elements [including the application of bridging bone graft and non-segmental fixation devices]”). Posterior cervical decompression patients were selected by the primary CPT codes 63015 (“Laminectomy with exploration and/or decompression of spinal cord and/or cauda equina, without facetectomy, foraminotomy or discectomy [eg, spinal stenosis], more than 2 vertebral segments; cervical“), 63045 (“Laminectomy, facetectomy, and foraminotomy [unilateral or bilateral with decompression of spinal cord, cauda equina and/or nerve root(s) ), (eg, spinal or lateral recess stenosis)], single vertebral segment; cervical”). From the above posterior cervical decompression cohort,
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
multilevel posterior cervical instrumentation and fusion patients were isolated using CPT code 22842 (“Posterior segmental instrumentation [eg, pedicle fixation, dual rods with multiple hooks and sublaminar wires]; 3 to 6 vertebral segments”). Patients undergoing urgent or emergent surgery or those with previous evidence of infection were excluded from the analysis. Among the variables available in ACS-NSQIP are patient characteristics including sex, age, height, and weight. Body mass index (BMI) was calculated from each patient’s height and weight. The ACS-NSQIP also includes information on medical comorbidities and American Society of Anesthesiologists (ASA) class. A modified Charlson comorbidity index (CCI) 8 was calculated for each patient based on the available comorbidity data. Such modified CCIs have been shown to be similar in efficacy to the original CCI,9,10 and the modified CCI employed in this study has been previously used with the ACS-NSQIP. 11 The comorbidities used to determine the modified CCI were directly available in the dataset and included (followed by corresponding point values): myocardial infarction (1), congestive heart failure (1), peripheral vascular disease or rest pain (1), transient ischemic attack or cerebrovascular accident (1), chronic obstructive pulmonary disease (1), diabetes mellitus (1), hemiplegia (2), end stage renal disease (2), ascites or esophageal varices (3), and cancer (6). The point values were summed for a total number, to which one point was added for each decade greater than 40 years of age.
Adverse Events
The ACS-NSQIP tracks patients for the occurrence of 23 individual adverse events that occurred within the first 30 postoperative days (including while in the hospital as well as after
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
discharge). The occurrence of any adverse event (AAE) was defined as the occurrence of any of the following: death, coma > 24 hours, on ventilator > 48 hours, unplanned intubation, stroke/cerebrovascular accident, thromboembolic event (deep venous thrombosis or pulmonary embolism), cardiac arrest, myocardial infarction, acute renal failure, sepsis, wound disruption, deep surgical site infection, return to the operating room, superficial surgical site infection, urinary tract infection (UTI), pneumonia, blood transfusion, progressive renal insufficiency, or peripheral nerve injury.
Postoperative length of stay
Postoperative length of stay was defined in the ACS-NSQIP as calendar days from operation to hospital discharge. Length of stay was treated as a continuous variable for analysis.
Readmission
Thirty-day readmission data in the NSQIP was first collected in 2011, so only patients who underwent surgery in 2011 or 2012 were included in the readmission analysis. For this study, readmission was defined as positive when a patient had an unplanned readmission one or more times within the first 30 postoperative days. In 2012, the ACS-NSQIP began collecting reasons for readmission and time to readmission.
Analysis
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
Statistical analyses were conducted using STATA® version 11.2 (StataCorp, LP, College Station, Texas, USA). All tests were two-tailed and the statistical difference was established at a two-sided α level of 0.05 (p < 0.05). Patients treated with laminoplasty or decompression and fusion were first compared by demographic and comorbidity variables using bivariate logistic regression. In order to control for selection bias between the laminoplasty and decompression and fusion procedure groups, propensity scores were used, with the propensity score defined as the conditional probability of receiving decompression and fusion based on the observed patient demographics and comorbidities. The propensity score has been extensively used in the literature to control for selection bias in nonrandomized studies.12-14 Propensity scores were calculated via multivariate logistic regression and propensity-adjusted p-values for preoperative patient demographics were reported. While propensity scores adjust for differences in measured variables between groups, it should be noted that they cannot adjust for differences in unmeasured variables such as mechanical instability, pathoanatomical alignment, or pain; hence, results could still be biased by these factors. There were three primary hypotheses in this study: That there was a difference between groups in (1) postoperative length of stay, (2) rate of any adverse events, and (3) rate of readmission. As these were the primary hypothesis tests, they were tested at the α = 0.05 significance level. The rates of any adverse events that occurred in laminoplasty and decompression and fusion cases were compared using bivariate and propensity-adjusted multivariate logistic regression, using laminoplasty cases as the reference. Propensity-adjusted multivariate logistic regression adjusted for age, sex, body mass index, ASA class, modified CCI, and selection bias between treatment groups. This procedure was then repeated for each of the individual adverse events; however, only individual adverse events with at least one event in
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
each of the laminoplasty and decompression and fusion groups were compared, and the Bonferroni adjustment to the p-value required for statistical significance was used for the multiple comparisons being conducted on the individual adverse events. For the 12 distinct individual adverse events being compared, the Bonferroni correction lowered the p-value cutoff for statistical significance for these tests from p
Perioperative outcomes after cervical laminoplasty versus posterior decompression and fusion: Analysis of 779 patients in the ACS-NSQIP database
Authors: Arya G. Varthi, M.D., Yale University School of Medicine Bryce A. Basques, B.S., Yale University School of Medicine Daniel D. Bohl, M.P.H., Yale University School of Medicine Nicholas S. Golinvaux, B.A., Yale University School of Medicine Jonathan N. Grauer, M.D., Yale University School of Medicine
Corresponding Author: Jonathan N. Grauer, MD Department of Orthopaedics and Rehabilitation, Yale University School of Medicine 800 Howard Avenue, New Haven, CT 06510 Tel: 203-737-7463, Fax: 203-785-7132, Email: [email protected]
Conflicts of Interest and Sources of Funding: No relevant conflicts of interest or sources of funding are reported.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
Study Design: Retrospective cohort study.
Objective: To compare the short-term outcomes for patients undergoing cervical laminoplasty versus posterior decompression and fusion for multilevel cervical pathology.
Summary of Background Data: There is conflicting data regarding the merits of cervical laminoplasty and posterior decompression and fusion for the treatment of multilevel cervical pathology.
Methods: Patients who underwent cervical laminoplasty or posterior decompression and fusion were identified in the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database from 2010 to 2012. Baseline patient characteristics were compared using bivariate logistic regression. Propensity-adjusted multivariate regressions were performed to assess differences in postoperative length of stay, adverse events, and 30-day readmission.
Results: A total of 779 patients were included in this study: 437 (56.1%) underwent cervical decompression and fusion and 342 (43.9%) underwent cervical laminoplasty. Decompression and fusion patients were found to be more comorbid at baseline than laminoplasty patients based
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
on increased American Society of Anesthesiologists (ASA) scores and Charlson Comorbidity Index (CCI). Propensity-adjusted multivariate analysis was used to control for differences in baseline patient characteristics, and found that compared to laminoplasty patients, decompression and fusion patients had increased length of stay (+1.2 days, p < 0.001), greater rates of any adverse event (OR = 1.7, p = 0.018), and were more likely to be readmitted (OR = 2.3, p = 0.028).
Conclusion: Posterior cervical decompression and fusion patients were found to have moderately worse short-term outcomes than laminoplasty patients. The information provided here can be used to inform patients and surgeons about the likely perioperative experience after they have made the decision to pursue one of these two procedures.
Key Words: cervical laminoplasty; cervical decompression and fusion; body mass index; NSQIP; length of stay; complications; readmission
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
Introduction
Cervical spinal stenosis is a concerning pathology that is the most common indication for multilevel posterior cervical decompression procedures.1 Several conditions can initiate a common pathway of facet arthropathy and disc degeneration, resulting in narrowing of the spinal canal. Two of the most common causes of myelopathy are cervical spondylosis and ossification of the posterior longitudinal ligament (OPLL).1 Clinically, patients with cervical myelopathy are often asymptomatic for long periods of time, followed by a stepwise decline in motor function.1 Mild myelopathy patients can be treated conservatively with physical therapy, anti-inflammatory medications and activity modification.2 However, once patients have motor weakness, difficulty with fine motor tasks and/or imaging demonstrating signal change within the spinal cord, surgical treatment is generally recommended.2 In patients with multilevel cervical stenosis and less than 10 degrees of cervical kyphosis, the two most widely accepted surgical options are currently cervical laminoplasty or posterior cervical decompression and fusion.3 There is conflicting data on which procedure is superior in terms of outcomes.4-7 When studying outcomes, there is a clear trend in spine research to turn to national databases to gather information and conduct studies with high patient numbers that are drawn from multiple institutions. The American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) is one such database that abstracts data from representative patients and multiple centers. Patient demographics, characteristics, perioperative course and 30-day follow-up are available for analysis from this database.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
The current study utilizes the ACS-NSQIP database to compare the demographics, adverse events, lengths of stay, and readmission rates for patients with multilevel cervical pathology treated with cervical laminoplasty versus decompression and fusion. Notably, as the ACS-NSQIP does not include information on mechanical instability, pathoanatomical alignment, or pain, there is potential selection bias. However, the results of this study can be used by patients and surgeons to have a more informed discussion about the likely perioperative experience for these two procedures.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
Methods
FDA Approval
In this study, posterior cervical instrumentation is discussed in general terms, however it is not approved by the Food and Drug Administration (FDA) for posterior cervical use.
Data Collection
We conducted a retrospective study using the ACS-NSQIP database. Patients who underwent cervical laminoplasty and multilevel posterior cervical decompression and fusion from 2010 to 2012 were selected using Current Procedural Terminology (CPT) codes. Laminoplasty patients were selected using CPT codes 63050 (“Laminoplasty, cervical, with decompression of the spinal cord, 2 or more vertebral segments”) and 63051 (“Laminoplasty, cervical, with decompression of the spinal cord, 2 or more vertebral segments; with reconstruction of the posterior bony elements [including the application of bridging bone graft and non-segmental fixation devices]”). Posterior cervical decompression patients were selected by the primary CPT codes 63015 (“Laminectomy with exploration and/or decompression of spinal cord and/or cauda equina, without facetectomy, foraminotomy or discectomy [eg, spinal stenosis], more than 2 vertebral segments; cervical“), 63045 (“Laminectomy, facetectomy, and foraminotomy [unilateral or bilateral with decompression of spinal cord, cauda equina and/or nerve root(s) ), (eg, spinal or lateral recess stenosis)], single vertebral segment; cervical”). From the above posterior cervical decompression cohort,
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
multilevel posterior cervical instrumentation and fusion patients were isolated using CPT code 22842 (“Posterior segmental instrumentation [eg, pedicle fixation, dual rods with multiple hooks and sublaminar wires]; 3 to 6 vertebral segments”). Patients undergoing urgent or emergent surgery or those with previous evidence of infection were excluded from the analysis. Among the variables available in ACS-NSQIP are patient characteristics including sex, age, height, and weight. Body mass index (BMI) was calculated from each patient’s height and weight. The ACS-NSQIP also includes information on medical comorbidities and American Society of Anesthesiologists (ASA) class. A modified Charlson comorbidity index (CCI) 8 was calculated for each patient based on the available comorbidity data. Such modified CCIs have been shown to be similar in efficacy to the original CCI,9,10 and the modified CCI employed in this study has been previously used with the ACS-NSQIP. 11 The comorbidities used to determine the modified CCI were directly available in the dataset and included (followed by corresponding point values): myocardial infarction (1), congestive heart failure (1), peripheral vascular disease or rest pain (1), transient ischemic attack or cerebrovascular accident (1), chronic obstructive pulmonary disease (1), diabetes mellitus (1), hemiplegia (2), end stage renal disease (2), ascites or esophageal varices (3), and cancer (6). The point values were summed for a total number, to which one point was added for each decade greater than 40 years of age.
Adverse Events
The ACS-NSQIP tracks patients for the occurrence of 23 individual adverse events that occurred within the first 30 postoperative days (including while in the hospital as well as after
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
discharge). The occurrence of any adverse event (AAE) was defined as the occurrence of any of the following: death, coma > 24 hours, on ventilator > 48 hours, unplanned intubation, stroke/cerebrovascular accident, thromboembolic event (deep venous thrombosis or pulmonary embolism), cardiac arrest, myocardial infarction, acute renal failure, sepsis, wound disruption, deep surgical site infection, return to the operating room, superficial surgical site infection, urinary tract infection (UTI), pneumonia, blood transfusion, progressive renal insufficiency, or peripheral nerve injury.
Postoperative length of stay
Postoperative length of stay was defined in the ACS-NSQIP as calendar days from operation to hospital discharge. Length of stay was treated as a continuous variable for analysis.
Readmission
Thirty-day readmission data in the NSQIP was first collected in 2011, so only patients who underwent surgery in 2011 or 2012 were included in the readmission analysis. For this study, readmission was defined as positive when a patient had an unplanned readmission one or more times within the first 30 postoperative days. In 2012, the ACS-NSQIP began collecting reasons for readmission and time to readmission.
Analysis
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
Statistical analyses were conducted using STATA® version 11.2 (StataCorp, LP, College Station, Texas, USA). All tests were two-tailed and the statistical difference was established at a two-sided α level of 0.05 (p < 0.05). Patients treated with laminoplasty or decompression and fusion were first compared by demographic and comorbidity variables using bivariate logistic regression. In order to control for selection bias between the laminoplasty and decompression and fusion procedure groups, propensity scores were used, with the propensity score defined as the conditional probability of receiving decompression and fusion based on the observed patient demographics and comorbidities. The propensity score has been extensively used in the literature to control for selection bias in nonrandomized studies.12-14 Propensity scores were calculated via multivariate logistic regression and propensity-adjusted p-values for preoperative patient demographics were reported. While propensity scores adjust for differences in measured variables between groups, it should be noted that they cannot adjust for differences in unmeasured variables such as mechanical instability, pathoanatomical alignment, or pain; hence, results could still be biased by these factors. There were three primary hypotheses in this study: That there was a difference between groups in (1) postoperative length of stay, (2) rate of any adverse events, and (3) rate of readmission. As these were the primary hypothesis tests, they were tested at the α = 0.05 significance level. The rates of any adverse events that occurred in laminoplasty and decompression and fusion cases were compared using bivariate and propensity-adjusted multivariate logistic regression, using laminoplasty cases as the reference. Propensity-adjusted multivariate logistic regression adjusted for age, sex, body mass index, ASA class, modified CCI, and selection bias between treatment groups. This procedure was then repeated for each of the individual adverse events; however, only individual adverse events with at least one event in
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of the article is prohibited.
each of the laminoplasty and decompression and fusion groups were compared, and the Bonferroni adjustment to the p-value required for statistical significance was used for the multiple comparisons being conducted on the individual adverse events. For the 12 distinct individual adverse events being compared, the Bonferroni correction lowered the p-value cutoff for statistical significance for these tests from p
