Elsevier

World Neurosurgery

Volume 138, June 2020, Pages e26-e34
World Neurosurgery

Original Article
Elixhauser Comorbidity Measure is Superior to Charlson Comorbidity Index In-Predicting Hospital Complications Following Elective Posterior Cervical Decompression and Fusion

https://doi.org/10.1016/j.wneu.2020.01.141Get rights and content

Objective

The predictive ability of Elixhauser Comorbidity Index (ECI) and Charlson Comorbidity Index (CCI) have been compared in orthopedic and gastrointestinal surgery; however, their predictive ability for complications secondary to spine surgery and posterior cervical decompression and fusion (PCDF) specifically is understudied. This study examines the predictive ability of the ECI and CCI for complications and morbidity following PCDF.

Methods

ECI and CCI were retrospectively computed for all PCDF cases in the National Inpatient Sample database from 2013 to 2014 and complications or morbidity were identified. C-statistics were used to analyze ECI and CCI predictive ability in a range of complications and compared with a base comorbidity model that included age, sex, race, and primary payer.

Results

PCDF was performed in 46,700 hospitalizations between 2013 and 2014. The complications for which ECI was found to be a significantly better predictor included airway complications (69.16% superior to CCI), hemorrhagic anemia (79.04% superior), cardiac arrest (72.39% superior), pulmonary embolism (83.01% superior), sepsis (62.44% superior), septic shock (78.90% superior), urinary tract infection (63.53% superior), death (74.28% superior), any minor complication (75% superior), any major complication (133% superior), and any complication at all (63.72% superior). The complications for which neither the ECI Index nor the CCI proved superior were acute kidney injury, myocardial infarction, cerebrovascular accident, deep vein thrombosis, pneumonia, wound dehiscence, and superficial surgical-site infection following PCDF.

Conclusions

ECI showed superior predictive ability to the CCI in predicting 8 of the 18 complications that were analyzed and inferior in none.

Introduction

Posterior cervical decompression and fusion (PCDF) is a surgery used to treat a range of cervical spine pathologies, including herniation refractory to conservative treatment, spondylosis, cervical fractures, some malignancies, and degenerative disk disease.1, 2, 3 PCDF has been indicated in an increasing number of patients, including those of advanced age and with more significant comorbidity burdens.4,5 A recent study showed that PCDF accounted for 12.35% of all cervical fusions cases.6 Specifically, the posterior approach is indicated for cases with posterior lesions compressing the spinal cord, lesions extending past 3-disc levels, or those with the presence of significant dural ossification.1 Predicting postoperative complications is of particular importance because the posterior approach has been shown to have greater in-hospital and 30-day complication rates in comparison with the anterior approach.6,7 With a theoretically greater risk of wound dehiscence or surgical-site infection (SSI) because the patient may have a harder time protecting surgical wounds on the back of the neck, predicting these and other dangerous medical complications is essential for optimizing outcomes in this surgical population.8 Finding ways to predict surgical complications following cervical fusion is important for assessing provider performance and patient outcomes as well as for providing critical information as it pertains to reimbursement schedules and cost management for the health system as a whole.9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19

Nationwide, cervical fusion procedures have become significantly more common in recent years, with about 400,000 being done each year between 2000 and 2010, incurring costs in the hundreds of millions of dollars.20 As the costs have risen, PCDF and cervical fusion surgery in general have been identified by the Center for Medicare and Medicaid Services (CMS) as a potential focus for implementation of mandatory bundled payment policies in an attempt to shift the basis for reimbursement from quantity of surgeries to quality of outcomes.21,22 Part of the formula for reducing cost, of course, lies in reducing length of stay, emergency department visits, and readmissions. However, predicting risk of postoperative complications is also crucial.

As it stands today, the 2 most prominent indices for comorbidity burden used in surgical research are the Elixhauser Comorbidity Index (ECI) and the Charlson Comorbidity Index (CCI). The CCI was first developed in 1987 and uses 19 different inputs to provide a score quantifying comorbidity burden. It is employed widely in spine surgery.18,23,24 It has been shown to be an effective predictor of both patient mortality and healthcare expenditure.24, 25, 26, 27, 28 ECI is a more recent index, introduced in 1998, and is composed of 31 comorbidities, of which only 7 are shared in common with CCI.29 Publications by Menendez et al.26,30 have shown that the ECI outperforms CCI in predicting inpatient mortality for both cervical spine fractures and other orthopedic operations. However, these studies have not teased out the predictive ability of the ECI versus CCI for perioperative complications in spine surgery, nor in PCDF specifically. Yet, it follows that different procedures may have drastically different comorbidity–complication relationships. Therefore, this study endeavors to use nationally representative data to quantify the predictive ability of ECI and CCI for PCDF perioperative complications.

Section snippets

Methods

This study was completed using the Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS) data from 2013 to 2014. The database was queried for any case in which patients received primary PCDF, as classified by the International Classification of Diseases, 9th Revision (ICD-9) with the code 81.03. Exclusion criteria included having an anterior approach with the ICD-9 code 81.02. For each of the cases, ECI and CCI component comorbidities were documented by matching to the

Results

The query of the NIS database yielded an estimated 46,700 hospitalizations for PCDF between 2013 and 2014. In total, 36,800 patients were found to have no complications, whereas 9900 had at least 1 complication. The 2 cohorts were not significantly different in terms of sex. Significant differences between the 2 cohorts were found in age, race, payer, length of stay, Elixhauser Comorbidity Score, and Charlson Comorbidity Score. On average, patients with a complication were 0.53 years older (P <

Discussion

Our study found that Elixhauser was a significantly better predictor in 11 of the 18 categories studied. Further, it was found to be an excellent predictor in 10 of 18 categories, acceptable in 4, and poor in 4 (hemorrhagic anemia, pneumonia, superficial SSI, and any minor complication). The difficulty with predicting minor complications (AUC of 0.70) likely was due to poor predictive ability for hemorrhagic anemia and superficial SSI. However, there was no category in which CCI or baseline

Conclusions

In this study, we demonstrate that ECI is superior to CCI in 8 of 18 categories and inferior in none. Elixhauser is a significant predictor of complications in 14 of 18 subanalyses. Thus, at a time when CMS is considering bundled payments for cervical spine fusions, a patient's ECI should be considered. Further study is needed to analyze the relative value of each of the 31 Elixhauser comorbidities and examine the predictive ability of these models on health care cost data.

CRediT authorship contribution statement

Samuel Z. Maron: Conceptualization, Investigation, Resources, Writing - original draft. Sean N. Neifert: Writing - review & editing, Visualization, Software, Formal analysis. William A. Ranson: Writing - review & editing, Conceptualization. Dominic A. Nistal: Writing - review & editing, Resources. Robert J. Rothrock: Writing - review & editing, Resources. Peter Cooke: Writing - review & editing, Resources. Colin D. Lamb: Writing - review & editing, Project administration. Samuel K. Cho:

References (45)

  • C.A. Kuhns et al.

    An outcomes analysis of the treatment of cervical pseudarthrosis with posterior fusion

    Spine (Phila Pa)

    (2005)
  • T. Sugawara

    Anterior cervical spine surgery for degenerative disease: a review

    Neurol Med Chir (Tokyo)

    (2015)
  • K.J. Song et al.

    Current concepts of anterior cervical discectomy and fusion: a review of literature

    Asian Spine J

    (2014)
  • J. Di Capua et al.

    Elderly age as a risk factor for 30-day postoperative outcomes following elective anterior cervical discectomy and fusion

    Glob Spine J

    (2017)
  • C. Weiler et al.

    Age-related changes in human cervical, thoracal and lumbar intervertebral disc exhibit a strong intra-individual correlation

    Eur Spine J

    (2012)
  • J.K. Yue et al.

    Risk factors for 30-day outcomes in elective anterior versus posterior cervical fusion: a matched cohort analysis

    J Craniovertebr Junction Spine

    (2017)
  • S.G. Memtsoudis et al.

    Increased in-hospital complications after primary posterior versus primary anterior cervical fusion

    Clin Orthop

    (2011)
  • J.P.Y. Cheung et al.

    Complications of anterior and posterior cervical spine surgery

    Asian Spine J

    (2016)
  • M.A. Cairns et al.

    Refining risk adjustment for the proposed CMS surgical hip and femur fracture treatment bundled payment program

    J Bone Joint Surg Am

    (2018)
  • M. Pine et al.

    Enhancement of claims data to improve risk adjustment of hospital mortality

    JAMA

    (2007)
  • K. Phan et al.

    Effect of preoperative anemia on the outcomes of anterior cervical discectomy and fusion

    Glob Spine J

    (2017)
  • A.S. Narain et al.

    Impact of body mass index on surgical outcomes, narcotics consumption, and hospital costs following anterior cervical discectomy and fusion

    J Neurosurg Spine

    (2018)
  • Cited by (29)

    • Surgical site infection in thoracic and lumbar fractures: incidence and risk factors in 11,401 patients from a nationwide administrative database

      2023, Spine Journal
      Citation Excerpt :

      Multiple studies have shown ECI to be a better predictor than CCI of inpatient mortality [32] and perioperative complications [33,34] following cervical spine surgery. Interestingly, ECI was not found to be a good predictor of superficial SSI in a study on elective posterior cervical decompression fusion patients [34]. Fourteen of the 17 CCI comorbidity categories (not including age, myocardial infarction, and leukemia) are included in ECI.

    • Alcohol Abuse, Morbid Obesity, Depression, Congestive Heart Failure, and Chronic Pulmonary Disease are Risk Factors for 90-Day Readmission After Arthroscopic Rotator Cuff Repair

      2022, Arthroscopy, Sports Medicine, and Rehabilitation
      Citation Excerpt :

      Pearson’s chi-square was used to compare patient demographics, and multivariate binomial logistic regression was used to calculate odds ratios (OR) on patient-related risk factors for 90-day readmissions. Risk factors assessed were those comorbidities, which comprise the Elixhauser-Comorbidity Index (ECI), which has been shown to be superior to the Charlson-Comorbidity Index (CCI) for observational studies (Table 1).14 A P value less than .05 was considered statistically significant.

    View all citing articles on Scopus

    Conflict of interest statement: John M. Caridi has received consulting fees from Zimmer Biomet. All other authors have no disclosures.

    View full text