Safety of Outpatient Single-Level Cervical Total Disc Replacement

A Propensity-Matched Multi-Institutional Study

Patawut Bovonratwet, BS; Michael C. Fu, MD, MHS; Vineet Tyagi, MD; Nathaniel T. Ondeck, MD, MHS; Todd J. Albert, MD; Jonathan N. Grauer, MD

Disclosures

Spine. 2019;44(9):E530-E538. 

In This Article

Results

Patient Population

After applying the inclusion and exclusion criteria, a total of 1985 patients who underwent outpatient or inpatient single-level CTDR were identified for analysis. Of these cases, 373 (18.79%) were performed as outpatient procedures and 1612 (81.21%) were performed as inpatient procedures (Table 1).

Patients who underwent outpatient CTDR procedures tended to be male (58.18% vs. 52.48%, P = 0.047) and had a lower percentage of patients with hypertension (18.50% vs.24.13%, P = 0.020) (Table 1, central columns). In addition, although not statistically significant, patients in the outpatient CTDR cohort tended to have a higher percentage of ASA ≤ 2 (83.38% vs.79.15%, P = 0.069). These variances in patient baseline characteristics may be the result of a selection bias for those considered for outpatient CTDR procedures. Propensity score matching was employed to reduce this bias in patient selection. Once patients were matched, there were no longer any statistical differences in patient baseline characteristics and comorbidities (P > 0.05 for all) (Table 1, right columns).

Perioperative Complications and 30-day Readmission

Before propensity score matching, there were no differences in rates of both individual complications and aggregated adverse events that occurred anytime during the 30-day perioperative period between outpatient and inpatient CTDR (P > 0.004 for all) (Table 2). There were no occurrences of wound complications in the outpatient cohort, while there were seven occurrences of wound complications (five superficial wound infection or dehiscence and two deep surgical site infection or organ space surgical site infection) in the inpatient cohort, but this was not statistically significant (P = 0.360). In addition, there were no differences in rates of individual or aggregated adverse events that occurred postdischarge between the two cohorts (P > 0.004 for all) (Table 3 and Table 4). Of note, there was no difference in the rate of 30-day readmissions between the two groups (Table 2).

After propensity score matching was performed, similar results were observed. There were again no differences in rates of both individual complications and aggregated adverse events that occurred anytime during the 30-day perioperative period (P > 0.004 for all) (Table 2 and Table 4). There were again no occurrences of wound complications in the outpatient cohort, while there were five occurrences of wound complications (three superficial wound infection or dehiscence and two deep surgical site infection or organ space surgical site infection) in the matched inpatient cohort, but this was not statistically significant (P = 0.340). The reasons for return to the operating room in the outpatient CTDR cohort include repair of retinal detachment (CPT code 67108), which is unrelated to the index surgical procedure, and displacement of cervical intervertebral disc without myelopathy (ICD-9: 722.0). Further, there were again no differences in rates of individual or aggregated adverse events that occurred postdischarge between the two matched cohorts (P > 0.004 for all) (Table 3 and Table 4). Similarly, there was no difference in the rate of 30-day readmissions between the two matched cohorts (Table 2). The reasons for the two readmissions in the outpatient CTDR cohort include displacement of cervical intervertebral disc without myelopathy (ICD-9: 722.0) and brachial neuritis or radiculitis (ICD-9: 723.4).

Relative Risks for Perioperative Aggregated Adverse Events

In the analysis of aggregated adverse events that occurred anytime during the 30-day perioperative period, Poisson regressions with robust error variances revealed no differences in risk for SAEs [RR = 0.818, 99.60% confidence interval (CI) = 0.126–5.295, P = 0.757] and AAEs [RR = 0.450, 99.60% CI = 0.076–2.653, P = 0.195] between the outpatient CTDR group relative to the matched inpatient CTDR group (Figure 1). RR for MAEs was not calculated, as no MAEs occurred in the outpatient cohort.

Figure 1.

Adverse events listed at the left. Diamonds central to the horizontal lines indicate relative risks. Horizontal lines denote the 99.60% confidence intervals of those relative risks, per Bonferroni correction. The vertical line indicates a relative risk of 1. Therefore, horizontal lines that cross the vertical line indicate relative risks that are not statistically significant. Relative risk for MAE is not calculated, as no MAE occurred in the outpatient CTDR group.

In the analysis of aggregated adverse events that occurred following hospital discharge, Poisson regressions with robust error variances revealed no differences in risk for SAEs (RR = 3.000, 99.60% CI = 0.169–53.122, P = 0.271), AAEs (RR = 0.750, 99.60% CI = 0.078–7.257, P = 0.715), and 30-day readmission (RR = 0.830, 99.60% CI = 0.083–8.288, P = 0.815) between the outpatient CTDR group relative to the matched inpatient CTDR group (Figure 2). RR for postdischarge MAEs was not calculated, as no postdischarge MAEs occurred in the outpatient cohort.

Figure 2.

Adverse events listed at the left. Diamonds central to the horizontal lines indicate relative risks. Horizontal lines denote the 99.60% confidence intervals of those relative risks, per Bonferroni correction. The vertical line indicates a relative risk of 1. Therefore, horizontal lines that cross the vertical line indicate relative risks that are not statistically significant. Relative risk for MAE is not calculated, as no MAE occurred in the outpatient CTDR group.

Subanalysis Comparing Perioperative Complications of Outpatient CTDR Versus Outpatient Single-level ACDF

After the inclusion and exclusion criteria, a total of 2616 patients who underwent outpatient single-level ACDF were identified as a further comparison group. Of these, 1119 patients were then propensity score matched to 373 outpatient single-level CTDR cases. There were no differences in any patient baseline characteristics between these two matched groups. Lastly, statistical analysis of matched cohorts revealed no significant differences in rates of both individual and aggregated adverse events between outpatient single-level CTDR versus matched outpatient single-level ACDF.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....