2.1 Study design

This retrospective database review was performed using the commercially available PearlDiver (PearlDiver Inc., Colorado Springs, Colorado, USA; www.pearldiverinc.com) patient records database. This study was exempt from institutional review board approval, as all queried data was deidentified and Health Insurance Portability and Accountability Act (HIPAA) compliant.

2.2 Setting

The database contains all Mariner private payer, Medicare, and Medicaid patients records for the years 2010–2022, searchable by International Classification of Diseases (ICD) 9th and 10th Edition codes as well as by Current Procedural Terminology (CPT) codes.

2.3 Participants

The study population included patients who underwent a one-level CDR (CPT-22856, CPT-0090T) or one-level PCF (CPT-63020, CPT-63045) from 2017 to 2022. Only patients who had a diagnosis of cervical radiculopathy (ICD-9-D-7234, ICD-9-D-7292, ICD-10-D-M4722, ICD-10-D-M4723, ICD-10-D-M5010, ICD-10-D-M5011, ICD-10-D-M5012, ICD-10-D-M50121, ICD-10-D-M50122, ICD-10-D-M50123, ICD-10-D-M5013, ICD-10-D-M5412, ICD-10-D-M5413) concurrently or within one year prior to their CDR or PCF procedure were included. Patients who had diagnoses of cervical myelopathy, additional vertebral segment decompression, previous surgeries, fractures, trauma, or infection were excluded. The CDR cohort was matched to the PCF cohort in a 1:1 ratio by age range, gender, and Charlson Comorbidity Index (CCI) score. All patients had a minimum of two years follow-up.

2.4 Variables

Patients in both groups were evaluated for 90-day medical complications: arrhythmia, myocardial infarction, cardiac arrest, atelectasis, respiratory failure, pleural effusion, pulmonary embolism, deep vein thrombosis, cerebrovascular accident, delirium, seizure, ileus, perforated ulcer, electrolyte imbalance, renal failure, urinary retention, pneumonia, sepsis, and urinary tract infection. Surgical complications included dysphagia, dysphonia, dural tear, nerve root injury, vertebral artery injury, transfusion, nerve root motor deficit, nerve root sensory deficit, cervical radiculopathy, spinal cord deficit, seroma, hematoma, deep wound infection, and superficial wound infection. 90-day emergency department visit, hospital readmission, and cervical reoperation surgery rates were also assessed.

2.5 Data sources/management

All variables were derived from uniformly coded claims within the PearlDiver Mariner database, ensuring consistent ascertainment across cohorts. Diagnostic and procedural codes were used to define inclusion criteria, exposures, and outcomes, allowing direct comparability of complication, revision, and utilization measures between CDR and PCF groups.

2.6 Bias

To minimize selection bias, cohorts were matched 1:1 by age, sex, and CCI. Standardized ICD and CPT coding ensured consistent case identification, and multivariable regression was used to adjust for residual confounders such as smoking and depression.

2.7 Study size

All eligible patients meeting inclusion criteria between 2017 and 2022 were captured from the PearlDiver database. After applying exclusions and performing 1:1 matching, the final study population consisted of 9008 patients (4504 per group).

2.8 Quantitative variables

Continuous variables such as age and CCI were analyzed as means with standard deviations, while categorical comorbidities were expressed as frequencies and percentages. BMI was grouped into clinically relevant categories (30–40 and > 40) to reflect obesity risk stratification.

2.9 Statistical method

Pearson χ2 test was used to assess for differences in demographics and preexisting comorbidities. Multivariable logistic regression was used to determine the independent effects of CDR and PCF on the postoperative outcomes after adjusting for demographic factors and pertinent comorbidities. All statistical analyses were conducted using the research query interface provided by PearlDiver Bellwether. The common statistical metrics, including frequencies, mean values, odds ratios (ORs), and 95 % confidence interval (CI), were collected by the Bellwether system using the R statistical package. Statistical significance was set at p < 0.05.

3.1 Participants

In this study, a total of 9008 patients were studied, split evenly between the CDR and PCF groups.

3.2 Descriptive data

The patient demographics were examined and compared in Table 1. Demographic variables including age, sex, and comorbidities including BMI, smoking history, chronic obstructive pulmonary disease, peripheral vascular disease, diabetes mellitus, hyperlipidemia, hypertension, hypothyroidism, congestive heart failure, coronary artery disease, renal disease, depression, and dementia are presented in Table 1. Of note, smoking was significantly less prevalent within patients in the CDR group (25.0 %) than PCF group (25 % vs 28 %, p = 0.002). Depression rates also exhibited a non-statistically significant decrease in the CDR group (33.1 %) than the PCF group (33.1 % vs 35.0 %, p = 0.056).

3.3 Main results

There were no statistically significant differences in medical complications between the two groups. Table 2 demonstrates a multivariate logistic regression analysis of 90-day medical complications between the CDR and PCF groups, adjusted for smoking and depression status.

Regarding 90-day surgical complications, there were multiple statistically significant values determined. Rate of dysphagia was significantly higher in the CDR group than the PCF group (2.2 % vs 1.1 %, OR = 2.27; 95 % CI: 1.59–3.28; p < 0.001). The rate of dysphonia was also significantly higher in the CDR group (0.4 %) than the PCF group (0.2 %), with an adjusted OR of 3.95 (95 % CI: 1.83–9.81; p = 0.001).

The rate of dural tear was significantly lower in the CDR group (0.0 %) than the PCF group (0.3 %), with an adjusted OR of 0.13 (95 % CI: 0.02–0.45; p = 0.006). Rate of wound dehiscence was also significantly lower in the CDR group (0.5 %) than the PCF group (1.4 %), with an adjusted OR of 0.36 (95 % CI: 0.22–0.58; p < 0.001). Moreover, rate of seroma was significantly lower in the CDR group (0.1 %) than the PCF group (0.4 %), with an adjusted OR of 0.20 (95 % CI: 0.06–0.52; p = 0.003). Wound infections were also less prevalent in the CDR group. Deep wound infection rates were significantly lower in the CDR group (0.4 %) than the PCF group (1.7 %), with an adjusted OR of 0.23 (95 % CI: 0.13–0.38; p < 0.001). Superficial wound infection rates were significantly lower in the CDR group (0.2 %) than the PCF group (0.6 %), with an adjusted OR of 0.31 (95 % CI: 0.14–0.62; p = 0.002) (Fig. 1). 90-day cervical revision surgery rate was significantly lower in the CDR group (0.6 %) than the PCF group (2.2 %), with an adjusted OR of 0.26 (95 % CI: 0.17–0.40; p < 0.001) (Fig. 1).

Close

Fig. 1 90-Day Surgical Complication Rates. Comparison of complication rates within 90 days following cervical disc replacement (CDR, red bars) and posterior cervical foraminotomy (PCF, blue bars). Bars represent the percentage of patients experiencing each complication, with p-values indicating statistical significance between groups.

Export to PPT

Similarly, 90-day emergency department visit rates were lower in the CDR group (9.5 %) than the PCF group (10.9 %), with an adjusted OR of 0.88 (95 % CI: 0.76–1.01, p = 0.074). Additionally, 90-day hospital readmission rates were significantly lower in the CDR group (1.3 %) than the PCF group (2.7 %), with an adjusted OR of 0.47 (95 % CI: 0.23–0.64, p < 0.001) (Fig. 2).

Close

Fig. 2 90-Day Readmission Rates. Comparison of emergency department (ED) visits and hospital readmissions within 90 days following cervical disc replacement (CDR, red bars) and posterior cervical foraminotomy (PCF, blue bars). Bars represent the percentage of patients experiencing each event, with p-values indicating statistical significance between groups.

Export to PPT

3.4 Presentation

4.1 Key results

Cervical radiculopathy is one of the most prevalent spinal conditions in the United States, affecting approximately 83.2 per 100,000 people.¹⁹ Over the past decade, treatment strategies have evolved, with Cervical Disc Replacement (CDR) and Posterior Cervical Foraminotomy (PCF) emerging as key surgical interventions that avoid fusion-associated complications.³ Our study aimed to compare these two procedures by analyzing patient demographics and evaluating 90-day postoperative medical and surgical complications.

Our findings indicate that both CDR and PCF effectively alleviate cervical radiculopathy, with only 1–2 % of patients experiencing persistent symptoms postoperatively, and no significant difference between surgical treatments. This result aligns with previous literature demonstrating comparable clinical outcomes between these surgical options ²¹. However, patients undergoing PCF experienced significantly higher rates of cervical revision surgery and hospital readmission within 90 days. In this study, cervical revision was broadly defined using CPT codes for cervical arthrodesis, hardware removal, laminectomy, laminotomy, and corpectomy below the C2 level within two years of the index procedure. This comprehensive, claims-based definition may capture an extensive range of subsequent cervical interventions, potentially inflating revision rates relative to more granular studies.

Prior studies offer valuable comparisons. Padhye et al. also reported higher revision rates for PCF at 2-year follow-up using a more granular chart review methodology.²⁰ Similarly, Ng et al. observed significantly higher short-term (30-day) reoperation rates for PCF compared to CDR and ACDF, mainly attributed to perioperative complications, further aligning with our conclusions.²¹ Conversely, Nayak et al., using the PearlDiver database and an earlier patient cohort (2007–2016), found no consistent significant differences in reoperation rates between PCF and CDR at 30 days, 3 months, and 1 year, except at 6 months when CDR demonstrated higher rates than PCF.²² The discrepancy between Nayak et al.’s results and our findings could potentially be explained by differences in ICD and CPT code selection methodology to broadly define potential revision surgeries.²² Additionally, Nayak et al.’s study included patients who received one- or two-level CDR and PCF, whereas our study focused exclusively on single-level CDR versus single-level PCF.²² Differences in revision rates may also reflect improvements in surgical techniques, technological advancements in prosthesis design, and enhanced patient management practices since Nayak et al.'s earlier cohort study.^17,18 Overall, existing literature across the last two decades generally supports our conclusion that PCF is associated with higher revision rates compared to CDR, as found in Fang et al.’s meta-analysis study.²³

Despite advancements in PCF techniques over the past decade, including increased utilization of CT navigation and endoscopic approaches, CDR appears to be the superior surgical option regarding revision rate and hospital readmission. Moreover, the PCF group exhibited a nonsignificant but trending increase (p = 0.074) in the rate of emergency room visits within 90-days. This trend in recent years could be explained by developments in CDR hybrid prosthesis technology, leading to greater preservation of native range-of-motion at the index and adjacent levels.²⁴

For CDR, the most common surgical complications were dysphagia and dysphonia. This finding aligns with previous literature, which has shown a well-established pattern of these complications with an anterior approach to cervical spine surgery.²⁵ Anatomically, the anterior path typically requires manipulation of the trachea and esophagus to visualize and access the cervical vertebrae.²⁶ On this route, many nerves are also passed, such as the recurrent laryngeal nerve, which can lead to vocal cord dysfunction if injured.²⁷ Consequently, these established risks should be thoroughly considered when choosing between anterior (CDR) and posterior (PCF) surgical routes. Notably, our demographic analysis revealed smokers were less likely to undergo the anterior CDR approach, likely due to their increased susceptibility to dysphagia and dysphonia resulting from tobacco use.^28,29 However, tobacco usage confers a detrimental impact on tissue repair and wound healing: complications that were still elevated in the PCF group after accounting for the effect of smoker status. Thus, for patients who smoke, careful consideration and preoperative counseling regarding the increased risk of wound complications with PCF is essential. Additionally, patients with a predisposition to esophageal cancer, those with motor deficits, and patients with autoimmune conditions affecting the neck, such as Eosinophilic Esophagitis, can be further encouraged to use a posterior approach.

Apart from the increased prevalence of dysphagia and dysphonia in the CDR group, the PCF group demonstrated a significantly poorer perioperative complication profile. Patients undergoing PCF exhibited higher incidences of dural tears, wound dehiscence, seromas, and both deep and superficial wound infections. The causes of these complications are likely related to the route of surgery. The posterior approach requires more extensive dissection and manipulation of the paraspinal muscles, thus leading to increased tissue trauma, infection, and wound complications.^30,31 Studies have illustrated that nuchal thickness can correlate with wound complications, providing a potential explanation for the findings of our study.³² We did not find a statistically significant difference in BMI between groups; however, over twenty percent of overall sampled patients had a BMI over 30. Based on previous literature, the anterior approach for CDR provides similar levels of medical and surgical complications across various levels of obesity.³³ However, obesity has been described as a risk factor for morbidity in operations involving a posterior cervical spinal approach.³⁴ Thus, in the context of PCF, elevated BMI and increased nuchal thickness may contribute to a higher incidence of wound complications, including dehiscence, seroma formation, and wound infections. Increased adipose tissue in the posterior cervical region may create a larger surgical dead space and prolong wound healing, predisposing patients to postoperative complications.³⁵ Additionally, excessive fat deposition can impair vascular supply to the surgical site, reducing tissue perfusion and increasing the risk of infection.³⁶ Given these findings, preoperative BMI optimization may be particularly beneficial for patients undergoing PCF, as weight reduction could mitigate wound-related morbidity and improve overall surgical outcomes.

There were two unexpected findings in our study. Our results demonstrated significantly higher incidences of dural tears in the PCF group compared to patients who underwent CDR. While this perioperative complication difference has not been previously described in the literature,¹⁶ this observation aligns with recent studies highlighting elevated dural tear rates in endoscopic posterior cervical foraminotomy procedures.³⁷ Historically, management of dural tears required open surgical repair.³⁸ However, recent advancements such as dural sealants and patch techniques have enabled effective management through minimally invasive approaches.^39,40 Although the optimal therapeutic strategy remains debated, careful consideration of emerging minimally invasive treatment options, such as dural sealants and patches, can mitigate the necessity for transitioning to open surgery. Continued investigation into minimally invasive management approaches remains essential for enhancing patient outcomes and reducing procedural complications.

4.2 Limitations & strength

There are several limitations of note for this study. PearlDiver, a private analytics database by Humana, includes patients aged 65 and older, with a concentration in the Southern United States, and does not consist of a randomly sampled population.⁴¹ Furthermore, clerical mistakes and annual updates to ICD-10 code definitions can lead to coding errors, changing results.⁴² To enhance the statistical power of the PCF group, we did not differentiate between microscopic, endoscopic, or open PCF. There may be differences in perioperative complication rates between these surgical subtypes of PCF; however, the current literature has described these subtypes as having similar complication profiles.⁴³

Moreover, the lack of granularity in the PearlDiver database does not allow for the evaluation of severities of complications. However, we accounted for this limitation by analyzing emergency department visits and hospital readmission rates as surrogate markers for complication severity. The lack of granularity also prevents assessment of estimated blood loss (EBL), a crucial variable influencing surgical outcomes such as renal complications and hypotension. Furthermore, PearlDiver's structure limits differentiation of revision procedures, restricting our ability to ascertain whether revisions involved solely the index level or additional vertebral levels. Lastly, the PearlDiver database does not allow us to control for several factors that may have guided surgical decision making, such as baseline severity of disease or radiographic findings. Future case-controlled, prospective single-center studies could be better suited to clarify potential associations between PCF and CDR in cervical radiculopathy patients, and more detailed studies could help better guide choices for surgeons and their patients.