Study design, data sources, and setting

This retrospective observational study used data from two national registries: the Swedish National Registry for Spine Surgery (Swespine) and the Swedish Myeloma Registry (SMR). The study population included all patients recorded in Swespine who underwent spinal surgery for metastatic disease secondary to MM between 2011 and 2023. Patients whose primary tumor was unknown at the time of surgery but later confirmed to be MM were also included.

Cross-linkage with the SMR was performed to verify diagnoses and obtain supplementary clinical data. For patients who underwent multiple spinal procedures, only the index surgery was included. In total, 207 patients were identified in both the Swespine and SMR registries. Seven patients were excluded (Fig. 1) due to Swespine records indicating surgery for spinal metastases from prostate cancer.

Fig. 1

Flowchart of Patient Inclusion

The Swedish national registry for spine surgery (Swespine)

Swespine was established in 1993 and expanded in 2006 to include metastatic spinal disease. It currently receives data from approximately 98% of spinal surgery units in Sweden, including all centers performing tumor surgery, and has consistently demonstrated high validity and completeness [15, 16]. The registry records details on primary tumor type, surgical procedures, metastatic location, neurological function, complications, and patient-reported outcomes at 6 weeks postoperatively.

Data retrieved for this study included patient age at the time of surgery, date of surgery, surgical approach and technique, and spinal lesions associated with metastases. Neurological function was assessed using the Frankel grading system. Baseline walking ability and pain levels were patient-reported preoperatively, with follow-up assessments conducted 6 weeks after surgery through digital or postal questionnaires.

The Swedish myeloma registry (SMR)

The (SMR), maintained by the Swedish Myeloma Group, includes all cases of MM diagnosed in Sweden since 1 January 2008, with an estimated coverage of 99% [17]. The registry systematically collects comprehensive demographic, clinical, and outcome data, including age at diagnosis, sex, Eastern Cooperative Oncology Group (ECOG) performance status, and date of diagnosis.

The SMR also records key laboratory parameters, including hemoglobin (Hb), M-protein (M-component), serum β2-microglobulin (Sβ2M), and serum free light chain (sFLC) levels with kappa/lambda ratio. The date of death is recorded for all patients. At diagnosis, patients are categorized into clinical subgroups—symptomatic myeloma, solitary skeletal plasmacytoma, extraskeletal plasmacytoma, or asymptomatic myeloma—and staged according to the International Staging System (ISS).

For this study, data extracted from the SMR included age and date at diagnosis, primary treatment regimen, date of death, and relevant laboratory parameters (Hb, M-component, Sβ2M, and sFLC). Disease category and ISS stage at primary diagnosis were also recorded.

Data linkage

Data linkage between Swespine and the SMR was performed using the Swedish personal identity number (PIN). This unique identifier enables precise cross-referencing of individual patient records across national registries [18, 19].

Independent variables

Disease staging was performed according to the ISS, which stratifies patients into three stages based on Sβ2M and serum albumin levels:

Stage I

Sβ2M < 3.5 mg/L and serum albumin ≥ 3.5 g/dL.

Stage II

Neither Stage I nor Stage III (Sβ2M < 3.5 mg/L with serum albumin < 3.5 g/dL, or Sβ2M 3.5–5.5 mg/L regardless of serum albumin level).

Stage III

Sβ2M > 5.5 mg/L.

Hb level and age at diagnosis were obtained from the SMR. Age at surgery was extracted from Swespine.

Ambulation status

Frankel grades were dichotomized into non-ambulatory (Frankel grades A–C) and ambulatory (Frankel grades D–E).

Synchronous and Metachronous Disease: Synchronous disease was defined as spinal surgery performed within 90 days of the primary MM diagnosis. This interval was selected to capture spinal manifestations occurring at or closely related to the time of diagnosis while accounting for potential delays in diagnostic work-up, referral, and surgical planning in routine clinical practice.

Metachronous disease was defined as spinal surgery performed more than 90 days after the primary MM diagnosis. To evaluate the robustness of this definition, predefined sensitivity analyses were performed using alternative cut-offs of 0, 30, and 60 days. The results were consistent across all thresholds, supporting the robustness of the 90-day definition.

The date of death was retrieved from the SMR, which receives mortality data from the National Cause of Death Registry.

Outcome measures

The primary outcome was postoperative survival, defined as the interval from the date of spinal surgery to the date of death from any cause. Secondary outcomes included patient-reported functional mobility and pain level. These outcomes were assessed preoperatively and at 6 weeks postoperatively to evaluate changes in ambulation and pain intensity.

Statistical analysis

Descriptive statistics for continuous variables were presented as means with standard deviations (SDs) or medians with ranges, as appropriate. Categorical variables were summarized as numbers (n) and percentages (%). Group comparisons were performed using Student’s t-test or the Mann–Whitney U test for continuous variables, depending on data distribution, and the chi-square test for categorical variables.

Survival analyses were conducted using the Kaplan–Meier method, with death defined as the event. Overall survival was calculated from the date of diagnosis, and postoperative survival was calculated from the date of surgery. Survival curves were compared using the log-rank test. A multivariable Cox proportional hazards model was constructed to identify independent predictors of mortality. Results are reported as hazard ratios (HRs) with 95% confidence intervals (CIs).

Based on clinical relevance, ISS stage and timing of spinal involvement were included as surrogate markers of disease burden and biological behavior. In contrast, patient age and preoperative ambulatory status were included as established prognostic factors. Detailed information on cytogenetic abnormalities, systemic treatment regimens, and treatment response was not available in the linked registry data and, therefore, could not be included in the analyses.

Missing data were handled using complete-case analysis. The multivariable Cox proportional hazards model included only patients with complete data for all covariates entered into the model (n = 152). No imputation of missing values was performed. Although multiple imputation was considered, the linked registry data lacked sufficient auxiliary variables to support a robust imputation model, particularly for missing ISS stage and patient-reported follow-up outcomes. Therefore, complete-case analysis was considered the most appropriate and transparent analytical approach.

The proportional hazards assumption was assessed using Schoenfeld residuals, and no significant violations were detected. Potential multicollinearity between haemoglobin (Hb) and ISS stage was also assessed, and no significant multicollinearity was identified. A p-value < 0.05 was considered statistically significant. All analyses were performed using IBM SPSS Statistics, version 29 (IBM Corp., Armonk, NY, USA) and R.