MORTALITY RISK EVIDENCE FOLLOWING VERTEBRAL COMPRESSION FRACTURES

WERE VCF PATIENTS AT HIGHER RISK OF MORTALITY FOLLOWING THE 2009 PUBLICATION OF THE VERTEBROPLASTY "SHAM" TRIALS?¹

STUDY OBJECTIVE

Balloon kyphoplasty (BKP) and vertebroplasty (VP) are associated with lower mortality risks than non-surgical management (NSM) of vertebral compression fractures (VCF). VP versus sham trials published in 2009 sparked controversy over its effectiveness. Rates of VP and BKP have decreased since 2009. In this study, the authors hypothesized that lower BKP/VP usage would lead to a greater mortality risk for VCF patients.

This study aims to evaluate if the mortality and morbidity of VCF patients during 2010-2014 had been impacted by the diminished use of BKP and VP during this period compared with the 2005-2009 period of higher BKP and VP usage.

PATIENT POPULATION

Total: 2,077,944 patients

• BKP (n=261,756)
• VP (n=117,232)
• NSM (n=1,698,956)

STUDY DESIGN AND FOLLOW-UP

The authors conducted an analysis of 100% Medicare claims data for 2005-2014. Patients were included in this study if they were 65 years of age or older, had a newly diagnosed VCF, and did not have a VCF in the previous 12 months. Patients were stratified into BKP, VP, and NSM cohorts.

Patients were excluded if they were enrolled in an HMO plan, were not enrolled in Parts A and B of Medicare, or had a fusion within 12 months of their VCF diagnosis (NSM cohort) or between their diagnosis and their BKP/VP procedure.

OUTCOMES MEASURED

The primary and secondary endpoints included in this study were: mortality, morbidity, readmission, hospital length of day (LOS), and discharge status. Survival and morbidity were analyzed by the Kaplan-Meyer method and compared between patient cohorts treated with NSM, BKP, or VP using multivariate Cox regression with adjustment for propensity score and other factors. Propensity score was derived for the probability of undergoing BKP or VP using logistic regression conditional on age, gender, race, census region, socioeconomic status, comorbidities, type of fracture, diagnosis of osteoporosis, fracture location, site of service and physician specialty of initial VCF diagnosis, and year.

PERFORMANCE (EFFECTIVENESS) RESULTS

Mortality: Based on the analysis of the 100% Medicare claims data for 2005-2014, the authors found that there was a 10% decline in vertebral augmentation (either BKP or VP) volume. Specifically, the vertebral augmentation volume rate was 14% in 2014 compared to 24% in 2009 (p<0.001). When comparing time periods, the propensity-adjusted mortality risk for VCF patients was 4% (95% CI: 3%-4%; p<0.001) greater in 2010-2014 than 2005-2009.

The overall 10-year mortality risk for VCF patients was 85.1%.

As shown below, the propensity-adjusted 10-year mortality risk analysis of the data from 2005-2014 by treatment cohort revealed that individuals treated with NSM had a significantly higher mortality risk than those treated with BKP and VP; and those treated with BKP had a significantly lower mortality risk than those treated with VP.

Comparison of propensity-adjusted 10-year mortality risk:

• 24% higher mortality risk for NSM vs. BKP, p<0.001
• 8% higher mortality risk for NSM vs. VP, p<0.001
• 13% lower mortality risk for BKP vs. VP, p<0.001

Morbidity: At one year, outcomes with at least 10% greater risk for NSM than BKP patients were cardiac complications (19%; 95% CI: 17%-21%; p<0.001) and pneumonia (23%; 95% CI: 22%-24%; p<0.001).

Note: The cause of death is unknown in the dataset, but various morbidities and mortality with pneumonia diagnosed in the 90 days prior to death were used to provide some insight into the health status leading to expiration.

Length of Hospital Stay and Discharge Status: While patients treated with NSM had a relatively shorter length of hospital stay (mean 5.2 days) compared to BKP (5.4 days) and VP (6.6 days; p-values < 0.001), NSM patients also had a relatively higher risk of discharge to a skilled nursing facility (48%) compared to BKP (31%) and VP (39.6%) (p-values <0.001). NSM also had the lowest relative rate of discharge to their home (33.7%) compared to BKP (56.9%) and VP (47%) (p-values <0.001).

Note: This study was unable to determine the criteria for referrals to BKP and VP compared to NSM. Because the NSM cohort was discharged to nursing facilities at a significantly higher rate than BKP and VP cohorts, the present analysis could not accurately assess the LOS between the three cohorts.

Length of stay was slightly longer for hospitalized BKP, but BKP’s having a higher likelihood to be discharged to home appears to reflect a shifting of the NSM patients from inpatient facilities to other facilities and does not reflect recovery of patients.

CONCLUSIONS

Based on this analysis of Medicare claims data from over 2 million elderly patients, the authors concluded that the five-year period following 2009 saw a steep reduction in vertebral augmentation volume and patients who were afflicted with VCFs and treated between 2010-2014 had significantly higher mortality risk than patients treated in the preceding 5-year period.

MORBIDITY AND MORTALITY AFTER VERTEBRAL FRACTURES: COMPARISON OF VERTEBRAL AUGMENTATION AND NON-OPERATIVE MANAGEMENT IN THE MEDICARE POPULATION²

STUDY OBJECTIVE

Edidin and colleagues sought to test the following hypotheses: (1) vertebral compression fracture patients in a balloon kyphoplasty cohort would experience no difference in mortality and morbidity as compared to patients in a conservatively managed ('nonoperated') cohort; (2) vertebral compression fracture patients in a vertebroplasty cohort would experience no difference in mortality and morbidity compared to patients in a conservatively managed cohort; and (3) vertebral compression fracture patients in a balloon kyphoplasty cohort would experience no difference in mortality and morbidity compared to patients in a vertebroplasty cohort.

PATIENT POPULATION

Total: 1,038,956 patients

• BKP (n=141,343)
• VP (n=75,364)
• NSM (n=822,249)

STUDY DESIGN AND FOLLOW-UP

Retrospective database review of Medicare claims for the 2005–2009 time period of vertebral compression fracture patients with up to 4 years of follow-up.

OUTCOME MEASURES

The prevalence of morbidities was evaluated based on the coding of corresponding ICD-13 9-CM diagnosis and procedure codes, as well as relevant CPT-4 codes for specific morbidities and subsequent treatment.

The mortality, morbidity, and subsequent treatment risks were estimated by the Kaplan-Meier method.

Differences in these outcomes were assessed by multivariate Cox regression, adjusting for comorbidities and other covariates. Subgroup analysis was performed for all patients with pathologic (or osteoporotic) vertebral compression fractures.

Propensity score matching analysis was performed to address selection bias through potential misspecification and account for potential uneven distribution of measured covariates.

PERFORMANCE (EFFECTIVENESS) RESULTS

In the non-matched cohorts, the unadjusted Kaplan-Meier risk of mortality at 4 years was 49.4%, 46.2%, and 41.8% for the nonoperated, vertebroplasty, and balloon kyphoplasty cohorts, respectively. The adjusted 4-year mortality risk for the nonoperated cohort was found to be 55.0% (95.0% Confidence Interval [CI], 53.0%-56.0%; P<0.001) and 25.0% (95.0% CI, 23.0%-26.0%; P<0.001) higher than the balloon kyphoplasty and vertebroplasty cohorts, respectively.

The balloon kyphoplasty cohort also had a 19.0% lower adjusted 4-year mortality risk than the vertebroplasty cohort (adjusted hazard ratio [AHR]=0.81, 95.0% CI, 0.70-0.82; P<0.001).

After propensity matching, the Kaplan-Meier risk of mortality at 4 years was still found to be greater for the nonoperated cohort with a greater adjusted 4-year mortality risk than those in the propensity matched balloon kyphoplasty (AHR=1.62, 95.0% CI, 1.601.64; P<0.001) and vertebroplasty (AHR=1.30, 95.0% CI, 1.28-1.33; P<0.001) cohorts.

The adjusted risk of mortality at 4 years continued to be lower by 17.0% (AHR=0.83, 95% CI, 0.81-0.85; P<0.001) for the propensity matched balloon kyphoplasty cohort over the vertebroplasty cohort. After propensity matching, patients in the nonoperated cohort had greater adjusted 4-year mortality risk (with any pneumonia diagnosis or with pneumonia as principal diagnosis in the prior 90 days) than those in the balloon kyphoplasty and vertebroplasty cohorts. The adjusted risk of mortality was also lower for the propensity matched balloon kyphoplasty cohort over the vertebroplasty cohort.

The mortality findings were generally similar for the subgroup of osteoporotic vertebral compression fracture patients before and after propensity matching. For example, after propensity matching, the nonoperated osteoporotic vertebral compression fracture cohort had a 70.0% higher adjusted 4-year mortality risk (AHR=1.70, 95.0% CI, 1.67-1.74; P<0.001) than the balloon kyphoplasty osteoporotic vertebral compression fracture cohort.

After propensity matching, the balloon kyphoplasty osteoporotic vertebral compression fracture cohort also had a 17.0% lower adjusted 4-year mortality risk (AHR=0.83, 95.0% CI, 0.80-0.87; P<0.001) than the vertebroplasty osteoporotic vertebral compression fracture cohort. These comparisons between the nonoperated and balloon kyphoplasty cohorts and between the balloon kyphoplasty and vertebroplasty cohorts were consistent for the 3 subgroups analyzed— the overall osteoporotic vertebral compression fracture patients, osteoporotic vertebral compression fracture patients who survived at least 1 year, and non-cancer osteoporotic vertebral compression fracture patients.

CONCLUSIONS

Edidin and colleagues were unable to support the null hypotheses that vertebral compression fracture patients in the balloon kyphoplasty, vertebroplasty, and nonoperated cohorts would experience no difference in mortality and morbidity. Instead, the 4-year results showed that vertebral compression fracture patients in the Medicare population who received kyphoplasty and vertebroplasty experienced lower mortality and morbidity than vertebral compression fracture patients who received conservative management, even after using propensity matching analysis to account for potential treatment selection bias because treatment was not randomly assigned.

Edidin and colleagues also found that vertebral compression fracture patients in the balloon kyphoplasty cohort also experienced an additional 4-year survival benefit with fewer morbidities overall than the vertebroplasty cohort. These survival benefits for balloon kyphoplasty over vertebroplasty were also evident for the subgroups of osteoporotic vertebral compression fracture patients, including those who survived at least 1 year and those with no cancer diagnosis.

The reasons underlying these apparent differences in patient outcomes among the treated and untreated cohorts and the causality cannot easily be addressed using the Medicare data alone. Despite this, Edidin and colleagues found higher 4-year mortality risk with pneumonia diagnosed in the prior 90 days, as well as greater risk of pneumonia, which provides some insight into possible causes of death.

SURVIVAL AND COST COMPARISON OF KYPHOPLASTY AND PERCUTANEOUS VERTEBROPLASTY USING GERMAN CLAIMS DATA³

STUDY OBJECTIVE

Examine the overall survival and treatment costs from a third-party-payer perspective for patients with osteoporotic vertebral compression fractures treated by vertebral augmentation or conservative treatment in Germany.

PATIENT POPULATION

Total: 3,607 patients

• BKP (n=441)
• VP (n=157)
• NSM (n=3,009)

STUDY DESIGN AND FOLLOW-UP

Observational study with a follow-up time of up to 5 years.

OUTCOME MEASURES

Claims data for 2005 to 2010 from a major health insurance fund were used.

Mortality risk differences between the operated (balloon kyphoplasty, percutaneous vertebroplasty) and nonoperated cohorts were assessed by using Cox regression.

Operated patient cohort was established by propensity score matching and adjusting for covariates.

For the matched operated patients with osteoporotic vertebral compression fractures, survival was estimated by using Kaplan-Meier method.

PERFORMANCE (EFFECTIVENESS) RESULTS

The operated cohort was 43.0% less likely to die than the nonoperated cohort in the 5-year study period (hazard ratio=0.57; P<0.001).

Patients who received balloon kyphoplasty had higher 60-month adjusted survival rate (66.7%) than those who received percutaneous vertebroplasty (58.7%)(P=0.68).

Cumulative 4-year mean overall costs after first diagnosis were lower for the balloon kyphoplasty patients (percutaneous vertebroplasty: €42,510 versus balloon kyphoplasty: €39,014).

Initial upfront higher costs driven by surgical treatment for patients who received balloon kyphoplasty are offset by considerable pharmacy costs in patients who received percutaneous vertebroplasty.

There were differences between the values of painkiller consumption (percutaneous vertebroplasty: €3321 versus balloon kyphoplasty: €2224).

CONCLUSIONS

Results suggest a higher overall survival rate for operated than nonoperated patients with osteoporotic vertebral compression fractures and indicate a potential survival benefit for patients who received balloon kyphoplasty compared with patients who received percutaneous vertebroplasty. The reasons merit further investigation.

Total costs were lower after 4 years for patients who received balloon kyphoplasty versus percutaneous vertebroplasty due to less consumption of pharmaceuticals.

IMPACT OF NONOPERATIVE TREATMENT, VERTEBROPLASTY, AND KYPHOPLASTY ON SURVIVAL AND MORBIDITY AFTER VERTEBRAL COMPRESSION FRACTURE IN THE MEDICARE POPULATION4

STUDY OBJECTIVE

Compare the survival rates, complications, lengths of hospital stay, hospital charges, discharge locations, readmissions, and repeat procedures for Medicare patients with new vertebral compression fractures that had been acutely treated with vertebroplasty kyphoplasty, or nonoperative modalities.

PATIENT POPULATION

Total: 68,752 patients

• BKP (n=22,817)
• VP (n=7,686)
• NSM (n=38,249)

STUDY DESIGN AND FOLLOW-UP

Retrospective database review with total follow-up of 129,783 person-years.

OUTCOME MEASURES

The primary parameter of interest was survival at 6 months, 1 year, 2 years, and 3 years.

Secondary parameters of interest were complications, lengths of hospital stay, hospital charges, discharge locations, thirty-day readmission rates, and repeat procedures. Complications during the index hospitalization and at six months included mortality, pulmonary embolism, deep-vein thrombosis, pneumonia, infection, decubitus ulcer, and neurologic compromise.

PERFORMANCE (EFFECTIVENESS) RESULTS

The estimated three-year survival rates were 42.3%, 49.7%, and 59.9% for the nonoperative treatment, vertebroplasty, and kyphoplasty cohorts, respectively.

The adjusted risk of death was 20.0% lower for the kyphoplasty cohort than for the vertebroplasty cohort (hazard ratio=0.80, 95% confidence interval, 0.77 to 0.84).

Patients who were managed nonoperatively were hospitalized longer (average, 7.38 days) than those who were managed operatively (vertebroplasty average, 5.73 days; kyphoplasty average, 3.74 days). Patients who were managed with kyphoplasty had much higher odds of being routinely discharged to home than those who were managed nonoperatively (59.9% compared with 24.3%, respectively; P<0.001).

Thirty-day readmission rates were high for all 3 cohorts, but patients in the nonoperative cohort were much more likely to return (rate of readmission, 61.9% for the nonoperative cohort, 52.4% for the vertebroplasty cohort, and 35.2% for the kyphoplasty cohort; P<0.001).

With regard to specific morbidities and subsequent treatment, Patients in the kyphoplasty cohort were the least likely to have had pneumonia and decubitus ulcers during the index hospitalization and at 6 months postoperatively; however, kyphoplasty was more likely to result in a subsequent augmentation procedure than was vertebroplasty (9.41% compared with 7.89%; P<0.001).

CONCLUSIONS

Vertebral augmentation procedures appear to be associated with longer patient survival than nonoperative treatment does.

Kyphoplasty tends to have a more striking association with survival than vertebroplasty does, but it is costly and may have a higher rate of subsequent vertebral compression fracture.

The results of this study suggest that the beneficial impact of minimally invasive surgery for vertebral compression fractures reaches beyond the acute phase and is associated with improvements in terms of post-discharge survival and morbidity.

MAJOR MEDICAL OUTCOMES WITH SPINAL AUGMENTATION VS CONSERVATIVE THERAPY⁵

STUDY OBJECTIVE

Compare major medical outcomes following treatment of osteoporotic vertebral fractures with spinal augmentation or conservative therapy. Additionally, evaluate the role of selection bias using preoperative outcomes and propensity score analysis.

PATIENT POPULATION

Total: 126,392 patients

• Vertebral Augmentation (n=10,541)
• NSM (n=115,851) 

STUDY DESIGN AND FOLLOW-UP

Retrospective cohort analysis of Medicare claims for the 2002 to 2006 period. Thirty-day and 1-year outcomes in patients with newly diagnosed vertebral fractures treated with spinal augmentation or conservative therapy were compared.

OUTCOME MEASURES

The main outcomes and measures included mortality, major complications, and health care usage.

With regard to hospital usage, McCullough and colleagues recorded general hospital admissions, intensive care unit (ICU) admissions, and discharges to a skilled nursing facility (SNF) during the year following vertebral fracture; however, they did not include hospital admissions during which spinal augmentation was performed.

Outcomes were compared using traditional multivariate analyses adjusted for patient demographics and comorbid conditions. Propensity score matching also was utilized to select 9,017 pairs from the initial groups to compare the same outcomes.

PERFORMANCE (EFFECTIVENESS) RESULTS

Mortality was significantly lower in the augmented cohort than among controls (5.2% versus 6.7% at 1 year; hazard ratio [HR]=0.83; 95.0% Confidence Interval [CI], 0.75-0.92); however, patients in the augmented group who had not yet undergone augmentation (preoperative subgroup) had lower rates of medical complications 30 days post fracture than did controls (6.5% versus 9.5%; odds ratio [OR]=0.66; 95% CI, 0.57-0.78), suggesting that the augmented group was less medically ill.

After propensity score matching to better account for selection bias, 1-year mortality was not significantly different between the cohorts. Furthermore, 1-year major medical complications were also similar between the groups, and the augmented group had higher rates of health care usage, including hospital and intensive care unit admissions and discharges to skilled nursing facilities.

CONCLUSIONS

McCullough and colleagues concluded that, after accounting for selection bias, spinal augmentation did not improve mortality or major medical outcomes and was associated with greater health care usage than conservative therapy, and that their results highlighted how analyses of claims-based data that do not adequately account for unrecognized confounding can arrive at misleading conclusions.

MORTALITY RISK FOR OPERATED AND NONOPERATED VERTEBRAL FRACTURE PATIENTS IN THE MEDICARE POPULATION6

STUDY OBJECTIVE

Evaluate the mortality risk for patients with vertebral compression fractures undergoing different treatment modalities. Specifically, Edidin and colleagues sought to test the following hypotheses: (1) The mortality risk for nonoperated (i.e., conservative treatment) vertebral compression fractures patients is greater than for those who undergo operative (i.e., kyphoplasty or vertebroplasty) treatment; and (2) the mortality risks for vertebral compression fractures patients who undergo either kyphoplasty or vertebroplasty are similar.

PATIENT POPULATION

Total: 858,978 patients

• BKP (n=119,253)
• VP (n=63,693)
• NSM (n=676,032)

STUDY DESIGN AND FOLLOW-UP

Retrospective database review of Medicare Claims for the 2005–2008 time period of vertebral compression fracture patients with up to 4 years of follow-up.

OUTCOME MEASURES

The mortality rate associated with vertebral compression fracture was evaluated by determining the date of death from the annual Medicare denominator file.

Survival of a vertebral compression fracture patient was calculated from the index diagnosis date until death or end of follow-up on December 31, 2008.

Survival of vertebral compression fracture patients in the 100.0% U.S. Medicare data set (2005–2008) was estimated by the Kaplan-Meier method, and the differences in mortality rates at up to 4 years were assessed by Cox regression (adjusted for comorbidities) between operated and nonoperated patients and between kyphoplasty and vertebroplasty patients.

An instrumental variables analysis was used to evaluate mortality-rate difference between kyphoplasty and vertebroplasty patients.

PERFORMANCE (EFFECTIVENESS) RESULTS

At up to 4 years of follow-up, patients in the operated cohort had a higher adjusted survival rate of 60.8% compared with 50.0% for patients in the nonoperated cohort (P<0.001) and were 37% less likely to die (adjusted hazard ratio [HR]=0.63, P<.001).

The adjusted survival rates for vertebral compression fractures patients following vertebroplasty or kyphoplasty were 57.3% and 62.8%, respectively (P<0.001).

The relative risk of mortality for kyphoplasty patients was 23.0% lower than that for vertebroplasty patients (adjusted HR=0.77, P<.001).

Using physician preference as an instrument, the absolute difference in the adjusted survival rate at 3 years was 7.29% higher in patients receiving kyphoplasty than vertebroplasty (P<.001), compared with a crude absolute rate difference of 5.09%.

CONCLUSIONS

Edidin and colleagues were able to utilize the 100.0% Medicare data set to provide a comparison of mortality risks in an extremely large vertebral compression fracture patient cohort and, as such, concluded that:

• The data set could be used for further analysis of other patient outcomes, as it is extremely useful for detecting small changes in outcomes owing to the sample size.
• Additional research would be needed to further evaluate the efficacy, causality, and benefits from various treatment modalities.