Retrospective national database study design.
This study was designed to determine whether acute percutaneous vertebral augmentation (PVA) alters morbidity compared with nonoperative management.
Osteoporotic vertebral compression fractures (OCFs) are common and represent a large economic and patient burden. Several recent studies have focused on whether PVA offers benefits compared with nonoperative treatment.
A retrospective cohort analysis was conducted using the Nationwide Inpatient Sample from 2015 to 2018. Patients with nonelective admissions for OCFs were identified using International Classification of Diseases (10th edition) codes. The exclusion criteria included age of less than 50 years, fusion and decompression procedures, and the presence of neoplasms and infections. Propensity score matching was implemented to construct 2:1 matched cohorts with similar comorbidities at admission. The patients were divided into the operative and nonoperative treatment groups. Univariate and multivariate regression analyses were performed to compare differences in in-hospital complication rates between the groups. All
We identified 14,850 patients in the operative group and 29,700 patients in the nonoperative group. In the multivariate analysis, operative treatment was associated with significantly lower rates of pneumonia (odds ratio [OR], 0.75;
Patients who undergo acute PVA for OCFs have lower rates of respiratory complications, cardiac complications, sepsis, and pressure ulcerations while having a higher risk of acute renal failure.
Osteoporotic compression fractures (OCFs) are the most common osteoporotic fracture among the aging population, comprising 27% of all osteoporotic fractures [
Recent literature has focused on treating OCFs. Operative management consists of percutaneous vertebral augmentation (PVA), including vertebroplasty and kyphoplasty, whereas nonoperative management consists of pain management and external orthoses [
Several studies have been conducted to elucidate the benefits of kyphoplasty versus nonoperative treatment for OCFs. Numerous studies have shown that kyphoplasty offers greater improvement in daily activities, quality of life, and vertebral deformity than conservative treatment [
However, the number of studies examining and comparing acute complication rates between operative and nonoperative management of these fractures in patients hospitalized for OCFs is limited. This study was designed to determine whether acute PVA alters morbidity compared with nonoperative management.
The Nationwide Inpatient Sample (NIS) is a national database within the Healthcare Cost and Utilization Project published annually and provides nationally representative information on over 7 million hospitalizations across the United States. The NIS is the only national database that uses discharge weights, allowing accurate estimations of nationwide incidence of diseases and procedures. Additionally, the NIS captures all patients, regardless of insurance status. The database uses International Classification of Diseases, 10th revision (ICD-10) codes for documenting all diagnoses and indicated procedures per hospital admission. Given the nonspecific nature of ICD-9 coding used in years before 2015, this study only analyzed NIS years 2015–2018 to use the additional granularity offered by ICD-10 coding. This study did not require Institutional Review Board approval as all patient data within the NIS is deidentified.
Patients who were nonelectively admitted for OCFs from the 2015 to 2018 NIS dataset were identified using ICD-10 codes. The patients were then divided into those who underwent operative treatment and those who received nonoperative treatment. Operative treatment was defined by the use of PVA, by either vertebroplasty or kyphoplasty. Patients aged less than 50 years, those who were surgically indicated for fusion and decompression procedures, and those with neoplasms and infections were excluded from the study. ICD-10 codes used in defining the diagnosis of osteoporotic compression fracture, the operative treatment group, and exclusion criteria are specified in
Nearest-neighbor propensity score matching 2:1 cohorts for age, total comorbidities, year of procedure, calculated risk of mortality at admission, and calculated illness severity at admission was performed. Univariate analysis was used to assess differences in demographic characteristics, costs, length of stay (LOS), and complication rates between the operative and nonoperative groups. Multivariate logistic regression controlling for significant comorbidities and demographic characteristics was performed to analyze the relationship between either operative or nonoperative OCF treatment and in-hospital complication rates. Significant comorbidities included deficiency anemias, diabetes status, complicated hypertension, fluid and electrolyte disorders, neurological disorders, obesity, renal failure, weight loss, and osteoporosis (
Cohort matching resulted in 14,850 patients who underwent operative treatment and 29,700 patients who underwent nonoperative treatment. The mean age of the patients in the operative (79.3% female) and nonoperative (82.9% female) treatment groups was 79.7 years and 79.6 years, respectively. All demographic and hospitalization characteristics are comprehensively outlined in
On univariate analysis, the operative group demonstrated significantly lower rates of respiratory complications, such as pulmonary embolism (PE) (odds ratio [OR], 0.68; 95% confidence interval [CI], 0.56–0.83;
On multivariate analysis, operative treatment continued to demonstrate a significantly decreased risk of complications as previously observed. Significantly lower rates of pneumonia (OR, 0.75; 95% CI, 0.67–0.84;
Operative treatment was associated with a higher mean total cost of admission and mean LOS at $76,030 and 6.73 days, respectively, compared with nonoperative treatment at $46,257 and 5.54 days, respectively (both
In this study, we found that patients hospitalized for OCFs who undergo conservative treatment have an increased rate of acute in-hospital complications, including respiratory complications, cardiac complications, sepsis, and pressure ulcerations, compared with those who undergo acute PVA within 1 week. After PVA, patients are encouraged to mobilize and have been shown to have decreased pain [
Our findings are supported by several studies that examined the morbidity and mortality benefits of kyphoplasty and vertebroplasty. Eddin et al. [
In a prospective study, Hoshino et al. have found that patients treated with kyphoplasty within 2 months of a painful OCF were less likely to have a decrease in activities of daily living [
Conversely, several publications refute these beneficial outcomes. In a randomized controlled trial, Firanescu et al. [
This study has several limitations. It is a national database study, so it lacks granularity on the patient level. Because of this, we could not report on fracture characteristics or radiographic parameters. An inherent selection bias exists when choosing patients for operative versus nonoperative management. A subset of patients selected for nonoperative management are likely too sick for operative treatment, despite having operative indications. While we matched our cohorts based on several comorbidities and controlled for many potential confounding factors, it is possible that we did not exclude this subset of patients. Additionally, although we believe that early mobilization may contribute to the improved morbidity benefit of acute PVA, we did not have data that support this theory because it is not recorded in the NIS database. Finally, the NIS database is an inpatient-only database, so we could not capture outpatient procedures or long-term complications following initial hospitalization.
Patients who undergo acute PVA for OCF have lower rates of respiratory complications, cardiac complications, sepsis, and pressure ulcerations, while having a higher rate of acute renal failure. This adds to the body of literature that supports that PVA offers morbidity and mortality benefits compared with nonoperative management.
No potential conflict of interest relevant to this article was reported.
ICD-10 codes used to define vertebral compression fracture, operative treatment, and exclusion criteria
Diagnosis | ICD-10 codes |
---|---|
Osteoporotic vertebral compression fracture | M80.08XA, M80.88XA |
Operative treatment | 0PU43JZ, 0PU44JZ, 0QU03JZ, 0QU04JZ, 0PS43ZZ, 0PU43JZ, 0QS03ZZ, 0QU03JZ, 0PU43JZ, 0PU44JZ, 0QU03JZ, 0QU04JZ |
Exclusion criteria | |
Fusion procedures | 0SG3070, 0SG3071, 0SG307J, 0SG30A0, OSG30AJ, 0SG30J0, 0SG30J1, 0SG30JJ, 0SG30K0, 0SG30K1, 0SG30KJ, 0SG3370, 0SG3371, 0SG337J, 0SG33A0, 0SG33AJ, 0SG33J0, 0SG33J1, 0SG33JJ, 0SG33K0, 0SG33K1, 0SG33KJ, 0SG3470, 0SG3471, 0SG347J, 0SG34A0, 0SG34AJ, 0SG34J0, 0SG34J1, 0SG34JJ, 0SG34K0, 0SG34K1, 0SG34KJ |
Spinal decompression procedures | 00NX0ZZ, 00NX3ZZ, 00NX4ZZ, 00NY0ZZ, 00NY3ZZ, 00NY4ZZ |
Neoplasm | C412, M8450XA, M8450XD, M8450XG, M8450XK, M8450XP, M8450XS, M8458XA, M8458XD, M8458XG, M8458XK, M8458XP, M8458XS, M8460XA, M8460XD, M8460XG, M8460XK, M8460XP, M8460XS, M8468XA, M8468XD, M8468XG, M8468XK, M8468XP, M8468XS, C419, C720, C729, C7940, C7949, C7951, C7952, C7989, C799, C800, C801, C802, C768, D166, D168, D169, D334, D334, D321, D337, D339, D329, D367, D369, D480, D487, D489, D492, D497, D4989, D499 |
Infection | M4620, M4621, M4622, M4623, M4624, M4625, M4526, M4627, M4628, M4630, M4631, M4632, M4633, M4634, M4635, M4536, M4637, M4638, M4639, M4640, M4641, M4642, M4643, M4644, M4645, M4546, M4647, M4648, M4649, M4650, M4651, M4652, M4653, M4654, M4655, M4556, M4657, M4658, M4659, G061, G062, G07, A1781 |
ICD-10, International Classification of Diseases, 10th version.
Prevalence of preexisting comorbidities at admission for operative and nonoperative treatment groups
Comorbidities | Operative treatment | Nonoperative treatment | |
---|---|---|---|
Presence of at least one comorbidity | 14,525 (97.8) | 29,050 (97.8) | 1 |
Deficiency anemias | 795 (5.4) | 1,955 (6.6) | 0.023 |
Chronic blood loss anemia | 130 (0.9) | 265 (0.9) | 0.936 |
Alcohol abuse | 340 (2.3) | 765 (2.6) | 0.413 |
Rheumatoid arthritis/collagen vascular disease | 1,385 (9.3) | 2,850 (9.6) | 0.683 |
Congestive heart failure | 3,615 (24.3) | 7,200 (24.2) | 0.917 |
Chronic pulmonary disease | 5,400 (36.4) | 10,575 (35.6) | 0.482 |
Coagulopathy | 965 (6.5) | 2,120 (7.1) | 0.262 |
Depression | 2,785 (18.8) | 5,405 (18.2) | 0.523 |
Diabetes (uncomplicated) | 1,745 (11.8) | 2,575 (8.7) | <0.001 |
Diabetes with chronic complications | 1,760 (11.9) | 3,140 (10.6) | 0.069 |
Drug abuse | 360 (2.4) | 780 (2.6) | 0.569 |
Hypertension (uncomplicated) | 7,570 (51.0) | 14,315 (48.2) | 0.013 |
Hypertension (complicated) | 4,070 (27.4) | 7,410 (25.0) | 0.012 |
Hypothyroidism | 3,730 (25.1) | 7,200 (24.2) | 0.365 |
Liver disease | 615 (4.1) | 1,350 (4.6) | 0.381 |
Lymphoma | 320 (2.2) | 610 (2.1) | 0.753 |
Fluid & electrolyte disorders | 5,225 (35.2) | 11,875 (40.0) | <0.001 |
Metastatic cancer | 175 (1.2) | 475 (1.6) | 0.118 |
Other neurological disorders | 1,620 (10.9) | 3,920 (13.2) | 0.002 |
Obesity | 1,260 (8.5) | 1,940 (6.5) | 0.001 |
Paralysis | 130 (0.88) | 380 (1.3) | 0.091 |
Peripheral vascular disorders | 1,555 (10.5) | 3,370 (11.4) | 0.214 |
Psychoses | 85 (0.57) | 265 (0.89) | 0.107 |
Pulmonary circulation disorders | 1,085 (7.3) | 2,430 (8.2) | 0.148 |
Renal failure | 2,945 (19.8) | 5,195 (17.5) | 0.007 |
Solid tumor without metastasis | 595 (4.0) | 1,355 (4.6) | 0.227 |
Peptic ulcer disease (without bleeding) | 160 (1.1) | 305 (1.03) | 0.825 |
Valvular disease | 1,760 (11.9) | 3,270 (11.0) | 0.237 |
Weight loss | 1,710 (11.5) | 4,460 (15.0) | <0.001 |
Acquired immunodeficiency syndrome | 5 (0.03) | 35 (0.12) | 0.211 |
Osteoporosis | 450 (3.0) | 1,220 (4.1) | 0.012 |
Smoking | 1,400 (9.4) | 2,855 (9.6) | 0.779 |
Values are presented as number (%). All variables with corresponding
Denotes significance.
Mean demographic characteristics of operative and nonoperative treatment groups
Characteristic | Operative treatment (n=14,850) | Nonoperative treatment (n=29,700) | |
---|---|---|---|
Age (yr) | 79.69 | 79.62 | 0.745 |
Length of stay (day) | 6.73 | 5.54 | <0.001 |
No. of diagnoses | 16.32 | 16.78 | <0.001 |
No. of procedures | 2.97 | 0.78 | <0.001 |
No. of ICD-10-CDM External Cause of Morbidity | 0.47 | 0.53 | 0.296 |
No. of days from admission to procedure | 3.34 | 2.31 | <0.001 |
Total charge (USD) | 76,030 | 46,257 | <0.001 |
Sex (female) | 11,780 (79.3) | 24,630 (82.9) | <0.001 |
Race | 0.001 | ||
White | 12,785 (88.3) | 24,305 (84.9) | |
Black | 280 (1.9) | 825 (2.9) | |
Hispanic | 750 (5.2) | 1,730 (6.04) | |
Asian or Pacific Islander | 395 (2.7) | 1,100 (3.8) | |
Native American | 50 (0.35) | 100 (0.35) | |
Other | 215 (1.5) | 570 (2.0) | |
Weekend admission | 3,595 (24.2) | 7,460 (25.9) | 0.076 |
Discharge Disposition |
<0.001 | ||
Routine | 2,990 (20.2) | 6,450 (21.7) | |
Transfer to short-term hospital | 100 (0.67) | 605 (2.04) | |
Other transfers (nursing, intermediate, etc.) | 8,180 (55.1) | 15,570 (52.5) | |
Home health care | 3,455 (23.3) | 6,340 (21.4) | |
Against medical advice | 10 (0.07) | 120 (0.4) | |
In-hospital death | 100 (0.67) | 600 (2.0) | |
Transfer in status |
0.011 | ||
Not transferred in | 13,500 (91.4) | 26,510 (89.6) | |
Transferred in from different acute care hospital | 715 (4.8) | 1,565 (5.3) | |
Transferred in from another type of health facility | 555 (3.8) | 1,505 (5.1) | |
Transfer out status |
<0.001 | ||
Not transferred out | 6,555 (44.2) | 13,510 (45.5) | |
Transferred out to a different acute care hospital | 100 (0.67) | 605 (2.0) | |
Transferred out to another type of health facility | 8,180 (55.1) | 15,570 (52.5) | |
Primary payment method | 0.051 | ||
Medicare | 13,315 (89.7) | 26,375 (88.9) | |
Medicaid | 320 (2.2) | 910 (3.1) | |
Private insurance | 995 (6.7) | 1,935 (6.7) | |
Self-pay | 80 (0.54) | 165 (0.56) | |
No charge | 0 | 35 (0.12) | |
Other payment | 140 (0.94) | 210 (0.71) | |
Patient location | <0.001 | ||
Central counties of metro areas >1 million | 3,775 (25.4) | 8,475 (28.6) | |
Fringe counties of metro areas >1 million | 4,305 (29.0) | 7,210 (24.3) | |
Counties in metro areas 250,000–1,000,000 | 3,135 (21.13) | 6,130 (20.7) | |
Counties in metro areas 50,000–250,000 | 1,700 (11.5) | 3,045 (10.3) | |
Micropolitan counties | 1,085 (7.3) | 2,805 (9.5) | |
Not metropolitan or micropolitan | 840 (5.7) | 2,000 (6.7) | |
Median household income | 0.678 | ||
0–25th percentile | 3,065 (20.9) | 6,030 (20.6) | |
26–50th percentile | 4,025 (27.4) | 7,780 (26.6) | |
51–75th percentile | 4,040 (27.5) | 8,100 (27.6) | |
76–100th percentile | 3,550 (24.2) | 7,395 (25.2) | |
Location/teaching status of hospital | <0.001 | ||
Rural | 895 (6.03) | 3,190 (10.7) | |
Urban, non-teaching | 3,630 (24.4) | 6,890 (23.2) | |
Urban, teaching | 10,325 (69.5) | 19,620 (66.1) | |
Hospital bed size capacity | <0.001 | ||
Small | 2,460 (16.6) | 6,460 (21.8) | |
Medium | 4,805 (32.4) | 8,510 (28.7) | |
Large | 7,585 (51.1) | 14,730 (49.6) | |
Hospital region | <0.001 | ||
Northeast | 2,305 (15.5) | 6,540 (22.02) | |
Midwest | 5,595 (37.7) | 7,730 (26.03) | |
South | 5,295 (35.7) | 8,135 (27.4) | |
West | 1,655 (11.1) | 7,295 (24.6) |
Values are presented as number or number (%).
Denotes significance.
Not included in multivariate analysis.
Comparison of complication rates between operative and nonoperative treatment groups using univariate and multivariate analysis with corresponding odds ratios and
Complications | Operative treatment | Nonoperative treatment | Univariate analysis | Multivariate analysis | ||
---|---|---|---|---|---|---|
|
| |||||
OR (95% CI) | OR (95% CI) | |||||
Deep vein thrombosis | 160 (1.1) | 380 (1.3) | 0.84 (0.70–1.01) | 0.067 | 1.11 (0.85–1.47) | 0.446 |
| ||||||
Pulmonary embolism | 135 (0.91) | 395 (1.3) | 0.68 (0.56–0.83) | <0.001 |
0.77 (0.57–1.04) | 0.089 |
| ||||||
Cerebral infarct | 100 (0.67) | 325 (1.1) | 0.61 (0.49–0.77) | <0.001 |
0.44 (0.32–0.62) | <0.001 |
| ||||||
Pneumonia | 1,100 (7.41) | 3,345 (11.3) | 0.63 (0.59–0.68) | <0.001 |
0.75 (0.67–0.84) | <0.001 |
| ||||||
Acute respiratory failure | 1,120 (7.5) | 2,625 (8.8) | 0.84 (0.78–0.91) | <0.001 |
0.84 (0.74–0.96) | 0.009 |
| ||||||
Myocardial infarction | 110 (0.74) | 640 (2.2) | 0.34 (0.28–0.42) | <0.001 |
0.20 (0.15–0.26) | <0.001 |
| ||||||
Cardiac arrest | 30 (0.2) | 115 (0.39) | 0.52 (0.35–0.78) | 0.001 |
0.58 (0.33–1.03) | 0.061 |
| ||||||
Acute heart failure | 805 (5.4) | 1,940 (6.5) | 0.82 (0.75–0.89) | <0.001 |
0.80 (0.69–0.91) | 0.001 |
| ||||||
Cardiogenic shock | 15 (0.1) | 75 (0.25) | 0.40 (0.23–0.70) | 0.001 |
0.23 (0.10–0.53) | 0.001 |
| ||||||
Acute renal failure | 2,265 (15.3) | 4,175 (14.6) | 1.10 (1.04–1.16) | 0.001 |
1.19 (1.08–1.31) | <0.001 |
| ||||||
Urinary tract infection | 2,700 (18.2) | 5,775 (19.4) | 0.92 (0.88–0.97) | 0.001 |
1.02 (0.94–1.11) | 0.581 |
| ||||||
Sepsis | 430 (2.9) | 1,990 (6.7) | 0.42 (0.37–0.46) | <0.001 |
0.39 (0.34–0.45) | <0.001 |
| ||||||
Septic shock | 95 (0.64) | 325 (1.09) | 0.58 (0.46–0.73) | <0.001 |
0.50 (0.37–0.67) | <0.001 |
| ||||||
Pressure ulcer | 355 (2.4) | 1,020 (3.4) | 0.69 (0.61–0.78) | <0.001 |
0.71 (0.60–0.85) | <0.001 |
| ||||||
Radiculopathy | 335 (2.3) | 540 (1.8) | 1.25 (1.09–1.43) | 0.002 |
1.10 (0.89–1.36) | 0.363 |
Values are presented as number (%), unless otherwise stated.
OR, odds ratio; CI, confidence interval.
Denotes significance.