Back to Program
A Simple Venous Thromboembolism Risk Score for Ambulatory Surgery Patients
Christopher Pannucci, MD MS, Amy Shanks, MS, Marc Moote, PA-C, Vinita Bahl, DMD, Norah Naughton, MD, Peter Henke, MD, Sachin Kheterpal, MD MBA, Paul Cederna, MD, Darrell Campbell, MD PhD.
University of Michigan, Ann Arbor, MI, USA.
An increasing proportion of plastic and reconstructive surgery is being performed in the outpatient setting. Factors which contribute to venous thromboembolism (VTE) risk after outpatient surgery are unknown. We used the National Surgical Quality Improvement Program (NSQIP) database to examine risk factors for VTE after outpatient surgery and empirically derive a VTE risk-scoring model.
NSQIP is a prospective database of surgical patients with 30-day outcomes. Inclusion criteria for this analysis were age ≥18, surgery classified as “outpatient”, and length of stay equal to zero days.
Independent variables included known VTE risk factors (Table 1). Age, operative time, and body mass index were transformed to categorical variables to facilitate risk-score creation.
Trained NSQIP clinical nurses collect risk factor and adverse event data using medical record review. Mandatory nurse-patient contact on post-operative day 30 identifies complications treated at other hospitals. NSQIP defines deep vein thrombosis (DVT) as venous clots requiring either systemic anticoagulation or IVC filter. Pulmonary embolus (PE) is defined as an obstructing pulmonary arterial clot. Imaging is required for DVT or PE diagnosis. The primary study outcome was VTE, generated as a composite of DVT and/or PE.
Multivariable logistic regression identified independent risk factors. ß-coefficients for independent predictors were used to derive a weighted risk-scoring model; this was compared to the unweighted risk-scoring model using the c-statistic.
A total of 168,518 patients met inclusion criteria. DVT incidence was 0.1% (172 patients), PE was 0.04% (38 patients) and VTE was 0.12% (210 patients). Of patients with VTE, 1 in 10 (0.013% overall) had both DVT and PE. Based on primary CPT codes, over 39,000 patients (23% of total population) included in the analysis had operations which are commonly performed by plastic surgeons.
Independent predictors of VTE included arthroscopic surgery, current pregnancy, active cancer, and invasive venous procedure. When compared to the reference group, age 41-60, age>60, BMI>40, and operative time>120 minutes were also independent predictors (Table 1). The model accounted for 80% of the variability in VTE (c-statistic 0.800). The average time-to-event for both DVT and PE was post-operative day 10 ± 7.
C-statistic for weighted risk scores (0.76 ± 0.02) was significantly higher than the unweighted risk score (0.72 ± 0.02). The weighted risk-score model is shown in Figure 1.
30-day VTE risk can be quantified in the outpatient surgery population using a simple risk-scoring model. Aggressive chemoprophylaxis may be considered in patients with higher risk. However, further research is necessary to examine the risks, benefits, and cost of chemoprophylaxis after outpatient surgery.
|Risk Factor||Adjusted Odds Ratio (95% CI)||p value|
|Male gender||1.04 (0.77-1.40)||0.819|
|General anesthesia||1.38 (0.95-2.00)||0.091|
|Arthroscopic surgery||4.87 (2.88-8.21)||<0.001|
|Abdominal laparoscopy||1.15 (0.72-1.82)||0.555|
|Current pregnancy||8.91 (1.11-71.24)||0.035|
|Active cancer||5.38 (2.33-12.41)||<0.001|
|Congestive heart failure||3.73 (0.43-32.21)||0.231|
|Diabetes requiring medication||0.86 (0.51-1.45)||0.567|
|Central vascular disease||1.40 (0.84-2.33)||0.192|
|Peripheral vascular disease||0.73 (0.17-3.14)||0.642|
|Current smoker||0.99 (0.67-1.46)||0.940|
|Renal failure on dialysis||0.70 (0.16-3.09)||0.636|
|Prior operation within 30 days||1.09 (0.40-2.99)||0.872|
|Invasive venous procedure||13.42 (9.56-18.84)||<0.001|
|Body mass index|
|Total operative time|
Back to Program