- Email: [email protected]
Caprini Risk Model Decreases Venous Thromboembolism Rates in Thoracic Surgery Cancer Patients
Caprini Risk Model Decreases Venous Thromboembolism Rates in Thoracic Surgery Cancer Patients Helene M. Sterbling, MA, Amy K. Rosen, PhD, Krista J. Hachey, MD, Niru S. Vellanki, BS, Philip D. Hewes, MD, MPH, Sowmya R. Rao, PhD, Emma Pinjic, MPH, Hiran C. Fernando, MBBS, and Virginia R. Litle, MD School of Medicine, Department of Surgery, and Division of Thoracic Surgery, Boston University School of Medicine, Boston; VA Boston Healthcare System, Center for Healthcare Organization and Implementation Research, Boston, Massachusetts; Department of Surgery, University of Rochester Medical Center, Rochester, New York; and Department of Surgery, Inova Fairfax Medical Campus, Falls Church, Virginia
Background. Extended postoperative chemoprophylaxis is effective in reducing venous thromboembolism (VTE) among general surgical patients. We hypothesized that implementation of the Caprini risk assessment model (RAM) would reduce VTE rates among patients undergoing lung and esophageal cancer surgery. Methods. The Caprini RAM, consisting of patient risk stratification and extended postoperative chemoprophylaxis with low molecular weight heparin, was implemented on the thoracic surgery service at Boston Medical Center in July 2014. Patients undergoing lung and esophageal cancer resections were enrolled in the postintervention group beginning in July 2014. Provider and patient adherence to treatment protocol was audited. Venous thromboembolism and adverse bleeding events were monitored for 60 days postoperatively. A preintervention control group including esophagectomy and lung cancer resection patients (January 2005 to June 2013) was used for VTE rate comparison. Exclusion
criteria included chronic anticoagulation and presence of filters. Results. There were 302 lung and esophageal cancer resection patients in the preintervention cohort, and 64 thoracic cancer resections in the postintervention group. The overall VTE rates for preintervention and postintervention cohorts were 7.3% (22 of 302) and 3.1% (2 of 64), respectively (p [ 0.28). Provider adherence to Caprini RAM score assignment was 100%, whereas patient adherence to treatment was 97.4%. There were no adverse bleeding events. Conclusions. This study demonstrates a trend toward decreased symptomatic VTE after Caprini RAM implementation, as demonstrated among high-risk cancer patients. The absence of bleeding complications and high provider and patient adherence to VTE RAM support the safety and feasibility of a VTE prevention protocol in thoracic surgery patients.
V
esophageal cancer resection patients can reach 20% and 14%, respectively [4, 5]. National quality leaders, such as the Agency for Healthcare Research and Quality, consider VTE events to be indicators of potentially preventable safety events and emphasize reducing these events as well as strict monitoring of VTE trends across hospitals [6]. The American College of Chest Physicians guidelines for VTE prophylaxis are used to facilitate perioperative VTE prevention; however, there is currently no consensus
enous thromboembolism (VTE) events, including pulmonary embolism and deep vein thrombosis, are potentially preventable complications of surgical interventions yet remain important sources of morbidity in postsurgical patients. Cancer patients are at increased risk for VTE, and VTE events are identified in nearly one fifth of cancer patients [1]. Although the exact pathophysiology of this heightened thrombotic propensity is not fully understood, studies show that factors such as cancer type, grade, metastases, chemotherapy, and surgical intervention all affect VTE risk in cancer patients [2, 3]. The post-VTE mortality rate among lung and
(Ann Thorac Surg 2018;-:-–-) Ó 2018 by The Society of Thoracic Surgeons
Dr Fernando discloses a financial relationship with CSA Medical and Galil.
Accepted for publication Oct 9, 2017. Presented at the Fifty-third Annual Meeting of The Society of Thoracic Surgeons, Houston, TX, Jan 21–25, 2017. Address correspondence to Dr Litle, Division of Thoracic Surgery, Department of Surgery, Boston University, 88 E Newton St, Collamore Bldg, Rm 7380, Boston, MA 02118; email: [email protected].
Ó 2018 by The Society of Thoracic Surgeons Published by Elsevier Inc.
The Appendices can be viewed in the online version of this article [https://doi.org/10.1016/j.athoracsur.2017.10.013] on http://www.annalsthoracicsurgery.org.
0003-4975/$36.00 https://doi.org/10.1016/j.athoracsur.2017.10.013
2
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
on the use of extended VTE chemoprophylaxis among thoracic surgeons [7, 8]. The Caprini risk assessment model (RAM) identifies patient-specific risk factors, each associated with a weighted number of points (Fig 1). Based on their score, patients are stratified into VTE risk categories: low-risk patients (score of 0 to 4) require no extended VTE prophylaxis, whereas patients with moderate risk (score of 5 to 8) and patients with high risk (score of 9 or more) receive as much as 10 and 30 days of extended prophylaxis, respectively [9, 10]. Cassidy and colleagues [11] demonstrated significant VTE rate reductions in vascular and general surgery after the implementation of the Caprini RAM at Boston Medical Center. We implemented the Caprini RAM in the thoracic surgery service in 2014 to determine whether it could help reduce VTE rates in high-risk lung and esophageal cancer surgery patients [11–13]. The aims of this study were to assess the feasibility and impact of the Caprini RAM on VTE rates among thoracic surgery cancer patients and to investigate the association between individual Caprini risk factors and VTE occurrence among these patients. We also sought to examine physician and patient adherence patterns to the Caprini RAM after resections.
Patients and Methods Comparative analysis of preintervention and postintervention cohorts for thoracic surgery cancer patients was performed. The preintervention group consisted of all patients who underwent esophagectomy or anatomic lung resections for cancer between January 1, 2005, and June 30, 2013; data for this group were collected retrospectively. Current procedural terminology codes were used to identify surgical patients, and the cases with International Classification of Diseases, ninth revision, codes for cancer were subsequently included in the study. The postintervention group included all prospectively enrolled postoperative patients within the Boston Medical Center thoracic surgery division between July 1, 2014, and September 30, 2016; data were collected prospectively. The installation phase, from July 2013 to the end of June 2014, was excluded from analysis. Recruitment and informed consent occurred during the first postoperative visit. Routine inpatient care included postoperative anticoagulation prophylaxis with three daily doses of subcutaneous unfractionated heparin and sequential compression device. Institutional Review Board approval was granted for both arms of the study. Inclusion criteria comprised a diagnosis of primary lung or esophageal cancer, esophagectomy or anatomic lung resections, and documented postoperative follow-up of 60 days or more. Exclusions applied to patients receiving chronic anticoagulation therapy or with an inferior vena cava filter, to patients having more than one postoperative surgical intervention, and to patients lost to follow-up. The implementation of the Caprini RAM in our thoracic division started on July 1, 2014. Attending surgeons, residents, and physician assistants were educated
Ann Thorac Surg 2018;-:-–-
on the use of the RAM. Patients’ Caprini scores were calculated and individual patient’s extended prophylaxis durations determined. For inpatient stays exceeding the RAM-designated prophylaxis duration, routine inpatient VTE prophylaxis was administered until patient discharge. Outpatient Caprini VTE prophylaxis consisted of low molecular weight heparin injections (enoxaparin 40 mg subcutaneous injections once daily if body mass index less than 40, twice daily otherwise). Patients instructed to continue their VTE prophylaxis as outpatients were educated on the daily use of enoxaparin injections through one-on-one teaching sessions with nurses. The primary endpoint of our study was the occurrence of any symptomatic VTE event within a 60-day postoperative period, confirmed by either duplex ultrasonography for extremity deep vein thrombosis, or computed tomography pulmonary angiogram scan for pulmonary embolism. Patients were not screened for asymptomatic VTE events. All confirmed VTE events in the postintervention cohort were treated with therapeutic doses of intravenous heparin or inferior vena cava filter placement. Secondary endpoints included provider adherence to Caprini scoring, category assignment, and prescription of correct VTE prophylaxis duration for each patient, all assessed by periodic chart review. Patient adherence to the Caprini regimen was evaluated by individual interviews by research staff during the first postoperative visit. Only patients from the postintervention cohort were included in the adherence analyses. Summary statistics were obtained for all variables, and preintervention and postintervention group differences were tested with two-sided c2 tests of independence or Fisher’s exact tests and Student’s t tests or Kruskal-Wallis tests. Additionally, we obtained relative risk (RR) and 95% confidence interval (CI) from bivariate regression models fit to evaluate the association of each Caprini risk factor with the occurrence of VTE. Finally, to balance the distribution of patient characteristics between the preintervention and postintervention groups, we used propensity score with caliper matching to select 64 patients (similar to the postintervention group) from the preintervention group of 302 patients. The propensity score was computed from a logistic regression model to predict presence in the postintervention group; it included sex, histology, clinical N stage, surgical approach, Caprini risk categories, and length of stay. All analyses were conducted in SAS 9.4 (SAS Institute, Cary, NC). A two-sided p value less than 0.05 was considered to be significant.
Results In the original sample, a total of 366 lung and esophageal cancer patients undergoing surgical resection were included in the study. The preintervention and postintervention arms of this study included 302 and 74 surgical cancer patients (55 lung cancers; 19 esophageal cancer), respectively. One lung cancer patient was excluded for a
Ann Thorac Surg 2018;-:-–-
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
3
Fig 1. Caprini risk factors and score risk category association. (BMI ¼ body mass index; COPD ¼ chronic obstructive pulmonary disease; IBD ¼ inflammatory bowel disease; VTE ¼ venous thromboembolism.)
nonresection pleurodesis; an additional 6 patients were excluded for having nonprimary lung malignancies. Similarly, 3 patients were excluded from the esophageal cancer group for perioperative issues rendering them ineligible. Therefore, a total of 64 thoracic cancer patients were included in the postintervention group (Table 1). The postintervention cohort, compared with the preintervention cohort, had a higher occurrence of lung adenocarcinoma (80% versus 59%, p < 0.01), higher frequency of pN2 cancer (14% versus 6%) and pN3 cancer (3% versus 1%, p ¼ 0.05), lower median length of stay (7 days versus 9, p ¼ 0.01), and lower median risk score (8 versus 9, p ¼ 0.01). Postintervention patients also had a lower incidence of abnormal pulmonary function (30% versus 40%, p ¼ 0.03) and central venous access (11% versus 40%, p < 0.0001) than the preintervention group; however, they were more likely to have a major surgery of 6 hours or more (36% versus 24%, p ¼ 0.04). A lower proportion of patients in the postimplementation cohort was included in the high-risk category (47%) than in the preintervention cohort (54%), although differences were not statistically significant (p ¼ 0.33). A total of 22 VTE events occurred within 60 postoperative days in the preintervention cohort. Two postimplementation VTEs occurred, both in the first month of Caprini RAM implementation. There was a 58% reduction in postoperative VTE incidence from 7% (22 of 302) among the preintervention cohort to 3% (2 of 64) among the postintervention cohort, although this was nonsignificant (p ¼ 0.28; Table 2). No VTE events occurred in lung and esophageal cancer resection patients between August 2014 and September 2016.
Risk factor distribution by VTE and relative risk (95% CI) is displayed in Table 3 (complete risk factor distribution are given in Appendix A). The preintervention group had a higher risk of VTE than the postintervention group (RR 2.3, 95% CI: 0. 6, 9. 7, p ¼ 0.18). Compared with patients with lung cancer, patients with esophageal cancer had a higher risk of VTE (RR 2.8, 95% CI: 1.3 to 5.9). Factors associated with the highest risk of VTE developing were sepsis (RR 10.3, 95% CI: 5.1 to 20.6, p < 0.0001), central venous access (RR 4.6, 95% CI: 1.9 to 10.7, p < 0.001), and a high Caprini risk score (RR 4.4, 95% CI: 1.5 to 12.7, p < 0.01). Congestive heart failure (RR 5.4, 95% CI: 1.6 to 18.1, p ¼ 0.05) and history of VTE (RR 5.7, 95% CI: 2.01 to 15.6, p ¼ 0.01) were also associated with a high risk of VTE development; however, these results should be considered with caution given the small sample sizes (fewer than 10 patients) available for analysis. Results from the propensity score analyses suggested that all differences, except for surgical approach, were reduced in this propensity matched sample (the 64 patients matched between the preintervention and postintervention groups; Appendix B, Tables B1–B3). Standardized differences for the variables in the propensity score model in the original sample ranged from 0.17 (surgical approach) to 1.01 (histology); in the matched sample, they ranged from 0.10 (histology) to 0.22 (length of stay). Based on these results, we conducted similar analyses on this propensity matched sample. Results for the main variable of interest (preintervention or postintervention) were similar to that in the original sample (RR 1.5, 95% CI: 0.3 to 8.7, p ¼ 0.65). Results for certain risk factors (eg, varicose veins, sepsis) were similar to those in
4
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
Ann Thorac Surg 2018;-:-–-
Table 1. Baseline Patient Characteristics by Cohort Variable Age, years, mean SD Sex Male Female Body mass index 30 kg/m2 Lung cancer diagnosis Adenocarcinoma Squamous cell carcinoma Small cell carcinoma Other non-small cellb Esophageal cancer diagnosis Adenocarcinoma Squamous cell carcinoma Other or mixed Stagingc pN0 pN1 pN2 pN3 Lung resection Pneumonectomy Lobectomy Segmentectomy Esophagectomy Median length of stay, days
Preintervention (n ¼ 302)
Postintervention (n ¼ 64)
p Valuea
64.0 10.8
63.6 10.0
0.79 0.30
144 (47.7) 158 (52.3) 82 (27.2)
26 (40.6) 38 (59.4) 11 (17.2)
139 62 6 25
38 8 0 2
(59.9) (26.7) (2.6) (10.8)
0.10 0.09
(79.2) (16.7) (0) (4.2) 0.80
51 (72.9) 18 (25.7) 1 (1.4)
13 (81.3) 3 (18.8) 0 (0) 0.05
219 57 19 2
(73.7) (19.2) (6.4) (0.7)
44 9 9 2
(68.8) (14.1) (14.1) (3.1)
17 196 19 70
(7.3) (84.5) (8.2) (100) 9
4 39 5 16
(8.3) (81.3) (10.4) (100) 7
0.76
0.01
a b Based on two-sided c2 or Fisher’s exact test and Student’s t test or Kruskal-Wallis test. Other non-small cell: bronchial carcinoid tumor (postc intervention, n ¼ 1); large cell carcinoma (preintervention, n ¼ 12; postintervention, n ¼ 1); mixed or other (preintervention, n ¼ 16). American Joint Committee on Cancer (AJCC) pathologic staging, versions 6 (2005-2010), 7 (2010-2016).
Values are n (%) unless otherwise indicated.
the original sample, albeit with higher relative risks and wider confidence intervals. All postintervention patients (n ¼ 64) received a Caprini score by discharge day, demonstrating a 100% provider adherence to Caprini score assignment. Three patients received incorrect Caprini scores (Table 4). A total of 63 patients required extended postoperative VTE prophylaxis, with one case of incorrect duration of VTE
prophylaxis prescribed. A total of 48 patients were asked to complete their Caprini VTE prophylaxis as outpatients, and 46 patients were interviewed for compliance. Of those, 45 patients reported an adherence rate of 97.2%. None of the patients involved in the above errors had either VTE or bleeding events. Corrected Caprini scores and risk categories were used for the purpose of risk analyses.
Table 2. Venous Thromboembolism Events by Cohort Variable VTE events Deep vein thrombosis Pulmonary embolism Inpatient Outpatient VTE events by cancer type Esophageal Lung a
Preintervention (n ¼ 302) 22 9 13 18 4
(7.3) (40.9) (59.1) (81.8) (18.2)
VTE ¼ venous thromboembolism.
2 0 2 2
(3.1) (0) (100) (100) 0
p Valuea 0.28
1.00 1.00
10 (45.5) 12 (54.5)
Based on two-sided c2 or Fisher’s exact test and Student’s t test or Kruskal-Wallis test.
Values are n (%).
Postintervention (n ¼ 64)
1 (50) 1 (50)
Ann Thorac Surg 2018;-:-–-
5
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
Table 3. Selected Caprini Risk Factor Distribution by Cohort and Relative Risk for Venous Thromboembolism Events Caprini Points
Caprini Risk Factor Age category, years 40–59 60–74 75 Body mass index 30 kg/m2 Swollen legs, current Varicose veins Sepsis, <1 month Congestive heart failure Serious acute lung disease, <1 month Major open surgery, >45 minutes Laparoscopic surgery, >45 minutes Prior cancerb Present cancer Confined to bed, >72 hours Central venous access History of VTE Present chemotherapy Major surgery lasting >6 hours a
No VTE (n ¼ 342)
VTE (n ¼ 24)
113 161 62 85 21 25 9 4 48 221 36 69 337 39 110 6 66 83
4 17 3 8 5 6 8 2 10 19 3 7 24 6 17 3 11 12
RR (95% CI)a
p Value 0.08
1 2 3 1 1 1 1 1 1 2 2 2 2 2 2 3 3 5
For venous thromboembolism (VTE) development.
b
(33.6) (47.9) (18.5) (24.9) (6.1) (7.3) (2.6) (1.2) (14) (64.6) (10.5) (20.2) (98.5) (11.4) (32.2) (1.8) (19.3) (24.3)
(16.7) (70.8) (12.5) (33.3) (20.8) (25) (33.3) (8.3) (41.7) (79.2) (12.5) (29.2) (100) (25) (70.8) (12.5) (45.8) (50)
2.8 1.4 1.5 3.4 3.6 10.3 5.4 3.8 2.0 1.2 1.6 2.4 4.6 5.7 3.2 2.8
Ref (1.0–8.1) (0.3–5.8) (0.6–3.3) (1.4–8.5) (1.5–8.4) (5.1–20.6) (1.6–18.1) (1.8–8.1) (0.8–5.2) (0.4–3.8) (0.7–3.6) . (1.0–5.7) (1.9–10.7) (2.01–15.6) (1.5–6.8) (1.3–6.1)
0.37 0.02 0.01 <0.0001 0.05 <0.01 0.13 0.77 0.31 . 0.08 <0.001 0.01 <0.01 <0.01
Except nonmelanoma skin cancer.
Values are n (%). CI ¼ confidence interval;
Ref ¼ reference;
RR ¼ relative risk.
Comment Our study investigated the feasibility of implementing the Caprini RAM focusing specifically on thoracic surgery lung and esophageal cancer patients at high risk for postoperative VTE. We found that post-Caprini VTE rates dropped by 58% compared with the preintervention group (7% to 3%), with no VTE events occurring in thoracic cancer patients past the first month of the Caprini RAM implementation. There were no adverse bleeding events in the postintervention group, supporting evidence that the Caprini RAM can be used safely in thoracic surgery patients [14]. Furthermore, provider and patient adherence was high, with a small provider error rate. To our knowledge, this is the first English language study to specifically investigate the effect of Caprini RAM implementation in VTE-susceptible lung and esophageal cancer resection patients.
The use of the Caprini RAM in patients undergoing abdominal and pelvic surgery for malignancy or inflammatory bowel disease is associated with a significant reduction in postoperative outpatient VTE events, despite variation in provider adherence rates to the VTE prophylaxis prescription [15]. Schmeler and colleagues [16] investigated the implementation of 30 days of extended VTE prophylaxis for patients undergoing laparotomy for gynecologic malignancies and showed a significant decrease in 30-day postoperative VTE events. Patients reported a 79% patient compliance with outpatient treatment [16]. Our study demonstrated the feasibility of the Caprini RAM implementation, with a subsequent 58% decrease in VTE rates on the thoracic surgery service. Although our findings demonstrate a trend in VTE reduction and are not statistically significant, they support the clinical significance of the Caprini RAM
Table 4. Caprini Risk Assessment Model Errors and Nonadherence Cases Type of Noncompliance or Error Provider calculation of Caprini score Provider prescription of VTE prophylaxis Patient nonadherence RAM ¼ risk assessment model;
Source Missing risk factor; incorrect risk factor selection Novice Caprini RAM user Administration of outpatient VTE prophylaxis
VTE ¼ venous thromboembolism.
Consequence Incorrect Caprini score, risk category, and VTE prophylaxis duration Incorrect VTE prophylaxis duration prescribed Patient did not receive outpatient enoxaparin injections
6
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
implementation. Since the initial two VTE cases that occurred in the first month of the Caprini program rollout, no VTE events have occurred for 27 consecutive months, suggesting that an extended postoperative chemoprophylaxis program may be associated with the reduction in 60-day VTE rates in lung and esophageal cancer patients. A higher proportion of patients undergoing procedures of 6 hours or more was observed in the postintervention cohort. Kim and associates [17] demonstrated that duration of surgery was an independent risk factor for VTE events, with each additional hour spent in the operating room carrying significant higher odds ratios of VTE events. This research suggests that lengthy operations such as esophagectomies may inherently carry a higher risk for postoperative VTEs. In addition, Chew and colleagues [18] demonstrated a higher risk for VTE events in patients with metastatic disease at the time of diagnosis, indicating that our postintervention group may have been more susceptible to VTE events compared with the pre-Caprini cohort [18]. Despite a significant increase in VTE risk factors, such as lengthy surgical procedures and advanced metastatic disease, the prevalence of VTE events in our postintervention cohort was reduced by half within the first year of implementation. We also examined which Caprini risk factors were associated with VTE events in our cancer patients. Evidence of varicose veins, swollen legs, neoadjuvant chemotherapy, serious lung disease, and surgery 6 hours or more, all had significant associations with VTE development. Having a high Caprini risk score was also significantly associated with VTE development (RR 4.4, 95% CI: 1.5 to 12.7, p < 0.01). Modification of the Caprini risk factors could help stratify our thoracic surgery cancer patients in a more appropriate manner. Finally, the high adherence rates observed in our cohort reinforce the feasibility of the Caprini RAM in a high-volume, safetynet hospital. Limitations associated with our study warrant noting. First, retrospective studies inherently rely on preexisting medical documentation and selection bias. We attempted to mitigate the effects of selection bias by analyzing propensity score matched data, which showed similar results to our original sample analysis. Additionally, our study design included anatomic lung resection but did not consider wedge resection procedures, possibly missing a number of lung cancer patients. Our small postintervention sample size may have led to the occurrence of type II error, thus missing a potential association between Caprini RAM implementation and VTE reduction in thoracic cancer patients, particularly in our small post-Caprini esophageal cancer group of only 16 patients. Although we reported 100% provider adherence to assigning Caprini scores, several errors were noted. The first type of error occurred during Caprini score calculation, ultimately leading to underprescription of prophylaxis duration. To improve accurate use of the Caprini RAM, a nurse manager with VTE RAM ownership could be
Ann Thorac Surg 2018;-:-–-
designated, or an autopopulating Caprini score calculator could be designed to facilitate real-time score assignment [19]. An error-minimizing workflow also needs to be developed to reduce opportunities for prescription mistakes while using the RAM. Although no bleeding events were reported in this study, it is important to note that adverse side effects of anticoagulant VTE prophylaxis can present as thrombocytopenia, epistaxis, gastrointestinal bleeding, and wound hematoma. These presentations were not specifically investigated in this study and should be closely monitored in the clinical setting. A reduction in median hospital length of stay was noted between the postintervention (7 days) and preintervention (9 days) cohorts (p ¼ 0.01). For patients with short length of stay, pathology reports and cancer diagnoses may not be finalized at time of discharge, possibly contributing to a reduction in calculated Caprini score for those patients. In contrast, all preintervention patients had a “cancer” risk factor included in their RAM calculations owing to the retrospective nature of the data collection. These real-time errors and lack of finalized cancer diagnoses may explain differences in median Caprini scores observed between the preimplementation and postimplementation cohorts (9 versus 8 points, p ¼ 0.01), as well as the lower proportion of high-risk category patients in the postintervention group. Further research is needed to develop a thoracic surgery–specific modification of the Caprini RAM that would best address these patients’ characteristics. Suggested risk factors for a thoracic-specific Caprini RAM include cancer histology, stage, and pack-year smoking history [20]. Noting that 97.7% of the study’s esophageal cancer patients were stratified into the Caprini high-risk category, the possibility of assigning all eligible esophageal cancer resection patients to a 30-day postoperative low molecular weight heparin VTE prophylaxis regimen should be investigated. In conclusion, our study is the first of its kind to specifically investigate the effect of the Caprini RAM in lung and esophageal cancer resection patients at high risk for postoperative VTE events. Although high provider and patient adherence to the RAM methodologies were observed, continued efforts are needed to achieve consistent compliance throughout the VTE prevention process. The implementation of the Caprini RAM in the thoracic surgery service was feasible, leading to a 58% reduction in 60-day postoperative symptomatic VTE incidence among lung and esophageal cancer patients. Although the Caprini RAM may have helped the thoracic surgery team achieve 2 consecutive years of VTE-free care for high-risk cancer patients, multicenter studies with larger patient samples are still needed to establish the tangible effects of any VTE prevention efforts. It is an opportune time for the thoracic surgery community to apply a riskstratification program to guide duration of extended chemoprophylaxis and reduce postoperative VTE for our high-risk patients.
Ann Thorac Surg 2018;-:-–-
The authors wish to thank Juan A. Mu~ noz-Largacha, MD, Boston University School of Medicine, Division of Thoracic Surgery, Qi Chen, PhD, Department of Surgery, and Amanda Meister, Physician Assistant, Division of Thoracic Surgery, Boston Medical Center, for their assistance with this study. Cardinal Health Foundation Grant funded this research project. The Clinical and Translational Science Institute at Boston University is funded by the National Institutes of Health Clinical and Translational Science Award program, grant U54TR001012.
Audio Discussion: Audio of the discussion that followed the presentation of this paper at the STS Annual Meeting can be accessed in the online version of this article [https://doi.org/10.1016/j.athoracsur.2017. 10.013] on http://www.annalsthoracicsurgery.org.
References 1. Blom JW, Doggen CJM, Osanto S, et al. Malignancies, prothrombotic mutations, and the risk of venous thrombosis. JAMA 2005;293:715–22. 2. Tesselaar ME, Romijn FP, Van Der Linden IK, Prins FA, Bertina RM, Osanto S. Microparticle-associated tissue factor activity: a link between cancer and thrombosis? J Thromb Haemost 2007;5:520–7. 3. Qureshi W, Ali Z, Amjad W, et al. Venous thromboembolism in cancer: an update of treatment and prevention in the era of newer anticoagulants. Front Cardiovasc Med 2016;3:24. 4. Merkow RP, Bilimoria KY, McCarter MD, et al. Postdischarge venous thromboembolism after cancer surgery. Ann Surg 2011;254:131–7. 5. Trinh VQ, Karakiewicz PI, Sammon J, et al. Venous thromboembolism after major cancer surgery: temporal trends and patterns of care. JAMA Surg 2013;2115:1–7. 6. Maynard G. Preventing hospital-associated venous thromboembolism: a guide for effective quality improvement. 2nd ed. Rockville, MD: Agency for Healthcare Research and Quality, AHRQ publication no. 16-0001-EF; October 2015. 7. Gould MK, Garcia DA, Wren SM, et al. Prevention of VTE in nonorthopedic surgical patients. Antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest 2012;141(Suppl):227–77.
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
7
8. Zwischenberger BA, Tzeng C-WD, Ward ND, Zwischenberger JB, Martin JT. Venous thromboembolism prophylaxis for esophagectomy: a survey of practice patterns among thoracic surgeons. Ann Thorac Surg 2016;101:489–94. 9. Pannucci C, Bailey S. Validation of the Caprini risk assessment model in plastic and reconstructive surgery patients. J Am Coll Surg 2011;212:105–12. 10. Caprini JA. Risk assessment as a guide for the prevention of the many faces of venous thromboembolism. Am J Surg 2010;199(Suppl):3–10. 11. Cassidy MR, Rosenkranz P, McAneny D. Reducing postoperative venous thromboembolism complications with a standardized risk-stratified prophylaxis protocol and mobilization program. J Am Coll Surg 2014;218:1095–104. 12. Hachey KJ, Hewes PD, Porter LP, et al. Caprini venous thromboembolism risk assessment permits selection for postdischarge prophylactic anticoagulation in patients with resectable lung cancer. J Thorac Cardiovasc Surg 2016;151: 37–44. 13. Hewes PD, Hachey KJ, Zhang XW, et al. Evaluation of the Caprini model for venothromboembolism in esophagectomy patients. Ann Thorac Surg 2015;100:2072–8. 14. Hachey KJ, Sterbling H, Choi DS, et al. Prevention of postoperative venous thromboembolism in thoracic surgical patients: implementation and evaluation of a Caprini risk assessment protocol. J Am Coll Surg 2016;222:1019–27. 15. Najjar PA, Madenci AL, Zogg CK, et al. Implementation of a comprehensive post-discharge venous thromboembolism prophylaxis program for abdominal and pelvic surgery patients. J Am Coll Surg 2016;223:804–13. 16. Schmeler KM, Wilson GL, Cain K, et al. Venous thromboembolism (VTE) rates following the implementation of extended duration prophylaxis for patients undergoing surgery for gynecologic malignancies. Gynecol Oncol 2013;128: 204–8. 17. Kim JYS, Khavanin N, Rambachan A, et al. Surgical duration and risk of venous thromboembolism. JAMA Surg 2015;150: 110–7. 18. Chew HK, Wun T, Harvey D, Zhou H, White RH. Incidence of venous thromboembolism and its effect on survival among patients with common cancers. Arch Intern Med 2006;166: 458–64. 19. Schiro TA, Sakowski J, Romanelli RJ, et al. Improving adherence to best-practice guidelines for venous thromboembolism risk assessment and prevention. Am J Heal Pharm 2011;68:2184–9. 20. Mason DP, Quader MA, Blackstone EH, et al. Thromboembolism after pneumonectomy for malignancy: an independent marker of poor outcome. J Thorac Cardiovasc Surg 2006;131:711–8.
Background. Extended postoperative chemoprophylaxis is effective in reducing venous thromboembolism (VTE) among general surgical patients. We hypothesized that implementation of the Caprini risk assessment model (RAM) would reduce VTE rates among patients undergoing lung and esophageal cancer surgery. Methods. The Caprini RAM, consisting of patient risk stratification and extended postoperative chemoprophylaxis with low molecular weight heparin, was implemented on the thoracic surgery service at Boston Medical Center in July 2014. Patients undergoing lung and esophageal cancer resections were enrolled in the postintervention group beginning in July 2014. Provider and patient adherence to treatment protocol was audited. Venous thromboembolism and adverse bleeding events were monitored for 60 days postoperatively. A preintervention control group including esophagectomy and lung cancer resection patients (January 2005 to June 2013) was used for VTE rate comparison. Exclusion
criteria included chronic anticoagulation and presence of filters. Results. There were 302 lung and esophageal cancer resection patients in the preintervention cohort, and 64 thoracic cancer resections in the postintervention group. The overall VTE rates for preintervention and postintervention cohorts were 7.3% (22 of 302) and 3.1% (2 of 64), respectively (p [ 0.28). Provider adherence to Caprini RAM score assignment was 100%, whereas patient adherence to treatment was 97.4%. There were no adverse bleeding events. Conclusions. This study demonstrates a trend toward decreased symptomatic VTE after Caprini RAM implementation, as demonstrated among high-risk cancer patients. The absence of bleeding complications and high provider and patient adherence to VTE RAM support the safety and feasibility of a VTE prevention protocol in thoracic surgery patients.
V
esophageal cancer resection patients can reach 20% and 14%, respectively [4, 5]. National quality leaders, such as the Agency for Healthcare Research and Quality, consider VTE events to be indicators of potentially preventable safety events and emphasize reducing these events as well as strict monitoring of VTE trends across hospitals [6]. The American College of Chest Physicians guidelines for VTE prophylaxis are used to facilitate perioperative VTE prevention; however, there is currently no consensus
enous thromboembolism (VTE) events, including pulmonary embolism and deep vein thrombosis, are potentially preventable complications of surgical interventions yet remain important sources of morbidity in postsurgical patients. Cancer patients are at increased risk for VTE, and VTE events are identified in nearly one fifth of cancer patients [1]. Although the exact pathophysiology of this heightened thrombotic propensity is not fully understood, studies show that factors such as cancer type, grade, metastases, chemotherapy, and surgical intervention all affect VTE risk in cancer patients [2, 3]. The post-VTE mortality rate among lung and
(Ann Thorac Surg 2018;-:-–-) Ó 2018 by The Society of Thoracic Surgeons
Dr Fernando discloses a financial relationship with CSA Medical and Galil.
Accepted for publication Oct 9, 2017. Presented at the Fifty-third Annual Meeting of The Society of Thoracic Surgeons, Houston, TX, Jan 21–25, 2017. Address correspondence to Dr Litle, Division of Thoracic Surgery, Department of Surgery, Boston University, 88 E Newton St, Collamore Bldg, Rm 7380, Boston, MA 02118; email: [email protected].
Ó 2018 by The Society of Thoracic Surgeons Published by Elsevier Inc.
The Appendices can be viewed in the online version of this article [https://doi.org/10.1016/j.athoracsur.2017.10.013] on http://www.annalsthoracicsurgery.org.
0003-4975/$36.00 https://doi.org/10.1016/j.athoracsur.2017.10.013
2
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
on the use of extended VTE chemoprophylaxis among thoracic surgeons [7, 8]. The Caprini risk assessment model (RAM) identifies patient-specific risk factors, each associated with a weighted number of points (Fig 1). Based on their score, patients are stratified into VTE risk categories: low-risk patients (score of 0 to 4) require no extended VTE prophylaxis, whereas patients with moderate risk (score of 5 to 8) and patients with high risk (score of 9 or more) receive as much as 10 and 30 days of extended prophylaxis, respectively [9, 10]. Cassidy and colleagues [11] demonstrated significant VTE rate reductions in vascular and general surgery after the implementation of the Caprini RAM at Boston Medical Center. We implemented the Caprini RAM in the thoracic surgery service in 2014 to determine whether it could help reduce VTE rates in high-risk lung and esophageal cancer surgery patients [11–13]. The aims of this study were to assess the feasibility and impact of the Caprini RAM on VTE rates among thoracic surgery cancer patients and to investigate the association between individual Caprini risk factors and VTE occurrence among these patients. We also sought to examine physician and patient adherence patterns to the Caprini RAM after resections.
Patients and Methods Comparative analysis of preintervention and postintervention cohorts for thoracic surgery cancer patients was performed. The preintervention group consisted of all patients who underwent esophagectomy or anatomic lung resections for cancer between January 1, 2005, and June 30, 2013; data for this group were collected retrospectively. Current procedural terminology codes were used to identify surgical patients, and the cases with International Classification of Diseases, ninth revision, codes for cancer were subsequently included in the study. The postintervention group included all prospectively enrolled postoperative patients within the Boston Medical Center thoracic surgery division between July 1, 2014, and September 30, 2016; data were collected prospectively. The installation phase, from July 2013 to the end of June 2014, was excluded from analysis. Recruitment and informed consent occurred during the first postoperative visit. Routine inpatient care included postoperative anticoagulation prophylaxis with three daily doses of subcutaneous unfractionated heparin and sequential compression device. Institutional Review Board approval was granted for both arms of the study. Inclusion criteria comprised a diagnosis of primary lung or esophageal cancer, esophagectomy or anatomic lung resections, and documented postoperative follow-up of 60 days or more. Exclusions applied to patients receiving chronic anticoagulation therapy or with an inferior vena cava filter, to patients having more than one postoperative surgical intervention, and to patients lost to follow-up. The implementation of the Caprini RAM in our thoracic division started on July 1, 2014. Attending surgeons, residents, and physician assistants were educated
Ann Thorac Surg 2018;-:-–-
on the use of the RAM. Patients’ Caprini scores were calculated and individual patient’s extended prophylaxis durations determined. For inpatient stays exceeding the RAM-designated prophylaxis duration, routine inpatient VTE prophylaxis was administered until patient discharge. Outpatient Caprini VTE prophylaxis consisted of low molecular weight heparin injections (enoxaparin 40 mg subcutaneous injections once daily if body mass index less than 40, twice daily otherwise). Patients instructed to continue their VTE prophylaxis as outpatients were educated on the daily use of enoxaparin injections through one-on-one teaching sessions with nurses. The primary endpoint of our study was the occurrence of any symptomatic VTE event within a 60-day postoperative period, confirmed by either duplex ultrasonography for extremity deep vein thrombosis, or computed tomography pulmonary angiogram scan for pulmonary embolism. Patients were not screened for asymptomatic VTE events. All confirmed VTE events in the postintervention cohort were treated with therapeutic doses of intravenous heparin or inferior vena cava filter placement. Secondary endpoints included provider adherence to Caprini scoring, category assignment, and prescription of correct VTE prophylaxis duration for each patient, all assessed by periodic chart review. Patient adherence to the Caprini regimen was evaluated by individual interviews by research staff during the first postoperative visit. Only patients from the postintervention cohort were included in the adherence analyses. Summary statistics were obtained for all variables, and preintervention and postintervention group differences were tested with two-sided c2 tests of independence or Fisher’s exact tests and Student’s t tests or Kruskal-Wallis tests. Additionally, we obtained relative risk (RR) and 95% confidence interval (CI) from bivariate regression models fit to evaluate the association of each Caprini risk factor with the occurrence of VTE. Finally, to balance the distribution of patient characteristics between the preintervention and postintervention groups, we used propensity score with caliper matching to select 64 patients (similar to the postintervention group) from the preintervention group of 302 patients. The propensity score was computed from a logistic regression model to predict presence in the postintervention group; it included sex, histology, clinical N stage, surgical approach, Caprini risk categories, and length of stay. All analyses were conducted in SAS 9.4 (SAS Institute, Cary, NC). A two-sided p value less than 0.05 was considered to be significant.
Results In the original sample, a total of 366 lung and esophageal cancer patients undergoing surgical resection were included in the study. The preintervention and postintervention arms of this study included 302 and 74 surgical cancer patients (55 lung cancers; 19 esophageal cancer), respectively. One lung cancer patient was excluded for a
Ann Thorac Surg 2018;-:-–-
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
3
Fig 1. Caprini risk factors and score risk category association. (BMI ¼ body mass index; COPD ¼ chronic obstructive pulmonary disease; IBD ¼ inflammatory bowel disease; VTE ¼ venous thromboembolism.)
nonresection pleurodesis; an additional 6 patients were excluded for having nonprimary lung malignancies. Similarly, 3 patients were excluded from the esophageal cancer group for perioperative issues rendering them ineligible. Therefore, a total of 64 thoracic cancer patients were included in the postintervention group (Table 1). The postintervention cohort, compared with the preintervention cohort, had a higher occurrence of lung adenocarcinoma (80% versus 59%, p < 0.01), higher frequency of pN2 cancer (14% versus 6%) and pN3 cancer (3% versus 1%, p ¼ 0.05), lower median length of stay (7 days versus 9, p ¼ 0.01), and lower median risk score (8 versus 9, p ¼ 0.01). Postintervention patients also had a lower incidence of abnormal pulmonary function (30% versus 40%, p ¼ 0.03) and central venous access (11% versus 40%, p < 0.0001) than the preintervention group; however, they were more likely to have a major surgery of 6 hours or more (36% versus 24%, p ¼ 0.04). A lower proportion of patients in the postimplementation cohort was included in the high-risk category (47%) than in the preintervention cohort (54%), although differences were not statistically significant (p ¼ 0.33). A total of 22 VTE events occurred within 60 postoperative days in the preintervention cohort. Two postimplementation VTEs occurred, both in the first month of Caprini RAM implementation. There was a 58% reduction in postoperative VTE incidence from 7% (22 of 302) among the preintervention cohort to 3% (2 of 64) among the postintervention cohort, although this was nonsignificant (p ¼ 0.28; Table 2). No VTE events occurred in lung and esophageal cancer resection patients between August 2014 and September 2016.
Risk factor distribution by VTE and relative risk (95% CI) is displayed in Table 3 (complete risk factor distribution are given in Appendix A). The preintervention group had a higher risk of VTE than the postintervention group (RR 2.3, 95% CI: 0. 6, 9. 7, p ¼ 0.18). Compared with patients with lung cancer, patients with esophageal cancer had a higher risk of VTE (RR 2.8, 95% CI: 1.3 to 5.9). Factors associated with the highest risk of VTE developing were sepsis (RR 10.3, 95% CI: 5.1 to 20.6, p < 0.0001), central venous access (RR 4.6, 95% CI: 1.9 to 10.7, p < 0.001), and a high Caprini risk score (RR 4.4, 95% CI: 1.5 to 12.7, p < 0.01). Congestive heart failure (RR 5.4, 95% CI: 1.6 to 18.1, p ¼ 0.05) and history of VTE (RR 5.7, 95% CI: 2.01 to 15.6, p ¼ 0.01) were also associated with a high risk of VTE development; however, these results should be considered with caution given the small sample sizes (fewer than 10 patients) available for analysis. Results from the propensity score analyses suggested that all differences, except for surgical approach, were reduced in this propensity matched sample (the 64 patients matched between the preintervention and postintervention groups; Appendix B, Tables B1–B3). Standardized differences for the variables in the propensity score model in the original sample ranged from 0.17 (surgical approach) to 1.01 (histology); in the matched sample, they ranged from 0.10 (histology) to 0.22 (length of stay). Based on these results, we conducted similar analyses on this propensity matched sample. Results for the main variable of interest (preintervention or postintervention) were similar to that in the original sample (RR 1.5, 95% CI: 0.3 to 8.7, p ¼ 0.65). Results for certain risk factors (eg, varicose veins, sepsis) were similar to those in
4
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
Ann Thorac Surg 2018;-:-–-
Table 1. Baseline Patient Characteristics by Cohort Variable Age, years, mean SD Sex Male Female Body mass index 30 kg/m2 Lung cancer diagnosis Adenocarcinoma Squamous cell carcinoma Small cell carcinoma Other non-small cellb Esophageal cancer diagnosis Adenocarcinoma Squamous cell carcinoma Other or mixed Stagingc pN0 pN1 pN2 pN3 Lung resection Pneumonectomy Lobectomy Segmentectomy Esophagectomy Median length of stay, days
Preintervention (n ¼ 302)
Postintervention (n ¼ 64)
p Valuea
64.0 10.8
63.6 10.0
0.79 0.30
144 (47.7) 158 (52.3) 82 (27.2)
26 (40.6) 38 (59.4) 11 (17.2)
139 62 6 25
38 8 0 2
(59.9) (26.7) (2.6) (10.8)
0.10 0.09
(79.2) (16.7) (0) (4.2) 0.80
51 (72.9) 18 (25.7) 1 (1.4)
13 (81.3) 3 (18.8) 0 (0) 0.05
219 57 19 2
(73.7) (19.2) (6.4) (0.7)
44 9 9 2
(68.8) (14.1) (14.1) (3.1)
17 196 19 70
(7.3) (84.5) (8.2) (100) 9
4 39 5 16
(8.3) (81.3) (10.4) (100) 7
0.76
0.01
a b Based on two-sided c2 or Fisher’s exact test and Student’s t test or Kruskal-Wallis test. Other non-small cell: bronchial carcinoid tumor (postc intervention, n ¼ 1); large cell carcinoma (preintervention, n ¼ 12; postintervention, n ¼ 1); mixed or other (preintervention, n ¼ 16). American Joint Committee on Cancer (AJCC) pathologic staging, versions 6 (2005-2010), 7 (2010-2016).
Values are n (%) unless otherwise indicated.
the original sample, albeit with higher relative risks and wider confidence intervals. All postintervention patients (n ¼ 64) received a Caprini score by discharge day, demonstrating a 100% provider adherence to Caprini score assignment. Three patients received incorrect Caprini scores (Table 4). A total of 63 patients required extended postoperative VTE prophylaxis, with one case of incorrect duration of VTE
prophylaxis prescribed. A total of 48 patients were asked to complete their Caprini VTE prophylaxis as outpatients, and 46 patients were interviewed for compliance. Of those, 45 patients reported an adherence rate of 97.2%. None of the patients involved in the above errors had either VTE or bleeding events. Corrected Caprini scores and risk categories were used for the purpose of risk analyses.
Table 2. Venous Thromboembolism Events by Cohort Variable VTE events Deep vein thrombosis Pulmonary embolism Inpatient Outpatient VTE events by cancer type Esophageal Lung a
Preintervention (n ¼ 302) 22 9 13 18 4
(7.3) (40.9) (59.1) (81.8) (18.2)
VTE ¼ venous thromboembolism.
2 0 2 2
(3.1) (0) (100) (100) 0
p Valuea 0.28
1.00 1.00
10 (45.5) 12 (54.5)
Based on two-sided c2 or Fisher’s exact test and Student’s t test or Kruskal-Wallis test.
Values are n (%).
Postintervention (n ¼ 64)
1 (50) 1 (50)
Ann Thorac Surg 2018;-:-–-
5
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
Table 3. Selected Caprini Risk Factor Distribution by Cohort and Relative Risk for Venous Thromboembolism Events Caprini Points
Caprini Risk Factor Age category, years 40–59 60–74 75 Body mass index 30 kg/m2 Swollen legs, current Varicose veins Sepsis, <1 month Congestive heart failure Serious acute lung disease, <1 month Major open surgery, >45 minutes Laparoscopic surgery, >45 minutes Prior cancerb Present cancer Confined to bed, >72 hours Central venous access History of VTE Present chemotherapy Major surgery lasting >6 hours a
No VTE (n ¼ 342)
VTE (n ¼ 24)
113 161 62 85 21 25 9 4 48 221 36 69 337 39 110 6 66 83
4 17 3 8 5 6 8 2 10 19 3 7 24 6 17 3 11 12
RR (95% CI)a
p Value 0.08
1 2 3 1 1 1 1 1 1 2 2 2 2 2 2 3 3 5
For venous thromboembolism (VTE) development.
b
(33.6) (47.9) (18.5) (24.9) (6.1) (7.3) (2.6) (1.2) (14) (64.6) (10.5) (20.2) (98.5) (11.4) (32.2) (1.8) (19.3) (24.3)
(16.7) (70.8) (12.5) (33.3) (20.8) (25) (33.3) (8.3) (41.7) (79.2) (12.5) (29.2) (100) (25) (70.8) (12.5) (45.8) (50)
2.8 1.4 1.5 3.4 3.6 10.3 5.4 3.8 2.0 1.2 1.6 2.4 4.6 5.7 3.2 2.8
Ref (1.0–8.1) (0.3–5.8) (0.6–3.3) (1.4–8.5) (1.5–8.4) (5.1–20.6) (1.6–18.1) (1.8–8.1) (0.8–5.2) (0.4–3.8) (0.7–3.6) . (1.0–5.7) (1.9–10.7) (2.01–15.6) (1.5–6.8) (1.3–6.1)
0.37 0.02 0.01 <0.0001 0.05 <0.01 0.13 0.77 0.31 . 0.08 <0.001 0.01 <0.01 <0.01
Except nonmelanoma skin cancer.
Values are n (%). CI ¼ confidence interval;
Ref ¼ reference;
RR ¼ relative risk.
Comment Our study investigated the feasibility of implementing the Caprini RAM focusing specifically on thoracic surgery lung and esophageal cancer patients at high risk for postoperative VTE. We found that post-Caprini VTE rates dropped by 58% compared with the preintervention group (7% to 3%), with no VTE events occurring in thoracic cancer patients past the first month of the Caprini RAM implementation. There were no adverse bleeding events in the postintervention group, supporting evidence that the Caprini RAM can be used safely in thoracic surgery patients [14]. Furthermore, provider and patient adherence was high, with a small provider error rate. To our knowledge, this is the first English language study to specifically investigate the effect of Caprini RAM implementation in VTE-susceptible lung and esophageal cancer resection patients.
The use of the Caprini RAM in patients undergoing abdominal and pelvic surgery for malignancy or inflammatory bowel disease is associated with a significant reduction in postoperative outpatient VTE events, despite variation in provider adherence rates to the VTE prophylaxis prescription [15]. Schmeler and colleagues [16] investigated the implementation of 30 days of extended VTE prophylaxis for patients undergoing laparotomy for gynecologic malignancies and showed a significant decrease in 30-day postoperative VTE events. Patients reported a 79% patient compliance with outpatient treatment [16]. Our study demonstrated the feasibility of the Caprini RAM implementation, with a subsequent 58% decrease in VTE rates on the thoracic surgery service. Although our findings demonstrate a trend in VTE reduction and are not statistically significant, they support the clinical significance of the Caprini RAM
Table 4. Caprini Risk Assessment Model Errors and Nonadherence Cases Type of Noncompliance or Error Provider calculation of Caprini score Provider prescription of VTE prophylaxis Patient nonadherence RAM ¼ risk assessment model;
Source Missing risk factor; incorrect risk factor selection Novice Caprini RAM user Administration of outpatient VTE prophylaxis
VTE ¼ venous thromboembolism.
Consequence Incorrect Caprini score, risk category, and VTE prophylaxis duration Incorrect VTE prophylaxis duration prescribed Patient did not receive outpatient enoxaparin injections
6
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
implementation. Since the initial two VTE cases that occurred in the first month of the Caprini program rollout, no VTE events have occurred for 27 consecutive months, suggesting that an extended postoperative chemoprophylaxis program may be associated with the reduction in 60-day VTE rates in lung and esophageal cancer patients. A higher proportion of patients undergoing procedures of 6 hours or more was observed in the postintervention cohort. Kim and associates [17] demonstrated that duration of surgery was an independent risk factor for VTE events, with each additional hour spent in the operating room carrying significant higher odds ratios of VTE events. This research suggests that lengthy operations such as esophagectomies may inherently carry a higher risk for postoperative VTEs. In addition, Chew and colleagues [18] demonstrated a higher risk for VTE events in patients with metastatic disease at the time of diagnosis, indicating that our postintervention group may have been more susceptible to VTE events compared with the pre-Caprini cohort [18]. Despite a significant increase in VTE risk factors, such as lengthy surgical procedures and advanced metastatic disease, the prevalence of VTE events in our postintervention cohort was reduced by half within the first year of implementation. We also examined which Caprini risk factors were associated with VTE events in our cancer patients. Evidence of varicose veins, swollen legs, neoadjuvant chemotherapy, serious lung disease, and surgery 6 hours or more, all had significant associations with VTE development. Having a high Caprini risk score was also significantly associated with VTE development (RR 4.4, 95% CI: 1.5 to 12.7, p < 0.01). Modification of the Caprini risk factors could help stratify our thoracic surgery cancer patients in a more appropriate manner. Finally, the high adherence rates observed in our cohort reinforce the feasibility of the Caprini RAM in a high-volume, safetynet hospital. Limitations associated with our study warrant noting. First, retrospective studies inherently rely on preexisting medical documentation and selection bias. We attempted to mitigate the effects of selection bias by analyzing propensity score matched data, which showed similar results to our original sample analysis. Additionally, our study design included anatomic lung resection but did not consider wedge resection procedures, possibly missing a number of lung cancer patients. Our small postintervention sample size may have led to the occurrence of type II error, thus missing a potential association between Caprini RAM implementation and VTE reduction in thoracic cancer patients, particularly in our small post-Caprini esophageal cancer group of only 16 patients. Although we reported 100% provider adherence to assigning Caprini scores, several errors were noted. The first type of error occurred during Caprini score calculation, ultimately leading to underprescription of prophylaxis duration. To improve accurate use of the Caprini RAM, a nurse manager with VTE RAM ownership could be
Ann Thorac Surg 2018;-:-–-
designated, or an autopopulating Caprini score calculator could be designed to facilitate real-time score assignment [19]. An error-minimizing workflow also needs to be developed to reduce opportunities for prescription mistakes while using the RAM. Although no bleeding events were reported in this study, it is important to note that adverse side effects of anticoagulant VTE prophylaxis can present as thrombocytopenia, epistaxis, gastrointestinal bleeding, and wound hematoma. These presentations were not specifically investigated in this study and should be closely monitored in the clinical setting. A reduction in median hospital length of stay was noted between the postintervention (7 days) and preintervention (9 days) cohorts (p ¼ 0.01). For patients with short length of stay, pathology reports and cancer diagnoses may not be finalized at time of discharge, possibly contributing to a reduction in calculated Caprini score for those patients. In contrast, all preintervention patients had a “cancer” risk factor included in their RAM calculations owing to the retrospective nature of the data collection. These real-time errors and lack of finalized cancer diagnoses may explain differences in median Caprini scores observed between the preimplementation and postimplementation cohorts (9 versus 8 points, p ¼ 0.01), as well as the lower proportion of high-risk category patients in the postintervention group. Further research is needed to develop a thoracic surgery–specific modification of the Caprini RAM that would best address these patients’ characteristics. Suggested risk factors for a thoracic-specific Caprini RAM include cancer histology, stage, and pack-year smoking history [20]. Noting that 97.7% of the study’s esophageal cancer patients were stratified into the Caprini high-risk category, the possibility of assigning all eligible esophageal cancer resection patients to a 30-day postoperative low molecular weight heparin VTE prophylaxis regimen should be investigated. In conclusion, our study is the first of its kind to specifically investigate the effect of the Caprini RAM in lung and esophageal cancer resection patients at high risk for postoperative VTE events. Although high provider and patient adherence to the RAM methodologies were observed, continued efforts are needed to achieve consistent compliance throughout the VTE prevention process. The implementation of the Caprini RAM in the thoracic surgery service was feasible, leading to a 58% reduction in 60-day postoperative symptomatic VTE incidence among lung and esophageal cancer patients. Although the Caprini RAM may have helped the thoracic surgery team achieve 2 consecutive years of VTE-free care for high-risk cancer patients, multicenter studies with larger patient samples are still needed to establish the tangible effects of any VTE prevention efforts. It is an opportune time for the thoracic surgery community to apply a riskstratification program to guide duration of extended chemoprophylaxis and reduce postoperative VTE for our high-risk patients.
Ann Thorac Surg 2018;-:-–-
The authors wish to thank Juan A. Mu~ noz-Largacha, MD, Boston University School of Medicine, Division of Thoracic Surgery, Qi Chen, PhD, Department of Surgery, and Amanda Meister, Physician Assistant, Division of Thoracic Surgery, Boston Medical Center, for their assistance with this study. Cardinal Health Foundation Grant funded this research project. The Clinical and Translational Science Institute at Boston University is funded by the National Institutes of Health Clinical and Translational Science Award program, grant U54TR001012.
Audio Discussion: Audio of the discussion that followed the presentation of this paper at the STS Annual Meeting can be accessed in the online version of this article [https://doi.org/10.1016/j.athoracsur.2017. 10.013] on http://www.annalsthoracicsurgery.org.
References 1. Blom JW, Doggen CJM, Osanto S, et al. Malignancies, prothrombotic mutations, and the risk of venous thrombosis. JAMA 2005;293:715–22. 2. Tesselaar ME, Romijn FP, Van Der Linden IK, Prins FA, Bertina RM, Osanto S. Microparticle-associated tissue factor activity: a link between cancer and thrombosis? J Thromb Haemost 2007;5:520–7. 3. Qureshi W, Ali Z, Amjad W, et al. Venous thromboembolism in cancer: an update of treatment and prevention in the era of newer anticoagulants. Front Cardiovasc Med 2016;3:24. 4. Merkow RP, Bilimoria KY, McCarter MD, et al. Postdischarge venous thromboembolism after cancer surgery. Ann Surg 2011;254:131–7. 5. Trinh VQ, Karakiewicz PI, Sammon J, et al. Venous thromboembolism after major cancer surgery: temporal trends and patterns of care. JAMA Surg 2013;2115:1–7. 6. Maynard G. Preventing hospital-associated venous thromboembolism: a guide for effective quality improvement. 2nd ed. Rockville, MD: Agency for Healthcare Research and Quality, AHRQ publication no. 16-0001-EF; October 2015. 7. Gould MK, Garcia DA, Wren SM, et al. Prevention of VTE in nonorthopedic surgical patients. Antithrombotic therapy and prevention of thrombosis, 9th ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest 2012;141(Suppl):227–77.
STERBLING ET AL CAPRINI RAM IN THORACIC CANCER PATIENTS
7
8. Zwischenberger BA, Tzeng C-WD, Ward ND, Zwischenberger JB, Martin JT. Venous thromboembolism prophylaxis for esophagectomy: a survey of practice patterns among thoracic surgeons. Ann Thorac Surg 2016;101:489–94. 9. Pannucci C, Bailey S. Validation of the Caprini risk assessment model in plastic and reconstructive surgery patients. J Am Coll Surg 2011;212:105–12. 10. Caprini JA. Risk assessment as a guide for the prevention of the many faces of venous thromboembolism. Am J Surg 2010;199(Suppl):3–10. 11. Cassidy MR, Rosenkranz P, McAneny D. Reducing postoperative venous thromboembolism complications with a standardized risk-stratified prophylaxis protocol and mobilization program. J Am Coll Surg 2014;218:1095–104. 12. Hachey KJ, Hewes PD, Porter LP, et al. Caprini venous thromboembolism risk assessment permits selection for postdischarge prophylactic anticoagulation in patients with resectable lung cancer. J Thorac Cardiovasc Surg 2016;151: 37–44. 13. Hewes PD, Hachey KJ, Zhang XW, et al. Evaluation of the Caprini model for venothromboembolism in esophagectomy patients. Ann Thorac Surg 2015;100:2072–8. 14. Hachey KJ, Sterbling H, Choi DS, et al. Prevention of postoperative venous thromboembolism in thoracic surgical patients: implementation and evaluation of a Caprini risk assessment protocol. J Am Coll Surg 2016;222:1019–27. 15. Najjar PA, Madenci AL, Zogg CK, et al. Implementation of a comprehensive post-discharge venous thromboembolism prophylaxis program for abdominal and pelvic surgery patients. J Am Coll Surg 2016;223:804–13. 16. Schmeler KM, Wilson GL, Cain K, et al. Venous thromboembolism (VTE) rates following the implementation of extended duration prophylaxis for patients undergoing surgery for gynecologic malignancies. Gynecol Oncol 2013;128: 204–8. 17. Kim JYS, Khavanin N, Rambachan A, et al. Surgical duration and risk of venous thromboembolism. JAMA Surg 2015;150: 110–7. 18. Chew HK, Wun T, Harvey D, Zhou H, White RH. Incidence of venous thromboembolism and its effect on survival among patients with common cancers. Arch Intern Med 2006;166: 458–64. 19. Schiro TA, Sakowski J, Romanelli RJ, et al. Improving adherence to best-practice guidelines for venous thromboembolism risk assessment and prevention. Am J Heal Pharm 2011;68:2184–9. 20. Mason DP, Quader MA, Blackstone EH, et al. Thromboembolism after pneumonectomy for malignancy: an independent marker of poor outcome. J Thorac Cardiovasc Surg 2006;131:711–8.