Experience with inferior vena cava filter placement in patients undergoing open gastric bypass procedures

Article Outline

• Abstract
• Methods
  • Placement of IVC filter
  • Study design
    • Group 1: retrospective review
    • Group 2: change in clinical practice
    • Group 3: change in clinical practice
  • Statistical analysis
• Results
  • Group 1: retrospective review
  • Group 2: change in clinical practice (prospective review)
  • Group 3: change in clinical practice
  • Overall morbidity and mortality
• Discussion
• Author Contributions
• Acknowledgment
• References
• Copyright

Objective

Patients undergoing open gastric bypass (OGB) for morbid obesity are at significant risk for pulmonary embolism (PE) despite the use of subcutaneous heparin injections and sequential compression devices. Prophylactic preoperative inferior vena cava (IVC) filter placement may reduce this risk. We report our experience with simultaneous IVC filter placement and OGB in an operating room setting.

Methods

From July 1999 to April 2001, 193 patients (group 1) underwent OGB. Eight patients had prophylactic intraoperative IVC filters placed for deep vein thrombosis, PE, or pulmonary hypertension. From May 2001 to January 2003, 181 patients (group 2) underwent OGB. There were 33 IVC filters placed for body mass index (BMI) greater than 55 kg/m² in addition to the above-mentioned criteria. To confirm observations made in group 1 and 2 patients, from July 2003 to May 2005, 197 patients (group 3) underwent OGB, and patients with a BMI greater than 55 kg/m² (n = 35) were offered IVC filter placement. Group 3A (n = 17) consented to IVC filter placement, and group 3B (n = 18) did not.

Results

Fifty-eight IVC filters were placed (100% technical success rate) with an increase in operating room time of 20 ± 5 minutes. In group 1, the eight patients with IVC filters had a BMI greater than 55 kg/m². There were four PEs (3 fatal and 1 nonfatal) in the other 185 patients, all which occurred in patients with BMIs greater than 55 kg/m². In group 2, there were no PEs. The perioperative PE rate in these patients was reduced from 13% (4/31; 95% confidence interval [CI], 1.1%-25.7%) to 0% (0/33; 95% CI, 0%-8.7%). Perioperative mortality was reduced from 10% (3/31; 95% CI, 0%-20.0%) to 0% (0/33; 95% CI, 0%-8.7%). There were no pulmonary emboli or deaths related to PE in group 3A patients. Group 3B patients had a 28% PE rate (two fatal and three nonfatal) and an 11% PE-related death rate. None of the remaining patients in group 3 had a PE.

Conclusions

Intraoperative IVC filter placement for the prevention of PE in morbidly obese patients undergoing OGB is feasible. We observed a significant reduction in the perioperative PE rate when a BMI greater than 55 kg/m² was used as an indication for IVC filter placement despite the use of subcutaneous heparin injections and sequential compression devices.

There has been a recent resurgence of interest in morbid obesity surgery.¹, ², ³, ⁴ Morbid obesity is defined as a body mass index (BMI) greater than 40 kg/m² or 100 pounds over ideal body weight and is a major national health concern.⁵ A recent national consensus concluded that the number of obese individuals in the United States has increased from 12% in 1990 to almost 20% in 1998.⁶, ⁷ A separate National Institutes of Health consensus panel concluded that surgery is a viable option in patients with a BMI greater than 40 kg/m² and in patients with a BMI greater than 35 kg/m² and significant medical comorbidities or those who have failed a major nutritional weight loss program.⁶, ⁷

Initial reports on surgical weight loss procedures concluded that morbidly obese patients with concomitant sleep apnea syndrome or pulmonary hypertension have a higher incidence of perioperative pulmonary embolism (PE).⁸ These observations were never formally evaluated in prospective, randomized trials, nor were they addressed in ongoing investigations of bariatric surgical weight loss procedures. The relatively high PE rate in our own initial experience with surgical weight loss procedures prompted us to review our data for these patients.

PE is a known but preventable cause of perioperative death that may occur at any time, from during surgery to several weeks after surgery. However, to date, there has been no formal consensus or recommendations for the prevention of PE in this young, high-risk group of patients despite the growing popularity of these operations. Currently accepted practices for the prevention of PE include perioperative subcutaneous heparin injections and sequential compression devices (SCDs). Inferior vena cava (IVC) filter placement has been advocated in patients with a history of deep venous thrombosis (DVT), PE, or pulmonary hypertension. Although prophylactic IVC filter placement may be beneficial, the logistics are both numerous and challenging.

Our objectives in this study were to describe an intraoperative technique of IVC filter placement in a high-risk group of morbidly obese patients undergoing the open Roux-en-Y gastric bypass procedure and to determine whether IVC filter placement during open gastric bypass (OGB) reduces the incidence of perioperative PE despite the use of subcutaneous heparin injections and SCDs.

Back to Article Outline

Methods 

Placement of IVC filter 

All patients having IVC filters placed at the time of open Roux-en-Y gastric bypass surgery had a BMI greater than 55 kg/m² and were placed on a Skytron (Skytron, Grand Rapids, Mich) Heavy Duty 6500 fluoroscopic table, which was reversed. A cephalad extension was added to the “head” (original foot) of the table to permit imaging of the patient’s abdomen. Imaging was performed by using a portable GE (General Electric, Salt Lake City, Utah) OEC 9800 digital fluoroscope. A transfemoral percutaneous puncture was made with an 18-gauge needle, and a Magic Torque (Boston Scientific, Watertown, Mass) wire was fluoroscopically guided into the IVC. The needle was then exchanged for a 6F, 7F, or 12F sheath supplied by the Trapease (Cordis, Warren, NJ), Simon-Nitinol (Nitinol Medical Technologies, Boston, Mass), Greenfield (Boston Scientific), or Bard Recovery (Nitinol Medical Technologies) IVC filter systems. Vena cavography was performed to measure IVC diameter, confirm patency, and identify the confluence of the iliac veins. Selective left and right renal venography was then performed by using a Cobra 1 (C1) catheter (Cook, Bloomington, Ind). A Trapease (n = 35), Simon-Nitinol (n = 9), Greenfield (n = 2), or Bard Recovery (n = 12) filter was then deployed into the infrarenal IVC under fluoroscopic control. Completion venography after IVC filter deployment confirmed the filter position and the patency of the IVC and renal veins.

Study design 

We reviewed our bariatric surgery program’s early results from its inception in July 1999 through April 2001. A retrospective analysis was performed on 193 patients (group 1; average BMI, 51 ± 7 kg/m²) undergoing the open Roux-en-Y gastric bypass procedure. During this initial 22-month period, indications for IVC filter placement included a history of DVT, PE, or pulmonary hypertension (mean pulmonary artery pressure >40 mm Hg) as measured by noninvasive echocardiography or Swan-Ganz catheterization. These indications for IVC filter placement were based on prior work by Sugerman et al.⁸ On the basis of these indications, eight patients (4.1%) had IVC filters placed at the time of operation.

From May 2001 to January 2003, a prospective analysis was performed on 181 patients (group 2; average, BMI 51 ± 6 kg/m²) undergoing the open Roux-en-Y gastric bypass procedure. In this next 21-month period, according to our experience with group 1, a BMI greater than 55 kg/m² or a history of DVT, PE, or pulmonary hypertension was used as the indication for IVC filter placement. A BMI greater than 55 kg/m² was added as an additional indication for IVC filter placement because despite our implementation of prior work by Sugerman et al,⁸ four patients still had PEs. Thirty-three patients (18%) had IVC filters placed at the time of operation in this group.

From July 2003 to May 2005, 197 patients (group 3; average BMI, 51 ± 7 kg/m²) underwent OGB, and patients with a BMI greater than 55 kg/m² (n = 35) were given the option of IVC filter placement on the basis of our prior clinical experience. Group 3A patients (n = 17; average BMI, 63 ± 5 kg/m²) underwent IVC filter placement with OGB, and group 3B patients (n = 18; average BMI, 63 ± 5 kg/m²) underwent OGB only, without IVC filter placement.

All patients in both groups had routine preoperative and postoperative lower extremity venous duplex examination. One bariatric surgeon and three vascular surgeons performed all of the OGB procedures and IVC filter placements, respectively, in groups 1 and 2. Two bariatric surgeons and one vascular surgeon performed all of the OGB procedures and IVC filter placements, respectively, in group 3 patients.

Additionally, all of the patients received SCDs, thromboembolic devices (TEDs), and weight-adjusted subcutaneous heparin (50 U/kg) injections before surgery and every 12 hours after surgery until they were ambulating more than 4 h/d. Routine perioperative pulmonary angiography, spiral computed tomographic scanning, and ventilation/perfusion scanning were not performed unless the patients had clinical sequelae suggestive of a PE. PEs were documented by spiral computed tomography, ventilation/perfusion scan, or autopsy within the perioperative period (30 days after surgery).

Statistical analysis 

For group 1 and 2 patients, data analysis was performed with a 95% confidence interval (CI) by using SPSS Software (SPSS Inc, Chicago, Ill). This data analysis was used to demonstrate that a BMI greater than 55 kg/m² was an additional risk factor for having a PE during OGB in addition to those previously described by Sugerman et al.⁸ It was also used to demonstrate the dramatic decrease in PE when IVC filters are placed despite the use of subcutaneous heparin injections and SCDs at the time of OGB in patients with a BMI greater than 55 kg/m².

The Fisher exact test was used to compare groups 1, 2, 3A, and 3B because all groups of patients had similar risk factors, hypercoagulable profiles, and comorbidities. These differences were calculated only by using patients with a BMI greater than 55 kg/m² and did not include the remaining cohort of patients with lower BMIs. The only difference in these two groups of patients was whether an IVC filter was placed at the time of OGB. Statistical significance was defined as P < .05.

Back to Article Outline

Results 

All patient and group characteristics were similar (Table I). Fifty-eight IVC filters were successfully placed in the infrarenal IVC in 58 patients (100% technical success rate). The average additional operating room time for IVC filter placement was 20 ± 5 minutes. The average fluoroscopic time was 3 ± 1 minutes, and the average amount of contrast used was 30 ± 10 mL. There was minimal blood loss, and there were no intraoperative complications. The average length of follow-up was 2.5 years (range, 1-42 months).

Table I. Patient and group characteristics^⁎

Variable	Group 1	Group 2	Group 3	Group 3A	Group 3B
n	193	181	197	17	18
Female	133	127	142	14	15
Male	60	54	55	3	3
BMI, kg/m2 (mean ± SD)	51±7	51±6	51±7	63±5	63±5
PE/DVT	5	3	2	1	1
PULM HTN	3	2	2	1	1
Sleep apnea	33	28	21	11	10
Venous stasis	5	7	5	3	2
DM	53	47	50	17	18
HTN	54	48	47	17	18

BMI, Body mass index; PE/DVT, history of pulmonary embolism or deep venous thrombosis; PULM HTN, pulmonary hypertension; DM, diabetes mellitus; HTN, hypertension.

Group 1: retrospective review 

In the initial 22-month period, 8 of the group 1 patients had a preoperative IVC filter placed for a history of DVT (n = 3), PE (n = 2), or documented pulmonary hypertension (n = 3). Again, these indications for IVC filter placement were based on prior work by Sugerman et al.⁸ None of these eight patients receiving IVC filters had a perioperative PE, and all had a BMI greater than 55 kg/m². Four patients in group 1 did not meet criteria for IVC filter placement and developed a fatal (n = 3) or nonfatal (n = 1) PE. All four patients had a BMI greater than 55 kg/m² and had risk factors, hypercoagulable profiles, and other comorbidities similar to others in the group. All four patients received subcutaneous heparin injections and SCDs. No other confounding variables other than a BMI greater than 55 kg/m² were identified to explain the PEs. There were 31 patients in group 1 with a BMI greater than 55 kg/m². These patients had a 10-fold risk for PE after OGB (relative risk, 10.2; 95% CI, 5.8-18) as compared with patients with a BMI less than 55 kg/m² undergoing OGB.

Group 2: change in clinical practice (prospective review) 

In the subsequent 21-month period, 33 of the group 2 patients had a BMI greater than 55 kg/m². All 33 patients received an IVC filter at the time of OGB on the basis of our observations in group 1 patients. Five patients with a BMI greater than 55 kg/m² also had a history of DVT (n = 2), PE (n = 1), or pulmonary hypertension (n = 2). There were no patients with a BMI less than 55 kg/m² and a history of DVT, PE, or pulmonary hypertension. None of the patients in group 2 undergoing OGB had a perioperative PE.

The perioperative PE rate in patients undergoing Roux-en-Y gastric bypass surgery was reduced from 13% (4/31; 95% CI, 1.1%-24.7%) to 0% (0/33; 95% CI, 0%-8.7%) when a BMI greater than 55 kg/m² was used as an indication for IVC filter placement (Table II). In addition, PE-related death was reduced from 10% (3/31; 95% CI, 0%-20.0%) to 0% (0/33; 95% CI, 0%-8.7%; Table II).

Table II. PE-related morbidity and mortality rates of patients with BMI greater than 55 kg/m²

PE, Pulmonary embolism; BMI, body mass index; IVC, inferior vena cava.

One postoperative IVC thrombosis occurred 4 months after Trapease IVC filter placement. Two postoperative localized, insertion-site DVTs occurred 3 months after filter placement and were treated with anticoagulation.

Group 3: change in clinical practice 

In a separate 22-month period, the 17 patients with a BMI greater than 55 kg/m² (group 3A) who underwent OGB and IVC filter placement had no perioperative PEs. This was in striking contrast to the 18 patients with a BMI greater than 55 kg/m² (group 3B) undergoing OGB only without IVC filter placement, who had a 28% perioperative PE rate and an 11% PE mortality rate (P < .05; Table III). Two patients in group 3 died from causes related to the gastric bypass procedure; one died from an anastomotic leak, and the other died from an internal hernia.

Table III. PE-related morbidity and mortality rates of patients with BMI greater than 55 kg/m²

PE, Pulmonary embolism; BMI, body mass index; IVC, inferior vena cava.

Overall morbidity and mortality 

The overwhelming benefit of IVC filter placement in patients with a BMI greater than 55 kg/m² undergoing OGB is minimized by the overall benefit IVC filter placement has on the entire cohort. Overall morbidity and mortality for group 2 patients were reduced by 2.3% and 1.7%, respectively, when compared with group 1 patients (Table IV).

Table IV. PE-related and overall morbidity and mortality rates

PE, Pulmonary embolism.

Back to Article Outline

Discussion 

Morbid obesity is a multisystemic disease. It affects every organ system, including the cardiovascular, pulmonary, endocrinologic, renal, integumentary, neurologic, and musculoskeletal systems. Surgical weight loss procedures can reverse many, if not all, of these comorbidities.⁶ With the growing prevalence of morbid obesity in the United States, there has been a resurgence of interest in its surgical correction.¹, ², ³, ⁴ This surgery is typically performed in relatively young, but high-risk, patients, who often expect immediate weight loss without any surgical complications. Unfortunately, because of the high-risk nature of the patient population, a significant number of associated complications can occur that result in significant morbidity and disappointment for the patients and their families.

PE is one of the most devastating complications, with an incidence of 2% to 4% in morbidly obese patients undergoing OGB.⁹ Despite the use of subcutaneous heparin injections, SCDs, and prior indications for IVC filter placement by Sugerman et al,⁸ a significant number of PEs and deaths from PEs was observed in our practice until we began placing IVC filters in patients undergoing OGB with a BMI greater than 55 kg/m². The PE rate was dramatically reduced in the patients with a high body mass index—again, those with a BMI greater than 55 kg/m².

This benefit was not as profound when the entire cohort of morbidly obese patients was analyzed. In fact, the overall morbidity and mortality of all morbidly obese patients undergoing OGB surgery was minimally reduced. There is no doubt, however, that a subgroup of morbidly obese patients with high BMIs, ie, those with a BMI greater than 55 kg/m², improve their morbidity and mortality remarkably when an IVC filter is placed.

Despite the benefits, the logistics of IVC filter placement in this group of patients are both numerous and challenging. First, the patients’ obesity prevents their positioning on standard angiographic tables. Standard tables in interventional radiology or cardiac catheterization suites hold patients up to 350 to 400 pounds, whereas the Skytron Heavy Duty 6500 operating room table holds patients up to 800 pounds. In addition, the percutaneous femoral puncture is cumbersome because of the massive abdominal pannus. We circumvented this problem by using tincture of benzoin and silk tape to retract the abdominal pannus for improved access to the groin. Another possible access site for IVC filter placement is the internal jugular vein, which can be localized by using ultrasonography. This approach was not used in these patients because most had previous Swan-Ganz catheter and/or central venous pressure lines placed. Because of the patients’ body habitus, the fluoroscopic images obtained are not as clear as those performed in relatively thin patients, but the images provide adequate visualization of the IVC, iliac, and renal veins for filter placement. Selective catheterization of the left and right renal veins ensured precise IVC deployment and helped improve our visualization. By performing simultaneous IVC filter placement and OGB, one eliminates the need for an additional admission or day of hospitalization.

Over this 65-month study period, several types of IVC filters were used. Most patients in this series received filters with low-profile introducer sheaths. A lower-profile introducer sheath reduces the time needed for manual compression and decreases the risk of hematoma formation. It may also reduce the incidence of insertion-site DVT. The long-term patency and PE rate may help clarify the role of retrievable IVC filters in this population. More recently, we have used the Tulip (Cook), Optease (Cordis), and Bard Recovery (Nitinol Medical Technologies) retrievable filters in these procedures, with success.

IVC filter placement is not a completely benign procedure. Early in our experience, there were two patients with postoperative DVT on the side of venous puncture for IVC filter insertion, and there was one IVC thrombosis 4 months after surgery. Both patients with postoperative DVT presented 3 months after IVC filter placement with lower extremity pain and edema and were diagnosed by using duplex sonography. These patients were both treated with intravenous heparin and subsequent warfarin therapy. Neither patient developed a PE. The patient with IVC thrombosis presented 1 week after the onset of symptoms and was found to have a compartment syndrome in both lower extremities necessitating bilateral fasciotomies. This patient subsequently died after 2 months of hospitalization from complications related to the gastric bypass procedure.

This study demonstrates the efficacy and feasibility of intraoperative IVC filter placement for the prevention of PE in morbidly obese patients undergoing OGB surgery. The procedure may be performed safely with a high technical success rate and minimal additional operating room time. We observed a significant reduction in the perioperative PE rate when a BMI greater than 55 kg/m² was used as a criterion for IVC filter placement. We recommend the use of BMI greater than 55 kg/m² as an indication for prophylactic IVC filter placement in addition to the use of subcutaneous heparin injections, SCDs, and the prior criteria described by Sugerman et al⁸ in patients undergoing OGB surgery. We encourage further prospective, randomized studies if there is still question about the benefit and if other institutions are willing to participate.

Back to Article Outline

Author Contributions 

Back to Article Outline

The authors thank Neal Cayne, Carlos Timaran, Reese Wain, Larry Scher, Taylor Reed, Harrie Kurvers, Karen Gibbs, Pratt Vemulapalli, Julio Teixeira, Soo Rhee, and Aimee Goodwin.

Back to Article Outline

References 

1. MacGregor MI, Bock AJ, Ball WC. Topics in clinical medicine: serious complications and sudden death in the Pickwickian syndrome. Johns Hopkins Med J. 1970;189:279–295
   • View In Article
2. MacLean LD, Rhode BM, Sampalis J. Results of the surgical treatment of obesity. Am J Surg. 1993;165:155–161
   • View In Article
   • MEDLINE
   • CrossRef
3. Mason EE, Ito CC. Gastric bypass. Ann Surg. 1969;170:329–339
   • View In Article
   • MEDLINE
   • CrossRef
4. Mokdad AH, Serdula MK, Deitz WH. The spread of the obesity epidemic in the United States 1991-1998. JAMA. 1999;282:1519–1522
   • View In Article
   • MEDLINE
   • CrossRef
5. National Institutes of Health Consensus Development Conference Panel. Gastrointestinal surgery for severe obesity. Ann Intern Med. 1991;115:956–961
   • View In Article
   • MEDLINE
6. Payne JH, DeWind LT. Surgical treatment of obesity. Am J Surg. 1969;118:141–147
   • View In Article
   • MEDLINE
   • CrossRef
7. Popkin BM, Udry JR. Adolescent obesity increases significantly in second and third generation US immigrants: the National Longitudinal Study of Adolescent Health. J Nutr. 1998;128:701–708
   • View In Article
   • MEDLINE
8. Sugerman HJ, Baron PL, Fairman RP, Evans CR, Vetrovec GW. Hemodynamic dysfunction in obesity hypoventilation syndrome and the effects of treatment with surgically induced weight loss. Ann Surg. 1998;207:604–613
   • View In Article
   • MEDLINE
   • CrossRef
9. Sugerman HJ, Kellum JM, Engle KM, et al. Gastric bypass for treating severe obesity. Am J Clin Nutr. 1992;55(Suppl):560–566
   • View In Article