laboratory and animal investigations Transbronchial Biopsy in the Presence of Profound Elevation of the International Normalized Ratio* David A. Brickey, DO; and Dennis P. Lawlor, MD

Study objective: To identify a level of coagulopathy, reported as the international normalized ratio (INR), that predicts hemorrhage following transbronchial forceps biopsy (TBBx) in an animal model. Design: Crossover blinded study using Yucatan mini-swine (Sus scrofa). Setting: Tertiary medical center with a dedicated animal research facility. Study design: A two-stage study. In stage 1, flexible fiberoptic bronchoscopy with TBBx was performed to establish the amount of bleeding in animals with normal coagulation systems. Animals then were administered escalating dosages of warfarin to obtain one of several increased INR levels. The endpoint of stage 1 was defined as the INR that resulted in a blood loss of > 100 mL in > 50% of the study animals. In stage 2, all the animals were to be anticoagulated to the INR level determined in stage 1. Topical and systemic measures would then be administered in an attempt to decrease postprocedure hemorrhage, and the results were recorded. Results: Eighteen animals were enrolled in the study. Despite INR levels > 10, no animals developed a hemorrhagic complication of the transbronchial forceps biopsy (TBBx). Eleven animals had INRs > 7. Four animal deaths were recorded, with three animal deaths attributed to nonpulmonary hemorrhage, each due to a ruptured ovarian cyst. One death was anesthesia related. Stage 2 of the study was not performed due to the extreme INR levels reached in the animals during stage 1 and to the lack of a procedure-related complication. Conclusions: Our study suggests that INR elevation does not correlate with an increased risk of bleeding following TBBx in this animal model. (CHEST 1999; 115:1667–1671) Key words: bleeding; fiberoptic bronchoscopy; hemorrhage; international normalized ratio; procedure complications; transbronchial biopsy Abbreviations: TBBx 5 transbronchial forceps biopsy; PT 5 prothrombin time; INR 5 international normalized ratio

is the most frequently encountered comB leeding plication and is the most distressing and difficult management problem for the bronchoscopist. Transbronchial lung biopsy has been used by a number of investigators to explore a variety of diffuse and *From the Pulmonary/Critical Care Medicine Department Wilford Hall Medical Center, 59th Medical Group, Lackland Air Force Base, TX. Financial support was given through the office of the United State Air Force Surgeon General and was distributed through the Clinical Investigation Facility at Wilford Hall Medical Center. Manuscript received October 30, 1998; revision accepted February 24, 1999. Correspondence to: David A. Brickey, DO, Pulmonary/Critical Care Medicine Department, Mike O’Callaghan Federal Hospital, 4700 Las Vegas Blvd, Nellis AFB, NV 89191-7007; e-mail: [email protected]

localized pulmonary disorders.1– 4 As these investigators reported their experiences with the use of transbronchial forceps biopsy (TBBx), it was recognized that the bleeding complication rate was increased over that of flexible fiberoptic bronchoscopy For editorial comment see page 1492 alone. There have been several case reports of fatal pulmonary hemorrhage following TBBx.5– 8 Despite the widespread use of TBBx in patients with underlying hematologic abnormalities or potential bleeding diathesis, specific recommendations have not been published. Zavala9 suggested the following coagulation guidelines that would prohibit the use of CHEST / 115 / 6 / JUNE, 1999

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21920/ on 01/22/2017

1667

TBBx: a prothrombin time (PT) . 16 s, a BUN . 30 mg/dL, and a platelet count , 50,000 cells/mL. These guidelines are based on a retrospective review of bleeding complications in a small number of patients. There are no published guidelines defining the level or type of coagulopathy that indicate a statistically significant risk of bleeding following TBBx. The purpose of the study was to define an international normalized ratio (INR) level that indicates an increased risk of a bleeding complication following TBBx and to document the efficacy of topical and systemic therapies to prevent this complication.

Stage 1: Following a protocol approved by the Institutional Animal Care and Use Committee, 18 Yucatan mini-swine (Sus scrofa) were enrolled in the study. The weight of the study animals ranged from 32 to 53 kg, and the animals had an estimated blood volume of 3.0 to 4.8 L. The animals entered the study after a 14-day quarantine period. All animals had baseline hematologic and coagulation parameters measured (Table 1). The bronchoscopist was blinded to all laboratory data. Animals were premedicated with acepromazine, 1.1 mg/kg, ketamine, 15 to 20 mg/kg, and atropine, 0.02 mg/kg. Intubation was performed with 7.5-mm endotracheal, high-/low-pressure cuffed tubes (Mallinckrodt, Inc; St. Louis, MO) following administration of 6.6 mg/kg thiopental sodium IV as an induction agent. Anesthesia was maintained with isoflurane, 1.5 to 2.5%. A certified nurse anesthetist under the direction of a veterinarian administered the anesthesia. Oxygen saturation, BP, and heart rate and rhythm

were monitored throughout the procedure. Mechanical ventilation (Model 2B Anesthesia Machine; Markomed; Telford, PA) was initiated and maintained at 10 mL/kg tidal volume at 8 to 10 breaths per minute. Oxygen saturation was maintained at . 92% throughout the procedure. Each of the animals underwent TBBx through a flexible fiberoptic bronchoscope (model P-30; Olympus America Inc; Long Beach, CA). TBBx using 2.2-mm alligator forceps (Microvasive forceps; Boston Scientific Corp; Natick, MA) was directed with fluoroscopy, and a specimen was taken from the right lower (caudal) lobe. A minimum of three and maximum of six samples were obtained. The TBBx tissue specimen was reviewed by a pathologist to confirm adequate sampling of pulmonary parenchymal tissue. A collection chamber (Medical Marketing Group, Inc; Decatur GA) was attached to the suction port of the bronchoscope, and all bronchial secretions were recorded. No solutions were administered through the bronchoscope during the procedure to avoid false measurement of endobronchial hemorrhage. The animals were recovered and monitored for the occurrence of hemoptysis or other complications, including pneumothorax, significant hemoglobin desaturation (defined as , 85% by pulse oximetry for a period of . 5 min), or death. Animals were monitored for a minimum of 4 h for any hemoptysis. Fluoroscopy was performed on all animals at the time of repeat tissue sampling. Any abnormalities were recorded. Serial hematocrits were measured to assess for occult pulmonary hemorrhage in the study animal. All deaths were investigated with a postmortem examination to evaluate for a procedure-related complication. After the establishment of bleeding amounts in the control group, the animals were treated with warfarin with the intent of increasing the level of INR. Warfarin was initially administered at 0.08 mg/kg as recommended by Cromeens et al.10 Warfarin dosage was adjusted to obtain one of several goal ranges for INR, which are listed in Table 2. PTs were documented using the INR. Blood analysis was performed using an analyzer (model MLA 900c; Hemoliance; Pleasantville NY), and thromboplastin-D levels were measured (Pacific Hemostasis; Huntersville NC), yielding an International Sensitivity Index of 1.97. The INR was prolonged by daily administration of warfarin. Each of the study animals was treated with warfarin until an INR of 2.1 to 3.0 was obtained. The animals then underwent bronchoscopy with TBBx. Significant bleeding was defined as $ 100 mL of blood-filled secretions. If there was no significant bleeding noted following TBBx, warfarin dosage was titrated to obtain a higher INR level. Repeat bronchoscopy with TBBx was performed at escalating INR levels until a significant level of bleeding was found in $ 50% of the study animals. One hundred milliliters was estimated to account for 2 to 3% of the blood volume of a study animal. Stage 2: After establishing a threshold INR that resulted in significant endobronchial bleeding (blood loss $ 100 mL) following TBBx in $ 50% of animals, anticoagulated animals would be randomized to preventive strategies (bronchoscopically administered topical epinephrine vs IV vitamin K administration).

Table 1—Baseline Laboratory Evaluation

Table 2—Swine Groups According to INR Levels

Tests

Groups

CBC Platelet count BUN/creatinine PT INR Partial thromboplastin time

2.1–3.0 3.1–4.0 4.1–5.0 5.1–6.0 6.1–6.9 $7

Materials and Methods Study Design The Yucatan mini-swine (Sus scrofa) was selected for the study based on its chest shape, its size, and the similarity of its pulmonary vascular system to that of humans. Previous anticoagulation studies using warfarin have been used in this animal model, and the swine’s coagulation system has been demonstrated to be similar to that of humans. The study was planned as a two-stage crossover study with animals serving as their own controls. In the first stage, the goal was to determine an INR level that predicted an increased bleeding risk following TBBx. In the second stage, each of the study animals would be administered warfarin to obtain the INR level determined in stage 1. The animals then would be randomized to receive topical or systemic treatment to enhance endobronchial hemostasis, and TBBx would be repeated. Methods

1668

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21920/ on 01/22/2017

Laboratory and Animal Investigations

Statistical Analysis

significant pulmonary hemorrhage, or other condition found on postmortem exam to classify the death as procedure related. Three animals died due to nonpulmonary hemorrhagic complications. The cause of death in all three animals was exsanguination due to hemorrhage from ruptured ovarian cysts. The INRs in these animals were elevated to 4.7, 11.2, and 13.6. None of the four animals had any significant hemorrhaging at the site of TBBx. The largest amount of hemoptysis, 65 mL., followed TBBx in an animal with an INR of 22.3. The remaining 14 animals had # 10 mL of hemoptysis and recovered without any ill effects following TBBx. There was pathologic confirmation of parenchymal lung tissue in 97% of TBBx samples. Each animal had a minimum of two TBBx samples demonstrating parenchymal tissue.

This study had a two-stage design. The first stage was to establish a reproducible threshold level of anticoagulation at which pathologic levels of bleeding ($ 100 mL suctioned, bloodfilled endobronchial secretions) develop after TBBx. The second stage would have used preventive interventions to reduce the amount of bleeding associated with TBBx. With an a level at 0.05 and a b level at 0.8, assuming that a bleeding threshold had been established at which $ 50% of anticoagulated animals would have pathologic levels of bleeding with TBBx, a sample size of 15 animals would be needed to establish that a therapy that reduced bleeding by 75% (ie, blood loss of 100 mL that is reduced to 25 mL) was statistically significant in reducing bleeding following TBBx. Three additional animals were included in the study size to allow for potential animal losses. Clinically significant endobronchial bleeding was defined as $ 100 mL of endobronchial bleeding. This is consistent with previously published guidelines.11

Results Discussion

Eighteen animals were enrolled in the study. Despite extreme elevations in INR levels (. 7) in 11 animals, no animals developed a hemorrhagic complication due to TBBx (Table 3). Stage 1 of the study was discontinued when no significant hemorrhaging was noted. Stage 2 of the study was not performed due to the failure to identify an INR level that indicated a significant risk of hemorrhage following TBBx. Three animals had an INR between 5 and 6. Four animals died during the study. The death of one animal was attributed to an anesthetic complication. The animal was found dead 4 h after the initial bronchoscopy. There was no pneumothorax,

The Yucatan mini-swine was chosen for this study for the following reasons: (1) the size of the animal allowed intubation with a 7.5-mm endotracheal tube that would accommodate a bronchoscope with a channel diameter sufficient to allow the use of 2.2-mm biopsy forceps; (2) the shape and size of the animal approximated that of a small adult patient; (3) the previous use of this animal in cardiac stent and anticoagulation studies; and (4) similarities between swine and human coagulation systems.12–17 As in previous reviews,1,6,7,9 preoperative bleeding studies did not predict the incidence of bleeding

Table 3—Indications for Hemorrhagic Complications* Animal No.

Baseline

1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 Blood loss, mL (average)

x a x x x x x x x x x x x x x x x x 0.5

INR 2–3

INR 3–4

INR 4–5

INR 5–6

INR 6–7

INR . 7

x x

x (22.3) x (35.4) x (9.8)

x x x x x

x x x

x (7.3) x (16.9) e

x x x x x

e (13.6) e (11.2) x (12.9) x (18.1) x (10.6) x (13.5)

x 8.5

0

3.8

x 6.7

10.9

*x 5 indicates fiberoptic bronchoscopy with TBBx performed in the INR range; e 5 animals with cause of death attributed to exsanguination from ovarian cyst bleeding; a 5 animal death attributed to anesthesia. Values in parentheses indicate INR level. CHEST / 115 / 6 / JUNE, 1999

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21920/ on 01/22/2017

1669

following TBBx. Our study shows that even with profound elevation of the INR there was no threshold level that was predictive of significant endobronchial bleeding. In all of the study animals whose deaths were due to exsanguination the cause of death was due to bleeding from sites unrelated to the TBBx site. Kosack and Brath18 reviewed the utility of coagulation parameter screening. They found in a review of 305 patients undergoing bronchoscopy with biopsy that 35 patients had procedure-related bleeding, only 3 of whom had abnormal results of coagulation studies. Their conclusion was that the majority of patients who have bleeding have normal coagulation parameters. When reviewing the incidence of bleeding or hemoptysis following bronchoscopy, multiple series showed that the majority of bleeding events occurred in patients with normal coagulation parameters. The incidence of bleeding varied from 1.9% to 9 to 10% in published series of patients.19 –25 There are no published studies outlining the risks of bleeding following TBBx with intensification in levels of coagulopathy. There are published studies of patients with liver disease regarding the bleeding risks of invasive procedures. In a study of 200 consecutive patients undergoing laparoscopic liver biopsy, Uewe26 found no correlation of measured variables, such as bleeding time, PT, or platelet count, with the directly observed liver bleeding time. That patient group had an average PT that was only slightly prolonged (1 to 2 s) beyond that of the control group. However, in patients with the most abnormal PTs (40 to 50 s) or prothrombin activity as low as 10% of the normal value, there was no correlation with a prolonged liver bleeding time. In patients who had prolonged bleeding during liver biopsy, there were as many cases with abnormal PTs or low platelet counts as there were with normal coagulation profiles. For other invasive procedures, such as paracentesis and thoracentesis, mild coagulation abnormalities did not correlate with an increased risk of bleeding. In a study of 608 patients, PTs or partial thromboplastin times # 2 times the midpoint of normal or a count of , 50,000 platelets did not preclude either paracentesis or thoracentesis.27 Overall, the frequency of bleeding complications requiring RBC transfusion was 0.2%. In open surgical procedures in which bleeding can be visualized and surgical hemostasis applied, the risk of bleeding complications is considered low unless the PT is . 1.5 times normal. This determination is based on studies of patients receiving oral anticoagulants who tolerate surgery and have PT levels in the therapeutic range.28,29

This study suggests that elevated INR measurements may overestimate the bleeding risk associated with TBBx. The implication would be that patients might be undergoing open lung biopsy unnecessarily when they could tolerate TBBx safely or may not be undergoing TBBx at all for fear of potential bleeding. This study shows that TBBx is well tolerated even with pronounced elevation in the INR in this animal model. Despite extreme INR elevations, this study failed to define an INR level that predicted a statistically significant risk of hemorrhage following TBBx. This study documented that animals with INRs in the range of 2 to 3 underwent TBBx without any significant transbronchial bleeding. These findings may support the safe use of TBBx when the INR is within this therapeutic range. This study provides potentially useful clinical information, however, there are also possible limitations. Malignant or inflammatory states may be associated with areas of increased vascularization and could result in increased hemorrhaging following TBBx. This study was performed in animals with normal lung parenchyma, and the findings may not be applicable to TBBx in areas of abnormal parenchyma. Although Yucatan mini-swine and humans have similar coagulation systems, future investigations should be directed at extending this study to humans before general acceptance of TBBx in patients with elevated INRs. The risk of procedurerelated bleeding from other types of isolated coagulopathy, (ie, uremia and thrombocytopenia) are of interest. The additive effect of multiple coagulation abnormalities warrants further clinical investigation. ACKNOWLEDGMENTS: The authors would like to thank David L. McGlasson, and the 59th Medical Wing Clinical Investigations Facility for their support. The efforts of the Pulmonary/Critical Care Medicine Research Department, specifically Margie Sullivan, Tim Houlihan, and Ray Knauer, were greatly appreciated. We are indebted to Drs. Ken Olivier and Stephen Derdak for their support and assistance in editing this manuscript.

References 1 Hanson RR, Zavala DC, Rhodes ML, et al. Transbronchial biopsy via the flexible fiberoptic bronchoscope results in 164 patients. Am Rev Respir Dis 1976; 114:67–72 2 Ellis JH. Transbronchial lung biopsy via the fiberoptic bronchoscope. Chest 1975; 68:524 –532 3 Ackart RS, Forem DR, Clayton RJ, et al. Fiberoptic bronchoscopy in outpatient facilities. Arch Intern Med 1982; 143:30 –31 4 Prakash UBS, Offordk P, Stubbs SE. Bronchoscopy in North America, the ACCP Survey. Chest 1991; 100:1668 –1675 5 Credle WF, Smiddy JF, Elliott RC. Complications of fiberoptic bronchoscopy. Am Rev Respir Dis 1974; 109:67–72 6 Suratt PM, Smiddy JF, Gruber B. Death and complications associated with fiberoptic bronchoscopy. Chest 1976; 69:747– 751 7 Preira W, Kovant DM, Snider GL. A prospective cooperative

1670

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21920/ on 01/22/2017

Laboratory and Animal Investigations

8

9 10

11 12

13

14

15 16

17 18

study of complications following flexible fiberoptic bronchoscopy. Chest 1978; 73:813– 816 Flick MR, Wasson K, Dunn LJ, et al. Fatal pulmonary hemorrhaging after transbronchial lung biopsy through a fiberoptic bronchoscope. Am Rev Respir Dis 1975; 111:853– 856 Zavala DC. Pulmonary hemorrhage and fiberoptic transbronchial biopsy. Chest 1976; 70:584 –588 Cromeens DM, Rodgers GP, Minor ST. Warfarin sodium for anticoagulation of atherosclerotic miniature swine. J Invest Surg 1990; 3:141–145 Burgher LW. Complications and results of transbronchoscopic lung biopsy. Nebr Med J 1979; 64:247–248 Badimon L, Badimon J, Galvez A, et al. Influence of arterial damage and wall shear rate on platelet deposition: ex vivo study in a swine model. Arteriosclerosis 1986; 6:312–320 Badimon L, Badimon J, Turitto VT, et al. Platelet thrombus formation on collagen type I: a model of deep vessel injury. Circulation 1988; 78:1431–1438 Steele PM, Chesebro JH, Badimon L, et al. Balloon angioplasty: natural history of the pathophysiological response to injury in a pig model. Circ Res 1985; 57:105–112 Lam JY, Chesebro JH, Steele PM, et al. Is vasospasm related to platelet deposition? Relationship in a porcine preparation of arterial injury in vivo Circulation 1987; 75:243–249 Fuster V, Chesebro JJ. Aortocoronary artery vein-graft disease: experimental and clinical approach for the understanding of the role of platelets and platelet inhibitors. Circulation 1985; 72(suppl V):V-65–V-69 Bowie EJW, Own CA, Zollman PE, et al. Tests of hemostasis in swine: normal values and values in pigs affected with von Willebrand’s disease. Am J Vet Res 1973; 34:1405–1411 Kosack EA, Brath LK. Do screening coagulation tests predict

19 20 21 22 23 24

25 26 27 28 29

bleeding in patients undergoing fiberoptic bronchoscopy with biopsy? Chest 1994; 106:703–705 Dreisin RB, Albert RK, Talley PA. Flexible fiberoptic bronchoscopy in the teaching hospital yielding complications. Chest 1978; 74:144 –149 Anderson HA, Fontana RS. Transbronchoscopic lung biopsy for diffuse pulmonary diseases: technique and results in 450 cases. Chest 1972; 62:125–128 Joyner LR, Scheinhon DJ. Transbronchial forceps lung biopsy through the fiberoptic bronchoscope. Chest 1975; 67: 532–535 Mitchell DM, Emerson CJ, Collins JV, et al. Transbronchial lung biopsy with the fiberoptic bronchoscope: analysis of results in 433 patients. Br J Dis Chest 1981; 75:258 –262 De Fenyl O, Capron F, Lebeau B, et al. Transbronchial biopsy without fluoroscopy: a five year experience in outpatients. Thorax 1989; 44:956 –959 Cordasco EM, Mehta AC, Ahmad M. Bronchoscopically induced bleeding: a summary of nine years of Cleveland Clinic experience and review of the literature. Chest 1991; 100:41– 47 Blasco LH, Garrido VV, Delgado MN. Safety of the transbronchial in outpatients. Chest 1991; 99:562–565 Uewe K. Bleeding after liver biopsy does not correlate with indices of peripheral coagulation. Dig Dis Sci 1981; 26:388 –393 McVay PA, Toy PTCY. Lack of increased bleeding after paracentesis and thoracentesis in patients with mild coagulation abnormalities. Transfusion 1991; 31:164 –171 Littman JK, Brodman HR. Surgery in the presence of the therapeutic effect of dicumarol. Surg Gynecol Obstet 1955; 101:709 –714 Rustad H, Myhre E. Surgery during anticoagulant treatment. The risk of increased bleeding in patients on oral anticoagulant treatments. Acta Med Scand 1963; 173:115–119

CHEST / 115 / 6 / JUNE, 1999

Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21920/ on 01/22/2017

1671