Original Article

J Vasc Access 2014 ; 15 ( 4): 278 -285 DOI: 10.5301/jva.5000200

Quantification of fibrin in blood thrombi formed in hemodialysis central venous catheters: a pilot study on 43 CVCs Thabata Coaglio Lucas1, Francesco Tessarolo2,3, Patrizia Veniero4, Elvira D’Amato2, Iole Caola5, Giandomenico Nollo2,3, Rudolf Huebner1, Giuliano Brunori4 Department of Mechanical Engineering, Universidade Federal de Minas Gerais, Belo Horizonte - Brazil Department of Industrial Engineering & BIOtech Center for Biomedical Technologies, University of Trento, Trento - Italy 3 Healthcare Research and Innovation Program, Bruno Kessler Foundation, Trento - Italy 4 Division of Nephrology, S. Chiara Hospital, Trento - Italy 5 Department of Medicine Laboratory, Azienda Provinciale per i Servizi Sanitari di Trento, Trento - Italy 1 2

ABSTRACT Purpose: Fibrin deposition and thrombotic occlusion represent a serious cause of access dysfunction in hemodialysis central venous catheters (CVCs). The aim of this work was to define and apply a method for imaging and quantifying fibrin in thrombi formed into the side holes of CVCs. Methods: Forty-three CVCs removed from a cohort of dialyzed patients were analyzed in this pilot study. Hematoxylin and eosin and a modified Carstair’s staining were applied on permanent thrombus sections. Fluorescence microscopy and image analysis were performed to quantify the fibrin amount. Results: Highly fluorescent areas were invariably associated with fibrin by Carstair’s method. The deposition of concentric layers of fibrin and erythrocytes was easily identified by fluorescence microscopy, showing growth features of the thrombus. Fibrin amount in diabetic patients was significantly higher than that in nondiabetic patients with median (interquartile range) values of 51% (47-68%) and 44% (30-54%), respectively (p=0.032). No significant difference in fibrin content was found by grouping data according to catheter type, permanence time, insertion site and dialysis vintage. Higher variability in fibrin values was found in thrombi from CVCs removed after 1-15 days compared with 16-60 days. A trend of an increase in fibrin amount in thrombi was noted according to blood platelet count at CVC insertion. Conclusions: The analytical method presented here proved to be a rapid and effective way for quantifying fibrin content in thrombi formed on CVCs with potential application in future clinical studies. Key words: Central venous catheter, Fibrin, Histology, Renal dialysis Accepted: October 15, 2013

INTRODUCTION Tunneled and nontunneled central venous catheters (CVCs) perform a primary role to correct acute metabolic derangements caused by renal dysfunction, an important ancillary role during grafts and fistulae maturation in patients with chronic kidney disease and an additional venous access for patients during the breaking-in of peritoneal dialysis (1, 2). Proper catheter management is aimed at preserving lumen patency and minimize the risk of infection. Among noninfective complications, vascular access failure due to thrombus formation is a common reason for hospitalization in hemodialyzed patients (2-4). This issue has increased the interest of clinicians and manufacturers for ameliorating CVC performance and lowering the risk of thrombus formation to finally provide an adequate and reliable dialysis 278

for both short- and long-term catheterization. The National Kidney Foundation guidelines from the ‘Dialysis Outcomes Quality Initiative’ set the lower limit to 300 mL/min for extracorporeal blood flow to properly treat the patient without excessive duration of the procedure (3-5). Insufficient blood flow is a signal of catheter malfunction, frequently related to thrombotic obstructions (1, 2, 4, 6, 7). Intraluminal catheter thrombi occur when a blood clot partly or completely occludes the lumen or catheter holes. Different catheters are equipped with a number of side holes of various sizes and positions on the catheter wall for maximizing blood flow, but simultaneously inducing vortexes and shear stresses that facilitate thrombus formation. Fibrin deposition during the initial thrombus formation and mature phase with lumen clogging are both important steps to characterize with respect to thrombotic tissue morphology and composition.

© 2014 Wichtig Publishing - ISSN 1129-7298

Lucas et al

Fibrin is composed of fibrinogen, albumin, lipoproteins and coagulation factors. A fibrin sheet around the CVC is documented to form within 24 hours from CVC insertion (3, 4, 6, 7). The fibrin sheath attracts platelets and coagulation factors and promotes leukocyte adherence. Further layers of fibrin are then deposited on the CVC surface, creating a fibrin network and plaques able to trap large amounts of erythrocytes. This mechanism can result in the clogging of CVC side holes and lumens. We recently proposed scanning electron microscopy and two-photon laser scanning microscopy as effective techniques for imaging and characterizing the superficial composition of thrombi in CVC (8). However, subsurface layers of thicker thrombi (i.e., >100 μm in thickness) remain hidden to this analysis. The purpose of this work was to define a method for imaging and quantifying all fibrin layers in thrombi formed into the side holes of hemodialysis CVC removed from patients. Specific histochemical techniques, fluorescence imaging and image analysis will be described and applied to a set of CVC thrombi collected from an observational clinical trial. MATERIALS AND METHODS CVC collection and thrombus processing A total of 43 CVC (30 nontunneled CVC, MedCOMP/ HEMO-CATH®, Harleysville, PA, USA, and 13 tunneled CVC, Palindrome™, Covidien, Mansfield, MA, USA) were collected from 36 patients (patients’ median age (range) 63 (33-83) years; male:female 23:13; 97% Caucasians; 33% diabetic) recruited in the period from August 2011 to May 2013 at the Nephrology Division of the Trento Hospital in Italy during a prospective observational study. The study was approved by the Ethics Committee for Research Involving Human Subjects and Clinical Trials (Process 04/2010) of Trento, Italy. A data collection form with essential information about patients and device was filled by the attending nephrologist at the time of catheter removal. Prothrombin time (PT), partial thromboplastin time (PTT) and blood platelet count were obtained at the time of CVC insertion and removal. Catheters were prevalently placed in the jugular (56%) or subclavian (40%) vein and occasionally in the femoral vein (4%). Eighty percent of the catheters were used to perform a 3 hour long hemodialysis procedure thrice a week. Median (range) insertion time was 16 (2-43) days for nontunneled CVC and 140 (26-987) days for tunneled CVC. CVCs were removed both for infective and noninfective reasons (21% for suspected or confirmed catheter-related infection, 72% at elective removal, 7% for malfunctioning). Dialysis vintage was also computed as the length of time in months that the patients spend on dialysis up to the CVC insertion. Thirty-four CVCs were used to perform the first dialysis procedure on the patient; seven CVCs

Fig. 1 - Nontunneled (A) and tunneled (C) representative catheter with thrombi occluding the proximal venous side holes. Dashed lines and dotted lines indicate, respectively, the transversal scalpel cuts for separating the catheter shaft and for realizing the two sample halves. Samples obtained after transversal sectioning of the same catheters are presented in parts B (nontunneled CVC section) and D (tunneled CVC section). CVC = central venous catheter.

were placed in patients with a dialysis vintage ranging from 2 to 72 months. All CVCs were rinsed in sterile saline immediately after removal from patient to reduce excess blood loss and then fixed in 4% buffered formaldehyde. A photographic documentation of the catheter was carried out and the proximal 10 cm shaft section was separated from the catheter body with a scalpel after a minimum fixation time of 48 hours. The catheter segment for histological analysis was cut in two halves by transversal sectioning through the center of the catheter hole by a scalpel. This procedure produced two equivalent samples available for histological processing and characterization according to the protocols described below. Catheter tips from a representative tunneled and nontunneled CVC are reported in Figure 1, showing the sample processed for histology. Histochemical staining Two samples obtained from each catheter were processed routinely for obtaining permanent sections. Briefly, the samples were dehydrated by immersing in ascending hydroalcoholic solutions, 100% ethanol and finally 100% xylene. After paraffin embedding, 5 µm thick sections were cut with a rotary microtome. At this stage, CVC slices were carefully removed from each section because of poor adherence to the glass slide, leaving only the thrombus slice for the following histochemical staining. Orientation of the thrombus section was easily identifiable by the specific ‘D’ shape of the thrombus portion inside the lumen. Two consecutive sections were stained, respectively, with hematoxylin-eosin (H&E) stain and a modified Carstairs’ stain for comparison. The section was stained for 10 minutes in Mayer’s hematoxylin and 30 seconds

© 2014 Wichtig Publishing - ISSN 1129-7298

279

Fibrin quantification in CVC thrombus

in eosin. Because platelets and fibrin within a thrombus cannot be easily differentiated by H&E staining, we adapted the original Carstairs’ staining method (9) to identify platelets and fibrin in the thrombus sections. Staining times and fixation solutions were optimized for a better differentiation of thrombus components. Briefly, paraffin sections were hydrated, placed in 5% ferric alum for 5 minutes, rinsed in running tap water, stained in Mayer’s hematoxylin for 5 minutes and then rinsed again in running tap water. Slides were placed for 4 minutes in picric acid-orange G solution (20 mL saturated aqueous picric acid, 80 mL saturated picric acid in ethanol and 0.2 g orange G) and then rinsed in distilled water. They were then placed in ponceau-fuchsin solution (0.5 g acid fuchsin, 0.5 g ponceau 2R, 1 mL acetic acid and distilled water to 100 mL) for 2 minutes and then rinsed in distilled water. The slides were treated with 1% phosphomolybdic acid for 3 minutes and then rinsed in distilled water. They were finally stained with aniline blue solution (1 g aniline blue in 100 mL 1% acetic acid) for 30 minutes, decolored in 1% aqueous acetic acid for 3 minutes, rinsed in several changes of distilled water, dehydrated, cleared and then mounted with acrylic medium. In sections obtained from thrombi fixed for 48 hours or longer, Carstairs’ method produces differential staining of fibrin (bright red), platelets (gray-blue to navy), collagen (bright blue), muscle (red) and red blood cells (RBCs) (yellow). Figure 2 shows fibrin layers, platelets and erythrocytes in a representative sample from a CVC. Carstairs’ stained slices were observed in bright field transmitting light microscope (DMIL, Leica, Wetzlar, Germany) at a magnification of 40×, 100× and 200×. Corresponding H&E stained slides were observed both in transmitted white light and under green light for evidencing red fluorescence among components of the thrombus tissue. Fluorescence images from H&E stained sections were compared with images obtained from Carstairs’ stained sections to identify the fluorescent thrombus components and to check the specificity of the fluorescence signal. Fluorescence microscopy and image analysis for quantification of thrombus composition Three additional sections were obtained at different sample depth and stained with H&E stain. H&E stained slices were used for quantifying fibrin amount by fluorescence microscopy. One high-resolution image (2592× 1944 pixels, RGB 16 bit) per slice was obtained by using a 4× magnification objective (Leica C PLAN, numerical aperture=0.10). The excitation light was obtained from the green line (546 nm) of a mercury lamp and red fluorescence was collected by a charge-coupled device color camera (DMIL 420, Leica). Acquired images were imported into Image J software (NIH, Bethesda, MD, USA) and converted to 8-bit gray scale. Binary threshold was applied twice: a first 280

Fig. 2 - Transversal section of a thrombus formed into the side hole of a nontunneled central venous catheter and occluding the whole lumen section. Fibrin layers (bright red) and red blood cells (yellow) are well differentiated at both low (A) and higher magnification (B and C). Platelets appear in gray-blue to navy at high magnifications and are often correlated to fibrin fibers (C). Carstairs’ stain: A) 40×, B) 100×; C) 200×.

low gray value threshold identified the total area occupied by the thrombus section and a second medium gray value threshold identified highly fluorescent tissue within the section as described in Figure 3. The total number of white

© 2014 Wichtig Publishing - ISSN 1129-7298

Lucas et al

Fig. 3 - Images from a representative thrombus section subjected to quantitative image analysis. A) H&E stained section in white transmitting light. B) Red fluorescence image obtained from illuminating with green light the same H&E stained section. C) Binary image obtained by thresholding the fluorescence image at a low gray value. White area corresponds to the whole thrombus section. D) Binary image obtained by thresholding the fluorescence image at a high gray value. White area corresponds to the highly fluorescent part of the thrombus section, corresponding to fibrin. Original magnification: 40×. H&E = hematoxylineosin.

pixels in the two binary images accounted, respectively, for the total thrombus area and the highly fluorescent area. Means and standard deviations of the ratio between highly fluorescent and total thrombus area were eventually computed for each sample. Clinical data and statistical analysis Quantitative data of fibrin amount obtained from image analysis were statistically analyzed for correlations with blood coagulation parameters (PT, PTT and blood platelet count) by Spearman rank test. Data were also grouped according to CVC type (tunneled or nontunneled), permanence time of the catheter into the vein of the patient (0-15 days; 16-60 days; more than 60 days), diabetic patient status (yes or no), patient dialysis vintage (catheter used for starting dialysis or inserted after more than 1 month of dialysis) and catheter insertion site (jugular, subclavian or femoral vein). Descriptive statistics of groups were expressed as median and interquartile range (IQR). Distributions of fibrin amount value in each group were checked for differences by MannWhitney or Kruskal-Wallis test for two or more than two independent groups, respectively. Statistical significance was considered for p