Contamination of Joints with Tissue Debris and Hair after Arthrocentesis: The Effect of Needle Insertion Angle, Spinal

Needle Gauge, and Insertion of Spinal Needles with and without a Stylet
Kevin Wahl1 , DVM, Stephen B. Adams1 , DVM, MS, Diplomate ACVS, and George E. Moore2 , DVM, PhD, Diplomate ACVIM & ACVPM
1 Veterinary Clinical Sciences, School of Veterinary Medicine, Purdue University, West Lafayette, IN and 2 Comparative Pathobiology, School of Veterinary Medicine, Purdue University, West Lafayette, IN

Corresponding Author: Stephen B Adams, DVM, MS, Diplomate ACVS, School of Veterinary Medicine, Purdue University, Lynn Hall, 625 Harrison Street, West Lafayette, IN 47907 E-mail: adamss@purdue.edu Received December 2010 Accepted December 2011 DOI:10.1111/j.1532-950X.2011.00969.x

Objective: To assess fetlock joint contamination with tissue debris and hair after arthrocentesis. Study Design: Experimental. Animals: Fetlock joint tissues (n = 10 horses). Methods: Soft tissue aps including the joint capsule were dissected from the dorsal fetlock joints of 7 anesthetized horses leaving an intact proximal base. Needles inserted through the tissue aps were ushed into tissue cell culture plates and examined for debris. Studies were repeated on excised fetlock tissue preparations after being stored for 5 days. Variables included gauge and type of needle, insertion of spinal needles with and without a stylet, angle of insertion, length of hair, and ante- and postmortem needle insertion. Tissue fragments collected from 3 horses were cultured for bacteria. Results: Compared to 20 g disposable needles inserted perpendicularly through unclipped skin, the odds ratios (ORs) for hair contamination were signicantly greater for 20 g spinal needles without a stylet, and signicantly less for 22 g spinal needles inserted with a stylet and for angled insertion of disposable needles. Tissue contamination OR was signicantly less for 20 g spinal needles inserted without a stylet, angled insertion, and clipped hair. Bacteria were isolated from 2.6% of tissue fragments. Conclusions: Angled needle insertion reduces joint contamination with tissue and hair. Spinal needles should be inserted with a stylet in place and 22 g spinal needles are preferable to 20 g spinal needles. Joints may be contaminated with bacteria after routine surgical preparation of skin.

Centesis of synovial structures for diagnostic analgesia, synovial uid analysis, and administration of therapeutic medications is common in equine practice. Septic arthritis may occur as a complication after joint injections. In a retrospective study (192 horses) admitted for treatment of septic arthritis or tenosynovitis, 22% of infections were caused by joint injection.1 Many joint infections after therapeutic injection are attributed to improper skin disinfection or contamination of needles, medications, or gloves; however, skin cannot be completely sterilized or disinfected and viable microorganisms may be introduced into joints after arthrocentesis. A previous study on the effect of needle size and type, reuse of needles, insertion speed, and removal of hair on joint contamination after injection showed signicant contamination with hair and tissue debris after centesis.2 Tissue contamination was identied in 91% of 1260 injections and

hair contamination in 30% when all needle sizes and injection techniques were combined; however, many questions were unanswered. Some equine joints such as the cofn and pastern joints are commonly injected with the needle inserted at an oblique angle to the skin that causes the needle to pass through more tissues. Spinal needles in sizes other than 20 g are used for injections in clinical practice. In the previous study, hair debris was noted to be caught between the needle and the stylet of 20 g spinal needles.2 The effects of insertion of spinal needles without the stylet on tissue and hair contamination are unknown. A limitation of the previous study was that all trials were performed on postmortem tissues that may have more friable tissue and weaker hair attachments because of autolysis. Glaser et al reported 100% of needle insertions using 18 g spinal needles with the stylet in place produced skin fragments identiable in the knee joints of people having arthroscopic surgery, and

Veterinary Surgery 41 (2012) 391398

Copyright 2012 by The American College of Veterinary Surgeons

391

Contamination of Joints with Tissue Debris after Arthrocentesis

Wahl et al.

2 of 6 (33%) of those tissue specimens submitted for polymerase chain reactions had evidence of bacterial DNA.3 The amount of tissue and hair contamination after needle insertions through viable equine joint tissues is unknown. Our purpose was to determine the effect of needle insertion angle, use of spinal needles without a stylet, size of spinal needle, and use of stored tissue preparations compared with live joint tissues on contamination of joints with tissue debris and hair after arthrocentesis. We hypothesized that: (1) angling the needle 45 to the skin surface (angled insertion) will increase joint contamination compared with insertion of the needle 90 to the skin surface (straight insertion); (2) insertion of spinal needles with the stylet in place will increase joint contamination compared with insertion of spinal needles without a stylet; (3) 20 g spinal needles will increase the risk of contamination of the joint compared with 22 g spinal needles; and (4) postmortem samples will not increase tissue debris or hair contamination compared with live joint tissues. A nal objective was to determine the prevalence of bacteria in tissue samples obtained from needle punctures through surgically prepared fetlock joint tissues.

MATERIALS AND METHODS

Viable full thickness tissue aps or 10 10 cm free full thickness joint capsule tissue specimens were created from the dorsal aspect of 16 fetlock joints (7 horses) admitted for euthanasia for reasons unrelated to fetlock joint disease. Horses were sedated by administration of xylazine (1.1 mg/kg intravenously [IV]) and butorphanol (0.02 mg/ kg IV) followed by a bolus of guaifenesin (5% solution administered IV to effect). Ketamine (2.2 mg/kg IV) and diazepam (0.03 mg/kg IV) were administered to induce anesthesia, which was maintained with isourane in oxygen. Horses were positioned in dorsal recumbency. Hair covering the dorsal aspect of 6 fetlock joints was clipped using an electric clipper with size 40 clipper blades (Turbo A5 Single Speed Clipper, Oster Professional Products, McMinnville, TN). Ten other fetlock joints were left unclipped. For viable tissue aps an Esmarch tourniquet was applied at the level of the mid metacarpus or metatarsus and the skin over the dorsal aspect of the fetlock joint was scrubbed for 1 minute with a surgical hand brush using chlorhexidine gluconate soap and rinsed with sterile saline solution to remove any surface debris and loose hair. A 10 cm horizontal incision was made at the mid pastern region through the skin and underlying tissues down to the proximal phalanx and 15 cm longitudinal incisions were extended from each side of the horizontal incision proximally on each side of the fetlock joint through all tissues to the level of the distal metatarsal/metacarpal bones to create a tissue ap. The tissue ap was dissected proximally immediately adjacent to the bones to include the dorsal fetlock joint capsule, extensor tendons, and subcutaneous tissues to create a full thickness pedicle with a proximal attachment and an intact blood supply. The skin and joint capsule were

lavaged with saline (0.9% NaCl) solution to remove any debris and the skin was left wet. Towel clamps placed at the distal corners of the tissue aps were used to keep the tissue aps taut during needle insertion. Needles were inserted through the full-thickness tissue preparation and for each individual puncture the needle was ushed with 1 mL sterile saline into 1 well of a sterile 24-well tissue culture plate (Becton Dickinson Labware, Franklin Lakes, NJ). Each plate was used for only 1 set of variables. Needles were used once. Needles were inserted at 90 to the skin surface (straight) or at 45 to the skin surface (angle) and attempts were made to separate needle puncture sites to avoid entering previous skin defects. No attempt was made to standardize the direction of the angled needle insertion in relation to the direction of the hair. Forty-eight punctures were made on each tissue preparation. All needles were inserted with a rapid thrust of the needle through the skin starting 2 cm above the skin surface. Horses were euthanatized with an overdose of pentobarbital while anesthetized immediately after collection of samples from viable tissue aps. Six full-thickness skin, subcutaneous tissue, tendon, and joint capsule tissue specimens (10 cm 10 cm) were harvested from the dorsal aspect of unused fetlock joints from 4 of the 7 horses as previously described.2 Harvested specimens were sealed in plastic bags and immediately stored in a refrigerator at 3 C for 5 days. Needle injection trials on these specimens were performed using a mounting frame.2 Tissue aps were secured on the mounting frame under tension, which we subjectively tried to make the same as the tension on the skin aps of the antemortem trials. Each tissue preparation was used for 48 needle punctures. Needle punctures were performed with the same technique as used in the antemortem studies. Wells in the tissue culture plates were examined for presence of tissue debris and hair debris as previously described with 4 magnication using an inverted microscope.2 Hair debris was classied as long, short, hair root, and clumps. Tissue fragments had to be easily seen with the microscope to be classied as present. Variables evaluated were gauge and type of needle (20 g sharp disposable needles [Kendall Monoject hypodermic needle aluminum hub; Tyco Healthcare Group LP, Manseld, MA]); 20 g and 22 g disposable spinal needles with and without a stylet [Kendall Monoject diamond point spinal needles with metal hub, Tyco Healthcare Group LP]), length of hair (unclipped and clipped), angle of insertion (90 , 45 ), and antemortem and postmortem needle insertion. Each trial had 48 individual punctures except for 3 trials in which preliminary data indicated a lack of statistical power. Forty-eight additional punctures were done for these 3 trials to increase statistical power and minimize type II error. A total of 768 individual needle punctures were evaluated from 7 horses in which 13 different combinations of variables were tested. Each combination of variables was tested on tissue preparations from 1 horse except for variables with 96 individual needle punctures that were tested on tissue preparations from 2 horses.

392

Veterinary Surgery 41 (2012) 391398

Copyright 2012 by The American College of Veterinary Surgeons

Wahl et al.

Contamination of Joints with Tissue Debris after Arthrocentesis

For the 2nd part of the study, the entire fetlock region was clipped using an electric clipper with size 40 clipper blades (Turbo A5 Single Speed Clipper, Oster Professional Products) in 3 other anesthetized horses. The skin of the fetlock region was aseptically prepared using a 2-stage surgical preparation. The 1st stage was a 3-minute scrub with chlorhexidine gluconate using a sterile scrub brush with the operator wearing nonsterile exam gloves. This was followed with a sterile saline solution rinse. The 2nd stage consisted of a 3-minute scrub with chlorhexidine gluconate with sterile gauze sponges with the operator wearing sterile gloves followed by a sterile saline solution rinse. The dorsal aspect of the fetlock region was isolated with sterile drapes in standard aseptic fashion. An aerobic culture swab was drawn over the skin in a 360 circular motion to culture the skin. A total of 5 tissue preparations from the 3 horses were cultured with aerobic culture swabs. After the swabbing for culture, 10 cm 10 cm free full-thickness joint capsule tissue specimens were harvested aseptically and transferred to a sterile draped surgical table. The tissue was mounted to the sterile wood frame using sterile equipment. All debris from the harvesting procedure was ushed off the skin and joint capsule with sterile saline solution. Sharp disposable 16 g needles (Kendall Monoject hypodermic needle aluminum hub; Tyco Healthcare Group LP) were used to collect tissue plugs into sterile 24-well tissue culture plates. Needles were inserted 90 to the skin surface using a rapid thrust of the needle and discarded after each use. Grossly visible tissue plugs removed from the tissue culture plates were cultured. A total 144 punctures were performed resulting in 115 grossly visible tissue plugs that were cultured aerobically using blood agar plates. Assessment of bacteriology was done using qualitative analysis (growth or no growth). Statistical Analysis Bivariate associations were assessed using the Pearson 2 test for categorical data. Separate mixed-effects multivariate logistic regression models were used to evaluate the associations between needle type, angle, hair clipping, and ante/postmortem status on the presence of hair or of tissue in wells postinjection. Horse was modeled as a random effect, with a covariance matrix specied as uncorrelated random effects with common variance. Interaction between hair clipping and angled needle insertion was also evaluated. Odds ratios (ORs) and 95% condence intervals (95% CI) were calculated by exponentiation of the regression coefcients. The reference technique was 20 g disposable needles inserted straight (90 ) through unclipped antemortem tissue preparations. Statistical signicance was set at P < .05.

Table 1 Wells Contaminated with Tissue and Hair Debris Grouped by Needle Size and Type, Clipped and Unclipped Hair, Ante- and Postmortem, and Angle and Straight Needle Insertion Trial needle 20 g 20 g 20 g 20 g 20 g 20 g 20 g 22 g SpW 22 g SpWo 20 g SpW 20 g SpWo 20 g SpW 20 g SpWo Total Insertion C,A,S C,P ,S C,A,V C,P ,V U,A,S U,P ,S U,A,V U,A,S U,A,S U,A,S U,A,S U,P ,S C,A,S Total wells 48 48 48 96 48 48 48 48 48 96 48 96 48 768 Hair present (%) 19 24 9 40 14 24 4 5 19 52 34 41 38 323 (39.58) (50.00) (18.75) (41.67) (29.17) (50.00) (8.33) (10.42) (39.58) (54.17) (70.83) (42.71) (79.17) (42.06) Tissue present (%) 43 47 33 65 44 47 45 42 47 95 42 93 37 680 (89.58) (97.92) (68.75) (67.71) (91.66) (97.92) (93.75) (87.50) (97.92) (98.96) (87.50) (96.88) (77.08) (88.54)

Sp, spinal; C, clipped; U, unclipped; A, antemortem; P postmortem; W, , with stylet; Wo, without stylet; S, straight insertion; V, angled insertion.

RESULTS
Of 768 wells examined in the 1st part of the study, 680 (88.5%) were contaminated with tissue and 323 (42.0%)

were contaminated with hair (Table 1). Both tissue and hair debris were observed in 301 (39.2%) wells. Multiple hairs were embedded in clumps of tissue in 59 wells. Hair follicles were observed in 96 wells and long hairs were observed in 51 wells. Four hundred eighty punctures were performed on antemortem specimens with hair debris observed in 194 (40.4%) wells and tissue debris observed in 428 (89.2%) wells. Two hundred eighty-eight punctures were performed on postmortem specimens with 129 (44.8%) wells contaminated with hair and 272 (94.4%) wells contaminated with tissue. Tissue debris, hair debris, or both was observed in some of the wells of the tissue culture plates for trials with all needle sizes and all hair lengths when examined without aid of magnication. More than 39% of wells were contaminated with hair after use of 20 g disposable needles on clipped antemortem tissues with straight needle insertion and less than 9% of wells were contaminated with hair after use of 20 g disposable needles on unclipped antemortem tissues with angled needle insertion. Insertion of 22 g spinal needles with the stylet in place resulted in the fewest number of wells contaminated with hair debris when compared with all other spinal needle insertion trials (Table 1). All trials using 20 g spinal needles inserted with and without a stylet though unclipped hair resulted in long hairs being transferred through the tissues that were then deposited on the joint capsule as the needle was withdrawn. A total of 41 long hairs from 240 20 g spinal needle insertions were deposited on the joint capsule. These long hairs were not included in the numbers of hair debris counted from the wells of the tissue culture plates and thus were not included in the statistical calculations. There were no long hairs noted on the joint capsules from the 2 trials using 22 g spinal needles. Both 20 and 22 g spinal needles had similar numbers of wells contaminated with tissue debris. 92.7% of wells from 22 g spinal needle insertions through antemortem tissues and 90.6% of wells

Veterinary Surgery 41 (2012) 391398

Copyright 2012 by The American College of Veterinary Surgeons

393

Contamination of Joints with Tissue Debris after Arthrocentesis

Wahl et al.

Table 2 Multivariate Odds Ratios, 95% Condence Intervals, and P-Values for Joint Contamination with Hair Debris and Tissue after Needle Insertion Through Fetlock Joint Tissues Hair Category Needle gauge and type 20 disposable 20 spinal with stylet 20 spinal without stylet 22 spinal with stylet 22 spinal without stylet Length of hair coat Unclipped Clipped Angle of insertion Straight Angle Ante- and postmortem Antemortem Postmortem Angle and hair interaction Angle clip Odds ratio 1.00 1.48 4.83 0.21 1.21 1.00 1.32 1.00 0.17 1.00 1.36 3.49 95% conf Reference 0.922.39 2.718.58 0.080.59 0.592.47 Reference 0.812.16 Reference 0.060.51 Reference 0.951.94 1.0411.72 P-value Odds ratio 1.00 1.62 0.40 0.35 2.32 1.00 0.39 1.00 0.38 1.00 2.45 0.46 Tissue 95% conf Reference 0.465.71 0.170.92 0.111.10 0.2819.40 Reference 0.200.75 Reference 0.180.80 Reference 1.244.82 0.063.82 P-value

.103 <.001 .003 .608

.449 .031 .070 .439

.263

.005

.002

.011

.093 .043

.010 .473

from 20 g spinal needle insertions through antemortem joint tissues had tissue debris. Removing the stylet for insertion of 20 g and 22 g spinal needles increased the amount of hair contamination in tissue culture plates from 40.8% of wells for needles with the stylet to 63.2% of wells for needles without the stylet when data from both needle gauges are combined. ORs for contamination of the tissue culture wells were determined separately for hair debris and tissue debris (Table 2). Removal of the stylet in both 20 g and 22 g spinal needles increased the OR for hair contamination, although it was only signicant (OR 4.83, P < .001) for the 20 g spinal needle. The 22 g spinal needle inserted with the stylet signicantly reduced the OR for hair contamination by 79% (P = .003) and, while not signicant, there was a trend for reduction in OR for tissue debris (P = .070). There was a signicant reduction of tissue debris when the 20 g spinal needle was inserted without a stylet (OR 0.40, P = .031). Clipping the hair slightly and insignicantly increased the OR for hair contamination but signicantly decreased the OR for tissue contamination (OR 0.39, P = .005) and angled insertion of the needle signicantly decreased the OR for contamination with both hair and tissue debris (OR 0.17, P = .002; OR 0.38, P = .011, respectively). Compared with antemortem joint tissues, needle punctures through postmortem tissues signicantly increased tissue contamination (OR 2.45, P = .010) while hair contamination was increased, but not signicantly. There was a statistically signicant interaction between angled insertion of the needle and clipping the hair on hair contamination after needle insertion (Fig 1). Angling the needle for injection through unclipped tissues resulted in a further decrease in hair contamination compared to clipped tissues or straight needle insertion. Three of 115 (2.61%) tissue samples from the needle punctures cultured on the blood agar plates were positive

for growth. None of the aerobic skin culture swabs taken before needle insertions were positive for growth.

DISCUSSION
In horses, spinal needles are routinely used to inject the navicular bursa, shoulder, hip, and sacroiliac joints when needles with increased rigidity or increased needle length are needed. Twenty gauge spinal needles inserted with the stylet in place increased the OR for tissue and hair contamination by 62% and 48%, respectively, compared with the reference needle. When the 41 long hairs that were deposited on the underside of the joint capsule after insertion of the spinal needles are considered, the hair contamination was much worse. We hypothesized that the stylet was responsible for the increased levels of both hair and tissue contamination; however, 20 g spinal needles inserted without the stylet signicantly increased the OR for hair contamination by over 4 times while signicantly reducing the OR for tissue contamination. In a previous study,2 hair was observed to be caught between the stylet and the bevel of the needle during insertion of the needle through the joint capsule. Equine hair is about 60100m in diameter.4 The small diameter of equine hair may allow it to be trapped between the stylet and the bevel of the needle when the stylet does not t tightly within the needle; however, this mechanism does not explain the markedly increased OR for hair contamination after insertion of the 20 g spinal needles with the stylet removed. This result may be explained by the shape of the bevel of the spinal needles, which is much shorter and less oblique and has a more triangular-shaped tip than the bevel on the sharp disposable needles (Fig 2). The 20 g disposable needles we used were coated with silicone to reduce insertion force and needle drag by reducing entry friction.5 Literature regarding

394

Veterinary Surgery 41 (2012) 391398

Copyright 2012 by The American College of Veterinary Surgeons

Wahl et al.

Contamination of Joints with Tissue Debris after Arthrocentesis

Figure 1 Percentage of wells contaminated with hair for insertion of the needle perpendicular to the skin (straight) and angled needle insertion (hair contamination has been averaged for needle insertion variables from Table 1).

tissue coring from silicone-coated needles is scarce and this variable may have affected our results. We believe the bevel design is the major factor that contributes to increased hair contamination with 20 g spinal needles. Transfer of hair into joints with spinal needles may become a clinical problem. We have identied hair contamination after insertion of 20 g spinal needles into joints in clinical cases of horses undergoing arthroscopy after surgical preparation of clipped hair (Fig 3). In contrast to 20 g spinal needles, 22 g spinal needles inserted with the stylet in place signicantly reduced the OR for hair contamination and, while not statistically signicant, markedly reduced the OR for tissue contamination.

We speculate that the decrease in tissue and hair contamination may be because of a decrease in surface area on the cutting point of the needle, decreased size of the lumen of the needle, or that the smaller spinal needle has a tighter tting stylet with a smaller gap between the stylet and the needle (Fig 4). This speculation is supported by previous reports in people that larger gauge needles and spinal needles with poorly tting stylets produce higher levels of tissue contamination and that the size of the tissue fragment is proportional to the size of the lumen of the needle.68 One such study on skin coring demonstrated that 85% of skin punctures using a regular 18 g needle and 87% of skin punctures using a spinal needle resulted in coring of skin

Figure 2 Needle tips: 20 g spinal needle and 20 g disposable needle. Note the more oblique angle and triangular tip of the spinal needle (black arrow) compared with the less oblique angle and less steep angulation of the cutting tip of the disposable needle. Note the small gap between the stylet and the lumen of the 20 g spinal needle.

Figure 3 This 20 g spinal needle was inserted through a clipped area of skin to identify the location for instrument insertion for arthroscopic surgery. A short hair stub was drawn into the joint by the needle.

Veterinary Surgery 41 (2012) 391398

Copyright 2012 by The American College of Veterinary Surgeons

395

Contamination of Joints with Tissue Debris after Arthrocentesis

Wahl et al.

Figure 4 Disposable spinal needles (20 g, 22 g) demonstrating the t of the stylet within the hub of the needle. Note that the 22 g spinal needle appears to have a tighter tting stylet and a smaller cutting surface area than the 20 g spinal needle.

fragments.6 The skin fragments from the spinal needle were smaller than skin fragments from the regular needles and the size of the fragment was correlated with the t of the stylet in the needle, with well-tting stylets producing small particles of skin that usually were caught between the heel of the needle bevel and the stylet.6 Twenty-two gauge spinal needles lack rigidity and are difcult to insert accurately in dense or deep tissue. When possible, we recommend using 22 g spinal needles. When the stiffness of a 20 g or larger spinal needle is desirable, making a small stab incision with a scalpel blade followed by insertion of the spinal needle may reduce joint contamination; however, this practice has not been evaluated. The fetlock joint preparations we used were from normal horses without evidence of fetlock joint disease. Diseased joints may have thickened joint capsules from chronic inammation.9 Needles inserted into diseased joints may pass through a much greater distance of tissue; however, thickness of tissue may not affect joint contamination. Insertion of needles at a 45 angle, thereby passing the needle through a longer distance of tissue, signicantly decreased both tissue and hair contamination. These results may be caused by the oblique insertion angle of the cutting point of the needle as it penetrates the skin, with the needle passing between hair follicles instead of coring the follicles with the open end of the needle. Clipping the hair signicantly decreased tissue debris and showed a trend for increased hair debris compared to unclipped hair. The least amount of hair contamination occurred with angled insertion of 20 g disposable needles through unclipped hair, and the least amount of tissue contamination occurred with angled insertion of 20 g disposable needles through clipped hair. We cannot explain the reason clipping decreases tissue contamination. In light of this data and results from a previous study,2 we support arthrocentesis using sharp disposable needles through unclipped hair because we believe hair debris is likely to illicit a stronger foreign body reaction than tissue debris and hair roots and follicles are likely to contain bacteria.2

Several antiseptic protocols and techniques have proven acceptable for aseptic preparation of skin including the 2-stage technique we used.1012 Presence of unclipped hair in horses during aseptic skin preparation does not inhibit the ability of antiseptics to effectively reduce bacterial ora.13 Complete sterilization of the skin with or without hair is not possible, and the amount of bacterial contamination of joints after joint injection in horses is unknown. Our data suggest contamination with bacteria in properly disinfected skin occurs infrequently but not rarely. The bacteria likely came from hair follicles and sebaceous glands since we did not isolate bacteria from the surface of the surgically prepared skin. Our culture results may under-represent the true level of bacterial contamination in the clinical setting. Staphylococcus spp are one of the most common isolates in equine septic arthritis.1, 14 Several studies have shown that isolation of Staphylococcus spp, specically methicillin resistant S. aureus is enhanced by specic enrichment and plating media including salt-containing trypticase soy broth and CHROMagar S. aureus medium.15, 16 A greater percentage of positive cultures may have been obtained in our study if a selective media had been used to culture the skin plugs. The percentage of positive cultures in our study demonstrates that the risk of joint infection after arthrocentesis is low; however, viable bacteria can be inoculated into joints even with proper skin disinfection. Common procedure among some equine clinicians is to routinely add antibiotics to medications injected into joints. This practice may have some validity since some joint injections will introduce microorganisms into the joint; however, we believe the use of antibiotics is no substitution for proper aseptic technique. The previous study by Adams et al showed marked contamination with tissue and hair debris with all variables; however, the study used postmortem tissue preparations mostly within 24 hours of collection.2 Postmortem autolysis could have resulted in more frequent tissue or hair contamination than would have occurred in the live horse. The postmortem tissues used in the current study were all tested 5 days after harvesting the tissues from horses. We found that use of postmortem tissues signicantly increased the OR for tissue contamination but not for hair contamination at 5 days postmortem; however, the data for antemortem and postmortem fetlock preparations follow similar trends when looking at differences between needles or insertion techniques and tissue and hair contamination was marked in both antemortem and postmortem trials. When data for insertion of 20 g sharp needles is combined, 40.42% of wells were contaminated with hair and 89.17% of wells were contaminated with tissue on antemortem trials and 44.79% of wells were contaminated with hair and 94.44% of wells were contaminated with tissue on postmortem trials. We believe that comparison of different needles and insertion techniques may be valid on cadaver tissue preparations; however, additional studies using fresh postmortem tissues and tissues frozen and then thawed are warranted to validate the use of cadaver specimens. Determining the amount of hair and tissue contamination that may occur

396

Veterinary Surgery 41 (2012) 391398

Copyright 2012 by The American College of Veterinary Surgeons

Wahl et al.

Contamination of Joints with Tissue Debris after Arthrocentesis

in clinical practice may best be done by using antemortem preparations. Study limitations include the use of needles from a single manufacturer and thus a limitation of needle design. Different congurations of needle bevels and cutting edges may give different results. We used tissues from fetlock joints in horses without any obvious evidence of fetlock joint disease. Studies on other joints of the horse and other animal species may give different results from those reported here. We did not measure the surface tension of the tissue aps or the mounted excised tissue specimens, and thus minor differences in tension between tissue aps could have inuenced our results; however, in all trials we attempted to keep the skin consistently and uniformly taut to mimic skin in situ. Angled needle insertion was done in various directions in relationship to the hair. We do not know the effect of the hair and follicle direction on angled needle insertion. We performed this study to further explore needle insertion techniques and to extend the ndings of a previous study.2 Our results differ from the results of the previous study in certain areas. For example, previously, hair contamination was seen in only 7 of 60 wells (11.7%) after 20 g disposable needle punctures straight through unclipped postmortem tissue preparations; however, in the present study, hair contamination was seen in 24 of 48 wells (50%) using 20 g disposable needle punctures straight through unclipped postmortem tissue preparations. In another example, both studies showed an increased odds for hair and tissue contamination for 20 g spinal needles inserted with a stylet when compared with 20 g sharp needles, however, the increased ORs for both hair and tissue was statistically signicant in the only the rst study. This study was not a replication of the initial study and we did not attempt absolute comparisons between studies. The methods of the studies differed in many aspects. The current study evaluated the usage of antemortem tissue preparations (live horses), used all needles only once, and evaluated angled insertion of needles. The variations or changes in ORs and their statistical signicance between this and the previous study may be explained by confounding of some parameters of these studies such as the number of times needles were used, angled needle insertion, and antemortem tissue use. When comparing results from the 2 studies for just the postmortem tissue preparations, the most obvious difference between the methods was the amount of time the tissues were stored before testing. In the previous study, tissues were stored from 2 hours to 5 days; however, most tissues were used within 24 hours. All postmortem tissues were tested in the current study at 5 days. This time frame may have allowed for autolysis and signicant loosening of the hair. Repeatability and reliability of data in these 2 studies can only be determined by additional controlled studies in which all variables except individual horse are controlled. However, we believe the combined data allow us to make some recommendations applicable to clinical practice. Recommendations from a previous study on arthrocentesis include single use of needles, use of 18 g or smaller

sharp pointed needles, limited use of 20 g spinal needles and when used, spinal needles should be inserted through clipped hair, insertion of sharp pointed regular needles though unclipped hair, and clearing tissue and hair debris from the lumen of needles by withdrawing synovial uid before injection of drugs.2 Based upon the results of the current study, we believe 22 g spinal needles are preferable to 20 g spinal needles and should be used when needle exibility is not a limitation, and oblique (angled) insertion to the skin of sharp disposable needles is preferable to straight insertion when feasible. We also recommend insertion of all spinal needles with the stylet in place. Future studies could include evaluation of smaller gauge needles, comparison of silicone-coated needles to similar noncoated needles, evaluation of 20 g spinal needle insertion through a small stab incision, evaluation of different needle bevel designs and needles from different manufacturers, determination of the actual incidence of septic joints in horses after arthrocentesis and the associated risk factors, determining if recommendations made in this report will reduce the incidence of septic arthritis after joint injections, and determination of the mechanism for hair entrapment on spinal needles. High-speed video analysis of the needles penetrating the skin could aid in identication of how tissues and hair are brought into the joint by the needle.

REFERENCES
1. Schneider RK, Bramlage LR, Moore RM, et al: A retrospective study of 192 horses affected with septic arthritis/tenosynovitis. Equine Vet J 1992;254:436442 2. Adams SB, Moore GE, Mohammed Elrashidy, et al: Effect of needle size and type, reuse of needles, insertion speed, and removal of hair on contamination of joints with tissue debris and hair after arthrocentesis. Vet Surg 2010;39:667673 3. Glaser DL, Schildhorn JC, Bartolozzi AR, et al: The inadvertent introduction of skin into the joint during intra-articular knee injections: do you really know what is on the tip of your needle? Proc Am Acad Orthop Surg 2001;68:130131 4. Tregear RT: Hair density, wind speed, and heat loss in mammals. J Appl Physiol 1965;20:679801 5. Yanagihara M, Fujii T, Wakamatu N, et al: Silicone granuloma on the entry points of acupuncture, venipuncture and surgical needles. J Cutan Pathol 2000;27:301305 6. Charlebois PA: Coring: the unseen menace. Can Anaes Soc J 1966;13:585597 7. Gobson T, Norris W: Skin fragments removed by injection needles. Lancet 1958;2:983985 8. Little C: Skin fragments in end opening needles. CMAJ 1955;72:374 9. Ralphs JR, Benjamin M: The joint capsule: structure, composition, ageing and disease. J Anat 1994;184:503509 10. Rochat MC, Mann FA, Berg JN: Evaluation of a one-step

Veterinary Surgery 41 (2012) 391398

Copyright 2012 by The American College of Veterinary Surgeons

397

Contamination of Joints with Tissue Debris after Arthrocentesis

Wahl et al.

surgical preparation technique in dogs. J Am Vet Med Assoc 1993;39:392395 11. Osuna DJ, DeYoung DJ, Walker, RL: Comparison of three skin preparation techniques in the dog part 1: experimental trial. Vet Surg 1990;19:1419 12. Zubrod CJ, Farnsworth KD, Oaks L: Evaluation of arthrocentesis site bacterial ora before and after 4 methods of preparation in horses with and without evidence of skin contamination. Vet Surg 2004;33:525530 13. Hague BA, Honnas CM, Simpson RB, et al: Evaluation of skin bacterial ora before and after aseptic preparation of clipped and nonclipped arthrocentesis sites in horses. Vet Surg 1997;26:121125

14. Lapointe JM, Laverty S, Lavoie JP: Septic arthritis in 15 Standardbred racehorses after intra-articular injection. Equine Vet J 1992;24:430434 15. Safdar N, Narans L, Gordon B, et al: Comparison of culture screening methods for detection of nasal carriage of methicillin-resistant Staphylococcus aureus: a prospective study comparing 32 methods. J Clin Microbiol 2003;41:31633166 16. Gaillot O, Wetsch M, Fortineau N, et al: Evaluation of CHROMagar Staph. aureus, a new chromogenic medium, for isolation and presumptive identication of Staphylococcus aureus from human clinical specimens. J Clin Microbiol 2000;38:15871591

398

Veterinary Surgery 41 (2012) 391398

Copyright 2012 by The American College of Veterinary Surgeons