Safety of Arthrocentesis and Joint Injection in Patients Receiving Anticoagulation at Therapeutic Levels

Transcription

1 CLINICAL RESEARCH STUDY Safety of Arthrocentesis and Joint Injection in Patients Receiving Anticoagulation at Therapeutic Levels Imdad Ahmed, MBBS, a,b Elie Gertner, MD a,b a Department of Internal Medicine, Regions Hospital, St Paul, Minn; and b University of Minnesota Medical School, Minneapolis. ABSTRACT BACKGROUND: Arthrocentesis and joint injections are commonly performed for both diagnostic and therapeutic indications. Because of safety concerns, there is often reluctance to perform these procedures in patients who are receiving anticoagulation at therapeutic levels. This study was undertaken to determine the safety of arthrocentesis and joint injection performed by physicians from different disciplines in patients who are anticoagulated. METHODS: We conducted a retrospective review of 640 arthrocentesis and joint injection procedures performed in 514 anticoagulated patients between 2001 and We assessed the incidence of early and late clinically significant bleeding in or around a joint, infection, and procedure-related pain. We further compared the incidence of these complications in 456 procedures performed in patients with an international normalized ratio 2.0 or greater and 184 procedures performed in patients with an international normalized ratio less than 2.0. RESULTS: Only 1 procedure (0.2%) resulted in early, significant, clinical bleeding in the fully anticoagulated group. There was no statistically significant difference in early and late complications between patients who had procedures performed with an international normalized ratio 2.0 or greater and those whose anticoagulation was adjusted to an international normalized ratio less than 2.0. CONCLUSION: Arthrocentesis and joint injections in patients receiving chronic warfarin therapy with therapeutic international normalized ratio are safe procedures. There does not seem to be a need for reducing the level of anticoagulation before procedures in these patients Elsevier Inc. All rights reserved. The American Journal of Medicine (2012) 125, KEYWORDS: Anticoagulation; Arthrocentesis; Bleeding; Joint injection Many patients who are receiving anticoagulation therapy require arthrocentesis for diagnostic and therapeutic purposes. There is often reluctance to perform the procedure on the part of physicians and patients because of concern for bleeding. Various strategies may be used to reduce the risk of bleeding, including stopping anticoagulation a few days before the procedure, reversing anticoagulation on the day of the procedure, or avoiding the procedure. Reversing or stopping anticoagulation therapy may put the patient at risk for thromboembolic events. Although this is a common clinical problem, there are little published data on the safety Funding: HealthPartners Research Foundation Grant. Conflict of Interest: None. Authorship: Both authors had access to the data and played a role in writing this manuscript. Requests for reprints should be addressed to Elie Gertner, MD, 640 Jackson St 11107E, St Paul, MN address: elie.x.gertner@healthpartners.com of arthrocentesis in anticoagulated patients. 1,2 The aim of this study was to: determine the safety of arthrocentesis in a large group of patients receiving chronic oral warfarin therapy with an international normalized ratio 2.0 or greater; and to compare this with a large group of patients receiving chronic oral warfarin therapy who had procedures performed only when the international normalized ratio was less than 2.0. MATERIALS AND METHODS We performed a retrospective chart review of a total of 640 arthrocentesis and joint injection procedures performed in 514 consecutive patients at a single health plan and institution (HealthPartners Medical Group and Regions Hospital) between 2001 and All patients were receiving chronic warfarin therapy. Indications were those commonly seen in a hospital or clinic setting, including joint pain, joint effu /$ -see front matter 2012 Elsevier Inc. All rights reserved. doi: /j.amjmed

2 266 The American Journal of Medicine, Vol 125, No 3, March 2012 sion, or decreased range of motion. A total of 560 procedures (87%) were performed because of joint pain. The joints studied were the knees, shoulders, and hips. The majority of injections involved the knees and shoulders. There were only 13 hip injections. Procedures were performed by internal medicine or family medicine practitioners, emergency department physicians, physical medicine and rehabilitation physicians, rheumatologists, and orthopedic surgeons. There was no institutional standard of care with respect to joint procedures while receiving anticoagulation. One of the 2 following strategies for the periprocedural management of anticoagulation was generally followed by providers: CLINICAL SIGNIFICANCE 1. Warfarin therapy was continued to maintain an international normalized ratio in the therapeutic range ( 2.0) during the entire periprocedural period (group A). 2. Warfarin was discontinued 3 to 5 days before the procedure or the international normalized ratio was normalized by coagulation factors or vitamin K. The procedure was performed when the international normalized ratio was less than 2.0. Warfarin was restarted in the evening after the procedure. Patients did not receive any bridging therapy. (group B). Any prior antiplatelet therapy was routinely continued in all patients throughout the periprocedural period. International normalized ratio values were reported within 24 to 48 hours before the procedure, and other laboratory values reported were within 1 week before the procedure. Arthrocentesis and joint injection were performed by standard techniques. Shoulder joint procedures included anterior and posterior approaches to the glenohumeral joint and subacromial injections. Knee procedures could be performed from the medial or lateral aspect of the knee. Each individual practitioner performed the procedure per his/her usual regimen. Local anesthesia was used. The needle size used was 20 to 22 gauge. All procedures were done at the bedside without the use of any ultrasound device. Assessment All patients were evaluated for baseline characteristics, use of concomitant medication, comorbid medical conditions, and baseline laboratory studies. The following end points were assessed: clinically significant bleeding in or around the joint, infection of the joint, and pain in the joint requiring a physician visit. Definition of Major Procedural-Related Complication Procedure-related complications were defined as follows: Arthrocentesis and joint injection (shoulder, knee, hip) can be performed safely in patients who are receiving anticoagulation at therapeutic levels. There is no need to decrease the level of anticoagulation before such procedures. Clinically significant bleeding: bleeding during the periprocedural period that required reversal of the anticoagulation or prolonged hemostasis to stop bleeding or admission to the hospital or surgical evacuation of the joint or delay in hospital discharge. Infection of joint: any infection in the joint related to the arthrocentesis procedure. Pain requiring a physician visit: pain in the joint that was related to the procedure as documented by the physician. Chronic pain in the joint not related to the procedure was not included. All end points were assessed within 24 hours (early complications) and within 30 days of procedure (late). Charts abstraction was performed by 3 independent chart abstractors using standardized data-collection forms. Statistical Analysis By using descriptive statistics, we summarized covariates in the overall sample and by group. Between-group differences in potential confounders were assessed using Wilcoxon tests for continuous variables and Fisher exact test for binary variables. Complication rates were compared between groups by Fisher exact test. Receiver operating characteristic analysis was performed to assess the value of international normalized ratio as a predictor of complications. All tests were 2 sided, with Analyses were performed using SAS v9.1 (SAS Institute Inc, Cary, NC). RESULTS A total of 456 procedures were performed with an international normalized ratio 2.0 or greater (group A) and 184 procedures were performed with an international normalized ratio less than 2.0 (group B). Indications for the procedure were usually joint pain/joint effusion or restricted range of motion. Underlying rheumatologic comorbidities included rheumatoid arthritis, osteoarthritis, and gout. The reasons for anticoagulation were pulmonary embolism, deep vein thrombosis, mechanical heart valve, atrial fibrillation, and cerebrovascular disease. The procedures were performed by family physicians (36.2%), internal medicine physicians (35.1%), orthopedic surgeons (13.5%), rheumatologists (9.8%), physical medicine and rehabilitation physicians (3.8%), and emergency department physicians (1.6%). Ninety percent of patients had injections only, 4.4% of patients had aspiration only, and 5.6% of patients had both aspirations and injections. There were no significant differences in age, sex, body mass index, and concurrent use of antiplatelet agents between the 2 groups. Groups also were comparable among all medical comorbidities examined (diabetes mellitus, hypercoagulability, hypertension, liver failure, renal failure, and

3 Ahmed and Gertner Arthrocentesis on Anticoagulation 267 Table 1 Baseline Characteristics Group A (INR 2) (n 456) Group B (INR 2) (n 184) P Value Age Sex (female) (%) BMI (kg/m 2 ) Diabetes (%) Hypertension (%) Renal failure (%) Aspirin (%) Clopidogrel (%) Hemoglobin (mean SD) Platelet (mean SD) Creatinine (mean SD) INR international normalized ratio; BMI body mass index; SD standard deviation. smoking status). Group A (international normalized ratio 2.0) had a mean international normalized ratio of (range ), and group B (international normalized ratio 2.0) had a mean international normalized ratio of (range ). Table 1 shows the baseline characteristic of the 2 groups. There was no difference in the use of aspirin and clopidogrel between the 2 groups. There was no difference in the preprocedural platelet count, creatinine, and hemoglobin levels between the 2 groups. Table 2 shows the early and late complications in both groups. Four patients accounted for 5 early/late complications. We did not find any statistically significant differences in early and late clinically significant bleeding between the 2 groups. There was no difference in the early and late infection of the joints or pain in the joint requiring a physician visit. Table 3 shows the details of all patients with complications. One of 456 procedures in group A had a bleeding complication. The international normalized ratio at that time was 2.3. There was no bleeding into the joint postprocedure. One procedure in this group resulted in an infection. The international normalized ratio is this patient was 2.0. Three patients returned because of pain presumed to be related to the procedure; one of these was the patient with the clinically significant bleeding. There was no statistically significant difference in the overall complication rate between patients with an international normalized ratio 2.0 or greater (group A) and patients with an international normalized ratio less than 2.0 (group B) (P.708). Receiver operating characteristic (Figure 1) analysis showed that international normalized ratio offered modest value as a predictive instrument, with a c-statistic of Of interest, 103 of 456 procedures (22.5%) performed in group A were in patients with an international normalized ratio greater than 3, with the highest international normalized ratio being DISCUSSION Despite concerns about increased bleeding in the periprocedural period in patients who are receiving anticoagulation, recent studies show that pacemaker and defibrillator placement, 3 intramuscular injection of influenza vaccine, 4 and coronary angioplasty 5 are safe in fully anticoagulated patients. Joint injections and aspirations are commonly performed procedures, but there are little data about their safety in patients who are fully anticoagulated. This study suggests that these procedures in patients with uninterrupted warfarin therapy and therapeutic international normalized ratio is associated with minimal risk of periprocedural bleeding. There was no increased risk for overall complications. Indeed, 103 procedures were performed in patients with an international normalized ratio greater than 3.0, and there Table 2 Early and Late Complications Between Two Groups Complications Group A (INR 2) (n 456) Group B (INR 2) (n 184) P Value Clinically significant bleeding (early) 1 (0.2%) 0 NS Clinically significant bleeding (late) 0 0 NS Infection of joint (late) 1 (0.2%) 0 NS Pain of joint causing physician visit 3 (0.7%) 0 NS INR international normalized ratio; NS not significant.

4 268 The American Journal of Medicine, Vol 125, No 3, March 2012 Table 3 Details of Patients in Group A with Complications Complications Age Sex Antiplatelet Agents Bleeding and 70 M Aspirin 2.3 pain* Pain 77 F None 5.3 Pain 67 F None 3.3 Infection 74 F Aspirin 2.0 INR international normalized ratio. *Same patient. Pre-procedure INR were no complications. To our knowledge, this is the first study that compares 2 strategies of anticoagulation management in patients undergoing joint aspiration and injection on chronic warfarin therapy. There are 2 small previous studies of arthrocentesis in anticoagulated patients. In a study by Salvati et al, 2 arthrocentesis was performed in 15 patients receiving chronic warfarin therapy with an international normalized ratio range of 1.3 to 5.0. Two of 15 patients (13.3%) had hemarthrosis with an international normalized ratio of 3.8 and 5.0. In a prospective study by Thumboo and O Duffy, 1 there were no patient-reported bleeding events in 32 patients who underwent arthrocentesis (n 15) and soft tissue aspirations/injections (n 17) while receiving chronic warfarin therapy. Median international normalized ratio value was 2.6 (range ). Both studies concluded that arthrocentesis in patients receiving chronic warfarin therapy was associated with a low risk of hemorrhagic complications. The risk of clinically significant bleeding in our study group with therapeutic international normalized ratio was 0.2%, and there were no bleeds in patients with supratherapeutic international normalized ratio. As noted, previous studies have shown that placement of pacemakers and defibrillators, 3 intramuscular injection of influenza vaccine, 4 and coronary angiography 5 are safe in fully anticoagulated patients, but the risk of significant bleeding may be increased in the presence of underlying joint disease, 6 thrombocytopenia or use of antiplatelet agents, 7 and intensity of anticoagulation and presence of comorbidities. 8 LIMITATIONS Our data were derived from a single large center database. This observational study was nonrandomized, and data were analyzed retrospectively. Differences in techniques may have existed among the practitioners. As a result, patient selection and operator bias were difficult to avoid. Most of the procedures were injections, but 10% were aspirations or aspiration and injection. These limitations should be kept in mind in the interpretation of the findings of this study. Also, this study does not address the safety of procedures for patients who are receiving anticoagulation by other methods, such as direct thrombin inhibitors. However, this study likely does reflect the real world of current common practice across many disciplines. Sensitivity Specificity Figure 1 Receiver operating characteristic analysis.

5 Ahmed and Gertner Arthrocentesis on Anticoagulation 269 CONCLUSIONS Arthrocentesis and joint injections in patients receiving chronic warfarin therapy with a therapeutic international normalized ratio seem to be safe without an increased risk of bleeding complications. This approach simplifies the periprocedural management of anticoagulation and may lead to improved outcomes and reduced health care costs. References 1. Thumboo J, O Duffy JD. A prospective study of the safety of joint and soft tissue aspirations and injections in patients taking warfarin sodium. Arthritis Rheum. 1998;41: Salvati G, Punzi L, Pianon M, et al. Frequency of the bleeding risk in patients receiving warfarin submitted to arthrocentesis of the knee. Reumatismo. 2003;55: (abstract). 3. Ahmed I, Gertner E, Nelson WB, et al. Continuing warfarin therapy is superior to interrupting warfarin with or without bridging anticoagulation therapy in patients undergoing pacemaker and defibrillator implantation. Heart Rhythm. 2010;7: Raj G, Kuma R, McKinney P. Safety of intramuscular influenza immunization among patients receiving long-term warfarin anticoagulation therapy. Arch Intern Med. 1995;155: Ziakas AG, Koskinas KC, Gavrilidis S, et al. Radial versus femoral access for orally anticoagulated patients. Catheter Cardiovasc Interv. 2010;76: Wild JH, Zvaifler NJ. Hemarthrosis associated with sodium warfarin therapy. Arthritis Rheum. 1976;19: Basu D, Gallus A, Hirsh J, Cade J. A prospective study of the value of monitoring heparin treatment with the activated partial thromboplastin time. N Engl J Med. 1972;287: Levine MN, Raskob G, Landefeld S, Hirsh J. Hemorrhagic complications of anticoagulation treatment. Chest. 1995;108(Suppl 1):276S 290S.