Pseudoseptic Reactions to Hylan Viscosupplementation: Diagnosis and Treatment

Reply: We appreciate the opportunity to respond to the objections raised by Drs. Magilavy, McPherson and Polisson, and perhaps we may clarify some salient issues concerning our article.6 It is only through such discussion that one can define the appropriate clinical diagnosis and treatments for these reactions, and most appropriately serve and advise patients with osteoarthritic knee pain. Regarding postmarketing safety data, the opportunity rarely exists to examine large cohort data to identify a specific adverse event of any product or class, therefore, these data are important because they may include reports to the regulatory agency and published studies. Our review was restricted to published reports listed on Medline and the Summary of Safety and Efficacy from the premarket applications (PMAs) of the respective products approved for use in the United States. Because of the heterogeneity of patient populations and reporting, the differential for any clinical diagnosis often is not absolute, but should provide a well-defined working guideline for the clinician in their workup for a diagnosis and subsequent treatment. We think we provided such guidelines for differentiating the diagnosis of severe acute inflammatory reactions (SAIRs) or pseudosepsis,6 and we will again briefly outline the clinically significant reactions that should be differentiated, because their therapeutic courses are distinct and often can be contrary. As we stated in our article,6 all intraarticular injections can elicit an inflammatory reaction, including saline and steroids. Therefore, we never implied that an inflammatory reaction and effusion cannot occur with noncross-linked hyaluronan. However, the major distinction of a pseudoseptic reaction lies in the: (1) severity; (2) temporal relationship to the injection or course of injection; (3) duration and clinical course of the reaction; (4) cellularity and differential in the effusions; and (5) ruling-out of pseudogout or sepsis. Although we agree that the clinical presentation of pseudosepsis can be variable, to have characterized the details provided in our article as providing “no unique clinical features” is not exactly correct; we think the characteristics reviewed provide adequate guidance for physicians skilled in the diagnosis of musculoskeletal diseases to identify and appropriately treat such a reaction. We stated information provided in our article was based on published reports, and was limited to the three intraarticular hyaluronan products approved for use in the United States. Therefore, although the MedWatch reports summarized in the CDRH MAUDE database can provide important information regarding postmarketing safety of products, they generally are not considered published reports and do not appear in Medline searches. Because these reports are spontaneous from healthcare professionals, relatives, and patients, they often are fragmented and provide inadequate information to allow for adequate evaluation of reliable data. An earlier review7 of the safety profile of the products that attempted to consider this database could only characterize pseudoseptic reports as possible based on fulfilling two or three of the criteria, and also noted the significant limitations of these reports. Although the report by Roos et al12 may warrant additional evaluation, it is not related to our article as it is a report on a product that is not approved in the United States. Magilavy et al state that we misquoted the data presented in the Summary of Safety and Efficacy (SS&E) in the PMA of hylan GF-20. Although there certainly are conflicting statements in that SS&E document, under Chronic Toxicity in Primates on page 7, it states “Two animals developed low titers of antibodies to Synvisc (possibly due to contamination by gram-negative bacterial lipopolysaccharide), while five animals developed antibodies to chicken proteins.” We incorrectly used seven of 20 to arrive at our figure of 35%, as when we again reviewed the data, four of the 20 monkeys had died, so the actual figure actually may be seven of 16 or 44%. Additional support for a possible immunologic basis that might explain the difference between hylan G-F 20 and the two sodium hyaluronates, Hyalgan and Supartz, is actually substantial. As we stated,6 direct comparisons in published preclinical reports have shown that hylan G-F 20 can elicit an antibody response,2,13,14 passive cutaneous anaphylaxis,13 inflammatory infiltrates after repeat exposure,15 and granulomatous reactions,13 whereas neither of the naturally derived hyaluronates had a reaction. In clinical studies, antibodies to hylan or chicken proteins have been seen in the sera of a patient having a pseudoseptic reaction,11 and there are two case reports indicating that chronic granulomatous reactions can be associated with hylan G-F 20 injections.3,17 Furthermore, previous communications from a representative of the manufacturer of hylan G-F 20,4 and the recent labeling changes for hylan G-F 20, are consistent with the clinical manifestations outlined in our article. Allthough in our review of the literature, we found reactions to be associated only with hylan G-F 20 exposure, it is possible that similar reactions to naturally extracted sodium hyaluronate also may occur at a substantially lower frequency. However, in our search of reports published before 1995, there were none representing 8 years of cumulative worldwide postmarketing data for Hyalgan and Supartz (both first approved in 1987), during which approximately 30-40 million injections were given. In contrast, since 1995 (Synvisc was first approved in Canada in 1993 and first marketed in 1995; it was approved in the United States in 1997) there have been more than 150 million injections of sodium hyaluronate products without a single published report of pseudosepsis, and yet during the same time, there have been 14 reports of such reactions to hylan G-F 20, and there are two reports of granulomatous reactions. Since our literature search there have been six additional clinical reports regarding hylan G-F 20,1,5,8-10,16 and some of these have been comparative studies with sodium hyaluronate1,5 in which there were no reported pseudoseptic reactions to the naturally derived hyaluronan. Similarly, two additional preclinical reports that define a difference between cross-linked and noncross-linked products have been published.13,14 Collectively, the clinical and preclinical data are substantial enough to support a difference between cross-linked hylan G-F 20 and naturally derived sodium hyaluronate products. Magilavy et al acknowledged the possibility that such reactions cannot be totally ruled out with naturally derived hyaluronates, but if they do occur it would seem to be at a much reduced frequency. We remain committed to our statement that physicians should be aware of the potential for development of such a reaction. This is important to fully inform patients regarding risks and to avoid unnecessary surgery when sepsis rather than reaction might be suspected. We applaud the change in the labeling for hylan G-F 20 that does make some effort to describe these reactions. Victor M. Goldberg, MD Case Western Reserve University, Cleveland, OH Richard D. Coutts, MD University of California at San Diego, La Jolla, CA