A Sufficiency of Caution

(See the Major Article by Holzbauer et al. on pages 1201–8.)

Human rabies is very rare in the United States, but clinicians, public health officials, and patients often experience anxiety assessing risk for exposure to the rabies virus. Absent evidence of a known exposure (bite by a confirmed rabid animal), the fatal nature of the disease results in a series of unanswerable and hypothetical “what ifs” that often leads to post-exposure prophylaxis administration out of an “abundance of caution” rather than adherence to the already conservative published guidelines [1]. This zero-risk approach makes patients and clinicians more comfortable but may lead to over half of post-exposure prophylaxis regimens being administered unnecessarily [2]. This is not a safety problem, as the current vaccine and immunoglobulin products available in the United States are safe and have virtually no significant adverse reactions [3, 4]. It may be that, in some circumstances in which post-exposure prophylaxis is used, a fatal motor vehicle accident driving to the clinic for vaccine presents a higher risk than dying of rabies.

The case report by Holzbauer, et al [5], in this issue of Clinical Infectious Diseases, is not a situation of inappropriate prophylaxis; rather, it represents a textbook response to a true rabies exposure. An 84-year-old awoke to a bat biting his finger; an actual bite wound was inapparent (as it may often be [6]), but the man washed his finger with soap and water, as recommended. The silver-haired bat (Lasionycteris noctivagans), a species implicated in a number of US human rabies cases, was available for prompt testing, and was positive for rabies virus. Post-exposure prophylaxis was initiated three days after exposure and he received a complete 4 doses of vaccine on schedule and more than the recommended 20 IU/kg of human rabies immunoglobulin (actual dose reportedly 30.9 IU/kg according to potency testing of the immunoglobulin used) partially injected into the bite site with the remaining dose given intramuscularly [7]. Yet, 5 months later, he developed clinical rabies and succumbed to it.

It is notable that this case represents the first documented failure of rabies-post-exposure prophylaxis in the Western Hemisphere, and the authors are to be commended for the comprehensiveness of their investigation. Post-exposure prophylaxis was administered according to current US guidelines, using high quality biologics and no deficiencies in procedures or biologic products were identified. The prophylaxis failure was imputed to lack of response to rabies vaccine due to the presence of an unsuspected monoclonal gammopathy, with no evidence of vaccine-induced immune response in the patient, that is, unsuspected immunocompromise. The investigation was a tour de force and this conclusion is the “most parsimonious” on offer. If the patient had been recognized as immunocompromised at the time of exposure, a fifth dose of vaccine and testing for neutralizing antibody to guide further treatment would have been indicated.

This tragic case tells us that you can do all the right things and still have a bad outcome, a lesson medicine learns over and over again. The authors are in the unenviable position of making recommendations to mitigate the chance of another such event and suggest that “If an immunocompromising condition is suspected, err on the side of caution by obtaining an antibody titer.” But how can this recommendation be successfully operationalized? What would constitute “suspected” immunocompromise in a person presenting to urgent or emergency care for rabies post-exposure prophylaxis? There was nothing in this case patient's history except a number of age-related conditions, none of which suggested immunocompromise, and age itself has not been associated with poor vaccine response, as referenced in the report. How far should one go in ruling out immunocompromise in someone with no evidence of it or any indication that it should be suspected on the basis of a single case of post-exposure prophylaxis failure out of millions of courses of immunoglobulin and vaccine administered?

It has been estimated that an average of 47 000–55 000 people receive post-exposure prophylaxis each year in emergency departments in the United States with direct medical costs of $165–209 million [8,9], which may be associated with non-direct costs (lost wages, transportation, child care, etc) of up to 40%–50% of direct costs [10] and additionally diversion of emergency department effort. In many cases, post-exposure prophylaxis is avoidable or not indicated because the human-animal interaction did not meet the definition for a rabies exposure, the animal was available for quarantine or testing, or the animal tested negative [1]. Clinicians seeing patients with potential rabies exposure are less likely to assess them as needing post-exposure prophylaxis if consultation with the health department is obtained [1]; however, it is a common experience for public health officials to assess a situation as representing low or no risk for rabies exposure or to recommend delaying post-exposure prophylaxis pending testing of the animal, and to learn later that post-exposure prophylaxis was administered anyway.

The addition of serologic testing to even a portion of these visits increases both cost and healthcare utilization. Currently, testing for rabies virus neutralizing antibodies using the recommended rapid fluorescent foci inhibition test (RFFIT) is only available from 2 laboratories in the United States: the Centers for Disease Control and Prevention (CDC) and Kansas State University (KSU) Rabies Laboratory. The RFFIT technique is labor intensive and time consuming [11], and neither laboratory option is ideal; submitting samples to CDC is not easily integrated into standard specimen handling processes, and the standard testing turnaround time at KSU is reported by the laboratory to be 3–4 weeks, although special arrangements for stat testing are possible.

The “abundance of caution” approach already widely applied in decisions about initiating post-exposure prophylaxis suggests that serologic testing could become a default practice. It does not seem practical to use suspicion of possible unrecognized immunocompromise as a criterion for post-vaccination serologic testing. If there is a real gap in our rabies prevention strategies posed by this case, and there may not be, then perhaps the answer is to focus more attention on situations of undisputed exposure, such as an actual bite from an animal documented to be rabid, and to consider the consequences of failure of post-exposure prophylaxis in these instances only. Published data on response to rabies vaccine and the rarity of prophylaxis failure suggests that even this may go beyond a practical, cost-effective approach, but it would be more feasible than attempting to assess suspicion of clinically inapparent immune compromise in the urgent and emergency care settings in which rabies exposures are usually managed, and more expeditious and economical than widespread use of serologic testing to assess individual responses.

Cases of human rabies in the United States due to bat strain virus, without documented bite exposure, have occurred [12]; but these are rare events, and the absence of a reported bite does not suggest the lack of a bite; rather, it suggests an incomplete public health investigation or a case patient who is unable to report their exposure by the time the diagnosis is suspected. It should be clearly understood that non-bite exposures have not been associated with rabies risk, except for the overt situations of tissue and organ transplant and possibly airborne transmission in bat caves, and therefore every-day, non-bite exposures do not need prophylaxis. In Canada, a definitive official recommendation against post-exposure prophylaxis for “the bat in the room” was made when the risk for rabies was estimated to be 1 in 2.7 billion person-years [13, 14]. Since early in the human immunodeficiency virus (HIV)/AIDS pandemic, when making recommendations about potential HIV exposure, we have been definitive about casual contact not transmitting the virus and have not given in to considering undocumented, theoretical circumstances of risk even though they may exist. We should be definitive about what is and what is not a risk for rabies so that exposures can be credibly assessed, anxiety assuaged, and people with real exposures managed efficiently and effectively, accepting that extraordinary cases such as investigated and reported by Holzbauer et al may still occur.

Notes

Financial support. This work was not funded by grant.

References:

Author notes