Clarifying Indeterminate Results on the Rabies Direct Fluorescent Antibody Test Using Real-Time Reverse Transcriptase Polymerase Chain Reaction

Abstract

Objectives:

Each year, rabies virus infection results in the death of more than 50 000 persons worldwide. In the United States, the Centers for Disease Control and Prevention (CDC) reported 23 human rabies cases from May 1, 2008, through October 1, 2017. Although rabies testing in the United States is highly reliable, some specimens submitted to rabies laboratories do not have adequate tissues or may be substantially decomposed. In these instances, the specimen may be considered unsatisfactory for testing or produce indeterminate results using the gold standard direct fluorescent antibody test. The objective of this study was to evaluate the number of unsatisfactory samples or samples with indeterminate results that were positive for rabies virus after additional testing using real-time reverse transcriptase polymerase chain reaction (RT-PCR).

Methods:

In 2016, we retested all unsatisfactory specimens or specimens with indeterminate results using real-time RT-PCR. We further typed any sample that was real-time RT-PCR positive to identify the infecting rabies virus variant.

Results:

Of 210 retested unsatisfactory specimens or specimens with indeterminate results, 9 (4.3%) were positive for rabies. In each case, the animal was infected with a homologous rabies virus variant.

Conclusion:

These results confirm the recommendation by CDC and state public health laboratories that indeterminate results should be considered positive and justify the prompt treatment of exposed persons through an animal that is suspected to have rabies.

Indeterminate results on the rabies direct fluorescent antibody test (DFAT), the gold standard for rabies diagnosis, are a conundrum for rabies laboratories. In temperate climates, the combination of a greater number of people enjoying the outdoors, children on summer break, increased activity of animals, and presence of juvenile animals results in the busiest time of the year for rabies diagnostic laboratories. The busy season for rabies laboratories in temperate climates coincides with the warmest time of the year, and it is not uncommon for decomposed samples to be submitted for rabies testing. However, a sample may be considered unsatisfactory for DFAT for several reasons, including lack of or inappropriate type of tissue for testing, decomposition, head trauma, exposure of tissue to chemicals, and desiccation.¹

Because rabies is fatal in nearly all cases in all mammals, once clinical signs appear (typically between 3 weeks and 3 months after exposure), diagnostic tests must be conducted rapidly and with high sensitivity and specificity.² However, the sensitivity and specificity of the test can be affected by the integrity of the sample.^3,4 The DFAT relies on rabies protein structures (epitopes) to which fluorescently labeled anti-rabies antibodies bind. Decomposition, desiccation, or exposure to chemicals (eg, bleach, formalin) can cause structural changes in epitopes that result in less efficient binding of the diagnostic antibodies.⁵ Inadequate amounts or incorrect type of brain tissue can result in an unsatisfactory sample, because rabies virus is not distributed equally among tissues.^6,7 The brain stem and cerebellum of animals are highly susceptible to rabies virus infection, whereas the cortex and other areas of the brain may have little to no antigen during an early rabies infection.⁸ A complete cross section of the brain stem, along with the 3 lobes of the cerebellum or hippocampus, are required for the DFAT, because rabies may be present on only one side during centripetal and centrifugal dissemination.^6-9

The national rabies compendium¹⁰ recommends that all indeterminate rabies test results should be considered positive, and exposed persons and animals should be treated accordingly. The objectives of this study were to determine (1) whether this recommendation could be supported by evidence provided by samples that were indeterminate on the DFAT but tested positive for rabies by real-time reverse transcriptase polymerase chain reaction (RT-PCR), (2) the cycle threshold values of these samples, and (3) the animals most likely to have unsatisfactory samples for the DFAT.

Methods

The New York State Department of Health Rabies Laboratory, located in upstate New York, tests 6000 to 7000 rabies specimens annually. The Laboratory tests all specimens from New York State, with the exception of specimens from New York City. Histories are submitted with the specimens and contain information on demographic characteristics, species, age, and gender of the animal and type of exposure. The type of exposure, such as a bite, possible bite, exposure to saliva, or exposure to other bodily fluid, is also recorded. These data are maintained for rabies surveillance and epidemiological investigations. Although most histories are complete when submitted, some histories lack important data, including the type of exposure, location of the animal, and contact information. In these instances, the county health department is contacted to obtain additional information. If an animal tests positive for rabies, the county is contacted directly; otherwise, the county obtains results on a Laboratory Information Network. The rabies specimen history form can be viewed online.¹¹

All rabies DFAT positive and indeterminate samples are archived for future use. We compiled a list of all the rabies indeterminate samples received during 2016 and retested them using real-time RT-PCR. Lower cycle threshold levels demonstrate a higher amount of viral RNA in the samples, whereas higher cycle threshold values indicate less viral RNA. To evaluate the real-time RT-PCR results, we considered a cycle threshold level <40 to be positive and a cycle threshold level ≥40 to be below our limit of detection (LOD).

We used several criteria to deem samples as unsatisfactory for the DFAT during necropsy.

Decomposition

We deemed a sample to be decomposed if (1) the central nervous system (CNS) tissue was putrefied, (2) the tissue was discolored and/or liquefied, or (3) parts of the brain were not discernible or were completely missing because of loss through the foramen during decapitation (also, in bats or smaller animals, CNS tissue may be consumed by maggots or may be desiccated, which results in a dry, hard carcass with a gummy consistency [heat fixed] or no CNS tissue available).

Inappropriate Tissue

We deemed a sample to contain inappropriate tissue if we found no cerebellum or an inadequate amount of cerebellum or we did not find a complete cross section of the brain stem because of the collection and submission technique used. Tissue may also be considered unsuitable for DFAT because of intentional chemical fixation (formalin, alcohol) or unintentional chemical fixation (brain in crushed skull exposed to household chemicals).

Mutilation

We deemed a sample to be mutilated if it contained an unidentifiable, inadequate amount of required tissue; had no cerebellum and/or did not have a complete cross section of the brain stem because of method of death; or had been subject to postmortem processing techniques, including methods that resulted in trauma to the head, a crushed skull, and fragmented or destroyed CNS tissue.

Other Reasons

We deemed a specimen to be unsatisfactory if we found no specimen to test in the submitted container (eg, for a specimen identified as bat tissue, no bat tissue was present or the material present was misidentified as bat tissue [eg, leaves, moth, clump of hair]).

Samples may also be considered indeterminate during slide reading as a result of autofluorescing bacteria on the DFAT slide interfering with the microscopic examination of the tissue or if we found any deviation in the staining pattern or technique that was contrary to the criteria outlined in the national rabies compendium.¹⁰

Under our current standard laboratory procedure, which was in use during the study period (2016), we immediately deemed a specimen to be unsatisfactory when no tissue was available. If the specimen lacked the appropriate amount or type of tissue, the DFAT was performed on any available CNS tissue in the specimen, with the caveat that if the tissue was not positive, the result would be reported as indeterminate and not negative. Indeterminate results on otherwise normal-looking CNS tissue can occur when unresolvable autofluorescing bacteria or other nonspecific staining creates an artifact that may imitate or mask a weak positive result.

We studied unsatisfactory samples processed from January 1 through December 31, 2016. During that period, we processed 210 unsatisfactory samples using real-time RT-PCR. When a specimen was deemed unsatisfactory because of a lack of available tissue, the skull was swabbed with a FLOQSwab (Copan Diagnostics Inc, Murrieta, CA). We placed the swabs into 2-mL tubes containing 1 mL of Eagles Growth Media and 1.4-mm ceramic beads (Omni International, Kennesaw, GA). We added a small piece of the brain tissue (4-5 mm in diameter) from specimens deemed unsatisfactory with tissue available into a 2-mL screw cap tube along with 1 mL of Eagles Growth Media to make an approximately 10% brain suspension. We placed the tubes into a –80°C freezer for archiving purposes and processed sample batches as appropriate. We vigorously vortexed each sample before freezing and immediately before lysis.

We used an automated extraction protocol and modified duplex real-time RT-PCR assay to test rabies samples, as described in Dupuis et al¹² and originally developed by Nadin-Davis et al.¹³ Briefly, we extracted RNA by using the QIAsymphony, an automated extraction and template addition instrument, and DSP Virus/Pathogen mini kit per manufacturer’s instructions (Qiagen, Hilden, Germany). We lysed 200 μL of a 10% brain suspension or swab transport media into 430 μL of freshly prepared lysis buffer. We spiked the lysis tubes with a known quantity of an exogenous green fluorescent protein (GFP) transcript as an internal control. We heated the lysis tubes at 65°C in a water bath for 15 minutes to inactivate the rabies virus before processing. We loaded the samples onto the QIAsymphony, processed, eluted in 110 μL elution buffer, and either immediately added them to the real-time RT-PCR reaction or stored them at –80°C for future testing.

To assess the presence of rabies virus RNA and to check for possible PCR inhibition, we used a duplexed real-time RT-PCR assay. The first primer/probe set targeted a highly conserved region in the nucleoprotein gene of all rabies virus variants. Details on specificity, sensitivity, and assay optimization are described in Dupuis et al.¹⁰ The second primer/probe set targeted an artificially synthesized GFP transcript as described in Tavakoli et al.¹⁴ Appropriate amplification of the GFP transcript ensured extraction efficiency and lack of PCR inhibitors. We added all oligonucleotides (Integrated DNA Technologies, Coralville, IA) at optimized concentrations into the qScript One-Step qRT-PCR, Low-ROX kit (Quantabio, Beverly, MA). We used the ABI 7500 FAST (Thermo Fisher Scientific, Waltham, MA) platform with the following cycling conditions: 5 minutes at 50°C, 30 seconds at 95°C, and then 45 cycles of 15 seconds at 95°C followed by 60 seconds at 50°C. We manually adjusted the baselines and thresholds to analyze the run. We repeated samples that yielded unexpected results from extraction or original tissue when possible. We considered samples to be below the LOD if the rabies virus cycle threshold level was ≥40, indicating a lack of rabies RNA, or below the LOD by our assay. If the GFP cycle threshold level was also ≥40, we considered the sample inhibited and unsatisfactory.

Results

In 2016, the New York State Department of Health received 7199 samples of tissues for rabies testing, of which 250 (3.5%) were deemed unsatisfactory for DFAT. The most common reason was decomposition, followed by mutilation, inappropriate tissue, and other (Table 1). Samples deemed unsatisfactory during necropsy were most commonly the result of decomposition or mutilation, leaving no CNS tissue to test. However, some samples were also deemed unsatisfactory during microscopic examination, when large amounts of bacteria or nonresolvable, nonspecific staining was present in the tissue (supplementary data available upon request).

Table 1.

Unsatisfactory specimens received for rabies virus testing at the New York State Department of Health Rabies Laboratory, 2016^a

Reason Sample Was Considered Unsatisfactory	Unsatisfactory Samples, No. (%) (N = 250)
Decomposition	144 (57.6)
Mutilation	92 (36.8)
Inappropriate or insufficientb tissue	11 (4.4)
Other	3 (1.2)

^a Of the 7199 specimens received, 250 (3.5%) were considered unsatisfactory for the direct fluorescent antibody test, the gold standard for rabies virus testing.

^b A full cross section of the brain stem and a section of all 3 lobes of the cerebellum are required for a sample to be considered satisfactory for the direct fluorescent antibody test.

Of the 250 unsatisfactory samples, 210 were processed for real-time RT-PCR. Of these 210 samples, 9 (4.3%) were positive for rabies RNA (Table 2). Of these 9 samples, 7 were deemed unsatisfactory during necropsy because of an inappropriate tissue; 3 samples were mutilated, and 4 were decomposed. Two were deemed unsatisfactory because of unresolved nonspecific staining observed during microscopic examination. Seven of the 9 unsatisfactory samples were from bats. The cycle threshold levels of the 9 samples ranged from 17.8 to 29.1. The 2 samples described as having unresolved nonspecific staining had cycle threshold levels of 26.4 and 29.6.

Table 2.

Specimens submitted to the New York State Department of Health Rabies Laboratory that were considered unsatisfactory for the DFAT but tested positive by real-time RT-PCR in 2016

Date Received	Species	Reason Sample Was Considered Unsatisfactorya	When the Sample Was Deemed Unsatisfactory	Comments	Exposures	Ct Level of Rabies Virusb	Ct Level of Rabies Virus Variantc	Ct Level of GFPb
April 13	Big brown bat	Other	During microscopic examination	Unresolved nonspecific staining	None	29.1	26.4	29.6
May 2	Raccoon	Decomposition	During necropsy	No central nervous system	Animal	25.4	24.2	33.3
May 3	Raccoon	Mutilation	During necropsy	Crushed	Both animal and human	18.0	15.9	35.8
August 9	Big brown bat	Mutilation	During necropsy	Crushed	Human	20.7	20.1	35.6
August 16	Big brown bat	Decomposition	During necropsy	Maggots	Animal	21.1	19.9	32.3
August 16	Little brown bat	Decomposition	During necropsy	Advanced decomposition	Animal	23.3	22.2	31.3
August 22	Big brown bat	Decomposition	During necropsy	Advanced decomposition	Animal	27.0	24.7	32.6
August 31	Big brown bat	Mutilation	During necropsy	Crushed	Animal	23.1	17.7	30.8
September 9	Big brown bat	Other	During microscopic examination	Unresolved nonspecific staining	Bat in living space	29.6	25.2	29.8

Abbreviations: Ct, cycle threshold; DFAT, direct fluorescent antibody test; GFP, green fluorescent protein; RT-PCR, reverse transcriptase polymerase chain reaction.

^a A full cross section of the brain stem and a section of all 3 lobes of the cerebellum are required for a sample to be considered satisfactory for the DFAT, the gold standard for rabies virus testing.

^b Cycle threshold is the point at which the amplification curve crosses the threshold, indicating the presence or absence of an amplicon based on the fluorescence given off during the cycle. The earlier a sample crosses the threshold, the greater the amount of RNA present in the sample. A Ct level of ≥40 (ie, a Ct level of 18.0) indicates a sample that contains a considerable amount of viral RNA, a Ct of 29.0 contains less RNA, and a Ct of <40 indicates a sample from which rabies RNA cannot be detected.

^c All rabies-positive animals were infected with a homologous rabies virus variant.

Of the 250 unsatisfactory samples, 52 (20.8%) involved a known human exposure, such as a bite or contact with saliva, and 61 (24.4%) involved contact with at least 1 domestic animal. Eighty-eight (35.2%) involved an unknown exposure to a child, a sleeping individual, or a person otherwise incapable of ruling out an exposure with a bat in the room (data available from authors). Eight of the 9 real-time RT-PCR–positive specimens involved direct or indirect contact with a human or domestic animal (ie, a dog rolling on a dead raccoon, contact with raccoon saliva, an animal owner removing a bat from the animal’s mouth, and dogs catching or playing with bats). The animal from which 1 of the 9 real-time RT-PCR–positive specimens originated was a big brown bat (Eptesicus fuscus) that had bitten a person. Another specimen involved a person using bare hands to touch a container in which a bat had been, but we did not have enough information to consider this incident an exposure. Most (n = 145, 58.0%) of the 250 unsatisfactory samples came from bats.

Discussion

Several factors can render a rabies specimen unsatisfactory, including inappropriate storage conditions (exposure to direct sunlight, excessive heat), delay in submission, lack of cooling material during specimen transport, and the way the animal was killed (trauma, gunshot, glue boards, exposure to chemicals). It was not surprising that most of the unsatisfactory samples were from bats. Based on our laboratory experience, and supported by previous research, bats decompose more rapidly than larger animals and, in New York State, are frequently submitted to laboratories for testing with trauma to the head.¹⁵ In addition, because of their small size and ability to enter tight spaces, bats may not be found immediately after death. Whenever possible, persons who have experience in humane capture, euthanasia, and specimen preparation should be part of the rabies specimen collection and submission process. To aid in situations in which no experienced person is available, state rabies laboratories should provide detailed information on their websites on how to properly collect specimens, along with 24-hour, 7-day contact information for state specialists.

The gold standard for rabies virus testing, the DFAT, uses fluorescein isothiocyanate–labeled antibodies directed toward rabies antigen (more specifically, rabies nucleocapsid protein epitope).¹⁶ The breakdown of rabies antigen due to degradation, regardless of the reason, can substantially decrease the antigenic material to which the antibody binds, possibly resulting in a false-negative result.⁵ In addition, the use of unverified biologics may result in false-negative results. As with the introduction of any new reagent into a diagnostic protocol, the introduction of new antibodies for rabies diagnosis, whether developed commercially or internally, requires thorough validation and quality control to ensure that the product is appropriate for locally circulating lyssaviruses and will be appropriate for lyssaviruses that may circulate in the future.¹⁷ With a nearly 100% fatality rate, a false-negative result can be a deadly laboratory mistake.

Several publications have described the use of real-time RT-PCR as a potential replacement for an indeterminate DFAT result.^3–7 As with any newly introduced diagnostic technique, extensive testing and validation is required before implementing real-time RT-PCR in a rabies laboratory. To ensure accurate and repeatable results, the use of real-time RT-PCR as a diagnostic tool for decomposed tissue requires evaluating a large number of samples in various degrees of decomposition, from various species, infected with various rabies virus variants. The inclusion of a beta-actin control may provide a method to determine extraction efficiency but cannot control for the varying amounts of RNA in tissues.¹⁷

Unfortunately, samples that are unsatisfactory for the DFAT because of inappropriate tissue are inherently unsatisfactory for PCR testing. When a sample is substantially decomposed or mutilated, there is often no tissue, an inadequate amount of tissue, or the wrong type of tissue (eg cortex) to make appropriate DFAT microscope slides. Sometimes, because of the sensitivity of PCR, a simple swab of the cranial cavity can often be enough to yield a positive result. However, negative results on the DFAT cannot always be unequivocally confirmed via real-time RT-PCR. During necropsy of unsatisfactory samples, scientists and PCR technicians may not be confident about the tissue they are swabbing. Also, because rabies virus does not infect neurological tissue uniformly, without the appropriate amount and type of tissue, the results of real-time RT-PCR may be unreliable.

Conclusion

Rabies deaths are rare in the United States because of the positive effects of vaccination laws for domestic animals, education, effective biologics, diagnostic laboratories, and public health investigations. Rabies exposures are and should continue to be handled conservatively. Despite the high fatality rate of rabies, it is a vaccine-preventable disease when postexposure prophylaxis is provided promptly and correctly. Although specimen collection protocols have advanced, unsatisfactory samples will continue to challenge rabies laboratories because of factors beyond the laboratory’s control. When the DFAT is performed on tissue deemed to be “unsatisfactory if negative,” a positive test has diagnostic value, but a negative test does not. Our results support the mandate by state public health facilities and by the Centers for Disease Control and Prevention’s national compendium to consider these specimens as rabies positive and treat exposed persons accordingly.