Urine Samples for Rabies RNA Detection in the Diagnosis of Rabies in Humans

SIR—Elsewhere, we reported the use of nucleic acid sequence-based amplification (NASBA) for the detection of rabies RNA in the diagnosis of human rabies before death [1]. We were able to detect rabies RNA in samples of saliva and/or CSF obtained from all 4 patients with rabies we studied. Although the NASBA method is relatively simple, it is often difficult to obtain saliva specimens by voluntary spitting or swabbing, because most of the patients are usually dehydrated and not cooperative. Lumbar puncture cannot always be performed safely in a severely agitated patient and may not always yield RNA [1]. An indwelling urinary catheter is inserted in all patients suspected of having rabies at our institution; therefore, urine samples can be obtained throughout the illness. We report the NASBA results with use of urine samples. From November 2000 through September 2001, we obtained specimens from 4 patients with rabies. None of the patients received postexposure rabies prophylaxis. Three patients with furious rabies had been bitten by stray dogs 24–60 days before the onset of symptoms. Durations of survival ranged from 5 to 13 days. One male patient (patient 1) had an unusual presentation with bilateral arm paralysis and loss of pinprick sensation in both arms. Quadriplegia followed, and coma developed within the next 2 days. We could not obtain a good history of rabies exposure. This patient survived for 13 days. Postmortem examination of brain samples revealed that they were positive for rabies by use of direct fluorescent antibody testing, mouse inoculation testing, and NASBA. All samples were kept at 4°C for 2 h (for 3 patients) and 24 h (for 1 patient), until studies for the presence of the rabies nucleocapsid gene were conducted. All patients except the patient with the unusual presentation (patient 1) had detectable rabies RNA in samples of saliva and/or urine. Patient 1 had no RNA detectable in 2 saliva samples (obtained on days 5 and 6 after the onset of symptoms) and in CSF and urine samples (1 sample each obtained on day 5 after onset). Saliva and urine samples obtained from patient 2 on day 2 after onset and from patient 3 on day 4 after onset were positive for rabies RNA. A CSF sample obtained at the same time from patient 3 was negative for rabies RNA. Of the 6 samples (3 saliva and 3 urine) obtained from patient 4 on days 4, 5, and 10 after onset of symptoms, only 1 saliva sample, which was obtained on day 4 after onset, was positive for rabies RNA. A CSF sample obtained from this patient on day 4 was also negative. These data confirmed our earlier findings that not every sample can be expected to test positive for rabies RNA. Therefore, serial tests should be performed on saliva, urine, and CSF samples whenever possible. Urine and saliva testing yielded the highest proportion of positive results in this small series (urine testing, 2 of 4 patients; saliva testing, 3 of 4 patients), as compared with testing of CSF samples (none of the samples obtained from 3 patients yielded positive results). Frozen sections of nuchal skin samples are not routinely obtained at our hospital for fluorescent testing techniques, which precluded us from using these samples for RNA detection [2]. Previous analyses using immunohistochemistry failed to demonstrate rabies virus antigen in kidney specimens [3]. Nevertheless, our study suggests that urine samples may be another practical source for rabies RNA detection. We thank Drs. Terapong Tantawichien and Anongrat Tiawilai, for referring patients, and Henry Wilde, for his valuable criticism of this letter.