Sirs: A thirty-one year old male was admitted to the Department of Medicine in Yan Chai Hospital with fever and headache. He had history of fever and sore throat for one week. His family doctor made a diagnosis of acute pharyngitis and gave him a course of amoxicillin. Few days later, he developed bilateral hearing loss and increasing early morning headache, exacerbated by cough. His temperature was 38 °C. There was no sign of meningeal irritation. Scattered blisters were noted over the left side of his neck along the distribution of cervical nerve C3 and C4. Bilateral hearing loss was present. Audiometry test revealed neurosensory hearing loss in both ears. His complete blood count, renal and liver functions, chest radiograph, electrocardiogram and CT of the brain were normal. Lumbar puncture revealed clear CSF with an opening pressure of 200 mm H2O, white cell count 240 WBC per mm3 (polymorphs 10 %, lymphocyte 90 %), protein 1.8 g/L, glucose 1.9 mmol/L. Simultaneous serum glucose was 6.9 mmol/L. CSF bacterial antigen test, cryptococcal antigen, fungal culture, PCR for TB and herpes simplex virus (HSV) DNA were negative. Brain MRI showed nonenhancing signal change in the lower pons and medulla, which was hypodense on T1-weighted images, hyperintense on T2-weighted and FLAIR images. The features were compatible with rhomboencephalitis (Fig. 1). IV acyclovir 10 mg/kg q8h was administered for 14 days. His fever came down on day 3 after the treatment began. His blisters and headache resolved gradually. Follow-up lumbar puncture showed normalization of CSF sugar with improvement in CSF protein and white cell counts. CSF viral titer revealed a 4-fold rise for IgG against varicella (pre-treatment titer 1: < 4, follow-up titer 1:16). PCR varicella-zoster DNA in his CSF before treatment was positive. He had mild residual neurosensory hearing loss and reported occasional tinnitus six months after the encephalitis. This case is a reminder that despite being a rare entity, a strikingly low CSF glucose can occur in herpes zoster-associated encephalitis. It can be potentially confusing to the attending clinician by mimicking other common infectious disease such as tuberculosis. Among 2 case reports in the past [1, 2], anti-tuberculous therapy was given to one patient before serological confirmation of varicella zoster meningoencephalitis. With the help of PCR assays, an early diagnosis can be achieved through detection of VZV DNA in CSF as soon as 24 hours after the onset of symptoms [3]. Compared with antibody indices [4], the intrathecal production of antibody is rarely detected before day 5 after onset of neurological symptoms, limiting the use of antibody index for early diagnosis of CNS infection. Currently available VZV DNA PCR assay methods have high specificity and do not cross-react with herpes simplex infection. The diagnostic sensitivity, however, ranges from 29 % to 100 % [5]. Therefore, we suggest that simultaneous analysis of CSF for both varicella-zoster virus DNA and antibody level should be studied in all suspected cases of herpes-zoster associated encephalitis. The presence of either or both in CSF in the typical clinical setting helps the clinician to establish the diagnosis of herpes-zoster associated encephalitis. This approach is particularly useful in Asian regions such as Hong Kong where the prevalence of mycobacterium tuberculous infection is high. Our case illustrated that, though rare, hypoglycorrhachia can occur in herpes zoster associated encephalitis, which has a mortality rate of 4 % to 30 % [6, 7] and residLETTER TO THE EDITORS