Objectives: To study the occurrence of relapse of herpes simplex encephalitis (HSE) and to find out whether soluble activity markers in cerebrospinal fluid (CSF) indicate direct viral or immune- mediated events. Methods: A consecutive series of 32 adult survivors of HSE were followed to determine the incidence of clinical relapse of HSE. Four patients had neurological deterioration interpreted as relapsing HSE. Four non- relapsing HSE cases were selected as matched controls. Fifty- nine batched, paired CSF and serum samples from the eight HSE patients were analysed for soluble activity markers,predominantly cytokines and mediators (interferon- γ , soluble CD8, tumour necrosis factor- α , and interleukin- 10), amount of HSV- DNA and markers of glial and neuronal destruction (neurofilament protein, glial fibrillary acidic protein, S- 100- β , and neuron specific enolase). Results: Relapse of HSE was diagnosed in 3 of 26 (12% ) acyclovir- treated patients (5 episodes during 6.1 years of follow- up) and in 1 of 6 vidarabine- recipients. All relapses occurred from 1 to 4 months after acute HSE, except for a second relapse after 3.3 years in one patient. Computer tomography at relapses revealed few abnormalities apart from those found during the primary disease. Intravenous acyclovir and corticosteroids were given for 7- 21 days in all the relapse patients. All relapse patients seemed to recover to the prerelapse condition. HSV- DNA was demonstrated in CSF in all patients during the acute stage but not in any of 13 CSF samples taken during relapse phases. The HSV viral load during the acute stage of HSE was not higher or of longer duration in the relapsing patients than in the non- relapsing HSE controls. The levels of sCD8 were increased in nearly all CSF samples tested with peaks of sCD8 at one month of acute HSE.Inall episodes of relapse, sCD8 peaks were detected during the first week at high levels. CSF levels of neuron- specific enolase, S- 100 and glial fibrillary acidic protein were markedly lower at relapse than at the acute stage of HSV- 1 encephalitis. Conclusion: The lack of demonstrable HSV DNA in CSF, the lack of acute CSF signs and the lack of signs of neural and glia cells destruction indicate that a direct viral cytotoxicity is not the major pathogenic mechanism in relapse. Instead, the pronounced CSF proinflammatory immunological response and the relative lack of CSF anti- inflammatory cytokine IL- 10 response suggest immunologically- mediated pathogenicity. Objectives: To study the occurrence of relapse of herpes simplex encephalitis (HSE) and to find out whether soluble activity markers in cerebrospinal fluid (CSF) indicate direct viral or immune-mediated events. Methods: A consecutive series of 32 adult survivors of HSE were followed by the determination of the incidence of clinical relapse of HSE. Four non-relapsing HSE cases were selected as matched controls. Fifty-nine batched, paired CSF and serum samples from the eight HSE patients were analysed for soluble activity markers, predominantly cytokines and mediators (interferon- γ, soluble CD8, tumor necrosis factor- α and interleukin-10), amount of HSV- DNA and markers of glial and neuronal destruction (neurofilament protein, glial fibrillary acidic protein, Results: Relapse of HSE was diagnosed in 3 of 26 (12%) acyclovir-treated patients (5 episodes during 6.1 years o f follow-up) and in 1 of 6 vidarabine- recipients. All relapses occurred from 1 to 4 months after acute HSE, except for a second relapse after 3.3 years in one patient. Computer tomography at relapses revealed few abnormalities apart from those found during the primary disease. Intravenous acyclovir and corticosteroids were given for 7- 21 days in all the relapse patients. All relapse patients was recovered to the prerelapse condition. HSV-DNA was demonstrated in CSF in all patients during the acute stage but not in any of 13 CSF samples taken during relapse phases. The HSV viral load during the acute stage of HSE was not higher or longer longer in the relapsing patients than in the non- relapsing HSE controls. The levels of sCD8 were increased in nearly all CSF samples tested with peaks of sCD8 at one month of acute HSE. All episodes of relapse, sCD8 peaks were detected during the first week at high levels. CSF levels of neuron-specific enolase, S- 100 and glial fibrillar y aThe lack of demonstrable HSV DNA in CSF, the lack of acute CSF signs and the lack of signs of neural and glia cells destruction indicate that a direct viral, cytotoxicity is not the major pathogenic mechanism in relapse. Instead, the pronounced CSF proinflammatory immunological response and the relative lack of CSF anti- inflammatory cytokine IL-10 response suggest immunologically-mediated pathogenicity.