OBJECTIVE: To evaluate the relationship between the HSV-1 and -2 loads in BAL fluid (BALF) and clinical outcome. DESIGN: Retrospective study. SETTING: The general intensive care unit of the University Hospital Maastricht. PATIENTS: Five hundred and twenty-one BALF samples from 462 patients were included. Patients were divided into three groups; (1) patients admitted to the hospital <48 h before lavage (Community), (2) patients admitted to the ICU >48 h before lavage (ICU) and (3) the remaining patients (non-ICU group). INTERVENTIONS: No additional interventions were conducted. MEASUREMENTS AND RESULTS: HSV-1 and HSV-2 loads were determined by real-time polymerase chain reaction (PCR). HSV-1 DNA was detected in 4.3% (4/92) of samples in the community group, 15% (18/121) in the non-ICU group and in 32% (99/308) of the ICU group. In the age group <50 years HSV-1 DNA was less frequently isolated compared to the age group >or=50 years (16/129 (12%) versus 187/376 (25%), respectively, OR = 2.6; P < 0.001). HSV-1 loads of >10(5) genome equivalents (ge)/ml were associated with an increased 14-day in-hospital mortality compared to patients with a HSV-1 load <or=10(5) ge/ml in BALF (41 vs. 20%, respectively, P = 0.001). HSV-1 pneumonia was histologically proven in two patients with a HSV-1 load exceeding 10(5) ge/ml. CONCLUSIONS: HSV-1 occurred more in critically ill patients and high loads in BALF were associated with an increased mortality. The higher mortality observed in patients with HSV-1 load >10(5) ge/ml enforces its clinical relevance and necessitates to start randomized medical intervention studies.