The German study on sudden infant death (GeSID) is a multi-centre case-control study aiming at the assessment of etiological factors and risk factors of SIDS. This report describes the study design and the methods applied and presents some general findings. Between 1998 and 2001, 455 cases of sudden and unexpected death of infants aged between 8 and 365 days were recruited into the study. The study comprised at least 11 out of the 16 German states with 18 centres involved. In 1999 and 2000, 75% of all SIDS cases registered with the Federal Office of Statistics (ICD 10/R95, n=384) in the study area were recruited into the study (n=286). A standardised autopsy including extended histology, microbiology, virology, toxicology and neuropathology investigations was carried out. Of the parents 82% (n=373) agreed to fill in an extensive questionnaire containing 120 questions reflecting all important aspects of the infant's development. For each SIDS case, the parents of three living control infants were interviewed. These controls were matched for age, gender and region (n=1,118). The response rate of the controls was 58.7%. Data were linked with medical records obtained from obstetrics departments, the children's hospitals, and general practitioners. Death scene investigation was performed in 4 study areas (cases: n=64, controls: n=191). All cases were classified into one of 4 categories using defined criteria: 7.3% of the children were assigned to category 1 (no pathological findings: SIDS), 61.1% to category 2 (minor findings: SIDS+), 20.4% to category 3 (severe findings: SIDS+) and 11.2% to category 4 (findings which explained the death: non-SIDS). In case conferences the previous history and circumstantial factors were included and an extended category (E-cat.) was defined. The consideration of these factors for the final classification is of great importance in the causal explanation of some cases. An analysis of 18 main variables in cases of categories 1-3 (SIDS) compared to the cases of category 4 (non-SIDS) showed significant differences for the sleeping position, coughing the day before death and breast-feeding indicating that the cases of both groups should be separated for further analyses.