The aim of this study was to assess the relationships between accident mechanisms as well as initial findings and the long-term course of whiplash injury. A representative sample of 117 consecutive patients referred by primary care physicians was followed-up over 12 months. Fractures or dislocations of the cervical spine, head trauma and pre-existing neurological disorders were exclusion criteria. The interval between the accident and the baseline examination was 7.4 days (SD 4.2 days). Assessment included accident features (e.g. passenger position in the car, head restraint, head position, type of collision), initial symptoms (e.g. intensity and onset of pain, symptoms of neurological dysfunction, multiple symptom score), and signs (restricted neck movement, neurological deficits). At the 1-year examination, patients were divided into an asymptomatic and a symptomatic group and were compared with respect to accident features and baseline findings. Twenty-four percent of patients were still symptomatic after 1 year. Analysing accident mechanisms separately, rotated or inclined head position was the primary feature related to symptom persistence (P=0.005). The symptomatic group scored higher at baseline on the multiple symptom rating (P=0.004) and had a higher incidence of initial headache (P=0.004) and neurological symptoms (P=0.008) together with a higher intensity of headache (P=0.0002) and neck pain (P=0.0009). The following set of initial variables predicted persistence of symptoms at 1 year (logistic regression):​ intensity of neck pain (P=0.001) and headache (P=0.009), rotated or inclined head position (P=0.02), unpreparedness at the time of impact (P=0.01) and car stationary when hit (P=0.01). In conclusion, accident mechanisms and initial findings suggestive of more severe injury were significantly related to long-term persistence of symptoms after whiplash injury.