On-site tests based on immunoassay techniques are widely used for toxicologic screening analysis in patients with suspected poisoning. However, such assays usually have been validated using urine samples with known concentrations of the investigated substances. In the present investigation, on-site screening results were evaluated in a clinical setting. This was a retrospective study of patients with suspected poisoning from January to December 2003 in the emergency department of a tertiary urban hospital. Urine samples were analyzed using the Triage 8 panel and gas chromatography-mass spectrometry (GC-MS). A total of 111 patients were included (54 female, 57 male; average age 37.8±19.7 years). A total of 3.8% of the patients showed no symptoms, 45.2% minor, 24.0% moderate, and 26.9% serious symptoms. In 50 patients (45.0%), Triage 8 results corresponded well with GC-MS results. In 17 patients (15.3%), the Triage 8 results were confirmed by GC-MS, but additional substances were determined that could not be detected by the Triage 8 panel. A completely negative Triage 8 screening result was obtained in 23 patients (20.7%) who showed toxicologically relevant findings in GC-MS. In 21 patients (18.9%), Triage 8 results could not be confirmed by GC-MS. The analysis of the results in view of the patients' medical histories revealed that in 20 patients (18.0%), no relevant toxic substance could be detected. Additionally, 8 patients (7.2%) showed intoxication with alcohol, which could not be detected by the presently applied toxicologic screening investigations. On-site screening results in suspected poisoning were not very helpful in the present study because practically every second patient ingested substances that were not detectable by the Triage 8 device. In addition, every fifth result was not in line with GC-MS findings. On-site test findings should be interpreted very carefully, and in critical cases, a GC-MS screening should be performed.