Routine diagnosis of animal trichinellosis for food safety and trade relies on a method of artificial digestion to free Trichinella muscle larvae from meat for subsequent identification by microscopy. As part of a quality control system, the French National Reference Laboratory (NRL) initiated ring trials to determine the sensitivity of the test performed in the 72 routine diagnostic laboratories in France. A method was devised to obtain calibrated meat samples containing known numbers of capsules with Trichinella spiralis muscle larvae. This method was based on an incomplete artificial digestion of Trichinella-infected mice carcasses to allow the collection of intact Trichinella capsules. Capsules were placed into a meatball of 100 +/- 2 g of pork and horsemeat to produce proficiency samples. Three categories of samples were prepared: small (3 to 5 capsules), medium (7 to 10), and large (12 to 15). The sensitivity was expressed as the percentage of muscle larvae recovered from each proficiency sample. Reproducibility was tested with ring trials organized between two NRLs (France and Canada), and a reference sensitivity of 84.9% was established. National ring trials were then organized in France, with the 72 routine diagnostic laboratories each receiving four proficiency samples per session. After five sessions, an improvement in the digest test sensitivity was observed. Results at the fifth session indicated sensitivities of 78.60% +/- 23.70%, 81.19% +/- 19.59%, and 80.52% +/- 14.71% muscle larvae for small, medium, and large samples, respectively. This study supports the use of proficiency samples to accurately evaluate the performance of routine diagnostic laboratories that conduct digestion tests for animal trichinellosis diagnosis.