How a device-anatomy glossary and clinical data verification maintained exact alignment between a French IFU’s procedural text and its bilingual product drawing labels for a Class III retrievable nitinol filter.
The client
A manufacturer of percutaneous vascular intervention devices, producing a retrievable vena cava filter system indicated for the prevention of recurrent pulmonary embolism in patients for whom anticoagulation is contraindicated or has failed. The device uses a self-expanding nitinol structure and is available in both femoral and jugular delivery configurations, each requiring its own procedural steps and anatomy-specific terminology in the IFU.
The challenge
The IFU for a vena cava filter doesn’t just describe a procedure – it describes an anatomy.
The device has a specific structure: cranial and caudal retrieval components, lateral and extended tissue anchors, an ePTFE filtering mesh. Every one of those component names appears twice in the IFU – in the procedural text and in the annotated product drawing labels – and a mismatch between those two instances reads as a translation error at labeling review for a Class III implantable device.
The IFU also contained clinical study data written directly into the document – primary endpoint success rates, pulmonary embolism incidence figures, and retrieval success percentages – each of which required verification against source values before sign-off. The retrieval-system terminology added a further layer: French interventional radiology practice uses système de retrait for this component, which diverges from a literal translation, and any inconsistency across the IFU’s procedural and drawing-label sections would be identifiable to a reader familiar with the clinical field.
The solution
We created a device-anatomy glossary specific to the self-expanding nitinol filter design before translation began, covering both the femoral and jugular delivery approaches and establishing consistent French equivalents for every component name. The glossary anchored both the IFU body text and the product drawing labels to the same reference – eliminating the alignment risk at its source rather than resolving it during review.
For the clinical study data, each numerical value was verified against source figures before sign-off. This kind of outcome data is vital and can’t be carried across on the basis of translation memory alone, given its regulatory function in a Class III device label.
The results
- Consistency across the body of the IFU and the bilingual product drawing labels. The nitinol filter’s component names (cranial and caudal retrieval systems, lateral and extended tissue anchors, ePTFE filtering mesh) matched exactly across both document types.
- All reported clinical study data verified against source values before delivery, given their regulatory weight in a Class III device label.
- Validated French terminology such as système de retrait was established with consistent application across procedural text and drawing labels.
