The aim of this study was to compare two wound healing models in surgical patients with special reference to the intrapatient variability. Two pieces of expanded polytetrafluoroethylene tubing and one perforated silicone tube containing two polyvinyl alcohol sponges were subcutaneously inserted into 28 patients and were removed after 10 (9 to 11) days. The amount of hydroxy-l-proline in two segments of the same piece of expanded polytetrafluoroethylene tubing correlated significantly with each other. Similar results were found for proline. Proline content in each polyvinyl alcohol sponge correlated weakly with each other, whereas no correlation was found between hydroxy-l-proline content in the two polyvinyl alcohol sponges from the same patient. In addition, no correlation was found between leucine in the two adjacent polyvinyl alcohol sponges or between the hydroxy-l-proline/leucine ratio in these same sponges. No concordance could be detected between the averages of hydroxy-l-proline content in the two segments of expanded polytetrafluoroethylene tubing and the hydroxy-l-proline content in the two polyvinyl alcohol sponges from the same person, although an association between the averages of proline in the two pieces of expanded polytetrafluoroethylene tubing and proline content in the two polyvinyl alcohol sponges from the same individual was observed. The measurement error was greater for the polyvinyl alcohol sponges than for the expanded polytetrafluoroethylene tubing implants during the first 10 days after wounding. The expanded polytetrafluoroethylene tubing accumulated 2.5 times more hydroxy-l-proline per length unit than the polyvinyl alcohol sponge. We concluded that the expanded polytetrafluoroethylene implant rather than the polyvinyl alcohol implant is preferred for quantitation of subcutaneous collagen deposition in surgical patients because it accumulates more new tissue and has a lower variability.