Abstract
Purpose::
Standard Dk values for 7 available gas-permeable (GP) reference materials used for calibration were recently updated using 15 data sets from 10 different operators at 8 different sites with 3 types of instrumentation: 9 polarographic, 4 coulometric, and 2 gas-to-gas. Here, the Dk values in each data set were individually calibrated using 3 different strategies and the results of the strategies were compared.
Methods::
The updated standard Dk values were 8.8, 26.8, 51.8, 61.8, 101.0, 130.9, and 154.3 Fatt Dk units, respectively, for single lots of silafocon A, paflufocon C, itafluorofocon A, tolofocon A, siflufocon B, melafocon A, and tisilfocon A. Each Dk of a data set was calibrated using (1) a linear regression of the measured and standard Dk values of the rest of the materials in the data set; (2) a linear regression of the measured and standard Dk values of 2 materials in the data set having lesser and greater Dk, with exceptions of the greatest Dk for which both references were necessarily lesser, and the smallest Dk for which one reference was greater and the other was taken to be the zero point; and (3) a linear regression that involved the zero point and the Dk values of the reference material having the highest Dk.
Results::
All but 2 mean calibrated Dk values were within 1.5% of their corresponding reference values: the exceptions were -4.2% and -4.0% of the lowest Dk (silafocon) using calibration strategies 2 and 3. Only 2 more mean calibrated Dk values were outside ±1.0% of the references: +1.4% and +1.1% of the highest Dk (tisilfocon) again using strategies 2 and 3. Confidence intervals (95% based on ±1.96×SEM; n=15) were less than ±5% of the reference Dk with 5 exceptions: for the lowest Dk (silafocon) using all 3 calibration strategies (±14.6, 10.5, and 11.8%), and for the highest Dk (tisilfocon) using strategies 2 and 3 (±7.1 and 6.9%). In general, strategy 3 resulted in slightly larger confidence intervals.
Conclusions::
Calibrated Dk values were less reproducible at or below 10 Fatt Dk units for which multiple assessments may be necessary to obtain a 95% confidence interval of less than ±10% of the Dk value. Reproducibility of ±10% or better is desired for verification of the product tolerance of ±20% listed in ANSI Z80.20-2004 and ISO 18369-4:2006. Multiple assessments of the greater hyper-permeables such as tisilfocon may also be necessary using at least 2 references for calibration.
Keywords: contact lens • cornea: clinical science • cornea: epithelium