Application of the Standard Simulator in Quality Evaluation of the Diffusion Capacity of the Lung for Carbon Monoxide Instruments postprint

Abstract: Background　The accuracy of the diffusion capacity of the lung for carbon monoxide（DLCO）testing directly influences the clinical decision making of diagnosing and treating pulmonary diseases. However，a long-term use of the instrument results in performance drift and measurement error. At present，daily calibration tubes and biological human verification are commonly used in clinical practice，but their sensitivity is insufficient to detect potential errors in instruments. Therefore，it is urgent to explore more accurate and objective methods for instrument quality assessment. Objective　To simulate single-breath DLCO using a standard simulator；to determine the accuracy of the DLCO instruments，and to correct them based on the causes of instrumental errors；to observe the outcome and maintenance time after calibration，and to explore its application value in combination with biological human verification. Methods　A standard DLCO simulator（Hans Rudolph series 5560，U.S.A），with 3 concentrations of carbon monoxide mixture（0.08%，0.10%，and 0.13%）and 3 kinds of inhalation volume （1.5 L，3 L，and 4.5 L），was used to simulate single-breath DLCO testing. The absolute error between the DLCO testing value and the target value of less than 2 mL·min-1·mmHg-1 was considered as the acceptable range. The accuracy of 4 DLCO instruments（2 brands each with 2 models）and the cause of instrumental errors were examined. Then each instrument was carried out corresponding maintenance according to the error，followed by comparison of the instrumental accuracy before and after calibration，and within 3 months of calibration. Differences in instrumental measurements were verified by biological human verification. Results　The average absolute error of DLCO in the baseline state was higher than 2 mL·min-1·mmHg-1 in 50% （2/4）of the instruments. The sources of errors of each instrument varied，including damages to the instrument accessories，wrong operating methods，and air leakage of the fixed cylinder. After calibration，the average absolute error of DLCO of all instruments was significantly lower than 2 mL·min-1·mmHg-1（P<0.001），and maintained for 3 months. After calibration，the differences in biological human measurements between different instruments significantly decreased，especially the extreme difference and coefficient of variation（CV）of DLCO measurements（P<0.05）. Conclusion　There are large differences in the accuracy of the DLCO instruments after a long-term use in clinical practice，and there are large differences in measurements between different instruments. Using standard simulators to evaluate the quality of DLCO instruments is an effective and feasible approach，which can evaluate the accuracy of the instrument and make up for the shortcomings of daily calibration. Standard simulators are recommended to be regularly applied for quality evaluation and quality control.