The new ERS/ATS standards for DLCO testing were published in the January issue of the European Respiratory Journal. The article was published as open access and can be downloaded from the ERJ website.
The biggest difference between the new standards and those from 2005 is that they are now primarily oriented towards Rapid-response Gas Analyzers (RGA). The authors explicitly state that the new standards do not make older systems that use discrete alveolar sampling and slower gas analyzers obsolete, but many of the new suggestions and requirements for labs and manufacturers require systems with a RGA.
The differences between the 2017 and 2005 standards that I’ve been able to find include:
♦ Flow accuracy was not specified in the 2005 standard but is now required to be ± 2% over a range of ± 10 L/sec.
♦ Volume accuracy is now required to be ± 2.5% (± 75 ml) instead of ± 3.5%. Notably the 2005 standard included a ± 0.5% error in the calibrating syringe. The accuracy of the 3-liter syringe is now stated separately. In the 2005 standard volume accuracy was over an 8-liter range. No volume range is specified in the 2017 standard.
♦ RGA response time (analyzer rise time) had not previously been specified but is now required to be ≤150 milliseconds. Sample transit time was discussed but no specific recommendations were made. Sample transport issues such as Taylor dispersion, gas viscosity and turbulence at gas fittings was also discussed and although it was suggested that manufacturers attempt to minimize these effects no specific recommendations were made.
♦ Analyzer linearity for both RGA and discrete sample systems has been relaxed to ± 1.0% in the 2017 standards from ± 0.5% in the 2005 standards.
♦ CO analyzer accuracy for both RGA and discrete sample systems is now specified as ≤10 ppm (which is ±0.3% of 0.3% CO). It was previously specified as ± 0.0015% (which is ± 0.5% of 0.3% CO).
♦ Interference from CO2 and water vapor for both RGA and discrete sample systems is now specified as ≤10 ppm error in CO (when CO2 and water vapor are ≤5%). Interference was recognized as a problem in the 2005 standard but error limits were not specified.
♦ Digital sampling rate was not discussed or specified in the 2005 standards. It is now specified as a minimum of ≥100 hz with a resolution of 14 bits. A 1000 hz sampling rate is recommended.
♦ Analyzer drift is specified in the 2017 standard as ≤ 10 ppm for CO and 0.5% of full scale for the tracer gas over 30 seconds. The 2005 standard specified drift as 0.5% of full scale over 30 seconds and did not differentiate between CO and tracer gases. The 2017 standards recommends that manufacturers provide a test mode to test drift.
♦ Barometric pressure sensor accuracy is now required to be within ±2.5%. This was not previously specified.
♦ Manufacturers of RGA systems are now required to include the following features:
- Monitor and report end-expiratory tracer gas and carbon monoxide concentrations and to alert the operator if the washout from previous testing is incomplete.
- Compensation for end-expiratory gas concentrations prior to test gas inhalation in the calculation of VA and DLCO.
- Ensure proper alignment of gas concentration signals and the flow signal (although notably specifications for this are not included).
- Measure anatomic dead-space using the Fowler method.
- Display a graph of exhaled gas concentration versus volume (not time) to confirm the point of dead-space washout and to report the amount of manual adjustment if this was done.
- Measure VA using all of the tracer gas data from the entire maneuver in the mass balance equation.
- Report the DLCO adjusted for the change in PAO2 due to barometric pressure.
♦ Further recommended (but not required) RGA options include:
- Ability to input simulated digital test data and compute DLCO, VA, TLC and Vb with ± 2% accuracy expected.
- Report the DLCO adjusted for change in PAO2 due to PACO2 with ± 2% accuracy.
♦ The maximum inspiratory pressure for demand valves has been reduced to <9 cmH2O from <10 cm H2O.
♦ Machine deadspace for adult testing has decreased to 200 ml from 350 ml. There was a further recommendation that machine deadspace should be smaller for children and patients with a VC <2.0 L but no specific requirements were made.
♦ Daily volume calibration must now be performed three time with a 3-liter syringe with using varying flow rates between 0.5 and 12.0 L/sec (injection times 0.5 – 6.0 seconds). Accuracy was not previously specified and now must be <2.5% error.
♦ Timer accuracy was specified in the 2005 standards. There are no timer specifications in the 2017 standards.
♦ Flow sensor zeroing prior to testing is now required.
♦ Gas analyzer linearity must now be checked monthly. The 2005 standards specified every three months. Manufacturers are urged to automate this process.
♦ A monthly calibration syringe leak test is now required.
♦ Both biological and calibration syringe QC testing are now required weekly. Previously either biological or calibration syringe QC were to be preformed weekly.
♦ Previous calibration syringe QC required the measured inspired volume (VI) to be “~3.30 L”. 2017 standards require accuracy to be ± 300 ml of VI * {STPD to BTPS conversion factor}. [Please note that there is a typo in the paragraph specifying this on page 8, line 8 in the 2017 standards where VA was substituted for VI].
♦ Changes in biological QC requiring action have been relaxed to a >12% change or >3 ml/min/mmHg (whichever is larger) from a simple >10% change. The 2017 standards also state that a mean of 6 prior tests should be used for this while the 2005 standards merely stated “from previous values”. Manufacturers were urged to developed automated QC processes.
♦ Calibration and QC logs can now be kept in a digital file folder.
♦ It was recommended that deep breaths during the pre-test tidal breathing period should be avoided in the 2005 standards but this was not included in the 2017 standards.
♦ The maximum acceptable time for exhalation to RV has been increased from 6 seconds to 12 seconds.
♦ Target VI has changed from ≥85% of the patient’s largest VC to ≥90%. The 2017 standards state however, that a VI of ≥85% of the patient’s largest VC is acceptable if the VA is within 200 ml or 5% (whichever is larger) of the patient’s highest VA from acceptable DLCO maneuvers.
♦ The 2017 standards now recommend that with RGA systems the exhalation following breath-holding should continue to RV in order to calculate VA using a mass-balance equation. Total expiratory time for discrete sample systems (washout and sample collection time) is still ≤4 seconds but is specified as ≤12 seconds in RGA systems.
♦ The DLCO test gas mixture is now required to contain 21% O2. The 2005 standards discussed a range of FIO2’s from 0.17 to 0.21 but only recommended that “inspired oxygen partial pressure values similar to the reference set used in the interpretation” be used.
♦ The required interval between tests (4 minutes minimum, 10 minutes for patients with severe obstruction) now includes the recommendation that the tracer gas concentration at end-exhalation (prior to the inhalation of the test gas mixture) should be ≤ 2% of the inspired concentration.
♦ The 2017 standards now recommend that the end-exhalation concentrations of CO (prior to inhalation of the test gas mixture) be used to adjust DLCO tests for CO back-pressure, to calculate COHb and to compensate for the effects of water vapor and CO2 on gas analyzers.
♦ The 2017 standards discuss the effect that prior testing (spirometry, bronchodilators and N2 washouts) have on DLCO and states that:
- bronchodilators are unlikely to affect DLCO and may therefore be used prior to DLCO testing
- prior spirometry efforts may affect DLCO but this has not been proven and therefore makes no recommendations against performing spirometry prior to DLCO testing
- sufficient time for alveolar O2 levels to return to normal is needed (2 times O2 wash-in time) after performing an N2 washout test. The standard recommends against performing N2 washout tests before DLCO testing but did not make this a requirement.
♦ The 2017 standards recommend that RGA systems calculate VA using mass-balance equations and this is discussed in detail (pages 17-19). This was not previously discussed nor was it an option.
♦ The equation for calculating anatomical deadspace using height (equation 20, page 16) is different from the one suggested in the 2005 standard (equation 10, page 728) but appears to be a restatement rather than being completely different.
♦ The 2017 standards discuss the measurement of anatomical dead space using the Fowler technique in detail. This was not previously discussed.
♦ The 2017 standards discuss flow and gas analyzer signal alignment in detail. This was not previously discussed.
♦ The 2017 standards discuss KCO with significantly more detail than in the 2005 standards. In particular although KCO is calculated from DLCO/VA it should be reported as KCO and not DLCO/VA.
♦ The 2017 standards discuss measuring the Phase III slope during exhalation as an index of ventilation inhomogeneity but does not specifically recommend it (probably a good idea since Phase III slopes during a single-breath N2 washout are usually obtained with low and constant expiratory flow rates which aren’t necessarily congruent with the expiratory flow rates during a DLCO maneuver and the differences between the manner in which these slopes are measured has not been studied).
♦ Repeatability between DLCO tests is now 2.0 ml/min/mmhg compared to 3.0 ml/min/mmHg in the previous standards.
♦ The 2017 standards include a suggested scoring/grading system for test quality based on inspired volume, breath-holding time and sample collection time (Table 3, page 22).
♦ Although the 2017 standards discusses the use of expired CO2 to estimate PAO2 it does not specifically recommend this.
♦ The 2017 standards have a new equation (38, page 24) used to correct DLCO for end-exhalation CO.
♦ The 2017 standards have new equations for altitude and barometric pressure correction (40 & 41, page 25) that are not a restatement of the previous equation (16, page 730).
♦ The 2017 standards include equations (42 & 43, page 25) to estimate barometric pressure at altitude that were not in the previous standards.
♦ The 2005 standards included equations to correct DLCO for alveolar volume (19 & 20, page 730) that are not discussed and not included in the 2017 standards.
♦ The DLCO results that are reported (Table 4, page 26) are now required to include:
- DLCO adjusted for barometric pressure
- DLCO LLN and/or Z-score
- VA LLN and/or Z-score
- KCO (instead of DLCO/VA)
- KCO LLN and/or Z-score
- Barometric pressure
- Average Breath-hold time
- Fowler anatomical dead space (RGA systems only)
- Single-breath TLC (RGA systems only)
- Test quality grade for acceptable maneuvers
- Reference values source
- Graphs of full maneuver.
- Graphs of exhaled gas concentrations versus volume (RGA systems only).
♦ The 2017 standards do not recommend any specific reference equations but note that many of the equations in use predate the 2005 standards. A short list (Table 5, page 27) of studies that were performed using the 2005 standards was included but it should be noted that the majority of the listed studies are pediatric and not adult.
Overall however, I am pleased that new DLCO standards have finally been released and that they are as comprehensive and forward looking as they are. I am pleased to see that many issues I had raised in previous blog postings have been addressed. I am also pleased that the bar has been raised on technical specifications, including calibration and quality control, and that care that has been taken make the standards as applicable to both old and new test systems.
There are two issues, however that I feel the need to comment on.
One of the biggest new recommendations is for patients on RGA systems to exhale to RV following the breath-hold period and that VA should be calculated with a mass-balance equation using the tracer gas concentrations that occur during the entire exhalation. DLCO is calculated however, using the CO gas concentrations from just the alveolar sample. Although the VA derived using mass balance will more likely approximate TLC, particularly when airway obstruction is present, other than the fact that this would raise the measured DLCO in patients with COPD the overall validity of this approach was not discussed.
I am still hard-pressed to understand why the 2017 standards continue to recommend that hemoglobin correction be performed on the predicted DLCO and not the observed DLCO, particular since the observed DLCO is corrected for PAO2, altitude and end-exhalation back pressures. The logic of this approach not only seems backwards but makes the comparison of trended DLCO results difficult. I read section of the new standards on hemoglobin correction carefully but no particular justification for this was put forward.
In the past I would have taken the release of the new DLCO standards as a sign that new standards for spirometry, lung volumes, methacholine challenge testing(?) and interpretation would also be released soon. The release of the previous standards in 2005 was troubled by confidentiality issues (early release of standards by insiders to some manufacturers) and this time around the ERS and ATS are playing their cards close to their vest. We can only wait and hope.
References:
MacIntyre N, et al. Series ATS/ERS task force: Standardisation of lung function testing. Standardisation of the single-breath determination of carbon monoxide uptake in the lung. Eur Respir J 2005; 26: 720-735.
Graham BL, et al. 2017 ERS/ATS standards fir single-breath carbon monoxide uptake in the lung. Eur Respir J 2017; 49: 1600016
PFT Blog by Richard Johnston is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License