Printer Friendly

Restrictions on spirometry.

Spirometry is well established as the "Gold Standard" to diagnose airway obstruction. Its ability to identify restriction, however, has not been well studied until recently. Classically, a restrictive ventilatory defect (restriction) is defined as a reduced total lung capacity (TLC). Knowing TLC is reduced can move the clinician to a take a different diagnostic approach for a patient. It would be nice to be able to identify this problem without the additional time and expense of measuring TLC directly. The interpretative approach taken by the ATS1 states that a low vital capacity, when there is no airway obstruction, allows one to infer the presence of a "restrictive ventilatory defect" (reduced TLC). The word "infer" was carefully chosen to convey a level of uncertainty. The next step in the ATS recommendation was to call for further testing if it was clinically indicated.

Two studies have recently looked at spirometry to see if it could be used to define restriction. Aaron studied 1,831 consecutive patients presenting to a university laboratory with orders for both spirometry and lung volume measurements. A spirometric pattern suggesting restriction (low vital capacity, no airway obstruction) was seen in 264 patients but the concurrent measurements of TLC showed reductions in only 58% of them. That's about half, not much better than the flip of a coin. In contrast, when the spirometric vital capacity was normal, 98% of the TLC measurements were also normal. They concluded that spirometry was very good at excluding restriction but it could not accurately predict reductions in TLC. In a following study from the same laboratory, Glady developed an algorithm to predict the presence of a restrictive ventilatory defect in individual patients. In her sample, restriction could be reasonably excluded if FVC is at least 85% of predicted, OR, if the patient has an FVC <85% in the presence of significant airway obstruction (defined as an FEV1/FVC ratio less than 55%).


How do we put these papers in perspective? First, both Aaron and Glady studied consecutive patients referred for spirometry and TLC measurements (the "gold standard" for restriction). This approach sets studies up for a possible "ascertainment bias". Ascertainment bias can occur when there is a systematic tendency in the choice of those who get the "gold standard" test and those who do not. In this case, it is possible the patients scheduled for TLC measurements had an increased likelihood of a restrictive process. If that were true--meaning ascertainment bias--was present, it could cause large false shifts in the estimation of the sensitivity and specificity of a diagnostic test. While Aaron and Glady did not address this problem, a third study of the spirometric restrictive pattern, now in press, did. Using a similar subject recruitment strategy, it evaluated the potential for ascertainment bias and found it was not present. That doesn't eliminate the possibility that it was present in the Aaron and Glady studies but it does make it much less likely.

The substantive finding of all three studies is that a restrictive pattern in spirometry (remember this means a low vital capacity in the absence of obstruction) predicts a low TLC only half the time. You may think that sounds kind of lame. Looked at another way, a test that changes the likelihood of an illness (or presence of a diagnostic pattern) by 50% is pretty impressive. In this case, it is not enough. That 50% possibility of restriction should probably lead to further testing depending on the clinical situation. The next test might well be a chest radiograph rather than another PFT.

Information about total lung capacity is also available in the second most commonly used pulmonary function test, the carbon monoxide diffusing capacity (DLCO). The alveolar volume (VA) measured in the single breath DLCO test is more or less an estimate of TLC. In fact, it is commonly--and inexactly--called the single breath TLC. If VA is to be used as an estimate of TLC, one needs to know how it performs. VA systematically underestimates true TLC and the underestimation increases with worsening airway obstruction. The Epidemiology Standardization Project defined the relationships: VA underestimates TLC an average of 400 ml in healthy subjects and 500 ml in patients with interstitial lung disease. For patients with airway obstruction, the average underestimation is 1.01 liters in those with mild COPD progressing to 2.78 liters in those with severe COPD. Punjabi published an algorithm that reasonably predicts TLC in patients with moderate to severe airway obstruction using measured VA and the FEV1/FVC ratio.

How could you use these studies in your laboratory? First, you could use the spirometry patterns that accurately predict that TLC is either normal or increased to reduce the number of TLC measurements; this would reduce the cost to patients. Second, a near normal single breath TLC (within 400-500 ml of normal) will reasonably exclude a reduced TLC; it will not exclude over inflation, a too large TLC. Third, you could consider using Punjabi's equation to predict a correct TLC. Using one or more of these would allow you to reasonably avoid TLC measurements in a large group of patients because you can accurately predict their TLC is either normal or too large. This would leave you a smaller pool of patients in whom exclusion of a restrictive pattern is not enough. Whether you measure TLC at this point depends on the clinical question being asked and the relevance of TLC measurement in answering the question.

I can see your heads spinning now. Trying to incorporate all this information with every patient could take more time than just doing the darn test and sorting out the data later. On the other hand, the patterns you see can be readily sorted and that means it could be patterned. Algorithms could be set up to allow pulmonary technicians to make the test choices while the patient is in the laboratory, thereby improving laboratory efficiency and reducing the number of unhelpful tests and/or the number of times patients have to return to the laboratory.

The methods used to determine the performance of spirometry in determining the restrictive pattern are an interesting extension of the noise and signal discussions we've had in previous articles. The results are especially important as a limitation as spirometry becomes more widely available in primary care settings and test quality and interpretative skills are more variable.

by Robert Crapo MD
COPYRIGHT 2004 Focus Publications, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2004 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Crapo, Robert
Publication:FOCUS: Journal for Respiratory Care & Sleep Medicine
Geographic Code:1USA
Date:Sep 22, 2004
Previous Article:Traveling by car with oxygen therapy.
Next Article:Dying in your sleep: thank evolution.

Related Articles
Breathing new into home care: Texas hospital remotely monitors the pulmonary function of post-transplant pediatric patients.
EasyOne Spirometry System.
EasyOne Spirometry System.
Which reference set for spirometry?
WTC five-year assessment: Herbert et al. respond.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |