Which do you use?
Our lab switched ~3 years ago to a composite set:
Spirometry: Eigen (3-5 yrs), Dockery/Wang (6,7 yrs), NHANESIII (8-80)
Lung Volumes/Diffusion: Rosenthal (6-18); Crapo (19+)
Resp Muscle Pressures: Wilson
Subscribe to:
Post Comments (Atom)
Dear Daniel,
ReplyDeleteThank you for pointing me to your blog, which looks very professional. I have one comment to make: You use disparate set of reference equations . I strongly recommend that you use Stanojevic 2009 (Am J Respir Crit Care Med Vol 180. pp 547–552, 2009) . The all-age equations not only are continuous from age 3-80 years, but they also properly model the scatter about the predicted values, which is not at all constant. It decreases for about 17% in very young children to about 10% during adolescence, and then increases again to about 17% in the elderly. I know that manufacturers do not make these predicted values available. So it is just a matter for end users to demand that they be made available , as they are the best that are available. This matters for clinical decision making. Wang has the disadvantage that, if a subject has a birthday, there is an inevitable discontinuity between predicted values just after birthday and one day before birthday. I am working very hard, in the framework of the Global Lungs Initiative (www.lungfunction.org) to supply all-age reference equations for various ethnic groups. I hope that they will become available early 2011.
All the best
Philip Quanjer
Dear Philip, thank you (as always) for your very insightful commenst. I have read the Stanojevic paper in the past and will revisit it. My recollections about why we have heretofore not used them (please correct me if I am wrong):
ReplyDelete> they are for Caucasian's only? in urban US cities such as ours, we have a decent sized population of African Americans and Hispanic americans.
> they are only for spirometry, and do not include pleth/diffusion
in addition, the CF Foundation utilizes Dockery/Wang for "peds" and then Knudson post-puberty (although we submit data to them in raw value format and they transform for their own internal statistics). The ATS has recommended NHANESIII/Hankinson for spirometry, although the ERS side was a bit more vague. When I first arrived in Pittsburgh, they were using Shoenberg (which also included a weight factor which appeared to give bizarre results in obese subjects!), and a bunch of years ago in Philadelphia, we were using Hsu until we made the change to this composite set.
If the issues of ethnicity can be addressed, I will revisit the Stanojevic equations.
(posted for Philip):
ReplyDeleteHi Daniel,
It is true that the Stanojevic equations are only for whites. That is why I am working so hard to derive equations for many ethnic groups. I hope they will become available early 2011. I am also working as hard I can to see to it that manufacturers will implement them by providing them with all sorts of information, even algorithms. And then, a common thing, the conservatism of organizations and individuals will mean that it takes some extra years before the equations are adopted.
Best wishes, Philip
Philip, I look forward to the multi-ethnic equations. I have to say, adding almost any kind of equation is trivial in the software we use (ComPAS, Morgan Scientific). If you have not seen that software yet, I encourage you to check it out! (www.morgansci.com)
ReplyDeleteI agree with Phillip that the Stanojevik equations make the most sense in terms of their continuous construction, but realize that they are themselves a synthesis of NHANES III, Rosenthal, Lebecque, and Corey, the latter 2 being fairly small studies. Also, they are limited to spirometry, and caucasians.
ReplyDeleteIn terms of practicality of use, they are not in the form of single equations, but require table lookups and equations to be executed. No manufacturer of PFT equipment yet allows for this to be done by the user.
I do suggest that the manufacturers should be looking at making these avaiable in their software. I have been using them to develop graphs of relationships between lung growth rate and morphometric growth rates in children to young adults.
In conclusion, the Stanojevic "set" is based on a statistical method to combine different sets into a continuous (and hopefully accurate) predited value curve. It is very promising, in my opinion.
Patrick, how feasible would these kind of lookup tables be to implement in ComPAS?
ReplyDeleteWhat DLCO/VA adjusted reference value do you use in your facility? I welcome replies from as many PFT labs as possible.
ReplyDeleteThanks
In Pittsburgh, we use diffusion predicteds from Rosenthal et al (Thorax, I think 1993). Our software (ComPAS, Morgan Scientific) allows one to select whether the PREDICTED value is adjusted (for Hb, COHb, MetHb) or the MEASURED value. The last ATS/ERS statement on DLCO did recommend adjusting the predicted value, so I am happy this supports that.
ReplyDeleteDaniel - I just spoke with Josh (Needleman) who I'm sure you know - I think our problem is the age of the patient. For example, I did a diffusion study on a 9-year-old with Beta Thalassemia Major, adjusted for Hemoglobin, and no reference value showed up for her DLCO/VA. Do you have a good reference value that will account for age, hemoglobin and alveolar adjustments?
ReplyDeleteThanks!
Diana
Diana, I can send you all our predicteds - send me your email address.
ReplyDeletesent to your email address. Thanks!
ReplyDeleteDaniel, what are your thoughts about using predicted values for spirometry for children who are of mixed race/ethnicity? Which predicted set would you use for a child, let's say for example, born to a Caucasian and African-American couple? Or Hispanic and African-American? Given that there is so much genetic complexity in these situations, does your lab follow any set of rules?
ReplyDeleteThat's a great question, Lokesh, and one for which there is no good answer. I think many guidelines recommend asking patients to self-identify race, but many might self-identify as mixed race, so that doesn't help. We don't have any firm rules. It certainly has been noted that ethnicity (and even upper/body segment ratios) is not the most conclusive predictor of lung function after height.
ReplyDeleteSome here at CHP have done genetic tests to look at markers of ethnic heritage, and I would wonder if you could use those in a way to "weight" the different ethnic predicted sets?