Monday, February 27, 2006

IPF recap

I thought it would be fun to highlight some of the previous cases. This is also useful because occasionally we'll get a comment long after the case was posted so it may get lost as new cases bump the old ones back.

Recently there was some interesting discussion about the treatment for IPF. There was the article on Acetylcysteine and pulmonary fibrosis (IPF). More often than not, we read a high impact journal and take what it finds at face value. As busy physicians we might not have the time or patience to read an article critically, and we sometimes "trust" that the peer reviewers read it critically for us. This may be especially true in a high impact journal like the New England Journal of Medicine. It seemed like a straight forward question: "where do I get acetylcysteine to prescribe for my patents?" But the comments highlight the fallacy of believing everything we read. As Jeff H pointed out, I think it's critical to point out that NAC was used IN ADDITION to Azathioprine and Prednisone. There are no trials that have evaluated NAC as monotherapy. Moreover, there was no "placebo" group included in the NEJM paper. So, there are NO DATA that oral NAC, as monotherapy, is any better (or worse) than nothing. Doug (and others) highlighted this, pointing out that NAC may have merely had a protective effect on the harmful effects of prednisone. And jgiustino's comments are interesting: All I can say is that this was a bad study! The data from patients that actually completed the study did not show ANY significant change in lung function between the two groups. The authors then needed to resort to "last observation carry forward" to show significance ....this is a highly controversial statistical method. I am very surprised that this has not been questioned by more people. I think that this study proved that NAC prevent bone marrow toxicity from imuran (they did not need "last observation carry forward" to prove this. All else is very suspect as far as I'm concerned and I wish they had the foresight to get a larger patient population into the study so they would not have to resort to such statistical games.

The diagnosis of UIP can also be challenging. What's more important, a biopsy showing UIP or typical CT findings of honeycombing at the bases with minimal ground glass? What about a biopsy with UIP and previous CTs showing LOTS of ground glass? I posted a case where a patient had an open lung biopsy that was consistent with UIP but with chest CT's back 6 years showed that there has been lots of ground glass (and upper lobe to boot). My question was, does the biopsy trump the atypical CT findings, OR do the previous CT findings make us question the assumption that UIP makes the diagnosis of IPF? I'm not sure the question has been adequately answered but the comments shed some light on the controversy. More than one pointed out that the diagnosis of UIP is also dependent of the pathologist.
And if we have a diagnosis of UIP by biopsy, would we withhold treatment with immunosuppressives? What if the pathologist read it wrong and it is truly a steroid-responsive disease? One commented that this is also the best argument on why we should "treat" our UIP patients with a 3 to 6 month course of prednisone and imuran..not to treat UIP but to definatively rule out other treatable diagnosis."

So the same research group showing that UIP trumps all other biopsy specimens (in terms of prognosis) also found that the dignosis of UIP is pathologist-dependent. In other words, if you are not at a major university specializing in IPF, ignore the biopsy report and treat everyone with prednisone for 6 months to see if they get better!

But seriously, it sure is strange that a pathologist would have difficulty with the diagnosis. After all, all you need is a biopsy showing fibroblastic foci, areas of normal parenchyma, areas of fibrosis and honeycombing in the periphery, and a paucity of inflammatory cells. What's the problem?

6 comments - CLICK HERE to read & add your own!:

jgiustino said...

I think part of the problem may simply be sampling error. I do not have studies to prove that this occurs however. Open lung biopsies are smaller than many pulmonologist believe. They are also not taken from deep within the lung but from the lung periphery ( where IPF "lives")
Again, I have no solid evidence on this but I have spoken to our pathologists about this who tell me that if the clinical findings or CT findings do not fit with UIP then the patient should be treated.

With respect to the NEJM article, if you all do journal clubs, please review this. I think its a farce. If you do a search on "last observation carry forward" you will get a sense of how uncomfortable the statisticians are with this method. You simply cannot reliably create data that is not there. Also, to use the method you need to know if the data that was lost was randomly lost or not. The authors conclude that the data was randomly lost based merely on the fact that they could not find a correllation with the patient factors they collected. But does this imply the data was randomly lost. They have no idea whether it was or was not. It is possilbe that patient loss correllate with religion. Since they did not collect data on patient religion , there is no way to know if this was a factor. I was very troubled by the fact that the NEJM published this. But you should all read the article and make up your own minds. The first clue of trouble is that the stats section is twice as long as the results section....

Jeff H said...

Thanks for the comments--I think you raise important points regarding the NEJM-NAC paper. I'm checking the journal every week, waiting for these issues to be raised in a letter-to-the-editor.

With regards to UIP in general, I think it's crucial to not lose sight of the fact that UIP is not the same as IPF. UIP is a histologic/pathologic pattern, whereas IPF, by definition, is a clinical-radiographic-pathologic diagnosis. In other words, the diagnosis of IPF requires that alternative diagnoses, such as collagen-vascular disease, drug toxicity, pneumoconioses, granulomatous diseases, etc. be excluded. So your pathologists are correct--finding UIP on a biopsy without an appropriate clinical and radiographic correlation should raise doubt about a diagnosis of "IPF."

So, IPF is defined histopathologically by UIP, but not all UIP represents IPF...

Jennings said...

"So, IPF is defined histopathologically by UIP, but not all UIP represents IPF..."
But does all UIP's have the same prognosis (i.e. UIP from IPF vs UIP from a patient with rheumatoid lung)?

Jeff H said...

I think that the presence of UIP, regardless of the etiology, portends a worse prognosis than non-UIP. That said, a diagnosis of IPF (with, by definition, UIP), probably has a worse prognosis than UIP in non-IPF patients (i.e. IPF has a worse prognosis than RA-associated UIP); however, I'm not sure that that kind of study has ever been done.

With the decreasing frequency of surgical lung biopsy, I'm not sure that kind of study will ever be done.

Anonymous said...

I don't know if I should be posting this here, but I just wanted to tell you I appreciate your dedication to my disease. (I have RA and PF.)

My name is Shelly, I was a medical assistant back when I had a life (haha) so I am able to understand your posts pretty well. (My Pulmonologist just loves that about me.)

I just started NAC, I'm on my 3rd bottle, I take 500mg 3x per day. I also take Prednisone, Enbrel, Azathioprene, Bosentin (Tracleer for PAH). I am taking a PFT and an HRCT in the next few weeks to see if there is anymore progression. I'll then take my findings up to my Dr. at University of Washington in Seattle. I see the director of lung transplantation there.

Do you think I should expect any changes in my cat scans from the NAC? Or do you think it is too soon to tell? Thanks for any opinions you can give me! ~Shelly

Jeff H said...

Thanks for the comment. I think that your best source of information is your pulmonary physcician, as he will obviously know far more about you than we do.

NAC has not been studied in collagen-vascular disease associated pulmonary fibrosis. Moreover, I don't know of any study involving NAC that has looked at radiographic progression of pulmonary fibrosis as an outcome variable. Most studies of pulmonary fibrosis look at more quantifiable outcomes such as physiologic parameters and functional capacities such as 6 minute walk distance.