Variabilities in Viral Load Testing

[jmjm530]

Here is the transcript from slide 24 of Dr. Mitchel Shiffman's presentation on the Clinical Options web site. Very interesting stuff. Full slide presentation here: http://tinyurl.com/lfaev
Requires free registration.

Dr. Shiffman:

"Now one of the most important things to realize about HCV RNA assays is there's some inherent variablities in these assays. These assays only measure virus within a sensitivity of plus or minus half a log unit. That means there's a one log unit variation in their ability to measure HCV RNA.

What does that mean?

That means as shown here in this graph that somebody with a one million level of hep c virus is really about the same level of virus as somebody with ten million.

The assay really can't detect the difference between those, they're kind of in the same ballpark. And the Same goes for levels of about 100,000 to 1,000,000. Again, those are fairly comparable.

So, when you're measuring hep c virus and following a patient, and their virus levels suddenly goes up from a million to five million, the virus really hasn't gone up that much it's the assay is not able able to detect the difference between those two and they're really i"n the same ballpark.

You really need to see changes in assays levels of greater than at least one log, either a ten fold changes to be clinically meaningful because of the limitation of these assays."

###

[GoofyDad]

<a href="http://jcm.asm.org/cgi/content/full/40/6/2031">This bDNA paper may be of interest</a>. I didn't study it closely, but I think the take away is that bDNA is highly accurate for quantification, and it's accuracy diminishes at lower viral loads. My doc generally runs two tests: Bayer bDNA (for quantification) and TaqMan TMA (to detect presence/absence of virus).  

I think we had a chart floating around that showed specificity for the various VL tests. May have been posted by Bill1954.

[jmjm530]

Again, I don't know your tx history, but the standard protocol for geno 2's is still 24 weeks with some studies suggesting a shorter 16 or 12 week protocol depending on whether you're doing Peg Intron or Pegasys. I believe Goofydad recently reposted those studies but to summarize from memory, in order to qualify for the shorter course, you should be non-detectible at week 4 and probably have been on weight-based ribavirin per the study. Also, I believe the shorter course is not recommended for anyone with cirrhosis. The complete study, however, might be more useful.

All the best.

-- Jim

[NYgirl]

Well at least in here we've never really counted on vl for very much BUT it does make you really wonder = since they do all tell us that a low VL is a good precursor to SVR...just how low are we in fact.

Good post thanks as always!

[FlGuy]

Kind of makes you wonder about the studies that provide hints or 'predictors' based on starting viral load. Or, how close a person is to undetectable in the early weeks of tx.  NYGirl : thinking of those early pcrs you had that were in the neighborhood of 400 vl.

[jmjm530]

NY: Well at least in here we've never really counted on vl for very much
---------------
Respectfully disagree. Viral load does account for a lot. Shiffman's presentation just adds more info on how it may or may not be interpreted, nor does it IMO challenge the statistical model of low pre-tx viral load being a positive predictor of SVR, as the lower you go in VL, the less the variance becomes in terms of IUs per ML.

Another thing I got from the presentation was perhaps some clarification on what we sometimes term here as "partial" response. Say when someone's viral load drops during tx from 10 million to 5 million. In fact, the viral load may not have dropped at all since it's only half a log drop which could be attributed to test variance. That may in part be why they like to see a two-log drop or more over a 12-week period, or at least a one log drop per month as I've also seen it put to know that the drugs are definitely working.

-- Jim

[PasoPerson]

In my case, the Hep C was found after my NP noticed my liver enzymes were "a little high" and ordered additional testing.  My VL was about 80,000.  Two weeks later, at the hepatologist's office, my VL was about 6,000.  Three months later, pre-treatment baseline VL was 1.3 million.  After 24 weeks it was 66,000.  Then switched to Infergen, in 4 weeks VL was 2400.  To me, I was jumping all over the charts.  So I am very happy to read Dr. Shiffman's report.  THANK YOU.

- Carolyn

[jmjm530]

Carolyn,

I really don't know the protocols/expectaions for Infergen, so really can't comment on your situation. Just want to say that I don't think Shiffman's presentation is in conflict with any of the "stopping" rules, whether they be the older "two logs in 12 week" rule, or the "non-detectible in 12 week" rule or some of the newer guidelines that suggest we try for non-detectible even earlier. These are all statistical models that by nature cannot predict any particular individuals's future results but they can give us *odds* which can be very important in making treatment decisions.

-- Jim

[FlGuy]

This topic is thought-provoking and a little disconcerting.  With al the vagueness around hcv, I tend to cling to the numbers and when the numbers ger a little slippery, it makes for a little uneasiness. Now I have another decision.  My relapse vl in May was 2 million. I re-start in 4 weeks and had planned on another pcr pre-tx as a baseline.  Not wanting to do unneeded expensive tests, I wonder if it's medically necessary before re-starting.  I probably will anyway, because in the back of my mind I'm sorting of still hoping against hope that that one (May) was a false positive, even those I had the alt/ast's to sort of match the relapse. And, I'd hate to face 48 weeks with a small voice saying 'was it a false positive'.  Guess I'll just roll up my sleeve again.  Interesting subject though.

[NYgirl]

I think that if Dr. Schiffman's info is correct - than How in the world are we judging things like 2 log drops in the first place?

If it's kind of a subjective type thing and can fluctuate so WILDLY...then how can any of the tests/odds be really determined to be anything except speculation or a good guess?

I've never really counted on VL for very much of anything.  I started with a LOW VL at 568,000.  It didn't help me get to UND early or anything.  But now thinking it could have been 50,680,000 the very next day or something - it does make me wonder.

Maybe that one "low" day at 568k was just a fluke?

You know what I am trying to say?

VL doesn't really matter all of that much to me in any aspect.  I am more a Stage/Grade/Geno believer if anything.

But that's just me.

[EquineLori]

Hi, Jim!  I've been in CA settling my Mom's estate with my brothers for the last couple of weeks.  I came home for 2 days to have the all important PCR blood drawn.  As you may remember, I think I'm a geno 2 or 3 with a beginning viral load of just barely under 25,000,000.  However poor my blood tests in re. to wbc,rbc,anc, etc., my viral load came back UND with quant. of 10 to 100,000,000.  I'm going to have to postpone my apt. with Dr. Glombicki to go back to CA.  Have a new apt. with him 9/25.  Jim, with the info. you posted on this thread and a thread from mkandrew on 9/2 re. first phase viral response vs. the problems of second phase when IFN must work to kill the virus in the liver itself, how long must I treat?  I plan to ask Dr. G what the relapse rate is for my geno type.  However much I hate tx and don't wish to tx one day longer than is necessary, I also really can't afford to relapse.  I feel this is my one and only chance.  John has come out of stage 4 cancer remission and should something horrible happen to him, I don't believe I'd tx again.  Jim, what should I do.  I need your sage advice!  Thank you, Lori

[Mister beagle bailey]

I agree your 558k could be a fluke just like mine was 318k and now post tx vl is 4 million.  So what are we to believe.

Bob

[FlGuy]

You have any details yet on trial sites and stuff?  If your doc is going to help you, is this the NY doc?  After months of looking at clinicaltrials.gov, I get the idea that for most hcv trials they require that you be off ifn/riba for 6 months so the results are not shaded by previous tx.  I guess you get to your 6 months post tx in December?  If the folks going through the vx trails are representative of trials in general, the studies will keep pretty close track of you for testing.  If the site is in NY you'll probably be seeing more of the Empire State if you get in the trial.  Good luck with the yesterday pcr, just don't get your hopes up too much.  Sometimes low expectations can soften a blow.  Enjoy you non-tx time Beagle.  These last several months have been great for me, hate to see it end. But, you gotta do what ya gotta go.

[Mister beagle bailey]

Don't worry I don't have my hopes up about the yesterday PCR.  Doc wanted to also recheck my AST and ALT, as 4 weeks ago they were ALT 13 and AST 24. What ever it is it is, my mind is already moving on to the next step.  Haven't heard from the co. yet about trial but yes your right they want you 6 months post tx before starting a trial, which I'll be in Dec.  I thought I might be a good subject for the trial given my thalassemia and the new drug replaces riba and it's sxs of anemia.  Where ever the trial is I'll travel there if they take me.  LOL

When is your new start date?

Beagle

[jmjm530]

NY: I think that if Dr. Schiffman's info is correct - than How in the world are we judging things like 2 log drops in the first place?
----------------------------
I don't see any conflict. What I get from Schiffman's presentation is that something like a two-log drop IS significant, but anything less than a log probably is not. To me, this is consistent with the so-called "stop" rules or the later extended treatment rules. Response to viral load and pre-tx viral load have been shown to be important in various studies. To oversimply -- BIG changes in VL are clinically significant, small changes may not be. Isn't that what all the log rule studies seem to tell us?

Lori,

Congratulations on being non-detectible but I don't remember your history or what point in treatment you're in. You might want to post a new thread later and maybe addressing it to Goofdad who is a geno 2 and has posted several studies before regarding treatment lengths. As to the Shiffman study I just posted, I don't think this should affect any of the treatment protocols but simply explains why small (under a log) changes may not be as relevant as some people imagine. Forgot what MKAndrew said about
first and second phase response, but I do remember him mentioning timing treatment with low VL, but that's a somewhat different topic.

[FlGuy]

Begin again Friday October 13.  Sort of an appropriate day/ date.  It's get me through most of hurricane season, so there shouldn't be much crazziness. So, I'm ready.  See my pals at Quest tomorrow for a lot of the blood tests, but the pcr is a couple of weeks away.

[EquineLori]

Ok, Jim I've been gone and read threads most current down.  Just read the one from Northstar/lilmmoma and read about Mr. Beagle. Shi#!!!  My answer was there.  Thank you, anyway, Lori

[EquineLori]

Shite, Jim we were cross posting.  Thank you.  I'll finish with Calif., see Dr. G. and post in a few weeks.  Again, thank you.  And my love to all.

[friole]

I just this morning finally finished listening to this slide show and the point Jim was making here was very interesting to me as well.  If you are tested before you treat and the results are 10,000,000 and you are tested again right before you treat and the results are 1,000,000 it doesn't mean a dang thing.  The test can be up to a half log off in either direction.

Another interesting tidbit from the slide presentation is how they test for the genotype.  They have a celuloid strip and on it they put all the antigens for all the different types of genotypes.  Then they drop the virus on it (after it has been amplified -- run thru the PCR test) and the virus sticks to the antigen.

[jmjm530]

My previous post was was a cross post too :)

The best of luck.

-- Jim

[FlGuy]

Although no longer recruiting :
http://www.clinicaltrials.gov/ct/show/NCT00093093?order=24  maybe there are some points of contact that you can make at one or more of these trial sites.  Pehaps there is something to be learned from a trial coordinator or mabe there's some continuing investigation that's going on that isn't published yet.

[NYgirl]

Said:  That means as shown here in this graph that somebody with a one million level of hep c virus is really about the same level of virus as somebody with ten million.



That is what I was trying to say...not that a VL isn't important in anspect to findout log drops etc. BUT it sounds to ME like the numbers aren't really all that accurate in terms of MILLIONS differentiation at times.

if it fluctuates that greatly sometimes...how do you really KNOW if it was a two log drop or a ten log drop really?

[jmjm530]

First of all your calculation in a previous post was wrong. A one log variance using 568,000 IU/ml is 5,580,000 not 50,680,000. But since the variance is for both directions, in reality the variance would be closer to 2,800,000 from 568,000.

Second, Shiffman isn't saying that the viral load fluctuates that wildly, what he's saying is that the tests being are not sensitive enough to make clinical decisions based on readings of less than one log. Again, this is consistent with all of the clinical studies that use log drop to determine treatment protocols.

And that's why when you went to see Dr. J, he said he wouldn't have treated you further if you weren't non-detectible by week 24 and why he extended your treatment because you weren't non-detectible by week 12.

I don't have a print out of the slide set, but I believe there's a caption on the page that says something like "Think in terms of logs, not VL in IU/ml". Again, this is what all the studies seem to suggest, so I don't see anything new or contradictory here, but rather among some other things an explanation of why small (under 1 log) decreases or increases in viral load may not be as signifcant as some think.

-- Jim

[GrandOak]

Gee thanks.  That means my base VL of 72 million was could be anywhere from 7.2 to 720 million.  Can I buy a vowel and go for the lower value? ;-P

[jmjm530]

GO: Gee thanks. That means my base VL of 72 million was could be anywhere from 7.2 to 720 million.
-------------------------------------

According to my take of Shiffman, probably no clinical difference between the two readings you mentioned. Can you think of any? After all,  both examples are considered "high" and the "two-log" rule is considered outdated by many with "non-detectible" being substituted. BTW I'm not making this stuff up, just passing along information from a well-respected hepatologist --  so please don't shoot the messenger, especially since this isn't even bad news. LOL.