A look back at PIVOT … with additional data for context

Five years ago now, in May 2011, Wilt first presented the results of the PIVOT trial at the annual meeting of the American Urological Association in Washington, DC. A bit over a year later, the final results were published in the New England Journal of Medicine.

Recently, in European Urology, Barbosa et al. have published a new paper questioning whether it is appropriate to generalize the results of the PIVOT trial to all men being treated for prostate cancer within the Veterans Administration system here in the USA (and indeed to all men getting treated in the USA for localized prostate cancer).

So first let’s go back in time and make sure we remember what it was that Wilt and his colleagues actually showed from the original PIVOT trial, which was three basic facts:

  • “Surgery did not reduce mortality more than observation in men with low PSA or low-risk prostate cancer.”
  • The trial results “suggest a benefit from surgery in men with higher PSA or higher risk disease.”
  • Compared to observation, surgery offered “reductions in all-cause and prostate cancer mortality that were not significant and less than 3 percent in absolute terms over 12 years.”

In other words, over a median 10-year follow-up period in a cohort of men among whom 48 percent died of all causes during the trial, there was a very limited survival benefit for radical prostatectomy compared to observation. And it is worth remembering that the average age of the men in the PIVOT trial was 67 years.

So what have Barbosa et al. now added to the equation?

Barbosa et al. set out to examine the patterns of overall and prostate cancer-specific survival of all men diagnosed with prostate cancer and initially treated within the VA health system between 1995 and 2001, which corresponds roughly to the same period in which men were being enrolled into the PIVOT trial.

To do this they compiled an analytic cohort of 35,954 patients, all diagnosed with clinically localized prostate cancer at the VA within the relevant time period and who met similar enrollment criteria to those treated in the PIVOT trial (i.e.,  an age at diagnosis ≤ 75 years and a clinical stage of T2 or lower).

Here is what they found for the men in their analytic cohort:

  • Their average (mean) age was 65.9 years (1.1 years younger than the PIVOT patients).
  • The average (median) follow-up was 161 months (41 months longer than the PIVOT patients).
  • Of the total patient cohort,
    • 22.5 percent of patients were treated with surgery.
    • 16.6 percent were treated with radiotherapy.
    • 23.1 percent were treated with androgen deprivation therapy (ADT).
    • Which implies that 37.8 percent of patients were simply observed in some manner.
  • Average (median) survival  was
    • 14 years for the entire cohort
    • 17.9 years for the men initially treated by surgery
    • 12.9 years for the men initially treated by radiation therapy
  • Roughly a third of the patients died of all causes within 10 years of diagnosis (as compared to 48 percent of men in the PIVOT trial).

Barbosa et al. politely state that

This finding sounds a note of caution when generalizing the mortality data from PIVOT to VHA patients and those in the community.

What they really mean by that is that their data suggest that men treated surgically in their cohort had a significantly longer survival (by nearly 6 years) than the men treated surgically in the PIVOT trial.

It is always difficult to interpret data like these. There is clearly an editorial commentary on the Barbosa et al. paper (by Cole et al.) in the same issue of European Urology, but the text of that editorial commentary is only available to subscribers (or those with a spare $35.95).

The original PIVOT study was a randomized clinical trial … but it enrolled far fewer men than had originally had been intended. Conversely, the Barbosa data set is a retrospective analysis based on men whose treatment reflected the choices of the patients in concert with their physicians. The men getting first-line radiation therapy were almost certainly older than  those given first-line surgery, and the men getting first-line androgen deprivation would largely have had metastatic disease at the time of diagnosis, so we aren’t comparing apples to apples. It is also easy to argue that the men who agreed to be randomized in the PIVOT trial may well have been less than excelent candidates for surgery — because the ones who thought surgery would cure them would probably never have agreed to be randomized to the PIVOT trial in the first place!

6 Responses

  1. A bit off topic, but I’ve often wondered how much of a bump up in co-morbidity survival occurs after a diagnosis of cancer. Hard to hide high blood pressure when you go from a visit every several years to several times a year.

  2. So if we are comparing apples to oranges, do I just take the Balbosa data at face value without comparing it to PIVOT?

  3. Maria:

    No. You can’t do that. What the Balbosa data tell us is that you have to be careful about taking the PIVOT data at face value. … But for different reasons you need to be careful about taking the Balbosa data at face value too. The truth — if there is a “truth” — is liable to be somewhere in between.

  4. Well what the heck am I comparing? Just that they found greater success from RP in basically the same cohort?

  5. But bottom line — I can take from these data that surgery is more successful for survival than radiotherapy? (I swear I thought I posted a response hours ago!)

  6. Maria:

    (1) In answer to your first comment, the answer is yes … All that these data tell you are that the retrospective analysis shows a more positive benefit for surgery than the randomized PIVOT study did in a similar cohort of patients. Whether that result is “real” or not is unknowable.

    (2) In answer to your second comment, no, you can not take from these data that surgery is more successful for survival that radiation therapy. There have never been data from a large, well-structured, randomized trial that compared the outcomes of radiation therapy and surgery in a well-defined cohort of men … but hopefully there will be soon when we get the data from the ProtecT study ongoing in the UK. In a non-randomized cohort like this there is an inevitable selection bias regarding which patients got surgery and which patients got radiation therapy, and it is highly likely that during the timeframe referred to the patients who got radiation therapy were older, and the type of radiation therapy being used was of far lower quality that they type of radiation therapy being used today.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: