SEARCH

SEARCH BY CITATION

The prostate cancer trifecta neatly encapsulates the idea that radical prostatectomy outcomes are multidimensional and competing. What we want for our patients is not just cure of cancer but recovery of urinary and erectile function; attempts to improve cure rates, by wider resection, can compromise functional recovery just as minimizing the risk of impotence or incontinence, by a more conservative resection, may increase the risk of recurrence.

The problem with the trifecta comes when it is used as a statistic, with authors calculating a specific trifecta rate. For example, Shikanov et al. [1] reported a trifecta rate of 76% at 2 years in a case series of 380 patients. Indeed, a recent systematic review reported on 12 separate attempts to calculate a trifecta rate [2], including, perhaps predictably, a trifecta nomogram [3], as well as an extension of the concept to a ‘pentafecta’, defined as the trifecta plus negative surgical margins and no surgical complications [4].

Despite its intuitive appeal, the trifecta is a poor measure of radical prostatectomy outcomes. There are several reasons why the trifecta should not be used as an endpoint in prostate cancer research. These are discussed below.

1. THE TRIFECTA IS STATISTICALLY INVALID

  1. Top of page
  2. 1. THE TRIFECTA IS STATISTICALLY INVALID
  3. 2. THE TRIFECTA IS MEANT TO INCLUDE THREE ENDPOINTS, BUT IS DRIVEN PREDOMINATELY BY ERECTILE OUTCOMES
  4. 3. THE TRIFECTA IS OVERLY INFLUENCED BY SHORT-TERM OUTCOMES
  5. 4. THE TRIFECTA GIVES EQUAL WEIGHTING TO EACH ENDPOINT
  6. CONFLICT OF INTEREST
  7. REFERENCES

Recurrence and functional recovery occur over time; in statistical terms, they are ‘time-to-event’ variables. There are two different types of time-to-event variable, failure and incidence, and it is not straightforward to combine them; yet this is exactly what is required by a trifecta analysis. A patient who recovers erectile and urinary function at 1 year, but recurs at 2 years, would be non-trifecta at 6 months, then go into trifecta at 12 months but then lose trifecta at 24 months. Exacerbating this problem is the fact that patients can lose and regain function repeatedly over time. For example, if we were to use a score of 24 on the International Index of Erectile Function-6 as the criterion for potency, one patient might have consecutive scores of 22, 25, 26, 23, 24, 23; if the patient were continent and recurrence-free, these scores would translate as non-trifecta, trifecta, trifecta, non-trifecta, trifecta, non-trifecta. By contrast, once a patient has recurred he is always considered to have recurred. So the trifecta mixes incidence and failure variables, with the former but not the latter being reversible.

2. THE TRIFECTA IS MEANT TO INCLUDE THREE ENDPOINTS, BUT IS DRIVEN PREDOMINATELY BY ERECTILE OUTCOMES

  1. Top of page
  2. 1. THE TRIFECTA IS STATISTICALLY INVALID
  3. 2. THE TRIFECTA IS MEANT TO INCLUDE THREE ENDPOINTS, BUT IS DRIVEN PREDOMINATELY BY ERECTILE OUTCOMES
  4. 3. THE TRIFECTA IS OVERLY INFLUENCED BY SHORT-TERM OUTCOMES
  5. 4. THE TRIFECTA GIVES EQUAL WEIGHTING TO EACH ENDPOINT
  6. CONFLICT OF INTEREST
  7. REFERENCES

More men recover urinary as compared with erectile function. In particular, whilst most patients who experience persistent incontinence are also impotent (∼90% in an informal analysis of data from the Memorial Sloan-Kettering Cancer Center), only a small proportion of patients who regain erectile function report pad use (∼10%). As a result, urinary outcomes have little effect on trifecta rates. It can also be shown that recurrence also has a smaller effect on trifecta than erectile dysfunction because it is less common. As a simple illustration, imagine that surgeon A has a long-term potency rate of 40% with 90% cancer control; surgeon B has slightly better potency (50%) but much poorer rates of cancer control (72%). Although these are very different outcomes, trifecta rates are identical: 40% × 90% = 50% × 72% = 36%.

3. THE TRIFECTA IS OVERLY INFLUENCED BY SHORT-TERM OUTCOMES

  1. Top of page
  2. 1. THE TRIFECTA IS STATISTICALLY INVALID
  3. 2. THE TRIFECTA IS MEANT TO INCLUDE THREE ENDPOINTS, BUT IS DRIVEN PREDOMINATELY BY ERECTILE OUTCOMES
  4. 3. THE TRIFECTA IS OVERLY INFLUENCED BY SHORT-TERM OUTCOMES
  5. 4. THE TRIFECTA GIVES EQUAL WEIGHTING TO EACH ENDPOINT
  6. CONFLICT OF INTEREST
  7. REFERENCES

Recovery of urinary and erectile function typically occurs in the first year or two after surgery; recurrence occurs over 10 years or more. Any surgeon still in practice reporting trifecta results will inevitably have more patients followed for 2 years than for 10, which means that trifecta results are driven more strongly by functional recovery than cancer recurrence. Surgeon A and B above have the same long-term trifecta rates; however, because surgeon B has better functional recovery rates, he will be able to report better trifecta outcomes.

4. THE TRIFECTA GIVES EQUAL WEIGHTING TO EACH ENDPOINT

  1. Top of page
  2. 1. THE TRIFECTA IS STATISTICALLY INVALID
  3. 2. THE TRIFECTA IS MEANT TO INCLUDE THREE ENDPOINTS, BUT IS DRIVEN PREDOMINATELY BY ERECTILE OUTCOMES
  4. 3. THE TRIFECTA IS OVERLY INFLUENCED BY SHORT-TERM OUTCOMES
  5. 4. THE TRIFECTA GIVES EQUAL WEIGHTING TO EACH ENDPOINT
  6. CONFLICT OF INTEREST
  7. REFERENCES

Trifecta outcomes would be the same for a surgeon with 50% potency rates and 90% freedom from recurrence as for a surgeon with rates of 90% for potency but 50% for recurrence. Patients might well trade-off an increase in recurrence rates for a decrease in the risk of impotence, but only if the latter was much larger than the former. Yes, patients do want to be recurrence-free and to recover erectile and urinary function; no, they do not consider each endpoint to be equally important.

In conclusion, the trifecta neatly encapsulates the goal of radical prostatectomy, but it is statistically invalid as an endpoint in clinical research and should not be reported as a measure of surgical outcome.

REFERENCES

  1. Top of page
  2. 1. THE TRIFECTA IS STATISTICALLY INVALID
  3. 2. THE TRIFECTA IS MEANT TO INCLUDE THREE ENDPOINTS, BUT IS DRIVEN PREDOMINATELY BY ERECTILE OUTCOMES
  4. 3. THE TRIFECTA IS OVERLY INFLUENCED BY SHORT-TERM OUTCOMES
  5. 4. THE TRIFECTA GIVES EQUAL WEIGHTING TO EACH ENDPOINT
  6. CONFLICT OF INTEREST
  7. REFERENCES