Wow, the fact the study showed at 30 days some ("non-statistically significant") death difference (2.1% vs 4.0%; p= 0.07), as well as reinfarction difference (0.8% vs 1.9%; p=0.11), with about 50 and 100 NNT, respectively; with only 1,000+ randomized patients; all that tells me that, maybe, with 2,000+ randomized "higher-risk" patients (I.e. diabetic, smokers, prior MI, etc.), the study could have met statistical significance for such a strong hard outcome with a good effect size (50 NNT to prevent death, incredible). Only 12% patients had diabetis, and 10% had prior MI, what if those patients had represented >50% the study population? In a 2,000+ sample size?
We have to remember that the P-value only tells us the % of the result being out of chance. I think 7% chance is low, which could be even lower with a higher power. What about the confidence intervals, how far away from being "narrow enough" where they? Let us think about it.
Maybe doing thrombus aspiration for STEMI in high-risk patients could in fact save more lives than we think. Who knows.
These are important points, Thank you Isaac. However, results of smaller trials should always be validated in larger well-powered trials as results are not always reproducible. Next, we will post the review of TASTE which asked similar question to TAPAS but was more than 6 times larger.
I just realized there was in fact statistical significance for cardiac death difference at 1 year. About 32 NNT! There were two versions, 30-day outcomes and 1-year outcomes. Different 1st authors, and different journals (NEJM vs Lancet).
- Citation: Vlaar PJ, et al. Cardiac death and reinfarction after 1 year in the Thrombus Aspiration during Percutaneous coronary intervention in Acute myocardial infarction Study (TAPAS): a 1-year follow-up study. Lancet. 2008 Jun 7;371(9628):1915-20.
Cardiac death at 1 year was 3.6% (19 of 535 patients) in the thrombus aspiration group and 6.7% (36 of 536) in the conventional PCI group (hazard ratio [HR] 1.93; 95% CI 1.11-3.37; p=0.020). 1-year cardiac death or non-fatal reinfarction occurred in 5.6% (30 of 535) of patients in the thrombus aspiration group and 9.9% (53 of 536) of patients in the conventional PCI group (HR 1.81; 95% CI 1.16-2.84; p=0.009).
32 NNT for mortality! That is huuuuge! Maybe longer duration appears to have increased the power since there was some mild signal at 30 days? Maybe aspiration does save some myocardium? Am I missing something here?
Maybe because TASTE had 24 hr tolerance so patients benefited less (vs 12 hrs in TAPA)? Or maybe because they had lower patients with smoking risk factor? Any other differences in patients baseline characteristics that could have changed the "risk spectrum" thereby increasing benefits in a higher-risk population?
There is a lot of learning. I love the discussion!
Great points, Isaac. Keep in mind that cardiovascular mortality is subject to bias specially when studies are not blinded.
Another important point to keep in mind, is that doing more analysis with different follow up times and different endpoints will increase the chance of type I error.
TAPAS was a good study but since it wasn't powered for clinical outcomes, we believe such studies should be followed by larger trials to validate their findings before adopting them into routine clinical practice.
We posted our review of TASTE (much larger trial) which was negative even at 1 year.
TASTE and TOTAL combined had almost 18,000 patients and both were negative. Both asked similar question to TAPAS (had 1,071 patients) but they were powered for clinical outcomes. This shows the importance of having adequately powered studies before changing practice.
Oh, That is right. Now things make much more sense. Thank you for taking the time to provide all these clarifications. I truly appreciate the teaching!
Haha giving an amuse bouche for TASTE. Cute 😉
Wow, the fact the study showed at 30 days some ("non-statistically significant") death difference (2.1% vs 4.0%; p= 0.07), as well as reinfarction difference (0.8% vs 1.9%; p=0.11), with about 50 and 100 NNT, respectively; with only 1,000+ randomized patients; all that tells me that, maybe, with 2,000+ randomized "higher-risk" patients (I.e. diabetic, smokers, prior MI, etc.), the study could have met statistical significance for such a strong hard outcome with a good effect size (50 NNT to prevent death, incredible). Only 12% patients had diabetis, and 10% had prior MI, what if those patients had represented >50% the study population? In a 2,000+ sample size?
We have to remember that the P-value only tells us the % of the result being out of chance. I think 7% chance is low, which could be even lower with a higher power. What about the confidence intervals, how far away from being "narrow enough" where they? Let us think about it.
Maybe doing thrombus aspiration for STEMI in high-risk patients could in fact save more lives than we think. Who knows.
These are important points, Thank you Isaac. However, results of smaller trials should always be validated in larger well-powered trials as results are not always reproducible. Next, we will post the review of TASTE which asked similar question to TAPAS but was more than 6 times larger.
I just realized there was in fact statistical significance for cardiac death difference at 1 year. About 32 NNT! There were two versions, 30-day outcomes and 1-year outcomes. Different 1st authors, and different journals (NEJM vs Lancet).
- Citation: Vlaar PJ, et al. Cardiac death and reinfarction after 1 year in the Thrombus Aspiration during Percutaneous coronary intervention in Acute myocardial infarction Study (TAPAS): a 1-year follow-up study. Lancet. 2008 Jun 7;371(9628):1915-20.
Cardiac death at 1 year was 3.6% (19 of 535 patients) in the thrombus aspiration group and 6.7% (36 of 536) in the conventional PCI group (hazard ratio [HR] 1.93; 95% CI 1.11-3.37; p=0.020). 1-year cardiac death or non-fatal reinfarction occurred in 5.6% (30 of 535) of patients in the thrombus aspiration group and 9.9% (53 of 536) of patients in the conventional PCI group (HR 1.81; 95% CI 1.16-2.84; p=0.009).
32 NNT for mortality! That is huuuuge! Maybe longer duration appears to have increased the power since there was some mild signal at 30 days? Maybe aspiration does save some myocardium? Am I missing something here?
- Link: https://pubmed.ncbi.nlm.nih.gov/18539223/
Maybe because TASTE had 24 hr tolerance so patients benefited less (vs 12 hrs in TAPA)? Or maybe because they had lower patients with smoking risk factor? Any other differences in patients baseline characteristics that could have changed the "risk spectrum" thereby increasing benefits in a higher-risk population?
There is a lot of learning. I love the discussion!
Great points, Isaac. Keep in mind that cardiovascular mortality is subject to bias specially when studies are not blinded.
Another important point to keep in mind, is that doing more analysis with different follow up times and different endpoints will increase the chance of type I error.
TAPAS was a good study but since it wasn't powered for clinical outcomes, we believe such studies should be followed by larger trials to validate their findings before adopting them into routine clinical practice.
We posted our review of TASTE (much larger trial) which was negative even at 1 year.
Another trial you can refer to is TOTAL which had over 10,000 patients and was also negative. https://www.nejm.org/doi/full/10.1056/NEJMoa1415098
TASTE and TOTAL combined had almost 18,000 patients and both were negative. Both asked similar question to TAPAS (had 1,071 patients) but they were powered for clinical outcomes. This shows the importance of having adequately powered studies before changing practice.
Oh, That is right. Now things make much more sense. Thank you for taking the time to provide all these clarifications. I truly appreciate the teaching!