Journal Club Highlights 2.8.18 (Cardiology) – CULPRIT-SHOCK

Screen Shot 2018-02-09 at 2.26.05 PM

We had another great Journal Club with Dr. Anurag Mehta presenting the CULPRIT-SHOCK trial comparing culprit vessel-only vs. immediate multi-vessel PCI in patients with acute myocardial infarction and cardiogenic shock.

Thank you to our faculty discussants: Dr. Khera (CUH), Dr. Das (Parkland), and Dr. Abdullah (VA).

See below for discussions from all 3 sites:


Q: Why was this study important to conduct? What is the background for this study?

  • CUH: 
    • Choosing Wisely – 5 Cardiology recommendations, including only doing PCI of culprit lesion in setting of AMI, BUT, this was not based on clear data, just retrospective
    • CULPRIT (not this trial CULPRIT-SHOCK) came out and patients did better with PCI of multivessel disease
    • There was no prior data in cardiogenic shock patients.
  • Parkland: 
    • Back in 2011, it used to be in the guidelines to ONLY intervene on the culprit lesion.
    • Then, PRAMI trial came out and made the guidelines.
    • This trial was influential and makes people more likely to intervene on culprit lesion only over immediate multivessel PCI
  • VA:
    • Cardiogenic shock complicating acute MI has HIGH mortality.
    • There were trials looking into this question in hemodynamically stable patients, but non looking at patients with cardiogenic shock.
    • It has been a class IIb recommendation – You can consider PCI of non-culprit vessels; however, the timing of these non-culprit vessel interventions hadn’t been established yet.
    • So, this was an important question to investigate.

Q: What was the population studied and which interventions were compared?

  • CUH: 
    • Population: acute MI + shock
    • Question: culprit-lesion only of multivessel interventions

Study design:

  • CUH: 
    • European, randomized open-label controlled trial
      • Randomization is meant to evenly distribute confounders. Did it work? Yes – Look at Table 1.
      • Open label? Impossible to blind this. Does introduce risk of bias because the investigator knows which outcome. However, in this case, cannot bias endpoint of death or renal replacement therapy (RRT). And reviewing the treatments received – they were pretty similar for medical management (statin, anti-platelets, etc.)
    • Inclusion criteria: Shock + acute myocardial infarction
      • How do you know if they enrolled what they said?
        • 1) Look at Table 1 criterial and clinical features
        • 2) Mortality consistent with prior studies in this case, so probably good
    • Exclusion criteria: looks reasonable, not stacking the deck and excluding certain patients
    • Intervention: 1-vessel vs. multi-vessel
    • Endpoints: composite of death + renal replacement therapy (RRT)
      • Death – definitely makes sense as an endpoint
      • RRT? – Contrast-induced nephropathy as a penalty. Is this a meaningful outcome? Short-term, possibly not, but long term HD needs unknown?
      • Somewhat a softer renal endpoint, but this will penalize the multivessel group more and is more conservative
      • Composites in cardiology often include positive and negative outcomes for net effect (i.e., recurrent acute MI + bleeding with antiplatelet agents)

Q: How do you know if the trial interventions are ethical?

  • CUH:
    • Is there a known answer? – there is not here
    • Your power and sample size need to be adequate. Set up study properly, or this is unethical to do because it won’t answer the question.
    • Need to do interim analysis during trials to ensure that you are not witholding something lifesaving from one group or causing high levels of death.

Comments on the statistics and power statement of the trial:

  • CUH:
    • Power statement for sample size determination – 38% PCI only, 50% multivessel:
      • Where are these numbers from? Go to the References.
      • Is this what they got? Pretty close to actual outcomes, but a little optimistic on the 12% ARR here with 25% RRR
      • Round up because people will drop out
    • When is power most important?
      • Null trials with negative results because then you need to scrutinize whether no difference was found because underpowered or really no difference exists.

Table 1 (Baseline Characteristics):

  • CUH:
    • Important for randomization, identifying confounders, and deciding whether a trial is generalizable to your patients.
    • Did this apply to our patients?
      • The study population (table 1) is similar to populations here and had similar treatment modalities for acute MR (i.e., rare thrombolysis), but less racial variability than here

Table 2 (Procedural Characteristics): 

  • CUH:
    • Notable for significant crossover. In general, crossover makes the difference smaller and biases toward the null. Since this study was positive, less important here.

Table 3 (Clinical Outcomes at 30 Days):

  • CUH:
    • Primary end point – actually are slightly off from estimates used to calculate sample size (55% and 45% vs. 50% and 38%), but still a significant decrease; However, ARR only ~10% and RRR ~20%.

Q: What are your thoughts on the Kaplan-Meier Curves and the separation pattern of the curves?

  • CUH:
    • The curves separate over the first 5 days for composite and renal failure, then have stable differences.
    • Death starts separating at 5-10 days and probably delayed from bleeding and renal failure (don’t seem to be dying on the table or immediately causing improvement in HF, which would cause crossover).
  • VA:
    • The early separation among groups suggest primary outcomes were happening within days of the procedure (rather than months, etc).
    • It seems like whatever negative effects of multivessel PCI have were happening within ~5-10 days of the procedure.

Q: What are the potential risks/benefits of immediate multi-vessel PCI?

  • VA:
    • Risks: higher contrast load, longer procedure time, more catheter manipulation, reperfusion injury, CTO procedures are higher risk
    • Benefits: theoretical benefit of decreased ischemia with multivessel revascularization

Q: What is the significance of crossover in studies and what were possible causes of crossover in this study?

  • VA:
    • Reasons for culprit –> MV crossover:
      • Sometimes it is difficult to just PCI one vessel. For example, if the lesion is located at a bifurcation with plaque extending into an adjacent vessel. Also, sometimes plaque can extend into other vessels during PCI of a single vessel.
    • Reasons for MV –> culprit only crossover:
      • One reason would be reaching the contrast limit, which was 300 cc, which is not too much, especially if you are dealing with more complicated lesions and CTOs.
    • What is the significance of cross over in studies?
      • Crossover makes it more likely that you will accept the null hypothesis because you get mixing of the two intervention groups (less likely to see a difference between the two arms).
  • Parkland:
    • If you have crossover, you increase the risk of trending toward the null and pushing towards a type 2 error. So, if you have crossover here and find a difference, this makes the findings more likely to be right.
    • If you have crossover in a non-inferiority trial, makes it more likely to be positive

Q: What is the difference between intention-to-treat, as-treated, and per-protocol analysis?

  • VA:
    • Intention to treat: Analyze people based on the group they were randomized to.
      • Benefit is that you maintain the randomization and reduce the bias/confounders
    • Per protocol: Usually the per protocol is less people because you have eliminated those that deviated from protocol (e.g., cross over)
    • As treated: Analyze people based on the intervention they received
    • All 3 analyses have their pros and cons – Usually for RCTs, most use intention to treat analysis, but their is additional information you can gain from looking at the per protocol and as treated analyses. In this trial, all 3 analyses were consistent, further strengthening their findings.
  • Parkland:
    • In inferiority trials, we do more per protocol.

Q: How might the management of chronic total occlusions (CTOs) in this study have influenced the results/interpretation?

  • VA: 
    • It’s good that they included CTOs in this study because it makes the study more generalizable to our real world populations.
    • However, CTOs are more complicated, longer procedures, with higher contrast load, etc.
  • Parkland:
    • This is a very interesting area. An artery that has been occluded for a while. Classic teaching is to leave this alone. There are a subset of people with refractory angina that CTO opening may be helpful. It depends on who you ask.
    • The general consensus is to only intervene on CTO for refractory angina
    • Opening CTOs in this trial may trend toward more harms in the multivessel disease group.

Q: What are differences between observational and randomized trials?

  • Parkland:
    • Observational provides an association but not causality.
    • RCTs CAN show causation.
    • The biggest issue in an observational study is “treatment indication bias”
      • Example: CABG vs. non-CABG observational study is skewed towards CABG because those that the surgeons want to perform CABG on may be “more healthy” than those surgeons do not want to intervene on.

Q: How has this study affected practice?

  • Parkland:
    • Back in 2011, it used to be in the guidelines to ONLY intervene on the culprit lesion.
    • Then, PRAMI trial came out and made the guidelines.
    • This trial was influential and makes people more likely to intervene on culprit lesion only over immediate multivessel PCI

Q: How did people die in this study?

  • VA: 
    • Neurologic deaths were higher in the multivessel PCI group
      • Maybe more manipulation of the catheters led to more strokes?
      • There was also more use of ECMO in the multivessel group – some studies report more neurological outcomes with use of ECMO.
      • The imbalance in use of ECMO also suggests the patients in the MV might have been sicker.



Cox’s Conference: AIDS-related Kaposi’s Sarcoma

In this week’s Cox’s Conference, Dr. Jennifer Wagner presented a case of AIDS-related Kaposi’s sarcoma to expert discussant, Dr. Bonnie Prokesch. Here, we will summarize the case presentation along with clinical pearls involving treatment and complications. Continue reading Cox’s Conference: AIDS-related Kaposi’s Sarcoma