Primary versus secondary PCI typically distinguishes episodes of PCI treatment that are emergent or elective, respectively. How researchers operationally define Primary PCI can impact how studies of variation of PCI quality of care, usually measured by its timeliness, impacts patient outcomes, like short term mortality. Operational definitions like "Primary PCI = Non-Transferred cases for PCI" fails to account for true ischemic time, the time since pain onset, which can blur acute episodes from resolved, non-acute episodes. The blurring of acute from non-acute episodes consequently obscures distinctions across acute versus non-acute patient outcomes, like short term mortality. The recent publication of letters criticizing a NEJM article claiming little survival benefit to AMI patients from reduced door to balloon time for "Primary PCI" highlights the importance of this definition, and none offers anything definitive, or anything better than "primary = non-transferred", as was used in the original research. I think some measure of ischemic time since pain onset would be much better, and it was available from the Cath PCI registry that was the data source for the original article. If they had used the available element to calculate ischemic time to PCI, what threshold would be a reasonable proxy for the time frame within which primary treatment might occur: 2 hours since pain onset; 6 hours, or 12 hours? What's reasonable and why? What are the clinical issues?
Data Letter to NEJM on Timely treatment for AMI with PCI
Hi Edwin,
While it is not stated clearly, I believe Menees et. al. used PCI indication (seq # 7035) to define Primary PCI ("Immediate PCI for STEMI") which corresponds to less than 12 hours from symptom onset with no intervening thrombolytic therapy, then used Admit source (#3010) to exclude transferred patients, and PCI Status (#7020) to exclude non-emergency cases. The use of PCI indication to identify primary PCI is consistent with the NCDR/ACC appropriateness classification code I believe implemented by the same group at Yale. https://www.ncdr.com/webncdr/cathpci/home/auc
I think you and Bates et. al. make an excellent point about the need to consider total ischemic time (in terms of associations with adverse outcomes) of which door-to-balloon time is not necessarily the largest component. Bates does a very nice job with the figure illustrating the patient,transport, and D2B components on a timeline.
Door to Balloon is however probably the most easily addressable component from the perspective of hospital QI initiatives, and as Menees states has been agressively targeted for reduction over the last few years with substantial improvements observed.
Primary PCI is performed on patients with acute coronary syndrome such as STEMI and NSTEMI. The PCI should be done ideally in less than 90 minutes (door-to-balloon) meaning the time from admission to PCI. 'Ad hoc PCI' is when an elective, urgent or semi urgent PCI is performed in IHD patients who are not infarcting at the time but is decided to proceed to a PCI during the diagnostic angiography procedure without waiting for further assessment by the heart team or other medical experts.
The issue I feel most strongly about is the sufficient clarity of operational definitions that would enable replication of important work, and as you might surmise from my view of what was provided in the original article, there was not sufficient detail to enable that. For that issue, I fault the reviewers first, then the editorial leadership, and lastly the authors.
Additionally, the sensitivity of the research methods to assess differential impact of PCI timing on mortality outcomes is weaker than the question deserves, where criteria that would use ischemic proxy times from the registry would have revealed more of the impact of timely PCI if it were tested against a more homogeneous cohort with ischemic time all less than 6 hours, or even 2 hours. That Ischemic time from onset can be calculated, which by the way, I feel the author's mistakenly claimed in their correspondence response, that that variable was not available. I can understand that they may not have had it due to case selection variables that may have been alternatively used for building the data set. ischemic time can be taken from the pain onset time and duration variables in the registry. It is the responsibility of the reviewers to insure that the method section is minimally sufficient to enable replication, and they missed the mark this time. Thank you very much for your comments..
- Badges
- Science topic
- Similar topics
- Medicine
- Epidemiology