Sinai EM Journal Club

This month’s journal club presentation began with what I believe was a discussion of blood clots in Cro-Mags before touching upon late 19th century versions of CPR, the landmark closed-chest cardiac massage paper, and eventually, a comparison of ROSC (return of spontaneous circulation) in real patients vs. as seen on television. Chad then led the group in a discussion of a new meta-analysis by Xin Li et al appearing in a recent issue of Resuscitation (2006: Vol 70, pp31-36) on the topic of CPR with and without thrombolytics.

Administering tPA to patients in cardiac arrest seems like a good idea — etiology is often (81%) PE or MI, and outcomes are so dismal (depending on city, appx 2-10% survive when the arrest happens outside the hospital, maybe 15% - 25% survive in-hopsital arrests) that it seems like it couldn’t hurt. So this paper tried to look at the evidence for this intervention, from 1966-2004, by doing what appears to be a cursory medline search of English-language papers. They found 121 papers, 9 of which met their criteria for cohort studies of cardiac arrested adults treated with CPR plus thrombolytics, with non-thrombolyzed controls. One paper lacked proper research indices and was excluded. That left us with seven retrospective studies, the largest of which was Lederer et al. (Resuscitation 2001; 50(1):71-76)  and one prospective study by Bottiger (Lancet 2001; 357:1583-5). All of these studies featured 100 mg of tPA administered during or within 15 minutes of CPR. All received heparin as well with some also receiving aspirin.

The authors then proceeded to compare outcomes — rate of ROSC +/- thrombolytics, 24 hour survival, survival to discharge, long-term neuro function, and morbitity from severe bleeding. Rate of ROSC was measured in four papers; one showed no statistical benefit but the meta-odds ratio for achieving return of spontaneous circulation with thrombolytics added to CPR was 2.57 (a significant different, with 95% CI from 1.76 to 3.74). The authors used some additional math to say this conclusion would require 26 negative articles to refute it (I think one good prospective study would do the trick, but hey, I’m just a blogger).

 Twenty-four hour survival rate was similarly superior among CPR+tPA patients compared to CPR alone; the odds ratio derived from three articles (two of which showed statistical difference) was 2.08 (95% CI 1.4-3.08).

Survival to discharge was assessed in four papers, three of which showed thrombolysis did nothing. And yet, the odds ratio these authors derived from the four papers was 2.0 (95% CI 1.23-3.27).

In two papers, neuro function was assessed on a CPC scale out to 6 months after ROSC — both papers showed benefit from thrombolytics and thus, the authors calculated an OR of 2.59 (95%CI of 1.70-3.95).

Finally, the risk of severe bleeding (bleeding into body cavities, intraparenchymal hematoma, ICH, or anything requiring transfusion) was assessed in 7 of the 8 papers included for meta-analysis. None individually showed significant risk of tPA (the odds ratios all crossed 1.0) but taken together the risk of major blood was 2.2 (95%CI 1.25-3.88). The authors note, however, the conclusion is vulnerable to publication bias.

Lots of problems with this analysis — from its limited search of the literature to its overreliance on Lederer’s trial (which reviewed 108 thrombolyzed arresters and 216 nonthrombolyzed, and was by far the biggest trial with most assessed endpoints, so it kept tilting the meta-analysis its way).

One might wish for some randomized, double-blind, placebo-controlled multicenter trial to give tPA on witnessed-arrest patients on CPR within 10 minutes in a standardized way, say, after the first vasopressor. One might further wish for endpoints such as 24h survival, 30-day survival, admission to hospital, ROSC, neuro performance, and bleeding complications from tenecteplase. In fact, one is wishing for TROICA, Bottinger’s latest project. The methods were described in the European Journal of Clinical Investigation (2005) 35, 315-323. Our spies in Europe report some good news from the yet-unpublished data — tPA doesn’t cause significantly higher bleeding! Yay.

On the downside, the TROICA study doesn’t seem to demonstrate benefit from adding tPA to witness-arrest CPR. But don’t despair — the European docs who staff their mobile-MICU ambulances were instructed to not enroll patients with likely PE, since they tend to give tPA over there very freely for clotbusting, and don’t want to deprive patients of any possible benefit. Thus, maybe their cohort of patients was skewed towards the primary arrhythmia bunch, an unrepresentative sampling of all-comers for arrest.

We’ll see. And we’ll take the good news that tPA doesn’t seem to hurt this group much (at least, not much more than they’re already hurt), when the paper comes out. In the meantime, using tPA on arrested patients will remain a difficult sell — only if you have a patient with known etiology of PE would I consider it (for MI, our times to cath lab are good enough to probably preclude tPA).