N Engl J Med. 2015 Jul 30;373(5):405-14. doi: 10.1056/NEJMoa1501969.
Therapeutic Hypothermia in Deceased Organ Donors and Kidney-Graft Function.
Free full text available at NEJM website
There is a good discussion, hosted on Medstro (you need an account to read all the discussion - but can log in via Twitter or Google etc), available at this link, in which a couple of the authors participate.
Apart from the paper itself, also check out the complete trial protocol, which has crucial details about the inclusion criteria, cooling methods, changes made in the protocol after trial began, the actual interim analysis etc
Tuesday Sep 22 2015 at 9 pm Eastern and 6 pm Pacific
Wednesday 8 pm BST, 12 noon Pacific and 3 pm Eastern
Storified versions of the Chats
Delayed graft function (DGF) - defined as need for dialysis within 7 days after a kidney transplant, has been well established as an important outcome - it is associated with worse long term graft outcomes. Among neurologically dead donors (ie not donation after cardiac death), the rates of DGF have been described to be up to 50%. Pretty high. The authors tried a novel intervention - therapeutic hypothermia of the donor - much before the transplant surgery occurs. Why would hypothermia work? One of the main mechanisms of injury for the transplant organ is ischemia-reperfusion. Many of the downstream consequences (activation of inflammatory cascade, apoptosis, free radical damage) are indeed temperature dependent. Ergo, hypothermia. This review, by Polderman, gives a nice overview of the pros and cos of therapeutic hypothermia.
RCT, where donors (note: not recipients) were randomized to standard (36.5 to 37.5 C) versus hypothermia (34 to 35 C) protocol. Donors were stratified according to organ procurement organization, donor status (standard versus expanded criteria) and whether donor had already received hypothermic treatment before declaration of death per neurologic criteria. Study was conducted on the West Coast, and funded by the Health Resources and Services Administration.
Exclusion criteria initially were:
- Subjects under the age of 18;
- body weight less than 50 kg;
- donation after cardiac death (DCD), as it does not fulfill the criteria for neurological death and donor management is inherently different; and
- potential thoracic organ donors, due to the concerns about the effect of hypothermia on cardiac performance
Final enrollment criteria were further refined:
- 18 years or older
- No DCD candidates
- Research authorization from first person authorization (FPA) or legal next of kin (LNOK) authorization
- If FPA from outside California, family resource coordinator (FRC) will note
- if research authorization included. If not, will approach LNOK for research authorization.
- No current hemodialysis or CKD/ESRD (based on biopsy or medical history)
- No split liver candidates ( based on transplant center request)
- Hemodynamic stability
- Neosynephrine</= 60mcg/min
- Dopamine </= 10mcg/kg/min
- Norepinephrine </= 10mcg/min
- Corrected coagulopathy
- INR </= 3x normal value
- Platelet >/= 50,000
Donors who were assigned to therapeutic hypothermia were either allowed to spontaneously reach a body temperature of 34 to 35°C or were cooled with the use of forced-air systems or passive-cooling devices. See figure below from Supp appendix for details of hypothermia methods used.
The primary efficacy outcome was DGF, defined per standard definition (need for dialysis within 7 days). The initial efficacy outcome was supposed to be a composite of terminal creatinine (in the donor) and DGF - but when it became apparent that DGF information was available for all recipients, this was changed.
The trial was designed to include 500 donors to have 90% power for 30% difference in DGF between groups. The primary analysis was an adjusted analysis that included covariates - donor creatinine at time of enrollment, donor age, organ procurement agency, donor type and cold ischemia times.
Of the 500 planned, 370 donors (180 in hypothermia and 190 in standard arm) had been randomized when the trial was halted after the interim analysis was done, for overwhelming superiority of the intervention. From these 370 donors, 583 kidneys had been transplanted. Donor characteristics are in table 1 (check out the last row - gives a taste of results to come). Recipient characteristics are in table 2 - and note, recipients were not randomized, and there are several concerning imbalances, especially cold ischemia times (and warm ischemia times are not available in a significant proportion). DGF occurred in 79/280 recipients in the intervention arm (28.2%) versus 112/286 in the standard arm (39.2%) - an impressive difference of about 11% absolute risk reduction (so NNT of 9!). The adjusted analysis - which was the primary outcome, is in table 3, which demonstrates an impressive OR of 0.62 (95% CI 0.43 to 0.92, p value 0.02) despite correction for covariates, including the aforementioned cold-ischemia times.
- Ethical issues: Informed consent was waived by the UCSF IRB since this was deemed non-human subject research (for donors) and minimal risk for recipients.
- How did hypothermia exactly work? What is the secret sauce - does the duration of hypothermia, or the exact temperatures achieved matter?
- Early stopping for efficacy is always a contentious issue, though it appears to be quite well done, using the O-Brien-Fleming boundary.
- Is DGF enough as an outcome, or would one want longer term, hard outcomes?
- Is this enough data (considering this seems a low-risk intervention) to start using it, or do we need confirmation with another RCT?
Join us on Sep 22/23 to look for answers to these questions and more!
Summary by Swapnil Hiremath.