Estimated reading time: 8 minutes
Back in March we reported on a pre-print trial of thromboprophylaxis of ICU patients. The results suggested that there was no benefit for therapeutic dose anticoagulation in COVID19 as a prophylactic strategy for critically ill patients with COVID19. This was surprising to many, who have seen high rates of venous thromboembolism in sicker pandemic patients.
At that time we also alluded to a rumour that the same platform study had found a different, and more positive effect, amongst ward based patients. No preprint was available for those results.
This week both trials have been published in the NEJM. The ICU trial is published here and the non-critically patient based trial here. There are no major changes to the critically ill subgroup results in the full paper, so we are happy with our previous take on that one. However, it’s time to have a good look at the slightly controversial results regarding moderately ill patients. The abstract is below, but as we always say, you should read the full paper before coming to any conclusions yourself.
However, as we’ve not reviewed the non-critically ill cohort it’s time to bring that question up to date. The abstract is below, but as we always say, you should read the full paper before coming to any conclusions yourself.
The three trials were the Randomized, Embedded, Multifactorial Adaptive Platform Trial for Community-Acquired Pneumonia (REMAP-CAP; NCT02735707), Accelerating Covid-19 Therapeutic Interventions and Vaccines-4 Antithrombotics Inpatient platform trial (ACTIV-4a; NCT04505774 and NCT04359277), and Antithrombotic Therapy to Ameliorate Complications of Covid-19 (ATTACC; NCT04372589). The investigators The three contributing trials to this multiplatform trial were the Randomized, Embedded, Multifactorial Adaptive Platform Trial for Community-Acquired Pneumonia (REMAP-CAP; NCT02735707), Accelerating Covid-19 Therapeutic Interventions and Vaccines-4 Antithrombotics Inpatient platform trial (ACTIV-4a; NCT04505774 and NCT04359277), and Antithrombotic Therapy to Ameliorate Complications of Covid-19 (ATTACC; NCT04372589). The investigators aligned their outcome measures, interventions and inclusion criteria across all three trials, to essentially provide a quicker result from a larger body of work. This is clever, but we should be clear that this methodology does not really increase the validity of results. This is not three trials with the same direction of result. It is three trials, pooled together, with a single result. (Ed – would this happen in non-COVID times I wonder).
Tell me about the patients.
This trial included hospitalised patients who were not on the ICU, or as they describe it ‘moderate disease’. There was some lack of clarity between the trials as to what this entailed though in broad terms, moderate disease severity was defined as hospitalization for Covid-19 without the need for ICU-level care.
ICU-level care was defined as the use of respiratory or cardiovascular organ support (oxygen delivered by high-flow nasal cannula, noninvasive or invasive mechanical ventilation, or the use of vasopressors or inotropes) in an ICU. In ACTIV-4a, investigators found that defining ICU-level care was tricky, so they settled on receipt of organ support, regardless of hospital setting to define ICU-level care. Patients who were admitted to an ICU but who did not get qualifying organ support were considered to be moderately ill.
So the definition revolved around what the patients got, and not so much on where they were. I think this is a sensible and pragmatic definition, particularly during a pandemic where critical care is often ‘without walls’ and relocated to theatre recovery areas, acute respiratory units etc..
Patients were also stratified according to d-dimer levels, but this was done to inform a prespecified subgroup analysis, not to guide intervention.
In terms of exclusions there was additional variation with time limits of 72 hours up to 14 days for inclusion into the trial between the three platforms.Patients with a clear reason for anticoagulation were excluded. It is also essential to recognise that all 3 trials excluded patients at risk of major bleeding – like all anticoagulation trials, this means the results may lack generalisability to a wider cohort of frail patients, or those with multiple comorbidities.
What about the interventions?
Patients were randomised centrally between treatments. There were two open label treatment arms
§ Usual care prophylactic dose anticoagulation (as per local protocols). It is worth noting that this was an intermediate dosing strategy in 26.5% of patients.
§ Therapeutic dose anticoagulation (as per local protocols equivalent to treatment of acute venous thromboembolism) delivered for up to 14 days or until recovery.
A range of heparin agents were used. Most were low molecular weight heparins administered subcutaneously.
What about the outcomes?
The primary outcome was organ support free days up to day 21. This is an ordinal scale which includes survival to hospital discharge, and days free of organ support to day 21 in survivors. A relevant, but interesting choice – it assumes that we are happy with a composite outcome and one which is subject to clinical decision. There are often cases where it is down to individual clinicians when they stop the high flow nasal oxygen/NIV, rather than fixed protocols.
They also followed patients up to 90 days and looked at a number of secondary outcomes.
What about the results?
The trial recruited 2219 patients with reasonable (but not fabulous) adherence to protocol. Adherence was notably better in the prophylaxis group.
With regard to the primary outcome, 1048 patients received usual-care thromboprophylaxis group of which 801 (76.4%) survived until hospital discharge without receipt of organ support during the first 21 days, as compared with 939 of 1171 patients (80.2%) in the therapeutic-dose anticoagulation group. So that’s a 3.8% difference (with confidence intervals ranging from 0.5% to 7.2%). Oddly though the paper reports a 4% difference which I can’t quite understand. So that sounds like a positive effect and would no doubt make a difference to patients and health services. In their Bayesian analysis this equates to a 98.6% probability that full anticoagulation is better.
Amongst the secondary outcomes the effects were less dramatic. Notably in terms of overall survival in hospital. There the difference was 92.7% for therapeutic anticoagulation vs. 91.8% for prophylaxis. In their Bayesian analysis this equated to an 87.1% likelihood of benefit in terms of survival.
When stratified by d-dimer level, those with a higher initial d-dimer were had a greater benefit from therapeutic anticoagulation as compared to thromboprophylaxis (Adjusted risk 5.1% vs. 1.4%) suggesting that d-dimer may be a marker for those who are most likely to benefit.
What about the analysis/stats?
Putting three trials together is not easy. In this trial they have analysed thromboprophylaxis in each patient as a separate event, rather than combine the three trials together as three separate entities as you might in a typical meta-analysis. A meta-analysis of individual patient data makes more sense in this setting.
The analysis is a Bayesian approach that essentially aims to determine the probability of the results found reflecting the truth.
Why the difference between ICU and non ICU patients?
This is one of the most interesting findings in these studies and we’d recommend reading the editorials and visiting the critical care reviews website to find out more. Perhaps the most plausible explanation is that ICU patients are too far down the inflammatory pathway for the additional anticoagulation to have an impact. We also know that critically ill patients have a higher risk of bleeding, so the balance of risk/benefit is always likely to be more delicate.
So what shall we do?
In the UK, national guidelines have already bought into these trial results. NICE guideline NG191 on COVID19 management has had a consensus recommendation in support of therapeutic dose anticoagulation for moderately ill patients, based on prior access to this ‘academic in confidence’ data. https://app.magicapp.org/#/guideline/L4Qb5n/section/L6yPVj
Whether this recommendation will now be upgraded following formal publication of the results remains to be seen. This trial was open label, used a composite outcome and had multiple exclusions. It is essentially a North American trial, with <10% participants recruited in a UK setting. It is also a potentially challenging therapeutic pathway to implement – your local thrombosis committee chair will I suspect be only too happy to tell you how hard it is to get standard prophylaxis prescribed reliably for hospital inpatients. Now we are talking about a pathway of treatment dose if you have COVID19, reduced if you get critically ill, re-escalated if you get diagnosed with a VTE, potentially with some intermediate dosing in between….? Very hard to implement reliably, absolutely ripe for prescribing errors and probably in need of further study within a pragmatic healthcare setting, We would suggest. In addition, other trial data recently published in the Lancet looking at therapeutic anticoagulation in hospitalised COVID19 patients, has not been so positive (https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(21)01203-4/fulltext). We have to say – the jury is still out a little on this.
Should we use d-dimer levels to guide decision making? We are not sure we know the answer to this either. NICE guidance is currently explicit that you should not do this, as the preceding body of evidence before this trial is very weak. And we suppose if you are doing a d-dimer in a patient with significant COVID19 disease and you find it to be more than double the normal limit, should you not consider whether your patient has an acute immunothrombosis or embolic clot burden? Should you not then image them to find out? And if they do have one, should you not treat this with therapeutic anticoagulation? We find we are doing this a lot locally – it helps guide individual assessments of risk/benefit in this uncertain area, and allows us to tailor patient care with shared decision making and up to date clinical information.
The authors of this multiplatform trial should be congratulated – it is a great piece of work, and signifies a real coming together of international academics to try and develop the evidence base in an important area. But it leaves us with more questions than answers really.
Dan Horner and Simon Carley
Therapeutic Anticoagulation with Heparin in Noncritically Ill Patients with Covid-19 The ATTACC, ACTIV-4a, and REMAP-CAP Investigators https://www.nejm.org/doi/full/10.1056/NEJMoa2105911
Therapeutic Anticoagulation with Heparin in Critically Ill Patients with Covid-19 The REMAP-CAP, ACTIV-4a, and ATTACC Investigators* https://www.nejm.org/doi/full/10.1056/NEJMoa2103417?query=featured_home
Critical Care Reviews website https://criticalcarereviews.com/
Simon Carley, “JC: Thromboprophylaxis on the ICU for COVID-19 patients. St Emlyn’s,” in St.Emlyn’s, March 24, 2021, https://www.stemlynsblog.org/jc-thromboprophylaxis-in-covid-19-patients-st-emlyns/.