This post covers a talk I was asked to give at the recent RCEM CPD conference in Belfast. A great event, and well hosted in spite of the LOC chair becoming indisposed by imminent fatherhood. Congrats Ian – this one’s for you…
I was asked principally to talk about the new British Thoracic Society guidelines for the outpatient management of pulmonary embolism1, but thought I would add a dusting of NICE guideline revision and a few cherries of recent literature, which should all directly impact how we provide care to these patients.
Why are we talking about this? Well, PE lends itself very well to the topic of ambulatory care. It is a potentially serious condition with significant morbidity and mortality, and it often presents insidiously. So we think about it. And we look hard for it. And this is good. But our increasing vigilance in tandem with better scanners, keener radiologists and patient awareness has resulted in a dilute pretest probability of around 5%. So we need to think carefully about how we balance resource use and time. And what we hope to achieve through hospital admission. That’s why it’s one of the top 5 conditions on the RCEM ambulatory toolkit2.
Emergency Medicine has been ahead of the game here I think. This paper was published in Manchester3 back in the days when I was an F3 just scratching my chin about the idea of a career in EM. The authors ambulated 425 patients with suspected PE attending the ED, with disease confirmed in 5% and an adverse event rate of <0.5%. How did you get on the pathway? HR <100 RR <20 SPO2 >95%. Simples.
There are lots of other options now. What do you do? Do you ambulate suspected disease? Do you ambulate confirmed disease? How do you decide? Are you confident in your pathway?
The British Thoracic Society noted variation in the management of this disease back in 2015 and felt a consensus multispecialty guideline would add value. They went about it well I think, inviting representation from the Society of acute medicine, RCEM, British Society of Haematology and ,many others including NICE methodologists. They also had the advantage of not being quite as hamstrung by randomised trial and economic data as NICE can be. The concept was to provide an evidence review and best practice consensus to support outpatient management with associated quality standards.
What is it not? This is not a diagnostic pathway. The NICE CG144 update released later this year will I suspected cover PERC, age adjusted d-dimer and all the other things that are crying out for national guidance. It is also not entirely about ambulatory care; there are sections here that refer to PESI484 and the idea of early discharge after a short period of inpatient care. It is also not a pathway for someone who worries you. I am a big fan of guidelines guide and clinicians decide, and you are the latter. If you think someone does not tick the box for ambulatory care but you can’t quite put your finger on why, then they don’t tick the box.
The guideline tries to embed this logic in the exclusion criteria. These are obvious but important. I for one am certainly seeing increasing problems with reluctance to admit patients where ambulatory care is feasible, even if they are sick or complex. These criteria offer a good line of defence against the overzealous medical registrar.
Key issues – risk stratification
What are the big take home points from this guideline? Well the first concerns risk assessment. This is the HOLY GRAIL OF VTE CARE if you like; a complex disease that can range from the probably doesn’t need any treatment, to the immediately life threatening, with very little in the way of physiological derangement in between. We need some help to decide where people sit on this spectrum.
How can we risk assess? Multiple scores have now been externally validated in data on >40,000 patients5. It is worth highlighting that these scores are different, but most have been designed to identify patients at higher risk of death, recurrence or haemodynamic collapse. As such they are screen in tools, assuming that the default is ambulatory care. This is ok provided you understand the design. The commonest tools in use are the PESI6, SPESI and HESTIA7 score. A low risk on any of these scores reduces the chance of adverse events to <2%.
The first thing to ask with regard to prognostic scoring systems, is whether they add value to gestalt. Surely I can decide who is low risk and who is high? Surely you can? Well, there is some emerging literature on the variation in gestalt by grade and interest. Sure enough, it’s been looked at for ambulatory PE within this paper from 20178. Here we see the SPESI outperform all clinicians across multiple attending and registrar grades with regard to sensitivity at ruling out a poor outcome. We also see a variety in low risk classification ranging from 20 to 40%. We can’t all be right….
The second thing to ask is if scores are good, then which score is best? This is a bit more goldilocks. PESI is complex but ordinal, and so can provide information on low risk, while also identifying those at higher risk for complication. SPESI is simple and dichotomous, with good prognostic characteristics but includes cancer as a variable that limits application to a degree. HESTIA purports to identify more patients as low risk, but again is dichotomous and has less external validation. There are papers like this one comparing 2 scores in small cohorts9, but no definitive evidence of superiority. The take home from the guideline is that any of these options are reasonable, as long as they are used in the manner intended and with insight into the pro’s and cons.
Key issues – advanced prognostic markers.
We don’t just have scores though. We have a plethora of advanced prognostic tools at our twitchy and tech loving fingertips. Want biomarkers? BNP10, NT BNP11, trop and HS trop for you madam. Want RV imaging? Perhaps some CT RV:LV ratio data12, focussed echo13 or PAP estimation sir. Want bleeding risk scores? I recommend HASBLED14 or VTE bleed15 to while away the rest of your evening….
But does any of this add benefit? David Jiminez has looked fairly extensively at the cumulative value of prognostic information; his study from 2014 suggests that SPESI outperforms the vast majority of these tests16. Essentially, with a validated score you are reducing the chances of an adverse event to <2%. Additional markers of PE severity may perhaps further reduce this risk, although potentially at the cost of reducing the number of overall patients suitable for outpatient management with only minimal further risk reduction. Further measurements may also add to time and cost. Are you sure you want the second course??
It is worth a specific mention that several of the respiratory physicians on this guideline felt negative biomarkers may have a role to play in identifying a small subset of patients who have PE and signs of RV strain on CT, but are still safe to consider ambulatory care. This is pretty individualised I think. It would certainly take an interesting set of circumstances for me to send home anyone with RV strain on the day of diagnosis.
Key issues – treatment and follow up
We should tackle follow up first really as it’s an interesting and under researched area. Whilst we are good at conducting research which proves ambulatory care in certain conditions does not kill people, we tend to be very bad at studying what information and support needs patients actually have. However, if we are going to believe the evidence here and apply it appropriately, then it seems clear we need to model the pathways within the trials.
What did the original and applicable randomized trials do? Quite a lot actually. The Aujesky17 and Beam studies18 provide some good data here, discharging patients with written and verbal guidance on the risks of recurrence, major bleeding and other adverse events. But not only this; they ensured routine follow up within the first week either by face to face appointment or by phone contact. Think about the value of this. How are you feeling? Are you taking your meds? Have you had any bleeding? Are you more breathless? Do you know who to call or where to come if you have a problem? This type of early safety netting and follow up is vital for expedient ambulatory care, and as such is endorsed within this guideline.
And then there is treatment. What should we use in suspected disease? What should we use in confirmed? What should we use in cancer patients? Well, this guideline is the first to endorse the use of a single agent DOAC pathway in the first 2 situations. If your hospital is anything like mine, I am sure there has been resistance to the idea of prescribing DOACs for patients with suspected disease. I have heard all combinations of licensing, compliance, reversibility and adverse event concerns. The same criticisms can easily be made of any form of LMWH. DOAC agents are now cheaper (in the short term) and arguably more effective in straightforward cases than other forms of anticoagulation. A recent SR/MA published in blood19 looking specifically at their performance in treatment of VTE, suggests the Factor Xa DOAC agents have a lower rate of major bleeding, Intracranial bleeding and fatal bleeding, with non inferior rates of VTE recurrence. Not bad for a few pills, when previously we required a series of injections and a district nurse.
What about DOACs compared to LMWH though? That’s what we have always used in suspected disease? Well, I would suggest that whenever you compare 2 treatments for a condition where 95% of patients do not actually have the condition, you’ll be unlikely to definitively prove any result. As such we just won’t get data on suspected disease. However, given it’s supposed improved clinical effectiveness in cancer patients, there is now data comparing LMWH to DOACs in this group for treatment of confirmed VTE. A SR/MA by a Canadian team published this year20 suggests the balance is a bit trickier, with lower rates of recurrence on DOAC therapy, but higher rates of bleeding. Not clear really.
As such, we need to look at practicality, cost, convenience. I think the take home is that there are multiple options. Offer your patients the choice. But remember the concerns and contraindications. And if you want to use a single agent DOAC pathway, your choice in suspected disease is limited to rivaroxaban, apixaban. The others require a 5 day LMWH lead in.
Key issues – special circs
Worth a nod to the specialist group comments within this guideline, which highlight a few key differences. Intravenous drug users are determined to be a higher risk group overall, both regarding compliance and risk for recurrence. As such an inpatient pathway is recommended.
Cancer patients are noted to be at slightly higher risk overall, and it should also be noted that SPESI automatically excludes everyone with cancer – therefore you might want to use HESTIA in this group to stratify risk. The mortality with cancer patients is significantly higher and that puts some people off the idea of ambulatory care in general. But I think I would ask you what you hope to achieve by admitting these patients to hospital. Is it for their benefit? Or for yours? Probably one for shared decision making.
Lastly, this is the first guideline I have seen that is clear about the fact that it still applies to pregnant patients. This is a real step forwards. Often, these patients are excluded from trials and guidelines using the rationale of safety and concerns on generalisability of the evidence. What this unfortunately leads to is an evidence free zone of people doing whatever they please. The committee were keen that pregnant patients could be offered the opportunities associated with outpatient care, but with reference to the caveats. First – risk scores cannot be applied to this group. Second – DOAC agents are contraindicated in this group. Third – MDT discussion should be encouraged, with regard to shared obstetric and haematology services.
The hunt is on (for the needle in the haystack)
Whilst we’re talking about suspected PE in pregnancy, it would be remiss of us not to highlight the recent progress in the literature. March 2019 saw the publication of ARTEMIS in the New England Journal of Medicine21; this followed the HTA funded Diagnosis of PE in Pregnancy (DIPEP) project22, published across a variety of haematology and obstetric journals last year; this followed Marc Righin23i’s earlier study in Annals of Internal Medicine and this followed lots of smaller studies and clinical practice reviews24. We must have the answer by now? Unfortunately not – these studies offer conflicting conclusions.
ARTEMIS was a multicentre prospective observational cohort study looking at 498 patients with suspected PE in pregnancy. These women were managed according to a pregnancy adjusted YEARS algorithim, which is a decision tool containing only 3 dichotomous variables – haemoptysis, signs of DVT and is PE the most likely diagnosis? After this you applied a probability adjusted cutpoint to your d-dimer level. Patients with a d-dimer below the individualized cutpoint were discharged without imaging. This strategy avoided CTPA in somewhere between 30-60% patients, no PEs were seen during 3 month follow up and a composite VTE event rate overall of 0.21% was noted, with a conclusion that this algorithm appeared safe and effective.
DIPEP was quite different. This was an observational cohort with added positive cases from UK Obstetric Surveillance Service. Approx 300 cases were prospectively recruited with supplemented cases of positive VTE. Only 6.5% of these patients were in the first trimester, compared to double that in ARTEMIS. The results from this study were less positive about the roles of clinical prediction or biomarkers; no prediction rule could be gleaned from the data and no biomarker exceeded an AUROC value of 0.7. The conclusions supported going straight to imaging without any messing around.
Interesting stuff. Why the disparity? Well, I suspect it’s a combination of low pretest probability, inclusion criteria, confounding and gestalt. The studies are quite different really. There are very few patients in the first trimester within DIPEP, whereas ARTEMIS highlights that this group are exactly where the highest efficiency of their algorithm was found. This discrepancy in results doesn’t really solve the problem, but does provide a further evidence base that you can look at in light of your current practice and lab assays. Personally, I am buoyed on by ARTEMIS; I think it probably highlights a group where gestalt is low who are suitable for d-dimer measurement. DIPEP reminds us that if you need to image or have any doubts/concerns – get on and do a CTPA. The background radiation is not as terrifying as we have always suspected.
How does this all help
I am sure many of you have been ambulating PE for years. But these guidelines and this recent literature provide support to some of the greyer areas and create national templates to support delivery of care that you know is right, but have been struggling to deliver. I think the key areas are as follows:
- There is now clear guidance supporting the use of risk scores and which ones are preferable
- Clear exclusion criteria to defend a decision to refer for inpatient care
- First guideline to advocate a single DOAC pathway as an option; this will help you with your medicines management committees.
- First guideline to even suggest that ambulating pregnant patients is an option, although complex as we have seen.
- First guideline to provide a template for follow up care
Quality standards are coming that will lend further support to development in this area. Need buy in, funding, staff or resources to deliver any of the above? Lever this national guidance and quality standards to get it.
Anything else to add?
There’s lots of other stuff out there to look at. NICE guidelines will be out for consultation this summer. More literature will come I am sure. And the BTS guidelines will drive care, which will create a lot of quality improvement work. Look at the papers highlighted above and try the following resources below if you want more.
Hope that helps. Good luck with your ambulating. Let these guidelines help you; but don’t let anyone tell you what to do. You are the clinician. You decide.
- 1.Howard LS. BTS guidelines for the initial outpatient management of pulmonary embolism: there’s no place like home. Thorax. June 2018:607-608. doi:10.1136/thoraxjnl-2018-211646
- 2.RCEM Guidance – Ambulatory Care. RCEM . https://www.rcem.ac.uk/RCEM/Quality-Policy/Clinical_Standards_Guidance/RCEM_Guidance.aspx?WebsiteKey=b3d6bb2a-abba-44ed-b758-467776a958cd&hkey=862bd964-0363-4f7f-bdab-89e4a68c9de4&RCEM_Guidance=3. Published 2019. Accessed April 20, 2019.
- 3.Hogg K. Outpatient diagnosis of pulmonary embolism: the MIOPED (Manchester Investigation Of Pulmonary Embolism Diagnosis) study. Emergency Medicine Journal. February 2006:123-127. doi:10.1136/emj.2005.027110
- 4.Moores L, Zamarro C, Gómez V, et al. Changes in PESI scores predict mortality in intermediate-risk patients with acute pulmonary embolism. Eur Respir J. June 2012:354-359. doi:10.1183/09031936.00225011
- 5.Elias A, Mallett S, Daoud-Elias M, Poggi J-N, Clarke M. Prognostic models in acute pulmonary embolism: a systematic review and meta-analysis. BMJ Open. April 2016:e010324. doi:10.1136/bmjopen-2015-010324
- 6.Pulmonary Embolism Severity Index (PESI) – MDCalc. MDCALC. https://www.mdcalc.com/pulmonary-embolism-severity-index-pesi. Published 2019. Accessed April 20, 2019.
- 7.Simplified PESI (Pulmonary Embolism Severity Index) – MDCalc. MDCALC. https://www.mdcalc.com/simplified-pesi-pulmonary-embolism-severity-index. Published 2019. Accessed April 20, 2019.
- 8.Jiménez Castro D, Barrios D, Morillo R, Nieto R, Guerassimova I, Gomez V. Clinical gestalt and the prognosis of pulmonary embolism. In: Pulmonary Circulation and Pulmonary Vascular Disease. European Respiratory Society; 2017. doi:10.1183/1393003.congress-2017.pa2357
- 9.Crobach MJT, Dolsma A, Donker ML, et al. Comparison of two methods for selection of out of hospital treatment in patients with acute pulmonary embolism. Thromb Haemost. 2013:47-52. doi:10.1160/th12-07-0466
- 10.Coutance G, Le Page O, Lo T, Hamon M. Prognostic value of brain natriuretic peptide in acute pulmonary embolism. Critical Care. 2008:R109. doi:10.1186/cc6996
- 11.Coutance G, Le P, Lo T, Hamon M. Prognostic value of brain natriuretic peptide in acute pulmonary embolism. Crit Care. 2008;12(4):R109. https://www.ncbi.nlm.nih.gov/pubmed/18721456.
- 12.Ghuysen A. Computed tomographic pulmonary angiography and prognostic significance in patients with acute pulmonary embolism. Thorax. November 2005:956-961. doi:10.1136/thx.2005.040873
- 13.Pavlidis A, Kallistratos M, Karamasis G, et al. Diagnosis and risk stratification in acute pulmonary embolism: the role of echocardiography. Rev Cardiovasc Med. 2013;14(1):56-65. https://www.ncbi.nlm.nih.gov/pubmed/23651987.
- 14.Kooiman J, van Hagen N, Iglesias del Sol A, et al. The HAS-BLED Score Identifies Patients with Acute Venous Thromboembolism at High Risk of Major Bleeding Complications during the First Six Months of Anticoagulant Treatment. Garcia de Frutos P, ed. PLoS ONE. April 2015:e0122520. doi:10.1371/journal.pone.0122520
- 15.Barco S, Konstantinides S, Klok F. External validation of the VTE-BLEED score for predicting major bleeding in stable anticoagulated patients with venous thromboembolism. Thromb Haemost. 2017:1164-1170. doi:10.1160/th16-10-0810
- 16.Jiménez D, Kopecna D, Tapson V, et al. Derivation and Validation of Multimarker Prognostication for Normotensive Patients with Acute Symptomatic Pulmonary Embolism. Am J Respir Crit Care Med. March 2014:718-726. doi:10.1164/rccm.201311-2040oc
- 17.Aujesky D, Roy P-M, Verschuren F, et al. Outpatient versus inpatient treatment for patients with acute pulmonary embolism: an international, open-label, randomised, non-inferiority trial. The Lancet. July 2011:41-48. doi:10.1016/s0140-6736(11)60824-6
- 18.Beam DM, Kahler ZP, Kline JA. Immediate Discharge and Home Treatment With Rivaroxaban of Low-risk Venous Thromboembolism Diagnosed in Two U.S. Emergency Departments: A One-year Preplanned Analysis. Hiestand BC, ed. Acad Emerg Med. June 2015:788-795. doi:10.1111/acem.12711
- 19.Li A, Garcia DA, Lyman GH, Carrier M. Direct oral anticoagulant (DOAC) versus low-molecular-weight heparin (LMWH) for treatment of cancer associated thrombosis (CAT): A systematic review and meta-analysis. Thrombosis Research. January 2019:158-163. doi:10.1016/j.thromres.2018.02.144
- 20.Li A, Garcia D, Lyman G, Carrier M. Direct oral anticoagulant (DOAC) versus low-molecular-weight heparin (LMWH) for treatment of cancer associated thrombosis (CAT): A systematic review and meta-analysis. Thromb Res. 2019;173:158-163. https://www.ncbi.nlm.nih.gov/pubmed/29506866.
- 21.van der Pol LM, Tromeur C, Bistervels IM, et al. Pregnancy-Adapted YEARS Algorithm for Diagnosis of Suspected Pulmonary Embolism. N Engl J Med. March 2019:1139-1149. doi:10.1056/nejmoa1813865
- 22.Goodacre S, Horspool K, Shephard N, et al. Selecting pregnant or postpartum women with suspected pulmonary embolism for diagnostic imaging: the DiPEP diagnostic study with decision-analysis modelling. Health Technol Assess. 2018;22(47):1-230. https://www.ncbi.nlm.nih.gov/pubmed/30178738.
- 23.Righini M, Robert-Ebadi H, Elias A, et al. Diagnosis of Pulmonary Embolism During Pregnancy. Ann Intern Med. October 2018:766. doi:10.7326/m18-1670
- 24.Simcox LE, Ormesher L, Tower C, Greer IA. Pulmonary thrombo-embolism in pregnancy: diagnosis and management. Breathe. December 2015:282-289. doi:10.1183/20734735.008815
- 25.Condliffe R, Elliot CA, Hughes RJ, et al. Management dilemmas in acute pulmonary embolism. Thorax. December 2013:174-180. doi:10.1136/thoraxjnl-2013-204667