Leaving hospital after a PE is the start of the part of treatment that prevents the next one. It has four pieces: a review at three months where continuing or stopping is decided; a short, specific search for a cause; testing that is only done when it would change that decision; and a check that the lungs have recovered. Each is described here, with what NICE says should happen and where the US guidance adds anything.
1. The three-month review
NICE requires that everyone who has had three months of anticoagulation (three to six months with active cancer) has the benefits and risks of continuing, stopping or changing discussed with them. It is a conversation, not a formality, and it should cover the following.
Was it provoked?
A provoked PE is one with a major transient trigger in the previous three months: surgery, trauma, significant immobility, pregnancy or the weeks after birth, or the combined pill or HRT. Unprovoked means none of those. The distinction drives everything that follows, because a provoked clot rarely recurs once the trigger has gone, and an unprovoked one recurs often: roughly one in ten in the first year off treatment, and around a third by ten years.
Provoked, and the trigger has gone
NICE: consider stopping at three months if the course has been uncomplicated. If treatment stops, the person should leave with written information on symptoms, a direct contact number for a thrombosis team, and out-of-hours details. A clot provoked by the pill or HRT counts as provoked; the hormone is stopped and an alternative found.
Unprovoked
NICE: consider continuing beyond three months, and tell people with a low bleeding risk that the benefits of continuing are likely to outweigh the risks. The decision is a balance of the recurrence risk against the bleeding risk and the person's own preferences. NICE says not to rely on a prediction score alone, but allows the HAS-BLED bleeding score as a guide and says to discuss stopping when it is 4 or more and cannot be improved.
Which drug, long term
Continue what has worked, if it is tolerated. If it is not, or circumstances have changed, NICE suggests switching to apixaban, which bled least in the trials. For people who decline to continue any anticoagulant, aspirin 75 or 150 mg is an option that roughly halves recurrence for a couple of years, well short of what an anticoagulant does.
The reduced dose. After six months, both apixaban (2.5 mg twice daily) and rivaroxaban (10 mg once daily) are licensed at half the treatment dose for extended prevention. In the trials that established this, AMPLIFY-EXT and EINSTEIN CHOICE, the half dose prevented recurrence about as well as the full dose in patients whose original clot was of low or uncertain recurrence risk, with major bleeding under 1% a year. A 2025 trial (RENOVE) tested the half dose in higher-risk patients and could not confirm it was as effective, although it bled less; for someone with a strong reason to continue — a second unprovoked clot, an antiphospholipid diagnosis — the full dose remains the safer assumption. NG158 itself does not name a dose; it says continue the current anticoagulant.
Every year thereafter
NICE: review general health, recurrence risk, bleeding risk and preferences at least once a year for anyone on long-term anticoagulation or aspirin. If that review is not happening, ask for it.
2. Looking for a cause: cancer
An unprovoked clot is sometimes the first sign of a cancer, and the natural instinct is to scan everything. The evidence says not to. In the SOME trial, adding a CT of the abdomen and pelvis to a basic assessment found no more cancers than the basic assessment alone, and around 4% of people with an unprovoked clot were diagnosed with cancer within a year either way. NICE's position follows: for an unprovoked DVT or PE, review the history, examine the person, and review the baseline bloods (full blood count, kidney and liver function, clotting). Do not offer further cancer investigations unless there are relevant symptoms or signs, in which case the suspected-cancer pathway applies. The US guideline reaches the same conclusion. What "examination" should mean in practice: chest, abdomen, breasts, prostate or testes, lymph nodes and skin, plus up-to-date participation in the national screening programmes for the person's age.
3. Looking for a cause: thrombophilia and genetic tests
A thrombophilia is a tendency to clot. Some are inherited — factor V Leiden, the prothrombin gene variant, and deficiencies of protein C, protein S or antithrombin — and one is acquired: antiphospholipid syndrome, in which antibodies make blood clot in both veins and arteries.
The tests are easy to order and hard to use well, for two reasons. First, the common inherited variants only modestly raise recurrence risk, so a positive result rarely changes a decision that the clinical picture has already made. Second, anticoagulants and the acute clot itself distort several of the results, so tests taken in hospital can be wrong. NICE's rules follow from that:
- Do not test anyone who is going to continue anticoagulation anyway: the result changes nothing.
- Do not test after a provoked clot.
- Consider antiphospholipid antibodies after an unprovoked clot if the plan is to stop anticoagulation, because a positive result is a reason to continue, and to use warfarin rather than a DOAC if the syndrome is "triple-positive".
- Consider inherited thrombophilia testing after an unprovoked clot in someone with a first-degree relative who has had a DVT or PE, if the plan is to stop.
- Do not routinely test relatives of someone found to have a thrombophilia.
The tests should be done at least a few weeks after the acute event, with specialist advice on timing and on any interruption of anticoagulation. The US guidance and the American Society of Hematology say the same thing in more words. If a clinic offers a "full thrombophilia screen" on the ward in the first week, the results will need repeating.
When a wider search is justified. A clot in an unusual place (arm, abdomen, brain veins), clots at a young age, clots in both veins and arteries, recurrent miscarriage, or a strong family history all justify a haematology referral. Tests that may then be worth doing include the JAK2 mutation (a bone-marrow condition that presents with clots), a check for paroxysmal nocturnal haemoglobinuria, and a fuller autoimmune screen. These are not routine after an ordinary first PE.
The every-test chapter lists each of these tests, what it detects, how common the result is, the timing rules that make most ward-ordered panels uninterpretable, and the tests that are sold but not worth doing.
4. Making sure the lungs recover
Around one person in three is still short of breath, tired or limited in exercise months after a PE. Most of this is deconditioning, anxiety and the after-effects of the illness, and it recovers with time and a graded return to activity. A small number have chronic thromboembolic pulmonary hypertension (CTEPH): the clot has not cleared but has scarred into the artery wall, and the pressure in the lung circulation rises. It affects roughly 3 in 100 people after a PE, it is progressive if untreated, and it is one of the few forms of pulmonary hypertension that can be operated on and cured.
The European guidance recommends a structured review at three to six months for everyone, asking about breathlessness and function. Anyone still symptomatic gets an echocardiogram; if that suggests raised pressure, a V/Q scan (which is more sensitive than CT for old, organised clot), and referral to one of the national pulmonary hypertension centres. NG158 is silent on this, so in the UK it depends on the local pathway. If you are still breathless at three months, ask for the echocardiogram. The US guideline includes the same follow-up recommendation.
5. Practical things that should be in place at discharge
- Written information on how and how long to take the anticoagulant, side effects, food and alcohol interactions, dental treatment, sport and travel, and pregnancy (NICE 1.5.1).
- An anticoagulant alert card, carried at all times (NICE 1.5.2).
- A named contact and out-of-hours number.
- The date of the three-month review, and who is doing it.
- If a hormone was the trigger: confirmation it has been stopped and what replaces it.
- If a filter was placed: the removal plan, in writing.
6. Where UK and US guidance differ
Very little in this chapter. Both stop after provoked clots, continue after unprovoked ones with shared decision-making, limit cancer investigation to examination and basic bloods, and confine thrombophilia testing to cases where it changes the plan. The US guideline is more explicit about the 3-to-6-month functional review and the CTEPH pathway; NG158 leaves that to the European guidance and local practice, which is why it is worth asking for rather than waiting for.