You don't need to be signed in to read BMJ Blogs, but you can register here to receive updates about other BMJ products and services via our site.

Venous thromboembolism

Primary Care Corner with Geoffrey Modest MD: Dvt recurrence in unprovoked dvts — HERDOO2 tool

24 Apr, 17 | by

​by Dr Geoffrey Modest

One perplexing issue in primary care is the appropriate duration of anticoagulation for people with unprovoked venous thromboses. A recent international study found that a specific clinical decision rule was effective in predicting recurrent DVT in women and could permit individualizing different therapies (see​).


— 2747 participants with a 1st unprovoked venous thromboembolism, VTE (either DVT with a noncompressible segment in the popliteal vein or more proximal leg veins and/or documented pulmonary embolism) who had completed 5 to 12 months of short-term anticoagulant treatment were followed prospectively from 44 healthcare centers in 7 countries (from North America, Europe, India, Australia), from 2008 to 2015.

— Mean age 54, 84% white, 75% on vitamin K antagonists for anticoagulation, VTE event was isolated DVT 41%/isolated PE 40%/DVT and PE 21%

— They used the HERDOO2 clinical decision rule: Hyperpigmentation, Edema, or Redness in either leg; D-dimer level ≥ 250 µg/L; Obesity with BMI ≥ 30; or Older age ≥ 65. D-dimer levels were drawn during anticoagulant treatment.

— Of these components: 24% had hyperpigmentation, edema or redness of leg/50% D-dimer ≥250 µg/ 32% >65 yo/ 43% BMI ≥30.

— Low risk patients (women with HERDOO2 score ≤1) were to discontinue anticoagulants (and almost all did); for high risk women and men it was left to the discretion of the clinicians and patients

— primary outcome was an adjudicated symptomatic major VTE


— of 1213 women, 631 (51.3%) were classified as low risk

— 17 who discontinued anticoagulants developed a recurrent VTE during 564 patient years of follow-up (3.0% per patient year)

— of 323 high risk women and men who discontinued anticoagulants, 25 had VTE during 309 patient years of follow-up (8.1% per patient year).

–7.4% in high risk women and 8.4% in high-risk men.

— of 1802 high risk women and men who continued anticoagulants, 28 had recurrent VTE during 1758 patient years of follow-up (1.6% for patient year)

— secondary outcomes:

–1 recurrent PE death (in a high-risk person who continued anticoagulation); risk of major bleeds was nonsignificant in any who stopped anticoagulation, and was 1.2% per patient year in men and high risk women who continued oral anticoagulants. 2 major bleeds were fatal.

–subgroup analyses: in women <50 yo (n=429) rate of recurrent VTE was 2.0% (not related to estrogen use) vs 5.7% in those >50 yo. No difference by country, type of index VTE, or type of anticoagulation


— patients with provoked VTE, such as after surgical procedure, have a 1% chance of VTE recurrence, whereas those with unprovoked VTE have a 10% chance in the 1st year after stopping short-term anticoagulants, 5% in the subsequent year, and 30% at 8 years. 3.6% of recurrent VTEs are fatal. Oral anticoagulation reduces the risk of recurrent VTE by 80-90%.

— The International Society on Thrombosis and Hemostasis suggest that it is safe to discontinue anticoagulants if the risk of recurrent VTE is <5% at one year after discontinuing treatment (with an upper bound of the 95% confidence interval being <8%).

— The HERDOO2 clinical decision rule has been found to be clinically effective in discriminating low risk versus high risk women, though not for men. This study was a large randomized trial in patients with unprovoked VTE.

— of note, over ½ of the women with unprovoked VTE in their study were low risk and could stop their anticoagulants (ie, less than the 5% cutpoint that they noted above)​. So, the potential effect of this decision rule is quite high for women.

— so, where does this HERDOO2 rule come from?? A study done in 2008 (see doi:10.1503/ cmaj.080493​ ) prospectively looked at 600 people with first unprovoked VTE and followed 18 months, finding an overall annual recurrent DVT rate of 9.3%. They focused on the 91 patients with confirmed recurrent DVTs to assess potential risk factors, and developed the HERDOO2 clinical rule, finding annual recurrent VTEs in 1.6% with scores ≤1 and 14.1% in those with higher scores.

— issues about generalizability:

–this study had only an 11.6 month followup (and the original study was only a bit longer), and, as per the above statisitics, lots of recurrent VTE events happen after the 1-year mark

–they excluded the few people with known high-risk thrombophilia (this is not routinely assess after a first event, so not sure why those patients had the test done and if this exclusion could affect the results)

–there were few non-white patients, and the risk of thrombophilia may vary by groups, though there are large deficits in our knowledge here, but there are some data suggesting that factor V Leiden and the prothrombin G20210A mutation are less common in African-Americans, though Black Africans in another study of patients who had strokes tended to have lower levels of protein S, protein C, and antithrombin III levels.

–subgroup analysis in the above study of women >50 yo had a higher VTE recurrence rate of 5.8% and would be good to see if this were a better cutpoint than the ≥ 65​ in the HERDOO2 algorithm

–continuing anticoagulants in the high risk groups was left to the discretion of the clinicians/patients, so unclear who the group was who continued or discontinued the meds and how that might skew those results.

— Overall, would be great to have another study of longer duration and including a more mixed group of patients, to assess generalizability of the results

so, bottom line: this study may well have far-reaching implications, given that a large number of women (not men) might be able to stop long-term (perhaps life-long) anticoagulation for unprovoked first VTE (including PEs, where the risk of a recurrent PE is higher). And, I would add the results of this study to my general gestalt in discussing the pros and cons of stopping anticoagulation. But, to me, this is still such a difficult clinical decision, with potentially life-threatening implications either way, that there should be another confirmatory study in a more mixed population of patients.

See here for a slew of articles on VTE, with my concerns about the novel anticoagulants (NOACs)


Primary Care Corner with Geoffrey Modest MD: Not anticoagulate symptomatic calf dvt

6 Dec, 16 | by EBM

By Dr. Geoffrey Modest MD

One somewhat confusing issue is whether to anti-coagulate a patient with a symptomatic calf deep vein thrombosis. A recent article addressed this issue (see .org/10.1016/ S2352-3026(16)30131-4 ).


  • Randomized double-blind controlled study of low-risk outpatients (no active cancer or previous thromboembolic disease), with a 1st symptomatic DVT in the calf
  • 259 patients from 23 centers in Canada, France, and Switzerland were randomized to low molecular weight heparin (LMWH, in this case nadroparin which is not available in the US) vs. subcutaneous placebo, and treated for 6 weeks. All patients were prescribed 30 mmHg compression stockings and followed for 90 days.
  • 50% women, mean age 52, 45% of the DVTs were in the posterior tibial or peroneal vein, 55% gastrocnemius or soleus. 6% were on daily low-dose aspirin, and 6% were on NSAIDs
  • Risk factors: 19% had family history of venous thromboembolic disease (VTE), 17% were on estrogen therapy, 10% recent surgery, 6% plaster immobilization, 13% travel, 9% bedrest
  • Calf DVT was determined if: there were non-compressible venous segments in the deep calf veins (posterior tibial, peroneal, anterior tibial) or muscular veins (gastrocnemius or soleus)
  • DVT extension was defined as ultrasonographically-proven extension to the popliteal, femoral, iliac, and cava veins
  • The primary outcome was a composite of proximal extension of the calf DVT, contralateral proximal DVT, and symptomatic pulmonary embolism at day 42. Prespecified secondary outcomes included the individual components of the primary outcome, assessed at day 42 and 90.


  • There was no significant difference between the groups in the composite primary outcome, which happened in 4 patients (3%) on LMWH and 7 patients (5%) on placebo
    • DVD extension was detected at day 3-7 in one patient in the LMWH group and 3 in the placebo group; 5 additional patients in the LMWH group and 4 in the placebo group were diagnosed during the 1st 42 days.
    • 2 patients had a nonfatal symptomatic pulmonary embolism, both in the LMWH group, at the 42 week checkup
    • No difference in primary outcome between patients with isolated muscular calf DVT vs those with peroneal or posterior tibial calf DVT
    • One additional venous thromboembolic events occurred between days 42 and 90, in a patient on placebo who developed a symptomatic nonfatal pulmonary embolism at day 89
  • Significant bleeding (1 major bleed, 4 others with clinically relevant bleeding) occurred in 5 patients (4%) in the LMWH group and none with placebo.
  • One patient was diagnosed with heparin-induced thrombocytopenia type 2 with LMWH


  • This study was pretty underpowered, given that they only enrolled 50% of the pre-specified sample size (the steering committee decided to discontinue the study because of poor enrollment). As a result, the current number of patients would have a 90% power detect a 70% risk reduction in the rate of the primary outcome, assuming an incidence of the primary outcome of 10% in the placebo group. However, an additional issue was that the actual number of events in the placebo arm was only 5.4% (vs the expected 10.0%); therefore, mathematically, it would have been necessary to include 1891 patients in each arm to have the initial scheduled statistical power.
  • There was an important study done a long time ago which fed into being more aggressive in treating distal DVTs, and it definitely changed my practice for many years (see Lagerstedt CI. Lancet 1985; 2: 515). This was a 3-month study of 51 patients with symptomatic calf-vein thromboses using open-label warfarin vs no medication, which found that in the non-anticoagulated group, 8 patients (29%) had DVT recurrences (6 symptomatic), 5 with proximal DVT extension and 1 with a pulmonary embolism, but 0% with warfarin (though in 3 of the 8 patients who had recurrences, they were high-risk having had a previous history of thrombosis).
  • One larger concern raised by this article is that sometimes we get way too much data from our diagnostic studies than we know how to deal with. This is certainly true in the MRI studies of low back pain (e.g., see Jensen MC, N Engl J Med. 1994:69,  where they did an MRI on 98 asymptomatic patients, finding that only 36% had a normal MRI, 52% had a bulge on at least one level, 27% had protrusion, and 1% extrusion. Another study of asymptomatic people found that 21% had spinal stenosis!!). In the DVT case we do have many studies over the years showing that following serial ultrasounds one week apart looking for proximal extension of a distal clot seems to work well for predicting who was at risk for thromboembolism. Whole-leg ultrasound was not available in the past, and we seemed to do quite well without it…  and, it turns out, its use leads us to anticoagulating more patients vs just checking for proximal DVT!!!

So, perhaps the safest algorithm, is just to look only for proximal DVT in low-risk patients (ie, avoid ordering the whole-leg imaging). Then repeating the ultrasound in one week if the first did not show a proximal DVT, providing only symptomatic care, including compression stockings as needed. And only anticoagulating if the second ultrasound shows proximal extension. This is consistent with the American College of Chest Physicians clinical practice guidelines (see Chest 2012; 141: e351S), which suggest that in patients with moderate pretest probability of a first DVT, to get D-dimer testing or proximal vein ultrasound, and to do whole-leg US only if patient unable to return for serial US testing. In those getting whole-leg ultrasound and finding an isolated distal DVT, “we [the guideline writers] suggest serial testing to rule out proximal extension over treatment”. And, though the above study was underpowered to elicit a clear conclusion, it was reassuring that only about 5% of the patients had a proximal extension and the number of significant adverse events was really low (3 of 259 in 90 days of followup, about 1%)

Primary Care Corner with Geoffrey Modest MD: Rivaroxaban and Increased Intracranial Hemorrhage vs. Dabigatran

6 Oct, 16 | by EBM

By Dr. Geoffrey Modest

A recent review compared the adverse effects of dabigatran and rivaroxaban, two of the NOACs (non-vitamin K antagonist oral anticoagulants) –see doi:10.1001/jamainternmed.2016.5954. This follows another report finding increased bleeding risks as well as MIs in post marketing case reports for dabigatran following its approval. This new report, funded by Medicare and the FDA, did find a significant increase in major bleeds from rivaroxaban over dabigatran.


  • Retrospective new-user cohort study of 118,891 patients with nonvalvular atrial fibrillation who were 65 years or older and enrolled in Medicare from 2011 to 2014. Patients were included if they had a diagnosis of atrial fibrillation or flutter and filled a first prescription for either drug in this time period. Excluded if they had less than six months of enrollment in Medicare, were younger than 65, were in a skilled nursing facility or nursing home.
  • 52,240 were on dabigatran and 66,651 were on rivaroxaban
  • 50% were 65 to 74-year-old/40% were 75 to 84-year-old, 47% female, 91% white/4% black, 33% with diabetes, 40% hyperlipidemia, 86% hypertension, 15% obese, 19% smoking, 15% with heart failure, 33% with CHADS2 score of 0 to 1, 40% with score of 2, 19% with score of 3; 10% with HAS-BLED score of 1, 54% score of 2, 29% score of 3.


  • 2537 had a primary outcome of thromboembolic stroke (n=306), intracranial hemorrhage (176), or major extracranial bleeding (1209) including major GI bleeding (1018) and mortality (846).
  • Rivaroxaban vs. dabigatran was associated with:
    • 19% reduction in thromboembolic stroke, HR= 0.81 (0.65 to 1.01). Nonsignificant. This translates to 1.8 fewer cases of thromboembolic stroke per 1000 person-years
    • 65% increase in intracranial hemorrhage, HR 1.65 (1.20 to 2.26).  AIRD (adjusted incidence rate difference): 2.3 excess cases per 1000 person-years
    • 48% increase in major extracranial bleeding, HR 1.48 (1.32 to 1.67). AIRD = 13.0 excess cases per 1000 person-years; including 40% increase in major gastrointestinal bleeding, HR 1.40 (1.23 to 1.59). AIRD of 9.4 excess cases per 1000 person-years
    • 15% increase in overall mortality, HR 1.15 (1.00 to 1.32), borderline significant at p=0.051, AIRD of 3.1 excess cases per 1000 person-years. In patients 75 years or older or with CHADS2 score greater than 2, the excess risk was significant (27% and 24% increases, respectively).
    • The excess rate of intracranial hemorrhage exceeded the attributed reduced rate of thromboembolic stroke (the latter actually didn’t reach statistical significance)


  • As I have mentioned in several prior blogs, I am very concerned about the widespread adoption and use of NOACs in non-valvular atrial fibrillation. I believe there was significant drug company malfeasance in both the design and presentation of the data. See blogs referenced at the end for more information. This is not to say that NOACs are decidedly bad, just that the studies on which their approval were based were seriously flawed, perhaps largely because of the drug company cover-ups.
  • As a point of reference, in a large meta-analysis of warfarin in patients with a fib, intracranial hemorrhage occurred in 6 patients on warfarin vs. 3 on control, with n= 2900); and extracranial hemorrhage with warfarin was less than 0.3% per year(see Hart RG. Ann Intern Med 2007; 146:857).
  • It was interesting that more cardiologists prescribed rivaroxaban than family practitioners. Not sure what this reflects. Sensitivity analysis did not show that the patients on rivaroxaban were sicker. Perhaps the cardiologists were aware of the post marketing studies on the harms of dabigatran. Perhaps they were more influenced by the ROCKET-AF study of rivaroxaban(which has been criticized for using defective INR machines, see blogs and NEWS FLASH below). It is important to note that there are no head-to-head comparisons of these different NOAC medications
  • It is perhaps significant that the curves for intracranial hemorrhage, major GI bleeding, and mortality diverged during the follow-up period, though there was convergence on the event rate for thromboembolic strokes (i.e.: increasing problems without increasing benefits).
  • One wonders if some of the increased incidence of bleeding with rivaroxaban is that it actually has the same half-life as dabigatran of about 12 hours, but is dosed at once a day, whereas dabigatran is dosed twice a day
  • The study is limited by several factors. It is a retrospective analysis, and cannot have the rigor of the conclusions of a well conducted randomized trial. Also the mean duration of on treatment follow-up was only four months, though they did have a reasonable representation of those on meds at least 6 to 8 months.
  • So, to me this huge acceptance of NOACs over warfarin (which certainly has its well-known drawbacks) seems more to reflect drug company shenanigans and heavy-duty marketing than clear benefit, and with attendant risks (though i have occasionally prescribed them, e.g. when patients were spending long periods of time in other countries without access to INR monitoring).

For blogs on NOACs, best to go to, which has blogs on the rivaroxaban study (ROCKET-AF) using  defective INR machines in those patients on warfarin, a huge Swedish study showing safety and efficacy of warfarin when patients were mostly in the appropriate therapeutic range, and a slew of articles on drug company malfeasance in promoting dabigatran when internal drug company memos for example showed that they knew that drug levels actually should be monitored (though this went against their main selling point: that one would not have to  monitor levels vs warfarin), as well as the study mentioned above about post-marketing serious adverse events of dabigatran, with the comment that “the Institute for Safe Medication Practices reported that dabigatran was responsible for more serious adverse events than 98.7% of all medications” [and it seems to be better than rivaroxaban, at least by the above study!!!!!]. I also continue to be concerned about the lack of easy access to reversal agents for the NOACs, especially in smaller hospitals outside of large academic centers.

*********NEWS FLASH********

  • There was a new investigative report by Deborah Cohen in BMJ last week (she seems to have spear-headed many of the really great reports in BMJ about drug company malfeasance), stating that the drug company (Janssen) actually knew that the INR testing devices were faulty in the ROCKET AF study (which single-handedly propelled rivaroxaban into blockbuster status), yet withheld this information from the FDA. It turns out that the increased intracranial bleeding with warfarin might have been related to the fact that 33% of actual laboratory readings of INR were above 4, yet the defective point-of-care machines gave an INR on the same sample of between 2-3 (see doi: 10.1136/bmj.i5131)

Primary Care Corner with Geoffrey Modest MD: Warfarin in Nonvalvular Atrial Fibrillation

13 May, 16 | by EBM

By Dr. Geoffrey Modest

Over the years, I have sent out several blogs about the drug company shenanigans/malfeasance in studies promoting NOACs (non-vitamin K antagonist oral anticoagulants) — See link at bottom. Here is a large study suggesting the benefits of warfarin (See doi:10.1001/jamacardio.2016.0199 ). The authors note that studies finding NOAC superiority were in comparison to warfarin where the times-in-therapeutic range (TTR) varied from 55.2% to 64.9%. In the current study researchers looked at the relative effectiveness of warfarin for patients with atrial fibrillation (AF) as it varied with TTR. Data are from a large Swedish registry.


  • Retrospective, multicenter cohort study of 40,449 patients
  • 40% women, mean age 73, mean CHA2DS2-VASc (see below) score 3.3, TTR<70% in 43%, hypertension 60%, heart failure 30%, renal failure 4%, excessive alcohol use 2%, history of falls 8%, prior major bleed 6%, MI 21%, diabetes 18%, stroke 19%, TIA 8%
  • 4311 patients also on aspirin with the warfarin, with the concomitant diseases/risk factors about the same as the overall cohort except that 43% had a prior MI (vs 17% just on warfarin)


  • Annual all-cause mortality 2.19% (2.07-2.31), intracranial bleed 0.44% (0.39-0.49)
  • Comparing those with TTR <70% vs >70% (all are annual rates)
    • All-cause mortality was 4.35% vs 1.29%
    • Any major bleed was 3.81 vs 1.61%, with intracranial bleed 0.72 vs 0.34%; GI bleed 1.26 vs 0.56%
    • Any thromboembolism was 4.41 vs 2.37%, MI 1.90 vs 0.98%, venous thromboembolism (VTE) 0.24 vs 0.09%, and arterial embolism 2.52 vs 1.41% [thromboembolism defined as: stroke, TIA, peripheral arterial emboli, venous thromboembolism, MI]
  • The role of INR variability (dividing those around the median of higher vs lower variability in their cohort): comparing high vs low
    • All-cause mortality was 2.94% vs 1.50%
    • Any major bleed was 3.04 vs 1.47%, with intracranial bleed 0.51 vs 0.38%; GI bleed 1.05 vs 0.50%
    • Any thromboembolism was 3.48 vs 2.46%, with MI 1.53 vs 0.96%, VTE 0.16 vs 0.11%, and arterial embolism 1.98 vs 1.51%
  • For those with TTR >70%, INR variability did not matter
  • For those on aspirin
    • All-cause mortality was 2.57% vs 2.13%
    • Any major bleed was 3.07 vs 2.04%, with intracranial bleed 0.62 vs 0.41%; GI bleed 1.16 vs 0.67%
    • Any thromboembolism was 4.90 vs 2.12%, with MI 1.53 vs 0.96%, VTE 0.19 vs 0.12%, and arterial embolism 2.72 vs 1.54%
  • For those with renal failure, intracranial bleed more than twice as common, with HR 2.25 (1.32-3.82)
  • The strongest predictor of intracranial bleeding was renal failure

So, a few issues:

  • These data, though not from a prospective, randomized study, do reflect more of a real-world community setting.
  • The results for patients in good control (TTR>70%) is actually better than in the “pivotal NOAC studies”.
  • In terms of the issue of using the combo of warfarin and aspirin:
    • I am very concerned about the increased serious adverse events with this combination (see , which looks at several observational studies on aspirin plus warfarin for patients with AF and CAD, all showing much higher risks (e.g., bleeding) without any clear benefit. The PREFER trial (see De Caterina, R. Heart 2014; 100: 1625) looked at a large European registry of patients with AF and found that 95% of those on dual antiplatelet and anticoagulation therapy did NOT have an “accepted indication” (i.e.: acute coronary syndrome or stenting –and, this is the current recommendation of the European Society of Cardiology guidelines in AF: anticoagulant therapy only except in these indications).
    • In the above Swedish study, only 2.6% of those on additional aspirin had a clear indication for this therapy. And this study really supported NOT using aspirin, given the higher mortality, higher MI/VTE/arterial embolism, and much higher bleeding risk (though, as noted, this was not an RCT, had many more patients with prior MI, so it is not so surprising that there was an increase in these thrombotic events. But twice the level of major bleeds???  And 50% more thrombotic events??? Is aspirin really useful?? I personally have stopped aspirin in my patients on warfarin, based on the prior blog)
  • So, my bottom line: this Swedish study confirms the utility of warfarin in patients with nonvalvular AF, though clearly differentiates its advantages in those in good control (INR in range >70% of the time) vs not. I do have trouble with the NOACs overall, given the noted drug company issues in my prior blogs and the lingering concern that if they are in an accident, reversal agents are not widely available. But I have occasionally used NOACs in patients who are leaving the country and cannot get their INR checked, or it is just too difficult to get the INR in range despite best efforts all around, often including home-based nursing care/checking the INR at home. That being said, the clear majority of my patients on warfarin are definitely in the >70% TTR category and seem to be doing just fine…. (though I do need to monitor them frequently, but, then again, most of them have significant cardiac and other morbidities, and my guess is that the increased monitoring/health system contact is actually a positive thing)

CHA2DS2-VASc (cardiac failure or dysfunction, hypertension, age 75 years [doubled], diabetes mellitus, stroke [doubled]–vascular disease, age 65-74 years, and sex category [female])

For older blogs, see: goes through the antithrombotic therapy guidelines but also reviews the history of drug company malfeasance with the NOACs, which really make me uneasy about using them…

Primary Care Corner with Geoffrey Modest MD: Rivaroxaban, Is It Really Better Than Warfarin for Non-valvular Atrial Fib?

16 Feb, 16 | by EBM

By Dr. Geoffrey Modest

The BMJ spotlight team (actually Deborah Cohen, associate editor) just released another investigation into poorly-regulated drug-company-supported studies which have led to large changes in clinical practice and large profits for drug companies (see BMJ 2016;352:i575). This one is a challenge to the ROCKET-AF study (see N Engl J Med 2011;365:883-91), which was the single study leading to FDA approval for rivaroxaban, a​ direct thrombin inhibitor, by showing that although it was as good as warfarin in patients with non-valvular atrial fibrillation in preventing strokes, there were fewer episodes of intracranial and fatal bleeds. Of note, rivaroxaban is now the world’s best-selling new anticoagulant. However, it turns out that the ROCKET-AF study relied on defective point-of-care INR machines which gave falsely low INR levels, thereby leading the researchers to increase the warfarin dosing and perhaps predisposing patients to more bleeding.


  • BMJ found that the INR devices used had been recalled in December 2014 because the FDA issued their highest level recall notice finding that these devices could give “clinically significantly lower” INR values. The device manufacturer Alere had received 18,924 reports of malfunction and 14 serious injuries, dating back to 2002 (prior to the ROCKET-AF trial starting)
  • September 2015: BMJ asked ROCKET investigators about the device recall. No direct response, just one from the drug company that they were unaware of the recall.
  • The European Medicines Agency (EMA, the European equivalent of the FDA) in April 2015 did not know these devices were used in the ROCKET-AF study, and called for an independent investigation into direct oral anticoagulants. The company claimed that the results of ROCKET-AF were still valid after conducting some sensitivity analyses. No independent investigation has happened to date.
  • But, as per Harlan Krumholz, a cardiologist at Yale, “the study should be considered of uncertain validity until a more thorough review can be  done”, and to have “an investigation by an independent group of experts to quickly determine if there are grounds for retraction [of the study]”.
  • In looking back at the original FDA approval for rivaroxaban, 2 primary reviewers had expressed concern about the warfarin control and recommended that the drug not be approved, stating “ROCKET provides inadequate information to assess the relative safety and efficacy of [rivaroxaban] in patients whose warfarin administration can be well-controlled”
  • BMJ asked the ROCKET researchers if those patients who had major bleeds had had laboratory INRs checked. None responded.
  • It turns out that at the 12-week and 24-week time-periods, there were simultaneous measurements of INR by device and a central lab, though this data has not been released.
  • There was a rebuttal by the investigators to the above BMJ charges (see DOI: 10.1056/NEJMc1515842). The investigators did post-hoc analyses using the FDA recall criteria [medical conditions that the FDA thought might lower the INR value of the device: anemia with hematocrit >30%, and conditions which raised fibrinogen levels (acute inflammation, chronic inflammatory conditions, severe infection, cancer or end-stage renal disease]. Results of the post-hoc analysis were mixed:
    • ​In those without recall conditions: no significant difference in strokes with rivaroxaban vs warfarin; more intracranial hemorrhages, fatal or critical organ bleeding with warfarin; more GI bleeding overall with rivaroxaban
    • In those with recall conditions (and presumably on unnecessarily higher doses of warfarin): decreased stroke in those on rivaroxaban but only in the per-protocol patients; no difference in intracranial hemorrhages, fatal/critical organ bleeding; more GI bleeding with rivaroxaban.

So, these articles bring up several issues:

  • It is a little unsettling to me that rivaroxaban was approved based on only one drug company-sponsored study
  • I’m not sure what to make of the post-hoc analysis by the investigators. Surprisingly, they seem to show fewer adverse events in those without the FDA-designated recall conditions, though interestingly in those patients with those conditions, rivaroxaban​seemed a bit worse on clinical bleeding. But are those recall conditions complete? Many older studies, for example, found a relationship between smoking and high fibrinogen levels (to the point that many posited that the main mechanism of smoking causing heart disease was mediated by fibrinogen, and that stopping smoking quickly reversed much of the increased CAD risk because smoking’s effect on fibrinogen levels reversed so quickly). Smoking was not on the recall list. Are there other unknown conditions which lead to aberrancies in the device INRs?Or are there other unknown biases which can happen in a post-hoc analysis? It would be interesting to know the relationship between the laboratory INR measurements at the 12 and 24-week times and the results of the point-of-care devices. And what the correlation was on an individual basis between these discrepancies and the positive and negative clinical outcomes. And, I would strongly support an independent review.
  • A prior BMJ issue and several other studies have suggested that dabigatran also may have many more adverse events than reported and that there were drug company shenanigans; in particular the drug company covered up evidence suggesting that there needed to be blood monitoring, despite the main selling point for the drug being that no drug monitoring was necessary, as opposed to warfarin and the need to follow INRs, (see As a result of the BMJ and other articles about dabigratan, at the end of 2015 both the EMA and FDA held meetings to see if there was a need to monitor blood levels of direct anticoagulants and adjust the dose as necessary to maximize benefits/minimize harms.
  • This raises the issue that the FDA historically has been less aggressive in regulating devices than meds. Typically, once one device is accepted, there is much less overview on similar products (new devices just need to be “substantially equivalent” or similar to ones already on the market). And, the issue of rivaroxaban highlights that approval of meds may well require studies using not-so-well-regulated devices which could potentially dramatically affect the results. And, this is compounded by the selection bias favoring the drugs: positive studies of drugs are published at a much higher rate than negative studies. And once a drug company has a positive study (as with ROCKET-AF), there is little incentive to do further confirmatory studies after a drug is approved.
  • So, yet again, I am very hesitant to prescribe these newer meds, given huge conflicts-of-interest, the large numbers of cases of drug companies hiding results they do not like, the decreasing authority and efficacy of the FDA in monitoring new drugs (e.g., see ), and the increasing reliance on post-marketing surveillance, which seems to happen only pretty rarely (e.g., see ). I continue to use warfarin as my preferred anticoagulant, and I use the close monitoring of INRs (which in rock stable patients, can actually be done on an every 2-3 month basis) as a means to check-in with these patients, who are typically medically-complex anyway.

Primary Care Corner with Geoffrey Modest MD: Antithrombotic Therapy Guidelines

19 Jan, 16 | by EBM

By Dr. Geoffrey Modest

The American College of Chest Physicians just came out with updated guidelines on antithrombotic therapy (see doi 10.1016/j.chest.2015.11.026). Details of those recommendations that are important for primary care providers (i.e., not including those critically ill, use of thrombolytic therapy, catheter-based thrombus removal):

  • They suggest using dabigatran, rivaroxaban, apixaban, or edoxaban over vitamin K antagonists (VKA, such as warfarin) for those with venous thromboembolic disease (VTE) and no cancer (all grade 2B recommendations, where Grade 2 recommendations are “weak recommendations”, based on “moderate” evidence = Grade B)
  • For those with VTE and cancer: low molecular weight heparin (LMWH) preferred over dabigatran/rivaroxaban/apixaban/edoxaban – i.e. NOACs  (non-vitamin K oral anticoagulants)  (grade 2C, where C=”low quality evidence”)
  • For those with VTE and on anticoagulants, do NOT use IVC filter — Grade 1B (strong recommendation)
  • Recommends against using compression stockings routinely to prevent post-thrombotic syndrome (PTS) — grade 2B [a large study did not find stockings decreased PTS or reduce leg pain, though they may be useful for chronic symptoms of PTS]
  • For those with subsegmental PE and no proximal DVT, they suggest clinical surveillance over anticoagulation if a low risk of recurrent DVT (grade 2C), but anticoagulation if high risk (also grade 2C) [2 issues here: these are often picked up by CT scan and can be false positives; and if real, are likely from small DVT and risk of recurrence or progression is small, but there are NO RCTs on this. reasonable to do bilateral leg ultrasounds, and upper extremity one if there is a central venous catheter]
  • For those with recurrent DVT while on non-LMWH therapy, then give LMWH (grade 2C); and if already on LMWH, then increase the dose (grade 2C)

In the body of the article, they note:

  • For those with proximal DVT or PE and nonsurgical transient risk factor, anticoagulate for 3 months
  • For patients with distal DVT provoked by surgery or nonsurgical transient risk factor, anticoagulate for 3 months [they note that there is a 15% risk of extension to the popliteal vein, and note that the decision to anticoagulate is not clear, with some providers suggesting anticoagulation, others only if there is proximal propagation in 1-2 week followup ultrasound. They also comment that a risk factor for extension is a positive D-dimer, especially if very high (not specified further), if thrombosis is multiple or >5cm long, if close to proximal vein, if active cancer, if history VTE, also if DVT not in the muscular veins of the soleus or gastrocs).
  • In those with unprovoked DVT (isolated distal or proximal DVT/PE), they recommend 3 months of anticoagulation. Re-evaluate the patient after 3 months for risk-benefit of continued therapy. Later in the text of the article, they comment: in those with isolated acute distal DVT and without severe symptoms or risk factors for extension (see above), do serial imaging of deep veins. If the thrombus extends, even within the distal veins, then anticoagulate
  • But, in those with a first VTE which is an unprovoked proximal DVT/PE and low to mod bleeding risk, give extended anticoagulation without stop date. If high risk of bleeding, then 3 months of anticoagulants (they do comment that d-dimer levels 1 months after stopping therapy might influence the decision to continue therapy)
  • Those with a second unprovoked VTE and low bleeding risk, extended therapy without stop date; high risk = 3 months. Re-evaluate periodically, such as annually
  • Those with DVT/PE and active cancer really should get anticoagulant therapy without stop date, but re-evaluate annually
  • In those with unprovoked proximal DVT/PE and the decision is to stop anticoagulants, you should give aspirin if no contraindication
  • In those with upper extremity DVT involving axillary or more proximal veins, anticoagulate
  • In terms of the choice of NOACs vs VKAs, they comment that the efficacy is similar though the risk of bleeding is less with NOACs (esp. intracranial), though in patients with atrial fibrillation, there is more GI bleeding with NOACs
  • Of note, of the 54 recommendations in the 30 statements in the article, “20 were strong and none were based on high quality evidence highlighting the need for further research” [my highlight]
  • Their estimates of risks of recurrent VTEs are:
    • VTE provoked by surgery: 3%  at 5 years
    • VTE provoked by non-surgical transient risk factor (e.g. estrogen therapy, pregnancy, leg injury, flight>8hrs,: 15% at 5 years
    • Unprovoked VTE: 30% at 5 years
    • Cancer-associated thrombosis: 15%/year
  • In terms of determining length of therapy (and leading to above recommendations):
    • An isolated distal DVT has 1/2 the risk of a proximal one for PE
    • A second unprovoked proximal DVT/PE increases the risk by an additional 50%
    • Categories for bleeding risk: [bleeding risk, per them, is similar to the HAS-BLED algorithm, including age (esp>75, though some increase >65), cancer, liver/kidney dz, diabetes, antiplatelet drugs, NSAIDs, alcohol, frequent falls]
      • Low (no bleeding factors): 0.8%/yr
      • Moderated (one bleeding factor): 1.6%/yr
      • High (2 or more bleeding factors): >6.5%/yr
    • ​A common issue is the patient with an unprovoked proximal DVT/PE without high risk of bleeding. They suggest that men have 75% increased risk of recurrence (the risk of recurrence in women is about 15% at 5 years) and those with positive D-dimer measured 1 month after stopping anticoagulation doubles the risk [I would add that there also was a study looking at D-dimer prior to stopping, finding that a raised level also predicted further VTEs. So I have been using both: D-dimer prior to stopping, and if normal, then 1 month after stopping. And I have been prescribing aspirin therapy in those who stop the anticoagulation, which has been further validated by 2 trials, with 35% reduced risk].
  • Outpatient treatment of acute PE: there are studies (not involving patients treated outside the hospital) finding that NOACs seem okay without initial heparin therapy

So, because of newer studies finding that NOACs seem to work well in patients with VTE, including in the setting of active cancer or recurrent VTE, and with the reported decrease in bleeding, the main change in the new guidelines over the previous ones is the elevation of NOACs over VKA’s. However, I would add that they do hedge on the quality of this data (none of these recommendations are Grade A) as noted above, and I am very concerned that many of those involved in the guidelines were involved in the NOAC studies with drug company support, and that there have been very real concerns about serious adverse events with NOACs and with drug company malfeasance/withholding damning data in their reporting (I am appending old blogs, some predating their posting on  BMJ, with drug company data itself suggesting that drug levels for dabagratan for example be monitored (one of the major selling points of NOACs is the lack of necessity to monitor levels, as INRs with VKAs).


1/12/12: Recent article which found an increase in MI/ACS (acute coronary syndromes) in meta-analysis of dabagatran use (see doi:10.1001/archinternmed.2011.1666). This was large meta-analysis (30,514 pts) on dabagatran in a variety of trials (2 on afib, 1 acute dvt prophylaxis, 1 ACS, 3 on short-term prophylaxis of DVT) comparing dabigatran to warfarin/enoxaparin or placebo, which found a small but significant increase in MI/ACS (30% relative risk increase, though only about 0.17% absolute increase from 0.79% to 1.19%, OR=1.33 (1.03-1.71, p=0.03). Not sure how to explain this since it works in preventing emboli in afib, though there was a paper reporting increased inflammatory markers with the use of dabagatran.

Brings up the all-to-frequent medical scenario. New drug comes to market. Works in preliminary studies (pretty much all funded by drug company). Drug makes sense to us (using a direct thrombin inhibitor should decrease thrombotic events, much easier to use than warfarin — no need for INR monitoring). BUT…. Subsequent studies start to find problems, especially when looking at other than the marker initially studied (thromboemboli in afib, preventing thrombosis in pts with joint replacement, acute DVT)


8/9/12: Recent editorial in annals on large number of adverse effects of dabigatran (see Annals Intern Med 2012: 157, 66). In its first 6 months, 505 adverse events from hemorrhage with 65 deaths, vs 176 for warfarin. In initial 6 weeks of marketing, “the Institute for Safe Medication Practices reported that dabigatran was responsible for more serious adverse events than 98.7% of all medications”. Reinforces importance not to jump on bandwagon and use new meds (esp given that vast majority of studies done by the pharmaceutical companies who stand to benefit….).  A recurrent theme.​


9/25/12: meta-analysis of 7 prospective randomized controlled trials see doi:10.1001/archinternmed.2011.1666

“For the period January 1, 2000, through December 31, 2011, the researchers identified 7 prospective randomized placebo-controlled clinical trials that met the study criteria, involving 31 286 patients. Based on the pooled results, the use of new-generation oral anticoagulant agents in patients receiving antiplatelet therapy after an ACS was associated with a dramatic increase in major bleeding events (odds ratio, 3.03; 95% CI, 2.20-4.16; P < .001). Significant but moderate reductions in the risk for stent thrombosis or composite ischemic events were observed, without a significant effect on overall mortality. For the net clinical benefit, treatment with new-generation oral anticoagulant agents provided no advantage over placebo (odds ratio, 0.98; 95% CI, 0.90-1.06; P = .57).”


7/28/14 blog (for some reason not in my BMJ blogs): the BMJ’s last issue had 6 articles on dabigatran (Pradaxa), showing the malfeasance of the drug company in promoting this agent [dabigatran is a direct thrombin inhibitor, specifically marketed as a competitor of warfarin for nonvalvular atrial fibrillation, though promoted for other indications as well, with the specific selling points based on a single key trial (RE-LY, see critique below) finding slightly lower incidence of major bleeding (16.4% vs 18.15%) without the costly and inconvenient monitoring of INRs required for warfarin – i.e., the high cost of the drug was more than justified by the savings from office visits and laboratory monitoring for warfarin. the following issues were discovered from the drug maker (Boehringer Ingelheim), largely through internal previously confidential documents:

  • The RE-LY trial was felt to have poor design and oversight, with a Canadian agency calling for “an independent audit of RE-LY to check for irregularities in conduct, sources of bias (this was an open-labeled, not blinded trial) and the cause of the unusually high incidence of intracranial hemorrhage in the warfarin arm” (thereby exaggerating the benefit of dabigatran), requesting patient level data to be made public. The FDA specifically was concerned about the likely high frequency of errors in the data set (misreporting of events), with subsequent FDA-mandated review by the company finding a further 3848 events. Of note, the increase in new MIs (n=270) found in the dabigatran​ group was dismissed as due to chance because of small numbers, but the increase in warfarin-associated intracranial hemorrhages was highlighted, even though there were many fewer of such events (n=155). the FDA found that the “blinded adjudicator of events” in fact was unblinded in 20% of the cases (e.g., by seeing identifying patient information)
  • ​The drug company has settled for $650M in litigation because the evidence provided by the company was incomplete.  The lawsuit was from fatal hemorrhage cases in the RE-LY trial, cases which it turns out were not included in the RE-LY trial’s initial compilation of adverse events, nor in the FDA-mandated recalculation, which it turns out was conducted by company scientists!!  Previously unreleased documents came to light through this litigation and freedom of information act
  • The drug company noted in internal documents that the risk of bleeding could have been reduced by 30-40% if plasma levels of dabigatranwere in fact monitored (company’s own unreleased data), and with appropriate dosage adjustment (i.e. undercutting their main selling point for the drug that there didn’t need to be drug level monitoring). Internal emails in the company released in the litigation showed that the company did not want this information released. Both the FDA and EMA (European medicines agency, the European FDA equivalent) specifically inquired about monitoring drug levels to improve safety, and one cardiologist, an FDA advisor, expressed concern that there was a 5-fold variability in plasma levels on the 150-mg dose. With regard to this marked plasma variability, one of the company’s own scientists commenting on the potential safety issues from not checking plasma levels was rebuffed by a company official noting in a 2012 internal email that “the publication [of this result] will [do] more harm than be useful for us, neither in the market but especially harmful in the discussions in the regulatory bodies”, and further, “the world is crying for this information — but the tricky part is that we have to tailor the message smart.”  And “this will make any defense of no monitoring [of plasma levels] to [health authorities] extremely difficult and undermine our efforts to compete with other [new oral anticoagulants].”
  • ​And, one of the concerns with bleeding with dabigatran​ is that there is no antidote, unlike vitamin K with warfarin overdoses [there now is a reversal agent available]


For 2 of the most relevant BMJ articles, see:  doi: 10.1136/bmj.g4517, and doi: 10.1136/bmj.g4670

  • The big issue here, as mentioned in several prior blogs including a few on dabigatran​ is that we are increasingly propelled into an era of privatization and deregulation, beginning with the Reagan presidency and leading to the situation now where almost all of the studies in major medical journals are funded by the drug companies (in the 1970’s the vast majority were publicly-funded), a large percentage of guidelines now are from specialty societies (as opposed to the recent past when the NIH led most of the guideline development) and many of the leading specialists are directly supported by drug companies, and where the cash-strapped FDA is increasingly making drug companies do further studies instead of independent groups. My own bias is that we should be very hesitant in the current climate to be early-adopters of new medications, unless there is a compelling need (for example, I think there is a compelling case for the new oral hep c drugs, or new and better-tolerated hiv meds)
  • Also, a recent report from 2015 which looked at Medicare claims from 2010-2011, looking at those with newly diagnosed atrial fibrillation who were put on dabigatan or warfarin within 60 days of the diagnosis (see JAMA Intern Med. 2015;175(1):18-24). They found that dabigatran (vs warfarin) was associated with
    • Higher risk of any bleeding event, with HR 1.30 (1.20-1.41)
    • Higher risk of major bleeding, with HR 1.58 (1.36-1.83)
    • Higher risk for GI bleeding, with HR 1.85 (1.64-2.07)
    • Though, lower risk of intracranial hemorrhage, with HR 0.32 (0.20-0.50)  [the incidence of intracranial bleeds was about 1/11th that of GI bleeding, though intracranial bleeds have more morbidity/mortality; no clear explanation of why this difference. Though see above about RE-LY trial and reporting issues]
    • And, the results of increased bleeding were especially in African-Americans and those with chronic kidney disease

See for more info

So, bottom line: I’m just not sure I trust the NOACs given all of the above. I have continued with warfarin, with no real issues/adverse events (with the one exception of a patient controlled on 1mg of warfarin, admitted to the hospital for chest pain, seen by cardiology, but patient had missed a dose of warfarin that day and her INR was a little low, and the cardiol fellow increased the warfarin to 2mg, noting it was “only 1 mg more”, and one week later she bled out from a GI hemorrhage. Now, I always check in with my warfarin patients when discharged from the hospital just to make sure things were not changed too much, and provide very close follow-up)

Primary Care Corner with Geoffrey Modest MD: Resume Anticoagulation After GI Bleed?

7 Jan, 16 | by EBM

By Dr. Geoffrey Modest

One of the many unanswered clinical questions is what to do with patients who have atrial fibrillation, are anticoagulated, but have a GI bleed. Do I just stop the anticoagulation? Should I bite my lip (but not too hard) and reinstate anticoagulation after the bleed? A recent observational study looked at this question, utilizing the great Danish clinical database (see BMJ 2015 Nov 16; 351:h5876​).


  • Danish cohort included all patients with atrial fibrillation (AF) from 1996-2012 who had a gastrointestinal (GI) bleed while on antithrombotic therapy. They then compared the patients who restarted therapy vs those who remained off therapy, beginning 90 days after the incident bleed
  • 4406 patients (mean age 78; 45% women; 24% on oral anticoagulation, 53% on antiplatelet agents, 19% on both; 25% on NSAIDs within 90 days of bleed, 15% on PPIs; 23% with prior stroke, 38% ischemic heart disease, 31% heart failure, 45% hypertensive, 20% vascular disease, 16% diabetic, 5% “alcohol misuse” as identified on admission for the GI bleed
  • Mean CHADS2 score of 2.1 (score of 1 means=moderate embolic risk, >=2 means mod-to-high risk), mean CHA2DS2-VASc score of 3.6 (score of >=2 means mod-to-high risk), mean HAS-BLED score of 2.6 (a composite of pro-bleeding factors such as hypertension, abnormal renal/liver function, labile INR, elderly, drugs/alcohol; where score >=3 means high risk of bleeding). Of note, there was really no difference in any of these scores in comparing those patients put back on oral anticoagulants or no meds. And no difference in whether they got gastroscopy or surgery, with 88% having gastric or duodenal ulcer, 12% gastritis, 2% GERD; and 5% with alcohol “misuse”, 90% on PPIs, 5% NSAIDs
  • 924 patients (27%) did not resume antithrombotic therapy (note: this is a large database study and not clear why the clinical decision was made to continue or stop therapy)

Results (I am only reporting results in those on single therapy with oral anticoagulant or antiplatelet agent):

  • In terms of absolute numbers, over a 2-year period, all-cause mortality of the whole group was really high at 40% (n=1745); thromboembolism was 12% (n=526); major bleeding was 17.7% (n=788) and recurrent GI bleed was 12.1% (n=546)
  • Comparing those who did not resume therapy with those who did:
    • For anticoagulant therapy (92% on vitamin-K antagonists):
      • All-cause mortality had HR=0.39 (0.34-0.46) – i.e. 61% reduction
      • Risk of thromboembolism had HR=0.41 (0.31-0.54) – i.e. 59% reduction
      • Risk of major bleed had HR=1.37 (1.06-1.77) – i.e. 37% increased risk, though a nonsignificant 22% increase in recurrent GI bleeds [HR=1.22(0.84-1.77)] perhaps because 90+% were on a PPI
    • For antiplatelet therapy (98% on aspirin):
      • All-cause mortality had HR=0.76 (0.68-0.86) – i.e. 24% reduction
      • Risk of thromboembolism had HR=0.76 (0.61-0.95) — also 24% reduction
      • Risk of major bleed (beginning 90 days after discharge from first bleed) had HR=1.25 (0.96-1.62 – i.e. nonsignificant 25% increased risk)
    • Subgroup analysis showed that as the CHA2DS2-VASc increased, there was associated decreased risk of all-cause mortality in those on anticoagulants (56% in those with CHA2DS2-VASc score of <2; 60% if 2-3; and 63% if >3); a HAS-BLED score >3 was associated with increased bleeding risk (41% if HAS-BLED score of <2; 64% if 2-3; and 57% if >3)
    • The single most effective change was taking patients who had been on an antithrombotic regimen (mostly aspirin) prior to the initial GI bleed and switching them to an oral anticoagulant (mostly vitamin K antagonist) after the bleed.

So, a few observations:

  • 25% of the patients with initial GI bleed were on NSAIDs. This is pretty clearly not a good thing, though hard to control completely given the easy availability of NSAIDs (I have had several patients on warfarin where I have explained the risk of NSAIDs, including naming the OTC brands, who subsequently used Advil or other OTC NSAIDs thinking they were okay. It probably is worth reiterating this message several times: by being OTC does not mean the drug is safe). Alcohol is another bleeding risk factor, which in those who drink excessively not only increases bleeding risk but also the risk of not taking anticoagulants correctly
  • I’m not sure what to make of the fact that many more of these patients were on antiplatelet agents (39% as single therapy) than anticoagulants (21% as single agents), given the high CHA2DS2-VASc scores overall (in general, the preferred therapy for those with CHA2DS2-VASc​ of 2 or more is oral anticoagulants)
  • Though they chose a waiting period of 90 days after the incident bleed to start their assessment of outcomes, there was no difference in sensitivity analysis if look at outcomes starting the day after the bleed
  • Although this was not a randomized controlled trial, there are several issues which I think make it still an important trial: this study involved a large number of patients drawn from a national registry which included all Danish residents and with linkage to pharmacy registries; there were minimal differences in the thrombotic risk (e.g. CHA2DS2-VASc score) or propensity to bleeding (HAS-BLED score) between the groups who resumed therapy and those not on therapy (i.e. no obvious selection bias); and there was a high absolute number of bad events overall and very high all-cause mortality benefit by reinstating anticoagulation. All of this suggests that there are reasonable grounds to restart oral anticoagulants, again with the proviso of trying to avoid NSAIDs and alcohol, and with close patient follow-up. Though it would be great to have a randomized controlled trial.

See for many blogs on AF.

Also, another from the Danish registry looking at patients without AF but with heart failure and high CHA2DS2-VASc score, showing a high incidence of thromboembolism (see ).

Primary Care Corner with Geoffrey Modest MD: Pulmonary Embolism Evaluation

16 Oct, 15 | by EBM

By Dr. Geoffrey Modest

The Clinical Guidelines Committee of the Am College of Physicians published a paper detailing what they consider to be “best practice advice” for the evaluation of patients with suspected acute pulmonary embolism –PE– (see doi:10.7326/M14-1772), focusing on what they see as the overuse of CT scans and plasma D-dimer tests.

Their advice:

  • Use validated clinical prediction rules to estimate the pretest probability of PE (e.g. Wells prediction rules or revised Geneva score, included below)
  • Patients who have low pretest probability of PE and who also meet all of the Pulmonary Embolism Rule-Out Criteria (PERC, included below) should NOT have a D-dimer done
  • Patients with an intermediate pretest probability of PE, or low pretest probability but do not meet all of the PERC, should have a high-sensitivity D-dimer done
  • In patients over 50yo, use age-adjusted D-dimer thresholds (agex10ng/ml, rather than a universal cutoff of 500 ng/ml)
  • In those with D-dimer levels below the above threshold, do NOT do any imaging (studies show way too many are done in ERs with no improvement in patient outcomes, though lots of radiation exposure and expense (see more reviews on excesses of radiation exposure)
  • In patients with high pretest probability of PE, go directly to CTPA (CT pulmonary angiography). Use ventilation-perfusion scans only if contraindication to CTPA or CTPA not available. Do NOT get D-dimer​ in addition

A few supportive points:

  • D-dimer has low specificity, but a normal high-sensitivity D-dimer in recent studies had a 99.5-100% sensitivity for excluding PE on CT scan
  • The PERC (which is not a screening tool for all patients but only those with signs/symptoms potentially suggestive of PE) in a large meta-analysis has found that the overall proportion of missed PEs was 0.3%, with pooled sensitivity of 97%. In those with score of “0”,  they state “the risk for PE is lower than the risks of testing”
  • Of note, pregnancy, heart failure or stroke are not part of the Geneva score because they do NOT add to its predictive performance
  • They do stress that alternative approaches may be okay: e.g., if someone has concerns for PE, are hemodynamically stable, and have lower extremity symptoms, it is reasonable to get leg ultrasound, and anticoagulate if positive (i.e., spare the radiation exposure, since the patient will be treated anyway). Though one issue is length of therapy here. Those who get a PE are at higher risk for a recurrent PE than those who just have a DVT. They comment that for patients with “cardiothoracic symptoms, the need for long-term anticoagulation can be determined after the initial treatment period”. I’m not sure what that means exactly. As I have mentioned in previous blogs, there are some interesting data supporting either checking D-dimer levels before stopping therapy or 3-4 weeks after stopping it (I actually do both), and in those with normal D-dimers one can stop therapy (this is not accepted universally as strategy, in part because those with normal D-dimers can still get PEs, though the studies suggest they are much less likely. I do discuss the risks and benefits of stopping anticoagulation with patients and try to make a joint strategy) — See​ for more detailed articles/critiques.



Wells prediction rules:


Revised Geneva score:




Primary Care Corner with Geoffrey Modest MD: Extended anticoag for PE

28 Jul, 15 | by EBM

By: Dr. Geoffrey Modest

JAMA published the PADIS-PE trial of patients having a first unprovoked pulmonary embolism (PE), randomized to stopping anticoagulation after 6 months vs continuing an additional 18 months (see JAMA. 2015;314(1):31-40​).


 –371 patients (mean age 58, 40% >65yo, 50% women, mean BMI 27, 45% with high bleeding risk per the Am College of Chest Physicians rating) with a first episode of unprovoked, symptomatic PE, initially put on 6 months of a vitamin-K antagonist, who were then randomized from 2007-2014 in 14 French centers to continued warfarin vs placebo for 18 months. Target INR = 2.0-3.0. Patients with known major thrombophilia were excluded. Blood was sent for thrombophilia workup on day 0.

–minor thrombophilia (heterozygous factor V Leiden or heterozygous G20210A prothrombin mutation, or factor VIII >90th %) was found in 24% and major thrombophilia (antithrombin deficiency, anticardiolipin antibodies >99th %, or homozygous factor V Leiden or combined thrombophilia) in 4%.

–primary outcome: composite of recurrent venous thromboembolism (VTE) or major bleeding at 18 months; secondary outcome: this composite at 42 months, each component of the composite, and death unrelated to PE or bleeding at 18 and at 42 months (24 months after the end of the study).


–Mean % of time in the target INR range was 69.1%.

–During the treatment period

–primary outcome in 6/184 patients in the warfarin group (3.3%, or 2.3 events per 100 person-years) vs 25/187 in placebo (13.5%, or 10.6 events per 100 person-years), with HR 0.22 (0.09-0.55, p=0.001).  

–in the warfarin group, 3 patients (1.7%) had a symptomatic recurrent VTE, all after discontinuation of warfarin. in the placebo group, 25 patients (13.5%) had symptomatic recurrence, only one of which was after stopping the placebo (p<0.001).

— major bleeding occurred in 4 patients in warfarin group and 1 in the placebo group (statistically nonsignificant)

–After treatment discontinuation

–composite outcome in 27 warfarin-treated (17.7%; 10.0 events per 100 person-years) and 17 placebo-treated (10.3%, 5.7 events per 100 person-years)

–in the warfarin group, 25 patients had symptomatic VTE (9.3 events per 100 person-years, all in the absence of anticoagulation, 4 of which were fatal). in the placebo group, symptomatic recurrence was in 14 patients (4.7 events per 100 person-years, all in the absence of anticoagulation)

–major bleeding occurred in 2 patients on warfarin group (1 while taking warfarin and 1 fatal event happened after stopping it), and 4 in the placebo group (all nonfatal; 2 while on warfarin)

–Review of the graphs showed a dramatic difference during the 18 months after randomization, and a gradual reduction in warfarin protection over the next 24 months of observation when all were off warfarin (see below for the primary endpoint)


​–After 42 months, composite outcome in 33 patients on warfarin (20.8%) and 42 on placebo (24.0%)  [HR 0.75 (0.47-1.18)] —  the benefit of extended warfarin Rx was not maintained

–No difference in rate of recurrent VTE, major bleeding, and unrelated deaths

–The risk of recurrent VTE was greatest in the first 6 months after stopping anticoagulation, then increased linearly by 4-5% per year.  the risk of major bleeding with continued warfarin was low, increasing by <2% per year

So, a few issues: 

–this study confirms that when anticoagulation is discontinued, the patient at is at risk for recurrent VTE, whether anticoagulation is for 3-6 month or longer. Also that the risk of major bleeding with anticoagulation is pretty low (confirmed by other studies)

–in the decision analysis of how long to anticoagulate, it is important to remember that the natural history of a venous thrombosis is different from a PE: the likelihood of a recurrent PE is much higher in the latter group, and the case-fatality rate is 4-fold higher than after a proximal DVT. In the above study, 80% of the recurrent VTE events in both groups were from  symptomatic PEs.

–if you and patient decide to stop anticoagulants, there are some potentially useful approaches: for other articles, including one on different and safer ways to stop anticoagulation by checking d-dimer levels, see here and especially here)

–if you decide to stop the anticoagulation, 2 studies have found some efficacy of low dose aspirin (about 30% decreased risk of aspirin vs placebo — better than placebo but not as good as continuing the anticoagulation).

–an (unfortunately) untested hypothesis might be: check for underlying thrombophilia; 6 months of anticoagulation; check d-dimer; in those who are d-dimer negative, offer aspirin and discontinue anticoagulation (after discussing risks/benefits with patient); also look to see if the underlying thrombophilia actually matters (and, if so, then preferentially continue life-long anticoagulation). Short of that, this study adds to other observational studies finding a high risk of recurrent PEs in those with initial symptomatic unprovoked PE, so my inclination (and what I have been doing) is to offer life-long anticoagulation (especially to low risk-of-bleeding patients), with a fallback to checking the d-dimer and changing to aspirin if all is okay.


Primary Care Corner with Geoffrey Modest MD: Bridge therapy for patients on warfarin and invasive procedures

10 Jun, 15 | by EBM

By: Dr. Geoffrey Modest 

A pretty common primary care conundrum is what to do with patients who are on warfarin but have impending surgery: stop the warfarin and hope they don’t get a pulmonary embolus, or do bridge therapy (there are several different ways to do the bridging: for those at high risk of venous thromboembolism –VTE, full-dose bridging is often done: stopping the warfarin 5 days prior to surgery, starting low molecular-weight heparin soon thereafter, stopping that for the surgery, and then restarting the heparin soon after the surgery (can be 1 day in those at low risk of post-op bleeding, longer if at higher risk); can do prophylactic low-dose heparin post-op; or can do intermediate (about 1/2 full-dose) for those with higher risk (eg VTE within the past month). Warfarin is resumed 12-24 hours post-op and usually takes 4-6 days to achieve target anticoagulation. A recent study in JAMA Internal Medicine looked at the outcome of 1178 patients in a retrospective cohort study in a large HMO (Kaiser Permanente Colorado) from 2006-2012 who were on warfarin for a VTE and had surgery with or without bridging (see doi:10.1001/jamainternmed.2015.1843​). They looked at the 30-day occurrence of clinically relevant bleeding, recurrent VTE, and all-cause mortality. Details:

 –1178 patients: mean age 66, 46% men, 56% on warfarin for deep venous thrombosis (dvt) and 44% for pulmonary embolism, 10 % with laboratory-confirmed thrombophilia (Protein S or C deficiency, antithrombin deficiency, antiphospholipid antibody syndrome, factor V Leiden, prothrombin 20210 mutation, increased factor VIII).

–Per the American College of Chest Physicians guidelines (AT9 guidelines, see CHEST 2012; 141(2)(Suppl):e326S–e350S), 79% of patients were considered low risk for VTE recurrence at the time of surgery, 18% moderate and 3% high risk (see below for risk stratification), and bridge therapy was done, respectively, in 29%, 34%. and 63% of the surgeries. 73% used therapeutic doses in the bridge therapy and 28% used prophylactic doses

–Procedures done: 37% GI endoscopies, 14% orthopedic, 10% spinal/intracranial, 9% abdominal. Of the surgeries, bridge therapy was performed more often for orthopedic and nonmajor abdominal or thoracic procedures


–clinically relevant bleeding within 30 days occurred in 15 patients (2.7%) in the bridge therapy group , but only in 2 patients (0.2%) without bridge therapy [(HR=17.2 (3.9-75.1)]

–high VTE risk group–5.6% with bridge therapy, 4.8% without. nonsignificant

–moderate VTE risk group–4.6% with bridge therapy, 0% without. p=0.004

–low VTE risk group–2.0% with bridge therapy, 0.1% without. p<0.001

–no significant difference in bleeding between those receiving therapeutic vs prophylactic doses in the bridge therapy

–no significant difference between groups in the rate of recurrent VTE in the bridge vs non-bridge groups (0 vs 3 patients, p=0.53)

–high VTE risk group–0% with bridge therapy, 0% without. nonsignificant

–moderate VTE risk group–0% with bridge therapy, 0.5% without. nonsignificant

–low VTE risk group–0% with bridge therapy, 0.2% without. nonsignificant

—no deaths in either group

General comments:

–it makes sense to stratify patients by surgical risk of bleeding: low risk (eg dental work)-continue with warfarin and therapeutic INR; moderate to high risk- stop the warfarin 5 days before the surgery (per AT9)

–the rationale for bridge therapy is two-fold: to decrease the amount of time the patient is not anticoagulated overall (around 7-10 days), but especially to do so in the high risk post-op period for VTE, since surgery leads to a more hypercoagulable state.

–the 2012 AT9 guidelines classifies the risk of recurrent VTEs in the post-op period as high (>10%/yr –eg, VTE within 3 months, severe thrombophilia), moderate (5-10%/yr –eg, VTE in past 3-12 months, non-severe thrombophilia such as heterozygous for factor V  Leiden or prothrombin gene mutation) and low (<5%/yr — eg, VTE>12 months previous and no other risk factors). This risk stratification is based on the annual risk of recurrent VTE in patients with mechanical heart valve, chronic atrial fibrillation or VTE who were either on no anticoagulation or less-effective treatment (eg aspirin).  ie, there is only indirect evidence to establish these risk categories for those with VTE on therapeutic anticoagulation.

So, this study is large but limited by being retrospective, and only a small % of the patients were high risk. However, it is impressive that bridge therapy made no difference in VTE occurrence in the much larger moderate and low risk groups, though there was a much higher risk of clinically relevant bleeding. We really need a prospective study, randomizing patients in the different risk categories to bridge therapy or not.  A prospective study could also sort out if there were other risk factors to consider which might favor bridge therapy, including those with a recent VTE (the risk is highest in the first 4 weeks post-VTE, 0.3-1.3%/day, decreasing to 0.03-0.2%/day over the next 4-12 weeks), or in those with several previous recurrences. also, we should note that this study did not include patients anticoagulated for atrial fibrillation or other indications (and it would be great to have good data on that… currently,  AT9 considers those in atrial fib with CHADS2 score of 0-2 as low risk, 3-4 as moderate, and 5-6 as high risk. Also high risk is if there is rheumatic heart disease or stroke/TIA within the past 3 months). But at least the current study gives some pretty strong support for withholding bridge therapy in those who are at least a few months out from their VTE event and who are in low or moderate VTE risk.


EBM blog homepage

Evidence-Based Medicine blog

Analysis and discussion of developments in Evidence-Based Medicine Visit site

Creative Comms logo

Latest from Evidence-Based Medicine

Latest from EBM