Anticoagulation for superficial vein thrombosis: NEJM reports important study and New York Times misses the boat as usual

First let's take a look at the study:

BACKGROUND

The efficacy and safety of anticoagulant treatment for patients with acute, symptomatic superficial-vein thrombosis in the legs, but without concomitant deep-vein thrombosis or symptomatic pulmonary embolism at presentation, have not been established.

...we assigned 3002 patients to receive either fondaparinux, administered subcutaneously at a dose of 2.5 mg once daily, or placebo for 45 days. The primary efficacy outcome was a composite of death from any cause or symptomatic pulmonary embolism, symptomatic deep-vein thrombosis, or symptomatic extension to the saphenofemoral junction or symptomatic recurrence of superficial-vein thrombosis at day 47.

RESULTS

The primary efficacy outcome occurred in 13 of 1502 patients (0.9%) in the fondaparinux group and 88 of 1500 patients (5.9%) in the placebo group (relative risk reduction with fondaparinux, 85%; 95% confidence interval [CI], 74 to 92; P less than 0.001). The incidence of each component of the primary efficacy outcome was significantly reduced in the fondaparinux group as compared with the placebo group, except for the outcome of death (0.1% in both groups). The rate of pulmonary embolism or deep-vein thrombosis was 85% lower in the fondaparinux group than in the placebo group (0.2% vs. 1.3%; 95% CI, 50 to 95; P less than 0.001).

The New York Times article, apparently echoing a related NEJM editorial in the same issue, criticized the authors for ignoring cost considerations. That's not entirely fair, because the authors of the paper raised the cost effectiveness red flag themselves by stating that the number needed to treat to prevent a major VTE was 88. You wouldn't even have to do the math to know that's pretty expensive.


But where the NYT article really blew it was in its total lack of perspective. A little background and a more in depth analysis of the study would have been helpful. According to the study findings the rate of progression to major VTE (DVT or PE) untreated was 1.3% at 47 days. That can be extrapolated to about 10% at one year---maybe a little less assuming that the risk wanes over time, but it's still quite a bit, and enough to be considered a state of extreme hypercoagulability by any standard. That finding shouldn't be surprising. Superficial vein thrombosis, when not caused by an IV catheter or varicose veins, is a known red flag for the hypercoagulable states that underly some forms of cancer and vasculitis.


One extremely important point the NYT piece ignored was that the NEJM study was not looking at therapeutic anticoagulation. The dose of fondaparinux used in the study, 2.5 mg daily, was the VTE prophylactic dose of fondaparinux. And the endpoints under discussion in this study are not resolution of the condition being addressed (superficial thrombosis) but prevention of more significant events. So in evaluating this study and comparing the results to what's already known it is important to recognize that this paper was an evaluation of VTE prophylaxis.


So from a cost effectiveness standpoint how does this use of fondaparinux stand up to other data on VTE prophylaxis? If you compare it with medical prophylaxis in hospitalized patients, it holds up quite well. Although pharmacologic VTE prophylaxis in hospitalized medical patients is widely accepted and emerging as a standard, the bang for the buck, when you look at published studies, is quite low. A large repository of studies on this topic can be found in the Chest guideline article on medical VTE prophylaxis. When examining this literature it is essential to note whether symptomatic or asymptomatic VTEs were reported. Symptomatic VTE represents the tip of the iceberg of the total VTE burden. The NEJM paper looked at only symptomatic events. Had they reported asymptomatic events based on screening via compression ultrasound or venography, as many prophylaxis studies have done, the baseline risk would have been much higher. The number needed to treat would have been substantially lower. That would have made the cost effectiveness analysis much more favorable. It's almost certain that the cost to prevent one event using such data would be below $50,000, which is the threshold for cost effectiveness by the usual analysis.


The essential question raised by this analysis is whether asymptomatic DVT is an “outcome that matters.” After all, it's more of a surrogate endpoint than a clinical one. If asymptomatic DVT is not an outcome that matters then you have to go back and question much of the vast literature that's been cited by guideline authors, not to mention guidelines themselves which recommend VTE prophylaxis in hospitalized patients.


For example, in this meta-analysis from Annals of Internal Medicine looking at VTE prophylaxis in medical patients the number needed to treat to prevent one symptomatic DVT was 233. And these were higher risk patients, with all 9 studies looking at select groups, and most studying patients that corresponded to those recommended for prophylaxis in current guidelines. Well, that doesn't make the NNT of 88 look so bad, does it?


Why did the New York Times leave this information out? The cynical side of me would say they wanted to frame this as a case of disease mongering by the pharmaceutical industry, but I'll give them the benefit of the doubt. It's probably more a matter of ignorance and sloppy reporting, as evidenced by their failure to take into account any of the background I've cited above.


A final point bears mention. The NEJM study provides information that can change practice without escalating costs because low molecular weight heparin (Lovenox) has gone generic. You wouldn't be on thin ice to extrapolate the study findings regarding fondaparanux to other pharmacologic agents for VTE prophylaxis, since studies of multiple agents have yielded only minor differences.


DB, who blogged about this study Friday, had a somewhat different take. He wonders whether insurance companies will pay for this treatment. He hopes they don't. The people who make these decisions for insurance companies, like the Obamacare czars, are somewhat challenged when it comes to interpreting clinical studies so who knows what they'll decide. But I think there are more things to consider than whether they'll pay for 45 days of Arixtra. If they don't, many patients can probably afford generic Lovenox out of pocket. And what about 5 days or so of Arixtra or a low molecular weight heparin followed by six to twelve weeks of warfarin? That alternative is plausible and deserves future study.