Daunted by the seeming impossibility of measuring and comparing hospitals on real outcomes (given our primitive state of data collection and heterogeneity in patient populations, among many other challenges), well-meaning bureaucrats and non-profiteering safety advocates like the Institute for Healthcare Improvement have created directives based on surrogate measures in an attempt to standardize and improve medical care. Despite their seeming authoritative base, these guidelines have almost without exception been constructed on an evidence base that's suggestive, but hardly sturdy. In the absence of anything more reality-based, though, payers and regulatory agencies have put their collective shoulder behind these "quality" guidelines, whose momentum has compelled the participation of physicians and created large-scale shifts in clinical practice.
Unintended consequences ensued: Holding hospitals accountable to a 4-hour window to deliver antibiotics for community acquired pneumonia led to misdiagnosis of CAP and over-prescription of antibiotics in emergency departments nationwide, with no proof of a subsequent survival benefit. (The 4-hour window emerged from a 2003 IDSA guideline that graded the recommendation B-III, or "moderate strength," based on expert opinion/clinical experience -- the lowest quality of evidence in their methodology table.)
Universal venous thromboembolism prophylaxis for hospitalized patients may be the next pillar of the wobbly quality edifice to crumble under the weight of evidence. After a systematic review of the literature, the American College of Physicians concludes in its just-released practice guideline that providing heparin prophylaxis to all patients (a Joint Commission recommendation) results in avoidable harm, from bleeding in patients at low risk for DVT/PE or at high risk for bleeding. Their recommendation to physicians is to perform a risk assessment on each patient and provide heparin prophylaxis where appropriate.
Methods: Lederle et al performed the Herculean task of reviewing 4,340 studies published 1950-2011, extracting 40 randomized trials evaluating DVT/PE prophylaxis, and combining them into multiple meta-analyses where possible.
- Heparin prophylaxis did not reduce total mortality, although there was a strong trend toward benefit (p=0.056 for ~6 deaths prevented per 1,000 patients treated overall).
- Heparin reduced the rate of pulmonary embolism, preventing about 3 PEs per 1,000 patients receiving prophylaxis (odds ratio 0.69 but with evidence of publication bias). These PEs were clinically evident, not overdiagnosed subsegmental PEs.
- However, finding PEs exacted a real cost, paid in blood: an absolute increase of 4 major bleeding events per 1,000 patients treated. ("Major bleeding" varied between studies, but meant at minimum 2 units pRBC transfused, and in some, meant fatal or life-threatening bleeding.)
- Intermittent pneumatic compression devices could not be evaluated sufficiently among all patients. However, among patients with stroke, those treated with lower-extremity compression devices had a troubling increase of 39 cases of lower-extremity skin damage per 1,000 treated.
- No significant differences in benefits or harms were detected between unfractionated heparin and low-molecular weight heparin, but there was a trend toward benefit for LMWH in preventing PEs. Heparin-induced thrombocytopenia was not measured as an outcome.
Clinical Takeaway: Use heparin except in low-risk patients, or who have a high risk of bleeding. How to estimate the risk of bleeding was not advised. This paper and entire discussion has all really been about general medical-ward patients, though.
As for the critically ill, they're arguably nearly all at moderate or high risk for both PE and bleeding. The risk for surveillance-detected DVT in the ICU is very high (5% even while receiving prophylaxis), and the risk of clinically significant PE in heparin-prophylaxed ICU patients is 1-2%, which is extremely high (per observations in the PROTECT trial). When does the bleeding risk from subcutaneous heparin outweigh its benefits of preventing a PE in the critically ill patient? Unless there's blood on the bed, hell if I know, but thromboprophylaxis to anyone without a strong contraindication should likely be the default, especially after this study suggested not doing so immediately on ICU admission could hurt people.
As always, reality is a murky place, but quality measures don't like murk. At my institution, risk-assessing heparin off a hospitalized patient's med list would result in the patient's room icon being highlighted RED (on a system of large centrally located flat screen LCDs they spent big bucks on, partly to comply with JCHAO). Everybody then secretly suspects that patient's doctor is a bad, uncaring person (I can't 100% prove this).
Lederle FA et al. Venous Thromboembolism Prophylaxis in Hospitalized Medical Patients and Those With Stroke: A Background Review for an American College of Physicians Clinical Practice Guideline. Ann Intern Med 2011;155:602-615.
Qaseem A et al. Venous Thromboembolism Prophylaxis in Hospitalized Patients: A Clinical Practice Guideline From the American College of Physicians. Ann Intern Med 2011;155:625-632.