An 89-year-old woman with dysuria is brought to the hospital from a home and found to have a urinary tract infection. Her past medical history includes mild cognitive impairment and hypertension. Her only medication is a thiazide diuretic. On examination she’s frail, hemodynamically stable, and is incidentally found to have a leg swelling with right greater than left. There’s pitting edema up to the right knee and mild pitting edema up to the left ankle. You’re concerned about the possibility of deep vein thrombosis (DVT).
How would you investigate her and manage her?
Venous thromboembolism (VTE) is of special importance in the elderly as geriatric patients have both a higher incidence of DVT and a higher mortality associated with pulmonary embolism than the general population. Elderly people tend to have all the usual risk factors associated with VTE but also face additional risk conferred from a higher incidence of comorbidities, immobility and the hypercoagulability associated with aging.
The diagnosis of DVT in an older patient can be challenging because these individuals are more likely to be asymptomatic or present atypically. Nevertheless, the diagnostic workup follows the standard algorithm for any case of suspected DVT. Age shouldn’t preclude the use of anticoagulation, and the management of DVT in the elderly follows the same principles as in the general population. But it’s important to realize that geriatric patients have a higher rate of bleeding complications and there’s a greater propensity to reach a supratherapeutic international normalized ratio (INR) level on low doses of anticoagulants. Furthermore, caution is needed when prescribing anticoagulants that are renally cleared (such as low molecular weight heparin [LMWH] and fondaparinux), given the frequency of renal dysfunction in this age group.
The aged have an elevated risk of developing deep vein thrombosis. The annual incidence rate of first DVT in individuals aged 60-74 years is estimated to be 2.3-2.9 per 1,000, and this doubles in those aged over 75.1 Older patients also have higher in-hospital mortality associated with pulmonary embolism and DVT, at 21% and 3%, respectively.2 Therefore, accurate diagnosis and management of DVT is critical in elderly patients.
The pathophysiologic mechanism underlying venous thrombosis was first described in the mid-19th century by Virchow as a triad of vascular endothelial injury, venous stasis, and alterations in blood constituents leading to hypercoagulability. This theory has been expanded upon to identify specific inherent and acquired risk factors for VTE. Major risk factors applicable to all patients and those unique to the elderly are listed in Table 1.
One explanation for the age-dependent rise in thrombotic risk is the higher prevalence of predisposing comorbid conditions. The presence of malignancy, either overt or occult, greatly increases risk of thrombosis. The cancers most commonly associated with VTE are pancreatic, gastrointestinal, ovarian, prostate and lung, most of which have a higher prevalence in the elderly.5-7 Chronic heart failure is also more common in older patients and is associated with hypercoagulability, likely because it often leads to venous stasis.8 Patients with right-sided heart failure and significant peripheral edema may be at particularly high risk. Other VTE risk factors with a higher prevalence in the elderly include: COPD, renal insufficiency, myeloproliferative disorders and neurological conditions with associated paralysis.9-11
Obesity is a significant risk factor for VTE in younger patients;12 but a more important and common risk factor in the geriatric population is frailty. A common definition of frailty is a combination of weight loss, poor grip strength, self-reported feeling of exhaustion, slow 15-foot walk time, and low physical activity.13 Studies have demonstrated that the frail elderly have a higher incidence of VTE independent of age, BMI and comorbidites.14
Prolonged immobility predisposes to VTE in patients of all ages. Because elderly patients are much more likely to experience immobility secondary to a medical condition, it would follow that this group is particularly vulnerable. Yet interestingly, while residing in a long-term care facility has been shown to be independently associated with VTE, being bedridden in a nursing home doesn’t appear to confer any additional thrombosis risk.15
Because the process of aging leads to abnormalities in the coagulation system, the elderly are in a constant prothrombotic state. Studies have shown that plasma levels of FVII, FVIII, FV antigen, fibrinogen, d-dimer, and homocysteine are significantly higher in this population than in younger control patients.16,17 It’s been established that elevated levels of these procoagulant factors, whether inherited or acquired, are strongly associated with thrombosis risk.16,18
With few exceptions, the diagnostic workup of an elderly patient follows the standard algorithm of any person with suspected DVT. Early diagnosis is key to reducing the risk of pulmonary embolism and short-term mortality. The American College of Physicians and the American College of Family Physicians have published a recent guideline for diagnosing DVT that can be applied to the elderly.19
But clinical assessment alone is insufficient for diagnosis and this may be especially true in older people. Compared with younger individuals, elderly patients are less likely to present with typical DVT symptoms of leg pain and difficulty walking.20 Asymptomatic DVT is also likely to be quite common as studies have shown that as many as 17.8% of patients over age 80 admitted for alternate reasons had DVT incidentally diagnosed.21
A validated clinical prediction rule, high-sensitivity d-dimer and/or ultrasound (US) have been suggested as important tools for DVT diagnosis. The Wells pretest probability model is the most studied and validated prediction tool.22,23 Using nine clinical characteristics, patients are stratified into those who are likely and unlikely to have a DVT. In a meta-analysis of 51 studies looking at the diagnosis of DVT, it’s been proven superior to clinical judgement alone. But the Wells model performed better in cohorts with a mean age under 60 years than in older patients,24 suggesting that its sensitivity and specificity may be slightly lower in the elderly. Nonetheless, the Wells score is still the best tool currently available for predicting DVT based on clinical parameters in the elderly.
In patients suspected of having DVT, d-dimers are a standard part of diagnosis. If negative, it can sufficiently rule out the need for further imaging in patients who score low risk with the Wells model. In the elderly, the sensitivity of d-dimer is high, as in younger patients, but specificity is lower, possibly due to increased comorbid conditions. There’s wide variation in the reported specificity of d-dimers in the elderly with venous thrombosis — from 5%-54% — and this is likely due to different patient populations.25-27 A negative result is still helpful in the elderly, but more false positives should be expected. It may be more useful in ambulatory patients suspected of DVT than in admitted patients, who are more likely to have other reasons for an elevated d-dimer.
Ultrasound is recommended in patients who have intermediate to high pretest probability or those with a positive d-dimer. It’s also been validated in large patient populations with a wide range of ages and has excellent sensitivity and specificity for diagnosing symptomatic above-knee DVT (94% and 94%, respectively).28
Venograms are rarely necessary, but should be considered if the initial US is negative and the index of suspicion is high.
As in younger patients, treatment of DVT in the elderly can be guided by the recent American College of Chest Physicians recommendations.29 For older patients not afflicted by cancer who suffer a DVT, usual treatment will initially include at least 5 days of therapeutic doses of parenteral LMWH, unfractionated heparin (UFH), or fondaparinux, a synthetic anti-Xa pentasaccharide. An oral vitamin K antagonist, most commonly warfarin, is added at the start of parenteral therapy and titrated to achieve an INR of 2-3. When the INR is ≥ 2 for ≥ 24 hours, the parenteral treatment is discontinued and the patient is maintained on oral anticoagulation alone. The duration of anticoagulation is usually three months for clots associated with a reversible risk factor or a first clot isolated to the calf veins. Longer periods of anticoagulation should be considered for unprovoked DVT or recurrent clots, with such decisions individualized to the patient’s unique clinical characteristics and preferences. For the majority of people with unprovoked proximal DVT who are tolerating anticoagulation, treatment is continued for at least six months, and indefinitely if patient preference allows.
The aged make up a subset of the patient population in important trials of anticoagulation for DVT. As a result, major guidelines aren’t specific to the elderly but do make note of unique aspects of their care.30 Older age shouldn’t preclude the use of anticoagulants, but some special considerations do apply.31 The issue most deserving of attention is the propensity of the elderly to bleed while on anticoagulants. Estimating bleeding risk is difficult. Quoted bleeding rates for elderly patients being treated for DVT are derived from trials of warfarin use for stroke protection in younger patients with atrial fibrillation. Under these circumstances, guidelines cite a 0.3-0.5%/yr increase in major bleeding due to warfarin over a baseline risk of 1%/yr.30 Two recent systematic reviews that focus on VTE estimate bleeding risk associated with warfarin at 2-3% over 1 year, and suggest that most bleeding events occur within the first three months of treatment.32-34 This compares to about 1-2% with LMWH and 2% with UFH.4,5,33 Therapeutic doses of fondaparinux carry a risk of major bleeding comparable to LMWH.35,36
Keep a sharp eye on the INR
Age is increasingly being recognized as an independent risk factor for bleeding, so these numbers may underestimate bleeding in the elderly.9,37-39 A large registry of patients being treated for VTE with LMWH then vitamin K antagonists found a statistically significant difference in major bleeding between the age groups (3.4% in those > 80 years, 2.1% in those < 80 years).9 Tight INR control to between 2-3 has been shown to minimize bleeding.40 Consequently, guidelines recommend that warfarin be initiated at a dose of ≤ 5 mg in the elderly (compared to 5-10 mg in younger patients) and subsequent dosing be based on INR response.30 In fact, over age 70, a majority of both men and women will become supra-therapeutic even at starting doses of 5 mg/day, and starting at 4 mg/day — with close attention to INR — may be a safer approach.41 It’s also imperative that the treating physician considers renal function when choosing an anticoagulant. Age is often associated with a decline in creatinine clearance and because both LMWHs and fondaparinux are excreted in the urine, they can accumulate if kidney function is compromised. While prescription guidelines specify that fondaparinux is not to be used in patients with a creatinine clearance < 30 ml/min, there’s no such criterion for the use of LMWH.31,42,43 If such agents are prescribed in older patients with even mild-to-moderate renal dysfunction, close anti-Xa level monitoring and dose adjustments are prudent.
Back to the case
Based on the clinical prediction rule, our patient is found to have a score of 1 for right calf swelling at least 3 cm larger than the left. This equates to an unlikely pretest probability of DVT. A highly sensitive d-dimer is ordered performed and returns positive. Venous US confirms a right popliteal DVT.
The patient is initiated on weight-adjusted subcutaneous tinzaparin and concomitantly on oral warfarin at a dose of 5 mg/day. Five days later her INR is 2.5 and her tinzaparin is discontinued 24 hours after reaching her INR target. She’s discharged back to her residence with ongoing warfarin therapy and instructions to follow up with her primary care physician for INR monitoring and counselling regarding duration of anticoagulation for her deep vein thrombosis.
For the most part, diagnosis and management in the elderly is similar to younger patients and the major guidelines can be applied with a few unique considerations. But the diagnosis can be more challenging, because elderly patients often present with atypical symptoms, and d-dimer is more likely to be positive in this age group for reasons other than DVT. Risk factors in the elderly include frailty, decreased mobility, and comorbid conditions such as heart failure and malignancy. Standard treatment is still anticoagulation with vitamin K antagonists, but bleeding rates are higher in the elderly than in younger patients. Good follow-up and tight INR control will minimize such incidents.
Lauren M. Gerard, MD, is an Internal Medicine Resident with the University of Toronto and is joining the Hematology Fellowship program at the University of Toronto in July 2011.
Robert Wu, MD, FRCPC is an assistant professor with the Department of Medicine at University of Toronto and general internist at the University Health Network (UHN).
Katina Tzanetos, MD, FRCPC is the education director of the post-graduate ambulatory clinic at the University Health Network and the medicine clerkship site director at the Toronto General Hospital for Year III medical students.