risk for recurrent thrombosis (Ap- pendix 1) (3). Although D-dimer has excellent negative predictive value, its positive predictive value is low (27.1%; 95% confidence interval: 16.4% to 40.0%) (4), making it likely that providing anticoagulation to all patients with a positive D-dimer will result in treating patients who are ac- tually at low risk of recurrent throm- boembolic disease. Our sensitivity analyses are very clear that assigning even a small number of low-risk pa- tients (as few as 2% in our models) to risks of longer-term anticoagulation markedly decreases the effectiveness of any treatment strategy; this sce- nario seems a possibility in ‘D-dimer only’ testing algorithms. Again, these questions will best be answered in randomized, prospective trials of testing and treatment strategies. Andrew D. Auerbach, MD, MPH Julie Hambleton, MD School of Medicine University of California, San Francisco 1. Bick RL, Baker WF. Antiphospholipid syn- drome and thrombosis. Semin Thromb He- most. 1999;25:333–350. 2. Prandoni P, Lensing AW, Prins MH, Ber- nardi E, Marchiori A, et al. Residual venous thrombosis as a predictive factor of recur- rent venous thromboembolism. Ann Intern Med. 2002;137:955–960. 3. Auerbach AD, Sanders GD, Hambleton J. Cost-effectiveness of testing for hyperco- agulability and effects on treatment strate- gies in patients with deep vein thrombosis. Am J Med. 2004;116:816 – 828. 4. Palareti G, Legnani C, Cosmi B, Valdre L, Lunghi B, et al. Predictive value of D-dimer test for recurrent venous thromboembo- lism after anticoagulation withdrawal in subjects with a previous idiopathic event and in carriers of congenital thrombophilia. Circulation. 2003;108:313–318. A 1-WEEK COURSE OF CORTICOSTEROIDS IN THE TREATMENT OF EOSINOPHILIC MENINGITIS To the Editor: Eosinophilic meningitis is com- monly found in Thailand and other Asian-Pacific countries. Angiostrongy- lus cantonesis, a nematode, is the most common cause of the disease (1– 4). Humans get these parasites by eating raw snails or by eating food contami- nated with snail excrement. A 2-week course of 60-mg pred- nisolone has been shown to relieve headache and reduce the number of lumbar punctures without any seri- ous side effects. Eighty percent of pa- tients were free of headache within 1 week of treatment, and there was no recurrent meningitis (1). We did a prospective cohort study to evaluate the efficacy of a 1-week course of cor- ticosteroid treatment in patients with eosinophilic meningitis. The diagnosis of eosinophilic men- ingitis was made clinically (headache with history of eating raw snails and a cerebrospinal fluid eosinophil count of 10%). All patients were evalu- ated for severity of headache by visual analogue scale (VAS) and routine laboratory tests. Prednisolone 60 mg/d in three divided doses was given to all patients who did not have a con- traindication to corticosteroid. We followed all patients clinically. We completely evaluated 52 pa- tients between August 2002 and March 2004. Forty-seven patients (90%) were symptom free within 1 week, with a mean duration of symp- toms of 4.8 days. Only 1 patient needed repeat lumbar puncture. Headache was improved after lumbar puncture, with a 63.29% decline in the VAS rating (Table). At follow-up, 8 patients (15%) had relapsed within 2 weeks (Table). Two patients needed no further treatment because the headaches were not se- vere (VAS, 2 and 3); 1 of these pa- tients recovered at day 12, the other at day 13. The other 6 patients needed additional treatment, such as lumbar puncture or prednisolone. All re- lapsed patients recovered by day 22. Minor adverse events, such as cushi- noid face, acne, and dyspepsia were reported. The mechanism to explain the symptoms of headache in eosino- philic meningitis has not yet been es- tablished. On autopsy, dead parasites were found in patients who had had severe eosinophilic meningoenceph- alitis (5,6); therefore, we believe that the presence of dead parasites causes meningeal inflammation. Prednis- olone can improve headache symp- toms by reducing the inflammatory process. A 1-week course of cortico- steroid showed the same beneficial ef- fect as a 2-week course in increasing the number of patients who felt free from headache. The relapsed patients may still have had residual inflamma- tory process. Because of the small number of subjects in our study, we could not assess the predictive factors of relapse. Therefore, we suggest that a 1-week course of corticosteroid treatment is also effective to treat eo- sinophilic meningitis, but that clini- cians should be aware of possible re- lapse. Table. Clinical Variables of the 52 Patients Variables Number (%), Mean  SD, or Percentage VAS at initial, mean 8.23  0.58 VAS after lumbar puncture Mean 3.02  0.55 Decrease (%) 63.29 Patients who recovered at 1 wk 47 (90) Patients who needed repeat lumbar puncture 1 (0.02) Relapsed patients 8 (15) Relapsed patients who needed additional treatment 6 (12) VAS = visual analog scale. Letters to the Editor 802 November 15, 2004 THE AMERICAN JOURNAL OF MEDICINE Volume 117