BRIEF REPORT

Lack of Efcacy of Prophylactic Platelet Transfusion for Severe Thrombocytopenia in Adults with Acute Uncomplicated Dengue Infection
David C. Lye,1 Vernon J. Lee,2,3 Yan Sun,4 and Yee Sin Leo1 Departments of 1Infectious Disease and 2Clinical Epidemiology, Tan Tock Seng Hospital, 3Biodefence Center, Ministry of Defence, and 4Clinical Project Management and Planning, National Healthcare Group, Singapore

Thrombocytopenia in dengue infection raises concerns about bleeding risk. Of 256 patients with dengue infection who developed thrombocytopenia (platelet count, !20 10 3 platelets/mL) without prior bleeding, 188 were given platelet transfusion. Subsequent bleeding, platelet increment, and platelet recovery were similar between patients given transfusion and patients not given transfusion. Prophylactic platelet transfusion was ineffective in preventing bleeding in adult patients with dengue infection. Acute dengue infection is endemic to many tropical countries, and its incidence is increasing globally [1]. Thrombocytopenia is common in acute dengue infection [2], and many experimental therapies, including corticosteroid treatment [3] and intravenous immunoglobulin treatment [4], have been used to treat severe thrombocytopenia in dengue hemorrhagic fever, because of the fear of potential bleeding. Although severe bleeding and coagulopathy occur in dengue shock syndrome (DSS) [5], the cause is multifactorial [6, 7] and is not caused by thrombocytopenia alone. Thrombocytopenia was not an independent predictor of severe bleeding in pediatric DSS [8]. Prophylactic transfusion with platelets and fresh frozen plasma was also not benecial in pediatric DSS [9]. Despite this, platelet transfusion was frequently given for severe thrombocytopenia to adults with acute dengue infection
Received 7 October 2008; accepted 5 January 2009; electronically published 17 March 2009. D.C.L. and V.J.L. contributed equally to this article. Reprints or correspondence: Dr. Vernon J. Lee, 42 How Sun Dr., Singapore 538611, Republic of Singapore (vernonljm@hotmail.com). Clinical Infectious Diseases 2009; 48:12625 2009 by the Infectious Diseases Society of America. All rights reserved. 1058-4838/2009/4809-0018$15.00 DOI: 10.1086/597773

in Singapore and Taiwan [2, 10]. We studied the impact of prophylactic platelet transfusion for severe thrombocytopenia in a large retrospective cohort of adults with acute dengue. Methods. We studied all patients with acute dengue in 2004 who were admitted to the Department of Infectious Diseases, Tan Tock Seng Hospital, Singapore [2]. Only patients who fullled the World Health Organization criteria for acute dengue (fever and 2 of the following symptoms: headache, eye pain, myalgia, arthralgia, leukopenia, rash, and bleeding) and who had results of laboratory diagnostic tests positive for dengue were included. All cases were classied as dengue fever, dengue hemorrhagic fever (all 4 criteria present: fever, thrombocytopenia [platelet count, !100 10 3 platelets/mL], bleeding manifestation, and plasma leakage), or DSS [5]. Patients with probable cases had acute dengue with positive results of serological analysis (Dengue Duo IgM & IgG Rapid Strip Test; Panbio), and those with conrmed cases had positive results of PCR. Medical chart review was performed to extract demographic characteristics, serial clinical and laboratory data, and treatment and outcome data. Severe thrombocytopenia was dened as platelet count !20 10 3 platelets/mL, on the basis of local clinical practice. Prophylactic platelet transfusion was dened as platelet transfusion without clinical bleeding, in contrast to therapeutic platelet transfusion with clinical bleeding. Clinical bleeding excluded petechiae. We excluded all patients with clinical bleeding before and on the day that their platelet count decreased to !20 10 3 platelets/mL. Patients whose platelet count decreased to !20 10 3 platelets/mL without clinical bleeding and who were given prophylactic platelet transfusion were compared with those who were not given prophylactic platelet transfusion. Prophylactic platelet transfusion was given on the basis of a clinicians assessment. Outcome measures consisted of clinical bleeding after prophylactic platelet transfusion, platelet count increment the day after transfusion, time to platelet count 150 10 3 platelets/mL, length of hospitalization, and death. The x2 test and Fishers exact test were used to compare categorical variables, and Students t test and the Mann-Whitney U test were used to compare continuous variables. All statistical analyses were performed using Stata, version 9.0 (Stata Corp), and tests were conducted with the signicance level at 5%. Corresponding percentages, 5th and 95th percentiles, ORs, and 95% CIs are reported. Results. In 2004, 1973 patients were admitted to our department at Tan Tock Seng Hospital who fullled the World Health Organization criteria for acute dengue and who had

Downloaded from http://cid.oxfordjournals.org/ by guest on June 5, 2012

1262 CID 2009:48 (1 May) BRIEF REPORT

positive results of laboratory diagnostic tests. Of the cases, 54% were probable, and 46% were conrmed. There were 1855 patients with dengue fever and 118 patients with dengue hemorrhagic fever; 10 of the patients with dengue hemorrhagic fever had DSS. Seven patients required admission to the intensive care unit, and 1 patient died; all 8 of these patients had dengue hemorrhagic fever. The median age was 32 years (5th95th percentiles, 1757 years), and 64% of patients were male. The median duration of illness at presentation was 5 days (5th95th percentiles, 3 7 days), and the median length of hospital stay was 4 days (5th 95th percentiles, 27 days). At presentation, clinical bleeding occurred in 220 patients (85% had gum bleeding, 16% had epistaxis, 5% had menstrual bleeding, and 1% had gastrointestinal bleeding). The median platelet count nadir for patients with dengue fever was 51 10 3 platelets/mL (5th95th percentiles, 1197 10 3 platelets/mL), compared with a median of 49 10 3 platelets/mL (5th95th percentiles, 990 10 3 platelets/mL) for patients with dengue hemorrhagic fever (P 1 .05). Platelet transfusion was given to 249 patients (12.6%) during hospitalization (either prophylactic or therapeutic platelet transfusion); 61 of these patients received therapeutic transfusion. Of the 1973 patients, 1666 had a platelet count nadir 120 10 3 platelets/mL, and 51 patients had bleeding and/or received platelet transfusion when the platelet count was 120 10 3 platelets/mL; these patients were excluded from further analysis. The remaining 256 patients without bleeding whose platelet count decreased to !20 10 3 platelets/mL were included in subsequent analysis, and 188 received prophylactic platelet transfusion. The baseline demographic and clinical variables on the day that the platelet count decreased to !20 10 3 platelets/mL for the 256 patients who either were or were not given prophylactic platelet transfusion are shown in table 1. The median platelet count for both patients given transfusion and patients not given transfusion was 15 10 3 platelets/mL (OR, 1.02; 95% CI, 0.94 1.09; P p .87). Patients given transfusion were more likely to have fever (temperature, 138 C), compared with patients not given transfusion (33% vs. 18%; OR, 2.30; 95% CI, 1.154.59; P p .02), and had signicantly higher systolic blood pressure (OR, 1.03; 95% CI, 1.011.06; P p .01) and pulse pressure (OR, 1.02; 95% CI, 1.001.05; P p .03). The median time from disease onset to hospital admission was 3 days (5th95th percentiles, 26 days) for patients given transfusion, compared with 3 days (5th95th percentiles, 16 days) for patients not given transfusion (P p .38). The median time from disease onset to platelet count !20 10 3 platelets/mL and platelet transfusion was 5 days (5th95th percentiles, 37 days) for patients given transfusion, compared with 4.5 days (5th95th percentiles, 3 8 days) for patients not given transfusion (P p .94 ). There were

no other signicant demographic or clinical differences between the 2 groups. The incidence of clinical bleeding was 6% among patients with platelet count 1150 10 3 platelets/mL, 12% among patients with platelet count of 100149 10 3 platelets/mL, 11% among patients with platelet count of 8099 10 3 platelets/mL, 10% among patients with platelet count of 5079 10 3 platelets/mL, 11% among patients with platelet count of 2049 10 3 platelets/mL, 13% among patients with platelet count of 1019 10 3 platelets/mL, and 0% among patients with platelet count !10 10 3 platelets/mL (P p .22, by test for trend). Among patients whose platelet count decreased to !20 10 3 platelets/mL, prophylactic transfusion was given to 188 patients at a median of 4 units of platelets (range, 112 units). Clinical bleeding subsequently occurred in 1 (0.5%) of 188 patients given prophylactic transfusion (severe gastrointestinal bleeding in a patient with known peptic ulcer disease who died due to disseminated intravascular coagulation and acute renal failure), compared with 2 (2.9%) of 68 patients not given transfusion (P p .17) (both instances of bleeding were mild and did not require additional intervention). The median platelet increment on the day after platelet transfusion in patients given transfusion (7 10 3 platelets/mL) was lower than the median platelet increment on the day after platelet count rst decreased to !20 10 3 platelets/mL in patients not given transfusion (11 10 3 platelets/mL; OR, 1.00; 95% CI, 0.981.01; P p .26). The median time to platelet count 150 10 3 platelets/mL was similar for patients given transfusion and patients not given transfusion (3 days; OR, 1.05; 95% CI, 0.791.39; P p .59). The median length of hospital stay was 6 days for patients given transfusion, compared with 5 days for patients not given transfusion (P p .09 ). One patient died in the group given transfusion, compared with none in group not given transfusion (P p 1.00). Discussion. Prophylactic platelet transfusion is given because of the fear of severe bleeding in patients with acute dengue and thrombocytopenia. Experimental use of corticosteroid treatment and intravenous immunoglobulin treatment [3, 4] and the overuse of platelet transfusion [2, 10] have been reported. Platelet transfusion was given to 50.3% of hospitalized patients with dengue in Taiwan [10] and 12.6% in Singapore [2]. Only 7 patients in our study had severe bleeding1 was in the group given transfusion, and the other 6 had a mean platelet count of 78 10 3 platelets/mL (range, 23111 10 3 platelets/mL) during the bleeding episode without ever reaching the platelet count nadir of !20 10 3 platelets/mL. Notably, a randomized study showed that intravenous immunoglobulin treatment did not hasten recovery from thrombocytopenia among patients with secondary dengue infection [4]. Our study further strengthens the evidence that thromboBRIEF REPORT CID 2009:48 (1 May) 1263

Downloaded from http://cid.oxfordjournals.org/ by guest on June 5, 2012

Table 1. Baseline demographic characteristics, clinical and laboratory data on the day the platelet count decreased to !20 103 platelets/mL, and clinical outcomes for patients with acute dengue infection who did or did not receive prophylactic platelet transfusion.
Patients given platelet transfusion (n p 188) 40 (2264) 144 (77) 4 (2) 124 (66) 60/86 (70) 11 (6) 18 (10) 9 (5) 0 (0) 46 (24) 13 27 1 17 21 17 14 27 (7) (14) (1) (9) (11) (9) (7) (14) Patients not given platelet transfusion (n p 68) 39 (2258) 45 (66) 2 (3) 46 (68) 24/32 (75) 2 (3) 5 (7) 1 (1) 0 (0) 22 (32) 3 12 0 4 6 5 3 8 (4) (18) (0) (6) (9) (7) (4) (12)

Variable Demographic characteristics Age, years Male sex Dengue diagnosis Dengue hemorrhagic fever Positive results of PCR Test results positive for IgG Preexisting medical conditions Diabetes mellitus Hypertension Hyperlipidemia Ischemic heart disease Clinical features Fever Headache Myalgia/arthralgia Eye pain Anorexia Nausea Vomiting Diarrhea Rash Temperature, C Temperature 38 C Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Systolic blood pressure !90 mm Hg Pulse pressure, mm Hg Pulse, beats/min Pulse !60 beats/min Abdominal tenderness Pleural effusion or ascites Laboratory results Hematocrit, % Hematocrit 50% Leukocyte count, 103 leukocytes/mL Leukocyte count !3.3 103 leukocytes/mL Platelet count, 103 platelets/mL Clinical outcomes Any bleeding Platelet increment the next day, 103 platelets/mL Time to platelet count 50 103 platelets/mL, days Length of hospital stay, days Death

P .54 .11 .66 .88 .65 .52 .80 .30 1.00

Downloaded from http://cid.oxfordjournals.org/ by guest on June 5, 2012

.26 .57 .56 1.00 .61 .82 .80 .57 .68 .09 .02 .01 .12 .27 .03 .84 .11 .57 1.00 .24 .17 .46 .32 .87 .17 .26 .59 .09 1.00

37.5 (36.639) 62 (33) 115 70 0 45 (95140) (5087) (0) (3070)

37.2 (36.539.4) 12 (18) 110 70 1 40 (95130) (5085) (1) (3060)

70 (5593) 18 (10) 3 (2) 0 (0) 45.7 (36.752.1) 23 (12) 3.4 (1.77.2) 86 (46) 15 (719) 1 (1) 7 ( 7 to 50) 3 (14) 6 (48) 1 (1)

70 (6094) 2 (3) 0 (0) 0 (0) 44.9 (35.450.3) 4 (6) 3.6 (1.78.2) 26 (38) 15 (819) 2 (3) 11 ( 4 to 41) 3 (15) 5 (47) 0 (0)

NOTE. For dichotomous variables, data are no. (%) of patients; for continuous variables, data are median (5th95th percentiles).

cytopenia in acute dengue infection does not correlate with bleeding risk. In pediatric DSS, thrombocytopenia did not predict severe bleeding in univariate analysis; the only 2 independent predictors of severe bleeding were shock and low hematocrit [8]. Another prospective pediatric study found that bleeding score did not correlate with platelet count [11]. Our large cohort of adults with acute dengue revealed that the incidence of clinical bleeding at hospital admission was independent of platelet count. The efcacy of prophylactic platelet transfusion and the threshold for transfusion is questionable. In a prospective study of acute myeloid leukemia, lowering the platelet threshold for prophylactic platelet transfusion from 20 10 3 platelets/mL to 10 10 3 platelets/mL did not increase the incidence of bleeding but signicantly reduced the number of patients who received platelet transfusion [12]. In pediatric DSS, prophylactic transfusion of platelets and fresh frozen plasma did not reduce bleeding or expedite platelet recovery; instead, it caused uid overload and prolonged hospitalization [9]. In addition, the improvement in platelet count was transient, lasting !5 h [9]. We have shown that prophylactic platelet transfusion did not improve relevant outcome measures, such as clinical bleeding, platelet increment, and platelet recovery. Our study has several limitations. Its retrospective cohort design does not allow for increased recruitment, and the overall number of episodes of clinical bleeding was small (3 [1%] of 256 patients). The lack of randomization may have resulted in treatment bias, although both the transfusion and the nontransfusion groups had similar baseline features. Our cohort of patients with platelet count !20 10 3 platelets/mL was limited to adult patients and included only 6 patients with dengue hemorrhagic fever, limiting the generalizability of our ndings. Additional research on the role of prophylactic platelet transfusion in a randomized study with a larger cohort is needed. Because our results revealed that there were no signicant benets of prophylactic platelet transfusion among adult patients with dengue, which are similar to results of other studies involving different cohorts, we no longer advocate prophylactic platelet transfusion given on the basis of platelet count for adults with acute uncomplicated dengue infection. This ap-

proach will save precious blood products and will reduce unnecessary patient exposure to transfusion risks.
Acknowledgments
Financial support. National Healthcare Group (small innovative grant SIG/05048); National Medical Research Council (individual research grant NMRC/1006/2005). Potential conicts of interest. All authors: no conicts.

References
1. World Health Organization. Impact of dengue. Available at: http:// www.who.int/csr/disease/dengue/impact/en/index.html. Accessed 29 September 2008. 2. Lee VJ, Lye DCB, Sun Y, Fernandez G, Ong AK, Leo YS. Predictive value of simple clinical and laboratory variables for dengue hemorrhagic fever in adults. J Clin Virol 2008; 42:349. 3. Leong KW, Srinivas P. Corticosteroid-responsive prolonged thrombocytopenia following dengue haemorrhagic fever. Med J Malaysia 1993; 48:36972. 4. Dimaano EM, Saito M, Honda S, et al. Lack of efcacy of high-dose intravenous immunoglobulin treatment of severe thrombocytopenia in patients with secondary dengue virus infection. Am J Trop Med Hyg 2007; 77:11358. 5. Dengue haemorrhagic fever: diagnosis, treatment, prevention and control. 2nd ed. Geneva: World Health Organization, 1997. 6. Hathirat P, Isarangkura P, Srichaikul T, Suvatte V, Mitrakul C. Abnormal hemostasis in dengue hemorrhagic fever. Southeast Asian J Trop Med Public Health 1993; 24(Suppl 1):805. 7. Wills BA, Oragui EE, Stephens AC, et al. Coagulation abnormalities in dengue hemorrhagic fever: serial investigations in 167 Vietnamese children with dengue shock syndrome. Clin Infect Dis 2002; 35:27785. 8. Lum LC, Goh AY, Chan PW, Abdel-Latif ME, Lam SK. Risk factors for severe hemorrhage in severe dengue infections. J Pediatr 2002; 140: 62931. 9. Lum LC, Abdel-Latif ME, Goh AY, Chan PW, Lam SK. Preventive transfusion in dengue shock syndromeis it necessary? J Pediatr 2003; 143:6824. 10. Lee MS, Hwang KP, Chen TC, Lu PL, Chen TP. Clinical characteristics of dengue and dengue hemorrhagic fever in a medical center of southern Taiwan during the 2002 epidemic. J Microbiol Immunol Infect 2006; 39:1219. 11. Krishnamurti C, Kalayanarooj S, Cutting MA, et al. Mechanisms of hemorrhage in dengue without circulatory collapse. Am J Trop Med Hyg 2001; 65:8407. 12. Wandt H, Frank M, Ehninger G, et al. Safety and cost effectiveness of a 10 109/L trigger for prophylactic platelet transfusions compared with the traditional 20 109/L trigger: a prospective comparative trial in 105 patients with acute myeloid leukemia. Blood 1998; 91:36016.

Downloaded from http://cid.oxfordjournals.org/ by guest on June 5, 2012

BRIEF REPORT CID 2009:48 (1 May) 1265