Received 3 June 2010 Revised 17 July 2010 Accepted 21 July 2010

Transcription

1 Review article 7 The efficacy of different dose intravenous immunoglobulin in treating acute idiopathic thrombocytopenic purpura: a meta-analysis of randomized controlled trials Yuan-Han Qin, Tian-Biao Zhou, Li-a Su, Feng-Ying Lei, Yan-Jun Zhao and Wei-Fang Huang The purpose of this study was to compare the effects of different dose intravenous immunoglobulin for treatment of acute idiopathic thrombocytopenic purpura. Randomized controlled trials (RCTs) comparing high-dose intravenous immunoglobulin () with low-dose intravenous immunoglobulin (low-ivig) for acute idiopathic thrombocytopenic purpura (ITP) were identified using a predefined search strategy. Effective rate, time of cessation of bleeding, time of platelet count beginning to rise, platelet count by the first week of treatment, the number of platelets after 2 weeks of treatment, time of platelet count to reach peak, peak value of platelet count after treatment, sideeffects and rate of developing into chronic ITP were extracted and compared by RevMan (The Cochrane Collaboration, Oxford, UK). Thirteen RCTs (646 patients) were identified. Meta-analysis showed that effective rate, time of cessation of bleeding, time of platelet count beginning to rise, platelet count by the first week of treatment, the number of platelets after 2 weeks of treatment, time of platelet count to reach peak, peak value of platelet count after treatment and rate of developing into chronic ITP were not statistically different between the two different treatment administrations. However, low-ivig was associated with a significantly reduced risk of side-effects {odds ratio (OR) 0.39 [95% confidence interval (CI) ]; P U 0.01]. In conclusion, low-ivig can be performed as effectively as without increasing the rate of developing into chronic ITP. Furthermore, the low-ivig regimen can have fewer side-effects than administration in patients with acute ITP. Blood Coagul Fibrinolysis 21:7 721 ß 2010 Wolters Kluwer Health Lippincott Williams & Wilkins. Blood Coagulation and Fibrinolysis 2010, 21:7 721 Keywords: acute idiopathic thrombocytopenic purpura, high-dose intravenous immunoglobulin, low-dose intravenous immunoglobulin, meta-analysis Department of Pediatrics, The First Affiliated Hospital of GuangXi Medical University, aning, GuangXi, China Correspondence to Tian-Biao Zhou, Department of Pediatrics, The First Affiliated Hospital of GuangXi Medical University, aning, GuangXi , China a126tianbiao@126.com Received 3 June 2010 Revised 17 July 2010 Accepted 21 July 2010 Introduction Idiopathic thrombocytopenic purpura (ITP), an autoantibody-mediated disorder affecting both children and adults, is one of the most common haemostasis disorders. The majority of patients with acute ITP run a course remitting within several weeks to 6 months [1]. Options for treatment of acute ITP include steroids, intravenous immunoglobulin (IVIG), anti-d immunoglobulin and no drug therapy during past decades [2]. In most cases, the life-threatening hemorrhage is the main reason for drug therapy [2]. Intravenous gamma globulin is widely used in clinic for treating patient suffering from ITP. Because of adverse effects of steroids and needing bone marrow aspiration before administration of steroids, IVIG is a substituted agent for steroids in recent years [2]. IVIG can induce a rapid response in platelet counts and stop bleeding quickly. However, IVIG administration brings some adverse effects such as fever, nausea, vomiting, headache, venous and arterial thrombosis [3,4]. The reason may be that IVIG infusion elevates blood viscosity, and the increased blood viscosity can cause a hypercoagulable state [4] and leads to adverse reaction. Side reaction of high-dose intravenous immunoglobulin () is more obvious than that of low-dose intravenous immunoglobulin (low-ivig) in clinical administration. In order to explore a suitable treatment schedule, the dosage of IVIG in ITP treatment was changed during the past decades. There was a controversy in the past about whether low- IVIG could get a similar effectiveness to for the patients with ITP. Furthermore, in some developing countries, such as China and some other Asian countries, the cost of a scheme such as 0.2 g/kg per day for 5 days is much less than that of 2.0 g/kg for a single dose (cost of medication much more than infusion cost). Therefore, this meta-analysis was conducted to explore whether a low- IVIG regimen could get a similar effectiveness as HD- IVIG administration for patients with ITP. There are some different characteristics between acute ITP and chronic ITP in the therapeutic standard strategies. Administration of IVIG is the fastest way to increase the platelet count for patients with acute ITP, but IVIG often must be repeatedly infused for chronic ITP [5]. In order to enhance the credibility of the ß 2010 Wolters Kluwer Health Lippincott Williams & Wilkins DOI: /MBC.0b0e

2 714 Blood Coagulation and Fibrinolysis 2010, Vol 21 o 8 evidence in our study, we only recruited the randomized controlled trials (RCTs) for acute ITP under treatment with immunoglobulin administration. For all the included literature, the total dose of 2 g/kg was regarded as the schedule, and less than 2 g/kg was regarded as low-ivig. Materials and methods Search strategy The relevant studies were screened from the search engines of PubMed and CBM-disc (China Biological Medicine Database) as of 31 December [(idiopathic thrombocytopenic purpura or ITP) and (high-dose intravenous immunoglobulin or low-dose intravenous immunoglobulin) and [therapy/narrow (filter)] was used in PubMed without language limitation. CBM-disc was searched using the terms [(idiopathic thrombocytopenic purpura or ITP) and (high-dose intravenous immunoglobulin) or (low-dose intravenous immunoglobulin) or (intravenous immunoglobulin)]. We also extended search spectrum to the related articles and the bibliographies of all retrieved studies. If multiple publications from the same study group occurred, we only recruited the most complete paper for analysis. Inclusion and exclusion criteria Inclusion criteria The inclusion criteria for the study are given below: (1) type: RCTs. (2) Object of the study: all patients who met the diagnostic criteria for acute ITP [6]. (3) Interventions: using low-ivig and for treatment, respectively. (4) Baseline information: comparable. Exclusion criteria The exclusion criteria for the study are given below: (1) IVIG total dose of high-dose group was less than 2 g/kg or that in low-dose group was more than 2 g/kg (total dose of 2 g/kg was regarded as the high-dose for ITP patient [7]). (2) Low-dose group or high-dose group including other drug. (3) The diagnostic criteria were not clear. Outcome measures Effective rate, time of cessation of bleeding, time of platelet count beginning to rise, platelet count by the first week of treatment, the number of platelets after 2 weeks of treatment, time of platelet count to reach peak, peak value of platelet count after treatment, side-effects and rate of developing into chronic ITP. Data collection Method of agreeing inclusion of studies was performed by two observers (Y.-H. Q., T.-B. Z.) independently according to predetermined inclusion criteria. Titles and abstracts were scanned first to make a list of possibly related literature and then full texts were obtained for those articles identified as either relevant or not clear, only RCTs fitting predefined inclusion criteria were included. Disagreements were resolved by other reviewers (L.-. S., F.-Y. L.). Appraising the quality of literature The Jadad scale [8] was used to score the quality of the included articles (Table 1) [9 21]. On the basis of scores, we subdivided them into the following two categories: the score below 3 was regarded as low quality and the 3 5 score was high quality (total score was 5). In order to get the detailed implementation process, we tried our best to contact the authors by or telephone (Y.-J. Z., W.-F. H.). Statistical analysis Statistical analysis was performed by RevMan The pooled statistics were calculated using the fixed effect model, but a random effect model was conducted if P value of heterogeneity test was less than 0.1. Results were expressed with odds ratios (OR) for dichotomous data and Table 1 The characteristics of trials Sample size Total IVIG dose (g/kg per day) Year Treatment group Control group Treatment group Control group Jadad score Blanchette et al. [9] g/kg per day), 1 day 1g/kg per day, 2day g/kg per day, 5day 0.4 g/kg per day, 5day 1 Xiang and Wang [11] g/kg per day, 5day 0.5 g/kg per day, 4day 1 Benesch et al. [11] g/kg per day, 2day 1g/kg per day, 2day 3 Wang et al. [] g/kg per day, 3day 0.4 g/kg per day, 5day g/kg per day, 5day 0.4 g/kg per day, 5day 3 Tang [] g/kg per day, 3day 2g/kg per day, 1day 2 Huang et al. [16] g/kg per day, 5day 0.4 g/kg per day, 5day 1 Cao and Ji [17] g/kg per day, 5day 0.4 g/kg per day, 5day 3 Zhu et al. [18] g/kg per day, 5day 0.5 g/kg per day, 4day 3 Xue et al. [19] g/kg per day, 5day 0.5 g/kg per day, 4day 3 Ge [20] g/kg per day, 1day 0.4 g/kg per day, 5day g/kg per day, 5day 0.4 g/kg per day, 5day 3 Treatment group had low-ivig and control group had. IVIG, intravenous immunoglobulin.

3 Meta-analysis of IVIG for acute ITP Qin et al. 7 weighted mean differences (WMD) for continuous data, and 95% confidence intervals (CI) were also counted. Heterogeneity between included studies was tested using the x 2 -test. If heterogeneity was present, we would try to check the reason out from aspects of study design and quality by using the methods of subgroup and sensitivity analysis. Funnel plot was drawn to assess the publication bias in our study. Results The search yielded 352 references, 16 in Pubmed and 336 in CBM-disc. Finally, RCTs included low-ivig and administration for ITP. One trial was excluded because the total dose of group was only 1 g/kg and the type of ITP was not in detail [22], and the patients in one study included acute ITP and chronic ITP [3]. Thirteen studies were included in our final review (Fig. 1). Eleven studies were published in Chinese [10,11, 21] and two in English [9,12]. The data of interest were extracted: effective rate, time of cessation of bleeding, time of platelet count beginning to rise, platelet count by the first week of treatment, the number of platelets after 2 weeks of treatment, time of platelet count to reach peak, peak value of platelet count after treatment, side-effects and rate of developing into Fig. 1 Articles retrieved for review: 352 chronic ITP. These studies contained 646 case series (low-ivig 332, 314) (Table 1). The baseline information of the literature was comparable. Effective rate The curative effect evaluation criteria developed by the second session of the ational Conference on Hematology in China [6] was referenced in our study: marked effect platelet count less than /l for 2 months, cessation of bleeding; well effect platelet count from 50 to /l for 2 months, cessation of bleeding; improvement increasing the platelet count, improving the bleeding symptoms; invalid no improvement of platelet and bleeding symptoms. The number of patients getting a good therapeutic effect was calculated as marked effect with well effect. Eight studies were included in our metaanalysis (469 patients) [10,14 18,20,21]. Pooled ORs was 1.00 ( ). Difference of effective rate was not statistically significant between low-ivig group and group (P ¼ 1). The P value of heterogeneity test was 0.97 (Fig. 2). Time of cessation of bleeding Data of time of cessation of bleeding were extracted from five studies (218 patients) [10,11,,,20], and the pooled WMD was 0.02 ( 0.35 to 0.38), suggesting there was no statistical significance between the two groups (P ¼ 0.94). The test for heterogeneity was negative with a P value of 0.93 (Fig. 3). Potentially relevant studies retrieved for more detailed evaluation: Potentially appropriate articles included in the systematic review: Studies included in meta-analysis: 327 articles excluded: no randomized controlled trial: 276 historical control: studies excluded: control group or treatment group with other drug: 5 treatment for chronic ITP: 5 2 articles excluded: total dose of treatment group did not fit the inclusion criteria: 1 patients included in acute ITP and chronic ITP: 1 Time of platelet count beginning to rise Three studies (2 patients) [,,20] reported the time of platelet count beginning to rise, and the pooled WMD was 0.04 ( 0.18 to 0.27). Difference was not statistically significant between low-ivig group and HD- IVIG group (P ¼ 0.71). The P value of the heterogeneity test was 0.89 (Fig. 4). Platelet count by the first week of treatment Three trials (102 patients) [14,19,21] were included in this analysis. Pooled WMD was 8.45 ( to 20.65), revealing there was no significant difference between low-ivig group and group (P ¼ 0.57). The test for heterogeneity was positive (P ¼ 0.67, Fig. 5). The number of platelets after 2 weeks of treatment Data of three RCTs (102 patients) [14,19,21] were useful for meta-analysis, and the pooled WMD was 5.44 (95% CI to 27.78), suggesting that there was no significant difference between the two groups (P ¼ 0.75). The test for heterogeneity was positive with a P value of 0.94 (Fig. 6). Flow chart for our studies. ITP, idiopathic thrombocytopenic purpura. Time of platelet count to reach peak Six studies [10,,18,21] (273 patients), reporting the index of time of platelet count to reach peak, were

4 716 Blood Coagulation and Fibrinolysis 2010, Vol 21 o 8 Fig. 2 n/ n/ 95% Cl 95% Cl Tang [] Cao and Ji [18] Huang et al. [16] Zhu et al. [18] Ge [20] / 12/ 9/10 /31 24/28 38/54 52/55 18/ 21/ 9/11 6/6 /30 24/27 37/53 49/53 19/ (0., ) 0.89 (0.12, 6.48) 0.49 (0.02,.92) 0.88 (0.27, 2.81) 0.75 (0., 3.72) 1.03 (0.45, 2.35) 1.41 (0.30, 6.65) 0.81 (0., 2.88) Total (95% Cl) (0.61, 1.63) Total events: 199 (low-ivig), 188 () Test for heterogeneity: Chi 2 = 1.82, df = 7 (P = 0.97), I 2 = 0% Test for overall effect: Z = 0.00 (P = 1.00) Favours low-ivig Favours Effective rate. Chi 2, x 2 -test; CI, confidence interval; df, degree of freedom; HD, high dose; I 2, heterogeneity index; IVIG, intravenous immunoglobulin; OR, odds ratio. reviewed. Pooled WMD was 0.2 ( 0.80 to 1.21). The difference was not statistically significant between low-ivig group and group (P ¼ 0.69). The P value of the heterogeneity test was 1 (Fig. 7). Peak value of platelet count after treatment For the analysis of peak value of platelet count after treatment, five randomized studies [10,,18] (2 patients) were counted. Pooled WMD was (95% CI to 21.97). Difference between two groups was not statistically significant (P ¼ 0.05). The test for heterogeneity was positive (P ¼ 1 Fig. 8). Side-effects Data were obtained from five studies [9,10,12,,18,21] (334 patients) and pooled to examine the side-effects of low-ivig treatment and administration. Pooled OR was 0.39 ( ). The difference was markedly significant between low-ivig group and group (P ¼ 0.01). The test for heterogeneity was positive with a P value of 0.28 (Fig. 9). Rate of developing into chronic idiopathic thrombocytopenic purpura Most studies did not report whether there were some patients developing into chronic ITP, and only three studies [9,12,19] (129 patients) were recruited. Pooled OR was 1.62 ( ). Difference between low-ivig group and group was not significant (P ¼ 0.32). The P value of the heterogeneity test was 0.57 (Fig. 10). Subgroup analysis according to age or race According to age of the patients in investigations, we divided the included studies into two groups, children group and adults group. In children group, eight studies [9,,16,19] were included for our analysis, and four studies [14,17,20,21] were recruited into our analysis for adults. One report [18], including children and adults, was excluded from our subgroup analysis. Pooled ORs or WMDs in children or adults were calculated, and we found results of effective rate, time of cessation of bleeding, time of platelet count beginning to rise, platelet Fig. 3 Xiang and Wang [11] Wang et al. [] Tang [] Ge [21] Total () (1.47) 4.30 (1.00) 3.70 (1.40) 3.40 (1.50) 2.42 (1.46) (1.44) 4.20 (1.00) 3.40 (1.30) 3.30 (1.30) 2.41 (1.50) ( 1.12, 0.56) 0.10 ( 0.79, 0.99) 0.30 ( 0.70, 1.30) 0.10 ( 1.30, 1.50) 0.01 ( 0.55, 0.57) ( 0.35, 0.38) Test for heterogeneity: Chi 2 = 0.83, df = 4 (P = 0.93), I 2 = 0% Test for overall effect: Z = 0.08 (P = 0.94) Favours low-ivig Favours Time of cessation of bleeding. Chi 2, x 2 -test; CI, confidence interval; df, degree of freedom; HD, high dose; I 2, heterogeneity index; IVIG, intravenous immunoglobulin; SD, standard deviation; WMD, weighted mean differences.

5 Meta-analysis of IVIG for acute ITP Qin et al. 717 Fig. 4 Wang et al. [] Tang [] Ge [21] (0.60) 1.50 (0.60) 2. (0.85) (0.50) 1.40 (0.50) 2.14 (0.81) ( 0.31, 0.51) 0.10 ( 0.45, 0.65) 0.01 ( 0.32, 0.30) Total () ( 0.18, 0.27) Test for heterogeneity: Chi 2 = 0., df = 2 (P = 0.89), I 2 = 0% Test for overall effect: Z = 0.37 (P = 0.71) ( 0.35, 0.38) Favours low-ivig Favours Time of platelet count beginning to rise. Chi 2, x 2 -test; CI, confidence interval; df, degree of freedom; HD, high dose; I 2, heterogeneity index; IVIG, intravenous immunoglobulin; SD, standard deviation; WMD, weighted mean differences. count by the first week of treatment, the number of platelets after 2 weeks of treatment, time of platelet count to reach peak, peak value of platelet count after treatment and rate of developing into chronic ITP in two groups were similar to the results of overall analyses. However, for the analysis of side-effects, four studies [9,10,12,] on children group and one study [21] for adults were only included. The pooled OR of side-effects suggested that there was no statistical difference between low-ivig and for children or adults (children, OR ¼ 0.53, , P ¼ 0.17; adults, OR ¼ 0.64, , P ¼ 0.64). However, the result was different from that of overall patients (OR ¼ 0.39, , P ¼ 0.01). We also conducted the subgroup analysis on the basis of race, but there were two studies [9,12] conducted for whites and we only recruited the indicators of side-effects and rate of developing into chronic ITP from the Caucasian reports into this meta-analysis. The P value of side-effects in Asian patients was 0.02, but 0. for whites (Asians, OR ¼ 0.22, ; whites, OR ¼ 0.56, ), and there were four included studies [10,,18,21] for Asians. For the analysis of rate of developing into chronic ITP, there was no marked difference between low-ivig and for Asians or Caucasian patients (Asians, OR ¼ 0.31, , P ¼ 0.48; whites, OR ¼ 1.98, , P ¼ 0.19), and we only recruited one report [19] into our analysis for Asians. It was difficult to draw a stable conclusion. However, the result was similar to that of overall patients (OR ¼ 1.62, , P ¼ 0.32). Testing for publication bias A funnel plot was drawn to assess publication bias. For most indicators, there were too few studies included. We only drew funnel plot for the effective rate in our metaanalysis. Published bias existed in our study (Fig. 11). Discussion We have endeavoured to provide enough information to help physician and stakeholders choose the most appropriate treatment for patients with acute ITP. This metaanalysis might be value for the doctor to choose a suitable scheme, especially in some countries where the cost of is much more than that of using low-ivig for 1 5 days. The evidence found in this systematic review stated that effective rate, time of cessation of bleeding, time of platelet count beginning to rise, platelet count by the first week of treatment, the number of platelets after 2 weeks of treatment, time of platelet count to reach peak, peak value of platelet count after treatment and rate of developing into chronic ITP were not statistically significant between low-ivig group and group, but the side-effects in administration group might be more prone to happen. This systematic review Fig. 5 Xue et al. [19] (92.40) 2. (53.51) (60.18) (1.90) (0.78) (59.02) ( 1.98, 49.78) 29. ( , 57.84) 1.32 ( 34.36, 31.72) Total () Test for heterogeneity: Chi 2 = 0.81, df = 2 (P = 0.67), I 2 = 0% Test for overall effect: Z = 0.57 (P = 0.57) Favours low-ivig Favours 8.45 ( 37.55, 20.65) Platelet count by the first week of treatment. Chi 2, x 2 -test; CI, confidence interval; df, degree of freedom; HD, high dose; I 2, heterogeneity index; IVIG, intravenous immunoglobulin; SD, standard deviation; WMD, weighted mean differences.

6 718 Blood Coagulation and Fibrinolysis 2010, Vol 21 o 8 Fig. 6 or sub-category Xue et al. [19] (1.6.40) (66.74) (80.81) (96.80) (97.46) (80.) ( 83.86, 73.26) ( 79.12, 49.30) 0.90 ( 45.54, 43.74) Total () Test for heterogeneity: Chi 2 = 0.12, df = 2 (P = 0.94), I 2 = 0% Test for overall effect: Z = 0.32 (P = 0.75) Favours low-ivig Favours 5.44 ( 38.65, 27.78) The number of platelets after 2 weeks of treatment. Chi 2, x 2 -test; CI, confidence interval; df, degree of freedom; HD, high dose; I 2, heterogeneity index; IVIG, intravenous immunoglobulin; SD, standard deviation; WMD, weighted mean differences. strongly supported that low-ivig could get a similar efficacy as and reduce the adverse effects. Acute ITP, as one of the most common bleeding disorders, may develop into serious spontaneous bleeding when the platelet count is under /ml [2]. Intracranial haemorrhage and gastrointestinal bleeding are the most serious complications, which can jeopardize the life of patient with acute ITP in a short time [2]. Therefore, how to increase the platelet count rapidly is much important in clinical administration. In the past decades, adrenal cortex hormones was used in the treatment of acute ITP, but there were some shortcomings that the elevated platelet was instable and haemorrhage symptom was easy to recur [ ]. As a substituted agent for steroids, IVIG can increase the platelet count rapidly and provide a positive effect to prevent the spontaneous bleeding for patient with acute ITP [,24]. However, the shortcoming of the IVIG regimen is that it is more expensive than other regimens and only stays in the body for a short time. During the treatment period of IVIG, the platelet count reaches the peak on the first week, begins to decline after the first week and falls to a minimum value at the third week [26]; the plasma level of immunoglobulin (IgG) restores to its original level [27]. It seems the platelet count may be related to the plasma level of IVIG. In other words, the dosage of infused IVIG may have a positive correlation with the therapeutic effect. Seifried et al. [28] found that there existed dose response relationship by analyzing the relation between IVIG concentration and platelet count in three patients with ITP, and time of platelet to reach peak or peak value of platelet count had a significant correlation with the concentration of IVIG in serum. On the contrary, Cohen [29] reported that low dose of IVIG had been proposed to be as effective as higher doses. Emmerich et al. [30] found that the onset and duration of response as well as the peak platelet count was independent to the doses of IVIG. If the reduced dose were as effective as high dosage, it would also reduce the cost of treatment in some developing countries such as China. In order to explore these problems, we conducted this meta-analysis. Our result showed that the differences of therapeutic effects between low-ivig treatment administration and treatment protocol were not statistically significant. The outcome predicted that the onset and duration of response of acute ITP might be independent to the plasma level of IVIG. In our study, the indicators of time of cessation of bleeding, time of platelet count beginning to rise and time of platelet to reach peak were a little favourable to the group (WMD ¼ 0.02, WMD ¼ 0.04 and WMD ¼ 0.20, respectively), although the differences Fig. 7 Wang et al. [] Tang [] Zhu et al. [18] (2.63) 8.40 (3.30) (7.10) 8.30 (3.00) (5.80) (6.40) Total () 142 Test for heterogeneity: Chi 2 = 0.12, df = 5 (P = 1.00), I 2 = 0% Test for overall effect: Z = 0.39 (P = 0.69) (2.53) 8.20 (3.10) 9.60 (8.80) 8.20 (2.90) (7.30) 9.35 (7.90) ( 1.37, 1.61) 0.20 ( 2.17, 2.57) 0.80 ( 5.52, 7.12) 0.10 ( 2.87, 3.07) 0.20 ( 2.30, 2.70) 0.75 ( 3.24, 4.74) ( 0.80, 1.21) Favours low-ivig Favours Time of platelet count to reach peak. Chi 2, x 2 -test; CI, confidence interval; df, degree of freedom; HD, high dose; I 2, heterogeneity index; IVIG, intravenous immunoglobulin; SD, standard deviation; WMD, weighted mean differences.

7 Meta-analysis of IVIG for acute ITP Qin et al. 719 Fig. 8 Wang et al. [] Tang [] Zhu et al. [18] (122.00) (114.00) 6.30 (105.30) (116.00) 9.00 (1.00) (7.00) (118.00) (143.10) (120.00) (143.00) ( 89.97, 59.97) ( , 72.27) ( , 78.85) ( , 97.95) 5.00 ( 54.22, 44.22) Total () Test for heterogeneity: Chi 2 = 0.14, df = 4 (P = 1.00), I 2 = 0% Test for overall effect: Z = 0.67 (P = 0.50) Favours low-ivig Favours ( 44.90, 21.97) Peak value of platelet count after treatment. Chi 2, x 2 -test; CI, confidence interval; df, degree of freedom; HD, high dose; I 2, heterogeneity index; IVIG, intravenous immunoglobulin; SD, standard deviation; WMD, weighted mean differences. were not statistically significant (P ¼ 0.94, P ¼ 0.71 and P ¼ 0.69, respectively). As we all know, these indexes are very important for preventing some serious complications such as intracranial bleeding. We speculate that the HD- IVIG regimen can improve the symptom of severe bleeding more quickly than the low-ivig [31] and prevent some life-threatening haemorrhages that endanger the health of patient with acute ITP. Therefore, an individualized dosing regimen should be applied according to patients morbid condition. A reasonable treatment protocol may get a good efficacy in individual patient. Side-effects are considered IVIG-related if they happen within 24 h of an infusion [3]. Some studies had reported the side-effect after using the IVIG [9,10,12,,18,21], such as headache, nausea, vomiting, fever and rash, but most of them were mild and happened to the patients after intravenous administration of. Increasing the infusion rates in a shorter duration could elevate the frequency of adverse events [32]. It seems that the plasma level of infused IVIG may partly be related to the occurrence of adverse reactions. In our study, the low- IVIG regimen could reduce the adverse effects in clinical administration. For the analysis of whether there were some patients developing chronic ITP, only three reports described [9,12,19]. The difference between low-ivig group and group was not statistically significant (P ¼ 0.32). The onset of patients with acute ITP becoming chronic ITP seemed to be independent of the dose of IVIG. However, there is not enough clinical evidence to confirm this at present. When we performed the subgroup analysis according to age or race, most of the results were similar to those of overall patients. However, the difference of sideeffects between low-ivig and in children or adults was not statistically significant, and they were similar to the result of whites. We only recruited a small number of studies into our investigation and it might affect the stabilization ofourresults.toacertain extent, it was difficult to draw a convincing conclusion from this subgroup analysis of age or race. In conclusion, administration of low-ivig may have a similar role as and can reduce the incidence of side-effects. It should be recommended within clinical practice guidelines and public health policies. Fig. 9 n/ n/ Blanchette et al. [9] Benesch et al. [12] Wang et al. [] Zhu et al. [18] 6/34 0/ 2/17 0/ 0/54 2/ 6/35 0/ 7/17 1/ 7/53 3/ (0.30, 3.60) ot estimable 0.19 (0.03, 1.11) 0.27 (0.01, 7.19) 0.06 (0.00, 1.02) 0.64 (0.10, 4.19) Totak () (0.18, 0.83) Total events: 10 (), 24 () Test for hererogeneity: Chi 2 = 5.04, df = 4 (P = 0.28), I 2 = 20.6% Test for overall effect: Z = 2.45 (P =.01) Favours low-ivig Favours Side-effects. Chi 2, x 2 -test; CI, confidence interval; df, degree of freedom; HD, high dose; I 2, heterogeneity index; IVIG, intravenous immunoglobulin; OR, odds ratio.

8 720 Blood Coagulation and Fibrinolysis 2010, Vol 21 o 8 Fig. 10 n/ n/ Blanchette et al. [9] Benesch et al. [12] Xue etal. [19] 7/34 5/17 0/ 4/35 3/17 1/ (0.53, 7.62) 1.94 (0.38, 9.88) 0.31 (0.01, 8.30) Totak () (0.62, 4.21) Total events: 12 (low-ivig), 8 () Test for hererogeneity: Chi 2 = 1.12, df = 2 (P = 0.57), I 2 = 0% Test for overall effect: Z = 0.99 (P = 0.32) Favours low-ivig Favours Rate of developing into chronic idiopathic thrombocytopenic purpura. Chi 2, x 2 -test; CI, confidence interval; df, degree of freedom; HD, high dose; I 2, heterogeneity index; IVIG, intravenous immunoglobulin; OR, odds ratio. Fig SE (log OR) Publication bias. OR, odds ratio; SE, standard error. Limitations Limitations of our review mainly concern the methodological quality of some recruited studies. Limiting factors include the absence of reporting adverse events and rate of patients developing chronic ITP in most studies. The follow-up was too short in most reports and only three studies [9,12,19] had reached 6 months duration. The number of the included investigations was small, and it was difficult to conduct a subgroup analysis according to age or race and draw a convincing conclusion. By composing a funnel plot for the indicator of effective rate, we found there existed publication bias in our study. The conservative assumption we made was that a number of small negative unpublished trials were missed, and our results might be biased toward showing a more positive effect than truly exists. The results of our meta-analysis may be robust on the basis of the fact that they had low statistical heterogeneity. Otherwise, the sample size in some studies was small [11,19] and the quality of methodology used in some reports was poor [10,11,,,16,20]. To a certain extent, it affected the conviction of our outcome. Acknowledgement The authors would like to gratefully acknowledge the most helpful comments on this study received from Professor Liang Rong, Department of Pediatricneonatology, Baylor College of Medicine, Houston, Texas, USA. References 1 Vranou M, Platokouki H, Pergantou H, Aronis S. Recurrent idiopathic thrombocytopenic purpura in childhood. Pediatr Blood Cancer 2008; 51: Shahgholi E, Vosough P, Sotoudeh K, Arjomandi K, Ansari S, Salehi S, et al. Intravenous immune globulin versus intravenous anti-d immune globulin for the treatment of acute immune thrombocytopenic purpura. Indian J Pediatr 2008; 75: Warrier I, Bussel JB, Valdez L, Barbosa J, Beardsley DS. Safety and efficacy of low-dose intravenous immune globulin (IVIG) treatment for infants and children with immune thrombocytopenic purpura. Low-Dose IVIG Group. J Pediatr Hematol Oncol 1997; 19: Paran D, Herishanu Y, Elkayam O, Shopin L, Ben-Ami R. Venous and arterial thrombosis following administration of intravenous immunoglobulins. Blood Coagul Fibrinolysis 2005; 16: Bussel JB. The use of intravenous gamma-globulin in idiopathic thrombocytopenic purpura. Clin Immunol Immunopathol 1989; 53:S Zhang Z. Blood disease diagnosis and efficacy of the standard. Beijing: Beijing Sci Press; 1998, pp Imbach P, Barandun S, d Apuzzo V, Baumgartner C, Hirt A, Morell A, et al. High-dose intravenous gammaglobulin for idiopathic thrombocytopenic purpura in childhood. Lancet 1981; 1: Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996; 17: Blanchette V, Imbach P, Andrew M, Adams M, McMillan J, Wang E, et al. Randomised trial of intravenous immunoglobulin G, intravenous anti-d, and oral prednisone in childhood acute immune thrombocytopenic purpura. Lancet 1994; 344: Xue TY. Comparison of the effects of different dose of gamma globulin on acute idiopathic thrombocytopenic purpura in children. Acad Med Xuzhou 2000; 20: Xiang YJ, Wang JL. Effectiveness of sub-ivig for idiopathic thrombocytopenic purpura. Zhejiang Clin Med 2002; 4: Benesch M, Kerbl R, Lackner H, Berghold A, Schwinger W, Triebl-Roth K, Urban C. Low-dose versus high-dose immunoglobulin for primary treatment of acute immune thrombocytopenic purpura in children: results of a prospective, randomized single-center trial. J Pediatr Hematol Oncol 2003; : Wang J, Zhang MF, iu JH. Clinical observation of the therapeutic effects of different dose intravenous gammaglobulin in treating children with severe idiopathic thrombocytopenic purpura (ITP). Chin J Coal Industry Med 2003; 1:80.

9 Meta-analysis of IVIG for acute ITP Qin et al Zhang SW. Observation on efficacy of - on severe idiopathic thrombocytopenic purpura. Mod Diagn Treat 2004; : Tang GH. Therapeutic effects of intravenous gammaglobulin administration for idiopathic thrombocytopenic purpura. Mod Med Hyg 2006; 22: Huang AM, Wang YC, Zhu HL. Clinical observation of the therapeutic effect of combination sub-dose gamma globulin with glucocorticoid in children with idiopathic thrombocytopenic purpura. Mod J Integr Tradit Chin West Med 2007; 16: Cao WB, Ji YH. Clinical observation of the curative effect of sub-standarddose gamma globulin plus glucocorticoid in patient with idiopathic thrombocytopenic purpura. Chin J Tradit Chin Med West Med Crit Care 2007; l: Zhu FY, Zhang Z, Cao ZH, Jiang ZP. Comparative study of intravenous gammaglobulin administration for thrombocytopenic purpura. Chin Contemp Med Sci 2007; 6: Xue AM, Qiao Y, Cao HZ. Observation of the effects of sub-dose intravenous gammaglobulin administration for idiopathic thrombocytopenic purpura in children. J Med Theor Prac 2007; 20: Ge SB. Analysis of combination therapy using sub-dose intravenous gammaglobulin in patient with idiopathic thrombocytopenic purpura in clinical practice. China Prac Med 2009; 4: Yang ZR. Treatment of severe idiopathic thrombocytopenic purpura with two different doses of gamma globulin. J Clin Intem Med 2009; 26: Imbach P, Wagner HP, Berchtold W, Gaedicke G, Hirt A, Joller P, et al. Intravenous immunoglobulin versus oral corticosteroids in acute immune thrombocytopenic purpura in childhood. Lancet 1985; 2: Rosthoj S, ielsen S, Pedersen FK. Randomized comparison of intravenous immunoglobulin and methylprednisolone pulse therapy in children with newly diagnosed idiopathic thrombocytopenia purpura. Danish ITP Group. Ugeskr Laeger 1998; 160: Blanchette VS, Luke B, Andrew M, Sommerville-ielsen S, Barnard D, de Veber B, et al. A prospective, randomized trial of high-dose intravenous immune globulin G therapy, oral prednisone therapy, and no therapy in childhood acute thrombocytopenic purpura. J Pediatr 1993; 1: Erduran E, Aslan Y, Gedik Y, Orhan F. A randomized and comparative study of intravenous immunoglobulin and mega dose methylprednisolone treatment in children with acute idiopathic thrombocytopenic purpura. Turk Pediatr 2003; 45: Ancona KG, Parker RI, Atlas MP, Prakash D. Randomized trial of high-dose methylprednisolone versus intravenous immunoglobulin for the treatment of acute idiopathic thrombocytopenic purpura in children. J Pediatr Hematol Oncol 2002; 24: egi VS, Elluru S, Sibéril S, Graff-Dubois S, Mouthon L, Kazatchkine MD, et al. Intravenous immunoglobulin: an update on the clinical use and mechanisms of action. J Clin Immunol 2007; 27: Seifried E, Pindur G, Stötter H, Wiesneth M, Rasche H, Heimpel H. Treatment of refractory chronic idiopathic thrombocytopenic purpura with high dose intravenous immunoglobulin. Blut 1984; 48: Cohen JI. Gamma globulin for idiopathic thrombocytopenic purpura. Engl J Med 1982; 307: Emmerich B, Hiller E, Woitinas F, Maubach PA, Riess H, erl C, et al. Dose response relationship in the treatment of idiopathic thrombocytopenic purpura with intravenous immunoglobulin. Klin Wochenschr 1987; 65: Godeau B, Caulier MT, Decuypere L, Rose C, Schaeffer A, Bierling P. Intravenous immunoglobulin for adults with autoimmune thrombocytopenic purpura: results of a randomized trial comparing 0.5 and 1 g/kg b.w. Br J Haematol 1999; 107: Stangel M, Hartung HP, Marx P, Gold R. Side effects of high-dose intravenous immunoglobulins. Clin europharmacol 1997; 20: