Cell Biochem Biophys DOI 10.1007/s12013-014-0265-1
ORIGINAL PAPER
A Meta-analysis of the Protective Effect of Recombinant Human Erythropoietin (rhEPO) for Neurodevelopment in Preterm Infants Huiping Wang • Lan Zhang • Yan Jin
Ó Springer Science+Business Media New York 2014
Abstract The purpose of this study is to assess the efficacy and safety of recombinant human erythropoietin (rhEPO) for improving neurodevelopment outcomes in preterm infants. According to the requirements of Cochrane systematic review, a literature search was performed among PubMed, EMBASE, Cochrane Central Register of Controlled Trials, Chinese Biomedical Literature Database, Chinese National Knowledge Infrastructure, Wan Fang Data, and VIP INFORMATION from the establishment of the database from January 1999 to December 2011. Quality assessments of clinical trials were carried out. Randomized controlled trials (RCTs) or quasi-RCTs with rhEPO in preterm infants were enrolled, and RevMan5.0 software was used for meta-analysis. Data extraction, quality assessment, and meta-analysis for the results of homogeneous studies were done by two reviewers. The trials were analyzed using weighted mean difference (WMD) for continuous data and odds ratio (OR) for dichotomous data, both kinds of data were expressed by 95 % CI. For homogenous data (P C 0.10), fixed effect model was calculated. Two RCTs and 3 quasi-RCTs including 233 preterm infants (119 of treatment group and 114 of control group) were included in the analysis. The results of quality assessment were that 1 study was A, 1 was B, and 3 were C. There was evidence of a significant effect of therapeutic rhEPO on the outcomes of MDI scores [WMD = 7.77, 95 % CI (3.49–12.06), P = 0.0004], PDI scores [WMD = 3.85, 95 % CI (0.62–7.09), P = 0.02] at 18–22 months and NBNA scores [WMD = 1.96, 95 % CI (1.56–2.37), P \ 0.00001] at 40 weeks of corrected gestational age. However, rhEPO had H. Wang (&) L. Zhang Y. Jin Department of Pediatrics, The Second Affiliated Hospital, Medical School of Xi’an Jiaotong University, Shaanxi, China e-mail: [email protected]
no effect on MDI \70 (OR = 0.70, 95 % CI 0.31–1.61), PDI \70 (OR = 2. 46, 95 % CI 0.94–6.45), cerebral palsy (OR = 1.08, 95 % CI 0.39–2.99), blindness (OR = 0.34, 95 % CI 0.01–8.56), and hearing loss (OR = 1.04, 95 % CI 0.06–17.15). There were no differences between groups with respect to the percentage of preterm infants with severe retinopathy of prematurity of stage III or above (OR = 1.30, 95 % CI 0.50–3.43), severe intraventricular hemorrhage of stage III or above (OR = 2. 91, 95 % CI 0.64–13.23), necrotizing enterocolitis (OR = 0.57, 95 % CI 0.13–2.54), and borderline personality disorder (OR = 1. 06, 95 % CI 0.50–2.26). The rhEPO treatment has beneficial effect on the neurodevelopment outcomes without severe adverse side effect in preterm infants. Keywords Recombinant human erythropoietin Preterm Neural system Meta-analysis
Introduction Neonatal brain injury, including commonly hypoxic-ischemic encephalopathy (HIE), periventricular leuko-malacia (PVL), cerebral-intraventricular hemorrhage (IVH), and hyperoxic brain damage, is an important cause of neonatal mortality and subsequent sequelae such as cerebral palsy, mental retardation, learning disability, and epilepsy [1–3]. Brain injury results in poor neurodevelopmental outcomes including cerebral palsy, mental retardation, hearing or visual impairment, and attention deficit hyperactivity disorder [4]. Although there is increasing evidence about underlying mechanisms and growing number of studies about treatment strategies, a safe and an effective treatment for both preterm and term infants regardless of the severity of the brain injury is urgently needed.
123
Cell Biochem Biophys
Erythropoietin (EPO), which was originally identified for its role in erythropoiesis, is widely used for the treatment of anemia in premature infants [5]. Surprisingly, EPO was found to play a variety of roles in modulation of the inflammatory response and has vasogenic effects [6]. Neuroprotection with EPO has been documented in spinal cord injury, traumatic brain injury, ischemic stroke, and perinatal asphyxia [7]. The neurodevelopmental outcomes of 20 premature infants treated with recombinant human erythropoietin (rhEPO) have been evaluated [8]. There was no adverse effect of EPO on neurologic outcome, cognitive outcome, or growth patterns and the rate of cognitive deficits. Later, developmental outcomes at 18–22 months’ corrected age in extreme low body weight (ELBW) infants treated with EPO were compared [9]. No difference between groups with respect to the percentage of infants with blindness, deafness or hearing loss, moderate to severe cerebral palsy was found. However, the limitation of this study was that the EPO treatment group included more IVH infants than that in the control group. The safety of EPO in preterm infants has been established by its high does administration in two trials. In a prospective trial of high-dose rhEPO in ELBW infants, the investigators compared 30 infants who were treated with high-dose rhEPO with 30 concurrent control subjects [10]. Early high-dose rhEPO is well tolerated by ELBW infants, causing no excess mortality. In another randomized, doubleblinded trial, the safety of administration of high-dose EPO (3,000 U/kg, 39) to preterm infants has been investigated [11]. No side effect of EPO such as IVH, ROH, and necrotizing enterocolitis (NEC) has been found [11]. Both of the above studies provide an important insight into the safety of preterm infants who received early high-dose rhEPO. These findings suggest that an early high-dose administration of rhEPO to preterm infants to improve neurodevelopmental outcome is feasible. Research findings from clinical studies regarding to the efficacy and safety of rhEPO for improving neurodevelopment outcomes in preterm infants are still controversial. Therefore, we conduct a meta-analysis aiming to assess the neuroprotective effect of rhEPO in preterm infants.
Materials and Methods Data Sources Studies were identified by extensively searching the PubMed, EMBASE, Cochrane Central Register of Controlled Trials, Chinese Biomedical Literature Database (CBM), Chinese National Knowledge Infrastructure (CNKI), Wan Fang Data, and VIP INFORMATION from January 1999 to December 2011. The search terms were (EPO or erythropoietin) and (preterm or premature or infants or
123
very low birth weight or extremely low birth weight) and (neurodevelopmental or brain or nerve). The language of the literatures was not restricted to English. In addition, references of the relevant articles were reviewed in order to identify additional studies not detected by the initial search. Study Selection Two reviewers (Yanping Mu and Xun Deng) independently searched literatures and examined relevant RCTs. Any disagreement about study selection or data extraction was resolved by consensus with the third reviewer (Yong Wang). For meta-analysis, all studies have to meet the following inclusion criteria: (1) A study described as RCT or quasiRCT. (2) Preterm infants (gestation \37 weeks). (3) Intervention measures: intravenous injection or hypodermic of rhEPO; total dose within the period of treatment [3,000 U kg-1; and administration of rhEPO started within the first 7 days of life and continued for at least 14 days. (4) Control group receiving a placebo or routine treatment. (5) Primary endpoints were MDI scores, PDI scores at 12–22 months or MDI scores \70, PDI scores \70; NBNA scores at 40 weeks of corrected gestational age; and serious nervous system sequel (cerebral palsy, blindness, or hearing loss). Safety indexes were severe retinopathy of prematurity (ROP) of stage III or above, severe IVH of stage III or above, NEC, borderline personality disorder (BPD). Non-RCT or quasi-RCT studies were excluded, as well as case reports, reviews with insufficient details to meet the inclusion criteria, abstracts in the proceedings of scientific conferences, and experimental trials; Premature infants with severe congenital malformation, hemolysis, anemia, polycythemia, serious infection, Intrauterine TORCH infection, congenital heart disease, and genetic metabolic disease were also excluded from the study. Data Extraction Two of the authors independently extracted data from the trials that met the inclusion criteria using. Authors would be contacted for missing data when necessary. For each trial, the following data were extracted: number of patients in each group; mean age and sex distribution of each group; gestational age; birth weight; drug regimen, including doses and treatment duration; and clinical outcome indicators (efficacy and safety of drugs). Assessment of Study Quality Quality assessment of the RCTs included in the meta-analysis was independently performed by the same reviewers according to Cochrane Handbook 5.0.1 and Juni et al. [12, 13], which assesses the descriptions of generation of random,
Cell Biochem Biophys Fig. 1 Studies selection process
allocation concealment, double blinding, and dropouts\withdrawals of the included trials. Each author rated the quality of the trials using Jadad grade (maximum grade = A; minimum grade = C; grade C B = good quality). Statistical Analysis Data were analyzed using Review Manager 5.1. Included articles were pooled and weighted. Odds ratio (OR) or weighted mean difference (WMD) and 95 % confidence interval (95 % CI) were calculated in a fixed-effects mode or in a random-effects model. Heterogeneity was assessed by calculating a v2 test and the quantity of heterogeneity was measured with I2 statistic. If heterogeneity (P \ 0.1 or I2 [ 50 %) was found among the trials, random-effects model would be chosen, otherwise fixed-effects model chosen. If heterogeneity was evident (I2 [ 70 %), the inferior quality study should be eliminated to analyze. A 2-tailed P value of less than 0.05 was deemed statistically significant.
Results Study Selection Process Our search yielded a total of 118 studies on recombinant human erythropoietin for neurodevelopment in preterm
infants. After reviewing each abstract or original publication and extracting data from the publications, five were finally selected [9, 14–17]. Studies including reviews (n = 22), non-RCTs (n = 35), foundational researches (n = 40), anemia studies (n = 6), full-term infants RCTs (n = 2) that do not meet the inclusion criteria of the metaanalysis (n = 8) were excluded. The details of the selection process were shown in Fig. 1. Study Characteristics The main characteristics of five included studies are presented in Table 1. Five studies involving 233 patients were ultimately confirmed that met the criteria for inclusion in the meta-analysis. All of the trials enrolled patients, two of which studied very low birth weight infant (VLBWI), three premature infants with gestational age\37 weeks and birth weight \2,500 g. And two studies adopted MDI and PDI scores, one bayley scale of infant development (BSID), one gessell scale, and three reported NBNA scores to assess clinical effective rates. The quality assessment of included studies is presented in Table 2. Two studies clarified adequate generation of random, one reported allocation concealment, two used double blinding and two reported numbers of dropouts\withdrawals. One was eventually assessed to be good in terms of methodology with a Jadad score A [12], one was B, the other three were C.
123
Cell Biochem Biophys Table 1 Basic data of included 5 literatures (x ± s) Study
Design
n1/n2
GA/weeks
Ohls [9]
RCT
51/51
26.3 ± 2.0/25.8 ± 1.7
Bierer et al. [14]
RCT
6/6
26.0 ± 1.1/26.9 ± 2.1
Wang [15]
quasiRCT
20/15
28–34
Wang et al. [16]
quasiRCT
20/20
31.5 ± 2.1/32.1 ± 2.5
He [17]
quasiRCT
22/22
29.4 ± 1.5/30.2 ± 1.3
Table 2 Quality assessment of included 5 literatures
Dosage/ U kg-1
Duration
Results
801 ± 139/783 ± 112
400
8–10 weeks
% MDI \70, % PDI \70
752 ± 150/810 ± 103
400
Until discharge, transfer death, or 35 completed weeks PMA
MDI, PDI, % MDI \70, % PDI \70
400
6 weeks
NBNA
1626 ± 338/1660 ± 255
250
4 weeks
MDI, PDI, NBNA
1135 ± 120/1075 ± 135
250
4 weeks
NBNA, Gesell
Birth weight/g
\2000
Study
Generation of random
Allocation concealed
Blind of outcome assessors
Loss of follow-up
Baseline similarity
Quality grade
Ohls [9]
Clear
Unknown
Yes
Yes
Statement
Similar
B
Bierer et al. [14]
Clear
Yes
Yes
Yes
Statement
Similar
A
Wang [15] Wang et al. [16]
Incomplete Incomplete
Yes Yes
Unknown Unknown
Unknown Yes
Statement Statement
Similar Similar
C C
He [17]
Incomplete
Yes
Unknown
Yes
Statement
Similar
Comparisons of Effectiveness Effect of rhEPO on MDI Two studies of Bierer et al. [14] and Wang et al. [16] provided specific data for analysis of MDI scores. MDI score in the rhEPO group [26 (50 %) of 52 patients] was higher than control group, while statistically significant difference was found [WMD = 7.77, 95 % CI (3.49–12.06), P = 0.0004] (Fig. 2). Data from two studies of Bierer et al. [14] and Ohls et al. [9] showed the incidence of MDI scores below 70 between the rhEPO group [51 (50 %) of 102 patients] and the control group [51 (50 %) of 102 patients] had no significant difference [OR = 0. 70, 95 % CI (0. 31–1. 61), P = 0.4] (Fig. 3). Effect of rhEPO on PDI Two studies of Bierer et al. [14] and Wang et al. [16] provided specific data for analysis of PDI scores. PDI score in the rhEPO group [26 (50 %) of 52 patients] was higher than control group, while statistically significant difference was found [WMD = 3.85, 95 % CI (0.62–7.09), P = 0.02] (Fig. 4).
123
Blind investigator
Data from two studies of Bierer et al. [14] and Ohls et al. [9] showed the incidence of PDI scores below 70 between the rhEPO group [51 (50 %) of 102 patients] and control group [51 (50 %) of 102 patients]. No statistically significant difference was found [OR = 2.46, 95 % CI (0. 94–6.45), P = 0.07] (Fig. 5). Effect of rhEPO on NBNA Scores Three studies [15–17] provided specific data for analysis of NBNA scores. We noted no significant heterogeneity between the trials (I2 = 0 %, P = 0.63). Meta-analysis of data using a fixed-effects model estimated NBNA scores in the rhEPO group. NBNA score in the rhEPO group [(52.1 %) of 119 patients] was higher than control group [(47.9 %) of 119 patients], while statistically significant difference was found [WMD = 1.96, 95 % CI (1.56–2.37), P \ 0.00001] (Fig. 6). Comparisons of Safety Adverse drug reaction data for severe ROP incidence, severe IVH (stage III or above) incidence, and BPD incidence in rhEPO group and control group were reported in
Cell Biochem Biophys
Fig. 2 Meta-analysis of MDI scores in rhEPO treatment and control groups
Fig. 3 Meta-analysis of MDI scores below 70 in rhEPO treatment and control groups
Fig. 4 Meta-analysis of PDI scores in rhEPO treatment and control groups
Fig. 5 Meta-analysis of PDI scores below 70 in rhEPO treatment and control groups
two studies [9, 14]. In the evaluable patients, the incidence of severe ROP (stage III or above) between the rhEPO group [58 (49.6 %) of the 117 patients] and the control group [59 (50.4 %) of the 117 patients] had no statistically
significant difference [OR = 1.30, 95 % CI (0.50–3.43), P = 0.59] (Fig. 7). The incidence of severe IVH (stage III or above) was 49.1 % (56/114) for rhEPO and 50.9 % (58/ 114) for comparator, while no statistically significant
123
Cell Biochem Biophys
Fig. 6 Meta-analysis of NBNA scores in rhEPO treatment and control groups
Fig. 7 Meta-analysis of the incidence of severe ROP (stage III or above) in rhEPO treatment and control groups
Fig. 8 Meta-analysis of the incidence of severe IVH (stage III or above) in rhEPO treatment and control groups
difference was found [OR = 2.91, 95 % CI (0.64–13.23), P = 0.17] (Fig. 8). And no statistically significant difference was found [OR = 1.06, 95 % CI (0.50–2.26), P = 0.87] (Fig. 9) between the rhEPO group [49.6 % (58/ 117)] and the control group [50.4 % (59/117)] about the incidence of BPD.
Discussion We conducted this meta-analysis to assess the efficacy and safety of rhEPO for improving neurodevelopment
123
outcomes in preterm infants. The MDI scores, PDI scores, and NBNA scores of rhEPO were higher than control group, and superior to control with statistically significant difference [MDI scores: WMD = 7.77, 95 % CI (3.49–12.06), P = 0.0004; PDI scores: WMD = 3.85, 95 % CI (0.62–7.09), P = 0.02; WMD = 1.96, 95 % CI (1.56–2.37), P \ 0.00001]. However, the incidence of PDI scores below 70 and the incidence of MDI scores below 70 had no significant difference between rhEPO and control group [PDI scores below 70 incidence: OR = 0. 70, 95 % CI (0.31–1.61), P = 0.4; MDI scores below 70 incidence: OR = 2.46, 95 % CI (0. 94, 6.45), P = 0.07]. The safety
Cell Biochem Biophys
Fig. 9 Meta-analysis of the incidence of BPD in rhEPO treatment and control groups
profile analysis regarding the incidence of adverse drug reactions had no significant difference between rhEPO and control group [severe ROP (stage III or above) incidence: OR = 1.30, 95 % CI (0.50–3.43; severe IVH (stage III or above): OR = 2.91, 95 % CI (0.64–13.23), P = 0.17; BPD incidence: OR = 1.06, 95 % CI (0.50–2.26), P = 0.87]. This meta-analysis indicated that rhEPO could be used to improve neurodevelopment outcomes in preterm infants. Recently, a long-term retrospective study firstly has confirmed the neuroprotective benefits of rEPO for ELBW infants at school age. ELBW infants receiving rEPO scored significantly better than untreated children in the overall developmental assessment (55 versus 39 % normally developed) as well as in the psychological examination (mean composite HAWIK-III IQ score, 90.8 versus 81.3). Also, this study confirms the neuroprotective effect of rEPO in ELBW infants with IVH. Furthermore, rEPO treatment works well even when given days after the onset of brain damage suggests a wider window of therapy. However, this study has a few limitations such as not being a randomized control trial and not a strict subgroup analysis for IVH mild or severe degrees [18]. Therefore, rhEPO may be recommended as a routine therapy for improving neurodevelopment outcomes in preterm infants in future, but more RCTs are needed to further study. However, some possible limitations in this meta-analysis should be acknowledged. First, our findings may be affected by the quality of trials included in the meta-analysis. Only one reported allocation concealment, two of the included trials were double blinding, and two with clarified adequate generation of random. Only one was assessed to be good in terms of methodology with a Jadad score A, three were C. Second, this meta-analysis is based on a relatively small number of RCTs and we acknowledge that using a limited number of studies raises the possibility of a second-order sampling error [19]. However, meta-analyses often include small numbers of studies; Higgins [19] evaluated 39 Cochrane reviews and found that 67 % of them included B5 studies and 20 % included B10 studies.
Therefore, more RCTs comparing rhEPO with control group for protection of neurodevelopment in preterm infants are needed to be carried out in future. Third, the eligibility criteria for inclusion of preterm infants were different from each other (Table 1), which might influence the obvious consistency of effects across those included studies and cause the between-study heterogeneity. Thus, to ensure uniformity in defining preterm infants’ characteristics and outcome measures, an individual patient data meta-analysis needs to be performed. Finally, the distinct differences in dose, timing, and pharmaceutical manufacturers of rhEPO used exist. In conclusion, although some limitations exist in this meta-analysis, based on the result of our meta-analysis, it is suggested that rhEPO is an effective and a relatively safe option for improving neurodevelopment outcomes in preterm infants. However, the long-term effects of rhEPO for neurodevelopment outcomes in preterm infants still needs to be further investigated.
References 1. Gonzalez, F. F., & Ferriero, D. M. (2009). Neuroprotection in the newborn infant. Clinics in Perinatology, 36, 859–880. 2. Volpe, J. J. (2003). Cerebral white matter injury of the premature infant-more common than you think. Pediatrics, 112, 176–180. 3. Felderhoff-Mueser, U., Bittigau, P., Sifringer, M., Jarosz, B., Korobowicz, E., Mahler, L., et al. (2004). Oxygen causes cell death in the developing brain. Neurobiology of Diseases, 17, 273–282. 4. McPherson, R. J., & Juul, S. E. (2008). Recent trends in erythropoietin-mediated neuroprotection. International Journal of Developmental Neuroscience, 26, 103–111. 5. Ghezzi, P., Bernaudinb, M., Bianchic, R., Blomgren, K., Brines, M., Campana, W., et al. (2010). Erythropoietin: Not just about erythropoiesis. Lancet, 375, 2142. 6. Gonzalez, F. F., & Ferriero, D. M. (2008). Therapeutics for neonatal brain injury. Pharmacology and Therapeutics, 120, 43–53. 7. McPherson, R. J. (2009). Can erythropoietin improve developmental outcomes for preterm infants? Pediatrics, 124, e805–e806.
123
Cell Biochem Biophys 8. Newton, N. R., Leonard, C. H., Piecuch, R. E., & Phibbs, R. H. (1999). Neurodevelopmental outcome of prematurely born children treated with recombinant human erythropoietin in infancy. Journal of Perinatology, 19, 403–406. 9. Ohls, R. K., Ehrenkranz, R. A., Das, A., Dusick, A. M., Yolton, K., Romano, E., et al. (2004). Neurodevelopmental outcome and growth at 18 to 22 months’ corrected age in extremely low birth weight infants treated with early erythropoietin and iron. Pediatrics, 114, 1287–1291. 10. Juul, S. E., McPherson, R. J., Bauer, L. A., Ledbetter, K. J., Gleason, C. A., & Mayock, D. E. (2008). A phase I/II trial of high-dose erythropoietin in extremely low birth weight infants: Pharmacokinetics and safety. Pediatrics, 122, 383–391. 11. Fauchere, J. C., Dame, C., Vonthein, R., Koller, B., Arri, S., Wolf, M., et al. (2008). An approach to using recombinant erythropoietin for neuroprotection in very preterm infants. Pediatrics, 122, 375–382. 12. Juni, P., Altman, D. G., & Egger, M. (2001). Systematic reviews in health care: Assessing the quality of controlled clinical trials. British Medical Journal, 323, 42–46. 13. Jadad, A. R., Moore, R. A., Carroll, D., Jenkinson, C., Reynolds, D. J., Gavaghan, D. J., et al. (1996). Assessing the quality of reports of randomized clinical trials: Is blinding necessary? Controlled Clinical Trials, 17, 1–12.
123
14. Bierer, R., Peceny, M. C., Hartenberger, C. H., & Ohls, R. K. (2006). Erythropoietin concentrations and neurodevelopmental outcome in preterm infants. Pediatrics, 118(3), 635–640. 15. Wang, Y. H. (2005). Effects of erythropoietin on anemia and nervous system of premature infants. Journal of Medical Forum, 26(7), 49–52. 16. Wang, Y. H., Zhu, C. L., Cheng, X. Y., Hu, S. H., Feng, H. Q., & Wang, X. Y. (2006). Effects of recombinant human erythropoietin on neurobehavioral development in premature infants. Chinese Journal of Practical Pediatrics, 12(1), 59–60. 17. He, J. S., Huang, Z. L., Yang, H., Weng, K. Z., & Zhu, S. B. (2008). Early use of recombinant human erythropoietin promotes neurobehavioral development in preterm infants. Chinese Journal of Contemporary Pediatrics, 10(5), 586–588. 18. Neubauer, A. P., Voss, W., Wachtendorf, M., & Jungmann, T. (2010). Erythropoietin improves neurodevelopmental outcome of extremely preterm infants. Annals of Neurology, 67, 657–666. 19. Higgins, J., Thompson, S., Deeks, J., & Altman, D. (2002). Statistical heterogeneity in systematic reviews of clinical trials: A critical appraisal of guidelines and practice. Journal of Health Services Research and Policy, 7(1), 51–61.
ORIGINAL PAPER
A Meta-analysis of the Protective Effect of Recombinant Human Erythropoietin (rhEPO) for Neurodevelopment in Preterm Infants Huiping Wang • Lan Zhang • Yan Jin
Ó Springer Science+Business Media New York 2014
Abstract The purpose of this study is to assess the efficacy and safety of recombinant human erythropoietin (rhEPO) for improving neurodevelopment outcomes in preterm infants. According to the requirements of Cochrane systematic review, a literature search was performed among PubMed, EMBASE, Cochrane Central Register of Controlled Trials, Chinese Biomedical Literature Database, Chinese National Knowledge Infrastructure, Wan Fang Data, and VIP INFORMATION from the establishment of the database from January 1999 to December 2011. Quality assessments of clinical trials were carried out. Randomized controlled trials (RCTs) or quasi-RCTs with rhEPO in preterm infants were enrolled, and RevMan5.0 software was used for meta-analysis. Data extraction, quality assessment, and meta-analysis for the results of homogeneous studies were done by two reviewers. The trials were analyzed using weighted mean difference (WMD) for continuous data and odds ratio (OR) for dichotomous data, both kinds of data were expressed by 95 % CI. For homogenous data (P C 0.10), fixed effect model was calculated. Two RCTs and 3 quasi-RCTs including 233 preterm infants (119 of treatment group and 114 of control group) were included in the analysis. The results of quality assessment were that 1 study was A, 1 was B, and 3 were C. There was evidence of a significant effect of therapeutic rhEPO on the outcomes of MDI scores [WMD = 7.77, 95 % CI (3.49–12.06), P = 0.0004], PDI scores [WMD = 3.85, 95 % CI (0.62–7.09), P = 0.02] at 18–22 months and NBNA scores [WMD = 1.96, 95 % CI (1.56–2.37), P \ 0.00001] at 40 weeks of corrected gestational age. However, rhEPO had H. Wang (&) L. Zhang Y. Jin Department of Pediatrics, The Second Affiliated Hospital, Medical School of Xi’an Jiaotong University, Shaanxi, China e-mail: [email protected]
no effect on MDI \70 (OR = 0.70, 95 % CI 0.31–1.61), PDI \70 (OR = 2. 46, 95 % CI 0.94–6.45), cerebral palsy (OR = 1.08, 95 % CI 0.39–2.99), blindness (OR = 0.34, 95 % CI 0.01–8.56), and hearing loss (OR = 1.04, 95 % CI 0.06–17.15). There were no differences between groups with respect to the percentage of preterm infants with severe retinopathy of prematurity of stage III or above (OR = 1.30, 95 % CI 0.50–3.43), severe intraventricular hemorrhage of stage III or above (OR = 2. 91, 95 % CI 0.64–13.23), necrotizing enterocolitis (OR = 0.57, 95 % CI 0.13–2.54), and borderline personality disorder (OR = 1. 06, 95 % CI 0.50–2.26). The rhEPO treatment has beneficial effect on the neurodevelopment outcomes without severe adverse side effect in preterm infants. Keywords Recombinant human erythropoietin Preterm Neural system Meta-analysis
Introduction Neonatal brain injury, including commonly hypoxic-ischemic encephalopathy (HIE), periventricular leuko-malacia (PVL), cerebral-intraventricular hemorrhage (IVH), and hyperoxic brain damage, is an important cause of neonatal mortality and subsequent sequelae such as cerebral palsy, mental retardation, learning disability, and epilepsy [1–3]. Brain injury results in poor neurodevelopmental outcomes including cerebral palsy, mental retardation, hearing or visual impairment, and attention deficit hyperactivity disorder [4]. Although there is increasing evidence about underlying mechanisms and growing number of studies about treatment strategies, a safe and an effective treatment for both preterm and term infants regardless of the severity of the brain injury is urgently needed.
123
Cell Biochem Biophys
Erythropoietin (EPO), which was originally identified for its role in erythropoiesis, is widely used for the treatment of anemia in premature infants [5]. Surprisingly, EPO was found to play a variety of roles in modulation of the inflammatory response and has vasogenic effects [6]. Neuroprotection with EPO has been documented in spinal cord injury, traumatic brain injury, ischemic stroke, and perinatal asphyxia [7]. The neurodevelopmental outcomes of 20 premature infants treated with recombinant human erythropoietin (rhEPO) have been evaluated [8]. There was no adverse effect of EPO on neurologic outcome, cognitive outcome, or growth patterns and the rate of cognitive deficits. Later, developmental outcomes at 18–22 months’ corrected age in extreme low body weight (ELBW) infants treated with EPO were compared [9]. No difference between groups with respect to the percentage of infants with blindness, deafness or hearing loss, moderate to severe cerebral palsy was found. However, the limitation of this study was that the EPO treatment group included more IVH infants than that in the control group. The safety of EPO in preterm infants has been established by its high does administration in two trials. In a prospective trial of high-dose rhEPO in ELBW infants, the investigators compared 30 infants who were treated with high-dose rhEPO with 30 concurrent control subjects [10]. Early high-dose rhEPO is well tolerated by ELBW infants, causing no excess mortality. In another randomized, doubleblinded trial, the safety of administration of high-dose EPO (3,000 U/kg, 39) to preterm infants has been investigated [11]. No side effect of EPO such as IVH, ROH, and necrotizing enterocolitis (NEC) has been found [11]. Both of the above studies provide an important insight into the safety of preterm infants who received early high-dose rhEPO. These findings suggest that an early high-dose administration of rhEPO to preterm infants to improve neurodevelopmental outcome is feasible. Research findings from clinical studies regarding to the efficacy and safety of rhEPO for improving neurodevelopment outcomes in preterm infants are still controversial. Therefore, we conduct a meta-analysis aiming to assess the neuroprotective effect of rhEPO in preterm infants.
Materials and Methods Data Sources Studies were identified by extensively searching the PubMed, EMBASE, Cochrane Central Register of Controlled Trials, Chinese Biomedical Literature Database (CBM), Chinese National Knowledge Infrastructure (CNKI), Wan Fang Data, and VIP INFORMATION from January 1999 to December 2011. The search terms were (EPO or erythropoietin) and (preterm or premature or infants or
123
very low birth weight or extremely low birth weight) and (neurodevelopmental or brain or nerve). The language of the literatures was not restricted to English. In addition, references of the relevant articles were reviewed in order to identify additional studies not detected by the initial search. Study Selection Two reviewers (Yanping Mu and Xun Deng) independently searched literatures and examined relevant RCTs. Any disagreement about study selection or data extraction was resolved by consensus with the third reviewer (Yong Wang). For meta-analysis, all studies have to meet the following inclusion criteria: (1) A study described as RCT or quasiRCT. (2) Preterm infants (gestation \37 weeks). (3) Intervention measures: intravenous injection or hypodermic of rhEPO; total dose within the period of treatment [3,000 U kg-1; and administration of rhEPO started within the first 7 days of life and continued for at least 14 days. (4) Control group receiving a placebo or routine treatment. (5) Primary endpoints were MDI scores, PDI scores at 12–22 months or MDI scores \70, PDI scores \70; NBNA scores at 40 weeks of corrected gestational age; and serious nervous system sequel (cerebral palsy, blindness, or hearing loss). Safety indexes were severe retinopathy of prematurity (ROP) of stage III or above, severe IVH of stage III or above, NEC, borderline personality disorder (BPD). Non-RCT or quasi-RCT studies were excluded, as well as case reports, reviews with insufficient details to meet the inclusion criteria, abstracts in the proceedings of scientific conferences, and experimental trials; Premature infants with severe congenital malformation, hemolysis, anemia, polycythemia, serious infection, Intrauterine TORCH infection, congenital heart disease, and genetic metabolic disease were also excluded from the study. Data Extraction Two of the authors independently extracted data from the trials that met the inclusion criteria using. Authors would be contacted for missing data when necessary. For each trial, the following data were extracted: number of patients in each group; mean age and sex distribution of each group; gestational age; birth weight; drug regimen, including doses and treatment duration; and clinical outcome indicators (efficacy and safety of drugs). Assessment of Study Quality Quality assessment of the RCTs included in the meta-analysis was independently performed by the same reviewers according to Cochrane Handbook 5.0.1 and Juni et al. [12, 13], which assesses the descriptions of generation of random,
Cell Biochem Biophys Fig. 1 Studies selection process
allocation concealment, double blinding, and dropouts\withdrawals of the included trials. Each author rated the quality of the trials using Jadad grade (maximum grade = A; minimum grade = C; grade C B = good quality). Statistical Analysis Data were analyzed using Review Manager 5.1. Included articles were pooled and weighted. Odds ratio (OR) or weighted mean difference (WMD) and 95 % confidence interval (95 % CI) were calculated in a fixed-effects mode or in a random-effects model. Heterogeneity was assessed by calculating a v2 test and the quantity of heterogeneity was measured with I2 statistic. If heterogeneity (P \ 0.1 or I2 [ 50 %) was found among the trials, random-effects model would be chosen, otherwise fixed-effects model chosen. If heterogeneity was evident (I2 [ 70 %), the inferior quality study should be eliminated to analyze. A 2-tailed P value of less than 0.05 was deemed statistically significant.
Results Study Selection Process Our search yielded a total of 118 studies on recombinant human erythropoietin for neurodevelopment in preterm
infants. After reviewing each abstract or original publication and extracting data from the publications, five were finally selected [9, 14–17]. Studies including reviews (n = 22), non-RCTs (n = 35), foundational researches (n = 40), anemia studies (n = 6), full-term infants RCTs (n = 2) that do not meet the inclusion criteria of the metaanalysis (n = 8) were excluded. The details of the selection process were shown in Fig. 1. Study Characteristics The main characteristics of five included studies are presented in Table 1. Five studies involving 233 patients were ultimately confirmed that met the criteria for inclusion in the meta-analysis. All of the trials enrolled patients, two of which studied very low birth weight infant (VLBWI), three premature infants with gestational age\37 weeks and birth weight \2,500 g. And two studies adopted MDI and PDI scores, one bayley scale of infant development (BSID), one gessell scale, and three reported NBNA scores to assess clinical effective rates. The quality assessment of included studies is presented in Table 2. Two studies clarified adequate generation of random, one reported allocation concealment, two used double blinding and two reported numbers of dropouts\withdrawals. One was eventually assessed to be good in terms of methodology with a Jadad score A [12], one was B, the other three were C.
123
Cell Biochem Biophys Table 1 Basic data of included 5 literatures (x ± s) Study
Design
n1/n2
GA/weeks
Ohls [9]
RCT
51/51
26.3 ± 2.0/25.8 ± 1.7
Bierer et al. [14]
RCT
6/6
26.0 ± 1.1/26.9 ± 2.1
Wang [15]
quasiRCT
20/15
28–34
Wang et al. [16]
quasiRCT
20/20
31.5 ± 2.1/32.1 ± 2.5
He [17]
quasiRCT
22/22
29.4 ± 1.5/30.2 ± 1.3
Table 2 Quality assessment of included 5 literatures
Dosage/ U kg-1
Duration
Results
801 ± 139/783 ± 112
400
8–10 weeks
% MDI \70, % PDI \70
752 ± 150/810 ± 103
400
Until discharge, transfer death, or 35 completed weeks PMA
MDI, PDI, % MDI \70, % PDI \70
400
6 weeks
NBNA
1626 ± 338/1660 ± 255
250
4 weeks
MDI, PDI, NBNA
1135 ± 120/1075 ± 135
250
4 weeks
NBNA, Gesell
Birth weight/g
\2000
Study
Generation of random
Allocation concealed
Blind of outcome assessors
Loss of follow-up
Baseline similarity
Quality grade
Ohls [9]
Clear
Unknown
Yes
Yes
Statement
Similar
B
Bierer et al. [14]
Clear
Yes
Yes
Yes
Statement
Similar
A
Wang [15] Wang et al. [16]
Incomplete Incomplete
Yes Yes
Unknown Unknown
Unknown Yes
Statement Statement
Similar Similar
C C
He [17]
Incomplete
Yes
Unknown
Yes
Statement
Similar
Comparisons of Effectiveness Effect of rhEPO on MDI Two studies of Bierer et al. [14] and Wang et al. [16] provided specific data for analysis of MDI scores. MDI score in the rhEPO group [26 (50 %) of 52 patients] was higher than control group, while statistically significant difference was found [WMD = 7.77, 95 % CI (3.49–12.06), P = 0.0004] (Fig. 2). Data from two studies of Bierer et al. [14] and Ohls et al. [9] showed the incidence of MDI scores below 70 between the rhEPO group [51 (50 %) of 102 patients] and the control group [51 (50 %) of 102 patients] had no significant difference [OR = 0. 70, 95 % CI (0. 31–1. 61), P = 0.4] (Fig. 3). Effect of rhEPO on PDI Two studies of Bierer et al. [14] and Wang et al. [16] provided specific data for analysis of PDI scores. PDI score in the rhEPO group [26 (50 %) of 52 patients] was higher than control group, while statistically significant difference was found [WMD = 3.85, 95 % CI (0.62–7.09), P = 0.02] (Fig. 4).
123
Blind investigator
Data from two studies of Bierer et al. [14] and Ohls et al. [9] showed the incidence of PDI scores below 70 between the rhEPO group [51 (50 %) of 102 patients] and control group [51 (50 %) of 102 patients]. No statistically significant difference was found [OR = 2.46, 95 % CI (0. 94–6.45), P = 0.07] (Fig. 5). Effect of rhEPO on NBNA Scores Three studies [15–17] provided specific data for analysis of NBNA scores. We noted no significant heterogeneity between the trials (I2 = 0 %, P = 0.63). Meta-analysis of data using a fixed-effects model estimated NBNA scores in the rhEPO group. NBNA score in the rhEPO group [(52.1 %) of 119 patients] was higher than control group [(47.9 %) of 119 patients], while statistically significant difference was found [WMD = 1.96, 95 % CI (1.56–2.37), P \ 0.00001] (Fig. 6). Comparisons of Safety Adverse drug reaction data for severe ROP incidence, severe IVH (stage III or above) incidence, and BPD incidence in rhEPO group and control group were reported in
Cell Biochem Biophys
Fig. 2 Meta-analysis of MDI scores in rhEPO treatment and control groups
Fig. 3 Meta-analysis of MDI scores below 70 in rhEPO treatment and control groups
Fig. 4 Meta-analysis of PDI scores in rhEPO treatment and control groups
Fig. 5 Meta-analysis of PDI scores below 70 in rhEPO treatment and control groups
two studies [9, 14]. In the evaluable patients, the incidence of severe ROP (stage III or above) between the rhEPO group [58 (49.6 %) of the 117 patients] and the control group [59 (50.4 %) of the 117 patients] had no statistically
significant difference [OR = 1.30, 95 % CI (0.50–3.43), P = 0.59] (Fig. 7). The incidence of severe IVH (stage III or above) was 49.1 % (56/114) for rhEPO and 50.9 % (58/ 114) for comparator, while no statistically significant
123
Cell Biochem Biophys
Fig. 6 Meta-analysis of NBNA scores in rhEPO treatment and control groups
Fig. 7 Meta-analysis of the incidence of severe ROP (stage III or above) in rhEPO treatment and control groups
Fig. 8 Meta-analysis of the incidence of severe IVH (stage III or above) in rhEPO treatment and control groups
difference was found [OR = 2.91, 95 % CI (0.64–13.23), P = 0.17] (Fig. 8). And no statistically significant difference was found [OR = 1.06, 95 % CI (0.50–2.26), P = 0.87] (Fig. 9) between the rhEPO group [49.6 % (58/ 117)] and the control group [50.4 % (59/117)] about the incidence of BPD.
Discussion We conducted this meta-analysis to assess the efficacy and safety of rhEPO for improving neurodevelopment
123
outcomes in preterm infants. The MDI scores, PDI scores, and NBNA scores of rhEPO were higher than control group, and superior to control with statistically significant difference [MDI scores: WMD = 7.77, 95 % CI (3.49–12.06), P = 0.0004; PDI scores: WMD = 3.85, 95 % CI (0.62–7.09), P = 0.02; WMD = 1.96, 95 % CI (1.56–2.37), P \ 0.00001]. However, the incidence of PDI scores below 70 and the incidence of MDI scores below 70 had no significant difference between rhEPO and control group [PDI scores below 70 incidence: OR = 0. 70, 95 % CI (0.31–1.61), P = 0.4; MDI scores below 70 incidence: OR = 2.46, 95 % CI (0. 94, 6.45), P = 0.07]. The safety
Cell Biochem Biophys
Fig. 9 Meta-analysis of the incidence of BPD in rhEPO treatment and control groups
profile analysis regarding the incidence of adverse drug reactions had no significant difference between rhEPO and control group [severe ROP (stage III or above) incidence: OR = 1.30, 95 % CI (0.50–3.43; severe IVH (stage III or above): OR = 2.91, 95 % CI (0.64–13.23), P = 0.17; BPD incidence: OR = 1.06, 95 % CI (0.50–2.26), P = 0.87]. This meta-analysis indicated that rhEPO could be used to improve neurodevelopment outcomes in preterm infants. Recently, a long-term retrospective study firstly has confirmed the neuroprotective benefits of rEPO for ELBW infants at school age. ELBW infants receiving rEPO scored significantly better than untreated children in the overall developmental assessment (55 versus 39 % normally developed) as well as in the psychological examination (mean composite HAWIK-III IQ score, 90.8 versus 81.3). Also, this study confirms the neuroprotective effect of rEPO in ELBW infants with IVH. Furthermore, rEPO treatment works well even when given days after the onset of brain damage suggests a wider window of therapy. However, this study has a few limitations such as not being a randomized control trial and not a strict subgroup analysis for IVH mild or severe degrees [18]. Therefore, rhEPO may be recommended as a routine therapy for improving neurodevelopment outcomes in preterm infants in future, but more RCTs are needed to further study. However, some possible limitations in this meta-analysis should be acknowledged. First, our findings may be affected by the quality of trials included in the meta-analysis. Only one reported allocation concealment, two of the included trials were double blinding, and two with clarified adequate generation of random. Only one was assessed to be good in terms of methodology with a Jadad score A, three were C. Second, this meta-analysis is based on a relatively small number of RCTs and we acknowledge that using a limited number of studies raises the possibility of a second-order sampling error [19]. However, meta-analyses often include small numbers of studies; Higgins [19] evaluated 39 Cochrane reviews and found that 67 % of them included B5 studies and 20 % included B10 studies.
Therefore, more RCTs comparing rhEPO with control group for protection of neurodevelopment in preterm infants are needed to be carried out in future. Third, the eligibility criteria for inclusion of preterm infants were different from each other (Table 1), which might influence the obvious consistency of effects across those included studies and cause the between-study heterogeneity. Thus, to ensure uniformity in defining preterm infants’ characteristics and outcome measures, an individual patient data meta-analysis needs to be performed. Finally, the distinct differences in dose, timing, and pharmaceutical manufacturers of rhEPO used exist. In conclusion, although some limitations exist in this meta-analysis, based on the result of our meta-analysis, it is suggested that rhEPO is an effective and a relatively safe option for improving neurodevelopment outcomes in preterm infants. However, the long-term effects of rhEPO for neurodevelopment outcomes in preterm infants still needs to be further investigated.
References 1. Gonzalez, F. F., & Ferriero, D. M. (2009). Neuroprotection in the newborn infant. Clinics in Perinatology, 36, 859–880. 2. Volpe, J. J. (2003). Cerebral white matter injury of the premature infant-more common than you think. Pediatrics, 112, 176–180. 3. Felderhoff-Mueser, U., Bittigau, P., Sifringer, M., Jarosz, B., Korobowicz, E., Mahler, L., et al. (2004). Oxygen causes cell death in the developing brain. Neurobiology of Diseases, 17, 273–282. 4. McPherson, R. J., & Juul, S. E. (2008). Recent trends in erythropoietin-mediated neuroprotection. International Journal of Developmental Neuroscience, 26, 103–111. 5. Ghezzi, P., Bernaudinb, M., Bianchic, R., Blomgren, K., Brines, M., Campana, W., et al. (2010). Erythropoietin: Not just about erythropoiesis. Lancet, 375, 2142. 6. Gonzalez, F. F., & Ferriero, D. M. (2008). Therapeutics for neonatal brain injury. Pharmacology and Therapeutics, 120, 43–53. 7. McPherson, R. J. (2009). Can erythropoietin improve developmental outcomes for preterm infants? Pediatrics, 124, e805–e806.
123
Cell Biochem Biophys 8. Newton, N. R., Leonard, C. H., Piecuch, R. E., & Phibbs, R. H. (1999). Neurodevelopmental outcome of prematurely born children treated with recombinant human erythropoietin in infancy. Journal of Perinatology, 19, 403–406. 9. Ohls, R. K., Ehrenkranz, R. A., Das, A., Dusick, A. M., Yolton, K., Romano, E., et al. (2004). Neurodevelopmental outcome and growth at 18 to 22 months’ corrected age in extremely low birth weight infants treated with early erythropoietin and iron. Pediatrics, 114, 1287–1291. 10. Juul, S. E., McPherson, R. J., Bauer, L. A., Ledbetter, K. J., Gleason, C. A., & Mayock, D. E. (2008). A phase I/II trial of high-dose erythropoietin in extremely low birth weight infants: Pharmacokinetics and safety. Pediatrics, 122, 383–391. 11. Fauchere, J. C., Dame, C., Vonthein, R., Koller, B., Arri, S., Wolf, M., et al. (2008). An approach to using recombinant erythropoietin for neuroprotection in very preterm infants. Pediatrics, 122, 375–382. 12. Juni, P., Altman, D. G., & Egger, M. (2001). Systematic reviews in health care: Assessing the quality of controlled clinical trials. British Medical Journal, 323, 42–46. 13. Jadad, A. R., Moore, R. A., Carroll, D., Jenkinson, C., Reynolds, D. J., Gavaghan, D. J., et al. (1996). Assessing the quality of reports of randomized clinical trials: Is blinding necessary? Controlled Clinical Trials, 17, 1–12.
123
14. Bierer, R., Peceny, M. C., Hartenberger, C. H., & Ohls, R. K. (2006). Erythropoietin concentrations and neurodevelopmental outcome in preterm infants. Pediatrics, 118(3), 635–640. 15. Wang, Y. H. (2005). Effects of erythropoietin on anemia and nervous system of premature infants. Journal of Medical Forum, 26(7), 49–52. 16. Wang, Y. H., Zhu, C. L., Cheng, X. Y., Hu, S. H., Feng, H. Q., & Wang, X. Y. (2006). Effects of recombinant human erythropoietin on neurobehavioral development in premature infants. Chinese Journal of Practical Pediatrics, 12(1), 59–60. 17. He, J. S., Huang, Z. L., Yang, H., Weng, K. Z., & Zhu, S. B. (2008). Early use of recombinant human erythropoietin promotes neurobehavioral development in preterm infants. Chinese Journal of Contemporary Pediatrics, 10(5), 586–588. 18. Neubauer, A. P., Voss, W., Wachtendorf, M., & Jungmann, T. (2010). Erythropoietin improves neurodevelopmental outcome of extremely preterm infants. Annals of Neurology, 67, 657–666. 19. Higgins, J., Thompson, S., Deeks, J., & Altman, D. (2002). Statistical heterogeneity in systematic reviews of clinical trials: A critical appraisal of guidelines and practice. Journal of Health Services Research and Policy, 7(1), 51–61.
