Scolaris Content Display Scolaris Content Display

Intravenous immunoglobulin for preventing infection in preterm and/or low birth weight infants

Collapse all Expand all

Abstract

Background

Nosocomial infections continue to be a significant cause of morbidity and mortality among preterm and/or low birth weight (LBW) infants. Preterm infants are deficient in immunoglobulin G (IgG); therefore, administration of intravenous immunoglobulin (IVIG) may have the potential of preventing or altering the course of nosocomial infections.

Objectives

To use systematic review/meta‐analytical techniques to determine whether IVIG administration (compared with placebo or no intervention) to preterm (< 37 weeks' postmenstrual age (PMA) at birth) or LBW (< 2500 g birth weight) infants or both is effective/safe in preventing nosocomial infection.

Search methods

For this update, MEDLINE, EMBASE, CINAHL, The Cochrane Library, Controlled Trials, ClinicalTrials.gov and PAS Abstracts2view were searched in May 2013.

Selection criteria

We selected randomised controlled trials (RCTs) in which a group of participants to whom IVIG was given was compared with a control group that received a placebo or no intervention for preterm (< 37 weeks' gestational age) and/or LBW (< 2500 g) infants. Studies that were primarily designed to assess the effect of IVIG on humoral immune markers were excluded, as were studies in which the follow‐up period was one week or less.

Data collection and analysis

Data collection and analysis was performed in accordance with the methods of the Cochrane Neonatal Review Group.

Main results

Nineteen studies enrolling approximately 5000 preterm and/or LBW infants met inclusion criteria. No new trials were identified in May 2013.

When all studies were combined, a significant reduction in sepsis was noted (typical risk ratio (RR) 0.85, 95% confidence interval (CI) 0.74 to 0.98; typical risk difference (RD) ‐0.03, 95% CI 0.00 to ‐0.05; number needed to treat for an additional beneficial outcome (NNTB) 33, 95% CI 20 to infinity), and moderate between‐study heterogeneity was reported (I2 54% for RR, 55% for RD). A significant reduction of one or more episodes was found for any serious infection when all studies were combined (typical RR 0.82, 95% CI 0.74 to 0.92; typical RD ‐0.04, 95% CI ‐0.02 to ‐0.06; NNTB 25, 95% CI 17 to 50), and moderate between‐study heterogeneity was observed (I2 50% for RR, 62% for RD). No statistically significant differences in mortality from all causes were noted (typical RR 0.89, 95% CI 0.75 to 1.05; typical RD ‐0.01, 95% CI ‐0.03 to 0.01), and no heterogeneity for RR (I2 = 21%) or low heterogeneity for RD was documented (I2 = 28%). No statistically significant difference was seen in mortality from infection; in incidence of necrotizing enterocolitis (NEC), bronchopulmonary dysplasia (BPD) or intraventricular haemorrhage (IVH) or in length of hospital stay. No major adverse effects of IVIG were reported in any of these studies.

Authors' conclusions

IVIG administration results in a 3% reduction in sepsis and a 4% reduction in one or more episodes of any serious infection but is not associated with reductions in other clinically important outcomes, including mortality. Prophylactic use of IVIG is not associated with any short‐term serious side effects.

The decision to use prophylactic IVIG will depend on the costs and the values assigned to the clinical outcomes. There is no justification for conducting additional RCTs to test the efficacy of previously studied IVIG preparations in reducing nosocomial infections in preterm and/or LBW infants.

PICOs

Population
Intervention
Comparison
Outcome

The PICO model is widely used and taught in evidence-based health care as a strategy for formulating questions and search strategies and for characterizing clinical studies or meta-analyses. PICO stands for four different potential components of a clinical question: Patient, Population or Problem; Intervention; Comparison; Outcome.

See more on using PICO in the Cochrane Handbook.

Intravenous immunoglobulin for preventing infection in preterm and/or low birth weight infants

Infants may acquire infections while in the womb or in the hospital after birth, especially if they require intensive care. Such infections may cause serious illness or death. Transport of immunoglobulin (substance in the blood that can fight infection) from the mother to the fetus mainly occurs after 32 weeks' gestation, and infants do not begin to produce immunoglobulin until several months after birth. Theoretically, the adverse effects of infection could be reduced by the preventive administration of intravenous immunoglobulin. To date, approximately 5000 infants have been enrolled in studies conducted to evaluate the effects of prophylactic use of intravenous immunoglobulin on neonatal outcomes. Intravenous administration of immunoglobulin results in a 3% reduction in blood‐borne infection and a 4% reduction in serious infection. Intravenous administration of immunoglobulin is not associated with reductions in other important neonatal outcomes or in length of hospital stay. Most important, intravenous immunoglobulin administration does not have any important effect on mortality. Prophylactic use of IVIG is not associated with any short‐term serious side effects. From a clinical perspective, a 3% to 4% reduction in nosocomial infection without a reduction in mortality or other important clinical outcomes is of marginal importance.

Authors' conclusions

Implications for practice

IVIG administration results in a 3% to 4% reduction in sepsis/any serious infection but is not associated with reductions in mortality or other morbidities (NEC, IVH, length of hospital stay). Prophylactic use of IVIG is not associated with any short‐term serious side effects. The decision to use prophylactic IVIG will depend on the costs and the values assigned to the clinical outcomes.

Implications for research

A full economic evaluation and a clinical decision analysis that incorporates baseline risk for confirmed nosocomial infection, clinical outcomes and economic outcomes after prophylactic IVIG, as well as values attached to infections prevented, is needed.

There is no justification for conducting additional RCTs to test the efficacy of previously studied IVIG preparations in reducing nosocomial infection in preterm and/or LBW infants. It is possible that the IVIG preparations used in published studies did not contain the necessary antibodies to prevent infection. The use of preparations with known specific antibodies against the common pathogens in a specific neonatal intensive care unit might be more effective, and RCTs to test the effectiveness of such preparations may be justified. The results of these meta‐analyses should encourage basic scientists and clinicians to pursue other avenues to prevent nosocomial infection.

Background

Description of the condition

Although survival has improved for preterm and/or low birth weight (LBW) infants, nosocomial infection continues to be a significant cause of morbidity and mortality in this population. A 25% incidence of late‐onset infection has been reported in a cohort of 6911 very LBW infants who were admitted to 12 US centres and who survived beyond three days (Stoll 1996). Neonates in whom late‐onset sepsis developed were significantly more likely to die than those who were not infected (17% vs 7%; P < 0.0001) (Stoll 1996).

Description of the intervention

Intravenous immunoglobulin (IVIG) contains pooled immunoglobulin G (IgG) extracted from the plasma of more than 1000 blood donors. 

IVIG is frequently given to immunodeficient patients who have decreased antibody production capabilities. In immunodeficient patients, IVIG is administered to maintain adequate antibody levels to prevent infection and to confer passive immunity.

How the intervention might work

Maternal transport of immunoglobulins to the fetus mainly occurs after 32 weeks' gestation, and endogenous synthesis does not begin until about 24 weeks after birth, so the preterm infant is especially vulnerable to infectious sources in the neonatal intensive care unit (Baker 1990a). Mean serum levels of IgG are 400 mg/dL in infants at less than 32 weeks' gestational age (GA) compared with 1000 mg/dL in term infants (Hobbs 1967; Stiehm 1966). The idea of preventing nosocomial infection with IVIG is attractive, as administration of IVIG provides IgG that can bind to cell surface receptors, provide opsonic activity, activate complement, promote antibody‐dependent cytotoxicity and improve neutrophilic chemo‐luminescence (Baley 1988).

Why it is important to do this review

Administration of IVIG to LBW infants has been studied extensively. Numerous descriptive review articles, commentaries and editorials on the use of IVIG in neonates have been published, often by the same researchers. These papers have included several randomised controlled trials (RCTs), the authors' personal experience with IVIG and/or information about the preparation or dosing regimen of IVIG (Weisman 1986; Bortolussi 1986a; Bortolussi 1986b; Fischer 1986; Stiehm 1986; Baley 1988; Fischer 1988; Gonzalez 1989; Kyllonen 1989; Stabile 1989; Noya 1989; Johnston 1990; Fischer 1990a; Fischer 1990b; Fischer 1990c; Baker 1990a; Baker 1990b; Bussel 1990b; Hammarstrom 1990; Kliegman 1990; Stiehm 1990; Whitelaw 1990; Berger 1991; Hill 1991a; Hill 1991b; Irani 1991; Kliegman 1991; Magny 1991a; Rondini 1991; Haque 1992; Siber 1992; Weisman 1992; Hill 1993; Weisman 1993; Weisman 1994b; Wolach 1997). Salzer (Salzer 1991) presented (in abstract form only) the results of a meta‐analysis of seven studies and concluded that no significant reduction was seen in the incidence of sepsis in the treated group. In "Effective Care of the Newborn Infant", Baley and Fanaroff (Baley 1992) present overviews of RCTs that studied the administration of IVIG to neonates. They reviewed seven studies of the prophylactic use of IVIG that reported an outcome of sepsis and concluded, "The preliminary data generated in trials of IVIG are promising, but use of this treatment modality still needs to be considered experimental and [it] should only, as yet, be used under study conditions." Lacy and Ohlsson (Lacy 1995) included additional trials and concluded that routine administration of IVIG to preterm infants to prevent infection is not recommended. Jenson and Pollock (Jenson 1997) used slightly different inclusion criteria and, like Lacy and Ohlsson (Lacy 1995), noted heterogeneity among studies. They concluded, "this heterogeneity probably belies the minimal benefit, at most, of prophylactic IVIG". The results of a Canadian multidisciplinary consensus‐building initiative (Consensus 1997) have been published, and the use of IVIG for prophylaxis of neonatal nosocomial infection was considered inappropriate. This review provides an update of our previous review, which was last updated with no changes in 2010 (Ohlsson 2001; Ohlsson 2007) and was first published in 1998 (Ohlsson 1998).

Objectives

To use systematic review/meta‐analytical techniques to determine whether IVIG administration (compared with placebo or no intervention) to preterm (< 37 weeks' gestational age (GA) at birth)or LBW (< 2500 g birth weight) infants or both is effective/safe in preventing nosocomial infection.

Methods

Criteria for considering studies for this review

Types of studies

Studies in which preterm and/or LBW neonates were randomly assigned to receive IVIG or placebo or no intervention.

Types of participants

Preterm and/or LBW neonates.

Types of interventions

IVIG for the prevention of bacterial or fungal infection. Studies that were designed to evaluate the effects of IVIG on humoral immune markers were excluded, as were studies in which the follow‐up period was one week or less. Studies that assessed the effectiveness of IVIG for treatment of suspected or confirmed infection were excluded.

Types of outcome measures

Primary outcomes

  • Sepsis, one or more episodes (clinical signs and symptoms of sepsis and positive blood culture for bacteria or fungi).

Secondary outcomes

  • Any serious infection (clinical signs and symptoms in conjunction with positive cultures (bacteria or fungi) from normally sterile body fluids (blood, cerebrospinal fluid, urine obtained by catheterization or suprapubic tap) or from tissue at autopsy). As per this definition, cases of sepsis if reported separately were also included in any serious infection.

  • Necrotizing enterocolitis (NEC) diagnosed according to Bell's criteria (Bell 1978). For repeated episodes of sepsis, any serious infection and NEC, only one occurrence per infant was counted as an outcome.

  • Death from all causes.

  • Death from infection (including death from NEC).

  • Length of hospital stay.

  • Incidence of bronchopulmonary dysplasia (BPD), defined as an additional oxygen requirement (above room air) at 28 days of age or a requirement for assisted ventilation for reasons other than apnoea of prematurity.

  • Incidence of intraventricular haemorrhage (IVH), any grade, classified according to Papile (Papile 1983).

  • Incidence of IVH, grade 3 or 4, classified according to Papile (Papile 1983).

  • Reports on possible side effects as described by the authors.

Search methods for identification of studies

The search strategy used to identify studies adhered to the guidelines of the Cochrane Neonatal Review Group.

Electronic searches

MEDLINE was searched from 1966 to July 2007. EMBASE (Excerpta Medica online) was searched from 1980 to July 2007. The Cochrane Library, Issue 2, 2007, was searched. No language restrictions were applied. Ms Elizabeth Uleryk developed and applied an extensive search strategy (available upon request) for MEDLINE and EMBASE in February 2001 and September 2003. The same strategy was used in 2007.

In December 2009, we updated the search as follows: MEDLINE (search via PubMed), CINAHL, EMBASE and the Cochrane Central Register of Controlled Trials (CENTRAL; The Cochrane Library) were searched from 2003 to December 2009. Search term: immunoglobulin. Limits: human, newborn infant and clinical trial. No language restrictions were applied.

Searches on May 28, 2013, were conducted by Ms Coleen M. Ovelman, Trials Search Co‐ordinator, and Yolanda R. Brosseau, Managing Editor, the Cochrane Neonatal Review Group. Abstracts of the Pediatric Academic Societies (PAS) Annual Meetings from 2000 to 2013 were searched on the same day by one of the review authors (AO).

Searching other resources

The search was initiated by review of personal files and published meta‐analyses. The reference list of identified studies and subsequently retrieved articles was scanned for additional references.

Data collection and analysis

Data collection and analysis were done in accordance with the methods of the Cochrane Neonatal Review Group.

Selection of studies

The criteria used to select studies for inclusion in this overview were:

  • design: RCT in which treatment with IVIG was compared with a placebo or no intervention provided to a control group;

  • population: inclusion of preterm (< 37 weeks' gestational age) and/or LBW (< 2500 g) infants;

  • intervention: administration of IVIG for the prevention of bacterial/fungal infection during initial hospital stay (8 days or longer). (Studies that were primarily designed to assess the effects of IVIG on humoral immune markers were excluded, as were studies in which the follow‐up period was one week or less. Studies designed to assess the effectiveness of treatment with IVIG for suspected/established infection were excluded.); and

  • reporting: included at least one of the following outcomes: sepsis, any serious infection, death from all causes, death from infection, length of hospital stay, IVH, NEC or BPD;and descriptions of side effects.

The titles (and abstracts when available) in MEDLINE, EMBASE and The Cochrane Library printouts were reviewed by the two review authors. Any article that the review authors believed might meet the inclusion criteria noted above or should have its reference list searched was retrieved. Informal attempts were made to locate unpublished studies, and attempts were made to request additional information from authors of published studies. Additional information was obtained on one published study (Sandberg 2000).

All identified trials (excluding those that used IVIG for treatment) are listed in the tables of Included studies and Excluded studies.

Data extraction and management

Data abstraction forms were developed and were pilot‐tested to verify definitions of terms. The two review authors independently abstracted information on each study, and one review author (AO) checked for any discrepancies and pooled the results. Data abstraction included whether the study involved prophylaxis or treatment, the number of participants enrolled, the number of participants enrolled but later excluded, the time period and geographical location of the study, baseline characteristics of participants, inclusion/exclusion criteria, the preparation and dosing regimen of IVIG and placebo and length of follow‐up.

Information on outcomes and on the numbers of affected infants was abstracted. The total number of infants with sepsis (clinical signs and symptoms plus positive blood culture (bacteria or fungi)) and any serious infection (clinical signs and symptoms in conjunction with positive cultures (bacteria or fungi) from normally sterile body fluids) was abstracted, as was information on NEC, death from all causes and death from infection. Information on length of hospital stay and on incidence of BPD and IVH was collected. Information on probable infection was not collected, as the definitions used by different investigators were too variable.

Assessment of risk of bias in included studies

Assessment of the quality of included studies (excluding abstracts) was performed independently by JBL and AO, using criteria developed by the Cochrane Neonatal Review Group. These criteria included blinding of randomisation, blinding of the intervention, complete follow‐up and blinding of outcome measurement. For each criterion, three possibilities were identified: yes, can't tell and no. The assignment was not done with the assessors blinded to author, institution, journal of publication or results, as both assessors were familiar with most of the studies and with the typographical layout of the journals and would have knowledge of these even when blinded. In addition, the results sections of articles often include methodological information. After independent evaluation was performed, the two assessors discussed the results of each study, and any discrepancies were resolved. 

For the update in 2009, the following issues were evaluated and entered into the risk of bias table:

  • Sequence generation: Was the allocation sequence adequately generated?; 

  • Allocation concealment: Was allocation adequately concealed?; 

  • Blinding of participants, personnel and outcome assessors: Was knowledge of the allocated intervention adequately prevented during the study? At study entry? At the time of outcome assessment?; 

  • Incomplete outcome data: Were incomplete outcome data adequately addressed?;

  • Selective outcome reporting: Are reports of the study free of suggestion of selective outcome reporting?; and 

  • Other sources of bias: Was the study apparently free of other problems that could put it at high risk of bias?

Measures of treatment effect

The statistical package (RevMan 5.2) provided by the Cochrane Collaboration was used. Typical relative risk (RR) and typical risk difference (RD) with 95% confidence intervals (CIs) using the fixed‐effect model are reported. If a statistically significant reduction in RD was noted, the number needed to treat to benefit (NNTB) or harm (NNTH) was calculated.

Assessment of heterogeneity

Statistically significant between‐study heterogeneity was reported when identified, and the test for inconsistency (I2 statistic) was applied when statistically significant heterogeneity was noted (Higgins 2003). For this update, the following cut‐offs were used:

  • <25%: no heterogeneity;

  • 25% to 49%: low heterogeneity;

  • 50% to 74%: moderate heterogeneity; and

  • > 75%: high heterogeneity.

Data synthesis

Meta‐analysis was performed using Review Manager software (RevMan 5.2) as supplied by The Cochrane Collaboration. For estimates of typical RR and RD, we used the Mantel‐Haenszel method. For measured quantities, we used the inverse variance method. All meta‐analyses were done using the fixed‐effect model.

Subgroup analysis and investigation of heterogeneity

Future updates will consider post hoc subgroup analyses to explore any heterogeneity noted in the primary analysis.

Results

Description of studies

Included Studies
Details of the included studies are provided in the Table of Included Studies.

Nineteen studies, including approximately 5000 preterm and/or LBW infants, met inclusion criteria. These studies were performed in many countries (US, Italy, UK, Saudi Arabia, France, Thailand, Belgium, Turkey, Sweden and Austria). The amount of IVIG per dose varied from 120 mg/kg (Haque 1986) to 1 g/kg (Bussel 1990a). The number of doses varied from a single dose (Atici 1996, Haque 1986, Christensen 1989, Ratrisawadi 1991, Weisman 1994a) to seven doses (Stabile 1988).

Different IVIG preparations were used: Gammagard (Baker 1992); Sandoglobulin (Atici 1996, Bussel 1990a, Chirico 1987, Clapp 1989, Fanaroff 1994, Tanzer 1997, Van Overmeire 1993, Weisman 1994a); Gamimmune (Christensen 1989); Intraglobin (Conway 1990, Haque 1986, Ratrisawadi 1991); IgVena (Didato 1988); Biotransfusion (Magny 1991b); unnamed product (Spady 1994; Sandberg 2000 (study supported by Baxter AG, Austria)); Venogamma (Stabile 1988); Gammumine‐N (Chou 1998).

Excluded Studies
Six studies were excluded, as they included infants who were heavier or more mature at birth than was permitted by the inclusion criteria (Kinney 1991, Adhikari 1996); lacked information on outcomes (Kacet 1991; Malik 1990); lacked a randomised control group (Acunas 1994) or provided immunoglobulin intramuscularly (Monintja 1989).

Risk of bias in included studies

The assessment of individual studies is presented in the Characteristics of included studies table.

The methodological quality of the studies varied. Five studies were of high quality (Baker 1992, Christensen 1989, Clapp 1989, Fanaroff‐I 1994, Weisman 1994a) (i.e. complete follow‐up, blinding of randomisation, intervention and outcome measurement could be ascertained from the published reports). In the remaining 15 studies, elements of bias could not be excluded. The lack of a placebo in 10 studies (Atici 1996, Chirico 1987, Conway 1990, Didato 1988, Fanaroff 1994 (phase II), Haque 1986, Ratrisawadi 1991, Stabile 1988, Tanzer 1997, Van Overmeire 1993) precluded blinding of the caregivers. One study (Fanaroff 1994) included two phases, with phase I providing a placebo but not phase II. In several studies, blinding of randomisation was not clearly described (Chirico 1987, Magny 1991b, Ratrisawadi 1991, Stabile 1988). In the study by Sandberg (Sandberg 2000), an intention‐to‐treat analysis was not applied. One study (Spady 1994) has been published in abstract form only, and the quality therefore could not be fully assessed. The study by Bussel (Bussel 1990a) represents an interim analysis, with data lacking from a large proportion of the infants randomly assigned.

Effects of interventions

Nineteen studies met inclusion criteria. These included a total of approximately 5000 preterm and/or LBW infants and reported on at least one of the outcomes of interest for this systematic review. No new trial was identified in the literature search conducted in May 2013. One additional trial was identified in July 2007 (Lelik 2004). However, this trial enrolled infants at greater than 38 weeks' gestational age and with birth weight greater than 2500 g. The study was therefore excluded. No new studies were identified in the literature search conducted in September 2003.

For details of results, see Data and analyses. It should be noted that for most outcomes, the large study by Fanaroff (Fanaroff 1994) greatly influenced the summary statistics, with an assigned weight ranging from 42.7% for the outcome of any serious infection to 88.3% for the outcome of IVH grade 3 or 4.

IVIG VERSUS PLACEBO OR NO TREATMENT (COMPARISON 1)

PRIMARY OUTCOME

Sepsis, one or more episodes (Outcome 1.1) (Figure 1):

Ten studies (including 3975 infants) reported on the outcome of one or more episodes of sepsis per infant (clinical signs and symptoms of infection and positive blood culture). Only the study by Ratrisawadi (Ratrisawadi 1991) showed a statistically significant reduction in sepsis (RR 0.38; 95% CI 0.19 to 0.79). When all studies were combined, a statistically significant (P = 0.02) reduction in sepsis was noted (typical RR 0.85, 95% CI 0.74 to 0.98); typical RD ‐0.03, 95% CI ‐0.05 to 0.00; NNT 33, 95% CI 20 to infinity). Significant between‐study heterogeneity was observed for this outcome for both RR and RD (P = 0.02; I2 = 54%; moderate for RR, 55% moderate for RD).

Any serious infection, one or more episodes (Outcome 1.2) (Figure 2):

Sixteen studies (including 4986 infants) reported on one or more episodes of any serious infection (sepsis, meningitis, urinary tract infection). Four studies (Atici 1996, Baker 1992, Haque 1986, Ratrisawadi 1991) showed a statistically significant reduction in any serious infection. A statistically significant reduction was found when all studies were combined (typical RR 0.82, 95% CI 0.74 to 0.92; typical RD ‐0.04, 95% CI ‐0.06 to ‐0.02; NNTB 25, 95% CI 17 to 50). Statistically significant between‐study heterogeneity was observed for this outcome (P = 0.01 and I2 = 50% (moderate) for RR; P = 0.0006 and I2 = 62% (moderate) for RD).


Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.1 Sepsis, one or more episodes.

Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.1 Sepsis, one or more episodes.


Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.2 Any serious infection, one or more episodes.

Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.2 Any serious infection, one or more episodes.

Necrotizing enterocolitis (NEC), one or more episodes (Outcome 1.3):

Seven studies (including 4081 infants) reported on NEC (Bell's stage 2 or 3). One study (Fanaroff 1994) showed a borderline statistically significant increase in NEC (RR 1.26, 95% CI 1.00 to 1.59). When all studies were combined, no significant increase was evident (typical RR 1.08, 95% CI 0.89 to 1.32; typical RD 0.01, 95% CI ‐0.01 to 0.02). No statistically significant between‐study heterogeneity was observed for this outcome for RR (P = 0.14; I2 = 38% (low)), but it was observed for RD (P = 0.05, I2 = 52% (moderate)).

Mortality (all causes) (Outcome 1.4) (Figure 3):

Fifteen studies (including 4125 infants) reported on mortality from all causes. Two studies (Chirico 1987, Tanzer 1997) showed a statistically significant reduction in this outcome. When all studies were combined, no statistically significant reduction was noted (typical RR 0.89, 95% CI 0.75 to 1.05; typical RD ‐0.01, 95% CI ‐0.03 to 0.01). No statistically significant between‐study heterogeneity was observed for this outcome for RR (P = 0.22; I2 = 21% (none)) and for RD (P = 0.15; I2 = 28% (low)).


Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.4 Mortality (all causes).

Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.4 Mortality (all causes).

Mortality (infectious) (Outcome 1.5):

Ten studies (including 1690 infants) reported on mortality from infection. One study (Atici 1996) showed a statistically significant reduction in this outcome. The overall analysis showed no significant impact of IVIG prophylaxis on this outcome (typical RR 0.83, 95% CI 0.56 to 1.22; typical RD ‐0.01, 95% CI ‐0.03 to 0.01). No statistically significant between‐study heterogeneity was observed for this outcome for RR (P = 0.11; I2 = 40% (low)) and for RD (P = 0.08; I2 = 42% (low)).

Duration of hospitalisation (Outcome 1.6):

None of eight studies (including 3562 infants) reported a significant reduction in length of hospital stay after IVIG prophylaxis. The overall typical weighted mean difference was ‐2.1 days (95% CI ‐4.5 to 0.3). No statistically significant between‐study heterogeneity was observed (P = 0.67 and I2 = 0% (none) for both RR and RD).

Bronchopulmonary dysplasia (BPD) (Outcome 1.7):

In only one study, the outcome of BPD was defined and data were provided. Several authors failed to define the outcome of BPD, and others defined the outcome but did not provide data. In a small study (n = 115), Clapp (Clapp 1989) showed a trend towards increased BPD (RR 1.53, 95% CI 0.78 to 3.01; RD 0.10, 95% CI ‐0.06 to 0.25). In another small study (n = 61), Chou (Chou 1998) found similar results (RR 1.61, 95% CI 0.42 to 6.16; RD 0.06, 95%CI ‐0.11 to 0.23). When combined (n = 176), the typical RR was 1.55 (95% CI 0.85 to 2.84), and the typical RD was 0.09 (95% CI ‐0.03 to 0.20). No between‐study heterogeneity was observed for this outcome for RR (P = 0.95; I2 = 0% (none)) and for RD (P = 0.74; I2 = 0% (none)).

Intraventricular haemorrhage (IVH) any grade (Outcome 1.8):

Four studies (including 3176 infants) reported on IVH (any grade). Prophylactic IVIG did not have a statistically significant effect on this outcome (typical RR 1.02, 95% CI 0.88 to 1.19; typical RD 0.00, 95% CI ‐0.02 to 0.03). No statistically significant between‐study heterogeneity was observed for this outcome (RR, P = 0.39; I2 = 0.9% (none); RD, P = 0.39; I2 = 0.6% (none)).

Intraventricular haemorrhage (IVH) grade 3 or 4 (Outcome 1.9):

Two studies (including 3000 infants) reported on IVH grade 3 or 4. The typical RR was 1.01 (95% CI 0.85 to 1.21), and the typical RD was 0.00 (95% CI ‐0.02 to 0.03). Statistically significant between‐study heterogeneity was observed (RR, P = 0.09, I2 = 65% (moderate); RD, P = 0.08, I2 = 68% (moderate)).

A rise in serum IgG in the treatment group was noted in all studies that measured serum levels of IgG.

No major adverse effects of IVIG were reported in any of the studies.

Results from excluded studies (see Characteristics of excluded studies) were similar to those from included studies.

Discussion

The effectiveness of IVIG in preventing nosocomial infection in neonates has been well studied. To date, more than 5000 preterm and/or LBW neonates have been enrolled in trials from many different areas of the world. No new trial was identified for this update conducted in May 2013. One additional trial was identified for the update of the review conducted in July 2007 (Lelik 2004). However, the study included infants at > 38 weeks' gestation and > 2500 g birth weight. Therefore, the study was excluded.

The methodological quality of the included trials varied. Five studies were of high quality, but elements of bias could not be excluded in the other studies, mainly because of the fact that the intervention and the assessment of outcomes were performed unblinded to group assignment, or there was lack of complete follow‐up of all randomly assigned infants. IVIG caused increased levels of IgG in serum. No major side effects were noted.

A small but statistically significant reduction in the incidence of sepsis and of any serious infection was found. Statistically significant between‐study heterogeneity was observed for these outcomes. The heterogeneity might be explained in part by variable rates of sepsis and any serious infection in the control groups; differences in preparation, dose and/or dose schedule for IVIG; differences in causative organisms for nosocomial infection; differences in attention to other preventive measures for nosocomial infection and differences in other co‐interventions by place and over time. Some asymmetry was noted when funnel plots were performed for sepsis and any serious infection. For the two main outcomes-sepsis (one or more episodes) and any serious infection (one or more episodes)-moderate inconsistency between study results was noted (I2 54% and 50%, respectively).

No statistically significant differences were noted in mortality from all causes, mortality from infection, NEC, BPD or IVH. Results for these outcomes were centred around an RR of 1.0, with very narrow CIs indicating no trends in either direction. In none of the studies that provided data on IVH was there any assurance that all neonates were subjected to ascertainment of an IVH according to a preset schedule for ultrasonographic examination. A trend towards shortened duration of hospital stay was noted with IVIG treatment [weighted mean difference (WMD) ‐2.1 days, 95% CI ‐4.5 to 0.3 days]. The outcome of hospital stay is highly dependent on the GA at birth of the neonate, the availability of institutions providing Level II care to which the neonate can be transferred and the social situation of the family.

It is possible that the IVIG preparations used in these studies did not contain the necessary antibodies to prevent infection and that the use of preparations with known specific antibodies against common pathogens in a specific neonatal intensive care unit might be more effective (Weisman 1994b).

The benefits of 3.0% and 4.0% reduction in sepsis and in any serious infection, respectively, should be weighed against the costs and the values assigned to this outcome. No serious side effects have been reported from IVIG to date, but unknown long‐term risks of administration of blood products and the pain associated with establishing an intravenous route for IVIG should be taken into account.

Units with high nosocomial infection rates may want to compare and adjust their infection control policies to those settings with low rates by using benchmarking techniques. If the rates remain high after such measures are taken, use of IVIG might be justified. Prophylactic use of IVIG should be based on a full economic evaluation and a clinical decision analysis that incorporates baseline risk for serious nosocomial infection, both clinical and economic outcomes following prophylactic IVIG and values attached to infections prevented. Such analyses have not been performed.

Although differences in inclusion criteria are seen, as well as differences in the number of studies published at the time of the reviews and the number of statistical analyses performed, the results of our systematic review are close to those of three previous meta‐analyses (Lacy 1995, Jenson 1997, Ohlsson 1998). The results of these meta‐analyses should encourage basic scientists and clinicians to pursue other avenues to enhance the immune system of preterm and/or LBW infants and to prevent nosocomial infection.

Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.1 Sepsis, one or more episodes.
Figures and Tables -
Figure 1

Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.1 Sepsis, one or more episodes.

Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.2 Any serious infection, one or more episodes.
Figures and Tables -
Figure 2

Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.2 Any serious infection, one or more episodes.

Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.4 Mortality (all causes).
Figures and Tables -
Figure 3

Forest plot of comparison: 1 IVIG versus placebo or no treatment, outcome: 1.4 Mortality (all causes).

Comparison 1 IVIG vs placebo or no treatment, Outcome 1 Sepsis, one or more episodes.
Figures and Tables -
Analysis 1.1

Comparison 1 IVIG vs placebo or no treatment, Outcome 1 Sepsis, one or more episodes.

Comparison 1 IVIG vs placebo or no treatment, Outcome 2 Any serious infection, one or more episodes.
Figures and Tables -
Analysis 1.2

Comparison 1 IVIG vs placebo or no treatment, Outcome 2 Any serious infection, one or more episodes.

Comparison 1 IVIG vs placebo or no treatment, Outcome 3 NEC, one or more episodes.
Figures and Tables -
Analysis 1.3

Comparison 1 IVIG vs placebo or no treatment, Outcome 3 NEC, one or more episodes.

Comparison 1 IVIG vs placebo or no treatment, Outcome 4 Mortality (all causes).
Figures and Tables -
Analysis 1.4

Comparison 1 IVIG vs placebo or no treatment, Outcome 4 Mortality (all causes).

Comparison 1 IVIG vs placebo or no treatment, Outcome 5 Mortality (infectious).
Figures and Tables -
Analysis 1.5

Comparison 1 IVIG vs placebo or no treatment, Outcome 5 Mortality (infectious).

Comparison 1 IVIG vs placebo or no treatment, Outcome 6 Duration of hospitalisation.
Figures and Tables -
Analysis 1.6

Comparison 1 IVIG vs placebo or no treatment, Outcome 6 Duration of hospitalisation.

Comparison 1 IVIG vs placebo or no treatment, Outcome 7 Bronchopulmonary dysplasia.
Figures and Tables -
Analysis 1.7

Comparison 1 IVIG vs placebo or no treatment, Outcome 7 Bronchopulmonary dysplasia.

Comparison 1 IVIG vs placebo or no treatment, Outcome 8 Intraventricular haemorrhage any grade.
Figures and Tables -
Analysis 1.8

Comparison 1 IVIG vs placebo or no treatment, Outcome 8 Intraventricular haemorrhage any grade.

Comparison 1 IVIG vs placebo or no treatment, Outcome 9 Intraventricular haemorrhage grade 3 or 4.
Figures and Tables -
Analysis 1.9

Comparison 1 IVIG vs placebo or no treatment, Outcome 9 Intraventricular haemorrhage grade 3 or 4.

Comparison 1. IVIG vs placebo or no treatment

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

1 Sepsis, one or more episodes Show forest plot

10

3975

Risk Ratio (M‐H, Fixed, 95% CI)

0.85 [0.74, 0.98]

2 Any serious infection, one or more episodes Show forest plot

16

4986

Risk Ratio (M‐H, Fixed, 95% CI)

0.82 [0.74, 0.92]

3 NEC, one or more episodes Show forest plot

7

4081

Risk Ratio (M‐H, Fixed, 95% CI)

1.08 [0.89, 1.32]

4 Mortality (all causes) Show forest plot

15

4125

Risk Ratio (M‐H, Fixed, 95% CI)

0.89 [0.75, 1.05]

5 Mortality (infectious) Show forest plot

10

1690

Risk Ratio (M‐H, Fixed, 95% CI)

0.83 [0.56, 1.22]

6 Duration of hospitalisation Show forest plot

8

3562

Mean Difference (IV, Fixed, 95% CI)

‐2.12 [‐4.54, 0.30]

7 Bronchopulmonary dysplasia Show forest plot

2

176

Risk Ratio (M‐H, Fixed, 95% CI)

1.55 [0.85, 2.84]

8 Intraventricular haemorrhage any grade Show forest plot

4

3176

Risk Ratio (M‐H, Fixed, 95% CI)

1.02 [0.88, 1.19]

9 Intraventricular haemorrhage grade 3 or 4 Show forest plot

2

3000

Risk Ratio (M‐H, Fixed, 95% CI)

1.01 [0.85, 1.21]

Figures and Tables -
Comparison 1. IVIG vs placebo or no treatment