Oral immunoglobulin for preventing necrotizing enterocolitis in preterm and low birth weight neonates

Necrotizing enterocolitis (NEC) is the most common emergency involving the gastrointestinal tract occurring in the neonatal period. NEC is characterized by acute onset of intestinal inflammatory necrosis which exhibits as abdominal distension, gastrointestinal bleeding and pneumatosis intestinalis on abdominal X-ray (Tudehope 2005). The origin of the intramural gas has been presumed to be from bacterial fermentation from gas-producing bacteria and a substrate (milk) (Willoughby 1994). NEC is a disease of the newborn, which indicates that the pathogenesis is somehow linked to physiologic characteristics unique to the newborn intestine (Edelstone 1982). The majority of neonates with NEC are premature or low birth weight (Cikrit 1984). The pathogenesis of NEC appears to be multifactorial, with any unifying hypothesis of its cause and prevention remaining unproven (Stoll 1994a). NEC is reported to be due to contributory factors such as mucosal injury caused by ischemia, infection and intraluminal injury with subsequent circulatory, immunologic and inflammatory host responses to the injury (Stoll 1994b). It is now well established that an exaggerated release of mediators of inflammation induced by microbial factors such as bacterial endotoxin, plays an important role in the development of noxious sequelae that follow infection of the host with pathogenic micro organisms (Wolf 1994; Eibl 1994). An exaggerated release of mediators of inflammation has also been implicated in the pathogenesis of NEC

The reported mortality rate for NEC is between 20 to 25% (Stoll 1994a) and has not changed appreciably over the past several decades (Petrosyan 2009). Long-term outcome is less certain. At discharge, many of the neonates remain at significant risk of frequent and prolonged hospitalisations due to nutritional compromise or stricture as a consequence of NEC (Petrosyan 2009). Approximately 20 to 40% of neonates who develop NEC eventually require surgical intervention Hsueh 2003; Petrosyan 2009). This leads to increased resource utilisation and possibly impaired developmental outcome (Hintz 2005). Additional morbidity arises from parental emotional grief and financial costs (Simon 1994).

Immunoglobulins play an essential role in the body's immune system. Immunoglobulins are large glycoproteins that are secreted by plasma cells and function as antibodies in the immune response by binding with specific antigens. They attach to foreign substances such as bacteria, and assist in destroying them. There are five classes of immunoglobulins: IgA, IgD, IgE, IgG, and IgM. Oral immunoglobulin may provide a prophylactic effect against NEC because of its immunoprotective effect, or its heterologous antibodies against infection of the gastrointestinal tract (Wolf 1994).

It has been proposed that orally administered antibodies bind to the antigen at the level of the gastrointestinal mucosa, which leads to intra-luminal agglutination of potentially infectious pathogens and, thus, interfere with colonisation of the mucosal surface by infectious pathogens, and neutralises bacterial toxic factors or viral particles (Wolf 1994). IgA, being a secretory immunoglobulin, might be expected to be more efficacious in protecting the neonatal gastrointestinal tract than the more readily available IgG. Bauer 1992 reviewed three trials of prophylactic intravenous immunoglobulin administration and reported a borderline statistically significant reduction of NEC.

There have also been reports of the effectiveness of using oral immunoglobulins as prophylaxis against NEC in premature and low birth weight neonates (Wolf 1994). It has been proposed that oral immunoglobulins produce an immunoprotective effect in the gastrointestinal mucosa. However, there are concerns regarding the strength of the evidence of the effectiveness of the use of oral immunoglobulins. The authors have been unable to identify any previous systematic reviews on the use of oral immunoglobulin for the prevention of NEC.

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Electronic searches

Searching other resources

The procedures of the Cochrane Neonatal Review Group (CNRG) were followed throughout.

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Assessment of heterogeneity

Data synthesis

Subgroup analysis and investigation of heterogeneity

See: tables Characteristics of included studies; Characteristics of excluded studies.

Results of the search

Included studies

Excluded studies

Details of the methodologic quality assessments are given in the table 'Characteristics of Included Studies'.

Allocation (selection bias)

Blinding (performance bias and detection bias)

Incomplete outcome data (attrition bias)

Although the trials reported outcomes such as definite NEC, suspected NEC, death and need for surgery, the pre-specified outcomes 'length of stay in hospital, hospital readmission, total parenteral nutrition administration, growth and development in childhood, parental emotional and financial costs' were not reported in any of the studies. Death was reported as 'during and after the study period', but not reported as 28 days post-delivery, discharge or by one year (as listed in the prespecified outcome measures). It was not anticipated that several outcome measures would be reported 'after the study period'. The decision to report these outcomes was made post-hoc. No data were available for subgroup analysis other than with regard to class of immunoglobulin, IgG/IgA.

In this review of three trials (including a total of 2095 neonates), two of the studies (Rubaltelli 1991; Lawrence 2001) investigated the use of IgG (nil or trace IgA) and one study (Eibl 1988) investigated the use of an IgG/IgA combination. The administration of oral immunoglobulin did not reduce the incidence of definite NEC, suspected NEC, surgery related NEC, or death from NEC, either during or after the study period.

Definite NEC during study period (Outcome 1.1):

Three trials reported the incidence of definite NEC during the study period and there was no reduction in any trial or overall [typical RR 0.84 (95% CI 0.57 to 1.25), typical RD -0.01 (95% CI -0.03 to 0.01)].

Definite NEC after the study period (Outcome 1.2):

Two trials reported the incidence of definite NEC after the study period and there was no reduction in either trial or overall [typical RR 1.30 (95% CI 0.47 to 3.60), typical RD 0.00 (95% CI -0.01 to 0.01)].

Suspected NEC during the study period (Outcome 1.3):

One trial reported the incidence of suspected NEC during the study period and there was no statistically significant reduction [RR 0.84 (95% CI 0.49 to 1.46), RD -0.01 (95% CI -0.02 to 0.01)].

NEC related surgery during the study period (Outcome 1.4):

The two small trials reported the number of neonates requiring surgery during the study period. There was no statistically significant reduction in either trial or overall [typical RR 0.21(95% CI 0.02 to 1.75), typical RD -0.03 (95% CI -0.06 to 0.00)]

NEC related deaths during the study period (Outcome 1.5):

Three trials reported the incidence of NEC related deaths during the study period and there was no reduction in any trial or overall [typical RR 1.10 (95% CI 0.47 to 2.59), typical RD 0.00 (95% CI -0.01 to 0.01)].

NEC related deaths after the study period (Outcome 1.6):

One trial reported the incidence of NEC related deaths after the study period and there was no reduction [RR 1.98 (95% CI 0.18 to 21.81), RD 0.00 (95% CI 0.00 to 0.01)].

Subgroup analysis:

There were two studies (Rubaltelli 1991, Lawrence 2001) that investigated the use of oral IgG (nil or trace IgA). The Lawrence 2001 study was large compared to the Rubaltelli 1991 study and thus dominated the results. Oral IgG did not reduce the incidence of definite NEC during the study period [typical RR 0.95 (95% CI 0.63 to 1.42), typical RD 0.00 (95% CI -0.02 to 0.02)], suspected NEC [RR 0.84 (95% CI 0.49 to 1.46), RD -0.01 (95% CI -0.02 to 0.01)], need for surgery [RR 0.34 (95% CI 0.01 to 8.28), RD -0.01 (95% CI -0.06 to 0.03)] definite NEC after the study period [typical RR 1.30 (95% CI 0.47 to 3.60), typical RD 0.00 (95% CI -0.01 to 0.01)] or death due to NEC during the study period [typical RR 1.39 (95% CI 0.55 to 3.55), typical RD 0.00 (95% CI -0.01 to 0.01)].

There was only one study (Eibl 1988) that investigated the use of an oral IgA/IgG combination (73% IgA, 26% IgG). There were trends for this combination to reduce the incidence of definite NEC during the study period [RR 0.08 (95% CI 0.00 to 1.39), RD -0.07 (95% CI -0.12 to -0.01), NNT 14 (95% CI 8 to 100)], death due to NEC during study period [RR 0.21 (95% CI 0.01 to 4.25), RD -0.02 (95% CI -0.06 to 0.02)] and need for surgery [RR 0.15 (95% CI 0.01 to 2.82), RD -0.03 (95% CI -0.08 to 0.01)]. None of the results were statistically significant.

One of the studies (Lawrence 2001) reported an increased incidence of Heinz bodies in the experimental group receiving oral immunoglobulin. However, the proportion of neonates given blood transfusions was similar in the experimental and control groups (62.2% vs. 69.7%) suggesting that clinically important haemolysis did not occur. There were no other reported side effects from the administration of oral immunoglobulin.

Three trials were included in the review. The randomised trials by Rubaltelli 1991 and Eibl 1988 were small and outcome assessment was not blinded. The Eibl 1988 study also had a large number of post randomisation exclusions in both the experimental and control groups (59%). The study by Lawrence 2001 was a large randomised, placebo-controlled and double-blinded study. Rubaltelli 1991 and Eibl 1988 excluded neonates that received breast milk whereas, in the Lawrence 2001 trial 90% received breast milk. Breast milk has previously been reported to have a protective effect against NEC (Lucas 1990) and this was acknowledged by Lawrence 2001. Rubaltelli 1991 and Eibl 1988 used similar dosages of oral immunoglobulin, 500 mg/day and 600 mg/day, respectively. Lawrence 2001 used a higher dose of 1200 mg/kg/day. The larger dose of oral immunoglobulin does not appear to have produced a greater response.

Eibl 1988 and Rubaltelli 1991 administered the oral immunoglobulin within the first twenty four hours following birth. However, Lawrence 2001 did not administer the oral immunoglobulin until after the initiation of enteral feeding. Thus, in the Lawrence 2001 study, 31% of neonates did not start the treatment until the fifth day or later. The effect of the timing of the administration of immunoglobulin on the incidence of NEC is unknown. However, in clinical practice it would be difficult to administer oral immunoglobulins to neonates who were unable to tolerate fluids orally.

The trials by Lawrence 2001 and Rubaltelli 1991 used predominately IgG. The study by Eibl 1988 used an immunoglobulin mixture containing 73% IgA and 26% IgG. To date, there is no randomised trial of IgA alone in the prevention of NEC, and the question of whether IgA has a protective effect against NEC is unanswered.

Eibl 1988 studied neonates weighing between 800 to 2000 g. Rubaltelli 1991 studied neonates weighing less than 1500 g or less than or equal to 34 weeks gestation and Lawrence 2001 similarly studied neonates weighing less than or equal to 1500 g. The association between prematurity or low birth weight and NEC are well known. Despite the increasing survival rate of extremely low birth weight (ELBW) neonates, there are no published randomised studies exclusively targeting ELBW neonates. It would clearly be important to stratify the groups at risk by gestational age and weight.

In the Lawrence 2001 trial, 10 out of 43 cases of definite NEC in the treatment group and 12 out of 41 cases of definite NEC in the control group did not receive any of the trial solutions prior to their illness. In the Eibl 1988 trial the rate of exclusion of neonates after randomisation was high (59%). Only one of the trials performed sample size estimation (Lawrence 2001).

Generally, the trials were of good methodological quality.

Based on the available trials, the evidence does not support the administration of oral immunoglobulin for the prevention of NEC. There are no randomised controlled trials of oral IgA alone in the prevention of NEC.

Future trials should examine the effects of oral IgA in extremely low birth weight neonates < 1000 g. In addition to examining effect on NEC, consideration should be given to reporting outcomes such as length of stay in hospital, hospital readmissions, need for total parenteral nutrition administration, growth and development in childhood and parenteral emotional and financial costs in any future studies . Given an incidence of NEC in this population of 8%, 1000 patients would be required to show a 50% reduction in NEC at the 5% level (two tailed).

Risk of bias table

Risk of bias table

Risk of bias table

Eibl MM, Wolf HM, Furnkranz H, Rosenkranz A. Prevention of necrotizing enterocolitis in low-birth-weight infants by IgA-IgG feeding. New England Journal of Medicine 1988;319:1-7.

* Lawrence G, Tudehope D, Baumann K, Jeffery, H et al. Enteral human IgG for prevention of necrotising enterocolitis: a placebo-controlled, randomised trial. Lancet 2001;357:2090-4.

Lawrence GW, The Australian NEC Study Group, Baumann K, Swanson C. Controlled double blind trial of oral human IgG in preventing neonatal enterocolitis (NEC). In: Proceedings of the Australian New Zealand Perinatal Society. 1996:A80.

Rubaltelli FF, Benini F, Sala M. Prevention of necrotizing enterocolitis in neonates at risk by oral administration of monomeric IgG. Developmental Pharmacology and Therapeutics 1991;17:138-43.

Fast C, Rosegger H. Necrotizing enterocolitis prophylaxis: oral antibiotics and lyophilized enterobacteria vs oral immunoglobulins. Acta Paediatrica. Supplement 1994;396:86-90.

Richter D, Bartmann P, Pohlandt F. Prevention of necrotizing enterocolitis in extremely low birth weight infants by IgG feeding? European Journal of Pediatrics 1998;157:924-5.