Premature infants are often given glycerin enemas or suppositories to facilitate meconium evacuation and transition to enteral feeding. The purpose of this study was to assess the available evidence for this treatment strategy.
We conducted a systematic search of Medline, Embase, Central, and trial registries for randomized controlled trials of premature infants treated with glycerin enemas or suppositories. Data were extracted in duplicate and meta-analyzed using a random effects model.
We identified 185 premature infants treated prophylactically with glycerin enemas in one trial (n = 81) and suppositories in two other trials (n = 104). All infants were less than 32 weeks gestation and had no congenital malformations. Treatment was associated with earlier initiation of stooling in one trial (2 vs 4 days, P = .02) and a trend towards earlier meconium evacuation in another (6.5 vs 9 days, P = .11). Meta-analysis demonstrated no effect on transition to enteral feeding (0.7 days faster, P = .43) or mortality (P = 0.50). There were no reports of rectal bleeding or perforation but there was a trend towards increased risk of necrotizing enterocolitis with glycerin enemas or suppositories (risk ratio = 2.72, P = .13). These three trials are underpowered and affected by one or more major methodological issues. As a result, the quality of evidence is low to very low. Three other trials are underway.
The evidence for the use glycerin enemas or suppositories in premature infants in inconclusive. Meta-analyzed data suggest that treatment may be associated with increased risk of necrotizing enterocolitis. Careful monitoring of ongoing trials is required.
Many thanks to Dr. Viswanathan for his letter and interest in our study (1). Glycerin suppositories are used to treat premature infants in many neonatal intensive care units, including our own. These interventions are low-cost, appear to be low-risk, and anecdotally, appear to facilitate meconium evacuation and the transition to full enteral feeding (1).
At the bedside, the passage of meconium following the administration of glycerin suppositories suggests to many of us that this must be "a good thing" and associated with superior clinical outcomes. Our review demonstrated that the evidence for this practice is limited and inconclusive. The three trials published to date are small and affected by one or more other major methodological issues (2-4). As a result, these interventions may or may not be associated with either clinical benefit or harm, such as increased risk of necrotizing entercolitis. This is especially important to consider since glycerin suppositories have never been formally approved for use in this patient population.
We agree with Dr. Viswanathan that the numbers in this meta-analysis are small and do not carry enough statistical power to make specific recommendations. This has important implications that we acknowledged explicitly in our review: "The nature of the association between glycerin medications and necrotizing enetercolitis will become clearer as additional data become available. One possibility is that the apparent relationship is nothing more than a spurious correlation that will disappear with the inclusion of more data. The other explanation is that there is a real effect that has not yet become statistically significant."
In general, we would advise against meta-analyzing outcomes on a per protocol basis. In contrast to conventional intention-to-treat analyses, which analyze participants in the groups to which they were randomized, per protocol analyses are at high risk of selection bias due to an imbalance of known and unknown prognostic factors between groups. Furthermore, protocol violations often do not occur with the same frequency in each treatment arm, such as in the study by Haiden et al., where protocol violations were almost twice as frequent in the intervention group compared to controls (36% vs 21%) (2). Finally, even with the multiple post hoc modifications made by Dr. Viswanathan, the risk ratio in the revised analysis is still 2.00, which translates into a doubling of the risk of necrotizing enterocolitis with the use of glycerin medications. If this trend is eventually shown to be a statistically significant, we suspect most clinicians would stop using these interventions altogether.
The findings of our review should not change clinical practice. We do think, however, that all of us involved in the care of premature infants should think more critically about our beliefs and treatment paradigms, and be wary of developing treatment protocols that are not based on sound evidence. We would also suggest that the trends highlighted in our meta- analysis should inform future studies on the use of glycerin medications in premature infants.
We are now conducting a pilot study to assess the feasibility of a multicenter randomized trial on the effectiveness of glycerin suppositories among premature infants less than 32 weeks gestation (5). We have randomized 22 of 30 participants to date and look forward to sharing our results in the near future.
References
1. Livingston MH, Shawyer AC, Rosenbaum PL, et al. Glycerin Enemas and Suppositories in Preterm Infants: A Meta-analysis. Pediatrics. 2015;135(6):1093-1106
2. Haiden N, Jilma B, Gerhold B, et al. Small volume enemas do not accelerate meconium evacuation in very low birth weight infants. J Pediatr Gastroenterol Nutr. 2007;44(2):270-273
3. Khadr SN, Ibhanesebhor SE, Rennix C, et al. Randomized controlled trial: impact of glycerin suppositories on time to full feeds in preterm infants. Neonatology. 2011;100(2):169-176
4. Shinde S, Kabra NS, Sharma SR, Avasthi BS, Ahmed J. Glycerin suppository for promoting feeding tolerance in preterm very low birth weight neonates: a randomized controlled trial. Indian Pediatr. 2014;51(5):367-370
5. Walton JM. Glycerin Suppositories Used Prophylactically in Premature Infants (SUPP): a Pilot Randomized Placebo-controlled Trial. NCT02153606. https://clinicaltrials.gov/ct2/show/NCT02153606
Conflict of Interest:
None declared
I read with interest the meta-analysis of glycerin suppository and/or enema studies by Livingston et al.(peds.2015-0143). The authors have adequately acknowledged the major methodological issues with the 3 studies that were included in the meta-analysis (1-3). The study indicates that the evidence for routine prophylactic glycerine enemas/suppository in preterm infants is inconclusive.
I have major concerns with the conclusion that 'glycerine enemas/suppository treatment may be associated with increased risk of necrotising enterocolitis (NEC). The NEC incidence in their meta-analysis was 9/93 in treatment group and 3/86 in the control group (RR 2.72[0.76 to 9.81], p=0.13). However in Haiden et al, (1), all 3 cases of NEC happened in patients with protocol violations (i.e. all 3 NEC cases did not receive any glycerine enema). In light of the high rate of protocol violation (~25%), it would be more appropriate to include the per-protocol (PP) analysis results of their study in the meta-analysis (1). In Shinde et al (2), the 2 NEC cases reported in the treatment arm were 'stage 1' NEC which is traditionally excluded in most NEC studies because of its poor specificity. If we consider these factors in the analysis, the results would be as follows:
NEC events (Treatment)
NEC events (Controls)
Haiden et al, (PP analysis) (1)
0/27
0/31
Khadr et al, (2)
4/29
1/25
Shinde et al, (3)
0/21
1/21
Total NEC events
4/77
2/77
This analysis will give a RR 2.00[0.38 to 10.6], p=0.68. This suggests that the use of glycerine enema/suppository has no effect on the incidence of NEC.
Also, NEC was not the primary outcome in any of these studies. The meta-analysis shows primary outcome (time to reach full enteral feeds) was shorter in the treatment group, though not statistically significant. The studies also did not report any major adverse events related to treatment.
Pietz et al, reported one of the lowest incidence of NEC in very low birth weight infants in United States NICU (4). Their feeding protocol included regular use of glycerin suppositories if there is no bowel movement for 24 hours. The rationale was to prevent the development of intestinal distension, with has the potential to alter intestinal blood flow (5).
The occasional use of glycerine suppository in preterm infants with feeding intolerance associated with no bowel movements or no rectal gas shadows is very common in our practice, often with good results. I see no reason to change this practice based on the current available evidence. I think, the news of further RCT's in the pipeline is encouraging, and may help to address these issues in a more definitive way.
References
Haiden N, Jilma B, Gerhold B, et al. Small volume enemas do not accelerate meconium evacuation in very low birth weight infants. J Pediatr Gastroenterol Nutr. 2007;44(2):270-273
Khadr SN, Ibhanesebhor SE, Rennix C, et al. Randomized controlled trial: impact of glycerin suppositories on time to full feeds in preterm infants. Neonatology. 2011;100(2):169-176
Shinde S, Kabra NS, Sharma SR, Avasthi BS, Ahmed J. Glycerin suppository for promoting feeding tolerance in preterm very low birth weight neonates: a randomized controlled trial. Indian Pediatr. 2014;51(5): 367-370
Pietz J, Achanti B, Lilien L, Stepka EC, Mehta SK. Prevention of necrotizing enterocolitis in preterm infants: a 20-year experience. Pediatrics. 2007 Jan;119(1):e164-70
Meyers RL, Alpan G, Lin E, Clyman RI. Patent ductus arteriosus, indomethacin, and intestinal distension: effects on intestinal blood flow and oxygen consumption. Pediatr Res. 1991; 29:569-574
Conflict of Interest:
None declared