How Exclusive Does Exclusive Enteral Nutrition Need to Be to Be Effective?

See “6- and 12-Month Outcomes after 90:10 Enteral Nutrition Induction Therapy in Pediatric Crohn’s Disease” by Davidson et al on page 70.

Exclusive enteral nutrition (EEN) is a well-studied, first-line therapy for the induction of remission in mild-to-moderate pediatric Crohn disease (CD) (¹). Most studies show a remarkable 80%–85% remission rate, and some demonstrated superiority to steroids in inducing mucosal healing (odd ratio of ~5) (²). While effective, EEN can be associated with poor adherence and presents with multiple challenges (³). Mechanisms of action are yet to be fully elucidated but include the exclusion of “offending foods,” direct effects on host immune response, microbial modulation, and enhanced intestinal barrier function (⁴).

The “principle of exclusion” is important to the success of EEN: Johnson et al (⁵) demonstrated that when children with CD were randomized to receive completely exclusive EEN, compared with partial enteral nutrition (PEN; 50% of calories from normal diet), 50% of the EEN group achieved remission compared with 15% in the PEN group. In contrast, Levine et al (⁶) showed that a combination of PEN with the Crohn disease exclusion diet (CDED) is effective for induction of remission in mild-to-moderate CD, likely since in addition to formula, CDED was highly regulated (with exclusions), which is different from free diet. In fact, almost all diets that have shown success in CD mainly work by excluding processed and other specific foods. This is true for EEN, CDED, CD-Treat, specific carbohydrate diet (SCD), and anti-inflammatory diets (^6,7). PEN, however, is still a useful adjunct, for example, while awaiting the full benefit from other treatments, such as biologics (¹).

In this issue, Davidson et al (⁸) describe their retrospective experience with a “90:10 enteral nutrition induction therapy,” which we will call 90:10 enteral nutrition (EN), in 105 CD patients over 6 years. Only 42% of patients completed 8–12 weeks of 90:10 EN, with many requiring alternative induction therapy (steroid/anti-tumor necrosis factor). Eighteen continued a PEN protocol (80:20 and then 70:30), but the vast majority were either no longer on PEN or required additional therapy at 12 months. The authors emphasize the importance of further research in determining the utility of EEN as both an induction and maintenance therapy, and surmise if starting biologics earlier would provide patients with more benefit.

While this work, examining the long-term effects of this specific approach and PEN, is important and is derived from a relatively large cohort, there are further considerations when drawing conclusions from these data, which lead us to somewhat different conclusions.

First, while EEN is an effective and safe induction therapy, it is often not recommended as a sole maintenance therapy, mostly due to poor adherence. Once patients are in remission, immunomodulators or biologics are typically recommended for long-term management (¹). Therefore, the need for therapy change at 12 months does not mean that EEN (or 90:10 EN in this case) is not an effective induction therapy but rather questions whether long-term use of PEN is sustainable (with the possible exception of PEN as part of CDED). One should not expect an induction therapy (EEN) alone to have long-lasting effects in CD, and an alternative maintenance therapy will be required to maintain remission.

More importantly, this study demonstrates the importance of defining how exclusive EEN needs to be, that is, is 90:10 EN the same as EEN? In addition to the randomized controlled trial (RCT) by Johnson et al (⁵) mentioned above, showing that PEN is not as effective as EEN, we question whether instructing patients to consume 90% of calories from formula truly achieves this goal. In a study by Lee et al (⁹), patients in the PEN group (instructed to consume ~90% formula, with the rest as unrestricted diet) ultimately consumed 47% of their calories from table foods. Indeed, clinical remission rates in the study by Lee et al (⁹) were 50% for PEN vs 76% for EEN. The authors of this article do acknowledge that the institutional protocols of 10%, 20%, and 30% of calories from table foods introduced significant variability, depending on the family’s ability to calculate accurate calories and without documentation of how much patients actually consumed.

One of the factors that may have impacted the overall effectiveness of 90:10 EN is the low rates of therapy completion/adherence in this study (42% at 12 weeks), which is much lower than that observed in comparable studies (eg, 97.5% of patients tolerated the combination of PEN with CDED) (⁶); it is difficult to respond to a therapy if patients are not acutely receiving it (and given that this is a retrospective study, we do not know why the adherence was so low).

In the CDED RCT, 78 children with mild-to-moderate CD were randomizing to CDED plus 50% of calories from formula, followed by CDED with 25% PEN, or a group that received EEN followed by a free diet with 25% PEN (powered to assess tolerance, with secondary end points of remission at weeks 6 and 12) (⁶). In the CDED study, PEN was successful when combined with 50% exclusion diet [with superior adherence, tolerance (97.5% vs 73.6% with EEN), and steroid-free remission (75% vs 59% with EEN)] (⁶), again highlighting the importance of excluding potentially “offending foods” in achieving remission.

We agree that EEN can be a difficult therapy for patients, families, and care teams, and adherence to any diet can be challenging. Multidisciplinary team involvement is essential for success and might account for variable success rates of EEN and other diets. Patient selection and identifying barriers to adhering to a liquid diet are also important (¹⁰). Future studies to determine best supports for families would be helpful in improving adherence, as well as investigating how closely patients actually adhere to EEN.

In interpreting the findings of this study, we find further assurances to data, supporting the need for complete exclusion (in contrast to PEN, or even 90:10 EN) for the success of EEN. EEN continues to be effective in inducing remission in mild-to-moderate CD patients, as it not only promotes mucosal healing but is also beneficial from a nutritional and growth perspective.

Our main message in reading this article returns to how we started this commentary: EEN works for induction of remission, but PEN with a free diet is likely less effective due to the lack of exclusion. Either EEN or PEN are challenging as maintenance therapies for most patients (¹). This article has certainly contributed to knowledge on diet outcomes and highlights the importance of providing patients and families with a complete perspective on prognosis. It also highlights the importance of further research in dietary management of pediatric CD in both induction and maintenance phases.