Nutritional and Educational Intervention to Recover a Healthy Eating Pattern Reducing Clinical Ileostomy-Related Complications

The aim of this study was to evaluate a diet intervention implemented by our hospital in order to determinate its capacity to improve the eating pattern of patients with an ileostomy, facilitating the implementation new eating-related behaviors, reducing doubt and dissatisfaction and other complications. The study was conducted with a quasi-experimental design in a tertiary level hospital. The elaboration and implementation of a nutritional intervention consisting of a Mediterranean-diet-based set of menus duly modified that was reinforced by specific counseling at the reintroduction of oral diet, hospital discharge and first follow-up appointment. Descriptive, bivariate and multivariate analyses were performed. The protocol was approved by the competent Ethics Committee. The patients of the intervention group considered that the diet facilitated eating five or more meals a day and diminished doubt and concerns related to eating pattern. Most patients (86%) had a favorable experience regarding weight recovery and a significant reduction of all-cause readmissions and readmission with dehydration (p = 0.015 and p < 0.001, respectively). The intervention helped an effective self-management of eating pattern by patients who had a physical improvement related to hydration status, which, together with an improvement in weight regain, decreased the probability of readmissions.


Introduction
In the management of colorectal cancer (CRC) or inflammatory bowel disease (IBD), it may be necessary to perform an ileostomy, which is a skin exteriorization from an ileal segment to eliminate feces either temporarily or permanently [1].
On average, 50% of the cases present clinical complications, however, this percentage can rise to 96% during the first three post-operative weeks where hydroelectrolytic alterations account for 20-29% [2]. In this period, up to 16% of these patients will present high fecal outputs through the stoma (>2000 mL/24 h), which supposes higher risk of dehydration, electrolyte imbalance and malnutrition [2]. Thus, readmission 60 days after discharge is common in these patients [3].
Factors such as the surgery, the underlying pathology and other treatment options are related to the nutritional and hydroelectric changes reported. Micronutrient deficit, malabsorption of bile fats and salts, caloric malnutrition and hydroelectric and weight losses are

Intervention
The elaboration and implementation of this intervention included endocrinologists, nutrition technicians and stoma therapists. A specific oral diet with a Mediterranean pattern was designed due to its widely proved beneficial effects to health, and for being the characteristic eating pattern of this region, easing adherence. Liquid, initial, soft (taken during hospitalization) and basal (taken at home) variants were elaborated to allow its progressive and successful introduction. For each variant, autumn/winter and spring/summer models were elaborated, allowing the use of seasonal food products in its preparation. For basal diets, 14 full menus were designed for lunch and dinner, as well as examples for breakfast and teatime, which were handed in to the patients after discharge (Appendix B).
Furthermore, patients received an educational session three times in different moments (reintroduction of diet after surgery, at discharge and during first visit after discharge). All lasted around 30-35 min where the stoma therapist explained key information on nutrition and eating habits to avoid complications during the first session. The key messages were summarized and provided in writing to the patient in the first session (Appendix B) and served as a follow-up script for the remaining sessions when the stoma therapist received feedback on the habits implemented and difficulties or complications associated with the eating dimension, providing solutions and reinforcing positive habits.

Procedures
The design of the menus and the health education session, as well as their approval by the hospital management, lasted from October 2017 to April 2018. The implementation of the intervention began in April 2018, when the prospective recruitment of patients for the intervention group began, and lasted until July 2020. The patients were recruited after surgery according to inclusion criteria (adults with ileostomy who signed the informed consent form, followed-up in our unit and who stated their intention to follow the diet proposed). Subjects with sensory/cognitive constraints and non-Spanish speakers were excluded due to the risk of misunderstanding the indications.
The control group was recruited retrospectively during their medical review visits between January and July 2018, excluding patients who underwent surgery before 2016 to minimize the memory bias.
A 23-item survey was elaborated, where sociodemographic (age, gender, marital status, schooling level, occupation, family situation), clinical (type of ileostomy, centimeters of ileum removed, etiology, chemotherapy treatment, incidence of gastrointestinal symptoms in case of chemotherapy treatment, self-care level) and anthropometric (weight and BMI) variables were measured. Due to the absence of a validated scale to determine the self-care level in these patients, this is usually determined by the stoma therapist through observation and interviews with the patients, who were assigned autonomous, semi-dependent or dependent levels.

Outcomes
To assess outcomes related to eating pattern management a 5-point Likert scale was used to determine the difficulty to implement the diet recommendations and the usefulness of the intervention by the patients. Additionally, onset of gastrointestinal problems related to the diet, compliance of the diet guidelines received, emergence of doubts in relation to the food products to eat and to their preparation, and if they considered that the diet was adequate were collected as dichotomous (yes/no) questions. The number of meals eaten per day and evolution of weight were also collected. The control group patients were asked if they recalled having gained, lost or maintained weight after 1-3 months of the ileostomy. The intervention group was followed up obtaining BMI values at the time of the surgery (baseline), weight at discharge, and at the first and second scheduled appointments (7-14 days and 30-60 days after discharge, respectively). Readmissions at 60 days after discharge, including dehydration among causes, were collected from clinical digital records from our emergency department.

Data Analysis
Data were analyzed in the strictest confidentiality with IBM SPSS software v22 (IBM, Chicago, USA). A descriptive analysis by absolute and relative frequencies and median and interquartile range was applied as appropriate. Pearson's chi-square test was used to evaluate the association in qualitative variables. Wilcoxon signed-rank test and Kruskal-Wallis test were performed for comparison of means checking normality tests beforehand. Finally, multivariate models were implemented to elucidate possible confounding bias. Statistical significance was defined as obtaining a p-value under 0.05.

Ethical Aspects
The Declaration of Helsinki and Guides for a Good Clinical Practice were taken into account to conduct this study. Protocol was approved by the competent Ethics Committee. The patients were informed about the study objectives and dynamics prior to their inclusion. This information was offered together with the informed consent document on the follow-up visits for the control group patients, and after the surgery in the intervention group patients.

Results
The number of patients included was 253, with 117 in the control group (71% of all the patients seen during the inclusion period) and 136 in the intervention group, where, initially, 164 met the inclusion criteria, but there were 28 losses (5 refusals to participate, 2 problems with language, 3 relevant missing data, 6 high-debit ileostomy, 9 unfavorable life prognosis and 3 deaths).
Age ranged between 18-89 years old, distributed with a mean and standard deviation of 58.5 ± 17.5 and 59.3 ± 15.6 in the control and intervention groups, respectively, without differences between them (p = 0.854). Except for the educational level (p = 0.042), no statistically significant differences were found between the groups for the sociodemographic characteristics (Table 1).  Table 2 shows the descriptive analysis of the clinical, anthropometric and outcome variables. No significant differences related to etiology, type of ileostomy or autonomy in care were observed between groups. There was difference for the "centimeters of ileum removed", whose mean and standard deviation were 20.5 ± 43.1 and 10.7 ± 18.6 cm in the control and intervention groups (p = 0.016), respectively. The most frequent etiology was CRC (60.9%) followed by inflammatory disease (27.3%) and other causes (11.8%) where different types of trauma were included. However, no differences were observed between groups regarding etiology. On the other hand, statistically significant differences were found for all outcomes, always in favor of the intervention group. Patients in the control group were asked if they thought that such an intervention would have been useful for a better management of their eating pattern. Among them, 109/117 (93.1%) assessed possible usefulness with four and five points (Likert scale). These points were assigned by 133/136 (98%) of the intervention group patients. It was not possible to assess statistical differences due to that, in the control group, the question explored the need, and in the intervention group, it was confirmatory of usefulness.
In the logistic regression models the intervention was identified as a risk factor to consider that the diet was adequate, and as a protective factor for eating five meals or more a day, being concerned when preparing the meals and having doubts in relation to the diet. A linear regression model showed that the intervention maintains an inverse relationship with the difficulty to implement the recommendations (Table 3). Weight gain and weight loss was reported by 43 (36.75%) and 69 (58.97%) control patients, respectively, leaving only 5 (4.27%) with a stable weight after the surgery. In the intervention group (Table 4) significant differences were observed at discharge and follow-up appointments 1 and 2, although this only involved significant differences in the weight loss percentage in the last assessment coinciding with significant difference in days from baseline to this timepoint. Among them, 92% attended the first scheduled review appointment at 10.5 ± 4.8 days and 96% attended the second appointment at 42.5 ± 10.9 days (mean and SD). Regarding readmissions at 60 days after discharge, patients in the control group had a readmission rate of 29.7%, with the specific rate of readmission due to dehydration being 17.8%. This values for intervention group were 16.2% and 4.4%, respectively. A statistically significant reduction between groups was observed for theses outcomes (p = 0.015 and p < 0.001, respectively). Odd ratios (CI95%) for intervention regarding the control group were 0.46 (0.24-0.87) for total readmissions and 0.21 (0.07-0.56) for specific readmission due to dehydration.

Discussion
This is a quasi-experimental study that has assessed a nutritional intervention in patients with an ileostomy. Usually, these patients simply receive a list of non-recommendable food products and culinary techniques, previous to their nutritional assessment [12,17,[20][21][22][23]. However, we have not found papers that offer an instrument to mitigate the patient's uncertainties during the elaboration of meals and improve the self-management of their eating pattern while inducing improvements related to weight and fluid balance.

Patients' Profiles
It was similar (57.7% men and 59 ± 16.5 years old) to other Western countries with the predominance of CRC patients (55.6% in the control group and 65.4% in the intervention group, p > 0.05) [24,25].
Educational level was lower in the control group, which could affect the understanding of the recommendations. Although 52% of the subjects were considered as autonomous patients in self-care, only two individuals stated living alone and without the support of family members, implying that most of them would have their care needs covered.
The main etiology was malignant pathology (55.6% in the control group and 65.4% in the intervention group), as previously reported [26]. Although there was no statistical difference in the cancer prevalence between the groups, it was verified by the number of patients who underwent chemotherapy (p = 0.049); however, the rates of gastrointestinal symptoms were similar, excluding this as a possible bias to improve the eating pattern.

Outcomes
The multivariate models showed that the intervention is a highly protective factor regarding the patients' concern at the time of preparing the meals (OR: 0.05, 95%CI: 0.02-0.12), and the doubts related to the food products (OR: 0.08, 95%CI: 0.04-0.14). This is a positive result since patients with ileostomy feel anxiety, confusion and frustration in relation to how to address diet [13]. Thus, this could lead them to adapt recommendations to their lifestyles and preferences on their own [27], increasing the risk of doing so inadequately. An inverse relationship was observed between belonging to intervention group and having difficulties in implementing the recommendations. These benefits are corroborated with the results observed in Models 1 and 3 (Table 3). In addition, our research significantly reduced the proportion of patients with gastrointestinal problems related to diet (OR:0.06, 95%CI:0.03-0.37).
Reducing the amount of food eaten in each meal and increasing the number of meals per day is beneficial for these patients [28,29]. Thus, Mukhopadhyay recommended an oral diet with six-eight meals/day with a reduction in the amounts eaten, leading all patients to recover their pre-surgery weight within 3 months [10].
Intestinal adaptation favors a partial recovering of intestinal function. This process is conditioned by the presence of food and secretions and begins 2-3 months after surgery [5]. Observational studies including patients, showed weight gain with respect to their presurgery weight 12 and 8 weeks after the procedure [10,21]. Likewise, in a retrospective study 13.3% of the patients lost weight after surgery and 68% presented normal weight at reconstruction surgery [22]. Vasilopoulos observed severe weight loss (>3 kg) in 53.8% of patients 3 weeks after the surgery [17]. At 6 weeks, this percentage rose to 70% when severe weight loss was defined as that higher than 7.5% of the pre-surgery weight [23]. Our results show a weight loss >7.5% regarding pre-surgery weight in 28.7% of the intervention group patients at 12.2 ± 8.3 days (discharge). This rose to 35.3% at 22.4 ± 10.8 days (first appointment), meaning an improvement compared with those previously reported. At 54.8 ± 13.9 days (second appointment) this ratio decreased to 25%. These severe weight losses could be a consequence of the post-surgical anatomical and functional loss and the absence of the intestinal adaptation process.
Other authors used the day of ileostomy as baseline to assesses the weight gain. However, we adopted the weight at discharge because it involved the self-management of the eating pattern, allowing observation of the results of the diets and nutritional guidelines implemented. All the weight assessments were within 90 days after ileostomy, with a mean and standard deviation of 54.8 ± 13.9 days. In the intervention group, 95 (69.85%) patients presented higher weight at the second appointment (42.5 ± 10.9 days).
The multivariate analysis shows that the weight loss differences were associated with pre-surgical BMI (higher in overweight and obese patients) rather than differences in days to attend the appointment. This could be prompted by the previous need to lose weight for the patients, who took advantage to obtain a healthy weight according to the clinical recommendations.
These results evidence, as previously, that early and maintained implementation of an oral diet is successful in recovering a patient's weight and/or to steer their BMI into a healthy range [10,21].
The intervention involved a significant reduction in readmission rate for all causes and due to dehydration (p = 0.015 and p < 0.001, respectively). A recent systematic review that included 27,089 patients showed that the global incidence of 30-and 60-day readmission with dehydration were 5.0% (range 2.1-13.2%) and 10.3% (range 7.3-14.1%), respectively [30]. In our study, the 60-day readmission with dehydration rates were 4.4% for intervention and 17.8% for the control group. The rate of readmission with dehydration in our intervention group was significant lower than three studies (marked in bold) that reported similar results to our control group. However, our control group showed a more increased rate than most of the studies included, but our intervention reduced the rate so that no significant differences were observed for the group that received it (Table 5). These improvements could be related to the increase in the number of intakes observed in the intervention group, since these could also be accompanied by a higher water intake.

Strengths/Limitations
An experimental design to compare the intervention vs. standard care was not possible because the hospital management ordered the intervention to be applied to all patients, considering that it was beneficial enough to exclude it in a group of subjects. For this reason, we were forced to retrospectively choose a control group prior to the implementation of the protocol in the service. Although significant differences were observed regarding the number of cm of ileus removed (greater in the control group), they were a number small enough to have clinical repercussions regarding the nutrition and hydration of the patients, since good tolerance of the loss has been described up to the middle of the small intestine (usual length from 6-8 m) [5]; none of the included patients lost lengths higher than 350 cm.
Our design represents an advantage over most previous studies, which were observational and did not contemplate a multivariate analysis to control the confusion bias derived from obtaining differences between groups such as "cm removed" or "educational level". Moreover, our sample size is larger than those previously reported, and this is the reason why we consider that there was high statistical power.
The lower rate of readmissions to the emergency room due to dehydration could be attributed to the fact that the intervention group was discharged during the COVID-19 pandemic and patients were afraid to go to the hospital; however, in our country during the confinement primary care services delivered telehealth care only. In addition, dehydration in this type of patient is severe and requires intravenous fluid therapy and being admitted to the hospital, so we ruled out that the patients would stay at home.
Finally, our intervention is complex because through the menus and recommendations we work on the nutrients provided that are of interest to these patients (fiber or fat), types of food and eating-related behaviors, making it difficult to identify a specific factor responsible for the improvement of results. However, this is common in the clinical context, where, unlike a laboratory, not only one variable is controlled, but many with the intention of offering comprehensive care to the patient.

Conclusions
This study is pioneering in proving the effectiveness of a nutritional intervention consisting of a Mediterranean-diet-based set of menus duly modified to allow for a progressive and effective self-management of the eating pattern of patients with an ileostomy. This was complemented with recommendations provided in an initial educational session and reinforced in two subsequent sessions; in addition, a brief written guide on these recommendations was added as documentation for home.
The success of this complex and multiapproach intervention has been evidenced through physical improvements observed in weight recovery, an improved hydration and a lower proportion of patients with gastrointestinal problems, behavioral improvements that were reflected in an increase in the number of intakes, decreased number of doubts regarding food selection/preparation, and readmission rate. The patients' satisfaction level was reflected through the positive assessment of the usefulness of the intervention and the consideration of having followed an adequate diet.

Conflicts of Interest:
The authors declare no conflict of interest.

Appendix A. Information Managed by Patients in the Control Group
Recommendations in case of diarrhea or liquid stools: Liquid stools are common with ileostomies and are also a normal occurrence after chemotherapy or radiotherapy treatments. On the other hand, an ileostomy patient, like any other person, can experience diarrhea after having eaten a certain food.
The recommended foods in case of diarrhea are: • White and toasted bread with crude oil. Recommendations for flatulence and bad odors: For a person with an ostomy, having excessive flatulence or gases can be annoying since they will not be able to control the moment they are expelled. So, it is important to know what foods cause them in order to try to avoid them if it is of social interest. Gases and bad odors are reduced by incorporating into the diet:

1100-Nutrition management
As an ileostomy carrier you will need to make some changes to your eating habits. In this way, you will avoid complications such as stoma obstruction, in addition to ensuring that your diet is capable of providing you with the nutritional requirements that you need for your day-to-day life. Remember: • Eat a varied and balanced diet, this will help your speedy recovery, as well as the healing of the ostomy and the surgical wound. • Increase the number of daily meals, make a minimum of 5 or 6, reducing the amount eaten in the main meals. In general, in foods of plant origin, you should remove the skin, seeds, filaments or threads, and all those parts that do not decompose well during digestion, as they could obstruct the stoma. • Avoid nuts, popcorn, dried fruit, mushrooms, sweet corn, coconut, orange albedo, celery, bean sprouts, peas, and raw vegetables in general, as these are incompletely digestible foods and could obstruct the ostomy • It is advisable to take fruit in juice or puree, avoiding pulp and seeds.

•
In the usual way prepare vegetables, legumes and vegetables in puree.

•
Avoid all those foods that have caused gastrointestinal discomfort or intolerance.

Basal Diet:
Its main objective is to provide a balanced and varied diet to the ileostomy patient who has satisfactorily tolerated the progression of the diet. This avoids those foods that can obstruct the ostomy, in this sense, vegetables and legumes are presented crushed and sifted. The fruit included attends to its characteristics and digestibility.
Brief Nutritional Assessment: