Early feeding reduces length of hospital stay in patients with acute lower gastrointestinal bleeding: A large multicentre cohort study

No studies have compared the clinical outcomes of early and delayed feeding in patients with acute lower gastrointestinal bleeding (ALGIB). This study aimed to evaluate the benefits and risks of early feeding in a nationwide cohort of patients with ALGIB in whom haemostasis was achieved.

Although antacid therapy is effective for upper GIB (UGIB), there are no effective medications for preventing ALGIB and its recurrence [5].Eating a meal stimulates the gastrocolic response, which increases colonic myoelectrical and motor activity [7].
Therefore, feeding after ALGIB might result in rebleeding due to the irritation caused by colonic peristalsis and the passage of faeces.Fasting is the traditional treatment for GIB and is believed to stabilize clots and reduce the risk of rebleeding [8].Although resumption of feeding is a key step after GIB, no studies have examined the appropriate timing for refeeding after ALGIB, and no ideal timing is recommended in guidelines from the United States [9], Europe [10] and Asia [11].The few studies that have compared the outcomes of early versus delayed feeding in patients with GIB have been limited to the upper gastrointestinal tract [12][13][14][15][16][17][18].Several randomized controlled studies found that early feeding had no significant effect on the risk of rebleeding in patients with UGIB treated by endoscopic haemostasis [14,16,18,19] but that the hospital stay was significantly shorter in the early feeding group than in the delayed feeding group [14,19].However, the results of these studies in UGIB cannot be extrapolated to patients with ALGIB, which has a different pathology.Clarifying the optimal time for resumption of feeding may inform treatment decisions for ALGIB, contributing to improvement of clinical outcomes and patient satisfaction.
The aim of this study was to evaluate the benefits and risks of early resumption of feeding in patients with ALGIB by comparing clinical outcomes between early and delayed feeding groups using nationwide data in Japan.We also compared outcomes between subgroups of presumptive and definitive colonic diverticular bleeding (CDB), because CDB is the most common cause of ALGIB [20,21] and has a high frequency of rebleeding and severe bleeding requiring arterial embolization and colectomy [22].Furthermore, it is reported that the rebleeding rate after conservative treatment for presumptive CDB is higher than that after endoscopic treatment for definitive CDB [22].

Patients and study design
This retrospective multicentre cohort study, the COlonic DivErtcular Bleeding Leaders Update Evidence from multicentre Japanese study (CODE BLUE-J study), was conducted at 49 hospitals across Japan [21,23].The ethics committees and institutional review boards of all 49 participating hospitals gave their approval for this study to be performed using the opt-out method for consent [21] (Table S1).
A total of 10,342 adult patients who were emergently hospitalized for acute haematochezia between January 2010 and December 2019 were enrolled in the study.Patients were excluded if they did not undergo colonoscopy, underwent colonoscopy >120 h after presentation, underwent surgery, interventional radiology (IVR), or barium impaction therapy as the initial treatment, or were diagnosed with UGIB (Figure 1).Given that we also intended to evaluate the effect of feeding within 3 days after colonoscopy on clinical outcomes, we excluded patients in whom endoscopic haemostasis could not be completed, those who developed haematochezia before feeding, those who had been fasted for more than 4 days after colonoscopy, and those for whom there was no information on the date of feeding or rebleeding after colonoscopy.This left a cohort of 5910 hospitalized patients with acute haematochezia in whom haemostasis had been achieved and feeding had resumed within 3 days after colonoscopy for analysis who were then divided into an early feeding group and a delayed feeding group (Figure 1).Early feeding [12,13] was defined as resumption of feeding within 1 day or sooner after colonoscopy and delayed feeding [12][13][14]18] as within 2-3 days after colonoscopy.

Variables investigated
We assessed 45 clinical variables, including baseline characteristics such as age, sex, vital signs on admission, lifestyle factors, presenting symptoms, laboratory data, comorbidities, and medication use within 30 days of admission and reviewed in-hospital examination findings obtained from the electronic medical records and endoscopic databases, as previously reported [21,23].
We evaluated 19 comorbidities using the Charlson Comorbidity index [24] as well as hypertension and dyslipidaemia.Endoscopic data were collected, including timing of colonoscopy, stigmata of recent haemorrhage (SRH), diagnosis after first colonoscopy, and method of endoscopic haemostasis.Timing of colonoscopy was divided into early (≤ 24 h), elective (24-48 h), and late (48-120 h) [25].SRH was defined as active bleeding or a densely adherent

What does this paper add to the literature?
In patients with acute lower gastrointestinal bleeding in whom haemostasis was successfully achieved, there was no significant difference in the rebleeding rate between the early and delayed feeding groups.The hospital stay after colonoscopy was significantly shorter in the early feeding group.
clot despite vigorous irrigation or a nonbleeding visible vessel on colonoscopy [26,27].The diagnosis of presumptive CDB was based on the following: very low likelihood of a bleeding source other than a colonic diverticulum by colonoscopy with other tests showing negative results, including upper gastrointestinal endoscopy and small bowel endoscopy; and computed tomography (CT) visualization of contrast medium extravasation localized to the diverticulum [21].Definitive CDB was based on colonoscopic visualization of diverticulum with SRH [21].Basically, conservative therapy was selected for presumptive CDB, and endoscopic treatment was performed for definitive CDB.
The haemostatic procedure was chosen at the discretion of the endoscopist.

Clinical outcomes
The outcome of interest was the rebleeding rate after colonoscopy, whether during hospitalization or after discharge.Rebleeding was defined as a significant amount of fresh bloody or wine-coloured stool after admission [21,28].The secondary outcomes were need for IVR or surgery, in-hospital mortality, and length of hospital stay after colonoscopy.

Statistical analysis
Categorical data were compared using the χ 2 test or Fisher's exact test as appropriate.Continuous data were compared using the Mann-Whitney U test.Propensity score matching (PSM) was used to adjust for differences between the two treatment groups [29].
A logistic regression model was used for propensity score estimation, with delayed feeding as a function of the patient's baseline characteristics and endoscopic factors.The model included age ≥70 years, sex, and 36 factors found to be of at least borderline significance (p < 0.10) in univariate analysis (Table 1).We performed one-to-one PSM between the early and delayed feeding groups using the nearest neighbour method within a calliper width of 0.2 of the standard deviation of the logit of the propensity score.Before matching, the area under the receiver operating characteristic curve for propensity scores for delayed feeding was 0.731 (95% CI: 0.717-0.746).
We also compared the clinical outcomes between the early and delayed feeding groups in subgroups of presumptive and definitive CDB and in the same way as for the main analysis.

p-value
Fasting period after colonoscopy, days

RE SULTS Patient characteristics and clinical outcomes
The characteristics of all patients (n = 5910) and the propensity score-matched patients (n = 3016) are shown in Table 1.One-to-one PSM identified 1508 patient pairs from the early and delayed feeding groups who were closely matched for baseline characteristics.
There was no significant difference in the rebleeding rate within 7 days after colonoscopy between the early and delayed feeding groups (9.4% vs. 8.0%, respectively; p = 0.196) or in the rebleeding rate within 30 days after colonoscopy (11.4% vs. 11.5%;p = 0.909) (Table 2).There was also no significant between-group difference in the need for IVR, need for surgery, or in-hospital mortality.However, the median hospital stay after colonoscopy was significantly shorter in the early feeding group (5 vs. 7 days; p < 0.001).

Clinical outcomes in subgroups of presumptive and definitive CDB
The baseline characteristics of patients in the subgroups of presumptive and definitive CDB before and after PSM are shown in Tables S2 and S3, respectively.One-to-one PSM identified 862 patient pairs in the presumptive CDB group and 493 patient pairs in definitive CDB group who were closely matched for baseline characteristics.In patients with presumptive CDB, there was no significant difference in the rebleeding rate, need for IVR, need for surgery, or mortality between the early and delayed feeding groups (Table 3).
The median hospital stay after colonoscopy was significantly shorter in the early feeding group than in the delayed feeding group (5 vs. 7 days; p < 0.001) (Table 3).Clinical outcomes were similar between the patients with definitive CDB and all patients (Table 3).

DISCUSS ION
To the best of our knowledge, this is the first study to compare the clinical outcomes of early feeding with those of delayed feeding in patients with ALGIB.The three important findings of this study are as follows.First, there was no significant difference in the rebleeding rate, need for IVR, need for surgery, or mortality between the early and delayed feeding groups.Second, the hospital stay after colonoscopy was significantly shorter in the early feeding group.Third, the results were unchanged in the subgroups of presumptive and definitive CDB.Therefore, this study establishes that early feeding has the advantage of shortening the hospital stay without increasing the risk of rebleeding in patients with ALGIB.
Our findings that there was no significant difference in the rebleeding rate between the early and delayed feeding groups and that the hospital stay after colonoscopy was significantly shorter in the  These results suggest that early feeding is safe and effective for patients with ALGIB and confirmed haemostasis.The importance of early oral feeding has long been stressed in the surgical field.Several randomized controlled trials [30][31][32][33] have demonstrated that early oral feeding after elective surgery for colorectal cancer has advantages, including a shorter postoperative hospital stay and less risk of postoperative morbidity and mortality when compared with traditional postoperative oral feeding.A meta-analysis also showed that early oral feeding is associated with a significant reduction in length of hospital stay and total postoperative complications when compared with traditional oral feeding in patients who have undergone elective colorectal surgery, with no significant difference in the risk of anastomotic dehiscence, pneumonia, wound infection, vomiting, or mortality [34].Based on the findings of these studies, the American Society of Colon and Rectal Surgeons and Society of American Gastrointestinal and Endoscopic Surgeons strongly recommend resumption of a regular diet immediately after colorectal surgery [35].
However, while the evidence for early oral feeding after colorectal surgery has been established, that for ALGIB is still scarce.Further studies are needed to translate the results of this study to clinical practice and establish an evidence base for early feeding after ALGIB.
Early feeding was beneficial in reducing the length of hospital stay after colonoscopy in this study.An extended hospital stay TA B L E 2 Clinical outcomes in all patients and propensity score-matched patients.
Surgery need during hospitalization Mortality during hospitalization 0 ( Length of hospital stay after colonoscopy, days reduces quality of life (QOL) during hospitalization and increases medical expenditure.Early postoperative initiation of oral feeding is reported to improve postoperative QOL and decrease complications [36,37].Early feeding has the potential to improve not only cost-effective use of medical resources but also QOL during hospitalization, and further studies are necessary from these perspectives.
Clinical outcomes were also compared in the presumptive and definitive CDB subgroups according to whether feeding was early or delayed.Endoscopic haemostasis was performed in patients with definitive CDB but not in those with presumptive CDB.
Although we expected that the effect of resumption of feeding would be greater in patients with presumptive CDB in whom endoscopic haemostasis was not performed than in those with definitive CDB, there were no significant differences in clinical outcomes according to whether CDB was presumptive or definitive.These results suggest that delayed feeding does not improve clinical outcomes, regardless of whether or not endoscopic haemostasis is performed.
Interestingly, when comparing the baseline characteristics of patients before PSM (Table 1, Tables S2 and S3), the item extravasation on CT was significantly higher in the delayed feeding group in all three tables.This finding suggests that physicians tended to choose delayed feeding when there was a factor that reflects active bleeding with increased risk of rebleeding [38,39].Moreover, of the patients in this study, 43.8% (2586/5910) were in the delayed feeding group, suggesting that a large number of physicians still seemed concerned that colonic peristalsis due to early feeding may cause rebleeding, although the majority tended to opt for early feeding.
However, the present study showed that early feeding did not increase rebleeding.We believe that these results will provide useful information for physicians who choose delayed feeding out of concern over rebleeding.
This study had some limitations.First, it was retrospective and the possibility of unmeasured confounders cannot therefore be completely ruled out.However, we tried to minimize the risk of bias by carefully selecting cases and analysing them by PSM.
Second, our database lacks detailed information on the content of meals.Third, the criteria for hospitalization differed from facility to facility and were not standardized.However, 96.9% (5724/5910) of the patients in this study had an Oakland score [3] of ≥9, which is recommended for inpatient treatment in Europe and the United States.Therefore, we consider that the results of this study would be applicable to clinical practice in Europe and the United States.This study also has the following three strengths.First, it included a large cohort (n = 5910) with few missing values [21,23].Second, we collected detailed information on, for example, extravasation on computed tomography and endoscopic data, including timing of colonoscopy, type of SRH, diagnosis after colonoscopy, and method of endoscopic haemostasis.Furthermore, the baseline characteristics between the two groups were balanced using PSM.Third, we confirmed the haemostatic status of all subjects clinically and endoscopically, which allowed the effect of feeding on rebleeding to be assessed as precisely as possible.

CON CLUS ION
In conclusion, this large nationwide study establishes that early re- The statistical analysis was performed using the Statistical Package for Social Sciences version 22 (IBM Corp.).All statistical significance was set at p < 0.05.F I G U R E 1 Flow chart of patients in this study.TA B L E 1 Baseline characteristics of all patients and propensity score-matched patients.
sumption of feeding in patients with ALGIB has the advantage of shortening the hospital stay without increasing the risk of rebleeding.Further trials are needed to determine exactly when feeding should be resumed to improve the overall quality of care in patients with ALGIB.AUTH O R CO NTR I B UTI O N S Takaaki Kishino: Conceptualization; data curation; formal analysis; investigation; methodology; project administration; writing -original draft; writing -review and editing.Tomonori Aoki: Investigation; methodology; project administration; writingoriginal draft; writing -review and editing.Eiji Sadashima: Formal analysis; investigation; methodology; project administration; writing -original draft; writing -review and editing.Katsumasa Kobayashi: Data curation; writing -original draft; writing -review and editing.Atsushi Yamauchi: Data curation; writing -review and editing; writing -original draft.Atsuo Yamada: Data curation; writing -review and editing; writing -original draft.Jun Omori: Data curation; writing -review and editing; writing -original draft.Takashi Ikeya: Data curation; writing -review and editing; writing -original draft.Taiki Aoyama: Data curation; writing -review and editing; writing -original draft.Naoyuki Tominaga: Data curation; writing -review and editing; writing -original draft.Yoshinori Sato: Data curation; writing -review and editing; writing -original draft.Naoki Ishii: Data curation; writing -review and editing; writing -original draft.Tsunaki Sawada: Data curation; writing -review and editing; writing -original draft.Masaki Murata: Data curation; writing -review and editing; writing -original draft.Akinari Takao: Data curation; writing -review and editing; writing -original draft.Kazuhiro Mizukami: Data curation; writing -review and editing; writing -original draft.Ken Kinjo: Data curation; writing -review and editing; writing -original draft.Shunji Fujimori: Data curation; writing -review and editing; writingoriginal draft.Takahiro Uotani: Data curation; writing -review and editing; writing -original draft.Minoru Fujita: Data curation; writing -review and editing; writing -original draft.Hiroki Sato: Data curation; writing -review and editing; writing -original draft.Sho Suzuki: Data curation; writing -review and editing; writing -original draft.Toshiaki Narasaka: Data curation; writing -original draft; writing -review and editing.Junnosuke Hayasaka: Data curation; writing -original draft; writing -review and editing.Tomohiro Funabiki: Data curation; writing -original draft; writing -review and editing.Yuzuru Kinjo: Data curation; writing -original draft; writing -review and editing.Akira Mizuki: Data curation; writing -original draft; writing -review and editing.Shu Kiyotoki: Data curation; writing -original draft; writing -review

All patients (n = 5910) Propensity score-matched patients (n = 3016) Early feeding n = 3324 Delayed feeding n = 2586 p-value Early feeding n = 1508 Delayed feeding n = 1508 p-value
Data are shown as the number (percentage) or median (interquartile range).p-values <0.05 are shown in bold.Clinical outcomes in subgroups of presumptive and definitive colonic diverticular bleeding.
Data are shown as the number (percentage) or median (interquartile range).p-values <0.05 are shown in bold.