Daytime only or time restricted feeding in critically ill patients: A scoping review

Abstract

Objective:

A scoping review was performed to understand the extent and type of published evidence in relation to restricting feeding of critically ill patients to the daytime only, with a nighttime fasting period.

Introduction:

Time-restricted feeding has been shown to be beneficial to long-term health. Critically ill patients admitted to intensive care units (ICUs) are traditionally fed continuously. The potential benefits or harms of daytime only feeding in the critically ill are unknown.

Inclusion criteria:

Studies of critically ill patients, cared for in any critical care environment, where feeding was stopped for a minimum of 6 h overnight, with any primary outcome, were included.

Methods:

Using the JBI framework, a search of OVID Embase, OVID Medline, CINHAL, PROSPERO, The Cochrane database and Web of Science was performed in July 2023.

Results:

Fourteen studies that included 868 participants, published between 1989 and 2023 met the inclusion criteria and were reported on. The patient cohorts were from general or mixed ICUs, and neurosurgical and paediatric cohorts. Feed was either administered by bolus, in cyclic patterns or continuously. The overnight fasting times ranged from 6 to 12 h, with reported primary outcomes of feed intolerance, nutritional delivery, ketosis, gastric pH, ventilator associated pneumonia and circadian rhythms. Daytime only feeding was found to increase ketosis and lower gastric acidity.

Conclusions:

Daytime only feeding in the critically ill has been reported, but details of its potential harms or benefits are limited by inconsistently defined outcomes and study small sample sizes.

Introduction

Intermittent feeding (or intermittent fasting) is associated with a number of long-term health improvements such as reduced oxidative stress, autophagy, improved sleep, improved blood sugar control, reduced risk of cancer, improved dementia and psychiatric risks and improved overall mortality.^1–4 In addition, a period of fasting in animals has demonstrated protection against age-related diseases through promoting cellular stress resistance. ⁵ The mechanisms underlying these benefits include activation of ketogenesis and autophagy, inflammation and glucose control relating to time restricted feeding.^1,6,7 Feeding time also plays a role in regulating a normal circadian rhythm which controls regulate core physiological systems including sleep and gut function.^5,8

Critical illness describes a life-threatening condition that usually involves organ dysfunction or failure, leading to admission to an intensive care unit (ICU). The range of diagnoses and disease processes resulting in critical illness are enormous and heterogenous, however significant inflammation is common to many. A hypermetabolic state induced by the stress responses is also a significant part of the pathological process. Morbidity and mortality are high amongst critically ill patients, with a high incidence of long-term ill health in the survivors of critical illness. ⁹ Traditionally, critically ill patients receive nutrition continuously, either by the enteral or parenteral route. It is possible that the benefits of time-restricted feeding observed in the community population may have relevance to critically ill patients and could lead to tangible improvements in their survival and long-term health.

Providing nutrition to critically ill patients is focused on delivering macro- and micronutrients enterally in preference to parenterally and minimising underfeeding. This is in the context of a population where 55% are malnourished on admission to an ICU. ¹⁰ Critically ill patients who cannot eat, are either fed enterally via a feeding tube into the gastrointestinal tract (commonly the stomach) or parenterally via a central vein. When administered enterally this can be achieved through intermittent boluses or continuously via an infusion, either cyclically for set periods of time or continuously over 24 h. Enteral feeding in critically ill patients, when not provided continuously may be provided by bolus or cyclically. In bolus feeding, a set volume is given over set durations at defined intervals, for example, 500 ml over 10 min every 4 h. In cyclic feeding, it is delivered continuously for a set period of time and paused for a set period of time within a 24 h cycle.

The European Society for Clinical Nutrition and Metabolism (ESPEN) guidelines for the provision of nutritional support therapy in the adult critically ill patient 2021, ¹¹ do not make any recommendations in relation to timing of enteral nutrition. For adult critically ill patients, most ICUs choose to provide continuous enteral feeding, due to the ease of administration via a pump. This delivery pattern is also supported by concerns that gastrointestinal malabsorption in critically ill patients may lead to vomiting and aspiration pneumonitis, and to minimise the risk of hyper- or hypoglycaemia. Bolus feeding and non-continuous feeding in ICU patients are also practiced, and there is conflicting evidence as to which method delivers more calories and is better tolerated. ¹²

A preliminary search of MEDLINE, the Cochrane Database of Systematic Reviews and JBI Evidence Synthesis was conducted and no current or underway systematic reviews or scoping reviews on the topic were identified. No reviews in relation to daytime only feeding or night-time time restricted feeding in the critically ill were identified.

Review articles describing delivery methods of enteral feeding for the critically ill indicate that there is published evidence that describes the intervention of daytime only feeding and associated outcomes. A scoping review was performed to encompass the wide range of feeding patterns which provide a night-time fasting period, that have been studied in critically ill patients. The objective of this scoping review is to assess the extent of the published evidence relating to critically ill patients who have had no enteral nutrition delivered for a minimum period of 6 h overnight. To identify studies where this pause might be reasonably described as a night-time fast, a minimum of 6 h night-time pause was chosen to provide a consistent inclusion criterion for the published evidence.

Review question

In critically ill patients who are fed via enteral feeding tube, what are the differences in outcomes when providing nutrition in the day and not at night, as compared to feeding at anytime of day?

Context

Studies that specified that the patients were mechanically ventilated or critically ill requiring invasive organ support were included, regardless of the country of origin. Studies from all geographical areas and all age groups were included.

Methods

The scoping review was conducted and reported in accordance with the JBI methodology for scoping reviews. ¹³

Study eligibility criteria

The inclusion and exclusion criteria for studies are shown in Table 1. Conference abstracts and non-English language articles were not included. Opinion pieces were reviewed for references that may meet the inclusion criteria.

Table 1.

Eligibility criteria for studies to be included in the review.

Inclusion	Patients – Critically ill or ICU patients, co-morbid cohorts, paediatric ICU patients Intervention – enteral feeding at different time-based regimes e.g. bolus, intermittent or cyclic. All studies where the feeding regime allowed an overnight fasting period of greater than or equal to 6 h. Comparator – enteral feeding at anytime Outcome – Any
Exclusion	TPN/parenteral feed; post pyloric feeding. Non critically ill patients. We specifically looked for a minimum overnight fasting period of 6 h, which excludes fasting periods of <6 h. Daytime fasting e.g. for religious purposes One off fasting episodes or perioperative fasting Animal studies Conference abstracts, opinion papers Non-English language or no translation to English available papers.

Information sources

The search strategy aimed to identify both published studies and those registered but ongoing. Initial searches were made in OVID Medline and Embase to identify studies and review articles on the topic. This scoping review considered both experimental and quasi-experimental study designs including randomised controlled trials, non-randomised controlled trials, before and after studies and interrupted time-series studies. In addition, analytical observational studies including prospective and retrospective cohort studies, case-control studies and analytical cross-sectional studies were considered for inclusion. This review also considered descriptive observational study designs including case series, individual case reports and descriptive cross-sectional studies for inclusion. Qualitative studies were also considered. Opinion papers were reviewed for references that may meet the inclusion in this scoping review.

Search strategy

Databases

Six databases were systematically searched, being Embase via Ovid, Ovid Medline^®, Web of science, Cinahl, Cochrane and Prospero (NIHR).

Selection of sources of evidence

The text words contained in the titles and abstracts of relevant articles, and the index terms used to describe the articles were used to develop a full search strategy for report the name of the relevant databases/information sources (see Appendix 1). The search strategy, including all identified keywords and index terms, were adapted for each included database and/or information source. The reference lists of all included sources of evidence were screened for additional studies. Only studies published in English were included as there is no translation service available. Studies published since 1970 were included. There are many studies where bolus or intermittent feed is given every 2, 3 or 4 h. These studies were not included as these protocols do not allow a prolonged fasting for example, >6 h period during the night.

Keywords and search terms

The full search strategy is detailed in the Appendix 1. The following terms using Boolean operators were used; ICU patients, critically ill patients, time restricted feeding, fasting, intermittent fasting, night-time fasting, daytime feeding and bolus feeding.

Data charting and process

Following each database search, all identified citations were collated and uploaded into Rayyan citation management system and duplicates. Titles and abstracts were screened against the inclusion criteria, by two independent reviewers and included in the review. Any disagreements in selection were resolved through review of the publications in question and discussion between the two reviewers. The results of the search and study inclusion process are presented in a Preferred Reporting items for Systematic Reviews and Meta-analyses extension for scoping review (PRISMA-ScR) flow diagram (See Figure 1).^14,15

Figure 1.

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flowchart.^14,15

Data extraction and items

Data were extracted from the articles included in the scoping review using the data extraction tool developed by the reviewers. The data fields were author, year of publication, country of origin, study aim or purpose, study design, population and sample size, intervention, control, nighttime fasting period and outcomes relating to the interventions.

Results

Data analysis and presentation

The initial search of six databases produced 918 records and a further 20 were identified from citations or references. 332 records were screened and 66 in total were assessed for eligibility. Fourteen studies published between 1989 and 2023 met the eligibility criteria for the review. Table 2 details the studies included in the review.

Table 2.

Summary of studies included in the review.

Author	Year	Origin	Type	Patients, number and cohort	Intervention	Night fast time	Primary outcome
Saito et al. 28	1989	Japan	Observational comparative study	18, N	Day feed or night feed	12	Serum cortisol
Nishimura et al. 27	1992	Japan	Observational comparative study	18, N	Day feed or night feed	12	Temperature peak
Bonten et al. 19	1996	USA	Unblinded single centre RCT	60, G	18 h feed	6	pH and VAP
Rhoney et al. 20	2002	USA	Retrospective cohort study	152, N	18 h bolus feed	6	Feed intolerance
Tamowicz et al. 22	2007	Poland	Single centre prospective randomised study	40, G	18 h feed	6	pH and VAP
Maurya et al. 21	2011	India	Single centre randomised study	40, N	18 h feed	6	Respiratory quotient
Kadamani et al. 16	2014	Lebanon	Single centre randomised study	30, G	4–6 hourly bolus	4–6	Feed intolerance
Tavares de Aranjo 18	2014	Brazil	Unblinded single centre randomised observational study	41, G	18 h feed	6	Feed intolerance
Nasiri et al. 17	2017	Iran	Triple blinded single centre RCT	40, G	18 h feed cyclic or bolus	6	Feed intolerance
Mahran et al. 23	2019	Egypt	Single centre, prospective parallel RCT	50, G	16 h bolus feed	8	Intraabdominal pressure /Feed intolerance
Zhu et al. 24	2020	China	Single centre RCT	78, G	16 h bolus feed	8	Feed intolerance
Ren et al. 25	2021	China	Non inferiority RCT	62, G	12 h bolus feed	12	Average glucose
Lee et al. 29	2022	Korea	Unblinded, single-centre, parallel-group, RCT	99, G	12–13 h bolus feed	11–12	Nutritional goals
Veldscholte et al. 26	2023	Netherlands	Open label single centre RCT	140, P	12–14 h feed	8–12	Ketosis; non inferiority of calories in

G: patient cohorts are General or mixed; N: neurosurgical, vegetative state or brain injury; P: neonatal and paediatric; RCT: randomised controlled trail.

Patients/populations

A total of 868 critically ill patients in 14 studies were included, from a range of countries including Lebanon, ¹⁶ Iran, ¹⁷ Brazil, ¹⁸ USA (x2),^19,20 India, ²¹ Poland, ²² Egypt, ²³ China (X2),^24,25 Netherlands, ²⁶ Japan (X2)^27,28 and Korea. ²⁹ Summary information for each study is shown in Table 2. Four studies^20,21,27,28 were ICU patients with brain injury alone and 10 were in a mixed ICU population.^{16–19,22–25,29,30} One study was in a paediatric/neonatal ICU with mixed diagnoses. ²⁶

Study designs

Of the 14 studies, 11 were randomised controlled trials, two were observational studies and one was a retrospective cohort study. All were unblinded to the clinicians by nature of the intervention, while blinding during analysis was mixed.

Interventions

The maximum duration of overnight fasting was between 6 and 12 h (Table 3).

Table 3.

Timing of feed delivery in intervention and control groups.

Pattern of feeding	Intervention arm feeding pattern	Control arm feeding pattern
Cyclic feeding e.g. continuous for a period of hours, then turned off for a period of hours	Maurya et al., 21 Rhoney et al., 20 Nishimura et al., 27 Saito et al., 28 Nasiri et al., 17 Tavaras de Araujo, 18 Bonten et al., 19 Tamowics, 22 Veldscholte et al. 26	Nishimura et al. 27 and Saito et al. 28 Cyclic overnight feeding in one control arm, continuous in a third arm
Bolus feeding e.g. bolus delivered every 6 h	Maurya et al., 21 Kadamani et al., 16 Mahran et al., 23 Zhu et al., 24 Ren et al., 25 Lee et al. 29	Maurya et al. 21 control 2 hourly bolus feeds with 8 h nighttime fast
Continuous 24 h feeding		Rhoney et al., 20 Nasiri et al., 17 Tavaras de Araujo, 18 Bonten et al., 19 Tamowics, 22 Veldscholte et al., 26 Kadamani et al., 16 Mahran et al., 23 Zhu et al., 24 Ren et al., 25 Lee et al., 29 Saito et al., 28 Nasiri et al. 17

Method of feed delivery in the Intervention group

The daytime feeding was given as an infusion (cyclic feeding)^{17–22,26–28,30} in nine studies and as boluses^{16,17,21,23–25,29} in the daytime in seven studies (Table 4)

Table 4.

Duration of maximum nighttime fasting, where 6 h is the minimum included in this review.

Duration of night fasting in hours	6	8	12
Number of completed studies	7	2	5 (including a paediatric study)
First author	Kadamani et al., 16 Nasiri et al., 17 Tavares de Araujo et al., 18 Bonten et al., 19 Maurya et al., 21 Tamowicz et al., 22 Rhoney et al. 20	Mahran et al., 23 Zhu et al. 24	Veldscholte et al., 26 Ren et al., 25 Nishimura et al., 27 Saito et al., 28 Lee et al. 29

Comparators

In 11 studies, the comparator was continuous 24 h enteral feeding. One study gave both the intervention and control groups an overnight fast of 8 h, where the intervention was five feeds at 4 h intervals in the day and the control was feed with 2 h intervals in the day. ²³ Two studies had comparators of both nighttime feeding in one group and continuous in another, making these three group studies.^27,28

Outcomes

The primary and secondary outcomes reported in the selected studies are summarised in Table 5.

Table 5.

Studies grouped by outcome that was studied.

Study group	Feed tolerance and bowel function	Nutrition goals achieved	Metabolic outcomes	Gastric pH and pneumonia	Circadian and other outcomes
Studies with this primary outcome	Kadamani et al., 16 Nasiri et al., 17 Rhoney et al., 20 Mahran et al., 23 Zhu et al. 24	Tavares, 18 Veldscholte et al. 26 Lee et al. 29	Maurya et al. 21 (RQ and REE), Veldsccholte, 26 Ren et al. 25 (glucose)	Tamowicz et al. 22	Nishimura et al., 27 Saito et al. 28
Studies with this secondary outcome	Veldscholte et al., 26 Bonten et al., 19 Ren et al., 25 Lee et al. 29	Rhoney et al. 20		Bonten et al. 19	Bonten et al. 19 (ICU mortality)
Overall finding due to daytime only feeding	Unable to conclude	Variable, less delivery if fasted >8 h.	More ketosis	Lower gastric pH	Unable to conclude

Feed intolerance

Feed intolerance was evaluated using gastric residual volumes (GRV), and clinical signs of the occurrence of vomiting, diarrhoea or constipation. There was no unifying definition used across all publications. The GRV used to signify intolerance was not uniform across included studies being 75 ml, ²⁰ 200 ml^16,17, 250 ml ²⁹ or >50% of the enteral feed provided in the last 4 h. ²⁶ Where a primary outcome of feed intolerance was used, three studies^16,17,23 found no significant difference between continuous and daytime feeding. In the two remaining studies with feed intolerance as the primary outcome, more feed intolerance was found in a study that used a GRV of 75 ml ²⁰ which is much lower than the other GVRs and makes these findings less likely to be valid. Significantly more diarrhoea was reported in one study ²⁴ and significantly less constipation in another. ¹⁶ When used as a secondary target, no difference^18,26,29 in feed intolerance was found in three further studies.

Overall, the heterogeneity of the definitions used to define feed intolerance prevents meaningful comparison across the included studies. However, no signal that daytime only feeding increases feed intolerance was detected.

Nutritional targets

There was no unifying definition of what the nutritional for example, caloric, protein or volume targets for critically ill patients were in the included studies. One study found no difference in achieving >60% of the prescribed calorific intake in a group of 41 patients. ¹⁸ Continuously fed patients were significantly more likely to achieve >80% of their target nutrition in a cohort of 99 patients. ²⁹ A feasibility trial of 140 critically ill children showed the nutritional intake was lower in the intermittently fed group than in the continuously fed group for the first 4 days, although this was not associated with an increase in feeding intolerance. ²⁶ The authors consider that feed delivery was principally limited by feed prescription and feed interruptions.

When used as a secondary outcome, a fasted at night group received more calories and protein but took longer to reach their nutritional goals in a study of 152 patients. ²⁰ There was no difference in the calorie intake between groups in a study of 78 patients where nutritional intake was a secondary target. ²⁴ The studies with reduced nutritional input had fasting times of 8–12 h ²⁶ and 11–12 h. ²⁹ The studies with no significant difference in nutritional delivery had fasting times of 6 h^18,20 and 8 h ²⁴ .

The lack of unifying predefined nutritional targets prevents meaningful comparison between included studies. For those when fasting times were greater than 8 h, nutritional delivery was reduced compared to those fed more frequently or continuously.

Metabolic outcomes

In an RCT of 62 patients with a 12 h fast ²⁵ or continuous feeding, the average glucose measurement was not significantly higher in either group, but there were more hyperglycaemia events in the continuously fed group. In the paediatric study of 140 children fasted for 8–12 h, there were no severe hypoglycaemic events and there was no difference in the rate of hypo- or hyperglycaemic events between the groups. In this study, neonates for example, aged <4 weeks, received supplementary glucose infusions at night. ²⁶ In the same study, ketone levels were significantly higher (median 0.4 (0.2; 1.0) vs 0.3 (0.1; 0.7) mmol/L, p < 0.001) in the fasted group. In the only study that measured respiratory quotient, no difference was found between the continuously fed and night time fasted groups. ²¹

Gastric pH and pneumonia

Both studies that measured gastric pH found a significant decrease in pH in the fasted group.^19,22 The incidence of ventilator associated pneumonia (VAP) were similar but there were more bacteria in the stomachs of those with VAP in the continuously fed group. ²²

Circadian rhythm

Two studies, each with 18 patients divided into three groups of six, compared continuous feeding with feeding for 12 h in the day or 12 h at night with daytime only light exposure. Outcomes of temperature ²⁷ and cortisol ²⁸ fluctuations were used, and a positive relationship with the time of feeding was reported. Both studies were small, making the validity of their findings questionable and the methodology makes observer bias possible in the temperature group. These studies are supportive of, but not confirmatory of evidence that the time of providing feed is associated with changes in body temperature and cortisol levels.

Quality of the published studies

Assessment of quality of the trials is beyond the reach of a standard scoping review. It should be noted that trials had populations of between 18 and 152 patients, with the majority being 40 and 60 patients. It is likely that these trials will be prone to both type 1 and type 2 errors.

Publication bias was not formally assessed in this scoping review.

Studies registered in this field, but unpublished

Six studies that meet the inclusion criteria were listed in a registry but remained in progress at the time of the search. Their primary outcomes included gut bacterial diversity, time to reach nutritional goals and nitrogen balance.

Discussion

Amongst the studies included in this review, there are a range of nighttime fasting periods and outcomes, with clear evidence that a nighttime fast increases ketosis and decreases gastric pH when compared to continuous feeding in the critically ill. There is a suggestion that nutritional targets are less likely to be met if the nighttime fasting period is greater than 8 h. The published studies are limited in terms of generalisability due to non-standardised, differing, definitions of clinical endpoints (e.g. feed intolerance) and by small sample sizes.

Studies included in this review

The results of the search yielded studies from 1989 onwards. Clinical practice has changed considerably in this time, and it is likely that studies from a more limited recent time frame would be more applicable to current practice. The relative rarity of the practice of daytime only feeding in the critically ill limits the available data and past study concepts may be revisited in a modern cohort with more robust methodologies. The paucity of evidence so far, gives no data to inform on differences in a fasting period on groups with different illness severity, body mass index, age or stage of recovery. All of these clinical scenarios are of interest, where a period of fasting may hold differing clinical implications.

Feed intolerance

Feed intolerance has been defined by ESPEN ¹¹ as a general term for when a patients is failure to reach enteral nutrition targets in addition to the presence of gastrointestinal symptoms. Many studies include the presence of raised GRV in the definition, but the volume is inconsistent across studies and it is also recognised by more recent authors to correlate poorly with feed intolerance. ³¹ The studies that have determined feed intolerance by using GRV, and with different volumes between studies, are hence difficult to interpret. In addition, using GRV, especially low volume GRV, will be likely to have affected the time limited feeding groups more because higher feed rates are used when feeding is given over a shorter for example, less than 24 h period. The feed rate per hour in these cohorts may be similar to the maximum GVR and if that GVR is considered ‘high’, feeding intolerance may be diagnosed.

Two studies^16,24 indicate intermittent feeding leads to more diarrhoea or less constipation, although a uniform assessment such as the Bristol stool chart has not been used. It is not clear if the difference in constipation is due to effects on the gastrocolic reflex on commencing feed, or differences in the timing of feeding. Future studies will benefit by sharing the same ESPEN definition of feed intolerance, making comparison of their data possible. Without this unified approach the relative effect of daytime feeding on feed tolerance is unknown.

Nutritional targets

Similarly, nutritional targets used across the included studies did not have a unifying definition. This makes comparisons and conclusions difficult and may be improved using ESPEN defined targets. In the studies where nutritional delivery was found to be reduced in the daytime only feeding group, the limitation to delivery imposed by the prescribed volume or quantity to be delivered is a modifiable factor for future work. For patients receiving the intervention of time restricted feeding of any kind, methods of optimising to quantity of feed by increasing the prescribed amount or adding the possibility of catch-up feeding following period when it is paused for procedures, may be considered.

Ketosis and gastric pH

A period of withholding enteral nutrition, or fasting, had two results that were not surprising following a period of fasting. A significantly higher level of ketone bodies ²⁶ and a more acidic gastric pH were demonstrated in the included studies. The increase in gastric acid was not proved to be protective against bacterial colonisation of the respiratory tract.^19,22 This lack of difference may be accounted for by the size of this study being below the number required to find a significant difference in VAP. Further markers of fasting have been measured following a single 12 h cessation of feeding in the critically ill, with associated changes in other markers of fasting. ³⁰ While this study did not meet the inclusion criteria of this scoping review, it is further evidence that fasting is an intervention that will modify fasting markers, which are of interest for future studies which explore potential benefits of fasting as an intervention.

Paediatric patients and nutrition

We included a study from a paediatric cohort in this review. ²⁶ Adult and paediatric patients are dissimilar, notably their physiology in which feeding patterns, nutritional needs, gut physiology, metabolism and energy reserves are very different to adult subjects. ³² The design of this study had important safety measures in place for neonates, where supplemental glucose infusions were added and the detail of the outcomes measures were adequate to assess for potential adverse outcomes as a result of the restrictive feeding intervention. Nutritional delivery was reduced in the intervention group; the authors noted this was within the limitations of what had been prescribed, and that future work can integrate methods of increasing the prescribed feed. Additionally, both groups received sufficient nutritional intake based on international nutritional guidelines. A nighttime fast of 12 h in a paediatric critically ill population was feasible and this study challenges concerns that such a fast might not be feasible in adults, who have more energy reserves. Replication of this feasibility study in an adult critically ill population with clinically significant endpoints may be considered, where the differences in adult gastric residual volumes, metabolism and adult ICU care provision may impact the outcome.

Feed as an influence on circadian rhythm

This scoping review identified two studies that had findings supportive of feeding time being a circadian influencer, but due to methodological features these are not confirmatory.^27,28

Significant studies, excluded from this scoping review

The theoretical benefit for protein metabolism and muscle synthesis are important to the critically ill, who commonly have significant muscle wasting during their acute illness and this in turn limits their rehabilitation and recovery. Fasting is therefore an attractive therapeutic intervention and outcomes relating to muscle mass and protein synthesis have been targeted by two studies.^33,34 Significantly, the fasting period of 4 h used in these studies excludes them from this review, leaving the question of what effects nighttime fasting has on muscle metabolism in the critically ill unanswered.

Strength and limitations

This scoping review encompassed a wide range of studies. In addition to using studies identified using several data base searches, it considered studies that were identified from the references of reviews. However, it has several limitations. The first is the design of the review, where a scoping review does not require a critical assessment of the included studies’ quality (e.g. sampling bias and statistical analysis). Secondly, the results of this review cannot describe the practice of daytime only feeding globally because the studies were from a limited range of countries. Thirdly, this review does not include the grey literature describing beyond that found via citations in the practice of daytime only feeding and related factors. Finally, we only included documents in English, which could eliminate the opportunity to explore findings from articles in languages other than English. Limitations relating to the available published work are study size, variability in the definitions used in classifying clinical outcomes, the use of different sub-populations of critically ill patients, the variation in the duration of the fasting period and the variation in bolus or cyclic feeding patterns in both the intervention and control arms.

Implications for research

The proof of concept of daytime only feeding is provided by the range of studies in this review. Adequately powered studies with clinically important outcomes are awaited. Key research questions designed with internationally recognised definitions of clinical outcomes should be addressed for example, feed intolerance, nutritional delivery and other clinically important outcomes. Key outcomes of mortality and significant causes of morbidity in the critically ill should be measured in the context of an adequately powered trial of daytime only feeding. The two circadian studies which were supportive of the time that feed is delivered being associated with changes in body temperature and serum cortisol were in small cohorts and used crude circadian outcomes. Larger, methodologically sound trials with appropriate circadian outcomes are indicated in the critically ill to investigate this.

Implications for clinical practice

The evidence from this review is inadequate to inform clinical practice. However, no clear signal of harm was identified in the presented studies, allowing continuation of the use of varied feeding patterns if this is locally established and found to be acceptable. A change from continuous to intermittent feeding in the critically ill should be informed by the methods that have been described in the studies identified in this scoping review. Priorities are providing adequate nutritional delivery and monitoring for feeding intolerance, while adapting local protocols based on clinical feedback. The optimal duration of fasting, where physiological benefits are balanced against limitations of providing the essentials of nutrition, is yet to be established.

Conclusion

Providing nutrition to critically ill patients in the daytime only as opposed to continuously has been studied in 14 relatively small trials. It has been shown to significantly increases serum ketones and decrease gastric pH. There are no adequately powered randomised trials of daytime only feeding compared to continuous feeding in critically ill adults to fully inform us with greater certainty of outcomes relating to this intervention.

Appendix 1

Search strategy

Ovid MEDLINE(R) and In-Process, In-Data-Review & Other Non-Indexed Citations <1946 to August 14, 2023>

• 1 “Intensive care patients”.mp.   3961

• 2 “critical illness”.mp. or exp Critical Illness/   45395

• 3 1 or 2   48712

• 4 “time restricted feeding”.mp. or Intermittent Fasting/   580

• 5 “bolus feeding”.mp.   131

• 6 “fasting”.mp. or exp Intermittent Fasting/or exp Fasting/   141492

• 7 “daytime feeding”.mp.   80

• 8 “night time fasting”.mp.   3

• 9 4 or 5 or 6 or 7 or 8   141858

• 10 3 and 9   211

COCHRANE = 100

Search Name:

Date Run: 15/08/2023 08:56:43

Comment:

ID Search Hits

• #1 “Critically ill patients”   6112

• #2 MeSH descriptor: [Critical Care] explode all trees   2676

• #3 “intensive care patients”   840

• #4 MeSH descriptor: [] explode all trees   0

• #5 “time restricted feeding”   154

• #6 “bolus feeding”   60

• #7 “daytime feeding”   1

• #8 “nighttime fasting”   6

• #9 “intermittent feeding”   71

• #10 “fasting”   41890

• #11 #1 OR #2 OR #3   8847

• #12 #5 OR #6 OR #7 OR #8 OR #9 OR #10   42053

• #13 #11 AND #12   100