Association between Hospital Prescribed Diets and Length of Stay, Re-Presentation, and Gastrointestinal Symptoms among Acute Uncomplicated Diverticulitis Patients: A Prospective Cohort Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Sir/Madam,

Thanks for inviting me to review the manuscript above. I give the authors opportunity to address the lack of explanation/discussion of the biological mechanism behind the apparent success of liberalized diet vs restricted diets. It would be interesting to see if subgroup analysis-based on the type of restricted diets would show anything different or would show for example that one of them is better or worse than the others. This finding might shed some light on which dietary factor/s missing from the specific prescribed diet or (what is there) that can be a contributing factor to their inferiority over the liberalized diets, directly or through gut microbial effects. This discussion can improve the manuscript and helps future studies.

Overall, the study limitations are very well recognized and no under- or overestimation of the findings exist. Indeed, this area requires further thorough research to ascertain the role of diets and dietary factors in diverticulitis.

Author Response

Comment 1: Thanks for inviting me to review the manuscript above. I give the authors opportunity to address the lack of explanation/discussion of the biological mechanism behind the apparent success of liberalized diet vs restricted diets.

Response: Thank you to the reviewer for the time and consideration given to this manuscript.

Regarding length of stay: The authors do not propose that a biological mechanism is responsible for a lower length of stay associated with the liberalised diet. Rather, we propose that the difference in length of stay between the two groups is due to health-service related factors. We have added to the discussion on the proposed mechanism:

“A plausible explanation for this is that diet tolerance is often a criterion for discharge; therefore, the explanation for the lower LOS associated with the liberalised diet may not be biological but rather a systems concern”[43]. This has implications for healthcare sustainability; if an intervention is contributing to increased LOS, it is important that the intervention improves outcomes for patients. In the case of diet restriction, there is no evidence of improved outcomes with this approach in patients with uncomplicated diverticulitis where oral food is tolerated.” (Page 7, paragraph 2)

Regarding re-presentation: There may be biological mechanisms of action resulting in the decreased risk of re-presentation in the liberalised diet group. We have expanded upon this in the discussion:

“Diverticulitis-related re-presentations were significantly associated with dietary re-striction in index admission, as well as an underweight BMI. There is no recognised mechanism of action in which a short-term dietary restriction during hospitalisation may directly influence the risk of diverticulitis recurrence at 6-months. However, the restriction may have indirectly increased the risk of diverticulitis-related representation. Imposed dietary restrictions in the general hospital population have led to distress, discomfort, and worsened diet quality [45]. In those with gastrointestinal disease, hospital-imposed dietary restrictions have been associated with a lasting fear of food [46-48].  A recent observational study found patients with a previous history of diverticulitis had significantly lower intake of calories, fibre, dietary vitamins A, C, D and E, and Oxygen Radical Absorbance Capacity index, compared to healthy controls [49]. Restricted diets are inadequate to meet patients’ nutritional requirements [47], contributing to the development or worsening of malnutrition [14] and may be a mechanism by which restricted diets were observed to have higher rates of readmission in this study. Malnutrition and inadequate dietary intake are risk factors for morbidity, mortality, and readmissions in many acute and chronic disease states [16]. However, as the current study did not assess patients for malnutrition or pre-existing unintentional weight loss, which is associated with significantly higher odds for diverticulitis readmission, [45] future research is required to explore the relationship between nutrition status, dietary restriction, and recurrence of diverticulitis..” (Page 8, paragraph 4)

Comment 2: It would be interesting to see if subgroup analysis-based on the type of restricted diets would show anything different or would show for example that one of them is better or worse than the others. This finding might shed some light on which dietary factor/s missing from the specific prescribed diet or (what is there) that can be a contributing factor to their inferiority over the liberalized diets, directly or through gut microbial effects. This discussion can improve the manuscript and helps future studies.

Response: This would indeed be very interesting and a worthwhile research question. Unfortunately, the data available in this cohort study would not support such an analysis.

Patients who received hospital-induced dietary restrictions received many changes to their diet codes within a few days, e.g., A patient would commence on ‘nil by mouth’ and progress to ‘clear fluids’ later that day, to then receive full fluids the next morning then perhaps go back to clear fluids that evening. Over the next two days, they might then be prescribed full fluids and then a low fibre diet.

This means that patients could not clearly be allocated to a specific type of restricted diet, rather only grouped as receiving a ‘restricted diet’, which is an umbrella term for many types of restricted diet codes.

We note that although we have described this practice of diet code progression in the methods (page 4, paragraph 2), we have not presented the results. Therefore, we have added the number of diet codes each patient was allocated during the admission to table 1 and briefly mentioned in text:

“Patients were observed to be prescribed from one to six diet codes during their admission, with the liberalised group having significantly fewer diet code changes throughout their admission (<.001; Table 1). The most common diet code progressions were clear fluids to low fibre (n=26 patients) followed by clear fluids to free fluids to low fibre (n=22 patients)..” (Page 5, paragraph 3)

In addition, we have added to the discussion the need for further research to elucidate dietary mechanisms producing the observed effects:

“Although this precluded the ability to control confounding factors, this study now provides sufficient justification and evidence to inform a future randomised controlled trial which will assist in identifying the diet-related mechanisms of action which produce the observed effects.” (Page 9, paragraph 2).

Comment 3: Overall, the study limitations are very well recognized and no under- or overestimation of the findings exist. Indeed, this area requires further thorough research to ascertain the role of diets and dietary factors in diverticulitis.

Response: Thank you for your endorsement of this topic area for future research.

Reviewer 2 Report

Comments and Suggestions for Authors

The liberalized diet is increasingly being used for uncomplicated diverticulitis. Larger controlled studies are lacking. This study aims to make a contribution to this.

First of all, the question that the authors need to ask themselves: It is unclear why the patient group was hospitalized . Uncomplicated diverticulitis is usually treated on an outpatient basis. The authors should comment on why their patients were hospitalized.

Acute uncomplicated diverticulitis" is defined as a limited local inflammatory reaction without evidence of a covered/open perforation. Acute complicated diverticulitis" is characterized by a covered or open perforation with the formation of abscesses or peritonitis. Imaging with CT was performed, a more precise specification of the diagnosis made with CT according to CDD is missing, especially since the CRP values in the groups differed significantly. The CRP level tended to correlate with complicated/perforated courses. Values > 50 mg/l reflect diverticulitis,

while a CRP > 200 mg/l raises suspicion of a perforation. This should be supplemented.

How was (symptomatic) uncomplicated diverticular disease (SUDD)

Or segmental colitis associated with diverticulosis (SCAD)?

The probability of diagnosing an adenoma or carcinoma during a colonoscopy after diverticulitis has occurred is significantly increased. Three large registry studies (Sweden, Taiwan, Denmark) and a meta-analysis are available on this question,

which unanimously show that diverticulitis is associated with an

associated with an increased risk of being diagnosed with colorectal carcinoma. The Lit is worth mentioning.

Apparently, the disease activity (CRP) was lower in the group of patients with a liberalized diet, so that the length of stay was naturally also lower. Was there a correlation here?

Author Response

Comment 1: The liberalized diet is increasingly being used for uncomplicated diverticulitis. Larger controlled studies are lacking. This study aims to make a contribution to this. First of all, the question that the authors need to ask themselves: It is unclear why the patient group was hospitalized. Uncomplicated diverticulitis is usually treated on an outpatient basis. The authors should comment on why their patients were hospitalized.

Response: Thank you for your comment; this raises a very important point which had not been sufficiently addressed in the study.

Unfortunately, as an observational study, the authors cannot comment upon the reasons for hospitalisation beyond that observed from medical record (i.e., admitted for management of acute uncomplicated diverticulitis). We do acknowledge that many patients with uncomplicated diverticulitis are outpatients; however, all patients in the study were deemed to require inpatient management by their treating surgeon for reasons unknown to the research team.

It is noteworthy that recruitment gradually reduced over the two years of the project which may have reflected a change in practice of surgeons at the study site to instead favour outpatient treatment where possible over time.

Further, although many health services now utilise a predominately outpatient model, data still supports an increasing number of admissions in the USA for acute, uncomplicated diverticulitis.

We have made the following changes to address this in the manuscript:

“Although preferably managed as outpatients, the rate of adults requiring acute care for uncomplicated diverticulitis is increasing globally” (Page 2, paragraph 1).

“Participants were recruited from January 2017 to March 2019 and outcome data collected until October 2019. The number of eligible participants decreased over the recruitment period which may have reflected a change towards outpatient rather than inpatient care.” (Page 3, paragraph 4)

“Results can also not be generalised to adults managed as outpatients rather than inpatients”. (Page 9, paragraph 2)

Comment 2: Acute uncomplicated diverticulitis" is defined as a limited local inflammatory reaction without evidence of a covered/open perforation. Acute complicated diverticulitis" is characterized by a covered or open perforation with the formation of abscesses or peritonitis. Imaging with CT was performed, a more precise specification of the diagnosis made with CT according to CDD is missing, especially since the CRP values in the groups differed significantly. The CRP level tended to correlate with complicated/perforated courses. Values > 50 mg/l reflect diverticulitis, while a CRP > 200 mg/l raises suspicion of a perforation. This should be supplemented.

Response: Unfortunately, the surgeons and practitioners at the observed site did not use the Modified Hinchey, or any other, classification system. The researchers were limited to the diagnoses and data available in the medical records: Diverticulitis, with complicated versus uncomplicated cases categorised based on the CT report. Patients were considered complicated if the CT report noted: perforation, localised abscess >5cm, or distant abscess. It is not within the scope of the researchers to make a precise specification of the diagnosis. We have updated the methods for readers to reflect this:

“Patients were excluded if they had complicated diverticulitis based on CT report (CT identified perforation, localised abscess >5cm, or distant abscess), required percutaneous drainage of abscess, laparoscopic lavage, or surgical resection during the index admission; were discharged within 48-hours of presentation; were pregnant; or unable to pro-vide informed consent.” (Page 3, paragraph 5)

Although CRP correlates with disease severity in the literature and in practice, and the two groups had different CRP levels, CRP was not associated with any outcome in this study. This has been replicated recently in a second cohort study (ref 54 as per below). We have added to the discussion and reporting on this:

“Patients on restricted diets had lower body weight, lower BMI, and higher CRP than those on liberalised diets; however, differences in BMI and CRP may not be clinically relevant. CRP was not associated with any outcome and therefore not included as a confounding variable in multivariable models.” (Page 6, paragraph 3)

“The significant difference between dietary restriction and baseline CRP may re-present a tendency to restrict diet in patients deemed more likely to experience complications, possibly in preparation for potential deterioration requiring surgical intervention. Despite the difference in baseline CRP between groups, CRP was not associated with outcomes suggesting that disease severity did not impact the difference in LOS nor rates of recurrence between groups. Supporting this, a recent cohort study also found that CRP was not predictive of recurrence in patients with acute, uncomplicated diverticulitis [54]. These findings need to be replicated in studies that can provide a more precise diverticulitis phenotype based on severity..” (Page 9, paragraph 1)

“Readers should not infer causation due to the observational nature of the study design and limited control and availability of confounding variables such as underlying diagnosis and disease classification.” (Page 9, paragraph 2)

Comment 3: How was (symptomatic) uncomplicated diverticular disease (SUDD) Or segmental colitis associated with diverticulosis (SCAD)?

Response: The underlying diagnosis for patients in this study was not available in medical records, and therefore patients could not be described as having SUDD or SCAD. The only data available were whether the patient had previously been diagnosed with diverticular disease (of any kind) and if they had experienced a previous hospital admission for diverticulitis. To clarify for readers, we have added to the manuscript:

“Age, sex, ethnicity, body mass index (BMI), medical and surgical history, smoking status, habitual alcohol intake, and previous episodes of diverticulitis were recorded at baseline (within 48 hours of admission) from the medical record and/or patient interview. No data were available to determine the type of underlying diverticular disease.” (Page 3, paragraph 6)

“^d Known diverticular disease of any classification.” (Footnote on Table 1)

“Readers should not infer causation due to the observational nature of the study design and limited control and availability of confounding variables such as underlying diagnosis and disease classification.” (Page 9, paragraph 2)

Comment 4: The probability of diagnosing an adenoma or carcinoma during a colonoscopy after diverticulitis has occurred is significantly increased. Three large registry studies (Sweden, Taiwan, Denmark) and a meta-analysis are available on this question, which unanimously show that diverticulitis is associated with an associated with an increased risk of being diagnosed with colorectal carcinoma. The Lit is worth mentioning.

Response: While not the focus of our research it is certainly noteworthy. We have added the following to the introduction:

“Malnutrition is common in hospitals, affecting 20–50% of patients and resulting in poor outcomes for both patients and health services [15,16]. Rates of malnutrition were under-estimated in patients admitted with uncomplicated diverticulitis, with a recent study identifying >60% of patients had severely impaired nutritional status and significantly greater length of stay (LOS) than their well-nourished counterparts [17]. Optimal nutritional status may also be a priority for patients with diverticulitis due to an association between acute diverticulitis and increased likelihood of being diagnosed with colorectal cancer [18,19].”

Comment 5: Apparently, the disease activity (CRP) was lower in the group of patients with a liberalized diet, so that the length of stay was naturally also lower. Was there a correlation here?

Response: There was no association with CRP with any outcome. We have clarified this for readers, please see response to comment 2 above.

Reviewer 3 Report

Comments and Suggestions for Authors

Very interesting and important topic, with potential to change common medical practice of treatment of very common disease - acute diverticulitis. This paper might iniciate more research on "normal diet" in this patients with along with already generally accepted "no antibitics needed" might suggest that patients with mild acute diverticulitis dont really need any treatment :)

Author Response

Response: Thank you for reviewing the paper and highlighting the importance of this research to inform evidence-based practice, with the potential to improve patient outcomes by avoiding detrimental and costly interventions.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

My questions were fully answered.