Insulin in Frail, Older People with Type 2 Diabetes—Low Threshold for Therapy

Round 1

Reviewer 1 Report

Diabetology-1718903-review report

I thank the journal for giving me an opportunity to review this article titled “Insulin in frail older people with type 2 diabetes – Low threshold for therapy” by Abdelhafiz et al. Overall, this review is relevant, and highlight important aspects of diabetes management and care for the frail elderly (focus on metabolic frail phenotypes, and the insulin + insulin analogue safety in managing these categories). However, despite some new information, there are major concerns that the authors are requested to correct/revise to redetermine:

• Several portions within the submitted copy is largely a spin-off from previous published review works by the first author (https://pubmed.ncbi.nlm.nih.gov/34151494/) with a very similar work also recently published in 2022 (https://pubmed.ncbi.nlm.nih.gov/35209779/). Sections of the introduction, methods, other subsections and some of the end-sections within the submitted paper can be deemed as “self-plagiarized” despite few reworded/restructured sentences. This is unacceptable and requires a completely new face-lift of the existing draft.
• The excessive details from previous published works, including the data presentation/writing style mar some of the key points the authors make. For example, wouldn’t the authors agree that sections on the use of insulin analogues and information in Figure 2 are vital, new information for readers but are majorly sandwiched between the overloaded details of clinical frailty and what is already known, published information? Why have the authors left it to the readers to painfully tease out all information? Suggestion: After rewriting a new, brief, and well-connected introduction, why not bring the focus of the reader directly to what the title indicates, and immediately start this review from line 208?
• Please do not include any speculative statements suggestive of clinical therapy that lacks empirical inferences or evidence from one’s own study to strongly back it. Inferring such statements via mere review analyses devoid of any scientific credibility is extremely risky to present as an empirically tested fact. For example, can the first author demonstrate any of his published clinical research study (pilot study/a clinical audit/multi-center, double blind cross-sectional studies on AM-SO metabolic phenotypes and frailty) to disprove what I would consider a purely speculative statement in lines 57-59 (Therefore, we believe…). While I am an ardent follower of some of the senior authors in the submitted paper, it is rather demoralizing when clinical reviews tend to pass off clinical approaches/course of action based on a mere review (a similar pattern observed in a different paper - https://pubmed.ncbi.nlm.nih.gov/34151494/) with absolutely no regard for evidence-based knowledge from one’s own clinical scientific investigations.
• Figure 1 PRISMA diagram is meaningless, this paper is neither a meta-analysis nor a clinical trial to justify methodologies with a flowchart. A thorough literature search on various search engines resulted in several different hits but the most relevant were just 5/6 articles which the authors tabulate for this review paper. Why is a flowchart justification needed for that? Also, the figure is of poor quality, formatting issues when in a .pdf file perhaps? Many of the words cannot be seen and the font is rather small.
• A clear definition of “old age” within the realms of diabetes-associated frailty and teasing out accurate information from the associated data is required. Frailty is indeed a heterogenous concept and increases to about 40% in patients who are greater than 80 years old, yet most of the inferences on clinical frailty for the current paper are drawn from studies where the mean age is less than 65 years. The authors may consider stratifying the data in terms of age and then focusing on specific studies (as is a known fact within the field that the information is absolutely sparce for the elderly with diabetes). The authors may like to consider the following papers:

https://pubmed.ncbi.nlm.nih.gov/19197079/: mean age 74.43 at baseline

https://pubmed.ncbi.nlm.nih.gov/22732903/: stratified (65–74 and ≥ 75 years old)

https://pubmed.ncbi.nlm.nih.gov/20536960/: ≥ 70 years old

• The authors may rephrase that the review will focus only on two end spectrums of the metabolic profile of frailty and the associated diabetes management (intermediary or other categories of the spectrum are not the focus of this review?). This also begs the question – have the authors probed in-depth (at all) into testing their theory on the efficacy of presence of these two metabolic frailty phenotypes? Is there any clinical research data to back their hypotheses that such subcategorization is better suited for a tailored therapy and greater chance at achieving precision medicine to manage T2D for the frail elderly?
• Line 19 – define “oldest old”? Not only in the abstract but throughout the manuscript.
• Details and management of the AM phenotype are well-described but not so for the SO phenotype? Please revise to include the same for the SO phenotype. The same is observed within the text in later part of the review. Why have authors not consolidated information on the SO phenotype and/or present some figure as how it was shown for AM?
• Line 27 - change “insulin NPH” to intermediate-acting insulin (if that’s what the authors wish to compare the insulin analogues against, also in lines 273-274)
• Introduction: Lines 63-66: any recommendations for the SO phenotype as the authors cover these two end spectrums of frailty? In addition, the authors need to specify the age range for the “older people with diabetes” (line 64).  
• Methodology

1. The inclusion criteria (line 79), for long-acting insulin analogues is ≥60 years, but how does this justify the concept of frailty and metabolic heterogeneity, more often progressively evident in the age group of >70-80 years? How were the authors able to deduce meaningful information for diabetes-induced frailty from studies presented in table 1 and 2 where the mean age is not even 65 years for many studies.
2. Table 1: First study by Tanaka et al., - would the authors not consider this study as biased as it focuses only on male participants when compared to the other listed studies?

Minor: 1) some language tweaks needed – such as, phrases in line 24 “emerging little evidence” is not very clear (either emerging or little evidence perhaps?).

Author Response

Insulin in frail older people with type 2 diabetes – Low threshold for therapy

Reviewer 1:

Many thanks for your comments and suggestions to improve this manuscript.

1. Repetition of previous work and suggestion to start this review from line 208.

We agree with your comments.  The introductory section seems to be too long and there is significant overlap with our previous work.  Therefore, we have concentrated the abstract, introduction, deleted the redundant parts and started, as suggested, from line 208, page 1-8.

2. Line 57, we believe…No evidence to support our claims.

We fully agree with your observation, there is no evidence in literature to support our views on the existence of metabolic differences across the spectrum of frail older people with diabetes.  It is largely speculative, as you have mentioned.  We have changed the word believe to the word suggest and also added a short paragraph in the future perspective section as a limitation of this review and highlighted that future research in this area is required, page 3&19.

3. Figure 1.

Yes, it is not meant to be a systematic review.  The figure was just for clarification of literature search but we have deleted it to avoid confusion. 

4. Frailty stratification by age.

We appreciate your point, but because we have deleted the introductory section of the manuscript and started from line 208, as suggested, the discussion of this part is not now required. 

5. The review focused only on the two ends of the metabolic spectrum of frailty.

Yes, we agree the manuscript focuses on the two ends of the spectrum, in fact, it focuses only on insulin therapy in the anorexic malnourished metabolic phenotype (one end of spectrum), highlighted in the introduction, page 3  We have mentioned the paucity of research, supporting this metabolic classification, as a limitation of the review in the future perspective section, page 19.

6. Oldest old.

Agree, we found this term confusing and not useful. We have deleted it from the abstract and across the manuscript. 

7. The SO phenotype.

We appreciate this, but we aimed in this review to suggest that insulin can be considered as an early option in those malnourished frail patients, we called AM phenotype.  This was the main purpose of the manuscript.  We did not expand beyond this point to keep the manuscript focused and concentrated on this objective.  There was no intention to write a comprehensive review and cover all hypoglycaemic therapy across the whole spectrum of frailty.  As you have mentioned, there is no clear evidence yet in the literature to suggest solid recommendations in this area.  There is however, some cover on the SO phenotype in the text as needed and also Figure 2 illustrates some suggestions on the treatment of this phenotype and insulin therapy remains (as it is in current clinical practice) a last resort after oral therapy in this phenotype, page 15&Figure 2.

8. Insulin NPH” to intermediate-acting insulin.

Changed as suggested.

9. SO introduction, specify age.

Please see the same point above.  We defined older people as aged ³60 years (added to the introduction) and this age limit was used in literature search. 

10. Age limit to >³60 years.

A valid point, but due to paucity of studies in literature in higher age groups.  We have included the available studies and we have to go down to the age of 60.  We have added this as a limitation in the future perspective section.  In addition, previous studies on the safety of long-acting insulin analogues were conducted on people aged <60 years and did not show benefits compared to intermediate acting insulins, but the benefits starts to emerge in studies conducted in those >60 years.  Although, is a relatively young age, but it shows some differences, page 11.

11. Tanaka et al study on only men.

Thanks, mentioned as a limitation of this study, page 9.

12. Emerging little evidence and other language issues.

Changed to emerging evidence and manuscript reviewed throughout, page 1.

Reviewer 2 Report

It is important to discuss how elderly persons with diabetes should be treated in the best and safest way so the purpose of the review is sound. However it needs major revision. Below are some examples:

There are a lot of repetitions for example line 36-37 “ This is largely due to the increase in life expectancy and the increasing ageing of the population” It is enough to write “This is largely due to the increasing ageing of the population” Line 40-41 again the authors write about increasing number of older people this sentence is unnecessary (As a result of increasing….). The same word is used multiple times in the same sentence  (line 50 hypoglycaemic, line 54 frailty, line 130-131 accelerated, line 135-136 promote etc).

Example of some mistakes: Line 38 Age of 65 years old , line 58  the long-acting insulin analogues (there are several).

The authors are not correctly describing insulin secretion and effects of insulin. One example:  They write (line 192-193) Insulin is normally secreted after ingestion of glucose..... . This is not correct there is basal insulinsecretion . It is also obvious that they do not have a clear picture of insulin resistance, insulin deficiency, effects in target organs.

Treatment options are not up-to-date. They write about  acarbose which is hardly ever used, but they don´t mention DPP4 inhibitors which according to guidelines can be used in fragile patients.They are not clarifying that some longacting insulin analogues should NEVER be given more than once per day and this are the ones not having a peak.

HbA1c is not the important tool to use when you evaluate treatment in this population .    

Author Response

Reviewer 2:

Many thanks for your comments and suggestions to improve this manuscript.

1. Word repetitions and language mistakes.

The introductory section has been shortened and language reviewed throughout the manuscript, page 1-8.

2. Insulin physiology.

Thank you for your observation, but we have removed this section as advised by the first reviewer.

3. a. DPP4 inhibitors.

We agree, we have mentioned this, highlighted in page 15 and Figure 2.

3. b. Acarabose.

Fully agree, it is an old agent and hardly used!  Reason to mention in text is just to demonstrate an example of some oral medications that are not suitable and to make the case for early introduction of insulin therapy in the AM phenotype.  However, it can be used, if tolerated, in the SO phenotype due to its weight loss properties.

3. c. Peakless insulin should be used once daily.

We have added this, page 16.

4. HbA1c not the focus of treatment.

We agree, we have added this in the goals of therapy section, page 17.

Round 2

Reviewer 1 Report

Minor corrections needed before acceptance. 

Comments for author File: Comments.pdf

Author Response

Many thanks for your detailed review and suggestions which have improved the manuscript significantly.  We have corrected the points you have raised.  In tables, we have clarified the status of type 2 diabetes in all the studies but we were not able to include all associated comorbidities due to space limitations.