Management of Diabetes and Hyperglycemia in the Hospital: A Systematic Review of Clinical Practice Guidelines Free

The prevalence of diabetes is rising in the U.S., with 15.8% of adults meeting diagnostic criteria in 2023 (1). Older adults make up a large proportion of people with diabetes, with a prevalence of 17.8% among those 40–59 years old and 27.3% among adults over the age of 60 years (1). Diabetes is particularly common among hospitalized patients, with more than 8 million hospitalizations in 2018 involving diabetes management, accounting for approximately one-third of hospitalizations (2). Furthermore, many patients without diagnosed diabetes may experience elevated blood glucose during hospitalization, related to either a new diabetes diagnosis or transient factors associated with acute illness. The importance of safely managing diabetes in the hospital has been demonstrated, with an increased risk of harm for a patient hospitalized with comorbid diabetes (3).

Clinical practice guidelines seek to synthesize best practices based on clinical evidence and expert opinion. To inform these, there is a body of evidence from examination of management of impaired glucose metabolism in hospitalized patients, including both observational studies and randomized clinical trials. Trials have been focused on comparison of blood glucose targets and treatment strategies to identify best practices for balancing and minimizing the risks of inpatient hyperglycemia and hypoglycemia. While informative and imperative, these studies have largely been focused on critical care populations (4) and some components of inpatient diabetes care are not accounted for, including what to do with oral diabetes medications during hospitalization and on discharge. Much of inpatient diabetes care is based on expert consensus and may differ across guidelines. Inconsistent guidance may contribute to variable inpatient practice patterns, which are commonly documented (5,6).

Thus, to identify areas of consensus, areas of disagreement, and gaps in guidance for inpatient diabetes care, we systematically reviewed inpatient diabetes management guidelines and synthesized recommendations on inpatient blood glucose monitoring, goals, and management. We also identified recommendations on hyperglycemia management in vulnerable populations such as older adults, across common inpatient clinical scenarios, and for transitions of care.

Steps for this systematic review included 1) relevant guideline search, 2) exclusion screening, 3) quality assessment, and 4) guideline synthesis. Details of the protocol for this systematic review were registered on International prospective register of systematic reviews (PROSPERO) (no. CRD42023449250 [https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023449250]).

To identify relevant guidelines, we searched MEDLINE PubMed using Medical Subject Headings (MeSH) terms of “diabetes” or “hyperglycemia” and “management” or “therapy” and publication type of “guideline.” The same PubMed search was done with the addition of the publication type of “systematic review” and limiting the journal search to highly cited general medicine and endocrinology journals: Annals of Internal Medicine, The BMJ, Diabetes Care, Diabetologia, Endocrine Reviews, Endocrinology, JAMA, JAMA Internal Medicine, The Lancet, The Lancet Diabetes & Endocrinology, and New England Journal of Medicine. Additionally, we searched the Guidelines International Network using the search terms “diabetes” and “guidelines,” in addition to pertinent domestic and international specialty society websites. Specialty society websites were searched where pertinent and according to affiliation with the American Medical Association (7) and consultation with experts. The search was initially conducted on 1 August 2023 and was updated on 14 August 2024.

We defined guidelines as statements that include recommendations intended to optimize patient care that are informed by a systematic review of evidence and an assessment of the benefits and harms of alternative care options (8). To be eligible, guidelines had to be published in English and include inpatient care recommendations for diabetes (Supplementary Fig. 1). Guidelines that were entirely derived from an existing guideline or were published prior to 2010 were deemed ineligible. The most recent version was included for guidelines with multiple versions. Multiple records from an organization for differing aspects of diabetes management that fell within inclusion criteria were grouped together as one guideline. As we were interested in the management of diabetes in hospitalized adults, we excluded guidelines specific to children or pregnant persons. We also excluded guidelines specific to diabetes emergencies such as diabetic ketoacidosis or hyperosmolar hyperglycemic state, which have previously been reviewed (9). These criteria were applied to each record by two blinded reviewers (L.M.W. and T.S.A.), and any differences in screening were reconciled by consensus.

To assess the quality of the guidelines selected for inclusion, we used the Appraisal of Guidelines for REsearch & Evaluation II (AGREE II) instrument (10–12). This instrument is comprised of 23 items falling into six domains: Scope and Purpose, Stakeholder Involvement, Rigour of Development, Clarity of Presentation, Applicability, and Editorial Independence. Each item is scored from 1 (strongly disagree) to 7 (strongly agree). Each guideline was assessed by three reviewers (L.M.W., B.X.W., and E.H.-K.). We calculated scaled scores for each domain by summing the domain’s item scores, subtracting the minimum possible score, and scaling the result as a percentage of the maximum minus the minimum possible score. Overall guideline quality was also scored from 1 to 7 and averaged, and whether the reviewer would recommend the guideline for use was indicated. We classified a guideline as of high quality on the basis of a threshold of 60% for the Rigour of Development domain scaled score, consistent with prior studies (13,14). Records from the same organization with similar methodologies were scored together, but if the methodologies differed, they were scored separately.

Two reviewers independently extracted guideline recommendations related to blood glucose management in the hospital and the circumstances that result in different recommendations—specifically dietary intake, age, clinical setting (e.g., perioperative, critical care), and diabetes technology. The data extracted for these topics included target blood glucose levels, glucose monitoring, and hyper- and hypoglycemia treatment for different populations. Recommendations on preparation for a safe transition to a posthospitalization setting were also extracted, specifically on discharge education, the adjustment of home antihyperglycemic medication regimens, provisions, and outpatient follow-up. When guidelines did not include inpatient recommendations on individual topics, we also examined outpatient guidelines by the same professional society, as these recommendations could be used to supplement inpatient guidance.

Ten guidelines met study inclusion criteria. We screened 348 unique records, of which 17 from 10 organizations, were included after full-text review (15–31) (Supplementary Fig. 1 and Table 1). Three organizations had multiple records on separate topics related to diabetes care; e.g., the International Diabetes Foundation had one guideline on general type 2 diabetes management and another on type 2 diabetes management specifically in older adults (23,24). U.S. organizations developed six of the guidelines, and Australia, Canada, Great Britain, and a multinational organization each developed one. Four guidelines were general, with sections on inpatient care, while the other six guidelines were specific to inpatient diabetes care, one of which was specific to intensive intravenous insulin infusion therapy (26). Three guidelines focused on older adults (22,23,29).

Five of the guidelines were rated as high quality, and overall quality scores ranged from 3.3 to 6.0 of 7.0 according to the AGREE II instrument (Supplementary Table 1). Individual domain scaled score ranges were as follows: scope and purpose 78%–100%, stakeholder involvement 20%–98%, rigor of development 17%–93%, clarity of presentation 54%–98%, applicability 38%–88%, and editorial independence 0%–97%. The American Diabetes Association, Diabetes Canada, and Endocrine Society guidelines were rated to be of the highest overall quality with a score of 6.0.

Eight guidelines were in part based on systematic reviews, while for the other four the methodology used to obtain evidence was not clearly stated. Despite these methodological limitations, data from all guidelines were extracted to capture real-world practice recommendations.

There was general consensus on which patients require blood glucose monitoring, how to monitor, and insulin treatment regimens across the 10 guidelines analyzed, while recommendations on blood glucose targets, management specific to inpatient older adults, and transitions of care were of variable presence and content.

All 10 guidelines recommended monitoring inpatient blood glucose for patients with known diabetes, and 9 guidelines also recommended monitoring for those with admission hyperglycemia (Table 2). Of these, six defined hyperglycemia according to a random admission blood glucose of >140 mg/dL. Another six guidelines additionally recommended initiating blood glucose monitoring if the patient is on glucocorticoids. There was agreement on an upper limit blood glucose goal of 180 mg/dL for the non–critically ill inpatient, while five guidelines recommended a lower limit of 100 mg/dL and three recommended a lower limit of 140 mg/dL. Recommended frequency of blood glucose monitoring was variable, with monitoring recommended at meals by five guidelines, with the addition of bedtime checks in three of those guidelines. Individualization of monitoring frequency was suggested, including three guidelines discussing the use of continuous glucose monitoring (CGM) in stable patients who are familiar with the technology. Eight guidelines recommended consulting diabetes specialists for inpatient management of hyperglycemia. Of these, six recommended consulting inpatient specialist individuals or teams, one recommended consulting diabetes educators, and one recommended consulting either teams or educators. In the nine guidelines that addressed treatment, insulin was consistently recommended, most often as basal, prandial, and correctional (n = 5) or basal and prandial (n = 3) (Table 3). The other guideline recommended scheduled insulin. Treatment strategies other than insulin were recommended in guidelines where the patient is stable and regularly eating (n = 5) or the patient’s diabetes is well managed (n = 3), with the recommendation of continuing outpatient oral medications, namely, dipeptidyl peptidase 4 inhibitors. Discouraged treatments included sodium–glucose cotransporter 2 (SGLT2) inhibitors (n = 2) and sliding scale insulin, or reactive insulin dosing in response to capillary blood glucose measures, without long-acting insulin (n = 6).

Guidelines provided additional management recommendations for patients in specific clinical situations (Supplementary Tables 2–6). Nine guidelines included discussion for patients in the intensive care unit, seven for patients in the perioperative period, seven for patients taking glucocorticoids, and six for patients taking nothing by mouth or using an insulin pump. For patients in the intensive care unit, the use of intravenous insulin was recommended in six guidelines, with no specific treatment modality recommended in the other three guidelines. Blood glucose monitoring every 30 min to 2 h was recommended in four guidelines. For patients in the perioperative period, the recommended upper limit blood glucose goal was 180 mg/dL in six guidelines and 140 mg/dL in one. The lower limit blood glucose goal ranged from 80 to 110 mg/dL, and target ranges differed by type of surgery in two guidelines. Treatment was also variable, with recommendations for both intravenous insulin and subcutaneous insulin. For patients taking nothing by mouth, recommendations included blood glucose monitoring every 4–6 h and generally discouraged using short-acting insulin while monitoring and instead encouraged favoring a basal insulin regimen. For patients using glucocorticoids, guidelines suggested typical insulin use with more frequent monitoring for adjustments as needed. Finally, for patients using insulin pumps, continued pump use was recommended where patients remain competent and there is appropriate supervision.

Inpatient recommendations for older adults were present in seven guidelines and were generally focused on factors often associated with older patients such as high clinical complexity and end of life care (Supplementary Table 7). Differences in monitoring were not often recommended; one guideline suggested blood glucose testing twice a day to every 3 days at the end of life. Less stringent blood glucose goals were recommended in five guidelines for those with limited life expectancy or frailty, while one guideline recommended the same blood glucose target range as for the general inpatient. Simplified treatment regimens were suggested in five guidelines, while one guideline recommended the same insulin regimen as for the general inpatient. In contrast to limited inpatient guidance, outpatient goals for older adults were clearly defined, with broad consensus on higher hemoglobin A_1c (HbA_1c) targets (7.0%–8.5% [53–69 mmol/mol]) and avoidance of high-risk medications (Supplementary Table 8).

Hypoglycemia recommendations were present in nine guidelines, but only four were inpatient specific (Supplementary Table 9). Moderate hypoglycemia was consistently defined according to blood glucose <70 mg/dL (n = 6) or <72 mg/dL (n = 3). Severe hypoglycemia was most often defined according to blood glucose <54 mg/dL (n = 3), though two guidelines defined it according to the moderate blood glucose cutoff with the addition of the patient requiring assistance to treat. Treatment consisted of ingestion of carbohydrate or glucose, or the use of intravenous glucose for patients taking nothing by mouth. For those without intravenous access, the use of intranasal or subcutaneous glucagon was recommended (n = 5). It was recommended that treatment regimens be reviewed and potentially modified after the occurrence of hypoglycemia episodes (n = 3), with an emphasis on avoiding or reducing medications with increased hypoglycemia risk such as sulfonylureas (n = 2) and insulins (n = 3).

For the nine guidelines with discussion of transitions of care, outpatient follow-up was recommended in all, though timing of follow-up was specified in only three guidelines and ranged from 1 week to 1 month (Table 4). It was recommended that follow-up be through primary care (n = 5) or a diabetes specialist or endocrinologist (n = 4). Return to home regimens was recommended from the day prior to discharge to the day after discharge in three guidelines. Four guidelines recommended considering changes to outpatient regimens based on HbA_1c (n = 3), blood glucose control in the hospital (n = 2), and new discharge medications such as corticosteroids (n = 1). Components of discharge education most often recommended were a review of medications (n = 8), blood glucose monitoring (n = 7), hypoglycemia prevention (n = 6), and nutrition (n = 6). Six guidelines recommended assessing the patient for self-management capabilities prior to discharge. Provision of medications at discharge was recommended in six guidelines, specifically insulin (n = 3). Provision of monitoring materials such as blood glucose test strips and lancets was recommended in five guidelines and hypoglycemia kits in two guidelines.

In this systematic review of clinical practice guidelines for inpatient blood glucose management, we identified 10 guidelines with recommendations on inpatient blood glucose targets, monitoring, and treatment. We identified multiple areas of consensus across guidelines, including the following: 1) glucose monitoring for patients with diabetes and patients with hyperglycemia on admission, 2) treating inpatient hyperglycemia with a basal and prandial insulin regimen (also called basal-bolus regimen), and 3) avoiding exclusive use of sliding scale insulin. We also identified areas of conflict in recommendations including target inpatient blood glucose ranges, the use of home and oral diabetes medications, and diabetes management provisions at hospital discharge. Finally, we identified areas with little guidance including how to transition patients back to home diabetes regimens, how to manage older adults, and how to manage hypoglycemia in the hospital setting. The lack of consensus present in current guidance may contribute to documented variable treatment patterns (5,6), as differences could reduce adherence to any one guideline.

Across guidelines, there was consensus on monitoring patients with capillary blood glucose checks at meals and bedtime; however, consensus may be shifting regarding the emerging technology of CGM. The topic of continued use of outpatient diabetes technology in the hospital, particularly CGM (32), was brought to the forefront of discussion during the coronavirus disease 2019 pandemic (33). The three most recent guidelines touched on CGM, largely recommending its use in patients already familiar with the technology and when comprehensive support is available. The Endocrine Society recommended CGM specifically in the context of patients at high risk for hypoglycemia (21), and the American Diabetes Association recommended CGM if the patient and inpatient diabetes team are educated in its use (17). The most prominent concerns about inpatient CGM use center around patient safety, as malfunctions or error in management can result in severe hyper- and hypoglycemia (33). As the use of this technology has become more widespread there have been clinical trials to assess its safety in the inpatient setting, with promising results of similar or better glycemic management compared with capillary blood glucose checks, and a reduced risk of hypoglycemia (34), though further research on implementation and protocols has been called for. As research is ongoing and CGM devices are not formally approved for inpatient use, concomitant point of care glucose testing is still practiced.

There were inconsistencies across guidelines on target blood glucose levels, with agreement of the top limit as 180 mg/dL but disagreement as to the lower limit. The data used to inform inpatient blood glucose goals are largely drawn from clinical trials in critical care units (4) rather than trials of general ward patients, further clouding the issue of a safe lower blood glucose limit.

One vulnerable population for whom inpatient guidance is especially lacking is older adults. When present, guidance was often focused on patients at high risk for hypoglycemia and those with high clinical complexity or limited life expectancy rather than explicitly for geriatric inpatients. This contrasts with outpatient diabetes guidelines, which provide contextualized recommendations for older adults and adults with geriatric syndromes including higher HbA_1c targets and avoidance of medications with higher hypoglycemia risks, in alignment with existing data (35). Outpatient guidance may not be transferable to the inpatient setting, however, where there are additional factors contributing to changes in blood glucose such as acute stress and changes to dietary intake, and HbA_1c is not the ideal metric.

Guidelines were in consensus on use of regimens of basal and prandial insulin and holding home oral medications in the hospital for most patients. Although it is common practice to hold oral medications and exclusively use insulin in the hospital, there is uncertainty as to whether this is best practice (36). The dosing flexibility of insulin compared with oral agents must be weighed against the increased risk of hypoglycemia with insulin, particularly for insulin-naïve patients with adequate blood glucose control at home. For the insulin naïve, guidance on insulin dosing in the transition from home oral antihyperglycemic medications to inpatient insulin was not detailed. Observational studies have illustrated similar glucose control outcomes with the use of oral antihyperglycemic medications in the hospital with or instead of insulin (37), and there is the potential for overtreatment and hypoglycemia with the switch to insulin (38). These factors were reflected in the guidelines, with some recommending resumption of home medications or the initiation of oral medications in stable, regularly eating patients. Still, there is concern for changes in patient condition and stability that may be better managed with insulin (38).

The transition of care from hospital to home was discussed in most guidelines but with varying levels of detail. Most guidelines did not provide a follow-up timeframe or recommend how to adjust outpatient antihyperglycemic medication regimens based on inpatient HbA_1c results. This transition from inpatient to home antihyperglycemic medication regimen is an important gap in guidance, including when to transition and whether changes should be made. There is variability in current practice, with an estimated 10%–16% of hospitalized older adults discharged with intensified diabetes regimens despite many already meeting outpatient glucose goals (6). Existing literature has noted the lack of guidance on the transitions of care topic (39) despite how crucial it is to maintaining blood glucose control after discharge. A lack of guidance is compounded by barriers to sufficient communication and follow-up including patient characteristics, such as poor health literacy, limited English proficiency, and un- and underinsured status, and systemic issues, such as long wait time for appointments. While the challenges surrounding the transition of care from hospital to home extend beyond what can be addressed in clinical practice guidelines, having clear, explicit recommendations is an important place to start.

In areas of limited consensus or data, greater clarity is needed. As current clinical trial data are largely from trials targeted toward critically ill patients, randomized clinical trials of both non–critically ill patients and populations at higher risk of hypoglycemia, such as older adults, are needed to inform practice. Consistent recommendations were found in areas with randomized controlled trials for the non–critically ill inpatient, such as optimal hyperglycemia management with insulin. As such, these studies not only would provide a robust evidence base for new recommendations but also could resolve disagreement in current guidance. Specific trials for vulnerable populations like older adults are warranted, as approaches have been proposed for differential care in the outpatient setting (35), without similar attention to differences in care needs in the inpatient setting. More recent guidelines, such as the American Association of Clinical Endocrinology and the American Diabetes Association guidelines, were more comprehensive in discussions of emerging practices such as CGM and use of SGLT2 inhibitors in the hospital. Further, guidelines may be in disagreement, as evidence continues to increase for practice, and not all guidelines have kept pace with evidence as it emerges. Revision between fully updated guidelines, such as the American Diabetes Association’s annual revisions (40), is a promising strategy to keep guidance up to date. Regardless, clinical practice guidelines more comprehensive in nature are needed, even if they were to rely on expert consensus where evidence is absent. As inpatient clinicians make decisions regarding glycemic management every day, they are in need of decision-making frameworks that address the unique issues posed by hospitalizations and care transitions beyond what is currently in the literature.

There are limitations to this study. When collecting guidelines, we excluded certain inpatient scenarios such as blood glucose management during pregnancy and in children, as there are typically separate guidelines for these populations. We also did not include guidance on diabetes emergencies. Our evidence base was also limited by the exclusion of non-English-language studies despite inclusion of internationally developed guidelines. Though we used a predefined search criteria and multiple search strategies, it is possible we missed relevant guidelines.

In sum, this systematic review identifies areas of consensus and areas for future guidelines to address, namely, blood glucose targets, managing diabetes in older adults in the hospital setting, and transitioning patients back to home diabetes regimens. Future clinical research can inform practice guidelines in prioritizing blood glucose target trials in non–critically ill inpatients and high-risk populations and the use of oral antihyperglycemic medications in the hospital.

Funding. This work was funded by the National Institute on Aging (NIA) (grant K76AG074878, principal investigator T.S.A.), National Institutes of Health (NIH). M.A.Sc. was additionally supported by NIA grant K24AG071906, E.M. by NIA grant K24AG035075, and M.A.St. by NIA grants K24AG049057, P30AG44281, and R24AG064025.

The National Institute on Aging had no role in the design or conduct of the study; collection, management, analysis, or interpretation of data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Duality of Interest. S.J.H. reports receiving grants from the Agency for Healthcare Research and Quality outside of the submitted work. M.A.St. reports receiving royalties from UpToDate. T.S.A. reports receiving grants from the American Heart Association, American College of Cardiology, and Boston Claude D. Pepper Older Americans Independence Center as well as personal fees from the American Medical Association and American Medical Student Association outside the submitted work. M.A.St. reports receiving grants from the NIH and honoraria from the American Geriatrics Society. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. L.M.W., B.X.W., E.H.-K., and T.S.A. researched and analyzed data. L.M.W. and T.S.A. wrote the first draft of the manuscript. All authors reviewed, edited, and approved the final version of the manuscript. L.M.W. and T.S.A. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Handling Editors. The journal editors responsible for overseeing the review of the manuscript were Elizabeth Selvin and Meghana D. Gadgil.