Title: Efficacy of basal-bolus insulin regimens in the inpatient management of non-critically

ill patients with type 2 diabetes: A systematic review and meta-analysis

Short title: Basal-bolus insulin in inpatient care

Authors: Merete B. Christensen, MD; Anders Gotfredsen, DMSc; Kirsten Nrgaard, DMSc

Affiliation: Department of Endocrinology, Copenhagen University Hospital Hvidovre,

Denmark

Corresponding author:

Merete B. Christensen

Department of Endocrinology

Kettegrds All 30

2650 Hvidovre

Denmark

Telephone: +45 23 81 12 64

E-mail: Merete.bechmann.christensen.01@regionh.dk

Fax: +45 38 62 61 34

Word count: 3246 words (excluding abstract and references)

Number of tables: 1 table

Number of figures: 3 figures

This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1002/dmrr.2885

This article is protected by copyright. All rights reserved.

Short summary of the manuscript: "Efficacy of basal-bolus insulin regimens in the

inpatient management of non-critically ill patients with type 2 diabetes: A systematic review

and meta-analysis" - # DMRR-16-RES-206.:

Hyperglycemia during hospitalization is associated with increased rates of complications and

longer hospital stays and various insulin regimes are used in the inpatient diabetes

management. In this systematic review and metaanalysis we aimed to assess the efficacy and

safety of basal-bolus insulin therapy in the management of hospitalized non-critically ill type

2 diabetes patients. Five RCTs and seven observational studies comparing basal-bolus insulin

(BBI) therapy to sliding scale insulin (SSI) therapy were included in the review. Meta-

analysis of RCTs showed significantly lower mean daily blood glucose with BBI than SSI,

but basal-bolus insulin therapy was associated with increased risk of mild hypoglycemia.

This article is protected by copyright. All rights reserved.

ABSTRACT:

Background: Hyperglycemia during hospitalization is associated with increased rates of

complications and longer hospital stays. Various insulin regimens are used in the inpatient

diabetes management of non-critically ill patients. We aimed to assess the efficacy and safety

of basal-bolus insulin therapy (BBI) by summarizing evidence from studies of BBI versus

sliding scale insulin therapy (SSI) in the management of hospitalized non-critically ill type 2

diabetes patients.

Methods: We searched MEDLINE, EMBASE, Scopus and the Cochrane Library for studies

comparing BBI therapy with SSI therapy in hospitalized non-critically ill patients with type 2

diabetes. Primary outcome was mean daily blood glucose (BG) during admission. Secondary

outcomes were incidence of hypoglycemia and length of hospital stay. Results of included

randomized controlled trials (RCT) were pooled and meta-analyzed to provide estimates of

the efficacy of BBI therapy.

Results: Five RCTs and seven observational studies were included in the review. Meta-

analysis of RCTs showed significantly lower mean daily BG with BBI than SSI. Mean

difference in daily BG between the two regimens ranged from 14 29 mg/dl. BBI therapy

was associated with increased risk of mild hypoglycemia (BG70 mg/dl, RR 5.75; 95% CI

2.79-11.83), (BG60 mg/dl, RR 4.21; 95% CI 1.6111.02) compared with SSI therapy. There

was no difference in risk of severe hypoglycemia (BG40 mg/dl) and no difference in mean

length of stay.

Conclusion: Basal-bolus insulin in the inpatient diabetes management results in significantly

lower mean daily BG than sliding scale insulin but is associated with increased risk of mild

hypoglycemia.

This article is protected by copyright. All rights reserved.

KEYWORDS: Inpatient management, hyperglycemia, type 2 diabetes, basal-bolus insulin

INTRODUCTION:

The prevalence of diabetes among hospitalized patients is high. One out of four hospitalized

patients has diabetes [13], thus treating hyper- or hypoglycemia represents an everyday

challenge in hospitals worldwide. Hyperglycemia during admission is associated with

increased rate of complications [3] and longer hospitals stays [4], hence the inpatient care of

diabetes patients accounts for a substantial proportion of total health costs [5].

The role of glucose control during hospitalization has been a subject of debate recently, as

has the choice of insulin regimens. The first landmark trials on glycemic targets were carried

out in intensive care units. Van den Berge et al. reported that intensive insulin treatment with

a glycemic goal of 110 mg/dl (6.2 mmol/L) reduced in-hospital morbidity and mortality [6]

however, subsequent trials failed to confirm this finding of reduced mortality with intensive

glycemic control in critically ill patients [7].

Only few randomized trials have studied glycemic targets in non-critically ill patients. Murad

et al. found in a systematic review and meta-analysis of randomized and observational studies

of non-critically ill patients with hyperglycemia that intensive glycemic control (using a

variable of insulin regimens) reduced the risk of hospital acquired infections but found no

effect on mortality [8]. A target glucose range of 140180 mg/dL (7.810.0 mmol/L) is now

recommended in the recent American Diabetes Association (ADA) guidelines for the

majority of critically ill and non-critically ill patients [9].

Insulin is the preferred treatment of hospitalized patients with sustained hyperglycemia and

several different insulin regimens are used worldwide. An insulin regimen with basal insulin

1-2 times daily and bolus insulin at the main meals is recommended by ADA for the inpatient

This article is protected by copyright. All rights reserved.

management of diabetes [9] and the sole use of sliding-scale insulin, i.e. fast-acting insulin as

correction insulin when blood glucose is above target, is strongly discouraged. However,

sliding scale insulin is still used in many countries. The reason for the persistent use of

sliding scale insulin regimens is unclear, and may simply be due to clinical inertia.

In this systematic review and meta-analysis we aim to assess the efficacy and safety of basal-

bolus insulin regimens for the inpatient management of non-critically ill patients with type 2

diabetes by summarizing all available evidence from randomized controlled trials (RCT) and

observational studies of basal-bolus insulin treatment versus sliding scale insulin.

METHODS:

Data Sources and Searches: We searched MEDLINE, EMBASE, Scopus and the Cochrane

Library using the following search terms: Type 2 diabetes, hyperglycemia, hospitalization,

inpatient, insulin, basal-bolus and sliding scale. The complete search for MEDLINE was

(((((("Diabetes Mellitus, Type 2"[Mesh:NoExp]) OR "Hyperglycemia"[Mesh:NoExp])) OR

((type 2 diabetes OR hyperglycemia)))) AND ((((("Hospitalization"[Mesh]) OR

"Inpatients"[Mesh]) OR "Hospitals"[Mesh])) OR (inpatient* OR hospitalization OR

hospitalisation OR "non-critically ill" OR non-intensive))) AND (((("Insulin"[Mesh]) OR

"Insulin, Long-Acting"[Mesh])) OR ((basal-bolus OR "insulin therapy" OR "sliding scale"))).

Manual searches included scanning of reference lists in relevant papers and conference

proceedings. Only studies published in English were included in this review.

Study Selection: Two authors (MC and KN) independently screened all titles and abstracts

identified through the search strategies. Studies included in this meta-analysis and review

were RCTs, observational studies and retrospective studies. We searched for studies

This article is protected by copyright. All rights reserved.

comparing basal-bolus insulin (BBI) treatment (combination of long-acting and pre-meal

fast-acting insulin) with sliding scale insulin (SSI) for the treatment of hyperglycemia during

hospitalization of non-critically ill patients with type 2 diabetes. SSI was defined as

subcutaneous insulin regimens using fast-acting insulin alone for the treatment of

hyperglycemia. Only studies with adult patients were included and the studies had to measure

the outcomes of interest. Studies were excluded if conducted in intensive care units or in

patients receiving parenteral nutrition.

Data extraction and Quality Assessment: Data were extracted using a standard data

extraction form. Two authors (MC and KN or MC and AG) independently extracted the data

and differences were resolved by consensus. If relevant data were not included in the

published study reports, the authors were contacted for further information. The risk of bias

in RCTs was assessed as recommended in the Cochrane Handbook for Systematic Reviews

of Interventions. Each domain (sequence generation, allocation concealment, blinding,

incomplete outcome, selective reporting and other) was scored as low risk, unclear or high

risk.

Data Synthesis and Analysis: Primary outcome was mean daily blood glucose. Secondary

outcomes were hypoglycemic events, mean length of stay and total daily insulin dose. Only

RCTs are included in the meta-analysis. Data from observational studies are presented as a

summary of findings and analyzed in a narrative synthesis.

Studies were analyzed using the mean and SD of change for the intervention period of each

study. For the meta-analysis we used both a fixed effect model and a random effects model.

Since no discrepancy was found between the two models, only results from the random

effects model are presented. Heterogeneity was assessed using the I2 statistics, which

This article is protected by copyright. All rights reserved.

indicates the proportion of variation in study results that is caused by between-study

heterogeneity rather than sampling error. Data analyses were performed using the software

Review Manager 5.3 (The Nordic Cochrane Centre, The Cochrane Collaboration,

Copenhagen, Denmark). Due to the small number of trials we did not conduct meta-

regression analyses, subgroup analyses to evaluate bias or assessments of publication bias or

small study effects.

RESULTS:

Study identification: Through electronic literature search 817 studies were identified. Title

and abstract was screened and 756 studies were considered irrelevant. The full texts of the

remaining 61 articles were reviewed. Twelve of these studies met the inclusion criteria and

were eligible for inclusion in the systematic review. The study selection process is illustrated

in Fig. 1.

Study characteristics and quality assessment: Five studies were RCTs [1014], two were

observational, prospective studies with historical controls [15,16] and five were retrospective

studies [1721]. Characteristics of the included studies are presented in Table 1, including

number and mean age of patients, mean admission blood glucose and mean HbA1C. All

studies were done in a non-ICU setting. Six were done in medical wards, three were done in

surgical wards and three reported results from both medical and surgical wards. Eight studies

included only patients with known history of type 2 diabetes. Two studies included a few

patients with type 1 diabetes and four studies also included patients with hyperglycemia

without known history of diabetes. Two studies did not specify type of diabetes. All studies

compared basal-bolus treatment with sliding scale treatment. In some studies sliding scale

This article is protected by copyright. All rights reserved.

insulin was used as stand-alone-therapy and in others it was used as correction insulin as add-

on to the patients preadmission therapy which in a few cases included basal insulin. Insulin

glargine was used as basal insulin in seven studies; NPH insulin was used in four studies and

in one retrospective study both types of insulin was used. The total daily starting dose of

insulin for basal-bolus treatment varied from 0.25 U/kg to 0.50 U/kg.

None of the studies were blinded since blinding of patients and study personnel was not

feasible in any of the included studies and all RCTs had a high risk of bias in at least one

domain. The retrospective studies were heterogeneous and had a potential risk of

confounding, however; lack of reported information made it impossible to assess the risk of

bias in these studies.

Glycemic control: Meta-analysis of mean daily blood glucose in the RCTs showed

significantly lower mean blood glucose in the BBI group than in the SSI group (Fig. 2). The

mean difference in daily blood glucose between the two treatment regimens ranged from 14

to 29 mg/dl (0.8 1.6 mmol/L). Heterogeneity (I2) between the studies was moderate (43%).

Overall basal-bolus therapy reduced the mean daily blood glucose significantly more than

sliding scale insulin therapy in 11 of the 12 included studies. Six studies reported data on the

number of patients achieving glycemic target [1013,15,19]. In all of them significantly more

patients treated with BBI had blood glucose values within the glycemic target range during

their hospital stay than patients treated with SSI (BBI range: 34% - 66%, SSI range: 23% -

50%).

Surgery patients had lower mean daily blood glucose than patients at medical wards,

however; patients at surgical wards generally had lower mean admission blood glucose and

lower mean HbA1C than medical patients.

This article is protected by copyright. All rights reserved.

Hypoglycemia: Meta-analysis of the risk of hypoglycemia showed that BBI therapy was

associated with increased risk of mild hypoglycemia (BG 70 mg/dl (3.9 mmol/L), RR 5.75;

95% CI 2.79-11.83), (BG 60 mg/dl (3.3 mmol/l), RR 4.21; 95% CI 1.6111.02) compared

with SSI therapy. No statistical heterogeneity was found (I2 = 1%). There was no significant

difference in the risk of severe hypoglycemia (BG 40 mg/dl (2.2 mmol/L)) between the two

treatment regimens (Fig. 3). However there were only few events of severe hypoglycemia (6

patients out of 633 patients), all of them occurring in the BBI group.

Overall the incidence of hypoglycemia in the included studies ranged from 2 % to 29 %. Six

studies reported significantly more episodes of hypoglycemia in patients treated with basal-

bolus therapy than in patients treated with sliding scale [13,1517,21,22]. Two retrospective

studies reported significantly more episodes of hypoglycemia in SSI-treated patients than in

BBI-treated patients [18,21]. The remaining four studies found no significant difference in

hypoglycemic episodes between the two treatment regimens [10,14,16,17]. The definition of

hypoglycemia was not consistent in all included studies. Some studies only reported episodes

of mild hypoglycemia and other studies reported episodes of both mild and severe

hypoglycemia. Three RCTs reported BG values 70 mg/dl (3.9 mmol/L) [1113], and four

RCTs reported BG values 60 mg/dl (3.3 mmol/L) and BG values 40 mg/dl (2.2 mmol/L)

[1013]. Schroeder et al. found no significant difference in hypoglycemic episodes between

the two treatment groups, but did not report the number of hypoglycemic episodes [14].

Length of stay: The meta-analysis found no overall difference in mean length of stay

between the two treatment regimens. Difference in mean length of stay varied from 0.1 to 0.5

days in the RCTs. Overall two studies found significantly shorter length of stay for BBI-

treated patients than SSI-treated patients [14,20] and one study found significantly shorter

length of stay for SSI-treated patients than BBI-treated patients [15]. One retrospective study

This article is protected by copyright. All rights reserved.

did not report length of stay and one study did not report difference in length of stay between

the two groups. The remaining 7 studies found no significant difference in length of stay

between the two regimens.

Infections/postoperative complications: Only two studies reported outcome on

postoperative complications. In the RABBIT 2 Surgery study basal-bolus treatment was

associated with significant reduction in postoperative complication as a composite including

wound infection, pneumonia, bacteremia, respiratory failure and acute renal failure [11]. In

contrast Schroeder et al. found no difference in postoperative complication between the BBI

group and the SSI group [14]. Due to the small number of studies no meta-analysis was done

for this outcome.

Total daily insulin dose: The total daily insulin dose was higher for the BBI groups than for

the SSI groups. Total daily insulin dose ranged from 12.5 to 55.6 U/kg/day for the BBI

groups and ranged from 3.1 to 35.2 U/kg/day for the SSI groups.

DISCUSSION:

In the present meta-analysis of RCTs we have demonstrated that basal-bolus insulin therapy

significantly reduces mean blood glucose in non-critically ill hospitalized patients with type 2

diabetes and/or hyperglycemia, compared to traditional sliding scale insulin. This result is

supported in all the observational studies, except one retrospective study. As a side effect the

meta-analysis however, showed a significant increase in the frequency of mild hypoglycemia

with basal-bolus treatment but no difference in incidence of severe hypoglycemia was found.

This article is protected by copyright. All rights reserved.

In the included studies, the majority of participants had type 2 diabetes while a few studies

also included some patients with type 1 diabetes or patients with hyperglycemia without a

diagnosis of diabetes. However, this heterogeneity might very well reflect the usual clinical

setting in general medical and surgical wards in most hospitals, why it should not be seen as

an obstacle to generalize and apply the results.

The insulin algorithm used for basal-bolus insulin treatment and adjustment varied in the

included studies. Insulin starting dose varied from 0.25 U/kg to 0.50 U/kg and different

titration algorithms were used. Both could have a significant influence on mean blood

glucose and the risk of hypoglycemia, but the heterogeneity of these algorithms makes it

impossible to evaluate or draw conclusions on the choice of basal-bolus insulin algorithm. To

our knowledge no studies or reviews have examined the effect of insulin starting dose or

insulin titration algorithm on achieved mean blood glucose during hospitalization. One study

indicated that electronic insulin orders improved mean blood glucose, but different insulin

titration algorithms were not compared [22].

Hyperglycemia as well as hypoglycemia during hospitalization is associated with increased

length of hospital stay. Increased glycemic variability has also been associated with increased

length of stay in patients with type 2 diabetes [23]. In this review we found no consistent

association between insulin regimen and length of stay. SSI treatment might result in greater

glycemic variability than BBI treatment, but none of the included studies have included

continuous glucose monitoring to analyze that aspect. Only two of the included studies

reported outcome on postoperative complications and only one of these found a benefit of

basal-bolus therapy. The impact of glycemic control on postoperative complications is

described in several other trials. In a recent randomized controlled trial of coronary artery

bypass grafting (CABG) patients by Umpierrez et al. intensive insulin therapy (target 100 -

140 mg/dl) did not significantly reduce perioperative complications after CABG surgery

This article is protected by copyright. All rights reserved.

compared to less intensive insulin therapy with target 141 180 mg/dl [24]. However,

Boreland et al. found in their systematic review that maintaining BG levels 200 mg/dl with

continuous insulin infusions in all stages of the perioperative period in cardiac surgery

patients with diabetes reduced the incidence of surgical site infection [25]. The association

between intensive glycemic control and postoperative complications in non-critically ill

patients has only been described in a few studies. In a systematic review and meta-analysis

Murad et al. found that intensive glycemic control was associated with decreased risk of

infections, a finding that was mainly derived from studies in surgical settings [8].

All studies included in this present review reported hypoglycemia as a secondary outcome;

however the definition of hypoglycemia varied between studies. There was a wide range in

the incidence of mild hypoglycemia. Only 3 % of BBI treated patients had an episode of

hypoglycemia in the RABBIT study by Umpierrez et al. In contrast 29% of BBI treated

patients experienced hypoglycemia in the study by Mader et al. The starting insulin dose and

the used insulin titration algorithm might have influenced the risk of hypoglycemia. None of

the included studies reported data on beta cell function and only three studies reported data

on diabetes duration. Diabetes duration was not reported in the RABBIT study, but all

patients were insulin nave. Mader et al. included patients with preadmission therapy with

OADs and/or insulin, and the patients had mean diabetes duration of 13 years. Long diabetes

duration is associated with decreased beta-cell function, which could make the patients more

susceptible to hypoglycemia during insulin treatment. Furthermore, some patients have a

higher risk of hypoglycemia than others. Factors associated with higher risk of hypoglycemia

include older age, decreased kidney function, low body weight and reduced nutritional intake.

To minimize the risk of hypoglycemia it is important to pay attention to these factors and

adjust insulin dose daily.

This article is protected by copyright. All rights reserved.

None of the included studies were powered to estimate mortality rate with the two different

insulin regimens, but mortality was reported as an outcome in two studies. In these two

studies only three events occurred (two in BBI group and one in SSI group).

The relationship between hypoglycemia and mortality in critically ill patients at intensive

care units have been described in several studies, but only a few have studied non-critically ill

patients at general wards. One retrospective cohort study by Turhin et al. analyzed data from

4,386 admissions of diabetic patients at general non-ICU wards. They defined hypoglycemia

as BG 50 mg/dl (2.8 mmol/L) and found that inpatient hypoglycemia was associated with

increased length of stay and increased inpatient mortality. Inpatient mortality was 2.96 % for

patients who had at least one hypoglycemic episode during the hospital stay versus 0.82 %

for patients who had no hypoglycemic episodes [26]. However this study did not differentiate

between spontaneous hypoglycemia and drug-associated hypoglycemia. Boucai et al found in

another retrospective cohort study of non-critically ill patients (with and without diabetes) at

general wards that only spontaneous and not drug-associated hypoglycemia (BG 70 mg/dl

(3.9 mmol/L)) was linked to increased mortality risk [27]. After adjustment for patient

comorbidities, the association between spontaneous hypoglycemia and mortality was

eliminated. This suggests that hypoglycemia could be a marker of disease burden rather than

a direct cause of death.

Our systematic review and meta-analysis has several strengths. It is the first review to

evaluate the efficacy and safety of basal-bolus insulin treatment compared to sliding scale

insulin and the first to include a meta-analysis of all RCTs. The strength of this review also

includes the comprehensive literature search which allows us to summarize all the available

information on this topic. Our review also has some limitations. Only few RCTs have been

published and the inclusion of observational studies increases the risk of bias and complicates

the comparison of included studies due to the heterogeneity in study design. The individual

This article is protected by copyright. All rights reserved.

studies included a relatively small number of patients, which made it impossible to do

subgroup analysis. Due to the relatively small number of RCTs included in the meta-

analysis, the results should be interpreted with caution.

The findings of the present study support the need for further trials addressing the problems

of hypoglycemia during hospitalization. In the future, computer-guided electronic decision-

aid devices could help to reduce the rate of hypoglycemia with the use of insulin-on-board

calculators, as described in the Gluco-Tab system [28]. The possibility to enter intake of

carbohydrates in the decision-aid device might also be helpful to guide insulin dosage to

patients with reduced nutritional intake. Another approach to reduce the risk of hypoglycemia

could be the use of the new ultra-long acting basal insulins, which in outpatient studies have

shown a reduced risk of nocturnal hypoglycemia [29]. We suggest that future studies could

evaluate these new basal insulins for the inpatient diabetes management and further study

insulin titration algorithms to address the problem of hypoglycemia.

In conclusion, basal-bolus insulin treatment for non-critically ill hospitalized patients with

diabetes efficiently reduces the mean daily blood glucose, but is associated with an increased

risk of mild hypoglycemia. There was no statistical significant difference in the risk of severe

hypoglycemia and no difference in length of hospital stay between the basal-bolus group and

the sliding scale group. Further randomized trials are needed to address the problem of

hypoglycemia.

This article is protected by copyright. All rights reserved.

ACKNOWLEDGEMENTS:

The authors would like to thank Bianca Hemmingsen for statistical assistance with the meta-

analysis. We also wish to thank Thomas Almdal for helpful comments on an earlier version

of the manuscript.

Funding: None

Duality of Interest: No potential conflicts of interest relevant to this article

Author Contributions: MC undertook the literature search. MC, AG and KN reviewed the

abstracts and full articles, extracted data and analyzed data. MC and KN wrote the

manuscript. All authors designed the study, contributed to the discussion and critically

reviewed the final manuscript. KN is the guarantor of this work and, as such, had full access

to all the data in the study and takes responsibility for the integrity of the data and the

accuracy of the data analysis.

This article is protected by copyright. All rights reserved.

REFERENCES:

1. Bach L, Ekinci EI, Engler D, Gilfillan C, Hamblin PS, MacIsaac RJ, et al. The high

burden of inpatient diabetes mellitus: the Melbourne Public Hospitals Diabetes

Inpatient Audit. Med J Aust. 2014;20:334338.

2. National Health Service GB. National Diabetes Inpatient Audit information relating to

health. 2013; Available from: http://www.hscic.gov.uk/catalogue/PUB13662/nati-diab-

inp-audi-13-nat-rep.pdf

3. Umpierrez GE, Isaacs SD, Bazargan N, You X, Thaler LM, Kitabchi AE.

Hyperglycemia: An independent marker of in-hospital mortality in patients with

undiagnosed diabetes. J Clin Endocrinol Metab. 2002;87:978982.

4. Whitston M, Chung S, Henderson J, Young B. What can be learned about the impact

of diabetes on hospital admissions from routinely recorded data? Diabet Med.

2012;29:11991205.

5. American Diabetes Association. Economic costs of diabetes in the U.S. In 2007.

Diabetes Care. 2008;31:596615.

6. van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, et al.

Intensive insulin therapy in critically ill patients. N Engl J Med. 2001;345:13591367.

7. Griesdale DEG, de Souza RJ, van Dam RM, Heyland DK, Cook DJ, Malhotra A, et al.

Intensive insulin therapy and mortality among critically ill patients: a meta-analysis

including NICE-SUGAR study data. CMAJ 2009;180:821827.

8. Murad MH, Coburn J a, Coto-Yglesias F, Dzyubak S, Hazem A, Lane M a, et al.

Glycemic control in non-critically ill hospitalized patients: a systematic review and

meta-analysis. J Clin Endocrinol Metab. 2012;97:4958.

This article is protected by copyright. All rights reserved.

9. American Diabetes Association. Standards of medical care in diabetes. Diabetes Care

2016;39(Supplement 1):s1109.

10. Umpierrez GE, Smiley D, Zisman A, Prieto LM, Palacio A, Ceron M, et al.

Randomized study of basal-bolus insulin therapy in the inpatient management of

patients with type 2 diabetes (RABBIT 2 Trial). Diabetes Care 2007;30:21812186.

11. Umpierrez GE, Smiley D, Jacobs S, Peng L, Temponi A, Mulligan P, et al.

Randomized study of basal-bolus insulin therapy in the inpatient management of

patients with type 2 diabetes undergoing general surgery (RABBIT 2 surgery).

Diabetes Care 2011;34:256261.

12. Umpierrez GE, Smiley D, Hermayer K, Khan A, Olson DE, Newton C, et al.

Randomized study comparing a Basal-bolus with a basal plus correction insulin

regimen for the hospital management of medical and surgical patients with type 2

diabetes: basal plus trial. Diabetes Care 2013;36:21692174.

13. Mader JK, Neubauer KM, Schaupp L, Augustin T, Beck P, Spat S, et al. Efficacy,

usability and sequence of operations of a workflow-integrated algorithm for basal-

bolus insulin therapy in hospitalized type 2 diabetes patients. Diabetes Obes Metab.

2014;16:137146.

14. Schroeder JE, Liebergall M, Raz I, Egleston R, Ben Sussan G, Peyser A, et al. Benefits

of a simple glycaemic protocol in an orthopaedic surgery ward: A randomized

prospective study. Diabetes Metab Res Rev. 2012;28:7175.

15. Roberts GW, Aguilar-Loza N, Esterman A, Burt MG, Stranks SN. Basalbolus insulin

versus sliding-scale insulin for inpatient glycaemic control: a clinical practice

comparison. Med J Aust. 2012;196:266269.

This article is protected by copyright. All rights reserved.

16. Prez A, Reales P, Barahona MJ, Romero MG, Miambres I. Efficacy and feasibility

of basal-bolus insulin regimens and a discharge-strategy in hospitalised patients with

type 2 diabetes--the HOSMIDIA study. Int J Clin Pract. 2014;68:12641271.

17. Harbin M, Dossa A, de Lemos J, Drummond I, Paty B, Taylor B. Evaluation of

Protocol-Guided Scheduled Basal-Nutritional-Correction Insulin Over Standard Care

for Vascular Surgery Patients. Can J Diabetes;2015:49.

18. Chen HJ, Steinke DT, Karounos DG, Lane MT, Matson AW. Intensive insulin

protocol implementation and outcomes in the medical and surgical wards at a Veterans

Affairs Medical Center. Ann Pharmacother. 2010;44:249256.

19. Patel GW, Roderman N, Lee KA., Charles MM, Nguyen D, Beougher P, et al. Sliding

scale versus tight glycemic control in the noncritically ill at a community hospital. Ann

Pharmacother. 2009;43:17741780.

20. Akhtar ST, Khalid M, Naqvi IH, Vaswaniess AS. Inpatient management of type 2

Diabetes Mellitus: Does choice of insulin regimen really matter? Pak J Med Sci.

2014;30:895898.

21. Zaman Huri H, Permalu V, Vethakkan SR. Sliding-scale versus basal-bolus insulin in

the management of severe or acute hyperglycemia in type 2 diabetes patients: A

retrospective study. PLoS One 2014 Jan;9:1-9.

22. Dicker D, Marcoviciu D, Beylin V, Belnik Y, Solun B. The Efficacy of Standardized

Electronic Orders In Controlling Blood Glucose Levels In Internal Medicine

Departments. J Diabetes Metab. 2015;6:2155.

23. Martin WG, Galligan J, Simpson S, Greenaway T, Burgess J. Admission blood

glucose predicts mortality and length of stay in patients admitted through the

emergency department. Intern Med J. 2015;45:916924.

This article is protected by copyright. All rights reserved.

24. Umpierrez G, Cardona S, Pasquel F, Jacobs S, Peng L, Unigwe M, et al. Randomized

Controlled Trial of Intensive Versus Conservative Glucose Control in Patients

Undergoing Coronary Artery Bypass Graft Surgery: GLUCO-CABG Trial. Diabetes

Care. 2015;38:1665-1672.

25. Boreland L, Scott-Hudson M, Hetherington K, Frussinetty A, Slyer JT. The

effectiveness of tight glycemic control on decreasing surgical site infections and

readmission rates in adult patients with diabetes undergoing cardiac surgery: A

systematic review. Heart Lung 2015;44:430440.

26. Turchin A, Matheny ME, Shubina M, Scanlon S V., Greenwood B, Pendergrass ML.

Hypoglycemia and clinical outcomes in patients with diabetes hospitalized in the

general ward. Diabetes Care 2009;32:11531157.

27. Boucai L, Southern WN, Zonszein J. Hypoglycemia-associated mortality is not drug-

associated but linked to comorbidities. Am J Med 2011;124:10281035.

28. Neubauer KM, Mader JK, Aberer F, Donsa K, Augustin T, Schaupp L, et al.

Standardized Glycemic Management with a Computerized Workflow and Decision

Support System for Hospitalized Patients with Type 2 Diabetes on Different Wards.

Diabetes Technol Ther. 2015;17:18.

29. Heller S, Mathieu C, Kapur R, Wolden ML, Zinman B. A meta-analysis of rate ratios

for nocturnal confirmed hypoglycaemia with insulin degludec vs. Insulin glargine

using different definitions for hypoglycaemia. Diabetic Medicine 2016;33:478-487.

This article is protected by copyright. All rights reserved.

 Table 1: Characteristics of included studies

Author/year Population and setting Sample size Mean Mean Mean Intervention Control Glycemic target Mean length
a) a) a)
Diabetes type (Intervention age HbA1C admission Regimen and insulin type Regimen and insulin (mg/dl) of stay
b) b)
Pre-admission treatment group/ (Years) (% BG Insulin starting dose type (days)
c) b)
Setting Control group) (mmol/mol)) (mg/dl) (TDD: Total daily dose) Insulin starting
instructions

Randomized controlled trials:

a) a) a)
Umpierrez et al Type 2 diabetes 65/65 56 13 8.8 (73) 227 65 BBI: Glargine x 1 daily + SSI: Regular insulin Fasting + premeal BBI: 5.2 6
b) b)
(2007) Insulin nave glulisine x 3 daily BG 140 mg/dl BG 140 SSI: 5.1 4
c) b)
General medical ward Starting TDD: 0.4 U/kg
at BG 140 -200 mg/dl, 0.5
U/kg at BG 201- 400 mg/dl
a) a) a)
Umpierrez et al Type 2 diabetes 104/107 58 11 7.7 (61) 190 92 BBI: Glargine x 1 + SSI: Regular insulin Fasting + premeal BBI: 9.1 7
b) b)
(2011) Diet, OAD or low dose Glulisine x 3 BG 140 mg/dl BG: 100 140 SSI: 9.4 13
b)
insulin (0.4 U/kg) Starting TDD: 0.5 U/kg
c)
General surgery ward
a) a) a)
Umpierrez et al Type 2 diabetes 144/74 59 11 8.4 (68) 204 84 BBI: Glargine x 1 + SSI: Regular insulin Fasting + premeal BBI: 5.9 5
b) b)
(2013)* Diet, OAD or low dose Glulisine x 3 BG 140 mg/dl BG: 100 140 SSI: 5.5 5
b)
insulin (0.4 U/Kg) Starting dose 0.5 U/kg
c)
General medical and surgery
services
a) a) a)
Mader et al Type 2 diabetes 37/37 68 11 8.7 (72) 197 54 BBI: Glargine x 1 + Standard care: SSI ( Fasting + premeal BBI: 7.5 5
b)
(2013) Diet, OAD or insulin Aspart x 3 pre-admission BG: 100 140 SSI: 7.0 4
c) b)
General internal medicine Starting dose 0.5 U/kg therapy). Insulin type
ward not reported
b)
NR
a) a) a)
Schroeder et al Known history of diabetes 35/30 70.5 N/A 170 BBI: NPH insulin x 1 + SSI: Actrapid 100 180 BBI: 7
b)
(2012) (type not specified) + patients Regular insulin x 3 BG 150 mg/dl SSI: 9.2
b)
with undiagnosed diabetes and Starting dose 0.25 U/kg
hyperglycemia (2 or more BG
180 mg/dl) b) Not reported
c)
Orthopedic surgery ward

This article is protected by copyright. All rights reserved.

Author/year Population and setting Sample size Mean Mean Mean Intervention Control Glycemic target Mean length
a) a) a)
Diabetes type (Intervention age HbA1C admission Regimen and insulin type Regimen and insulin (mg/dl) of stay
b) b)
Pre-admission treatment group/ (Years) (% BG Insulin starting dose type (days)
c) b)
Setting Control group) (mmol/mol)) (mg/dl) (TDD: Total daily dose) Insulin starting
instructions

Prospective trials with historical controls:

a) a) a)
Roberts et al Type 2 diabetes 124/96 74 12 8.0 (64) 197 75 BBI: Glargine x 1 + SSI: Actrapid 72 180 BBI: 6.8 1.6
b) b)
(2012) Diet, OAD or insulin Rapid-acting x 3 NR SSI: 5.9 2.2
c) b)
Medical and surgical ward Starting dose 0.3 U/kg
for patients with diet-
controlled diabetes, 0.4
U/kg for the rest
a) a) a)
Prez et al Type 2 134/62 73 10 8.3 (67) 211 70 BBI: Glargine x 1 + SSI: Regular insulin Premeal: 90 BBI: 8.8
b)
(2014) Diet, OAD or insulin Rapid-acting x 3 (some patients also 130 SSI: 10.7
c) b)
Medical and surgical Starting dose 0.4 U/kg received long-acting Postmeal: 180
insulin)
b)
NR

Retrospective studies:
a) a) a)
Harbin et al Type 1(17%) + type 2(83%) 46/41 69 9 NR 164 BBI:NPH x 2 + SSI: Insulin type not 90 - 144 BBI: 15.4 14
b)
(2015) Insulin OAD Regularx3 reported (some SSI: 13.0 11
c) b)
Vascular surgery unit Starting dose = pre- patients also received
admission TDD long-acting insulin)
b)
NR
a) a) a)
Chen et al Known history of diabetes, 89/409 69.7 7.5 (58) 195 90 BBI: Glargine/NPH x 1 + SSI: Insulin type not 80 - 150 BBI: 6.2 4
(2010) type unknown Aspart/regular x 3 reported SSI: 7.4 8.7
b) b)
NR NR
c)
Medical ward?
a) a) a)
Patel et al Type 1 + type 2 + 210/121 67.8 7.1 (54) F-BG: BBI: Glargine x 1 + SSI: Insulin type not 80 - 150 NR
(2009) undiagnosed diabetes 183 92 Aspart/regular x 3 reported
b) b)
NR NR
c)
Medical ward
a) a) a)
Akhtar et al Type 2 138/186 51 NR NR BBI: NPH + short acting SSI: Regular NR BBI: 7.8 1.9
b) b)
(2014)* Diet or OAD insulin NR SSI: 15.5 3.6
c)
Medical ward
a) a) a)
Zaman et al Type 2 159/179 NR 11.7 (104) 219 167 BBI: NPH + Actrapid SSI: Actrapid NR Mean for both
b) b)
(2014) NR NR groups: 7.9 6
c)
Medical ward

This article is protected by copyright. All rights reserved.

Legend: Mean values SD are given unless otherwise indicated. BG: Blood glucose, F-BG: Fasting blood glucose, BBI: Basal-bolus insulin,

SSI: Sliding scale insulin, NR: Not reported, OAD: Oral antidiabetic agents.

*These trials have three treatment arms; only BBI and SSI arms are described here.

This article is protected by copyright. All rights reserved.

Figure 1:

Legend: Flow diagram of study selection process.

This article is protected by copyright. All rights reserved.

Figure 2:

Legend: Meta-analysis of difference in mean blood glucose between basal-bolus insulin

(BBI) regimens and sliding scale insulin (SSI) regimens.

This article is protected by copyright. All rights reserved.

Figure 3:

Legend: Risk ratio of hypoglycemia. A: RR of mild hypoglycemia (BG 60 mg/dl). B: RR

of severe hypoglycemia (BG 40 mg/dl). BBI: Basal-bolus insulin, SSI: Sliding scale

insulin, NE: Not estimable.

This article is protected by copyright. All rights reserved.