ORIGINAL RESEARCH

Movement is Improvement: The Therapeutic Effectsof Exercise and General Physical Activity on GlycemicControl in Patients with Type 2 Diabetes Mellitus:A Systematic Review and Meta-Analysisof Randomized Controlled Trials

Sayed Z. A. Shah . Jawad A. Karam . Alam Zeb . Rafi Ullah .

Arif Shah . Ijaz Ul Haq . Iftikhar Ali . Haider Darain . Hong Chen

Received: November 10, 2020 /Accepted: January 15, 2021 / Published online: February 5, 2021� The Author(s) 2021

ABSTRACT

Introduction: Exercise is considered a corner-stone in achieving an optimized blood glucoselevel and reducing body weight, body massindex (BMI), and waist circumference. Thisstudy aimed to investigate and quantitatively

summarize the literature regarding the thera-peutic effects of exercise and general physicalactivity on glycemic control.Methods: A systematic review and meta-analy-sis of the literature on the therapeutic effects ofexercise on glycemic control in patients withtype 2 diabetes mellitus (type 2 DM) were con-ducted using electronic databases. Studies withan exercise intervention lasting more than8 weeks were included. Both qualitative andquantitative analyses were performed. Qualita-tive data were presented narratively intable form. Quantitative analysis was performedusing a random-effects model with a 95% con-fidence interval and a significance level of 0.05.The Physiotherapy Evidence Database (PEDro)scale and the Cochrane Risk of Bias 2 (RoB2)tool were used to assess the quality of evidenceand the risk of bias.Results: A total of 21,559 articles were identi-fied through different databases. Out of 21559studies, only 32 randomized controlled trialswere deemed eligible for inclusion in this study.The average exercise session was 45.15 min,while the average follow-up duration was21.94 weeks. The mean exercise frequencyaccording to our findings was 3.25 days/week.Almost all the studies reported decreases inglycated hemoglobin (HbA1c; P\ 0.0001),fasting blood glucose (P = 0.03), BMI (P = 0.04),and waist circumference (P = 0.007) after theexercise intervention.

S. Z. A. Shah � H. Chen (&)Department of Rehabilitation Medicine, TongjiHospital, Tongji Medical College, HuazhongUniversity of Science and Technology, Wuhan,People’s Republic of Chinae-mail: chenhong1129@hotmail.com;chenhong1129@hust.edu.cn

J. A. KaramDepartment of applied exercise science, ConcordiaUniversity Chicago, 7400 Augusta St., River Forest,IL 60305, United States

A. ZebSchool of Health Sciences, Peshawar, Pakistan

R. Ullah � A. Shah � I. U. Haq � I. AliParaplegic Centre, Phase#4, Hayatabad, Peshawar,Pakistan

H. DarainInstitute of Physical Medicine and RehabilitationSciences, Khyber Medical University, Hayatabad,Peshawar, Pakistan

H. ChenWHO Collaborating Center for Training andResearch in Rehabilitation, Tongji Hospital, TongjiMedical College, Huazhong University of Scienceand Technology, Wuhan, People’s Republic ofChina

Diabetes Ther (2021) 12:707–732

https://doi.org/10.1007/s13300-021-01005-1

Conclusion: Exercise plays an important role inoptimizing glycemic control and improvingquality of life (QoL), BMI, and waist circumfer-ence in type 2 DM patients. Exercise could be asafe adjunct therapy to medical treatments inthese patients.Registration: PROSPERO: CRD42020210816

Keywords: Diabetes mellitus, type 2; Exercise;Fasting blood glucose; Glycated hemoglobin;Meta-analysis; Physical activity; Systematicreview

Key Summary Points

Why carry out this study?

Exercise is considered a cornerstone in type 2DM management, but diabetes patients areincreasingly likely to lead a sedentarylifestyle.

This study investigated the effects ofdifferent exercise regimens on glycemiccontrol, anthropometric parameters, andquality of life for type 2 DM patients.

What was learned from the study?

Exercise significantly affects glycemiccontrol, induces beneficial anthropometricchanges, and enhances quality of life.

General physical activity and walking areconsidered low-impact activities, but theycan still benefit type 2 DM patients.

DIGITAL FEATURES

This article is published with digital features,including a summary slide, to facilitate under-standing of the article. To view digital featuresfor this article go to https://doi.org/10.6084/m9.figshare.13574684.

INTRODUCTION

Diabetes mellitus (DM), a very common medi-cal condition, is associated with decreased

production of insulin or an inability of the bodyto utilize it properly (types 1 and 2 DM,respectively) [1]. Type 2 DM is the most preva-lent form of diabetes, accounting for nearly90% of all cases of diabetes [2, 3]. According tostatistics from the International Diabetes Fed-eration (IDF), the estimated combined globalprevalence of type 1 and type 2 DM, bothdiagnosed and undiagnosed, was 463.0 millionadults aged 20–79 years in 2019. Based on this2019 estimation, a prevalence of 578.4 millionis projected by the year 2030, and if the presenttrend continues, 700.2 million adults aged20–79 years will be diabetic by 2045 [4]. Thereason behind this escalation of the prevalenceof diabetes could be sedentary lifestyles,increasing levels of obesity, aging populations,and diets that are high in calories and fat [5].

Body mass index (BMI) and waist hip ratio(WHR) are independent risk factors for type 2DM [6]. In developed countries, diabetes mostlyoccurs in the elderly, while in developingcountries, diabetes mostly affects people aged34–65 years [7]. It is often reported that a highproportion of patients with diabetes performonly a small amount of physical activity,including structured exercises [3]. Type 2 DM isa major cause of premature morbidity andmortality associated with cardiovascular disor-ders, kidney diseases, neuropathies, retinopa-thy, and amputation [5]. Along with othercomplications, diabetic patients may also sufferfrom depression, anxiety, denial, worries, andfrustration [1]. Poor health management ser-vices and a lack of awareness of diabetes canincrease diabetes-related mortality [8]. Diabetesmellitus has increased the socioeconomic bur-den on developing countries [9], and is consid-ered to be a prime reason for increasinghealthcare costs worldwide [10]. Diabetes drugsare expensive, and most patients fail to gethigh-quality treatment due to the expensivenature of the treatment, the poor allocation offunds for healthcare, and the lack of medicalreimbursement in developing countries [11].

Studies have shown that regular exercise candelay diabetes and associated complications,but it has been observed that most diabeticpatients are inactive. Generally speaking, dia-betic patients can attain a controlled level of

708 Diabetes Ther (2021) 12:707–732

blood sugar by adopting a healthy nutritionalplan, exercising regularly, and reducing bodyweight, although some may need to useexogenous insulin [12].

According to Gabriel and Zierath [13], theacute effects of exercise include enhanced glu-cose metabolism and improved insulin sensi-tivity. Blood glucose was found to be animportant substrate in the skeletal muscle oxi-dation process during strenuous exercise, andexercising increases glucose utilization withoutaffecting blood insulin levels. Exercise alsoincreases certain protein levels that are impor-tant for glucose homeostasis. The increase inthe level of glucose transporter type 4 protein(GLUT4) induced by exercise training enhancesskeletal muscle glucose uptake. This strategycould regulate blood glucose levels in the case ofinsulin resistance. Gabriel and Zierath alsonoted that performing exercise training forlonger increases the aerobic threshold and theconcentration of mitochondria in skeletalmuscles [13]. Exercise has long been consideredthe cornerstone of diabetes mellitus manage-ment, along with diet adjustment and drugtreatment. Aerobic exercise, which consists ofexercising large groups of muscles in a rhyth-mic, repeated, and continuous manner for atleast 10 min, is used in diabetes management.In strenuous short-duration exercise, musclesmainly use energy from the glycogen stores, butthe muscle glycogen stores are gradually deple-ted during long-duration exercise, so energy ismainly provided by glucose and free fatty acidsfrom adipose tissues [13, 14]. High-intensityexercise consumes glucose at a higher rate thanit is produced by the liver, so the blood sugarlevel drops [14]. It has been proven that exerciseincreases the sensitivity of muscle cells toinsulin and increases the activity of oxidativeenzymes [5]. Moreover, the overall effects ofexercise on carbohydrate metabolism are asso-ciated with improved insulin sensitivity [15].

The results of a randomized controlled trialof adults with type 2 diabetes that explored thebenefits of aerobic or resistance training aloneor combined indicated that both types of exer-cise training alone improved glycemic control,but the improvement was greatest with com-bined resistance and aerobic training [16].

The increase in glucose uptake by skeletalmuscles, enhanced insulin sensitivity, andimproved oxidative enzyme activity during exer-cise are the rationale for incorporating exercise asa therapeutic adjunct into the management oftype 2 DM. The purpose of this systematic reviewand meta-analysis was to investigate and quanti-tatively summarize the literature regarding thetherapeutic effects of exercise and general physi-cal activity on glycemic control, exercise-associ-ated anthropometric changes, and the quality oflife of type 2 DM patients.

METHODS

A systematic review and meta-analysis of ran-domized controlled trials was conductedaccording to the Preferred Reporting Items forSystematic Reviews and Meta-Analyses(PRISMA) 2009 guidelines [17]. This study wasregistered with PROSPERO (registration ID:CRD42020210816).

Study Objective

To explore and quantitatively summarize theliterature regarding the therapeutic effects ofexercise and general physical activity on gly-cemic control, exercise-associated anthropo-metric changes, and quality of life of type 2 DMpatients.

Study Question

Do exercise and general physical activity havebeneficial therapeutic effects on the glycemiccontrol, anthropometric parameters, and qual-ity of life of type 2 DM patients?

PICO

P (Population): Type 2 DM patientsI (Intervention): Exercise and general physical

activity or combinations of exercise with othertherapies

C (Comparison): Nonexercise/standard treat-ment/placebo, or different exercise regimengroups of type 2 DM patients

Diabetes Ther (2021) 12:707–732 709

O (Outcomes): HbA1c, fasting blood glucose(FBG), quality of life (QOL), BMI, weight, waistcircumference.

Search Strategy

The electronic databases AMED, Embase, HMIC,BNI, MEDLINE, PsycInfo, CINAHL, HEALTHBUSINESS, PubMed, Web of Science (WoS), andEBSCO were used to conduct a systematic liter-ature search. The limits applied were publica-tion year (2011–2020), English language studies,and human studies. Another method of identi-fying relevant studies was to screen the refer-ence lists of studies retrieved from thedatabases. We included studies up to the year2000 identified through screening the referencelists of the studies retrieved from the databases.A search alert was created for Web of Scienceand EBSCO to retrieve recently published stud-ies. The following keywords and MeSH termswere identified for every database individuallyand used during the searches: ‘‘physical activ-ity,’’ ‘‘exercise,’’ ‘‘diabetes mellitus,’’ ‘‘type 2.’’Another method of identifying relevant studieswas to screen the reference lists of the studiesretrieved from the databases.

Study Selection

Inclusion and Exclusion CriteriaThe study inclusion criteria were as follows: (1)randomized controlled trials (RCTs) that stud-ied the effects of exercise on glycemic control inpatients with type 2 DM and had a follow-upperiod of C 8 weeks; (2) participants of any ageand gender; and (3) quantitative measurementof glycemic control or anthropometric changes.The following studies were excluded: studieswith a diabetes mellitus type 1 population,single-arm studies, animal studies, ex vivo/in vitro studies, review studies, retrospectivestudies, and case studies.

Types of InterventionExercise, walking, or general physical activitycompared with the nonexercise control groupor the use of other treatments or exerciseinterventions. Studies combining other

therapies with exercise, walking, or physicalactivity were also included.

Types of ParticipantsStudies with type 2 DM patients were included;studies where all subjects were only at risk ofdiabetes were excluded.

Types of Outcome MeasuresThe primary study variable was HbA1c, andsecondary outcomes were FBG, QOL, BMI,bodyweight, and waist circumference.

Study Selection Process

The study selection process comprised threescreening phases: study title screening, abstractscreening, and, finally, full-text screening.There were two independent reviewers at eachscreening phase. The corresponding authorresolved any conflicts or disagreements betweenthe two reviewers. The search results wereuploaded to Rayyan QCRI (https://www.rayyan.qcri.org: Data Analytics, Qatar ComputingResearch Institute) for citation screening by twoindependent reviewers [18]. After screening thetitles, the abstracts of the selected titles werescreened. Finally, the full texts of the selectedstudies were screened in detail for relevancy.

Study Quality and Risk of Bias Assessment

Data extracted for external validity and report-ing quality assessment included study ID (firstauthor’s name along with the year and refer-ence), study design characteristics (data regard-ing experimental design and study groups,number of participants per group), participantdetails, intervention (types of exercise, inten-sity, frequency, and follow-up), control (controltype and control intervention details), primaryand secondary outcome measures, and infor-mation regarding adverse events.

The Physiotherapy Evidence Database(PEDro) scale was used for study quality assess-ment [19]. Two independent reviewers per-formed the study quality evaluation; anyconflicts/disagreements were resolved by com-paring the recorded data and through

710 Diabetes Ther (2021) 12:707–732

discussion with the corresponding author. ThePEDro scale has 11 quality assessment items (yesor no), of which 10 are used to calculate thefinal PEDro score (0–10). Sackett’s levels of evi-dence were used to simplify the study qualitylevels, and a study was considered to be level 1(higher quality) if it scored C 6 on the PEDroscale and level 2 if it scored\6 [20].

The risk of study bias was assessed by twoindependent reviewers using the Cochrane riskof bias (RoB 2) tool [21]. Every study was ratedwith the RoB 2 tool for selection bias, detectionbias, performance bias, outcome assessment,incomplete data, outcome reporting, and otherthreats to study validity identified by tworeviewers. The risk of bias was categorized aseither a low risk of bias (the domain was con-sidered adequate), a high risk of bias (thedomain was considered inadequate), or anunclear risk of bias (not enough information tomake a judgment about the bias). The RoB 2Excel tool was filled for individual studies, andthe results were entered into Cochrane ReviewManager 5.3 [22].

Data Extraction

For the study of primary and secondary vari-ables, postintervention means and standarddeviations (SD) as well as sample size detailswere retrieved by two independent reviewers.Any conflicts/disagreements were resolvedthrough discussion with the correspondingauthor and by comparing the recorded data.The Microsoft data extraction form includedthe following details of the included studies:study design, population details, target vari-ables, interventional procedures (study groups,exercise type, volume, and intensities), andstudy findings. The authors prepiloted the dataextraction form in some studies. Data wereextracted from the texts, tables, and figures ofthe included studies.

Data Synthesis

We performed both qualitative and quantitativeanalyses. The data extracted were summarizedin table form to effectively present the

characteristics and results of the included stud-ies. To study the primary and secondary vari-ables, a meta-analysis was performed using arandom-effects model.

Meta-Analysis

Quantitative analysis was performed for thestudy’s primary outcome measure (HbA1c) andthe secondary outcome measures (FBGs, BMI,weight, and waist circumference). The meansand standard deviations of outcomes from theexercise and control groups were comparedusing mean differences and a random-effectsmodel with a 95% confidence interval. The timepoints for data extraction regarding the primaryand secondary variables were not specified, butwe preferred time points of 8 weeks and above.For some studies where the standard deviations(SD) were not provided, we calculated SDs withan online calculator using the formula providedby Hozo et al. [23].

Effect sizes of 0.2, 0.5, and 0.8 were consid-ered small, moderate, and large, respectively[24]. I2 and the chi-square test were used toevaluate heterogeneity among the studies. Weconsidered an I2 value of 25% to indicate lowheterogeneity, 50% to represent moderateheterogeneity, and 75% to indicate highheterogeneity [25]. All the analyses were per-formed in Cochrane Review Manager 5.3 [22],and p\0.05 was taken to indicate statisticalsignificance.

Compliance with Ethics Guidelines

This study is based on previous studies; theauthors did not perform any clinical or pre-clinical experiment involving humans oranimals.

RESULTS

Study Selection

The research databases AMED, Embase, HMIC,BNI, MEDLINE, PsycInfo, CINAHL, HEALTHBUSINESS, PubMed, Web of Science (WoS), and

Diabetes Ther (2021) 12:707–732 711

EBSCO were used to conduct a systematic liter-ature search. The search retrieved 21,559 arti-cles. A total of 32 articles were also identifiedfrom the reference lists of other studies. 1448duplicates were identified using Rayyan QCRIand EndNote. Title and abstract screening bytwo independent reviewers resulted in theexclusion of a further 20,063 studies. Eightyarticles were deemed eligible for full-textscreening, while 48 were excluded for particularreasons. Finally, 32 articles were eligible forinclusion in the systematic review, and 26 inthe meta-analysis (see the PRISMA flow diagramin Fig. 1).

Study Characteristics

The average age calculated from 26 studies was57.36 years. Two studies recruited participantsin the age ranges 30–80 years [26] and 40–65years [27], respectively; another recruitedelderly patients above 50 years [28]. The averageexercise session was 45.15 min, while the aver-age follow-up duration was 21.94 weeks. Themean exercise frequency according to our find-ings was 3.25 days/week.

For all studies, the study population com-prised type 2 DM patients. Depending on thestudy, the participants were also obese [29],hypertensive [30], adults with coronary artery

Fig. 1 PRISMA flow diagram depicting the study selection process

712 Diabetes Ther (2021) 12:707–732

disease (CAD) [31], or newly diagnosed type 2DM patients [32], had received their diagnosis5–8 months earlier [26] or\5 years earlier [33],or were female participants only [34], sedentarypatients with 5 years of diagnosis [35], elderlypatients with peripheral neuropathy [28], oroverweight, dysregulated, and sedentarypatients [36]. Most of the studies recruited par-ticipants with HbA1c levels of 6.4–7.0[26, 30, 32, 37–44], followed by levels of C 6.5[27, 45, 46], 7.0–10.0 [34], 7.2–9.0 [28, 31,33, 47], 10.0–11.0 [48, 49], or 7.0–11.0 [36].

A variety of exercise interventions were uti-lized in the studies, but aerobic training (12studies) [27, 28, 30, 32, 39, 41, 44, 45, 49–52]and resistance training (RT) (14 studies)[27, 31, 36, 39–41, 44, 45, 47, 48, 52–55] werethe most common. More specifically, the exer-cises used in the studies were as follows: aerobicexercises: moderate-intensity continuous (MIC)training [30, 32], supervised structured aerobicexercise training (SSAET) [51]; anaerobic train-ing: 5 studies [32, 37, 55–57], high-intensityinterval training (HIIT) [30, 32, 37, 56], sprintinterval training (SIT) [32]; resistance training(RT): biodensity resistance exercise [40], pro-gressive resistance training (PRT) [53, 54];endurance training: 1 study [37]; combinedexercise training: combined exercise regimen[38, 44, 45, 47]; walking: [29, 31, 39, 43, 45, 46];general physical activity: [26, 31, 33, 42, 44];nonlocomotive physical activity [42]; yoga:[34, 43]; and nonspecific exercises [35].

The common outcome measures used bymost of the studies were as follows: HbA1c (allstudies), FBG (20 studies) [26–29, 32, 35–37,40, 42–44, 47, 48, 50–52, 56–58], BMI (17studies) [27, 29, 30, 34–38, 44, 45, 48, 50,51, 54–57], bodyweight (13 studies)[26, 27, 31, 33, 35, 36, 42, 44, 45, 48, 54, 56, 57],waist circumference (11 studies) [26, 29, 33,35, 41, 44, 45, 47, 48, 53, 54, 56], and QoL (5studies) [33, 36, 43, 48, 54].

The durations of interventions reported bythe included studies were as follows: 8–12 weeks[27, 28, 30, 35, 37, 38, 42, 43, 48, 55–57], 16–-25 weeks [36, 37, 44, 50, 51, 53], and 6 months

to 1 year [26, 29, 31, 39–41, 46, 52]. Three daysper week (3d/wk) [30, 32, 35–38, 41, 45–48,51, 53–55, 57] was the most common exercisefrequency, followed by 5d/wk, [39, 50, 56], 1d/wk [59], and 2d/wk [31, 34, 43], daily [29], 7visits/3 month [44], and 1d/5 weeks [44]. Mostof the studies were consistent in the duration ofan exercise session: 5–10 min [40], 20–45 min[30, 32, 35, 37, 38, 44, 47, 49, 50], 45–60 min[27, 31, 34, 36, 39, 43, 51, 55, 56], 270 min/wk[41], or 150 min/wk [33, 45, 52].

Progression and increasing intensity ofexercise were reported by 12 studies as follows:based on an increasing number of steps at eachvisit [29]; based on monitoring body weight andmaximum oxygen uptake (VO2 max) everythree months, progression from 70% to 80%HRR was observed [47]; based on increasedexercise intensity with every session [32]; basedon increasing the mean treadmill grade andMETs progressively [41]; based on a progressiveincrease in resistance [53, 56]; based on anindividualized elevation of exercise intensity(14–16 on the Borg scale) at week 12 [48]; basedon a 30-min increase in each phase of a five-phase program [51]; based on increasing non-locomotive physical activity [42]; based onprogression in the completion of 12 repetitionsduring the final set of exercises in two consec-utive sessions [45]; and based on the progres-sion of low-volume HIIT to high-volume HIITover 6–12 weeks [55]. A detailed description ofthe study characteristics is available in Table 1.

Study Quality and Risk of Bias

The selection, performance, and detection bia-ses were similar for most of the studies (Fig. 2,3). The quality and level of evidence of eachstudy are summarized in Table 2. The majority(26) of the studies were of high quality (level 1),but 6 studies were of lower quality (level 2)[33, 39, 42, 44, 45, 57].

Diabetes Ther (2021) 12:707–732 713

Table

1Characteristics

ofthestudiesincluded

inthisreview

Stud

y

no.

Reference

Participant

inform

ation

Group

sExercise

frequency/du

ration

/

volume/intensity/mod

e

Progression

Adverse

events

Pvaluea

Con

clusion

1Abdelbasset

etal.(2020)

[30]

Samplesize

=48

G1n=16,G

2n=16,

G3n=16

Sex:M

=27,F

=21

Age

=40–6

0years

G1.

MIC

G2.

HIIT

G3.

Control

40min

exercise

3times/

weekfor8weeks

Not

reported

Not

reported

\0.05

BothMIC

andHIIT

groups

show

eda

significant

decrease

in

BMI,IH

TG,visceral

lipids,ALT,H

bA1c,

andlipid

profile

2Brinkmann

etal.(2019)

[38]

Samplesize

=30

G1n=15,G

2n=15

Sex:M

=19,F

=11

Age

=60

±8years

G1.

Trainingin

the

overnight-fasted

state

G2.

Trainingin

thefed

state

30min

combined

endurance/strength

training

program

3

times/w

eekfor

8weeks

Not

reported

Not

reported

0.001

Exercisingin

either

the

fasted

orfedstate

resultedin

improvem

entsin

the

subjects’p

hysical

fitness,b

ody

composition,glucose,

andfastingtriglyceride

values

3Dasguptaet

al.

(2017)

[29]

Samplesize

=347

G1n=174,G2n

=173

Sex:M

=157,

F=190

Age

=60

±11

years

G1.3000

steps/dayover

a

year

G2.30–6

0min

ofactivity

daily

3000

steps/dayover

1year,3

0–60

min

of

activity

daily

Num

berof

steps

increased

witheach

visit

Not

reported

\0.05

HBA1c

andinsulin

sensitivitysignificantly

improved

4Magalhaesetal.

(2019)

[47]

Samplesize

=80

G1n=28,G

2n=25,

G3n=27

Sex:M

=38,F

=42

Age

=58.5

±7.7years

G1.

MCT

withRT

G2.

HIITwithRT

G3.

Control

group

5.0±

7.1min

MCT

group,33.1

±6.4min

HIITgroup3times/

wk

Exercisesessions

were

upgraded

every

3monthsfor

VO2max

Not

reported

[0.05

Nosignificant

effectson

glycem

iccontrolwere

reported,b

utthe

combination

ofMCT

withRTim

proved

body

composition

and

CRF

714 Diabetes Ther (2021) 12:707–732

Table

1continued

Stud

y

no.

Reference

Participant

inform

ation

Group

sExercise

frequency/du

ration

/

volume/intensity/mod

e

Progression

Adverse

events

Pvaluea

Con

clusion

5Vanroyet

al.

(2017)

[39]

Samplesize

=48,

G1n

=27,G

2n

=21

Sex:M

=27,F

=21

Age

=59.4

±8.2years

G1.

IC

G2.

CC

60min

walking,aerobic

andstrength

training

for8weeks

Not

reported

Not

reported

\0.05

HbA

1clevelsdecreased

significantly

6Winding

etal.

(2018)

[37]

Samplesize

=32

G1n=12,G

2n=13,

G3n=7

Sex:M

=19,F

=13

Age

=58

±8years

G1.

END

G2.

HIIT

G3.

Control

group

40min

END

or20

min

HIIT,3

days/w

k

Not

reported

Not

reported

\0.05

Fastingglucose,HbA

1c

levels,

andglycem

ic

variability

werereduced

intheHIITgroup

7Halvariet

al.

(2017)

[31]

Samplesize

=108

G1n=57,G

2n=51

Sex:M

=88,F

=20

Age

=63.1

±7.9years

G1.

Physicalactivity

G2.

Non-physicalactivity

60min

walking,

swim

ming,bicycling,

resistance

training

for

2days/w

k

Not

reported

Not

reported

0.001

Physicalactivity

intervention

was

effectivein

decreasing

glucoselevelsrelative

to

anon-PA

controlgroup

8Hangpinget

al.

(2019)

[40]

Samplesize

=300

G1n=200,

G2n=100

Sex:M

=122,

F=178

Age

=50–7

0Years

G1.

BioDensity

resistance

exercise

G2.

Control

group

5-10

min

for1day/wk

Not

reported

Not

reported

\0.05

BioDensity

resistance

training

device

lowered

fastingplasmaglucose

andHbA

1c

9Heiskanen

etal.(2017)

[32]

Samplesize

=26

G1n=13,G

2n=13

Sex:M

=16,F

=10

Age

=40–5

5Years

G1.

SIT

G2.

MIC

T

6sessions

in2weeks,

SIT:4–

69

30sof

all-o

utcycling,MIC

T:

40–6

0min

atthe

intensityof

60%

of

peak

workload

Boutsand

duration

were

increased

witheach

session

Not

reported

0.002

Fastingglucose,insulin

,

and2hglucose

remainedun

altered,but

HbA

1cdecreasedafter

training

Diabetes Ther (2021) 12:707–732 715

Table

1continued

Stud

y

no.

Reference

Participant

inform

ation

Group

sExercise

frequency/du

ration

/

volume/intensity/mod

e

Progression

Adverse

events

Pvaluea

Con

clusion

10Akinciet

al.

(2018)

[27]

Samplesize

=65

G1n=22,G

2n=21,

G3n=22

Sex:M

=18,F

=47

Age

=40–6

5years

G1.

Control

group

G2.

Supervised

exercise

group

G3.

Internet-based

exercise

50–6

0min

ofphysical

activity,exercise,

aerobicandresistance

3days/w

kfor8weeks

Not

reported

Not

reported

0.003

Supervised

group-based

andInternet-based

exercise

cansimilarly

improveglycem

ic

control,waist

circum

ference,and

qualityof

life

11Alonso-

Dom

ınguez

etal.(2019)

[46]

Samplesize

=204

G1n=102,

G2n=102

Sex:M

=111,

F=93

Age

=25–7

0years

G1.

Control

G2.

Intervention

(food

workshop,

5walks,

smartphone

application)

Onceaweekfor5weeks,

4km

walk

Not

reported

Not

reported

0.241

Improvem

entsin

postprandialglucose,

waistcircum

ference,

andsystolicBPwere

noted,

butthey

didnot

reachstatistical

significance(P

\0.05)

12DiLoretoet

al.

(2005)

[44]

Samplesize

=182

G1n=28,G

2n=27,

G3n=31,G

4n=27,

G5n=32,G

6n=34

Sex:M

=88,F

=94

Age

=62

±0.7years

Counselingsessionwith

atleast30

min

of

physicalactivity,

followed

by15-m

in

sessions

every3months

for2years

Physicalactivity

Not

reported

Not

reported

\0.05

Bodyw

eight,waist

circum

ference,heart

rate,fasting

plasma

glucose,serum

LDL,

HDLcholesterol,and

HbA

1cwereim

proved

bytheintervention

13Churchet

al.

(2010)

[ 41]

Samplesize

=262

G1n=41,G

2n=73,

G3n=72,G

4n=76

Sex:M

=97,F

=165

Age

=55.8

years

G1.

Control

G2.

Resistanceexercise

G3.

Aerobicexercise

G4.

Bothresistance

and

aerobicexercise

Resistancetraining

for

3days/w

k

Duringweeks

12and24,

exercise

dose

was

reduced

byone-third

Not

reported

0.03

Acombination

ofaerobic

andresistance

training

improved

HbA

1clevels

716 Diabetes Ther (2021) 12:707–732

Table

1continued

Stud

y

no.

Reference

Participant

inform

ation

Group

sExercise

frequency/du

ration

/

volume/intensity/mod

e

Progression

Adverse

events

Pvaluea

Con

clusion

14Balducciet

al.

(2012)

[52]

Samplesize

=606

G1n=303,

G2n=303

Sex:M

=351,

F=254

Age

=58.8years

G1. Exercise?

coun

selling

group

G2.

Counsellin

galone

Exercise150min/w

kin

2sessions

ofmixed

aerobicandresistance

Not

reported

Not

reported

0.002

Physicalactivity

improved

HbA

1c,B

MI,andwaist

circum

ferencein

subjectswithtype

2

diabetes

15Castaneda

etal.

(2002)

[53]

Samplesize

=62

G1n=31,G

2n

=31

Sex:M

=22,F

=40

Age

=66

±8years

G1.

PRT(progressive

resistance

training)

G2.

Control

group

45min

sessionfor

3days/w

k

Not

reported

3incidentsof

chestpain

in

subjectswith

CAD

\0.05

PRTwas

effectiveat

improvingglycem

ic

controland

someof

the

abnorm

alitiesassociated

withmetabolic

synd

romein

type

2

diabetes

adults

16Earnestet

al.

(2014)

[45]

Samplesize

=262

G1n=41,G

2n=72,

G3n=73,G

4n=76

Sex:M

=97,F

=165

Age

=30–7

5years

G1.Control

G2.

Aerobic

G3.

Resistancetraining

G4.

Aerobic?

resistance

training

Treadmill

for3–

5days/

wk,

10–1

2repetitions;

AER?

RES=2

sessions

ofRES/wk,

10–1

2repetitions

Not

reported

Not

reported

\0.02

Aerobicandaerobicplus

resistance

training

significantlyim

proved

metabolicsynd

rome

scores

inpatientswith

T2D

M

17And

rewset

al.

(2011)

[26]

Samplesize

=593

G1n=99,G

2n=248,

G3n=246

Sex:M

=385,

F=208

Age

=30–8

0years

G1.

Control

G2.

Intensivedietary

support

G3.

Intensivedietary

supportandactivity

30min

briskwalk

5days/w

k

Increasedover

5weeks

then

maintained

Not

reported

0�013

Anintensivediet

intervention

soon

after

diagnosisim

proved

glyaem

iccontrol;the

addition

ofan

activity

intervention

conferred

noadditionalbenefit

18Cox

etal.

(2016)

[33]

Samplesize

=47

G1n=21,G

2n=18

Sex:M

=18,F

=21

Age

=24–8

0years

G1.

Routine

care

G2.

GEM

intervention

Routine

activitiesof

daily

livingandfood

choices

Not

reported

Not

reported

0.01

GEM

intervention

ledto

significant

improvem

entsin

HbA

1c

Diabetes Ther (2021) 12:707–732 717

Table

1continued

Stud

y

no.

Reference

Participant

inform

ation

Group

sExercise

frequency/du

ration

/

volume/intensity/mod

e

Progression

Adverse

events

Pvaluea

Con

clusion

19Hem

mati

(2017)

[60]

Samplesize

=90

G1n=30,G

2n=30,

G3n

=30

Sex:M

=51,F

=39

Age

=50.6

years

G1.

FTF

G2.

TEG

G3.

CTG

30min

ofmoderate-to-

vigorous

intensity

aerobicexercise

Initially

twice/month

then

once/m

onth

for2months

Not

reported

0.21

Improvem

entsin

BMI,

HbA

1c,F

BS,

cholesterol,and

triglyceride

werenoted,

butwerenot

statistically

significant

20Hsieh

etal.

(2018)

[48]

Samplesize

=30

G1n=15,G

2n=15

Sex:M

=11,F

=19

Age

=35–6

5years

G1.

RTA

G2.

CG

RT

2times/w

k,walking

orcycling(60–

80%

aerobiccapacity)5

times/w

k

Not

reported

Not

reported

0.407

Supervised

RTexercises

improved

waist

circum

ference,fasting

glucoselevels,

andpeak

diastolic

bloodpressure;

nosignificant

changes

inoverallQoL

21Miyam

otoetal.

(2017)

[42]

Samplesize

=32

G1n=10,G

2n=12,

G3n=10

Age

=60.0

±3.1years

G1.

LPA

G2.

NLPA

G3.

Control

group

Weartriaxial

accelerometer

and

increaseLPA

byverbal

instructionfrom

PT

for12

weeks

Not

reported

Not

reported

0.684

There

weremarked

improvem

entsin

both

HbA

1candinsulin

sensitivity

22Mensberget

al.

(2016)

[36]

Samplesize

=34

G1n=18,G

2n=16

Sex:F=10,M

=24

Age

=56.5

years

G1.

GLP-1R

A

G2.

Placebogroup(PG)

Supervisespinning

and

resistance

exercises1h

3times/w

kfor16

weeks

Not

reported

Not

reported

\0.001

Acombination

of

supervised

exercise

training

andliraglutide

inducedmajor

improvem

entsin

glycem

iccontrol

718 Diabetes Ther (2021) 12:707–732

Table

1continued

Stud

y

no.

Reference

Participant

inform

ation

Group

sExercise

frequency/du

ration

/

volume/intensity/mod

e

Progression

Adverse

events

Pvaluea

Con

clusion

23Jibril(2017)

[35]

Samplesize

=60

G1n=30,G

2n=30

Sex:M

=28,F

=32

Age

=65.3

±8years

G1.

RE

G2.

CG

REfor12

weeks

of

20min

REthrice

or

threetimes/w

k

Not

reported

Not

reported

[0.05

The

PAintervention

asa

wholedidnotproduce

theexpected

effectsin

patientswithT2D

M,

buttimeeffectssuggest

that

certainaspectsare

effective

24Shakil-ur-

Rehman

etal.(2018)

[51]

Samplesize

=102

G1n=51

G2n=51

Sex:M

=55,F

=47

Age

=58

±8years

G1.

SSAET

G2.

CG

25weeks

ofSSAET

exercises,routine

medication,

and

dietaryplan

Not

reported

Not

reported

P\

0.05

SSAETwas

equally

effective,irrespective

of

gend

er,ascomparedto

thenonexercisecontrol

groupin

T2D

M

patients

25Sreedeviet

al.

(2017)

[34]

Samplesize

=124

G1n=41,G

2n=42,

G3n=41

Age

=30–6

5years

G1.

YG

G2.

PIG

G3.

CG

Yogafor60

min

2days/

wk,peer

mentorship

of13–1

4T2D

M

patientswith45-to

60-m

ineducational

classon

diabetes/w

k

Not

reported

Not

reported

0.5

Alonger

follow-upis

needed

toobservethe

effectsof

yoga

andpeer

mentorshipon

glycem

ic

control

26Cassidy

etal.

(2019)

[57]

Samplesize

=22

G1n=11,G

2n=11

Sex:M

=17,F

=5

Age

=60

±2years

G1.

HIIT

G2.

Control

group

36cyclesessions

over

12weeks,3

sessions/

wk

High-intensity

intervals

increasedby

10seach

week

Not

reported

0.03

HIITim

proved

glycem

ic

control

27Dixitet

al.

(2017)

[28]

Samplesize

=87

G1n=47,G

2n=40

Sex:M

=53,F

=34

Age

=50

years

G1.

Moderate-intensity

aerobicexercise

G2.

Control

group

Moderate-intensity

aerobicexercise

for8

weeks

Not

reported

Not

reported

\0.001

Moderate-intensity

aerobicexercise

enhanced

glycem

ic

control

Diabetes Ther (2021) 12:707–732 719

Table

1continued

Stud

y

no.

Reference

Participant

inform

ation

Group

sExercise

frequency/du

ration

/

volume/intensity/mod

e

Progression

Adverse

events

Pvaluea

Con

clusion

28Fayh et

al.(2018)

[49]

Samplesize

=30

G1n=15,G

2n=15

Sex:M

=12,F

=18

Age

=50.57±

6.38

years

G1.

Placeboplus

exercise

G2.

n-3PU

FAcapsules

plus

exercise

High-intensityexercise,

30min,twosessions

during

thestudy

Exerciseun

til

exhaustion

Not

reported

0.752

PUFA

n-3

supplementation

reducedtriglyceride

as

wellas

TRAPlevels

afterexercise,b

utits

effectson

HbA

1c,

inflammatory,and

oxidativestressmarkers

werenotsignificant

29Francoiset

al.

(2018)

[55]

Samplesize

=53

G1n=18,G

2n=16,

G3n=19

Sex:M

=19,F

=34

Age

=56

±9years

G1.

Milk

?HIIT

G2.

Macronu

trient

control?

HIIT

G3.

Placebo?

HIIT

12weeks

ofcardio

and

resistance-based

HIIT

thrice

weekly

Not

reported

Not

reported

\0.01

HIIT

stronglyim

proved

HBA1c

and

cardiovascular

risk

factors,so

thereis

potentialforreduced

cardiovascular

mortality

andmorbidity

30Bocket

al.

(2019)

[43]

Samplesize

=48

G1n=24,G

2n=24

Sex:M

=18,F

=30

Age

=32–7

4years

G1.

Lyengar

yoga

G2.

Program

ofstandard

exercise

Two60-m

insessions

weeklyfor12

weeks,

moderate-intensity

exercise

Not

reported

Musculoskeletal

pain

observed

\0.05

Yogaintervention

produced

improvem

entsin

blood

glucoseand

psychosocialmeasures

ofdiabetes

managem

ent

31Wycherley

etal.

(2014)

[54]

Samplesize

=84

G1n=33,G

2n=51

Sex:M

=49,F

=35

Age

=55.6

G1.Calorie-restricteddiet

CRD

G2.

CRD

?exercise

Moderateenergy-

restricted

diet

for16

weeks,8

weeks

resistance

exercise

3days/w

k

Not

reported

Not

reported

0.04

Lifestylemodification

and

participationin

structured

weight-loss

programswere

beneficialforthe

managem

entof

type

2

diabetes

720 Diabetes Ther (2021) 12:707–732

Table

1continued

Stud

y

no.

Reference

Participant

inform

ation

Group

sExercise

frequency/du

ration

/

volume/intensity/mod

e

Progression

Adverse

events

Pvaluea

Con

clusion

32Zhang

etal.

(2017)

[56]

Samplesize

=32

G1n=17,G

2n=15

Sex:M

=15,F

=17

G1.

Exercisegroup

G2.

CG

60-m

inaerobicbicycle

training

5times/w

k

for12

weeks

Not

reported

Not

reported

0.099

Aerobicexercise

significantlydecreased

serum

fetuin-A

levelsin

type

2diabetesmellitus.

There

was

no

significant

difference

in

fastingglucose,OGTT

2-hglucose,HbA

1c,

fastinginsulin

,and

HOMA-IRlevels

betweentheexercise

andcontrolgroups

Tableshow

sthecharacteristicsof

thestudiesincluded

inthissystem

aticandmeta-analysisreview

T2D

Mtype

2diabetes

mellitus,HIIT

high-in

tensityinterval

training,RT

resistance

training,MCT

continuous

moderate-intensitytraining,END

endurance,

MIC

moderate-intensity

continuous,SIT

sprint

intervaltraining,M

ICTmoderate-intensitycontinuous

training,w

kweek,M

male,Ffemale,nsamplesize,F

TFface-to-face

group,TEGtelephoniceducationgroup,

CTG

controlgroup,RTG

resistance

training

group,GLP-1R

Aglucagon-like

peptide-1receptor

agonist,REreboun

dexercise,SSA

Esupervised

structured

aerobicexercise

training,Y

Gyoga

group,PIGpeerintervention

group,CRDcalorie-restricted

diet,PAphysicalactivity,L

PAlocomotivephysicalactivity,N

PLAnonlocom

otivephysicalactivity,T

RAPtotalreactiveantioxidant

potential

aPvalues

areonlyforglycem

iccontrolvariables(H

bA1c)

Diabetes Ther (2021) 12:707–732 721

META-ANALYSIS

Outcomes

Effects of Exercise Intervention on GlycemicControlPostexercise Improvements in Glycated Hemo-globin (HbA1c) A total of 26 studies wereanalyzed in this segment. There were significantdifferences between the exercise and controlgroups regarding HbA1c improvement afterexercise intervention (Fig. 4a, P\0.0001,z = 4.28, I2 = 81%) [26–34, 37–42, 44, 46–49,51, 53, 54, 56, 57, 60]. Similarly, when the fourstudies [29, 32, 38, 49] with no nonexercisegroups were excluded, significant improve-ments in HbA1c were observed after exerciseintervention (Fig. 4b, P\ 0.0001, z = 3.81,I2 = 83%).

Postexercise Improvements in Fasting BloodGlucose (FBG) Fasting blood glucose wasreported by 13 studies. Analysis of FBG showeda significant difference between the exerciseand nonexercise groups (Fig. 4c, P = 0.03,

z = 2.21, I2 = 99%) [26, 27, 34, 35, 37, 42, 44,47, 48, 51, 56, 57].

Anthropometric ChangesEffect of Exercise Intervention on Body MassIndex (BMI) A comparison of 13 studies thatreported BMI changes after exercise interven-tion showed that BMI improved significantlywith exercise (Fig. 5a, P = 0.04, z = 2.04,I2 = 91%) [26, 29, 30, 34, 35, 41, 44, 47, 51, 54,56, 57, 60].

Effect of Exercise Intervention on BodyWeight A total of 11 studies were analyzed toassess the weight reduction after exercise inter-vention. No significant improvement in weightreduction was achieved following exerciseintervention, according to random-effectsmodels (Fig. 5b, P = 0.11, z = 1.60, I2 = 78%),but significant improvement was seen with afixed-effects model (P = 0.02) [26, 29, 31, 33,35, 42, 44, 48, 54, 56, 57].

Effect of Exercise Intervention on Waist Cir-cumference Ten studies reported waist cir-cumference changes following exercise

Fig. 2 Risk of bias graph. The review authors’ judgments about each risk of bias item are presented as percentages across allincluded studies

722 Diabetes Ther (2021) 12:707–732

intervention. A significant difference wasobserved among the studies regardingimprovement in waist circumference afterexercise intervention (Fig. 5c, P = 0.007,

z = 2.71, I2 = 91%) [26, 29, 33, 35, 41, 44,47, 48, 53, 56].

DISCUSSION

This systematic review and meta-analysis aimedto quantitatively examine the evidence regard-ing the therapeutic effects of exercise and gen-eral physical activity on glycemic control,exercise-associated anthropometric changes,and QoL in type 2 DM patients. Thirty-twostudies with 2140 participants who underwent awide variety of exercise interventions wereincluded in this systematic review and meta-analysis. The main findings of this study arethat various types of exercise can improve gly-cemic control and quality of life and can inducesignificant anthropometric changes in type 2DM patients.

Glycemic Control

Our meta-analysis of 26 studies [26–31, 33, 34,38–42, 44, 46–49, 51, 53, 54, 56, 57, 60] thatreported patient HbA1c levels revealed thatexercise intervention significantly improvedpatient HbA1c levels. However, these findingscould be limited by variation in the HbA1c levelset as an enrollment criterion in the includedstudies. This enrollment criterion was eitherHbA1c C 6.5 [27, 43–45], HbA1c 7–10 [34],HbA1c 6.4–7 [26, 30, 32, 37–44], HbA1c 7.2–9[28, 31, 33, 47], HbA1c 10–11 [36, 48, 49], orHbA1c 6.5–11 [27, 36, 45]. We can assume thatthere is a greater chance of a decrease in HbA1cin a population with higher HbA1c levels thanin a population with lower HbA1c levels. Simi-larly, a significant reduction in FBG wasobserved in this study, further verifying thetherapeutic efficacy of exercise.

Anthropometric Changes

In this study, significant improvements in BMI[26, 29, 30, 34, 35, 41, 44, 47, 50, 51, 54, 56, 57]and waist circumference [26, 29, 33, 35, 41, 44,47, 48, 53, 56] were observed with exerciseintervention. These anthropometric changes

Fig. 3 Risk of bias summary table that presents theauthors’ judgments about each risk of bias item for eachincluded study

Diabetes Ther (2021) 12:707–732 723

Table

2The

qualityandlevelof

evidence

ofeach

study

Stud

yCom

pliancewithPEDro

scalecriteria

Total

PEDro

scalescore/

evidence

level

Criterion

1

Criterion

2

Criterion

3

Criterion

4

Criterion

5

Criterion

6

Criterion

7

Criterion

8

Criterion

9

Criterion

10

Criterion

11

Abdelbasset

etal.(2020)

[30]

YY

YY

NY

YY

YY

Y9/level1

Brinkmann

etal.(2019)

[38]

YY

NY

NN

NY

YY

Y6/level1

Dasguptaet

al.

(2017)

[29]

YY

NY

YY

NN

NY

Y6/level1

Magalhaesetal.

(2019)

[47]

YY

NY

NN

NY

YY

Y6/level1

Vanroyet

al.

(2017)

[39]

YY

NY

NN

NY

NY

Y5/level2

Winding

etal.

(2018)

[37]

YY

YY

NN

NY

NY

Y6/level1

Halvariet

al.

(2017)

[31]

YY

YY

NN

NY

YY

Y7/level1

Hangpinget

al.

(2019)

[40]

YY

NY

NY

YY

YY

Y8/level1

Heiskanen

etal.(2017)

[32]

YY

NY

NN

YY

YY

Y7/level1

Akinciet

al.

(2018)

[27]

YY

YY

YN

NY

YY

Y8/level1

Alonso-

Dom

ınguez

etal.(2019)

[46]

YY

YY

NN

YY

NY

Y7/level1

724 Diabetes Ther (2021) 12:707–732

Table

2continued

Stud

yCom

pliancewithPEDro

scalecriteria

Total

PEDro

scalescore/

evidence

level

Criterion

1

Criterion

2

Criterion

3

Criterion

4

Criterion

5

Criterion

6

Criterion

7

Criterion

8

Criterion

9

Criterion

10

Criterion

11

DiLoretoet

al.

(2005)

[44]

YY

NY

NN

NY

NY

Y5/level2

Churchet

al.

(2010)

[41]

YY

NY

NN

NY

YY

Y6/level1

Balducciet

al.

(2012)

[52]

YY

NY

NN

YY

NY

Y6/level1

Castaneda

etal.

(2002)

[53]

YY

NY

NN

NY

YY

Y6/level1

Earnestet

al.

(2014)

[45]

YY

NY

NN

NY

NY

Y5/level2

The

tableshow

swhich

ofthe11

criteriaof

thePE

Dro

scaleweremet(Y)or

notm

et(N

)by

each

study;thetotalP

EDro

scalescorecorrespond

stothetotalnum

berof

thecriteriathat

weremet

bythestudy(notethat

criterion1was

excluded

from

thescoring,so

themaxim

umpossiblescorewas

10)

Diabetes Ther (2021) 12:707–732 725

726 Diabetes Ther (2021) 12:707–732

could maximize the wellbeing of diabetespatients and reduce the likelihood of obesity- orweight-gain-associated complications. Most ofthe studies reported a decrease in body weightafter the intervention; however, the findings ofthe meta-analysis were nonsignificant. We

hypothesize that weight reduction needs vig-orous exercises and long-term follow-up.

Quality of Life

Exercise improves wellbeing and enhances thequality of life. Two studies showed that exerciseled to improved psychological functioning [33],significant improvements in overall QoL, sexu-ality, mobility, anxiety and worry, diabetescontrol, and severity, and no significant change

bFig. 4 Glycemic control. a HbA1c exercise vs controlgroups. b HbA1c exercise vs nonexercise control groups.c Fasting blood glucose (FBG) exercise vs control groups

Fig. 5 Anthropometric changes due to exercise intervention. a Improvements in BMI (body mass index). b Improvementsin body weight. c Reduction in waist circumference

Diabetes Ther (2021) 12:707–732 727

in the social burden domain (D-39) [54]. A studyreported similar increases in QoL in two groups,one receiving liraglutide and exercise interven-tion and the other receiving placebo and exer-cise intervention. Despite some liraglutidetreatment-associated side effects, the QoL in theexercise plus liraglutide group was not inferior.These findings indicate the importance ofexercise in improving QoL [36]. Another studyreported that yoga led to a greater increase inQoL than standard exercise, but there was nosignificant between-group difference in QoL[43]. One study reported that the change in QoLfollowing exercise intervention was nonsignifi-cant; it was argued that this result was obtainedbecause the QoL questionnaire—the Audit ofDiabetes Dependent Quality of Life(ADDQoL)—may not have been appropriate forthe population studied [48]. However, onlythree of the studies that reported QoL had anonexercise group. When the study only com-pares exercise groups, all of the participantsmay undergo similar improvements in QoL. Acomparison between nonexercise and exercisegroups could offer greater insight into theeffects of exercise on QoL [33, 48, 54].

Different Exercise Regimens and TheirEffects on Blood Sugar and BodyComposition

Nutrition Plus ExerciseInterestingly, it was observed that nutritionalstatus (fasted overnight or fed) is not animportant factor in exercise interventions toimprove glycemic control in type 2 DM patients[38]. Diet plus exercise leads to a significantweight reduction [54]; moreover, glucose levelscan be strongly controlled by following anintensive diet plan after being diagnosed withtype 2 DM, such that the addition of physicalactivity (PA) may produce no additional bene-fits [26]. The GEM (glycemic load, exercise, andmonitoring blood glucose) program is uniqueand effective at modifying lifestyle and con-trolling HbA1c in diabetic patients [33].

Anaerobic TrainingEvidence has shown that HIIT significantlyimproves glycemic control (HbA1c) [30, 55, 57]and BMI [30]. Unsupervised HIIT [57] and HIITeven in lower training volume than enduranceexercise, has better effects than enduranceexercise on fitness and glycemic control [37].However, one study reported that SIT may beless beneficial for a diabetic heart than MICT,but MICT may reduce the HbA1c level morethan SIT [32]. Another study added that HIITwith RT did not affect glycemic control; how-ever, a year of combined MCIT and RT inter-vention improved body composition [47].

Aerobic TrainingAerobic exercise plays a key role in optimizingblood glucose and inducing anthropometricchanges. Different types of aerobic traininghave been shown to be effective in type 2 DMpatients. MICT significantly improves HbA1c[28, 30, 56], BMI [30], and body composition[47], and assists with weight loss [56]. Super-vised structured aerobic exercise training(SSAET) is effective at reducing BMI, FBG, andHbA1c [51]. Additionally, a study reported thatthe full benefits of aerobic exercise could beachieved with an energy expenditure of morethan 20 METs/hr/week [44].

Resistance TrainingResistance exercises are of paramount impor-tance to weight loss and glycemic control.Studies included in this review revealed signifi-cant glycemic control and anthropometricchanges [48] with resistance training (RT)interventions such as bioDensity resistancetraining [59], progressive resistance training(PRT) [53], and supervised and structured RT[48].

Combined Exercise TrainingCombining different exercise regimens couldresult in additional benefits. A combined aero-bic and RT approach was found to be better forHbA1c control than aerobic or resistance train-ing alone [41]. Moreover, aerobic and resistancetraining with exercise counseling significantlyimproved fitness, HbA1c, and weight loss [52].

728 Diabetes Ther (2021) 12:707–732

Supervised RT with HIIT exercise training andliraglutide induced major improvements inglycemic control along with robust weight loss[36]. Another study showed that a year of MCTwith RT improved body composition [47].

YogaA limited number of studies used yoga as anintervention for type 2 DM. Findings regardingthe effects of yoga on glycemic control are notconclusive due to the limited number of studiesthat included yoga. In one of the studies, yogasignificantly improved HbA1c as compared tostandard exercise [43], while it had incrementaleffects on glycemic control in another study[34]. Long-term studies of the benefit of yoga intype 2 DM are warranted.

Walking and Physical Activity (PA)General physical activity is generally consideredto be of less value than exercise interventionwhen it comes to weight loss and glycemiccontrol in type 2 DM patients. However, thefindings of the studies in the present review inrelation to general physical activity are inter-esting, with most of them showing that suchactivity improved glycemic control [29, 31, 39].Miyamoto et al. reported that nonlocomotivephysical activity (N-LPA) did not affect glucoseand fat levels [42].

A study reported that rebound exercise couldbe beneficial in type 2 DM [35]. Moreover, asupervised group-based or Internet-based exer-cise is more beneficial than simple counselingfor achieving glycemic control, reducing thewaist circumference, and improving quality oflife [27].

Limitations

A limitation of this study is the inability togroup the findings of different studies accordingto the exercise regimes used, such as resistancetraining, aerobic training, high-intensity inter-val training, and moderate-intensity continu-ous training. An ability to group the findings inthis manner would provide readers with a betterunderstating of the effects of a variety of exer-cise regimens. However, we have partly

mitigated this limitation by discussing eachexercise regimen separately.

CONCLUSIONS

Exercise is the cornerstone to good Glycemiccontrol in type 2 DM patients. Exercise trainingcan lead to significant anthropometric changesand enhance the quality of life. An exercise planinvolving 40- to 60-min sessions 3–5 days perweek or 150 min of exercise per week can sig-nificantly improve glycemic control, BMI, andwaist circumference. However, according to ourfindings, structured exercise performed as rarelyas 1–2 days per week can also reduce bloodglucose. Although considered to be less intenseactivities, general physical activity and walkingcould still help to reduce blood glucose.

ACKNOWLEDGEMENTS

Funding. No funding or sponsorship wasreceived for this study or the publication of thisarticle. The Rapid Service Fee was funded by theauthors.

Authorship. All named authors meet theInternational Committee of Medical JournalEditors (ICMJE) criteria for authorship for thisarticle, take responsibility for the integrity ofthe work as a whole, and have given theirapproval for this version to be published.

Disclosures. Sayed Zulfiqar Ali Shah, JawadAlam Karam, Alam Zeb, Rafi Ullah, Arif Shah,Ijaz Ul Haq, Iftikhar Ali, Haider Darain, andHong Chen declare that they have no conflict ofinterest.

Compliance with Ethics Guidelines. Thisarticle is based on previously conducted studiesand does not contain any studies with humanparticipants or animals performed by any of theauthors.

Data Availability. All data for this articleare presented in this manuscript.

Diabetes Ther (2021) 12:707–732 729

Open Access. This article is licensed under aCreative Commons Attribution-Non-Commercial 4.0 International License, whichpermits any non-commercial use, sharing,adaptation, distribution and reproduction inany medium or format, as long as you giveappropriate credit to the original author(s) andthe source, provide a link to the CreativeCommons licence, and indicate if changes weremade. The images or other third party materialin this article are included in the article’sCreative Commons licence, unless indicatedotherwise in a credit line to the material. Ifmaterial is not included in the article’s CreativeCommons licence and your intended use is notpermitted by statutory regulation or exceeds thepermitted use, you will need to obtain permis-sion directly from the copyright holder. To viewa copy of this licence, visit http://creativecommons.org/licenses/by-nc/4.0/.

REFERENCES

1. Ijaz S, Ajmal M. Experiencing type II diabetes inPakistan. Pak J Social Clin Psychol. 2011;9(1).

2. Sazlina S-G, Browning CJ, Yasin S. Promotingphysical activity in sedentary elderly Malays withtype 2 diabetes: a protocol for randomised con-trolled trial. BMJ Open. 2012;2(6):e002119.

3. Taylor JD, Fletcher JP, Tiarks J. Impact of physicaltherapist-directed exercise counseling combinedwith fitness center-based exercise training on mus-cular strength and exercise capacity in people withtype 2 diabetes: a randomized clinical trial. PhysTher. 2009;89(9):884–92.

4. International Diabetes Federation. IDF diabetesatlas, 9th edn. Brussels: IDF; 2019. https://www.diabetesatlas.org

5. Short KR, Vittone JL, Bigelow ML, et al. Impact ofaerobic exercise training on age-related changes ininsulin sensitivity and muscle oxidative capacity.Diabetes. 2003;52(8):1888–96.

6. Khan I, King H. Pakistan National Diabetes Surveyprevalence of glucose intolerance and associatedfactors in North West Frontier Province (NVVFP) ofPakistan. JPMA. 1999;49:206.

7. Rosenbloom AL, Joe JR, Young RS, Winter WE.Emerging epidemic of type 2 diabetes in youth.Diabetes Care. 1999;22(2):345–54.

8. Iqbal T, Rashid F, Saleem SA, Shah SA, Khalid GH,Ishtiaq O. Awareness about diabetes mellitusamongst diabetics. JRMC. 2013;17(2):294–6.

9. Qidwai W, Ashfaq T. Imminent epidemic of dia-betes mellitus in Pakistan: issues and challenges forhealth care providers. JLUMHS. 2010;9(3):112–3.

10. Akhter J. The burden of diabetes in Pakistan: theNational Diabetes Survey. JPMA. 1999;49(9):205.

11. Hakeem R, Fawwad A. Diabetes in Pakistan: epi-demiology, determinants and prevention. J Dia-betol. 2010;1(3):3.

12. Anderson RM, Funnell MM, Arnold MS, Barr PA,Edwards GJ, Fitzgerald JT. Assessing the culturalrelevance of an education program for urban Afri-can Americans with diabetes. Diabetes Educ.2000;26(2):280–9.

13. Gabriel BM, Zierath JR. The limits of exercisephysiology: from performance to health. CellMetab. 2017;25(5):1000–11.

14. Sigal RJ, Kenny GP, Wasserman DH, Castaneda-Sceppa C, White RD. Physical activity/exercise andtype 2 diabetes a consensus statement from theAmerican Diabetes Association. Diabetes Care.2006;29(6):1433–8.

15. Sigal RJ, Kenny GP, Wasserman DH, Castaneda-Sceppa C. Physical activity/exercise and type 2diabetes. Diabetes Care. 2004;27(10):2518–39.

16. Sigal RJ, Kenny GP, Boule NG, et al. Effects of aer-obic training, resistance training, or both on gly-cemic control in type 2 diabetes: a randomized trial.Ann Intern Med. 2007;147(6):357–69.

17. Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMAGroup. Preferred reporting items for systematicreviews and meta-analyses: the PRISMA statement.PLoS Med. 2009;6(7):e1000097

18. Ouzzani M, Hammady H, Fedorowicz Z, Elma-garmid A. Rayyan—a web and mobile app for sys-tematic reviews. Syst Rev. 2016;5(1):210.

19. Maher CG, Sherrington C, Herbert RD, MoseleyAM, Elkins M. Reliability of the PEDro scale forrating quality of randomized controlled trials. PhyTher. 2003;83:713–21.

20. Eng J, Teasell R, Miller W, et al. Spinal cord injuryrehabilitation evidence: method of the SCIRE sys-tematic review. Top Spinal Cord Inj Rehabil.2007;13:1–10.

730 Diabetes Ther (2021) 12:707–732

21. Higgins JPT, Altman DG. Assessing risk of bias inincluded studies. In: Higgins JPT, Green S, editors.Cochrane handbook for systematic reviews ofinterventions. Chichester: Wiley; 2008. p. 187–241.

22. Cochrane Collaboration. Review Manager, version5.3. Copenhagen: Cochrane Collaboration; 2014.

23. Hozo SP, Djulbegovic B, Hozo I. Estimating themean and variance from the median, range, and thesize of a sample. BMC Med Res Methodol.2005;5(1):13.

24. Leppink J, O’Sullivan P, Winston K. Effect size—large, medium, and small. Perspectives Med Educ.2016;5(6):347–9.

25. Lipsey MW, Wilson DB. Practical meta-analysis.Thousand Oaks: SAGE; 2001.

26. Andrews R, Cooper A, Montgomery A, et al. Diet ordiet plus physical activity versus usual care inpatients with newly diagnosed type 2 diabetes: theEarly ACTID randomised controlled trial. The Lan-cet. 2011;378(9786):129–39.

27. Akinci B, Yeldan I, Satman I, Dirican A, OzdinclerAR. The effects of Internet-based exercise comparedwith supervised group exercise in people with type2 diabetes: a randomized controlled study. ClinRehabil. 2018;32(6):799–810. PubMed PMID:29417832. eng.

28. Dixit S, Maiya A, Shastry B. Effect of moderate-in-tensity aerobic exercise on glycosylated haemoglo-bin among elderly patients with type 2 diabetes &peripheral neuropathy. Indian J Med Res.2017;145(1):129.

29. Dasgupta K, Rosenberg E, Joseph L, et al. Physicianstep prescription and monitoring to improveARTERial health (SMARTER): A randomized con-trolled trial in patients with type 2 diabetes andhypertension. Diabetes Obes Metab. 2017;19(5):695–704. PubMed PMID: 28074635.

30. Abdelbasset WK, Tantawy SA, Kamel DM, et al.Effects of high-intensity interval and moderate-in-tensity continuous aerobic exercise on diabeticobese patients with nonalcoholic fatty liver disease:a comparative randomized controlled trial. Medi-cine. 2020;99(10):e19471. PubMed PMID:32150108.

31. Halvari H, Healey J, Olafsen AH, et al. Physicalactivity and motivational predictors of changes inhealth behavior and health among DM2 and CADpatients. Scand J Med Sci Sports. 2017;27(11):1454–69. PubMed PMID: 27688002.

32. Heiskanen MA, Sjoros TJ, Heinonen IH, et al. Sprintinterval training decreases left-ventricular glucose

uptake compared to moderate-intensity continuoustraining in subjects with type 2 diabetes or predia-betes. Sci Rep. 2017;7(1):1–11.

33. Cox DJ, Taylor AG, Singh H, et al. Glycemic load,exercise, and monitoring blood glucose (GEM): aparadigm shift in the treatment of type 2 diabetesmellitus. Diabetes Res Clin Pract. 2016;111:28–35.

34. Sreedevi A, Gopalakrishnan UA, Ramaiyer SK,Kamalamma L. A randomized controlled trial of theeffect of yoga and peer support on glycaemic out-comes in women with type 2 diabetes mellitus: afeasibility study. BMC Complement Altern Med.2017;17(1):100.

35. Nuhu JM, Maharaj SS. Influence of a mini-trampo-line rebound exercise program on insulin resis-tance, lipid profile and central obesity inindividuals with type 2 diabetes. J Sports Med PhysFitness. 2017;58(4):503–9. PubMed PMID:28249384.

36. Mensberg P, Nyby S, Jørgensen PG, et al. Near-normalization of glycaemic control with glucagon-like peptide-1 receptor agonist treatment combinedwith exercise in patients with type 2 diabetes. Dia-betes Obes Metab. 2016;19(2):172–80. PubMedPMID: 27717126.

37. Winding KM, Munch GW, Iepsen UW, Van Hall G,Pedersen BK, Mortensen SP. The effect on glycaemiccontrol of low-volume high-intensity intervaltraining versus endurance training in individualswith type 2 diabetes. Diabetes Obes Metab.2018;20(5):1131–9. PubMed PMID: 29272072.

38. Brinkmann C, Weh-Gray O, Brixius K, Bloch W,Predel HG, Kreutz T. Effects of exercising beforebreakfast on the health of T2DM patients—a ran-domized controlled trial. Scand J Med Sci Sports.2019;29(12):1930–6. PubMed PMID: 31442336.

39. Vanroy J, Seghers J, Bogaerts A, Devloo K, De CockS, Boen F. Short- and long-term effects of a need-supportive physical activity intervention amongpatients with type 2 diabetes mellitus: a random-ized controlled pilot trial. PLoS One. 2017;12(4):e0174805. PubMed PMID: 28384237.

40. Hangping Z, Xiaona Q, Qi Z, et al. The impact onglycemic control through progressive resistancetraining with bioDensityTM in Chinese elderlypatients with type 2 diabetes: the PReTTy2 (Pro-gressive Resistance Training in Type 2 Diabetes)trial. Diabetes Res Clin Pract. 2019;150:64–71.

41. Church TS, Blair SN, Cocreham S, et al. Effects ofaerobic and resistance training on hemoglobin A1clevels in patients with type 2 diabetes: a random-ized controlled trial. JAMA. 2010;304(20):2253–62.

Diabetes Ther (2021) 12:707–732 731

42. Miyamoto T, Fukuda K, Oshima Y, Moritani T. Non-locomotive physical activity intervention using atri-axial accelerometer reduces sedentary time intype 2 diabetes. Phys Sportsmed. 2017;45(3):245–51. PubMed PMID: 28664755.

43. Bock BC, Thind H, Fava JL, et al. Feasibility of yogaas a complementary therapy for patients with type2 diabetes: the Healthy Active and in Control(HA1C) study. Complement Ther Med. 2019.42:125–31. PubMed PMID: 30670230.

44. Di Loreto C, Fanelli C, Lucidi P, et al. Make yourdiabetic patients walk: long-term impact of differ-ent amounts of physical activity on type 2 diabetes.Diabetes Care. 2005;28(6):1295–302.

45. Earnest CP, Johannsen NM, Swift DL, et al. Aerobicand strength training in concomitant metabolicsyndrome and type 2 diabetes. Med Sci SportsExerc. 2014;46(7):1293.

46. Alonso-Domınguez R, Garcıa-Ortiz L, Patino-Alonso MC, et al. Effectiveness of a multifactorialintervention in increasing adherence to theMediterranean diet among patients with diabetesmellitus type 2: a controlled and randomized study(EMID study). Nutrients. 2019;11(1):162.

47. Magalhaes JP, Judice PB, Ribeiro R, et al. Effective-ness of high-intensity interval training combinedwith resistance training versus continuous moder-ate-intensity training combined with resistancetraining in patients with type 2 diabetes: a one-yearrandomized controlled trial. Diabetes Obes Metab.2019;21(3):550–9. PubMed PMID: 30284352.

48. Hsieh P-L, Tseng C-H, Tseng YJ, Yang W-S. Resis-tance training improves muscle function and car-diometabolic risks but not quality of life in olderpeople with type 2 diabetes mellitus: a randomizedcontrolled trial. J Geriatr Phys Ther. 2018;41(2):65–76.

49. Fayh APT, Borges K, Cunha GS, et al. Effects of n-3fatty acids and exercise on oxidative stress param-eters in type 2 diabetic: a randomized clinical trial.J Int Soc Sports Nutr. 2018;15:18. PubMed PMID:29713249.

50. Hemmati Maslakpak M, Razmara S, Niazkhani Z.Effects of face-to-face and telephone-based family-oriented education on self-care behavior andpatient outcomes in type 2 diabetes: a randomizedcontrolled trial. J Diabetes Res. 2017;2017:8404328.PubMed PMID: 29359166.

51. Shakil-Ur-Rehman S, Karimi H, Gillani SA, Amjad I,Ahmad S, Yaseen A. Response to a supervisedstructured aerobic exercise training program inpatients with type 2 diabetes mellitus—does gendermake a difference? A randomized controlled

clinical trial. J Natl Med Assoc. 2018;110(5):431–9.PubMed PMID: 30129517.

52. Balducci S, Zanuso S, Cardelli P, et al. Changes inphysical fitness predict improvements in modifiablecardiovascular risk factors independently of bodyweight loss in subjects with type 2 diabetes partic-ipating in the Italian Diabetes and Exercise Study(IDES). Diabetes Care. 2012;35(6):1347–54.

53. Castaneda C, Layne JE, Munoz-Orians L, et al. Arandomized controlled trial of resistance exercisetraining to improve glycemic control in older adultswith type 2 diabetes. Diabetes Care. 2002;25(12):2335–41.

54. Wycherley TP, Clifton PM, Noakes M, BrinkworthGD. Weight loss on a structured hypocaloric dietwith or without exercise improves emotional dis-tress and quality of life in overweight and obesepatients with type 2 diabetes. J Diabetes Investig.2014;5(1):94–8.

55. Francois ME, Pistawka KJ, Halperin FA, Little JP.Cardiovascular benefits of combined intervaltraining and post-exercise nutrition in type 2 dia-betes. J Diabetes Complications. 2018;32(2):226–33. PubMed PMID: 29198993.

56. Zhang L-Y, Liu T, Teng Y-Q, et al. Effect of a12-week aerobic exercise training on serum fetuin-Aand adipocytokine levels in type 2 diabetes. ExpClin Endocrinol Diabetes. 2018;126(08):487–92.

57. Cassidy S, Vaidya V, Houghton D, et al. Unsuper-vised high-intensity interval training improvesglycaemic control but not cardiovascular auto-nomic function in type 2 diabetes patients: a ran-domised controlled trial. Diab Vasc Dis Res.2019;16(1):69–76.

58. Savikj M, Gabriel BM, Alm PS, et al. Afternoonexercise is more efficacious than morning exerciseat improving blood glucose levels in individualswith type 2 diabetes: a randomised crossover trial.Diabetologia. 2019;62(2):233–7. PubMed PMID:30426166.

59. Hangping Z, Xiaona Q, Qi Z, et al. The impact onglycemic control through progressive resistancetraining with bioDensity(TM) in Chinese elderlypatients with type 2 diabetes: The PReTTy2 (Pro-gressive Resistance Training in Type 2 Diabetes)trial. Diabetes Res Clin Pract. 2019;150:64–71.PubMed PMID: 30779971.

60. Hemmati Maslakpak M, Razmara S, Niazkhani Z.Effects of face-to-face and telephone-based family-oriented education on self-care behavior andpatient outcomes in type 2 diabetes: a randomizedcontrolled trial. J Diabetes Res. 2017:8404328.

732 Diabetes Ther (2021) 12:707–732