Literature retrieval and screening
A total of 134 literatures were obtained by preliminary examination, including 13 articles in Chinese, 121 articles in English. Import Endnote software, by find duplicates, excluding 46 duplicates, Of these ,7 in Chinese and 39 in English. Reading topics and abstracts, according to the exclusion criteria, Excluding 76 articles. Twelve papers were included in the full text, of which 2 were in Chinese, 10 articles in English; One paper lacks full text, to contact the author via email for supplements. The paper was written in Spanish and read after translation through Youdao. Six articles[11-16] were finally included in the SRs/MAs, 1 article[14] in Chinese, 5 articles[11-13, 15, 16]in English; The process and results of literature screening is shown in Fig.1.
Basic characteristics of inclusion and exclusion studies
The basic characteristics of the 6 included SRs/MAs studies are shown in Table 1. According to the year of publication, three papers in 2019[11-13], two in 2018[14, 15], and one in 2015[16]. According to the language, one in Chinese for MA study[14], five in English including two SRs/MAs[11, 13], one MA[15] and two SRs papers[12, 16]. According to type of original study included ,five randomized controlled trials (RCT)[11-15],one was included in the pre-/post-test(PPT) studies[16]; The number of original studies included ranged from five to 13; The sample size ranged from 84 to 345; The participants ranged in age from 15 to 70. The intervention group was HIIT.
Table 1 Basic characteristics of studies included
Included
Research
|
Included original research
|
Interventions
|
Evaluation tools
|
Fund support
|
Main conclusions
|
Registration
|
Type
|
Number
(sample size)
|
Age (years)
|
Test Group
|
Control group
|
LORA-
POZO 2019[11]
|
RCT
|
10(548)
|
43.1-70.1
|
HIIT
|
CON
LIT
M IT
|
PEDro
|
Yes
|
HIIT intervention, compared to MIT, LIT and CON, turns out to be effective in the improvement of the anthropometric conditions (bodyweight, BMI), cardiovascular conditions (systolic BP and VO2max) and metabolic conditions (HbA1c) in subjects with T2D.
2. Any firm conclusions about the effectiveness of HIIT on diastolic BP cannot be draw.
3. It is necessary to promote new lines of research in order to identify the most effective protocols according to HIIT’s frequency, session duration and rehabilitation program duration, as well as a detailed description of the exercises.
|
Yes
|
DA SILVA D E 2019[12]
|
RCT
|
5(324/182=142)
|
40-60 years
|
HIIT
|
C E
CT
M IT
C ON
|
CRH5.1.0
|
Yes
|
1. there is little evidence of the effectiveness of high-intensity exercise in reducing type 2 diabetes in randomized controlled trials; 2. The insufficient number of studies, the results of intervention cannot be accurately used as a basis for public guidance; 3. The results are inconsistent, not presenting much diversity between the training methods, and the different methods of volume and intensity
control training.
|
No
|
LIU J-X 2019[13]
|
RCT
|
13(345)
|
15-70 years
|
HIIT
|
CON
MICT
|
CCT
GRADE
|
Yes
|
1. HIIT is an effective strategy for improving cardiorespiratory fitness in patients with T2D, preferable to MICT. 2. Results related to other parameters associated with the prognosis of T2D, such as HbA1c, body weight, and BMI, were not conclusive. 3. Future studies should investigate the effects of HIIT in T2D patients through multicenter RCTs with large sample sizes over the long term.
|
No
|
CHEN B-L
2018[14]
|
RCT
|
8(275)
|
≧18 years
|
HIIT
|
MICT CON
|
CRH5.1.0
GRADE
|
Yes
|
There is no evidence showed that HIIT is more effective on blood glucose, blood fat and body composition in the patients with T2DM, which needs further research.
|
Yes
|
DE NARDI A T 2018[15]
|
RCT
|
5(120)
|
61.7± 6
|
HIIT
|
MICT
|
CRH5.1.0
GRADE
|
No
|
High-intensity interval training has the potential to be used as a treatment modality for individuals with prediabetes or T2D. This intervention induces cardiometabolic adaptations similar to those of MICT and provides greater benefits to functional capacity in patients with T2D. However, strong supporting research is further needed.
|
Yes
|
CURRY M 2015[16]
|
RCT
P PT
|
6(84)
|
Minimum mean age 40.2 years
|
HIIT
|
MICE
PPT
|
25P
|
No
|
1.HIIT and MICE do not consistently show significant differences. 2. More comprehensive evidence is available concerning the benefits of HIIT in patients with type 2 diabetes are needed. 3. it is suggested that future trials increase the duration of data collection (12–24 months) in order to yield more conclusive evidence. 4. The professionals who study the effects of exercise and type2 diabetes should build a consensus on standardized outcomes.
|
No
|
The control group included Continuous Exercise (CE), Continuous Training (CT), Moderate Intensity Training (MIT) and Moderate-Intensity Continuous Training (MICT), moderate-intensity continuous exercise (MICE), moderate Intensity interval Training (MMIIT) and Low Intensity Training (LIT), No Exercise or quiet and routine treatment (CON), etc. In terms of quality assessment, Cochrane evaluation tools are mainly used for bias risk assessment, and three of them adopted Cochrane Reviewers' Handbook Version 5.1 (CRH5.1)[12, 14, 15], One by Cochrane Collaboration's tool (CCT)[13]; The other two use Pedro scale (Pedro) [11]and 25-point Standardized Checklist(25P)[16]. The quality of evidence in three literatures was evaluated by GRADE[13-15]; Four noted fund support, of which three were funded[11, 13, 14], one stated that it was not funded[15], and all of the studies did not provide funding for inclusion in original study.
Six excluded studies and the reasons for exclusion are shown in Table 2. The reasons for exclusion are mainly as follows: participants of the study include other groups of people, such as healthy people, overweight people, sedentary people; there is no clear retrieval strategy, inclusion and exclusion criteria, and evaluation methods of literature quality.
Table 2 Excluded articles and reason for exclusion
Research
|
Exclusion grounds
|
DE OCA
2019[17]
|
narrative review, Participants of the study include other groups of people, such as healthy people, overweight people, sedentary people; The literature quality evaluation methods such as methodological research results were not used.
|
WORMGOOR S G 2017[18]
|
narrative review, literature quality evaluation methods such as methodological research results were not used.
|
FRANCOIS M E 2015[19]
|
There is no clear retrieval strategy, inclusion and exclusion criteria, evaluation method of literature quality.
|
LITTLE J P 2014[20]
|
There is no clear retrieval strategy, inclusion and exclusion criteria, evaluation method of literature quality.
|
JELLEYMAN C 2015[21]
|
The participants of the study include other groups of people, such as healthy people, overweight people, sedentary people.
|
Cheng-ji Han 2019[22]
|
There is no retrieval strategy, inclusion and exclusion criteria, evaluation method of literature quality.
|
As shown in Fig.2, A total of 21 original studies were included in six SRs/MAs. If the original study was published after or in the same year as the SRs/MAs , the total will not be included. Three original studies were included in all included SRs/MAs, as follows: MAILLARD F2016[23], MITRANUN W2014[24] and TERADA T2013[25]. HOLLEKIM-STRAND S M2014[26], KARSTOFT K2013[27], STOA EM2017[28], ALVAREZ C2016[29] and CASSIDY S2016[30] were included in three or more SRs/MAs. There are 13 original studies were included by only one SRs/MAs and it's on the edge of the figure, as follows: TERADA T2013[25], SHABAN N2014[31], GILLEN J B2012[32], LITTLE J P2011[33] were included in CURRY M,2015[16]. Two papers of TERADA T 2013[25, 34] are the same research. others are Pre and post control studies. ROBINSON E2015[35], KARSTOFT K2014[36] were included in CHEN B-L 2018. The participants included in ROBINSON E2015[35] are sedentary adults. KARSTOFT K2014[36] is the same study as KARSTOFT K 2013[27].WINDING KM2018[37], BELLIA A2017[38], RAMOS JS2016[39], LEE SS2015[40], BACKX K2011[41] were included in LIU J-X 2019[13]. Participants included in RAMOS JS2016[39], LEE SS2015[40] are persons with metabolic syndrome, Adolescent patients with T2DM. RUFFINO J S2017[42], BALDUCCI S2012[43] were included in LORA-POZO,2019[11]. In fact, the intervention of BALDUCCI S2012 is the moderate-to-high intensity exercise that performed aerobic training at 70% of predicted VO2max and resistance training at 60% of predicted 1-RM(Repetition Maximum)[43].
Evaluation results of AMSTAR2
The evaluation results of AMSTAR2 tools are shown in Table 3. The evaluation results of key items are “yes” as follows: item 2 is 33.3% (2/6), item 4 is 66.7% (4/6), item 7 is 16.7% (1/6), item 9 is 83.3 % (5/6), item 11 is 66.7%(4/6), item 13 is 16.7%(1/6), item 15 is 0(0/6). For other items, the evaluation result of "yes" is 100% for items 1, 5 and 8, item 6 is 83.3% (5/6), item 14 is 66.7% (4/6), item 16 is 50% (3/6), item 12 is 33.3% (2/6), item 3 and 10 are 0 (0/6). Using the AMSTAR2 tools, the final quality evaluation is extremely low.
Table3 Evaluation results of AMSTAR2 tools
Author (year)
|
1
|
2
|
3
|
4
|
5
|
6
|
7
|
8
|
9
|
10
|
11
|
12
|
13
|
14
|
15
|
16
|
Quality rating
|
LORA-POZO 2019
|
Y
|
Y
|
N
|
N
|
Y
|
Y
|
N
|
Y
|
Y
|
N
|
Y
|
Y
|
N
|
Y
|
N
|
Y
|
Very low
|
DA SILVA
DE 2019
|
Y
|
N
|
N
|
Y
|
Y
|
Y
|
N
|
Y
|
Y
|
N
|
N
|
N
|
N
|
N
|
N
|
N
|
Very low
|
LIU J-X 2019
|
Y
|
N
|
N
|
Y
|
Y
|
Y
|
N
|
Y
|
Y
|
N
|
Y
|
N
|
N
|
N
|
N
|
Y
|
Very low
|
CHEN B-L 2018
|
Y
|
PY
|
N
|
Y
|
Y
|
Y
|
N
|
Y
|
Y
|
N
|
Y
|
N
|
N
|
Y
|
N
|
N
|
Very low
|
DE NARDI
AT 2018
|
Y
|
Y
|
N
|
Y
|
Y
|
Y
|
N
|
Y
|
Y
|
N
|
Y
|
Y
|
Y
|
Y
|
N
|
Y
|
Very low
|
CURRY M
2015
|
Y
|
N
|
N
|
N
|
Y
|
N
|
Y
|
Y
|
Y
|
N
|
N
|
N
|
N
|
Y
|
N
|
N
|
Very low
|
Note: Y is Yes, PY is Partial Yes, N is No.
Risk of bias rating
Because this study is OoSRs, only to assess the risk of bias of SRs/MAs, the first phase can be skipped. The second phase focuses on identifying possible biases during the each process of production OoSRs, include 4 domains. The third phase is to judge the overall bias risk of SRs/MAs, as shown in Table 4, the results of phase three are all for high risk.
Table 4 Evaluation results of ROBIS tools included in the study (phases II and III)
Author (year)
|
Phase II
|
Phase III
|
DOMAIN1
|
DOMAIN2
|
DOMAIN3
|
DOMAIN4
|
Judging Risk of bias
|
LORA-POZO
2019
|
Low
|
High
|
Low
|
High
|
High
|
DA SILVA D E 2019
|
Uncertainty
|
High
|
Low
|
High
|
High
|
LIU J-X2019
|
Uncertainty
|
Low
|
Low
|
High
|
High
|
CHEN B-L
2018
|
Uncertainty
|
High
|
Low
|
High
|
High
|
DE NAR D AT
2018
|
Low
|
Low
|
Low
|
High
|
High
|
CURRY M 2015
|
Uncertainty
|
High
|
Low
|
High
|
High
|
The evaluation results of the quality of evidence
According to the outcomes list of all included SRs/MAs studies (number of original studies included), there are 79 outcomes (included in 21 original studies). As shown in Fig.3, Lora-Pozo2019 has 8 (10) items[11], DA SILVA DE 2019 has 13(5) items[12], LIU J-X2019 has 13(13) items[13], CHEN B-L 2018 has 56(8) items[14], DE NARDI AT2018 has 14(5) items[15], CURRY M2015 has 29(6) items[16], The bubble size in the figure reflects the ratio of the number of included outcomes to the number of included original studies.
As shown in Fig.4, Outcomes of Liu J-X2019[13]HIIT VS CON included 11 items, HIIT VS MICT included 13 items; Outcomes of CHEN B-L 2018[14] included 12 items; Outcomes of DE NARDI AT2018[15] included 12 items. The red color in the figure indicates that the statistical difference was significant between the two groups, and the bubble size indicates the absolute value of the effect amount. Based on the principles of careful evaluation and citation, Results of GRADE systematic review based on the above three SRs/MAs studies were 39 items. Among them, 29 items were low quality, eight items were very low quality and two items of intermediate quality. Low quality and very low quality evidence bodies account for 95% of the total.