Skip to main content

Table 4 Summary of findings table—GRADE assessment. Question: What is the effect of exercise training on cardiac autonomic function (cardiac autonomic reflex test variables) in type 2 diabetes mellitus?

From: Effect of exercise training on cardiac autonomic function in type 2 diabetes mellitus: a systematic review and meta-analysis

Certainty assessment

№ of patients

Effect

Certainty

Importance

No. of studies

Study design

Risk of bias

Inconsistency

Indirectness

Imprecision

Other considerations

Exercise

Control

Absolute

(95% CI)

E:I ratio (assessed with cardiac autonomic reflex tests)

2

Non-randomized studies

Seriousa

Seriousb

Not serious

Seriousc

All plausible residual confounding would reduce the demonstrated effect

49

48

MD 0.01 higher

(0.07 lower to 0.09 higher)

Low

Important

Valsalva ratio (assessed with cardiac autonomic reflex tests)

2

Non-randomized studies

Seriousa

Not serious

Not serious

Seriousc,d

All plausible residual confounding would reduce the demonstrated effect

49

48

MD 0.1 lower

(0.23 lower to 0.03 higher)

Moderate

Important

30:15 ratio (assessed with cardiac autonomic reflex tests)

2

non-randomized studies

Seriousa

Seriousb

Not serious

Seriousc,d

All plausible residual confounding would reduce the demonstrated effect

49

48

MD 0.01 lower

(0.09 lower to 0.07 higher)

Low

Important

E: I ratio (follow-up: mean 12 weeks; assessed with: cardiac autonomic reflex tests)

1

Randomized trial

Not serious

Not serious

Seriouse

Seriouse

None

28

28

MD 0

(− 5.59 lower to 5.59 higher)

Lowe,f

Important

Valsalva ratio (follow-up: mean 12 weeks; assessed with cardiac autonomic reflex tests)

1

Randomized trial

Not serious

Not serious

Seriouse

Seriouse

None

28

28

MD 0.36 higher

(0.1 higher to 0.62 higher)

Lowe,f

Important

30:15 ratio (follow-up: mean 12 weeks; assessed with cardiac autonomic reflex tests)

1

Randomized trial

Not serious

Not serious

Seriouse

Seriouse

None

28

28

MD 0.24 higher

(0.01 lower to 0.49 higher)

Lowe,f

Important

  1. CI Confidence interval, MD Mean difference
  2. aThe studies are non-randomized controlled trials. A likely bias due to confounding and selection bias may be suspected
  3. bThe intervention in both studies was different, which may have influenced the value
  4. cImprecision may be suspected as there are only two studies included in the analysis
  5. dImprecision may be suspected as there is a wide confidence interval
  6. eThe small sample cannot represent the population as a whole
  7. fSmall sample size