Skip to main content

Table 3 Summary of findings

From: Restrictive versus conventional ward fluid therapy in non-cardiac surgery patients and the effect on postoperative complications: a meta-analysis

Outcomes

Illustrative comparative risks* (95% CI)

Relative effect (95% CI)

No of participants (studies)

Quality of the evidence (GRADE)

Assumed risk

Intervention risk/risk difference

Conventional regimen

Restrictive regimen

Complications

Overall rate

Randomized controlled trials

RR 0.46 (0.23 to 0.95)

214 patients (3 studies)

      

Moderate1,2

455 per 1000

210 per 1000 (105 to 432 per 1000)

Non-randomized studies

 

645 patients (4 studies)

      

Low

585 per 1000

433 per 1000 (310 to 602 per 1000)

RR 0.74 (0.53 to 1.03)

 

Mortality

Randomized controlled trials

OR 0.51 (0.05 to 4.90)

214 (3 studies)

      

Low2

2 per 1000

1 fewer per 1000

(0 fewer to 10 fewer per 1000)

  

Non-randomized studies

  

      

Low

16 per 1000

5 fewer per 1000

(4 fewer to 22 fewer per 1000)

OR 0.30 (0.06 to 1.46)

537 patients (3 patients

Postoperative length of stay (PLOS)

Scale: days

Randomized controlled trials

   

The mean PLOS ranged across control groups from

7 to 7 days

The mean PLOS in the intervention groups was 0.60 higher (0.75 lower to 1.95 higher)

MD 0.60 (0.75 to 1.95)

62 patients (1 study)

      

Very low3

Non-randomized studies

   

The mean PLOS ranged across control groups from

7 to 12 days

The mean PLOS in the intervention groups was 1.81 lower (3.27 lower to 0.35 lower)

MD − 1.81 (− 3.27 to − 0.35)

439 patients (2 studies)

      

Very low3

  1. CI Confidence interval, OR Odds Ratio, MD Mean Difference
  2. *Assumed risk: mean baseline risks of the studies. The Risk difference (and its 95% confidence interval) is based on the assumed risk in the conventional fluid regimen group and the relative effect of the intervention (and its 95% CI)
  3. 1High risk for reporting bias in 3 out of 4 studies. One study terminated prematurely
  4. 2One included randomized trial terminated prematurely
  5. 3Inconsistency not explained. Therefore, we downgraded by one scale starting from low