Skip to main content

Table 2 Summary of findings for mortality outcomes [35]

From: Evaluation of the health and healthcare system burden due to antimicrobial-resistant Escherichia coli infections in humans: a systematic review and meta-analysis

Burden of disease measure

Type of antimicrobial resistance

Number of participants (studiesa)

Relative effect sOR (95% CI)

Absolute effect Risk difference (95% CI)

Certainty of the evidence (GRADE)b

Commentc

30-day mortality

Third-generation cephalosporin

31,934

(23 studies)

2.02 (1.66–2.46)

112 more deaths per 1000 (from 76 to 151 more)

Evidence to support upgrading due to strong association and no evidence to support downgrading

Quinolone

27,703 (8 studies)

1.49 (1.23–1.82)

58 more deaths per 1000 (from 28 to 93 more)

No evidence to support downgrading or upgrading

MDR

6506 (4 studies)

1.63 (1.54–1.71)

96 more deaths per 1000 (from 83 to 106 more)

No evidence to support downgrading or upgrading

All-cause mortality

Third-generation cephalosporin

40,623 (51 studies)

2.27 (1.92–2.70)

130 more deaths per 1000 (from 98 to 166 more)

Evidence to support upgrading due to strong association and no evidence to support downgrading

Quinolone

31,324 (16 studies)

1.72 (1.40–2.12)

82 more deaths per 1000 (from 48 to 121 more)d

No evidence to support downgrading or upgrading

MDR

6814 (5 studies)

1.63 (1.55–1.70)

92 more deaths per 1000 (from 81 to 100 more)

No evidence to support downgrading or upgrading

Bacterium-attributable mortality

Third-generation cephalosporin

327 (3 studies)

1.76 (0.84–3.70)

78 more deaths per 1000 (from 18 fewer to 225 more)

Downgraded due to serious inconsistency and imprecision. No evidence to support upgrading.

Quinolone

–

–

–

–

Not reported

MDR

–

–

–

–

Not reported

  1. sOR, summary odds ratio
  2. aAll studies were observational
  3. bGRADE assessment began at low instead of high, since studies were observational
  4. cDetails of GRADE assessment available in Additional file 23
  5. dRaw data not available from one study and therefore did not contribute to calculation of baseline risk