Skip to main content
. 2018 May 16;8(5):e020371. doi: 10.1136/bmjopen-2017-020371

Table 2.

Association of bariatric surgery with first incident AKI, stratified by length of follow-up

PY Events Rate per 1000 PY (95% CI) Crude RR
(95% CI)*
P values Adjusted RR
(95% CI)
P values
All patients
Day 1–30
  Unexposed 203 0 0
  Bariatric surgery 199 5 25.1 (10.5 to 60.4)
>Day 30
  Unexposed 7882 54 6.9 (5.2 to 8.9)
  Bariatric surgery 8061 34 4.2 (3.0 to 5.9) 0.62 (0.40 to 0.95) 0.027 0.37 (0.23 to 0.61) <0.001
All patients analysed by type of surgery§
Day 1–30
  Unexposed
  Gastric band
  Sleeve gastrectomy
  Gastric bypass
  Other
>Day 30
  Unexposed 7882 54 6.9 (5.2 to 8.9)
  Gastric band 4614 17 3.7 (2.3 to 5.9) 0.54 (0.31 to 0.93) 0.026
  Sleeve gastrectomy 728 <5 5.5 (2.1 to 14.6) 0.80 (0.29 to 2.21) 0.670
  Gastric bypass 2655 13 4.9 (2.8 to 8.4) 0.71 (0.39 to 1.31) 0.277
  Other 63 0
All patients over whole period of follow-up
  Unexposed 8085 54 6.7 (5.1 to 8.7)
  Bariatric surgery 8259 39 4.7 (3.5 to 6.5) 0.71 (0.47 to 1.07) 0.099 0.45 (0.28 to 0.72) 0.001

Unexposed refers to the propensity-matched comparison group.

*Poisson regression model.

†Wald test.

‡Poisson regression model adjusted for age at baseline, sex, calendar time, CKD at baseline, history of AKI, history of taking oral antidiabetics and BMI at baseline.

§No analysis for day 1–30 owing to sparse data.

¶Cell counts <5 have been suppressed to ensure anonymity.

AKI, acute kidney injury; BMI, body mass index; CKD, chronic kidney disease; PY, person years; RR, rate ratio.