Table 1.
Studies comparing the incidence of acute kidney injury between patients exposed and unexposed to intravenous contrast medium in various acute care settings
| Reference | Setting/design | Contrast group (patients) | Control group (patients) | Additional means to delineate the renal risk attributable to contrast medium | Comments |
|---|---|---|---|---|---|
| Polena et al. [67] |
ICU, Retrospective cohort (single-centre) |
N = 75 AKI: 18.6% |
N = 75 AKI: 2% |
None | |
| Tremblay et al. [68] |
Trauma centre; Retrospective cohort (single-centre) |
N = 56 AKI: 3% |
N = 39 AKI: 16% |
None | Proportion of ICU patients unclear |
| Oleinik et al. [69] |
ED, patients with intracerebral haemorrhage; Retrospective cohort (single-centre) |
N = 368 AKI: 6% OR 1.4 (95%CI 0.6–3.2) |
N = 130 AKI: 14% |
Multivariate regression | Proportion of ICU patients unclear |
| Lima et al. [70] |
Stroke patients; Retrospective cohort (single-centre) |
N = 575 AKI: 5% OR 0.42 (95%CI 0.24–0.71) |
N = 343 AKI: 10% |
Multivariate regression | Proportion of ICU patients unclear |
| Aulicky et al. [71] |
Stroke patients; Retrospective cohort (single-centre) |
N = 164 AKI: 3% |
N = 77 AKI: 4% |
Multivariate regression | Proportion of ICU patients unclear |
| Mc Gillicuddy et al. [72] |
Trauma centre, elderly trauma patients; Retrospective cohort (single-centre) |
N = 822 AKI: 1.9% |
N = 249 AKI: 2.4% |
None | Proportion of ICU patients unclear |
| Ng et al. [73] |
ICU patients (oncology); Retrospective cohort (single-centre) |
N = 81 AKI: 17% |
N = 81 AKI: 17% |
1-to-1 matching on baseline serum creatinine, SOFA score and age | |
| Cely et al. [19] |
ICU; Prospective cohort (single-centre) |
N = 53 AKI: 9.4% |
N = 53 AKI: 15% |
1-to-1 matching on baseline creatinine clearance, diabetes, mechanical ventilation, vasopressor use | |
| Sinert et al. [74] |
ED patients with normal renal function; Retrospective cohort (2 centres) |
N = 773 AKI: 5.7% |
N = 2956 AKI: 9.0% |
None | |
| Kim et al. [75] |
ICU, trauma patients; Retrospective cohort (single-centre) |
N = 389 AKI: 30% OR 0.99 (CI 95% 0.78–1.25) |
N = 182 AKI: 29% |
None | |
| Ehrmann et al. [20] |
ICU; Prospective cohort (2 centres) |
N = 146 AKI: 5.5% |
N = 146 AKI: 5.5% |
1-to-1 propensity score matching** | |
| Christ et al. [76] |
ICU patients after cardiac arrest; Retrospective cohort |
N = 89 AKI: 15.7% |
N = 53 AKI: 37.7% |
None | |
| Gao et al. [77] |
ICU; Retrospective cohort (single-centre) |
N = 474 AKI: 14.8% OR 1.66 (95% CI 0.72–3.90) |
N = 1,896 AKI: 12.4% |
Multivariate regression | |
| Sonhaye et al. [78] |
ED; Prospective cohort (single-centre) |
N = 620 AKI: 3% OR: 95%CI is missing |
N = 672 AKI: 2% |
Multivariate regression | Proportion of patients admitted to the ICU unclear |
| Heller et al. [79] |
ED patients admitted to the hospital; Retrospective cohort (single-centre) |
N = 6954 AKI: 8.6% |
N = 909 AKI: 9.6% |
None | 8% of patients admitted to the ICU |
| McDonald et al. [16] |
ICU; Retrospective cohort (single-centre) |
N = 1223 with eGFR > 45 AKI: 31% OR 1.21 (CI 95% 0.87–1.68) N = 285 with eGFR ≤ 45 AKI: 50% OR 0.88 (CI95% 0.75–1.05) |
N = 1223 with eGFR > 45 AKI: 34% N = 285 with eGFR ≤ 45 AKI: 45% |
1-to-1 propensity score matching** | An increased risk of dialysis was observed in patients with pre-contrast eGFR ≤ 45 ml/min/1.73 m2 |
| Hinson et al. [25] |
ED patients admitted to the hospital; Retrospective cohort (single-centre) |
N = 7,201 AKI: 6.8% OR 1.00 (95% CI 0.99–1.01) |
N = 10,733 AKI: 8.5% |
1-to-1 propensity-score matching** |
ED critical care designation in 9% of patients Proportion of patients admitted to the ICU unclear |
| Miyamoto et al. [17] |
ICU, patients with sepsis-associated AKI receiving continuous RRT; Retrospective cohort (national database) |
N = 3485 Composite outcome (in-hospital death or RRT dependence at discharge): 49.6% OR 0.98 (95% CI 0.88–1.07) RRT dependence at discharge: 4.4% OR 1.08 (95% CI 0.85–1.31) median duration of RRT: 4 [IQR 2–11] days |
N = 3485 Composite outcome (in-hospital death or RRT dependence at discharge): 50.2% RRT dependence at discharge: 4.1% median duration of RRT: 4 [IQR 2–11] days |
1-to-1 propensity score matching** | |
| Shih et al. [80] |
ICU (subgroup analysis), patients with CKD; Retrospective cohort (national database) |
N = 51 30-day RRT: 25.5% aHR 0.95 (95% CI 0.44–2.05) |
N = 176 30-day RRT: 25.6% |
Cox proportional hazard model adjusted for age, sex and comorbid conditions | Analysis of the Taiwan National Health Insurance Research Database |
| Goto et al. [81] |
ICU (patients with sepsis and AKI); Retrospective cohort (single-centre) |
N = 100 further deterioration of renal function = 34% | N = 100 further deterioration of renal function = 35% | 1-to-1 propensity score matching** | |
| Hinson et al. [82] |
ED, patients with sepsis; Retrospective cohort (single-centre) |
N = 1464 AKI: 7.2 OR 0.99 (95% CI 0.97–1.02) |
N = 2707 AKI: 9.6% |
1-to-1 propensity score matching** |
ED critical care designation in 4% of patients Proportion of patients admitted to the ICU unclear |
| McGaha et al. [83] |
Paediatric trauma centre, severely injured patients; Retrospective cohort (single-centre) |
N = 164 AKI: 7.3% |
N = 47 AKI: 8.5% |
None | 57% of patients admitted to the ICU |
| Williams et al. [18] |
ICU; Retrospective cohort (6-hospital health system) |
N = 2306 AKI: 19.3% OR 1.11 (95% CI 0.95–1.29) |
N = 2306 AKI: 18.0% |
1-to-1 propensity score matching** |
ICU intensive care unit, AKI acute kidney injury*, Contrast iodinated contrast media, OR odds ratio, SOFA sequential organ failure assessment, 95% CI: 95% confidence interval, eGFR estimated glomerular filtration rate, ED emergency department, RRT renal replacement therapy, IQR interquartile range, CKD chronic kidney disease, aHR adjusted hazard ratio, DRF deterioration of kidney function
*The definition for AKI may differ from one study to another. This may, in part, account for the between-studies discrepancies in the reported incidence for AKI
**Patients exposed and patients unexposed to contrast were matched on their propensity to be administered contrast. This approach aims at mimicking randomization in an observational study design