Skip to main content

Table 4 Clearance measurements of hemofiltration vs.

From: Hemofiltration compared to hemodialysis for acute kidney injury: systematic review and meta-analysis

  

Change in clearance hemofiltration vs hemodialysisa

Molecular substance

Number of trials; number of patients randomized

Effect estimate

95% confidence interval

P-value

Heterogeneity (I2)

Smaller molecules

Urea (60 Da)

4 [33, 36–38]; 49

+1%b

-2% to +3%

0.60

0%

Phosphate (95 Da)

1 [37]; 18

0%c

-4% to +4%

1.00

n/a

Creatinine (113 Da)

3 [33, 37, 38]; 43

+1.8%b

-0.4% to +4.1%

0.12

0%

Uric acid (168 Da)

2 [33, 37]; 28

+4%

+1% to +7%

0.01

0%

Larger molecules

Vancomycin (1.8 kDa)

1 [33]; 10

+18%

+8% to +28%

0.0003

n/a

β2-microglobulin (11.8 kDa)

2 [37, 38]; 33

+94%d

+78% to +112%

<0.0001

0%

IL-1 Receptor Agonist (16-18 kDa)

1 [24]; 12

+77%e,f

+24% to +153%

0.002

n/a

Retinol Binding Protein (21.2 kDa)

1 [37]; 18

+42%

+4% to +94%

0.03

n/a

IL-6 (26 kDa)

2 [24, 34]; 22

+6%f,g

-62% to +191%

0.91

89%

  1. hemodialysis
  2. aUsing RoM. bResults are unchanged if two trials either using different ultrafiltration rates [31] or comparing different doses of hemofiltration vs a mixture of hemodialysis and hemofiltration [21] are also included after accounting for the different flow rates (urea clearance +1% (95% CI -2% to +3%, P = 0.58, I2 = 0%, six studies [21, 31, 33, 36–38], 74 patients) and creatinine clearance +1.9% [95% CI -0.3% to +4.1%, P = 0.09, I2 = 0%, five studies [21, 31, 33, 37, 38], 68 patients)). cResults unchanged if data from the filtration-only vs dialysis-only groups in the sustained low efficiency RRT trial [28] are added: +31% (95% CI -35% to +165%, P = 0.45, I2 = 72%, two studies [28, 37], 30 patients). dOne of these studies [38] also measured adsorptive clearance of β2-microglobulin and found negligible (<1%) clearance occurred via adsorption to the filter and this was similar between hemofiltration and hemodialysis. eComparing higher-dose (2.5 L/h) groups only. Comparing lower-dose (1 L/h) groups in the same study [24] showed a smaller but still statistically significant increased clearance of IL-1 Receptor Agonist +28% (95% CI +10% to +50%%, P = 0.002). fThe medians and ranges reported in this trial [24] were converted to means and standard deviations using published guidelines [81]. gIncluding the higher-dose (2.5 L/h) groups for Morgera 2004 [24] only. Including also the lower-dose (1 L/h) groups for this trial [24] produced similar results -2% (95% CI -46% to +80%%, P = 0.96, I2 = 0%, two studies [24, 34], 34 patients]. Abbreviations: CI, confidence interval; Da, Daltons; I2, I2 heterogeneity measure; IL, interleukin; kDa, kiloDaltons; n/a, not applicable; RoM, ratio of means [17]; RRT, renal replacement therapy.
Back to article page

Critical Care

ISSN: 1364-8535