Table 1

Trace element concentrations independent of sample source

Element

Number

of

cohortsb

Number of

hemodialysis

participants

and control

participants

Pooleda

SMD

I2,

%

Range of

SMDs

Number of

studies

significantly

favouring

lower

concentrations

in hemodialysis

participants

Number of

studies

significantly

favouring lower

concentrations

in control

participants

Number of

studies with

non-significant

findings


Arsenic

3

110/54

-

96

(-1.23, 1.77)

1

1

1

Boron

4

69/494

-

96

(-5.21, 2.96)

1

1

2

Cadmium

6

722/968

2.07

98

(-1.01, 7.7)

1

5c

0

Chromium

11

330/635

0.84

95

(-0.66, 5.14)

0

6c

5

Cobalt

1

7/9

-

-

-

0

0

1

Copper

42

1712/1444

0.52

93

(-1.74, 5.28)

6

16c

20

Fluorine

1

7/8

-

-

-

0

1

0

Lead

14

1217/1751

0.11

96

(-2.89, 1.73)

4

5c

5

Manganese

8

399/522

-0.7

96

(-3.96, 0.29)

4c

0

4

Mercury

2

607/264

-

99

(-0.94, 0.49)

1

1

0

Molybdenum

1

14/59

-

-

-

0

1

0

Nickel

9

369/323

-

99

(-3.96, 4.67)

3

6

0

Selenium

46

1496/1443

-1.50

91

(-9.16, 1.97)

37c

1

8

Vanadium

5

112/137

3.07

87

(1.18, 6.28)

0

5c

0

Zinc

74

2515/2699

-1.61

95

(-8.99, 3.24)

56c

5

13


aPooled SMDs using the random-effects method; this estimate indicates the average effect of all included studies. Negative values indicate lower concentrations in the hemodialysis participants and positive values indicate lower concentrations in the control participants. Pooled SMDs are reported only when vote counting results are significant.

bSome studies counted more than once if results reported by strata, for example, women and men separately.

cVote counting sign test results were significant at the P ≤ 0.05 level. Applied to comparisons with a minimum of three studies.

SMD = standardized mean difference.

Tonelli et al. BMC Medicine 2009 7:25   doi:10.1186/1741-7015-7-25

Open Data