Table 3

Comparison of ACPA and RF tests in terms of positive and negative predictive values

CCP2

CCP3

MCV

RF

Reference

Number of patients (RA/control)

PPV (%)

NPV (%)

PPV (%)

NPV (%)

PPV (%)

NPV (%)

PPV (%)

NPV (%)

Luis Caro-Oleas and colleagues [54]

124/158

95.2

73.1

92.3

70.5

90.9

73.6

Coenen and colleagues [44]^d

133/165

89.7 to 91.4^a

87.1 to 89.1^a

76.7

88.2

80.0

87.2

Correia and colleagues [39]

86/90

92.6 to 96.7^b

72.9 to 78.7^b

91.9 to 94.9^b

74.4 to 74.6^b

80.3

72.4

Dos Anjos and colleagues [40]

70/88

91.7

84.7

90.6

87.2

Liu and colleagues [33]

170/136

95.5

66.8

93.7

77.4

82.0

69.9

Lutteri and colleagues [24]

120/170

87.7 to 96.2^c

76.6 to 78.3^c

92.6

78.1

77.9

79.5

Sghiri and colleagues [29]

170/309

91.1

86.3

66.0

84.7

58.6

79.9

Soos and colleagues [52]

119/118

97.6

74.4

90.0

78.8

80.2

74.0

Ursum and colleagues [48]

123/39

95.8

96.1

Van der Linden and colleagues [17]

201/424

67.1

79.0

64.0

80.0

56.3

79.2

61.7

77.8

Average

91.2

78.4

84.9

79.8

80.4

81.5

75.9

75.3

Control groups are not identical. ACPA, anti-citrullinated protein antibodies; MCV, mutated citrullinated vimentin; NPV, negative predictive value; PPV, positive predictive value; RA, rheumatoid arthritis; RF, rheumatoid factor. ^aThree CCP2 tests used. ^bTwo CCP2 tests and two cut-off values used. ^cSix CCP2 tests used. ^dCoenen and collaborators also used the citrullinated filaggrin test (CPA; Genesis, Littleport, UK): PPV, 85.5; NPV, 85.6.

Pruijn et al. Arthritis Research & Therapy 2010 12:203   doi:10.1186/ar2903