Ilc. Chapple et al., Prediction and diagnosis of attachment loss by enhanced chemiluminescent assay of crevicular fluid alkaline phosphatase levels, J CLIN PER, 26(3), 1999, pp. 190-198
The current study aimed to apply a novel enhanced chemiluminescence assay i
n the analysis of gingival crevicular fluid (GCF) alkaline phosphatase (ALP
) levels from patients with untreated adult periodontitis. 3666 sites in 25
patients were monitored prior to and after attachment loss was detected wi
th a Florida disc probe. Parameters assessed were, relative attachment leve
l, probing pocket depth, occurrence of bleeding on probing (single episode)
, GCF volume (mu l), total ALP levels (mu IU/30 s sample time) and ALP conc
entration (IU/l). After recruiting patients to the study, all measures were
taken at baseline and 3 months later, prior to the institution of non-surg
ical periodontal therapy at active sites. Thresholds for determining attach
ment loss were calculated using a modification of the tolerance method. The
mesio-buccal sites of all teeth had GCF samples collected. The size of ind
ividual patient thresholds used to define whether attachment loss had occur
red, was dependent upon the discomfort felt by that patient during electron
ic probing, with a positive correlation existing between discomfort on prob
ing(10 cm visual analogue scale) and threshold size (R=0.52,p<0.049). A tot
al of 274 sites (7.5%) experienced attachment loss of which 39 sites had GC
F samples available for analysis. Total ALP levels were significantly highe
r at baseline for sites that progressed to attachment loss than paired cont
rols (p<0.003), but all other parameters showed no differences (p>0.1). The
re were significant increases in total ALP levels and GCF volumes for activ
e sites between baseline and 3 month measures (p<0.01), but not for control
sites or lest site ALP concentration (p>0.8). The diagnostic accuracy for
GCF ALP as a predictor of future attachment loss (threshold 900 mu IU/30 s)
was 64%, with + ve and -ve predictive values of 62% and 68%. When a thresh
old of 1300 mu IU/30 s was selected for ALP as a marker of recent or curren
tly active disease, diagnostic accuracy and +ve/-ve predictive values were
77% and 77%/76%, respectively. These results indicate that total GCF ALP le
vels may serve as a predictor of future or current disease activity.