Can secretory immunoglobulin A in saliva predict a change in lung infection status in patients with cystic fibrosis? A prospective pilot study

Abstract

Background: Chronic lung infection with Pseudomonas aeruginosa is the main cause of mortality in patients with cystic fibrosis (CF). Sinus colonization with P. aeruginosa often precedes intermittent lung colonization, and intermittent colonization precedes chronic infection.When P. aeruginosa colonizes the sinuses, elevated immunoglobulin A (IgA) levels specific against P. aeruginosa can be detected in saliva. Therefore, we hypothesized that increasing levels of IgA in saliva can be detected before P. aeruginosa lung colonization.

Methods: Forty-nine CF patients free from lung colonization with P. aeruginosa or other Gram-negative bacteria (GNB) were included in this prospective study. Saliva and serum samples were collected and examined for IgA antibodies against P. aeruginosa with at least 6-month intervals between sequential samples.

Results: A total of 110 measurements of IgA in saliva were included. During a median of 8.5-month follow-up, 25 patients changed their lung infection status. We were able to construct a statistical model that for a given value of IgA in saliva, could predict the probability of a change in lung infection status within the next 8.5 months (median): p = 1 / (1 + exp(-(-0.9582 + 1.6518*IgA)). The model includes a prediction band where 95% of new measurements are predicted to fall within. The model, however, failed to reach statistical significance (P = 0.056 1-tailed), probably because of lack of power.

Conclusion: The saliva IgA model may predict a worsening in lung infection status presumably acting as a surrogate marker of P. aeruginosa bacterial sinusitis. The model may identify patients at risk of subsequent lung colonization and, thus, be a helpful clinical tool, but it should be tested in studies with larger sample sizes to evaluate its utility.

Keywords: IgA antibodies; Pseudomonas; sinusitis.

Figure 1

Change in lung infection status. Forty‐nine CF patients free from GNB lung colonization were included in this prospective study. Repeated saliva samples were obtained from each of the patients with at least a 6‐month interval. However, 21 samples from our prospective study in 20096 were also included, to increase power. Consequently, 110 measurements from 49 patients were included. In the 49 patients, we observed 61 events where patients could change in lung infection status. The numbers in Table 1 refer to these events. During a median of 8.5‐months follow‐up, 7 patients changed to intermittent lung colonization with P. aeruginosa (Pa + (i)), 9 patients changed to chronic lung infection with P. aeruginosa (Pa + (c)) and 9 patients changes status to lung colonization with other gram‐negative bacteria (GNB+)

Figure 2

Predicted probability of a change in lung infection status for a given s‐IgA saliva measurement. This model predicts a probability for a change in lung infection status for a given IgA OD value, P = 1 / (1 + exp(−(−0.9582 + 1.6518* within. IgA)). The prediction band indicates where 95% of new measurements are predicted to fall. The points in the figure show the predicted probabilities. Y axis is the probability of a change in lung infection status. X axis refers to the IgA level in saliva