Clinical outcomes associated with Staphylococcus aureus and Pseudomonas aeruginosa airway infections in adult cystic fibrosis patients

A total of 91 adult patients were followed at the Montreal Chest Institute CF clinic
during the study period, and 84 patients were included in this study. We analyzed
the microbiology of a total of 366 sputum samples collected during the study period
(average 4.3 samples per patient): 37 % of samples were positive for SA (including
n?=?4 for MRSA), and 70 % were positive for PA. The study patients were classified
into one of 3 groups based on their PA and SA airway infection status, namely “PA”
(n?=?50; 59 %), “SA only” (n?=?20; 24 %), and “no PA/SA” (n?=?14; 17 %) for patients infected with neither PA nor SA. The patients’ demographic
and clinical outcomes for each group are presented in Table 1. The identification of microorganisms other than PA and SA, namely Aspergillus and Stenotrophomonas maltophilia, in each of the 3 groups are summarized in Additional file 1: Table S1. The patients in the different infection groups were similar in age, sex
and BMI. There was a trend suggesting that “PA” patients were associated with worse
FEV
₁
and FVC % than “SA only” and “no PA/SA” patients. In univariate analyses, the infection
groups significantly differed in their exacerbation rate, clinical scores and CRP
levels. Across these clinical outcomes, “SA only” patients were similar to “no PA/SA”
patients, while “PA” patients experienced the highest rate of exacerbation, CRP and
lowest clinical score.

Table 1. Clinical characteristics of cohort patients based on infection status

To adjust for confounding factors that affect CF disease such the age and sex, we
next performed multivariable analyses. Although there was a trend suggesting that
“no PA/SA” patients had better FEV
₁
compared to “PA” patients using a linear regression model, we found no statistically
significant association between airway infection status and FEV
₁
, (Table 2). Next, we compared the exacerbation rate of the different airway infection groups
using a Poisson regression model, as presented in Table 3. Compared to “PA” patients, both “SA only” and those with “no PA/SA” were associated
with a significantly lower rate of pulmonary exacerbation. In the adjusted model,
“SA only” patients were associated with a relative risk of 0.2 (95 % CI [0.1-0.4])
compared to “PA” patients. We also examined several secondary outcomes, namely CRP
(Table 4) and the modified Matouk total clinical score (Table 5), a clinical subscore (which quantifies signs, symptoms and complications of CF disease).
The “SA only” patients were associated with reduced CRP levels (in both unadjusted
and adjusted models) and higher clinical scores (unadjusted model only) compared to
“PA” patients. Interestingly, patients with “no PA/SA” showed similar results to “SA
only” patients. Taken together, these analyses suggest that “SA only” patients overall
have significantly better clinical outcomes than “PA” patients.

Table 2. Association between infection status and FEV
₁
%

Table 3. Association between infection status and exacerbation rate

Table 4. Association between infection status and C-reactive protein

Table 5. Association between infection status and total clinical score

In all four crude models, “no PA/SA” and “SA only” groups had better total clinical
scores and subscores than the reference “PA only” group. These trends were also observed
in the adjusted models, although only the comparison between “no PA/SA” and “PA only”
groups remained statistically significant. Consistent with the observation that patients
infected with “no PA/SA” and “SA only” had less active disease than those infected
with “PA only”, their plasma CRP levels were lower in both crude and adjusted models.

Among PA infected patients in the “PA” group, 10 out of 50 were co-infected with both
PA and SA. The demographic and clinical characteristics of patients based on their
infection group, with the “PA only” and “PA?+?SA” groups separated, are presented
in the Additional file 1: Table S2, and both groups showed no significant differences (Table 6). In order to determine whether including the “PA?+?SA” patients to the “PA” group
had a significant effect on our results, we excluded the “PA?+?SA” patients and repeated
the univariate and multivariate analyses. As shown in the Additional file 1: Table S3 to S6, all results were similar. Furthermore, 4 out of 84 patients (4.8 %)
in our cohort were MRSA, and all were co-infected with both PA and SA. The results
of all univariate and multivariate analyses remained unchanged after exclusion of
the MRSA patients (data not shown).

Table 6. Clinical characteristics of cohort patients based on infection status (“no PA/SA”
excluded)