Summary
Rationale: Chronic mucoid Pseudomonas aeruginosa within the airway in cystic fibrosis (CF) patients can determine prognosis. Understanding the risk factors of mucoid P. aeruginosa acquisition may change how we deliver care. This study aims to evaluate whether presence of risk factors reported to predict disease severity including gender, CFTR genotype, bacterial organisms in airway cultures, and serum levels of vitamins A and E, albumin, C-reactive protein, alpha 1-antitrypsin, and immunoglobulins increased the risk of mucoid P. aeruginosa acquisition. Methods: Primary endpoint was age at first transition from negative to positive culture for mucoid P. aeruginosa. Cox proportional hazards regression with time-dependent covariates examined development of mucoid P. aeruginosa infection and its association with longitudinally measured serum biomarkers, pulmonary function, and culture results for other organisms. Results: Median ages at CF diagnosis and at first culture were 0.55 and 5.7 years, respectively. Median number of cultures/patient was 17. Of the 323 subjects, 150 developed mucoid P. aeruginosa during a median 8.1 years’ follow-up. In multivariate analysis, gender (relative hazard [RH] 0.55 for male vs. female, P=0.001), number of DF508 alleles (RH 1.66 for1 or 2 vs. 0, P=0.04), FEV1 % (RH 1.16 for 10% decrease, P=0.008), and most recent Staphylococcus aureus status (RH 0.24 for positive vs. negative, P< 0.0001) remained statistically significant. Conclusion: Female gender, number of DF508 alleles, decreased lung function, and lack of S. aureus on recent sputum culture are important risk factors for early detection of mucoid P. aeruginosa.
Keywords: cystic fibrosis, Pseudomonas colonization
INTRODUCTION
Cystic fibrosis (CF) lung disease is characterized by chronic infection by Pseudomonas aeruginosa,1,2 with reported prevalence increasing with age. Chronic P. aeruginosa within the CF airway is a well-known determinant of prognosis;3 several studies demonstrate associations between P. aeruginosa colonization and subsequent lung function decline.4–6 While not the first pathogen to colonize the CF lung, P. aeruginosa is the most common. Early infection with other pathogens, such as S. aureus, may prime the airway for later infection by P. aeruginosa.7 Respiratory tract infections occur in 15–30% of infants, 30–40% of children, 60% of adolescents, and 80% of adults with CF.6,8,9 Bronchoscopy data confirm P. aeruginosa colonization in ≤20% of CF children by age 2,10 but, Burns et al. 1 showed that 39 of 40 CF young children <3 years of age had evidence of P. aeruginosa by serial oral pharyngeal cultures or antibody. Despite extensive inflammatory response to ward off colonization, most CF patients become infected with a dominant organism, usually P. aeruginosa, which can no longer be eradicated. Once this organism switches to the mucoid phenotype, lung function decline increases. Thus, the course of P. aeruginosa infection in most CF patients has three distinct stages: no P. aeruginosa; initial nonmucoid P. aeruginosa; and mucoid P. aeruginosa colonization.1,6,11 Progressive pulmonary disease associated with chronic mucoid P. aeruginosa and airway inflammation appears to play a major role on the morbidity and mortality in CF patients,12 and successful strategies have been developed to reduce chronic colonization by active early eradication.4,13–15 However, elucidation of the risk factors for mucoid P. aeruginosa acquisition may lead to changes in management of patients with CF.
Epidemiologic studies of P. aeruginosa in CF provide some insight into timing and risk factors for nonmucoid and mucoid P. aeruginosa colonization.6,16 A recent prospective study by Li et al.6 showed that mucoid P. aeruginosa played a much greater role in CF lung disease than nonmucoid P. aeruginosa in 56 children diagnosed by newborn screening. An older retrospective study by Demko et al.17 reports the timing between initial nonmucoid P. aeruginosa cultures and the appearance of mucoid P. aeruginosa. P. aeruginosa colonization has been associated with meconium ileus, number of hospital admissions and early diagnosis of CF.18 Another study by Wang et al.19 suggests no difference in risk of P. aeruginosa acquisition between children diagnosed as newborns or in early childhood. Subsequent studies by Maselli et al.16 of P. aeruginosa acquisition in children identified by newborn screen show a positive association among female gender, homozygous DF508 mutation, and S. aureus isolation and early detection of P. aeruginosa (median age of acquisition 8.1 years). Long term use of oral antibiotics that have no in vitro activity against Pseudomonas and integration of CF infants with older CF patients were also associated with increased risk of P. aeruginosa infection.20–22 Finally, clinic exposures, aerosol use, lower level of mother’s formal education, and female gender were significantly associated with earlier acquisition of mucoid P.aeruginosa.17,23,24 However, none of these studies evaluated the development of mucoid P. aeruginosa infection and its association with serum biomarkers, S. aureus colonization, and lung disease progression.
With extensive clinical registry data collected at our center, we can evaluate whether factors that have been reported to predict lung disease severity increase the risk of P. aeruginosa acquisition. Of these, we analyzed gender, CFTR genotype, organisms in airway cultures, and serum levels of vitamins A and E, albumin, C-reactive protein (CRP), alpha 1-antitrypsin (AAT), and immunoglobulins.25,26 Hypothesizing that some of these biomarkers and/or colonization with a specific organism predict mucoid P. aeruginosa colonization, we explored risk factors for acquisition of P. aeruginosa in infants, children, and adults with CF followed at Children’s Hospital Boston. We report unique results on development of mucoid P. aeruginosa infection and its association with S. aureus and lung disease progression in patients with CF.
METHODS
Study Population and Data Collection
In a study approved by the Hospital’s Institutional Review Board, we examined all CF patients followed from 1993 to 2005 who were registered within a clinical and laboratory database at Children’s Hospital Boston. The diagnosis of CF was documented in the medical record by pilocarpine iontophoresis sweat test (sweat chloride >60 mmol/L). The primary endpoint was the age at first transition from culture-negative to culture-positive for mucoid P. aeruginosa. Accordingly, patients were selected for analysis if they had two or more culture results available in the database, the first of which was negative for mucoid P. aeruginosa. Those patients who were negative for mucoid P. aeruginosa at first culture include both subjects with cultures negative for any P. aeruginosa as well as subjects with cultures positive for nonmucoid P. aeruginosa. Biomarker measures analyzed included those described in the literature as being correlated with lung function including serum CRP, total serum IgE, total serum IgG, serum AAT, albumin, vitamin A and E levels, total white blood cell and neutrophil counts, and neutrophil percent of total white blood cell count.25,26 For each patient, we extracted all laboratory values and lung function test results from the medical record database at Children’s Hospital Boston. Specifically, for pulmonary function testing, microbiologic data, genotype, and biomarker analysis data were extracted from the laboratory databases based on a CF diagnosis and downloaded into an ORACLE database. A structured query language reporting tool was run to join the hospital-wide laboratory values requested with the CF patient population followed at Children’s Hospital Boston. Given the method for extracting data from various electronic sources and merging them, it was not possible to obtain an individual patient symptom history or medication information. However, as Children’s Hospital Boston is an accredited CF Care Center, patients received standard CF care as outlined by the CF Consortium guidelines. Most patients did not receive antistaphylococcal prophylaxis. Anti-pseudomonas antibiotics were routinely given to patients with two positive CF cultures (either documented by deep throat cultures or bronchoalveolar lavage) in an attempt to eradicate P. aeruginosa.
Sputum Culture Ascertainment
Sputum and deep throat swab specimens were cultured for bacterial pathogens. Colonization was defined as one positive microbiologic growth on culture. Cultures were performed in the Infectious Diseases Diagnostic Division at Children’s Hospital Boston. Samples were plated on selective media for isolation of important CF pathogens: Trypticase soy agar with 5% sheep blood, MacConkey agar, oxidative-fermentative polymyxin B-bacitracin lactose agar, Haemophilus isolation agar and mannitol salt agar. Identification of organisms was performed by standard techniques, including the use of biochemical panels. In addition to P. aeruginosa, other organisms which were selectively cultured and identified included S. aureus, Haemophilus influenzae, Stenotrophomonas maltophilia, Achromobacter xylosoxidans, Enterobac-teriaceae, and Burkholderia species in the B. cepacia complex.
CFTR Genotype Analysis
Genomic DNA isolated from each subject was evaluated for the presence of any of 1000 CFTR gene mutations (Genzyme, Cambridge, MA or Ambry Genetics, Aliso Viejo, CA) as part of their clinical evaluation.
Pulmonary Function Measurements
Forced expiratory volume in 1 sec (FEV1), forced vital capacity (FVC), and mid-peak flows (FEF25–75) were determined by standard spirometry meeting American Thoracic Society criteria; absolute values were converted to a percentage of the predicted volume expected for a healthy individual of the same age, sex, and height on the basis of the regression equations developed by Knudson et al.27,28
STATISTICAL METHODS
We defined the “baseline” level of longitudinally measured laboratory and pulmonary function test measurements for an individual as the most recent measurement made on or before the date of the first culture for that subject or, if there were no such measurements, as the earliest available measurement taken after the first culture. In analyses of age at first positive culture, follow-up time was measured from birth, but each subject entered the risk set at the time of the first culture. Median age at first culture was estimated by a product-limit type estimate of the baseline hazard function from a Cox proportional hazards model with no covariates and late entry into the risk set.
The analysis of risk factors for acquisition of mucoid Pseudomonas also used the Cox model but with covariates. In this analysis, highly skewed laboratory measurements were transformed to a logarithmic scale for analysis (absolute WBC, lymphocyte, neutrophil and eosinophil counts, and serum levels of AAT, CRP, IgA, IgG, and IgM). Zero values were replaced by small positive values when necessary before taking logarithms. Modeling of risk factors proceeded as follows: except for gender and genotype, all covariates were measured longitudinally; laboratory and PFT measurements were analyzed as time-dependent covariates, with the most recently measured value as the covariate. In a preliminary analysis, the six PFT measurements (absolute liters and percent predicted for each of FEV1, FVC, and FEF25–75) were considered as a group in order to reduce the number of these highly correlated variables to a single predictor. When considering culture results as risk factors, we constructed two time-dependent indicator variables: one for any prior lifetime history of a positive culture and another for the most recent result. By considering these covariates simultaneously in a model, we were able to test whether a lifetime history of a positive culture was associated with any risk in addition to “current” status. We checked whether the relative hazard for each covariate in our final model varied according to age (≤ 10 vs. > 10 to ≤20 vs. >20 years) by testing for interaction between the covariate and age using a likelihood ratio test.
In additional analyses exploring the relationship between S. aureus and P. aeruginosa, age-specific prevalences of positive cultures were calculated by generalized estimating equations to account for clustering by subject. (Each subject could contribute several observations within an age interval.) In addition, we calculated the within-subject odds ratio between positivity for these two organisms using conditional logistic regression. This analysis uses only within-subject information to estimate the odds ratio and controls for all nontime-dependent patient characteristics by using each subject as its own control.
SAS version 9.1 was used for all analyses. Two-sided P values are reported, with P values < 0.05 considered statistically significant.
RESULTS
Of 542 CF patients identified, 10 had no culture results available and 18 had only one, leaving 514 subjects with two or more cultures. Of these, 323 were negative and 191 were positive for mucoid P. aeruginosa at first culture. Table 1 shows the clinical characteristics of these two cohorts. Although similar by gender and genotype, those who were positive for mucoid P. aeruginosa at the first culture tended to be much older and have a longer history of CF before the first available culture result. The remainder of this analysis was restricted to the 323 patients who were negative for mucoid P. aeruginosa at first culture. This baseline was obtained at a median age of 5.7 years. The median number of cultures per patient was 17, collected over a median follow-up of 8.1 years. Table 2 shows characteristics of the CF cohort, who were negative for mucoid P. aeruginosa at first culture, at baseline. The median age at CF diagnosis was 0.55 years with 90th percentile 7.82 years. The cohort was 52% male. Forty-six percent of the patients were DF508 homozygous and 35% were heterozygotes. Thirty-four percent of the cohort had positive cultures for nonmucoid P. aeruginosa. Clinical serum marker levels and lung function measurements confirm that the values were performed at baseline and not during a pulmonary exacerbation. For laboratory values and lung function measurements, when a value measured at or before the first culture date was not available, the earliest measurement after the first culture was used as the baseline value. For the blood cell counts and other laboratory measurements, the baseline values were measured at or before the first culture in approximately 50–70% of subjects; in 75–90% of subjects the first measurement was no later than 12 months after the first culture. For lung function measurements, approximately 80% were measured at or before and 97% were measured no later than 12 months after the first culture. Table 3 shows the number and percentage of patients who were culture-positive for each organism at any time during follow-up. Also shown are the actuarial estimates of median age at first positive culture. For mucoid P. aeruginosa this median age is 11.3 years. Eleven patients died between ages 10 and 45 years (median 24 years).
TABLE 1.
Characteristics of Patients by Mucoid P. aeruginosa Status at First Culture
| Mucoid P. aeruginosa positive (N = 191) |
Mucoid P. aeruginosa negative (N =323) |
P-value | |
|---|---|---|---|
| Male gender | 96(52%) | 167 (52%) | 0.78 |
| # DF508 alleles1 | 0.69 | ||
| 0 | 34(20%) | 59 (19%) | |
| 1 | 65(38%) | 108 (35%) | |
| 2 | 71(42%) | 141 (46%) | |
| Age at first culture (years) | 18.9 (9.2, 34.4) | 5.7 (0.3, 21.0) | <0.001 |
| Years from diagnosis to first culture2 | 14.6 (2.5, 29.3) | 1.9 (0.0, 16.5) | <0.001 |
For categorical characteristics, data are n (%) and P values are from Fisher’s exact test. For continuous characteristics, data are median (10th percentile, 90th percentile) and P values are from the Mann– Whitney–Wilcoxon test.
N = 170 mucoid P. aeruginosa positive, N =308 mucoid P. aeruginosa negative with genotype data.
N = 170 mucoid P. aeruginosa positive, N = 300 mucoid P. aeruginosa negativewith known date ofdiagnosis.
TABLE 2.
Baseline Characteristics for 323 Patients Negative for MucoidP. aeruginosa at First Culture
| Characteristic | N | n (%) or median (10th, 90th)1 |
|---|---|---|
| Male gender | 323 | 167 (52%) |
| Number of DF508 alleles | 308 | |
| 0 | 59 (19%) | |
| 1 | 108 (35%) | |
| 2 | 141 (46%) | |
| Age at diagnosis (years) | 300 | 0.55 (0.02, 7.82) |
| Culture status at baseline | ||
| Nonmucoid P. aeruginosa positive | 323 | 110 (34%) |
| Staphylococcus aureus positive | 323 | 251 (78%) |
| Stenotrophomonas maltophilia positive | 323 | 2 (0.6%) |
| Burkholderia cepacia complex positive | 323 | 3 (0.9%) |
| Blood counts | ||
| WBC count (× 1,000/mm3) | 323 | 10.15 (6.53, 14.84) |
| Lymphocyte count (× 1,000/mm3) | 321 | 3.46 (1.73, 6.74) |
| Lymphocyte percent | 321 | 37.8 (18.0, 57.0) |
| Neutrophil count (× 1,000/mm3) | 321 | 4.6 (2.6, 9.1) |
| Neutrophil percent | 321 | 49.1 (30.3, 71.4) |
| Eosinophil count (× 1,000/mm3) | 321 | 0.19 (0.07, 0.47) |
| Eosinophil percent | 321 | 2.0 (0.8, 4.9) |
| Other markers | ||
| CRP (mg/dl) | 316 | 0.00 (0.00, 1.03) |
| AAT (mg/dl) | 321 | 151 (109, 216) |
| Total protein (g/dl) | 320 | 6.9 (5.7, 7.9) |
| Albumin (g/dl) | 323 | 3.8 (3.2, 4.2) |
| IgA (mg/dl) | 321 | 101 (26, 256) |
| IgG (mg/dl) | 321 | 781 (328, 1480) |
| IgM (mg/dl) | 321 | 93 (42, 190) |
| Vitamin A (mcg/dl) | 319 | 34.8 (16.0, 53.0) |
| Vitamin E (mg/L) | 319 | 7.6 (3.0, 14.3) |
| Lung function | ||
| FEV1 percent of predicted | 293 | 98 (66, 128) |
| FVC percent of predicted | 293 | 102 (76, 127) |
| FEF25–75 percent of predicted | 293 | 84 (28, 129) |
| FEV1 liters | 293 | 1.31 (0.79, 2.90) |
| FVC liters | 293 | 1.54 (0.93, 3.73) |
| FEF25–75 liters | 293 | 1.48 (0.62, 2.92) |
10th and 90th percentiles.
TABLE 3.
Number and Percent of Patients Who Tested Positive for Each Organism at Some Time During Follow-Up
| Organism | n (%) | Median age (years) at first positive |
|---|---|---|
| Nonmucoid P. aeruginosa | 262 (81%) | 2.9 |
| Mucoid P. aeruginosa | 150 (46%) | 11.3 |
| S. aureus | 310 (96%) | 0.46 |
| S. maltophilia | 50 (15%) | 28.6 |
| Burkholderia cepacia complex | 32 (10%) | Not reached |
Median age at first positive culture is also shown.
Table 4 depicts the univariate analysis of the association between each potential risk factor and the risk of acquisition of mucoid P. aeruginosa. Gender, decreased lung function, and decreased vitamin A levels were significantly associated with an elevated risk of mucoid P. aeruginosa. In addition, both a lifetime history and the most recent status of infection with S. aureus were associated with a lower risk of P. aeruginosa. Other characteristics that were marginally significant (P < 0.10) include presence of at least one DF508 allele, lifetime history of S. maltophilia infection, and the most recent serum levels of neutrophils, AAT, total protein, and vitamins A and E. These variables were retained for further consideration in multivariable models.
TABLE 4.
Univariate Proportional Hazards Models Predicting First Culture Positive for Mucoid P. aeruginosa
| Covariate | Comparison | RH | 95% CI | P-value |
|---|---|---|---|---|
| Gender | Male vs. female | 0.60 | (0.43, 0.83) | 0.002 |
| DF508 alleles | 1 or 2 vs. 0 | 1.54 | (0.98, 2.41) | 0.06 |
| Culture history | ||||
| S. aureus | Ever vs. never + | 0.38 | (0.25, 0.58) | <0.0001 |
| S. aureus | Recent + vs. – | 0.23 | (0.17, 0.32) | <0.0001 |
| BCC | Ever vs. never + | 1.32 | (0.61, 2.86) | 0.48 |
| BCC | Recent + vs. – | 0.64 | (0.16, 2.63) | 0.54 |
| S. maltophilia | Ever vs. never + | 1.58 | (0.97, 2.57) | 0.07 |
| S altophilia | Recent + vs. — | 1.24 | (0.44, 3.51) | 0.69 |
| Lung function | ||||
| FEV1 % | 10% points ↓ | 1.20 | (1.11, 1.30) | <0.0001 |
| FEV1 liters | 1 L ↓ | 1.75 | (1.32, 2.30) | <0.0001 |
| FVC % | 10% points | 1.19 | (1.09, 1.30) | 0.0001 |
| FVC liters | ↓ 1 L ↓ | 1.56 | (1.22, 2.01) | 0.0004 |
| FEF25–75 % | 10% points | 1.12 | (1.05, 1.18) | 0.0002 |
| FEF25–75 liters | ↓ 1 L ↓ | 1.40 | (1.16, 1.70) | 0.0005 |
| Blood counts | ||||
| WBC count | Twofold ↑ | 1.20 | (0.85, 1.68) | 0.30 |
| Lymphocyte count | Twofold ↑ | 1.13 | (0.83, 1.55) | 0.43 |
| Lymphocyte % | 10% points ↑ | 0.95 | (0.80, 1.13) | 0.55 |
| Neutrophil count | Twofold ↑ | 1.26 | (0.97, 1.63) | 0.09 |
| Neutrophil % | 10% points ↑ | 1.05 | (0.90, 1.22) | 0.57 |
| Eosinophil count | Twofold ↑ | 1.00 | (0.86, 1.16) | 0.99 |
| Eosinophil % | 1% point ↑ | 1.00 | (0.92, 1.09) | 0.95 |
| Other markers | ||||
| AAT | Twofold ↑ | 1.65 | (0.98, 2.77) | 0.06 |
| Albumin | 1 g/dl ↑ | 0.77 | (0.51, 1.17) | 0.22 |
| CRP | Twofold ↑ | 1.01 | (0.95, 1.08) | 0.70 |
| Total protein | 1 g/dl ↑ | 0.81 | (0.63, 1.03) | 0.09 |
| IgA | Twofold ↑ | 1.05 | (0.89, 1.25) | 0.55 |
| IgG | Twofold ↑ | 1.10 | (0.76, 1.59) | 0.62 |
| IgM | Twofold ↑ | 1.04 | (0.84, 1.31) | 0.70 |
| Vitamin A | 1 mcg/dl ↑ | 0.98 | (0.97, 0.99) | 0.008 |
| Vitamin E | 1 mg/L ↑ | 0.96 | (0.93, 1.00) | 0.08 |
RH, relative hazard; CI, confidence interval; BCC, Burkholderia cepacia complex; FEV1, forced expiratory volume in 1 sec; FVC, forced vital capacity; FEF25–75, mid peak flows.
As a preliminary step in multivariable modeling, we investigated the two S. aureus measures in adjusted analyses. When lifetime culture history and most recent culture status of S. aureus were considered together in a model, only the recent status was significant [recent positive, RH = 0.20, 95% CI (0.14, 0.30), P < 0.0001; lifetime history of any positive, RH = 1.35, 95% CI (0.83, 2.20), P= 0.23]. On the basis of these results, we considered only the most recent S. aureus status in further multivariable models.
Six different lung function measures were all highly significant in univariate analysis. We explored them as a group with a view toward reducing the number of correlated predictors before building multivariable models. For each of FEV1, FVC, and FEF25–75, when the pair of variables (e.g., percent predicted FEV1 and FEV1 liters) was considered together in a model, the liter variable was clearly the least significant (all P values > 0.16). We then considered the three percent predicted variables in pairs. After adjusting for percent predicted FEV1, the other variables both had P values > 0.54. Therefore, we retained only percent predicted FEV1 in further models.
Starting with the highly significant risk factors from the univariate analysis and preliminary modeling steps, we considered adding other candidate factors one at a time and arrived at the final model shown in Table 5. This model includes gender, genotype, percent predicted FEV1, and most recent S. aureus culture status. All of these predictors remained statistically significant when any of the other candidate predictors was added to this model, and none of the other predictors was significant (all P values ≥ 0.10).
TABLE 5.
Multivariable Proportional Hazards Models Predicting First Culture Positive for Mucoid P. aeruginosa
| Covariate | Comparison | RH | 95% CI | P-value |
|---|---|---|---|---|
| Gender | Male vs. female | 0.55 | (0.38, 0.79) | 0.001 |
| DF508 | 1 or 2 vs. 0 alleles | 1.66 | (1.02, 2.72) | 0.04 |
| FEV1 % | 10% point ↓ | 1.16 | (1.06, 1.26) | 0.0008 |
| S. aureus | Most recent + vs. – | 0.24 | (0.17, 0.34) | <0.0001 |
We also considered modeling DF508 genotype as a three-level variable (0, 1, 2 alleles). The relative hazards (RH) for 1 versus 0 alleles (RH = 1.63) and for 2 versus 0 alleles (RH = 1.69) were very similar, therefore the two-level variable used in our final model (RH = 1.66 for 1 or 2 alleles vs. 0 alleles) is preferred. In order to evaluate possible biases introduced by inclusion of subjects whose first available culture was at a late age, and subjects whose first available lung function test results were after their first culture results, we re-evaluated our final model restricted to two subgroups: subjects whose first available culture was at age <10 years, and subjects whose “baseline” lung function was measured at or before the date of their first available culture. The results were very similar to those shown in Table 5, suggesting that conclusions were not influenced by inclusion of subjects with missing early data. Finally, we checked for heterogeneity of the RH across age groups; none was found.
The observation that S. aureus colonization is associated with a lower risk of P. aeruginosa colonization was investigated further. Table 6 shows age-specific prevalence of these two organisms. The prevalence estimates take into account the clustering due to each subject’s possible contribution of several samples within a single age interval. From birth to adulthood, age-specific prevalences of mucoid P. aeruginosa increase from less than 5% to over 40%. S. aureus prevalence is increased slightly to just over 80% in the first 15 years, then decreases, especially after age 24. These trends are broadly consistent with the hypothesis that the presence of S. aureus is protective against mucoid P. aeruginosa but do not demonstrate a within-subject association. In a more direct comparison, we used conditional logistic regression to estimate the within-subject odds ratio (OR) between the two prevalences. The association was again consistent with the proportional hazards analysis: OR = 0.43, 95% CI (0.34, 0.53), P < 0.001. To check the possibility of residual confounding by age, we added age to the model. The OR changed only slightly, OR = 0.41, 95% CI (0.32, 0.51), P < 0.001.
TABLE 6.
Age-Specific Prevalences of Mucoid P. aeruginosa and S. aureus
| Prevalence |
|||||
|---|---|---|---|---|---|
| Age (years) | Number of Samples |
Number of subject |
Samples per subject |
Mucoid P. aeruginosa (%) |
S. aureus (%) |
| P. aeruginosa (%) | (%) | ||||
| 0–4 | 1,071 | 155 | 6.9 | 3.5 | 71.9 |
| 5–9 | 1,470 | 199 | 7.4 | 10.4 | 75.0 |
| 10–14 | 1,373 | 168 | 8.2 | 16.5 | 81.5 |
| 15–19 | 963 | 124 | 7.8 | 26.1 | 77.2 |
| 20–24 | 450 | 68 | 6.6 | 26.6 | 77.3 |
| 25–29 | 259 | 34 | 7.6 | 37.7 | 63.7 |
| 30–34 | 170 | 22 | 7.7 | 44.0 | 62.0 |
| ≥35 | 160 | 19 | 8.4 | 41.0 | 56.6 |
DISCUSSION
We examined clinical registry data for risk factors that may be associated with the initial detection of mucoid P. aeruginosa from a cohort of 323 CF patients followed at Children’s Hospital Boston. Our findings indicate that female gender, presence of at least one DF508 allele, lower percent predicted FEV1, and a negative S. aureus sputum culture were predictive of mucoid P. aeruginosa. The relative hazards for mucoid P. aeruginosa colonization in males versus females was 0.55, and the earlier acquisition of mucoid P. aeruginosa in females has been previously postulated to contribute to their diminished survival.17 These genotype and gender findings are consistent with results demonstrating that both of these characteristics are associated with initial detection of P. aeruginosa and poor outcomes in CF. Our findings confirm a gender difference for mucoid P. aeruginosa acquisition likely contributing to the more rapid lung function decline in females with CF.16,17,29,30 CF patients with 1 or 2 versus 0 DF508 mutations had an increased risk of mucoid P. aeruginosa. As a deficiency in CFTR protein in CF patients is thought to impair the clearance of P. aeruginosa,31,32 this finding is consistent with the decreased CFTR function evident in patients homozygous for DF508. Similarly, while there is genetic heterogeneity in patients homozygous for the DF508 mutation, patients with DF508 mutations generally have more severe lung disease than CF patients with either one DF508 allele or other non-DF508 genotypes.33 At any given age, among people who never tested positive for mucoid P. aeruginosa, those with lower lung function are more likely to convert to positive mucoid P. aeruginosa than those with higher lung function. While these results suggest that decreased lung function is a predecessor to mucoid P. aeruginosa, representative papers in the literature showed that mucoid P. aeruginosa may also be associated with a subsequent further decline in lung function.3,6 We also found that laboratory serum biomarkers did not predict mucoid P. aeruginosa acquisition in the multivariate analysis. Finally, we found that a CF patient’s airway does become colonized by progression from no P. aeruginosa, initial nonmucoid P. aeruginosa, and then mucoid P. aeruginosa,1,6,11 confirming prior studies. This progression could be due to the host’s process in mediating innate immunity to mucoid P. aeruginosa infection. An ongoing inflammatory response likely acts as a “double-edged sword” and suggests that: mediators of immunity and inflammation, when produced in a physiologically appropriate fashion in response to infection, mediate resistance; failure of this response leads to infection; and ongoing production of the same mediators contributes to immunopathology and lung function decline.
In addition and perhaps more importantly, we found that a positive S. aureus culture is associated with a lower risk of mucoid P. aeruginosa. This finding contradicts previous investigations that demonstrate an association between more frequent positive S. aureus cultures and existing infection with mucoid P. aeruginosa.16,34 However, the Patient Registry 1999 Annual Data Report from the Cystic Fibrosis Foundation8 showing a rise in P. aeruginosa, as S. aureus colonization diminishes, supports our results. We found that both a lifetime history of S. aureus infection and a recent history of S. aureus infection were associated with a lower risk of first acquisition of P. aeruginosa colonization. However, when both were considered together, only the recent history was significant. Coupled with the observations that S. aureus tends to be acquired early in life and maintains a high age-specific prevalence, this finding suggests that, if S. aureus has a protective effect, it is short-lived. It may be that transition of S. aureus status from positive to negative creates an opportunity for P. aeruginosa colonization. The conditional logistic regression analysis also supports this idea by demonstrating that, within an individual, a change in infection status from positive to negative for one organism tends to be associated with a change in the opposite direction for the other organism. Lending further support to these results, studies by Ratjen et al.20 and Stutman et al.21 found that patients receiving continuous anti-staphylococcal antibiotic therapy had a significantly higher rate of P. aeruginosa acquisition than patients receiving only intermittent or no antibiotic therapy. They concluded that continuous therapy with antistaphylococcal antibiotics directed against S. aureus increases the risk of colonization with P. aeruginosa. Although the use of continuous therapy with antistaphylococcal antibiotics remains controversial,35 our analysis also suggests that this strategy favors early colonization with P. aeruginosa.
While our study design and analyses are unique and some of our findings novel, our study has certain limitations. We, as well as others, used oropharyngeal (OP) and/or sputum cultures16,19,22,24 in our analyses, which may have imposed a selection bias as the younger patients, unable to expectorate, may not have provided as many samples as the older patients. However, as depicted in Table 6, the number of cultures per subject was fairly constant across age strata. Although mucoid P. aeruginosa is considered the predominant respiratory bacterial pathogen in adult CF patients with one particular study showing a prevalence rate of 92% in 16 year olds, our study shows lower prevalence rates, for example 26% in 15–19 year olds (Table 6). In comparison to the Li et al. paper referenced, we have a much larger sample size contributing to prevalence estimates with 124 patients contributing to the 15–19 age range compared with 13 patients age 16 in the Li paper. The lack of inclusion of a chest X-ray assessment either by a modified Shwachman or other scoring system could also be perceived as a limitation of this study especially in light of recent studies showing that chest X-ray changes may be more sensitive in detecting early pulmonary involvement in CF than spirometry.36 However, there is no agreed upon chest X-ray or CT scan scoring system among CF centers.
The limitations of our registry database include the taking of measurements at multiple time points for clinical reasons and not in the consistent method usual for study purposes. The historical database is also clustered toward more data on the sicker CF patients, as these patients would have more clinic visits and values in the database. These limitations should not affect the reproducibility of our results. The strengths of this database include having numerous measurements on a large cohort of CF patients, with multiple observations, followed at one center over a 13-year period. Collection of study subjects from a single center affords some control over variation in environmental factors, which are difficult to control in multicenter studies.
In conclusion, female gender, presence of at least one DF508 allele, decreased lung function, and a lack of S. aureus isolation on the most recent sputum culture are important risk factors for early detection of mucoid P. aeruginosa. Recognition of these risk factors should lead to earlier aggressive antibiotic and efficacious treatments to delay chronic infection with mucoid P. aeruginosa. Furthermore, our analysis suggests that prophylactic anti-staphylococcal antibiotics may favor earlier colonization with mucoid P. aeruginosa.
ACKNOWLEDGMENTS
We wish to thank Dr. Mary Ellen Wohl for the inspiration for this research. We acknowledge the assistance of the Cystic Fibrosis Center at Children’s Hospital Boston and its participants. We also wish to thank Jeanne Greeno, Data Warehouse Team Leader, and Paul O’Byrne, Manager of the Children’s Hospital ISD Knowledge Management Group, and Nancy Shotola, Director of the Pulmonary Function Laboratory at Children’s Hospital Boston, for their assistance with the collection of the study data. Dr. Alexander McAdam, Medical Director of the Infectious Disease Diagnostic Division at Children’s Hospital Boston, provided assistance with the CF cultures. Nathaniel Weller, Soma Datta, Brooke Schuemann, and Daniel Asher provided technical expertise and support. The authors also thank Drs. Michael Knowles, Thomas Martin and Diane Gold for critical reading of the manuscript.
Grant sponsor: National Heart, Lung and Blood Institute; Grant number: 5 K23 HL074202; Grant sponsor: Division of Pediatric Pulmonology, Children’s Hospital Boston; Grant number: 98404-14.
REFERENCES
- 1.Burns JL, Gibson RL, McNamara S, Yim D, Emerson J, Rosenfeld M, Hiatt P, McCoy K, Castile R, Smith AL, Ramsey BW. Longitudinal assessment of Pseudomonas aeruginosa in young children with cystic fibrosis. J Infect Dis. 2001;183:444–452. doi: 10.1086/318075. [DOI] [PubMed] [Google Scholar]
- 2.Henry R, Mellis C, Petrovic L. Mucoid Pseudomonas aeruginosa is a marker of poor survival in cystic fibrosis. Pediatr Pulmonol. 1992;12:158–161. doi: 10.1002/ppul.1950120306. [DOI] [PubMed] [Google Scholar]
- 3.Henry RL, Mellis CM, Petrovic L. Mucoid Pseudomonas aeruginosa is a marker of poor survival in cystic fibrosis. Pediatr Pulmonol. 1992;12:158–161. doi: 10.1002/ppul.1950120306. [DOI] [PubMed] [Google Scholar]
- 4.Hoiby N, Koch C. Cystic fibrosis 1. Pseudomonas aeruginosa infection in cystic fibrosis and its management. Thorax. 1990;45:881–884. doi: 10.1136/thx.45.11.881. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Konstan MW, Morgan WJ, Butler SM, Pasta DJ, Craib ML, Silva SJ, Stokes DC, Wohl ME, Wagener JS, Regelmann WE, Johnson CA. Risk factors for rate of decline in forced expiratory volume in one second in children and adolescents with cystic fibrosis. J Pediatr. 2007;151:134–139. doi: 10.1016/j.jpeds.2007.03.006. 139 e131. [DOI] [PubMed] [Google Scholar]
- 6.Li Z, Kosorok MR, Farrell PM, Laxova A, West SE, Green CG, Collins J, Rock MJ, Splaingard ML. Longitudinal development of mucoid Pseudomonas aeruginosa infection and lung disease progression in children with cystic fibrosis. JAMA. 2005;293:581–588. doi: 10.1001/jama.293.5.581. [DOI] [PubMed] [Google Scholar]
- 7.Lyczak JB, Cannon CL, Pier GB. Establishment of Pseudomonas aeruginosa infection: lessons from a versatile opportunist. Microbes Infect. 2000;2:1051–1060. doi: 10.1016/s1286-4579(00)01259-4. [DOI] [PubMed] [Google Scholar]
- 8.Cystic Fibrosis Foundation. Patient registry 1999 annual data report. [Google Scholar]
- 9.Rosenfeld M, Gibson RL, McNamara S, Emerson J, Burns JL, Castile R, Hiatt P, McCoy K, Wilson CB, Inglis A, Smith A, Martin TR, Ramsey BW. Early pulmonary infection, inflammation, and clinical outcomes in infants with cystic fibrosis. Pediatr Pulmonol. 2001;32:356–366. doi: 10.1002/ppul.1144. [DOI] [PubMed] [Google Scholar]
- 10.Armstrong DS, Grimwood K, Carlin JB, Carzino R, Olinsky A, Phelan PD. Bronchoalveolar lavage or oropharyngeal cultures to identify lower respiratory pathogens in infants with cystic fibrosis. Pediatr Pulmonol. 1996;21:267–275. doi: 10.1002/(SICI)1099-0496(199605)21:5<267::AID-PPUL1>3.0.CO;2-K. [DOI] [PubMed] [Google Scholar]
- 11.Fegan M, Francis P, Hayward AC, Davis GH, Fuerst JA. Phenotypic conversion of Pseudomonas aeruginosa in cystic fibrosis. J Clin Microbiol. 1990;28:1143–1146. doi: 10.1128/jcm.28.6.1143-1146.1990. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Davis P. Clinical pathophysiology and manifestations of lung disease. In: Yankaskas J, Knowles M, editors. Cystic fibrosis in adults. Philadelphia: Lippincott-Raven Publishers; 1999. pp. 45–67. [Google Scholar]
- 13.Gibson RL, Emerson J, Mayer-Hamblett N, Burns JL, McNamara S, Accurso FJ, Konstan MW, Chatfield BA, Retsch-Bogart G, Waltz DA, Acton J, Zeitlin P, Hiatt P, Moss R, Williams J, Ramsey BW. Duration of treatment effect after tobramycin solution for inhalation in young children with cystic fibrosis. Pediatr Pulmonol. 2007;42:610–623. doi: 10.1002/ppul.20625. [DOI] [PubMed] [Google Scholar]
- 14.Rosenfeld M, Ramsey BW, Gibson RL. Pseudomonas acquisition in young patients with cystic fibrosis: pathophysiology, diagnosis, and management. Curr Opin Pulm Med. 2003;9:492–497. doi: 10.1097/00063198-200311000-00008. [DOI] [PubMed] [Google Scholar]
- 15.Treggiari MM, Rosenfeld M, Retsch-Bogart G, Gibson R, Ramsey B. Approach to eradication of initial Pseudomonas aeruginosa infection in children with cystic fibrosis. Pediatr Pulmonol. 2007;42:751–756. doi: 10.1002/ppul.20665. [DOI] [PubMed] [Google Scholar]
- 16.Maselli JH, Sontag MK, Norris JM, MacKenzie T, Wagener JS, Accurso FJ. Risk factors for initial acquisition of Pseudomonas aeruginosa in children with cystic fibrosis identified by newborn screening. Pediatr Pulmonol. 2003;35:257–262. doi: 10.1002/ppul.10230. [DOI] [PubMed] [Google Scholar]
- 17.Demko CA, Byard PJ, Davis PB. Gender differences in cystic fibrosis: Pseudomonas aeruginosa infection. J Clin Epidemiol. 1995;48:1041–1049. doi: 10.1016/0895-4356(94)00230-n. [DOI] [PubMed] [Google Scholar]
- 18.Kerem E, Corey M, Stein R, Gold R, Levison H. Risk factors for Pseudomonas aeruginosa colonization in cystic fibrosis patients. Pediatr Infect Dis J. 1990;9:494–498. doi: 10.1097/00006454-199007000-00008. [DOI] [PubMed] [Google Scholar]
- 19.Wang SS, FitzSimmons SC, O’Leary LA, Rock MJ, Gwinn ML, Khoury MJ. Early diagnosis of cystic fibrosis in the newborn period and risk of Pseudomonas aeruginosa acquisition in the first 10 years of life: a registry-based longitudinal study. Pediatrics. 2001;107:274–279. doi: 10.1542/peds.107.2.274. [DOI] [PubMed] [Google Scholar]
- 20.Ratjen F, Comes G, Paul K, Posselt HG, Wagner TO, Harms K. Effect of continuous antistaphylococcal therapy on the rate of P aeruginosa acquisition in patients with cystic fibrosis. Pediatr Pulmonol. 2001;31:13–16. doi: 10.1002/1099-0496(200101)31:1<13::aid-ppul1001>3.0.co;2-n. [DOI] [PubMed] [Google Scholar]
- 21.Stutman HR, Lieberman JM, Nussbaum E, Marks MI. Antibiotic prophylaxis in infants and young children with cystic fibrosis: a randomized controlled trial. J Pediatr. 2002;140:299–305. doi: 10.1067/mpd.2002.121930. [DOI] [PubMed] [Google Scholar]
- 22.West SE, Zeng L, Lee BL, Kosorok MR, Laxova A, Rock MJ, Splaingard MJ, Farrell PM. Respiratory infections with Pseudomonas aeruginosa in children with cystic fibrosis: early detection by serology and assessment of risk factors. JAMA. 2002;287:2958–2967. doi: 10.1001/jama.287.22.2958. [DOI] [PubMed] [Google Scholar]
- 23.Farrell PM, Shen G, Splaingard M, Colby CE, Laxova A, Kosorok MR, Rock MJ, Mischler EH. Acquisition of Pseudomonas aeruginosa in children with cystic fibrosis. Pediatrics. 1997;100:E2. doi: 10.1542/peds.100.5.e2. [DOI] [PubMed] [Google Scholar]
- 24.Kosorok MR, et al. Comprehensive analysis of risk factors for acquisition of Pseudomonas aeruginosa in young children with cystic fibrosis. Pediatr Pulmonol. 1998;26:81–88. doi: 10.1002/(sici)1099-0496(199808)26:2<81::aid-ppul2>3.0.co;2-k. [DOI] [PubMed] [Google Scholar]
- 25.Fayon M. CF-emerging therapies: modulation inflammation. Paediatr Respir Rev. 2006;1:S170–S174. doi: 10.1016/j.prrv.2006.04.212. [DOI] [PubMed] [Google Scholar]
- 26.Levy H, et al. Inflammatory markers of lung disease in adult patients with cystic fibrosis. Pediatr Pulmonol. 2007;42:256–262. doi: 10.1002/ppul.20563. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Knudson RJ, Lebowitz MD, Holberg CJ, Burrows B. Changes in the normal maximal expiratory flow-volume curve with growth and aging. Am Rev Respir Dis. 1983;127:725–734. doi: 10.1164/arrd.1983.127.6.725. [DOI] [PubMed] [Google Scholar]
- 28.Ferris BG. Standardization of spirometry, 1994. Update. American Thoracic Society. Am J Respir Crit Care Med. 1995;152:1107–1136. doi: 10.1164/ajrccm.152.3.7663792. [DOI] [PubMed] [Google Scholar]
- 29.Davis PB. The gender gap in cystic fibrosis survival. J Gend Specif Med. 1999;2:47–51. [PubMed] [Google Scholar]
- 30.Schneiderman-Walker J, et al. Sex differences in habitual physical activity and lung function decline in children with cystic fibrosis. J Pediatr. 2005;147:321–326. doi: 10.1016/j.jpeds.2005.03.043. [DOI] [PubMed] [Google Scholar]
- 31.Pier GB. CFTR mutations and host susceptibility to Pseudomonas aeruginosa lung infection. Curr Opin Microbiol. 2002;5:81–86. doi: 10.1016/s1369-5274(02)00290-4. [DOI] [PubMed] [Google Scholar]
- 32.Pier GB, et al. Role of mutant CFTR in hypersusceptibility of cystic fibrosis patients to lung infections. Science. 1996;271:64–67. doi: 10.1126/science.271.5245.64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Johansen HK, Nir M, Hoiby N, Koch C, Schwartz M. Severity of cystic fibrosis in patients homozygous and heterozygous for delta F508 mutation. Lancet. 1991;337:631–634. doi: 10.1016/0140-6736(91)92449-c. [DOI] [PubMed] [Google Scholar]
- 34.Abman SH, et al. Early bacteriologic, immunologic, and clinical courses of young infants with cystic fibrosis identified by neonatal screening. J Pediatr. 1991;119:211–217. doi: 10.1016/s0022-3476(05)80729-2. [DOI] [PubMed] [Google Scholar]
- 35.Smyth A. Prophylactic antibiotics in cystic fibrosis: a conviction without evidence? Pediatr Pulmonol. 2005;40:471–476. doi: 10.1002/ppul.20305. [DOI] [PubMed] [Google Scholar]
- 36.Kosorok MR, et al. Acceleration of lung disease in children with cystic fibrosis after Pseudomonas aeruginosa acquisition. Pediatr Pulmonol. 2001;32:277–287. doi: 10.1002/ppul.2009.abs. [DOI] [PubMed] [Google Scholar]