Seasonality of acquisition of respiratory bacterial pathogens in young children with cystic fibrosis

Cystic fibrosis (CF), an autosomal recessive disorder, is characterized by chronic and recurring endobronchial infection, resulting in progressive structural lung disease and, generally, premature death. Beginning early in life, abnormal mucociliary clearance promotes colonization by a variety of bacteria [1, 2]. Haemophilus influenzae and methicillin-susceptible Staphylococcus aureus (MSSA) are frequently among the first organisms isolated from respiratory cultures [1]; however, their effect on clinical outcomes is debated [2]. Pseudomonas aeruginosa, a Gram-negative bacterium frequently first isolated from respiratory cultures in early childhood, is regarded as the most important CF-related respiratory pathogen [3, 4]. Other gram-negative organisms and methicillin-resistant S. aureus (MRSA) often follow, and appear to also adversely affect clinical outcomes [2, 5].

While risk factors for specific bacterial infections have been reviewed [2, 5, 6], little information is available regarding the incidence patterns of these infections. P. aeruginosa appears to be typically initially acquired from the natural environment; however, the sources for other organisms are far less clear. A more complete understanding of the incidence patterns of acquisition of other pathogens may provide an opportunity for the identification of potential aetiologic factors, as well as refinement of infection control strategies for these infections. Seasonal variation of infectious diseases is often observed [7] and increasing evidence suggests that many Gram-negative infections display such patterns [8, 9]. Previously, we reported differential seasonal acquisition of initial P. aeruginosa in young children with CF and no such variation for MSSA infections [10]. The purpose of this study was to describe and compare seasonal incidence for other common CF-related respiratory bacterial pathogens including MRSA, Stenotrophomonas maltophilia, Achromobacter xylosoxidans and Haemophilus influenzae, in young CF patients residing in the U.S.

A total of 4522 children met inclusion criteria and were included in the study population. During follow-up a total of 910 (20.1%) children acquired MRSA during 55,629 person-seasons of follow-up, 1161 (25.7%) children acquired S. maltophilia over 53,007 person-seasons, 228 (5.0%) acquired A. xylosoxidans over 60,688 person-seasons and 2148 (47.5%) of the children acquired H. influenzae during 42,508 person-seasons of follow-up. Among those acquiring each organism, the median age at acquisition of MRSA, S. maltophilia, A. xylosoxidans and H. influenzae, respectively, was 25 months (25th-75th percentiles: 14-42 months), 21 (13-35), 32 (18-48) and 20 (13-31) months. Overall, approximately 55% of individuals had a culture frequency of four per year.

Table 1 describes the demographic and clinical characteristics of the study population by acquisition status of each pathogen. Children acquiring each of the organisms were less likely to have been identified by newborn screening than those remaining infection-free. Those acquiring S. maltophilia and A. xylosoxidans were more likely to be Hispanic than those remaining free of these organisms, while the converse was true for MRSA and H. influenzae. The mean age of CF diagnosis was greater among children acquiring MRSA, A. xylosoxidans and H. influenzae than among those remaining free of those pathogens. Finally, patients who acquired MRSA, S. maltophilia and H. influenzae were more likely to have CF mutations resulting in minimal CFTR function.

Incidence of each of the pathogens, overall and by season, are presented in Table 2. H. influenzae incidence (50.5 per 1000 person-seasons) was highest overall, followed by A. xylosoxidans (21.9 per 1000 person-seasons), MRSA (16.4 per 1000 person-seasons), and S. maltophilia (3.8 per 1000 person-seasons). Peak incidence of MRSA (19.5 per 1000 person-seasons), A. xylosoxidans (4.9 per 1000 person-seasons), and H. influenzae (56.8 per 1000 person-seasons) was observed in the winter season, while S. maltophilia acquisition was highest in the summer season (23.6 per 1000 person-seasons).

Results of the Poisson regression evaluating the seasonality of each respiratory pathogen are presented in Table 3. Compared to winter, MRSA acquisition was less common in spring (IRR: 0.79; 95% CI: 0.65, 0.96) and summer (IRR: 0.69; 95% CI: 0.57, 0.84). A. xylosoxidans acquisition was also less likely in spring (IRR: 0.59; 95% CI: 0.39, 0.89). H. influenzae acquisition was less likely in summer (IRR: 0.88; 95% CI: 0.78, 0.99) and autumn (IRR: 0.78; 95% CI: 0.69, 0.88) seasons. No statistically significant seasonal differences were observed for S. maltophilia acquisition. Lastly, patterns of pathogen acquisition were similar within climate zones (Additional file 2).

In this study, seasonal variation was observed for rates of initial acquisition of MRSA, A. xylosoxidans, and H. influenzae in young U.S. children with CF, while no such variation was observed for S. maltophilia acquisition. Compared to winter season, MRSA acquisition was less likely in spring and summer and A. xylosoxidans acquisition was less likely in spring. Summer and autumn seasons were associated with decreased H. influenzae acquisition compared to the winter season. Strengths of this investigation included a large national cohort of young children with CF with frequent monitoring of respiratory microbiology.

Interestingly, we previously reported seasonal variations in P. aeruginosa acquisition in the same cohort of young CF patients; however, the seasonal patterns of acquisition differed, with higher incidence of P. aeruginosa in summer (IRR: 1.22; 95% CI: 1.07, 1.40) and autumn (IRR: 1.34; 95% CI: 1.18, 1.52) seasons compared to winter [10]. Also, while in the current analyses we observed seasonal variation in MRSA acquisition, in our prior study no seasonal variation in MSSA acquisition was observed.

The reason for these different seasonal patterns of pathogen acquisition is unclear. Seasonal factors can influence both host susceptibility and pathogen proliferation in the environment [7, 16‐18]. In CF patients, P. aeruginosa, a ubiquitous environmental organism, is typically acquired from the environment [19]. The source of other organisms is less clear but may include the indoor or outdoor environment, clinic or hospital settings and patient-to-patient transmission. We [20] and others [21] have shown that higher ambient temperatures increase the risk of P. aeruginosa acquisition among CF patients, which may partially explain our observation of a higher acquisition rate in the summer. On the other hand, viral respiratory infections, more prevalent in the winter both in CF patients and in the general population, are thought to increase airway susceptibility to bacterial colonization in CF patients [22‐24], which may explain the higher observed rate of acquisition of MRSA, A. xylosoxidans and H. influenza in the winter months.

MRSA colonization is associated with poorer clinical outcomes in CF patients [25, 26] and is now frequently cultured from the respiratory tracts of patients. The overall prevalence among U.S. CF patients has risen from approximately 2% in 2001 to 25% in 2012 and even in young children <2 and 2-6 years MRSA prevalence in 2012 was 11 and 18%, respectively [1]. In a multicenter study of the clonal distribution of MRSA isolates in children with CF, Champion and colleagues [27] reported that approximately 70% of all isolates were healthcare associated MRSA (HA-MRSA) strains. Further, recent studies have reported that HA-MRSA infections display a seasonal pattern in the general population, with peak rates occurring in winter months; this is in contrast to community acquired MRSA infections that peak in summer months [28, 29]. In the current investigation, MRSA incidence was highest in the winter season as would be consistent with HA-MRSA infections typically found in this population. Interestingly, we observed no seasonal variation in MSSA acquisition among young CF patients in our prior study [10]; previous investigations of seasonal variations of nasal colonization of MSSA in the general population have reported mixed results [30].

Due to rare occurrences of S. maltophilia and A. xylosoxidans respiratory infections in the general population, evaluation of seasonal patterns has been limited and precludes comparison to results obtained herein. To our knowledge seasonal acquisition rates of these pathogens have not previously been described in the CF population. Accordingly, Burkholderia cepacia, a Gram-negative bacterium ubiquitous in the natural environment, is now recognized as a clinically significant pathogen in CF patients [2]. In this study meaningful analyses of B. cepacia could not be performed due to the relatively few cases (n = 41) of B. cepacia in this young population. During follow-up a total of 18, 10, 10, and three cases of initial B. cepacia were reported in the spring, summer, autumn, and winter seasons, respectively.

The clinical significance of first isolation of bacterial pathogens from respiratory cultures in young CF patients is a topic of debate and likely varies by organism [5]. While initial P. aeruginosa colonization can spontaneously clear, it is likely to develop into chronic infection, which is nearly impossible to eradicate and is associated with poorer outcome and survival [31, 32]. Thus standard of care around the world is to attempt to eradicate initial P. aeruginosa regardless of symptoms. For other organisms, the natural history is not well described and antibiotic treatment decisions are not standardized [33, 34]. Nonetheless, a potential “window of opportunity” during seasons of higher risk may present an opportunity for more intensive monitoring or targeted interventions to minimize infection risk.

There are several limitations to the present investigation. First, the exact date of pathogen acquisition was unknown due to the non-acute, subclinical nature of these infections. In this study, culturing was generally performed at quarterly intervals that enabled a “natural” seasonal analysis of incidence patterns. Second, differentiation between subtypes of bacteria species could not be performed in this study. For example, the Registry does not contain information regarding SCCmec types or Panton Valentine leukocidin genes with which to differentiate community and healthcare-associated MRSA. Given trends in the general population, it is possible that seasonal acquisition differs by subtype. Similarly, H. influenzae subtypes data were not collected so seasonal patterns of these subtypes could not be evaluated. Third, our study was limited to children diagnosed with CF prior to two years of age with a maximum of six years follow-up; therefore the generalizability of results is limited to very young children. Fourth, data on viral infections was unavailable in the Registry and may play an important role in viral induced secondary bacterial infections resulting from respiratory tract damage. Finally, the majority of pathogens were identified through culturing of isolates obtained from oropharyngeal swabs. Rosenfeld and colleagues [35] previously reported a moderate sensitivity and specificity for identifying P. aeruginosa and MSSA by oropharyngeal swab compared to bronchoalveolar lavage in children with CF; however, the accuracy of oropharyngeal cultures for the presence of the pathogens evaluated herein in the lower respiratory tract is unknown. Therefore, results from this study more accurately reflect upper airway than lower airway pathogen acquisition.

In conclusion, seasonal variations of bacterial respiratory pathogen acquisition rates were observed in a large cohort of young children with CF residing in the U.S. Future studies of factors, such as macroenvironmental variables, that may be contributing to these observed incidence patterns are needed. Similarly, the impact of global warming on both local weather patterns and extreme weather events may impact seasonal acquisition of respiratory pathogens. These findings may then contribute to our understanding of the aetiology of these infections and may provide guidance for future infection control strategies.