A chromobacter xylosoxidans airway infection is associated with lung disease severity in children with cystic fibrosis.

Background: Despite the increasing prevalence of Achromobacter xylosoxidans lung infection in patients with cystic fibrosis (CF), its clinical pathogenicity remains controversial. The objective of this study was to evaluate the effects of this emerging bacterium on lung disease severity in CF children.
Methods: This case-control retrospective study took place in two French paediatric CF centres. 45 cases infected by A. xylosoxidans were matched for age, sex, CFTR genotypes and pancreatic status to 45 never-infected controls. Clinical data were retrieved from clinical records over the 2 years before and after A. xylosoxidans initial infection.
Results: At infection onset, lung function was lower in cases compared with controls (p=0.006). Over the 2 years prior to A. xylosoxidans acquisition, compared with controls, cases had more frequent pulmonary exacerbations (p=0.02), hospitalisations (p=0.05), and intravenous (p=0.03) and oral (p=0.001) antibiotic courses. In the 2 years following A. xylosoxidans infection, cases remained more severe with more frequent pulmonary exacerbations (p=0.0001), hospitalisations (p=0.0001), and intravenous (p=0.0001) and oral antibiotic courses (p=0.0001). Lung function decline tended to be faster in cases (-5.5% per year) compared with controls (-0.5% per year). Conclusions: This case-control study demonstrates that A. xylosoxidans occurs more frequently in the patients with the worse lung disease. Further studies assessing the pathogenicity of this emerging pathogen and international treatment recommendations are warranted.

Introduction

Cystic fibrosis (CF) is a severe autosomal recessive genetic disease caused by mutations in the CF transmembrane conductance regulator (CFTR) gene [1]. Lung disease remains the major cause of morbidity and mortality in CF, with progressive decline of lung function due to excessive airway inflammation associated with recurrent bacterial infections [2, 3]. While Staphylococcus aureus and Haemophilus influenzae are the most prevalent bacteria in the airways of young patients with CF, Pseudomonas aeruginosa predominates in later decades [4, 5]. Other opportunistic pathogens, including Achromobacter xylosoxidans, Burkholderia cepacia and Stenotrophomonas maltophilia, are increasingly detected in patients with CF [4-6].

A. xylosoxidans is a strict aerobic Gram-negative bacillus with broad natural resistance and frequent acquired resistance to antibiotics [7]. Although prevalence of A. xylosoxidans in the CF airways varies worldwide, it has increased over the last few decades. In 2017, the prevalence was recognised to reach 5.8% in the USA and 6.7% in France [4, 5]. The pathogenicity of A. xylosoxidans lung infection remains controversial [8-14] and there is no international recommendation concerning its management. Indeed, it is actually unclear whether antimicrobial treatments directed against this emerging pathogen alter the severity of CF lung disease [8-11]. Therefore, we conducted a retrospective study to evaluate the effects of A. xylosoxidans airway infection on lung disease severity in CF children.

Methods

Patients

This case–control retrospective study took place in two French paediatric CF centres, where 480 patients were registered: Hôpital Trousseau and Hôpital Necker Enfants Malades (Paris, France). The cases were defined as CF patients with at least one positive sputum culture with A. xylosoxidans during their clinical follow-up. A. xylosoxidans was identified in sputum cultures by MALDI-TOF (matrix-assisted laser desorption/ionisation time-of-flight) mass spectrometry analysis.

These cases were matched for age, sex, CFTR genotypes and pancreatic status with CF controls, for whom A. xylosoxidans had never been identified. Clinical data were retrieved from electronic patient records, supplemented when necessary with data from paper patient records. The database and data collection were approved by the French national data protection authorities (CNIL 908324 and CCTIRS 08.015bis) and each patient and/or their legal guardians were informed prior to entering their data into the database.

For each case, the date of initial infection by A. xylosoxidans defined time T0. Clinical data were subsequently collected over 2 years before and after T0, defining time T−24, T−12, T+12 and T+24. Lung function was evaluated by measurements of forced expiratory volume in 1 s (FEV1), expressed as percent predicted values using the Global Lung Function Initiative equations [15]. Rates of pulmonary exacerbations, of hospitalisations for pulmonary exacerbations, and of intravenous and oral antibiotic courses were obtained, as well as changes in airway bacterial colonisation.

Statistical analysis

Data were expressed as mean±sd for continuous variables and number (percentage) for categorical variables. The t-test was used to compare quantitative data and Fisher's exact test was used for categorical data comparisons. The differences were considered significant for p-values <0.05.

Results

Case analysis: CF children infected by A. xylosoxidans

Among the 480 patients followed in the two paediatric CF centres, 45 CF children (28 girls and 17 boys) had been infected by A. xylosoxidans at a median age of 11.5±4.9 years. 18 (40%) patients were homozygous for the CFTR F508del mutation and 42 (93%) were exocrine pancreatic insufficient. None of these patients were under a CFTR modulator at the time of data collection.

Over the 2 years before and after A. xylosoxidans initial infection (from T−24 to T+24), annual lung function decline reached 2.4% per year, with FEV1 % pred decreasing from 88±18% to 78.4±16% (p=0.04). In comparison with the year before A. xylosoxidans infection, during the year after infection, the following rates increased: annual rate of respiratory exacerbations (+1.8 per year; p=0.001), hospitalisations (+0.3 per year; p=0.006) and oral antibiotic courses (+1.7 per year; p=0.001). Although following the same trend, the increase in intravenous antibiotic courses was not significant (+0.4 per year; p=0.08).

Case–control comparison analyses

Each of the 45 cases was matched (age, sex, CFTR genotypes and pancreatic status) to 45 controls. Similar to the cases, none of these controls were under a CFTR modulator at the time of data collection. Cases and controls were compared in the 2 years before (table 1) and 2 years after (table 2) A. xylosoxidans acquisition. At infection onset, lung function was lower in cases compared with controls (FEV1 % pred 81.3±18% versus 94.2±16%, respectively; p=0.006) (figure 1).

TABLE 1

Case–control comparison over the 2 years prior to Achromobacter xylosoxidans initial infection

	Cases	Controls	p-value#
Subjects	45	45
FEV1 % pred	88.0±18	92.7±17	0.3
BMI z-score	0.03±0.9	−0.03±0.9	0.8
Annual rate of respiratory exacerbations	2.8±1.7	2.0±1.3	0.02*
Annual rate of hospitalisations	0.1±0.4	0	0.05*
Annual rate of oral antibiotic courses	3.7±1.8	2.4±1.6	0.001*
Annual rate of intravenous antibiotic courses	0.4±0.9	0.12±0.4	0.03*
Pseudomonas aeruginosa colonisation	64 (29)	24 (11)	0.0002*
MRSA colonisation	9 (4)	9 (4)	0.9

Data are presented as mean±sd or % (n) over the 2 years prior to A. xylosoxidans initial infection for the cases (between T−24 and T0). FEV1: forced expiratory volume in 1 s; BMI: body mass index; MRSA: methicillin-resistant Staphylococcus aureus. #: for the comparison of sex, age, CFTR genotypes and pancreatic status matched cases and controls. *: p<0.05.

TABLE 2

Case–control comparison over the 2 years following Achromobacter xylosoxidans initial infection

	Cases	Controls	p-value#
Subjects	45	45
FEV1 % pred	78.4±16	93.2±17	0.003*
BMI z-score	−0.1±1	−0.01±0.8	0.7
Annual rate of respiratory exacerbations	4.7±2.3	1.9±1.1	0.0001*
Annual rate of hospitalisations	0.5±0.7	0	0.0001*
Annual rate of oral antibiotic courses	5.3±2.4	2.2±1.4	0.0001*
Annual rate of intravenous antibiotic courses	1.3±1.6	0.2±0.5	0.0001*
Pseudomonas aeruginosa colonisation	51 (23)	29 (13)	0.04*
MRSA colonisation	16 (7)	2 (1)	0.05*

Data are presented as mean±sd or % (n) over the 2 years following A. xylosoxidans initial infection for the cases (between T0 and T+24). FEV1: forced expiratory volume in 1 s; BMI: body mass index; MRSA: methicillin-resistant Staphylococcus aureus. #: for the comparison of sex, age, CFTR genotypes and pancreatic status matched cases and controls. *: p<0.05.

FIGURE 1

Trajectories of forced expiratory volume in 1 s (FEV1) for cases and matched controls over 2 years before (T−24) and after (T+24) initial infection by Achromobacter xylosoxidans (T0).

Over the 2 years prior to A. xylosoxidans acquisition, the cases had more frequent pulmonary exacerbations (p=0.02), hospitalisations (p=0.05), and oral (p=0.001) and intravenous (p=0.03) antibiotic courses, and were more frequently colonised by P. aeruginosa (p=0.0002) (table 1). No difference was observed for colonisation with methicillin-resistant S. aureus (MRSA) before A. xylosoxidans acquisition.

In the 2 years following A. xylosoxidans initial infection, FEV1 % pred values were systematically lower in cases compared with controls (78.4±16% versus 93.2±17%, respectively; p=0.003 after 2 years) (table 2). The FEV1 % pred rate of decline between T−12 and T+12 was faster in cases, although not significantly (−5.5% per year in cases versus −0.5% per year in controls; p=0.14) (figure 1). Cases remained more severe, with more frequent pulmonary exacerbations (p=0.0001), hospitalisations (p=0.0001), and intravenous (p=0.0001) and oral (p=0.0001) antibiotic courses (table 2). Colonisation with P. aeruginosa was more frequent in cases (incidence 51% in cases versus 29% in controls; p=0.04), as was colonisation with MRSA (incidence 16% in cases versus 2% in controls; p=0.05) (table 2).

Discussion

This case–control study demonstrates that A. xylosoxidans occurs more frequently in the patients with the worse lung disease. Indeed, before the initial infection and in comparison with CF controls matched on age, sex and CFTR mutations, these children had a lower lung function, experienced more frequent respiratory exacerbations, and required more frequent hospitalisations and antibiotic courses (oral and intravenous).

We observed a higher prevalence of airway infections by A. xylosoxidans than that reported in the registries worldwide [4, 5]. We also found a prevalence of 10.6% in children with CF, whereas the French CF Registry reported a prevalence of 6.7% in the overall cohort in 2018 [5]. Although the prevalence shown in this study is higher than those reported in the French and US CF registries [4, 5], it is comparable to that found in recent studies realised in France [11], Italy [9] and Spain [10], suggesting a recent increasing prevalence worldwide, or at least in Europe. This increase may also be secondary to the improvement of bacteria detection in CF sputum. In 1998, Burns et al. [16] used standardised techniques for identification and susceptibility testing of CF specimens, and observed a prevalence of 8.7% of A. xylosoxidans lung infection in 595 American CF patients, whereas in the same year the US CF Registry reported a prevalence of only 0.5%. In our study, A. xylosoxidans was identified in sputum cultures by MALDI-TOF mass spectrometry analysis, a method applied routinely in microbiology laboratories for a decade. MALDI-TOF mass spectrometry is recognised to allow identification of rare pathogens in CF and might be involved in the increasing identification of A. xylosoxidans [17].

We observed that the children with CF infected by A. xylosoxidans had a more severe lung disease with worse lung function, and more frequent respiratory exacerbations, hospitalisations and antibiotic courses before and after A. xylosoxidans acquisition. These results are in accordance with several other studies, such as De Baets et al. [8] in 2007, Recio et al. [10] in 2018 and Tetart et al. [11] in 2019. A Brazilian retrospective case–control study also showed that infected cases were more frequently hospitalised in the 2 years following primary infection with A. xylosoxidans compared with non-infected controls [13]. The hypothesis could be that A. xylosoxidans acquisition is more frequent in CF patients who already have severe lung disease, the severity of which further increases after infection with this deleterious bacterium. This hypothesis is also supported by the more frequent P. aeruginosa airway colonisation in cases; P. aeruginosa is well known to be associated with accelerated lung function decline in CF patients [18]. Tetart et al. [11] underlined that co-isolation of P. aeruginosa with A. xylosoxidans is associated with a significantly faster annual decrease in FEV1 compared with patients colonised with A. xylosoxidans only. However, it is difficult to differentiate the specific influence of either bacterium, i.e. A. xylosoxidans and P. aeruginosa, as also highlighted by Hansen et al. [19]. Moreover, Hansen et al. [19] showed that chronic pulmonary inflammation, measured by cytokine production, was comparable in patients infected with A. xylosoxidans and patients infected with P. aeruginosa, underlying the deleterious pathogenicity of this pathogen.

These observations show the importance of A. xylosoxidans for the patient's prognosis. A 5-year modelling study has shown that each respiratory exacerbation had a detrimental effect on lung function, equivalent to a loss of 12% of FEV1 [20]. As such, the increasing frequency of exacerbations observed in our study after A. xylosoxidans acquisition is of great importance. In the same way, in a Canadian cohort of 1103 CF patients followed during 18 years it was observed that patients chronically infected with Achromobacter spp. had higher risk of death or transplantation than uninfected patients [14]. In our study, FEV1 % pred of cases also continued to be lower in the 2 years following the primary infection, compared with that of controls. Nevertheless, the real causal link between the decline in respiratory function and the presence of A. xylosoxidans in the airways is difficult to establish. Indeed, in our study, with the cases being already more severe before the primary infection, their lung function did not seem to decrease significantly after, compared with that of controls. Case–control studies have shown that A. xylosoxidans infection was associated with a significantly higher FEV1 annual rate of decline in the 3 years following primary infection [10, 11]. Another Danish study found the same tendency of worsened respiratory function with a faster decline in FEV1 for patients infected with A. xylosoxidans, with high levels of specific anti-A. xylosoxidans antibodies [21].

Efficacy of antibiotic treatments for this multidrug-resistant organism is still unclear and there are no treatment recommendations so far, in particular on the need to systematically prescribe antibiotic therapy at the time of discovery. According to several studies, the most active antibiotics against A. xylosoxidans would be piperacillin-tazobactam, meropenem and trimethoprim-sulfamethoxazole [22-25]. It was also shown that early treatment with inhaled antibiotics such as ceftazidime, colistin or tobramycin could prevent or at least postpone chronic A. xylosoxidans airway colonisation in patients with CF [24]. However, some studies have underlined a high rate of acquired resistance to many of these antibiotics [7, 26]. Thus, management recommendations for this multidrug-resistant organism appear crucial.

Our study has several limitations. First, it is a retrospective study and, despite the rigorous analysis of the medical files, some data were missing that could have resulted in recruitment bias. To limit this bias, we chose to conduct an observational study as well as a case–control analysis to increase the power of the analyses. Moreover, our study only involved two paediatric CF centres, which led to a selection bias with the impossibility of having a very large sample size. There is therefore a risk of loss of statistical power. Nevertheless, we were able to observe significant results.

To conclude, we have shown that A. xylosoxidans lung infection is associated with increased lung disease severity in children with CF. While this pathogen was considered as an infrequent bacterium infecting the airways of CF patients until recently, its incidence appears to be increasing. Larger prospective studies assessing the pathogenicity of this emerging pathogen as well as international treatment recommendations are urgently warranted.