Prevention of chronic infection with Pseudomonas aeruginosa infection in cystic fibrosis

Abstract

Purpose of review

This review provides an update on definitions of chronicity of infection, approaches to airway sampling to detect infection, strategies for Pseudomonas aeruginosa eradication, impact of cystic fibrosis transmembrane regulator protein (CFTR) modulators and future challenges for clinical trials.

Recent findings

Rates of P. aeruginosa have decreased over the past two decades with establishment of effective eradication protocols. Definitions of chronic P. aeruginosa infection have required adaptation for healthier populations. Although molecular (PCR) approaches to early P. aeruginosa detection are sensitive, to date, earlier diagnosis has not impacted on clinical outcomes. Despite eradication regimens, some people with early P. aeruginosa fail to clear their infection. Most people also experience a recurrence and eventual transition to chronic infection. Several recent studies sought to address this gap. CFTR modulators (predominantly ivacaftor) demonstrated reduced P. aeruginosa density, although infection may persist or recur demonstrating the need for continued antiinfective therapies in the modulator era.

Summary

Future studies of approaches to P. aeruginosa eradication will be complex due to expanded availability and ongoing competitive clinical trials of CFTR modulators. Studies to address optimal eradication therapy, particularly in adults, will be required, though adequate recruitment to power these studies may prove challenging.

INTRODUCTION

Over the past three decades, there have been dramatic improvements in outcomes for people with cystic fibrosis (CF), including improved lung function, nutritional status and survival (Fig. 1) [1]. The delivery of multidisciplinary clinical care in parallel with the implementation of newborn screening (NBS), effective pancreatic enzymes and nutritional replacement, and aggressive treatment of airway pathogens are all thought to have contributed to these improved outcomes [2].

FIGURE 1.

Median FEV₁% predicted for people with cystic fibrosis in the United States by age (1998–2018). Source of data: Cystic fibrosis patients under care at CF Foundation accredited care centre in the United States, who consented to have their data entered. Reproduced with permission from [7^■].

In the period, 2001–2012, when NBS was implemented progressively in the USA, incident cases of Pseudomonas aeruginosa decreased significantly, suggesting that earlier diagnosis and use of effective therapies may impact on the development of airway infection [3^■,4]. Yet, P. aeruginosa infection occurs in a significant proportion of young children by age 5 years. In the Australian CF Bronchoalveolar Lavage study, 53% of children who were diagnosed by NBS had at least one positive P. aeruginosa sample, though chronic infection was rare [5,6,7^■].

Patient registries have become the backbone of monitoring CF populations and have profiled the changing microbiology in CF [1,8-11]. The prevalence of P. aeruginosa infection increases with age; however, the prevalence in older children at the time of transition to adult care has decreased dramatically [12,13^■■] (Fig. 2). In parallel, prevalence of other bacterial pathogens has changed with reduced prevalence of Burkholderia cepacia complex, and increased Stenotrophomonas maltophilia, Achromobacter species, methicillin resistant Staphylococcus aureus infection (particularly in the USA) and nontuberculous mycobacteria (NTM) infection [1,8-11,14]. The links between deceased P. aeruginosa and changing prevalence of other bacterial pathogens are unclear. The observation that prevalence of P. aeruginosa varies considerably between CF centres supports the role of between centre benchmarking to assess consistency of antibiotic therapy approaches [8]. Genetically indistinguishably strains of P. aeruginosa have been widely reported in CF centres and thought to be because of patient-to-patient transmission [6,15]. Importantly, such strains frequently have greater antimicrobial resistance (AMR). Prevalence of such strains varies greatly between centres and between regions/countries [15], although are uncommon in younger children with CF [6].

FIGURE 2.

P. aeruginosa infection status based on Leeds criteria of people with cystic fibrosis in the United States by age (2017). Source of data: Cystic fibrosis patients under care at CF Foundation accredited care centre in the United States, who consented to have their data entered. Reproduced with permission from [7^■].

From 2008 to 2017, chronic P. aeruginosa prevalence was reduced from 61.3 to 46.3% in adults (≥18 years) and from 15.2 to 7.1% in children (<18 years) in the UK [8]. Utilizing the US CF Foundation Patient Registry (CFFPR), 15 504 patients with CF (≥13 years, median follow-up – 5 years), the annual incident rates of chronic P. aeruginosa infection fell from 14.3% in 2003 to 6.4% in 2012 [13^■■]. Some patients do not develop P. aeruginosa even in adult life. In a recent study, adults with CF who were never infected with P. aeruginosa had milder lung disease and were less likely to have two F508del-CFTR mutations than contemporaries who had previously had P. aeruginosa [16^■]. These findings may have implications for the early commencement of effective CFTR modulators in children.

This review provides an update of the evidence for the definitions of chronicity of infection, approaches to airway sampling to confirm presence of infection, strategies for P. aeruginosa eradication, predictors of treatment failure, the impact of CFTR modulators and the future challenges for clinical trials.

DEFINING P. AERUGINOSA INFECTION STATUS

Airway infection with P. aeruginosa progresses from initial acquisition to chronic airway infection over a period of years [7^■]. Intermittent detection of P. aeruginosa occurs during the transition to chronic infection. Defining stages of infection is important for treatment decisions, prognostication and as an outcome measure for clinical trials. The Leeds criteria is the most widely used definition of infection status [17] and defines stages of infection as chronic (>50% of cultures positive in 12 months), intermittent (≤50% of cultures positive); free (culture negative for 12 months) and never infected. Using this definition, children who were classified as chronically infected had lower lung function and worse illness and radiology scores compared with those in other groups [17].

The Leeds criteria however has several limitations. Developed initially in the paediatric population, it is not as reliable in adult populations [12,18]. It was also developed in the setting of routine monthly surveillance cultures, which is not the practice in all CF care centres [12,19]. Also, Leeds criteria are based only on microbiologic culture detection of P. aeruginosa, and does not consider other factors, including molecular approaches, clinical status, serology or bacterial phenotypes such as mucoidy. Pragmatic clinician-defined criteria for adults with P. aeruginosa were developed for clinical and clinical trial use and were compared with Leeds criteria in a recent publication [20^■]. Notably, clinicians were more likely to classify patients as chronically infected compared with Leeds criteria, potentially due to incorporation of additional clinical data particularly in nonexpectorating patients.

European consensus criteria (ECC) has also been used to assess P. aeruginosa status defining chronic infection as at least three positive cultures for P. aeruginosa over 6 months with evidence of infection or tissue damage (P. aeruginosa without evidence of infection is considered chronic colonization) [21]. Jonckheere et al. [22^■] compared ECC with genotyping of P. aeruginosa isolates to determine whether persistence of identical genotypes predicted chronic colonization. Genotyping results were consistent with ECC diagnosis of chronic colonization for 93% of patients, and predicted chronicity on average 9.3 months earlier, suggesting that genotyping may provide additional information in defining chronic infection.

Initial P. aeruginosa strains are predominantly nonmucoid and have an antibiotic-susceptible phenotype [23,24] with most children acquiring environmental stains [6]. To examine the relationship between chronic P. aeruginosa infection and the mucoid phenotype, Heltshe et al [7^■] examined data from more than 5000 children in the CFFPR (2006–2015) less than 2 years of age at diagnosis; children were followed to a median age of 5.5 years. Of those with a positive P. aeruginosa culture, 13% developed chronic infection and 17% cultured mucoid P. aeruginosa. Mucoidy preceded chronic infection in the majority and was associated with a higher risk of transition to chronic infection.

Boutin et al. [25^■] examined the utility of quantitative PCR (qPCR) detection of P. aeruginosa in distinguishing intermittent from chronic infection. Using a cross-sectional study design with one-time sample collection, higher abundance of P. aeruginosa in sputum, and to a lesser but still significant extent in oropharyngeal swabs, was associated with chronic infection. Abundance of P. aeruginosa was better at distinguishing chronic versus intermittent infections than mucoidy status, suggesting that qPCR could be an additional diagnostic tool to define chronic infection.

DIAGNOSTIC AND SAMPLING APPROACHES FOR P. AERUGINOSA

Sputum has been the routine clinical specimen to evaluate the airway microbiology in patients with CF. Improvements in pulmonary function render many patients unable to provide expectorated sputum samples [12]. This can be a particular challenge following completion of eradication therapy for a child with CF. Alternative approaches include collection of upper airway samples (oropharyngeal or cough swabs), which are relatively noninvasive and have a high negative predictive value (NPV) (90–96%) in diagnosing lower airway P. aeruginosa infection but suffer from low positive predictive values (PPVs) (10–79%) compared with bronchoalveolar lavage (BAL) cultures [26,27^■,28^■,29,30^■■] (Table 1). BAL samples allow earlier identification of lower airway pathogens [27^■,31,32^■], but are invasive and do not alter key outcomes (positive P. aeruginosa culture or HRCT findings) at 5 years of age when used to guide antibiotic therapy in preschool children with CF diagnosed by NBS [5]. Thus, a less invasive and yet accurate method to detect early infection with P. aeruginosa is urgently needed.

Table 1.

Sampling approaches for P. aeruginosa detection

	Ease	Invasiveness	Utility
Expectorated sputum	+++	−	Gold standard (when available)
Oropharyngeal swab	++	−	Compared with BALF: Sensitivity 44% [26], 23% [27■], 86% [28■] Specificity 95% [26], 91% [27■], 100% [28■] PPV 44% [26], 18% [27■], 100% [28■] NPV 95% [26], 94% [27■], 93% [28■] Compared to induced sputum [29]: Sensitivity 60% Specificity 97% PPV 50% NPV 95%
Nasal swab	++	−	Poor and variable sensitivity/specificity/PPV/NPV; not recommended [28■]
Induced sputum	+	+	Sensitivity 100% [28■] Specificity 100% [28■] PPV 100% [28■ NPV 100% [28■] May have improved sensitivity over BALF [30■■]
BALF	−	+++	Gold standard [28■]

Two approaches that have received recent focus are the use of P. aeruginosa serology and induced sputum. Several recent studies have examined the role of serology. In the study by Boutin et al. [25^■], detection of P. aeruginosa in throat swabs by PCR was associated with P. aeruginosa in sputum in 75% of cases. Performing P. aeruginosa qPCR on culturenegative throat swabs may further improve the diagnosis of lower airway infection in young children with CF, yet more validation is required. In a study from Toronto, children with CF with new P. aeruginosa infection had evidence of immune response to P. aeruginosa up to a year in advance of its culture [33^■]. Yet, the identification of qPCR positive results did not predict success of eradication therapy.

Up to 20% of patients transitioning to adult care do not produce spontaneously expectorated sputum leading to the suggestion of routine sputum induction for nonproductive children and adults [12,29]. Notably, two studies have considered this issue recently. In an analysis comparing nasal swabs and induced sputum with BAL samples in children with CF, the microbiological yield, specificity, sensitivity, PPV and NPV of cough swab and induced sputum were very similar in the nonexpectorating participants [28^■]. In the CF-SpIT study, induced sputum and cough swabs were compared with BAL in 124 children with CF undergoing 200 sputum induction (aged 6 months to 18 years) [30^■■]. This study showed that both sputum induction and six-lobe BAL provided independent information about lower airway pathogens when compared with the standard two-lobe BAL. The authors argued that induced sputum was a suitable option for symptomatic children before BAL, as it will identify the correct lower airway pathogen in most patients.

ERADICATION OF EARLY P. AERUGINOSA INFECTION IN CHILDREN

Treatment with antipseudomonal antibiotics (tobramycin or colistin) to eradicate initial or early P. aeruginosa infection is established standard of care in CF, with a single course of inhaled antibiotics most commonly recommended [34-38]. P. aeruginosa detection early in life generally suggests worse outcomes [39-41]. However, Petrocheilou et al [42] compared chest computed tomography (CT), lung function and BMI at age 6–7 years in children with and without a positive P. aeruginosa culture in the first year of life. They found no difference in this intensively treated population suggesting that early eradication approaches may attenuate the impact of P. aeruginosa, reinforcing the importance of early detection.

Although inhaled tobramycin solution (TIS) is effective in eradicating early P. aeruginosa, nearly 10–25% of people with early infection fail to eradicate, and recurrence rates are high with nearly 30% recurring within 18–27 months [36-38]. Blanchard et al. [43] evaluated the effectiveness of a stepwise approach to P. aeruginosa eradication including a second course of TIS followed by 14 days of intravenous antibiotics as well as TIS for children who remain culture positive. For children who were asymptomatic, eradication rates improved from 77% after one course of TIS to 89% after the three steps. Success rates were somewhat lower in children who were symptomatic at the time of new P. aeruginosa isolation (67%) [43].

Several studies have sought to address the high failure rate of initial eradication attempts. Ratjen et al. [44^■■] performed a double-blind, placebo-controlled study of children with CF less than 7 years old with early p. aeruginosa. Patients were randomized to TIS or placebo for 28 days with an optional cross-over at 35 days. Eradication rates at 29 days were significantly higher in the TIS group than in placebo (85% compared with 24%, P < 0.001), consistent with recommendations for eradication. For the cross-over, patients who were remained positive received open-label TIS, whereas those who were P. aeruginosa negative received either placebo or blinded TIS. Importantly, those who received placebo followed by TIS (either blinded or open-label), had lower rates of eradication than those who received early TIS (48% compared with 76%) suggesting that a delay, even of 35 days, may reduce the efficacy of eradication. As routine surveillance cultures are often collected on a quarterly basis, this finding may explain some of the eradication failures in clinical practice. A separate observational study found that individuals with fewer sputum samples surveyed and with multiresistant P. aeruginosa were more likely to fail to eradicate, also supporting that early detection may improve eradication rates [45].

The OPTIMIZE clinical trial sought to determine whether the addition of chronic azithromycin three times weekly to TIS would prolong the time to P. aeruginosa recurrence and decrease the risk of pulmonary exacerbations in children with early infection [46^■■]. The study was a multicentre, double-blind, randomized, placebo-controlled, 18-month study in children with CF (6 months to 18 years, n = 221) with early P. aeruginosa infection. Azithromycin or placebo was given in addition to standard TIS treatment. Enrolment was stopped early due to a reduction in pulmonary exacerbations of 44% in those who received chronic azithromycin compared with placebo. Eradication rates, recurrence and development of persistent P. aeruginosa infection (seen in 20–25% of participants) did not differ between treatment groups. Thus, the addition of chronic azithromycin at the time of early P. aeruginosa isolation appears to reduce pulmonary exacerbations, consistent with other studies [47,48], but does not increase P. aeruginosa eradication or reduce recurrence.

ERADICATION OF P. AERUGINOSA IN ADULTS WITH CYSTIC FIBROSIS

With increasing numbers of people transitioning to adult care without P. aeruginosa infection, evidence to support treatment decisions is now an important consideration for adult healthcare teams. Data to support approaches to eradication of P. aeruginosa in adults with CF are limited. Studies examining eradication approaches included small numbers of adults [36,37]. In the study published by Ratjen et al. [36], eight of 123 participants (6.5%) were 18 years or older and in the study led by Taccetti et al. [37], 27 of 223 participants (12.1%) were adults. The small numbers of adults included in these pivotal studies prevented subgroup analysis, to determine the success of eradication approaches in adults with CF. An uncontrolled single-centre study described a success rate in the adult population (79%) [49] similar to that seen in paediatric populations (63–93%) [36,37,50]. Importantly, this cohort studied included a significant number of patients with residual function CFTR mutations (in particular, the R117H mutation), which is associated with milder pulmonary phenotype and therefore therapies to attempt eradication of early p. aeruginosa infection may be more likely to be successful.

Consequently, it is not currently possible to assess the optimal regimen for eradication of P. aeruginosa in adults with CF. In the unpublished TORPEDO-CF study, intravenous antipseudomonal antibiotics were used for eradication of P. aeruginosa. The study, undertaken in the UK, randomized 286 patients, with investigators estimating that 25% would be adults [51]. Yet, only 15 (5.2%) of those enrolled were adults [51], underlining the challenges of recruiting this population. In addition to challenges recruiting to a clinical trial, there may be other factors dissuading patients from intensive therapy. In the adult with CF, conflicting commitments including employment and postschool studies may impact on their ability to participate in prolonged antipseudomonal therapies, particularly regimens including parenteral antibiotics [52]. The optimal regimen in adults is yet to be determined.

P. AERUGINOSA AND THE AIRWAY MICROBIOME

In patients with CF, the CF airway microbiome is perturbed and becomes progressively less diverse with increasing age and lung disease severity [53]. Importantly, the intermittent presence of P. aeruginosa does not appear to impact microbiome diversity, whereas chronic P. aeruginosa does, suggesting that early eradication therapy of P. aeruginosa may limit CF airway dysbiosis [54]. The prevalence and relative quantity of anaerobic bacteria cultured from sputum samples are also associated with milder lung disease and lower rates of antibiotic use [55^■]. The impact of altered microbiome in CF requires further study, although narrow biodiversity associated with P. aeruginosa predominance is an independent predictor of poorer prognosis [56^■].

CFTR MODULATORS AND P. AERUGINOSA

The discovery of CFTR modulators that directly impact the dysfunctional chloride channel has the potential to change the CF landscape [57,58]. Ivacaftor has now been approved in children with gating CFTR mutations who are 6 months or older, the dual combination of ivacaftor/lumacaftor approved for children who are homozygous for the F508del CFTR mutation who are 2 years and older, and tezacaftor/ivacaftor approved for children homozygous for the F508del CFTR mutation or with one residual function mutation who are 6 years and older [59-61]. The earlier use of CFTR modulators in children prior to chronic airway infection has led to interest in the potential for impact of CFTR modulators on the natural history of airway microbiology and in particular, P. aeruginosa. Effective CFTR modulators have the potential to enhance natural clearance, to delay first infection in very young children and to improve the effectiveness of eradication and reduce prevalence of chronic P. aeruginosa. Observational studies of children with CF identify CF genotype as a primary risk factor for P. aeruginosa acquisition [62]. Similarly, adults with CF never infected with P. aeruginosa are more likely to have genotypes with residual CFTR function [16^■]. Thus, improving CFTR function with modulators is likely to be critical in preventing P. aeruginosa infection.

The GOAL (G551D Observational) study demonstrated the real-world impact of ivacaftor on clinical outcomes including airway microbiology. Intriguingly, following the commencement of ivacaftor, there was a significant reduction in rates of P. aeruginosa (~20% reduction) when compared with 12 months prior to initiation of ivacaftor [63]. Of 151 participants, 26 out of 89 (29%) who were P. aeruginosa culture positive the year prior to ivacaftor were culture negative in the year following treatment. The odds of culturing P. aeruginosa were reduced by 35% (odds ratio 0.65; P < 0.001) [64].

In 12 patients with chronic P. aeruginosa, there was a very rapid decrease in the sputum P. aeruginosa density (CFU/ml) by day 2 of ivacaftor therapy, which persisted for the first year of treatment. Importantly, none of these patients became P. aeruginosa infective negative. On average, after 210 days of ivacaftor, P. aeruginosa CFU had increased to baseline. In a subgroup of patients in the GOAL study, induced sputum collected from 14 patients before and after ivacaftor therapy had reduced P. aeruginosa density and increased Staphylococcus aureus density and Prevotella species relative abundance [65].

Observational studies of the impact of combination CFTR modulators (including ivacaftor/lumacaftor and tezacaftor/lumacaftor) on the CF airway microbiology are underway, In a recent single-centre study from the UK, children receiving either ivacaftor or ivacaftor/lumacaftor (>80%) demonstrated delay in the acquisition of P. aeruginosa and S. aureus [66^■■].

ANTIMICROBIAL RESISTANCE AND P. AERUGINOSA

With improvements in management and life expectancy in CF, and growing concerns worldwide about the development of AMR, strategies to prevent and treat multidrug-resistant P. aeruginosa are needed. Recently, an international working group was formed to begin to address AMR in people with CF [67^■-70^■]. The detection of multidrug-resistant P. aeruginosa at initial isolation is associated with risk of eradication failure [45]. Rates of multidrug-resistant P. aeruginosa also increase with increasing age and disease severity [71]. A challenge with AMR in CF is the disconnect between antimicrobial susceptibility testing in the laboratory and clinical response to antimicrobial therapies [69^■,70^■]. Developing better laboratory approaches to predict treatment response is critical.

NEW ANTIBIOTIC COMPOUNDS OF TREATING P. AERUGINOSA

The recognition of growing rates of AMR and the limited investment of antibiotic development programmes over the past two decades have become global priorities [72-74]. Several approaches have shown promise including compounding of combinations of antibiotics (tobramycin and fosomycin) [75], and repurposing antibiotics in novel preparations with a focus on effective antibiotics for people with CF, in particular for inhalation to support local delivery of high concentrations of the antibiotic(s) [76,77] (Table 2). Several examples include nebulised levofloxacin [78,79], nebulised liposomal amikacin [80], nebulised liposomal and free ciprofloxacin compounds [81^■] and dry powder colistin [82]. To date, studies of efficacy have been limited and no studies have evaluated their role in eradication regimens.

Table 2.

Antibiotic compound development for P. aeruginosa

Drug	Preparation	Stage of development	FDA approval
Antibiotic
Levofloxacin	Solution	Phase II/III	Yes
Ciprofloxacin	Solution (free/liposomal)	Phase II/III	Not for CF
Amikacin	Liposomal solution	Phase II/III	Not for CF (NTM)
Colistin	Dry power	Phase II/II	No
Oligo-G	Inhaled	Phase II	No
Bacteriophage	Inhaled	Phase I/II	No
Gallium	Inhaled	Phase II	No

CHALLENGES IN UNDERTAKING ANTIBIOTIC TRIALS IN THE TREATMENT OF P. AERUGINOSA (INCLUDING ERADICATION)

As discussed above, there are number of published highly effective regimens for P. aeruginosa infection eradication in children with CF; therefore, the ability to demonstrate efficacy of new regimens will be challenging due to the very large numbers of participants required to achieve trial endpoints. One such study (TORPEDO-CF) included a parenteral combination (IV ceftazadime and tobramycin) with nebulised colistin compared with oral ciprofloxacin and nebulised colistin (standard therapy) [51], aimed to determine if additional parenteral antibiotics enhanced effectiveness without compromising safety and tolerability. The study demonstrated that IV therapy was not superior in achieving the primary outcome and there were no differences in secondary outcomes and the standard therapy arm was more cost-effective.

In the setting of the highly competitive and rapidly changing CFTR modulator development programmes, there are a number of challenges both in the design of and recruitment to antibiotic trials that have been recently reviewed [83^■]. There is also a potential reluctance by patients and healthcare teams to support recruitment and for sponsors to invest and undertake trials of antibiotics.

CONCLUSION

Despite recent decreases in P. aeruginosa infection in the CF population, detection remains common and highly variable centre to centre. Definitions of chronic P. aeruginosa infection are evolving and require a degree of pragmatism. As the health of CF population continues to improve, spontaneously expectorated sputum is more difficult to obtain for many children and even adults with CF. Sputum induction is promising but will have logistical challenges in delivery in the busy clinical setting when considering infection control issues and local resources. Current eradication protocols are based on nebulised antibiotics (tobramycin or colistin) with or without oral ciprofloxacin. Eradication protocols including intravenous antibiotics are not more effective. Effective CFTR modulator therapies are likely to impact on rates of P. aeruginosa infection and other common CF airway pathogens.

Despite success of eradication approaches, individuals still experience eradication failures, frequent recurrence of P. aeruginosa and progression to chronic infection. Improved management, potentially through strategies of earlier detection, more intensive second-line treatment approaches or novel antimicrobial approaches. Use of antimicrobials however will need to consider the impact on AMR particularly, as people with CF are expected to live longer, healthier lives.

Studies of better treatments will need to focus on the population of patients who fail to clear using standard protocols. Similarly, studies to address the efficacy of eradication protocols in adults with new-onset P. aeruginosa infection are required despite logistical challenges.

KEY POINTS.

• Definitions of P. aeruginosa infection status are evolving and require further validation.

• As the CF population becomes healthier, the proportion of patients who are able to produce spontaneously expectorated sputum decreases. Sputum induction is a suitable approach to assess the lower airway microbiology before bronchoscopy is considered.

• Eradication of P. aeruginosa is an established standard of care with children with CF, whilst the numbers of adults with CF who require such therapy is increasing.

• Eradication attempts are not always successful, and most people experience recurrence and progression to chronic infection over time.

• CFTR modulators decrease short-term P. aeruginosa density and lead to increased airway microbiome diversity, although long-term effects are less clear.

• Future studies need to evaluate those who fail to clear P. aeruginosa despite conventional eradication protocols and study therapeutic approaches to P. aeruginosa eradication in adults with CF.