Introduction

Pulmonary infections in cystic fibrosis (CF) have been recognized as the major cause of morbidity and mortality, leading to the irreversible failure of lung functions [1]. A succession of pathogens is classically described, starting with Haemophilus influenzae and Staphylococcus aureus, progressing to Stenotrophomonas maltophilia, Pseudomonas aeruginosa and Burkholderia spp. [2]. However, CF lungs are ecological niches with complex and atypical microbiota. The virulence of some unusual species in CF has not been clearly demonstrated, and their clinical interest remains unclear [3, 4]. The gram-positive branching and filamentous rods Nocardia spp. belong to these “uncertain” CF pathogens [5].

Nocardia spp. are ubiquitous worldwide-spread actinobacteria that can be isolated from soil, composting vegetation and water. After inhalation or direct inoculation, these bacteria can cause localized or disseminated human diseases (nocardiosis) with tissue destructions, abscesses, systemic disorders and metastatic lesions [5–7]. Among all these infections, 60–80 % are associated with pulmonary nocardiosis [8]. Symptoms of nocardiosis are unspecific and can be fever, thorax pain, increased sputum and night sweats. X-ray shows multiple lung infiltrates that are sometimes associated with necrosis [7]. Common risk factors for nocardiosis include immunosuppression, malignancies and severe lung disease (chronic obstructive pulmonary disease, bronchiectasis, emphysema…) [6, 9–12]. Though CF is associated with complex immunologic and metabolic disorders, lung damage as well as corticosteroid or transplantation-related immunosuppression, Nocardia spp. are rarely isolated in the sputum of CF patients, as suggested by the few number of cases reported in the literature. Indeed, Barrio et al. [5] described the isolation of Nocardia spp. in only 9 out of 387 patients (2.3 %) followed in 6 CF units in Spain. This low incidence of nocardiosis among CF patients may be explained by the fact that current laboratory protocols do not include specific processing of CF specimens for Nocardia isolation. Moreover, CF patients frequently receive several antimicrobial drugs against major pathogens that could prevent the development of Nocardia spp. in lungs either directly or by modifying the pulmonary ecosystem. Lastly, the CF population is mainly constituted of children, whereas nocardiosis is described preferentially in adults [6, 8]. However, the increasing life expectancy of CF patients could imply a potential increase in the incidence of Nocardia infections and supports the need of improving our knowledge of their pathogenicity in this clinical setting.

When Nocardia spp are isolated from CF sputum, their clinical significance and the corresponding patient management remain unclear. Determining the difference between Nocardia infections and colonizations is challenging since the availability of study cohorts is limited due to the low incidence of nocardiosis in CF. Moreover, nocardiosis can be under-diagnosed since the pulmonary symptoms are unspecific. In the literature, both infections and colonizations are described. Lastly, Nocardia spp. are rarely the only microorganism present in CF lungs [6, 7].

The objective of the present report was to describe in 3 French CF centers the clinical and microbiological features of 16 CF patients with sputum culture positive for Nocardia spp. in order to (1) analyze the impact of such bacteria on the lung status and (2) review the clinical practices. Finally, we propose guidelines for the clinical management of Nocardia infections in CF patients.

Patients and methods

Patients

CF patients from 3 French CF centers (Lyon, Montpellier and Toulouse) were included in our retrospective study. All had at least one sputum culture positive for Nocardia spp. between 2001 and 2011. For each patient, we recorded age, sex, clinical status, pulmonary function tests (PFT) results (forced vital capacity—FVC and forced expiratory volume in 1 s—FEV1), body mass index (BMI) at the time of Nocardia isolation and antimicrobial treatments (Table 1). Patients were classified as “patient with acute pulmonary exacerbation” as defined by Fuchs et al. [13] when they presented a combination of at least 4 of the following symptoms: change in sputum, new or increased hemoptysis, increased cough, increased dyspnea, malaise or fatigue or lethargy, temperature above 38 °C, anorexia or weight loss, sinus pain and tenderness, change in sinus discharge, change in physical examination of the chest, decrease in the pulmonary function by 10 % or more from a previously recorded value, or radiographic change indicative of pulmonary infection.

Table 1 Patients’ characteristics
Full size table

Microbiological data

In the hospital microbiological laboratories which analyzed the patients’ sputa, Nocardia strains were isolated by classical sputa analysis (culture during 5 days at 37 °C on various media, not specific of the Nocardia species). The associated pathogens and the total number of Nocardia positive sputa were recovered. At least one Nocardia isolate from each patient was sent to the French Observatory for Nocardiosis (FON, Lyon, France) which is a center specialized in the identification and antimicrobial susceptibility tests of Nocardia sp. Isolates from several sputa were sent for six patients (Table 2).

Table 2 Microbiological data and antimicrobial treatment
Full size table

Nocardia identification was performed by the sequencing of both 16S rRNA gene and hsp65 gene as recommended for the separation between phylogenetically close species [14, 15]. A 600-bp fragment of 16S rRNA gene and a 440-bp fragment of hsp65 gene were amplified by PCR according to the protocols established by Rodriguez-Nava et al. [16]. After sequencing, the sequence analysis was performed with BLAST software (nucleotide–nucleotide BLAST http://www.blast.ncbi.nlm.nih.gov/Blast.cgi) [17]. The identification at species level was done according to the similarity percentage with the reference species present in GenBank. The minimal similarity percentage requested for each identification was 99 %.

The antibiotic susceptibility testing was performed and analyzed using disk diffusion method on cation-supplemented Mueller-Hinton medium according to the CLSI standards M24-A2 and CA-SFM guidelines [18, 19]. The antibiotic disks used were amoxicillin (25 µg), amoxicillin–clavulanic acid (20/10 µg), ceftriaxone (30 µg), cefotaxime (30 µg), cefepime (30 µg), ciprofloxacin (5 µg), linezolid (30 µg), trimethoprim–sulfamethoxazole (1.25/23.75 µg), amikacin (30 µg), gentamicin (15 µg), tobramycin (10 µg), imipenem (10 µg), minocycline (30 µg), vancomycin (30 µg) and erythromycin (15 µg) (Bio-Rad, Marnes-la-Coquette, France).

Results

Eighteen patients were described as having sputum culture positive for Nocardia spp between 2001 and 2011. Two patients were excluded because of the lack of clinical and biological data. The remaining 16 CF patients (5 females and 11 males) had consistent recorded clinical data and were included in our study. Most of them carried a homozygote Δ508 mutation (6/16) or a heterozygote Δ508 mutation (5/16). Other mutations were W1282X/V232D, N1303K/W1282X and N1303K/T2640Del, whereas the CFTR genotypes of the two oldest patients (patients 10 and 16) were unknown. As expected, most of the patients (14/16 patients) were treated by oral, IV or nebulized antibiotics during the 3 months before Nocardia isolation. No patient was transplanted (Table 1).

The age of patients at first Nocardia isolation varied from 6 to 78 years (mean age = 22.68 years; median age = 15.5 years) with nine patients (56.3 %) younger than 18 years.

Concerning the clinical status at the first Nocardia isolation, 7 patients showed an acute pulmonary exacerbation (patients 5, 7, 10 and patients 12–15) (Table 1). Three of these seven patients (patients 7, 13, 14) had low spirometry results (FEV1 < 70 % pred) and 3 (patients 7, 10 and 12) had a BMI lower than 16.5. The nine patients without acute exacerbation had spirometry results close to (patient 6) or superior to 90 %. Only two of them (patients 4 and 6) had a BMI lower than 16.5. However, 5 of these 9 patients without such exacerbation (patients 1, 2, 8, 9, 11) presented at least 2 systemic (fever, loss of weight, asthenia) or respiratory changes (cough, increased sputum, modified chest X-ray) (Table 1). Notably, patient 9 was described with frequent exacerbations even though this was not the case at the time of the first Nocardia isolation. Moreover, patient 6 developed a pulmonary infection 2 weeks after the Nocardia isolation. Finally, patients 8 and 9 were described as having functional lung degradation between 1 and 3 years later, whereas Nocardia was still isolated from sputa (data not shown).

At the time of the first isolation, Nocardia was the sole pathogen in 6 patients, two patients with pulmonary exacerbation (patients 10 and 12), one patient (patient 1) with several recent respiratory changes (apparition of bronchial sounds and lobar opacity) and three patients with no more than one relevant symptom (patients 4, 6 and 16). In the 10 other cases, Nocardia was associated with other pathogens (S. aureus in 8 patients, S. maltophilia in 5 patients, P. aeruginosa in 4 patients and Aspergillus in 5 patients). Particularly, patient 7 was described as having bronchopulmonary aspergillosis.

Partial sequencing of 16S rRNA and hsp65 genes identified 5 species of Nocardia: N. cyriacigeorgica (n = 8), N. farcinica (n = 3), N. abscessus (n = 3), N. wallacei (n = 1) and N. veterana (n = 1). Patient 4 was co-colonized by two Nocardia species: N. abscessus and N. veterana (Table 2). For patient 13, the isolate sent to FON was overgrown by another organism and could only be identified at the genus level (Table 2). Most of the strains were isolated from at least two sputa (11/16 patients), and for 6 patients, more than twice (3–11 sputa). Colonization of the CF tract lasted from 1 month to up to 47 months except for patient 9 who still grows Nocardia in 2014 (Table 2).

Antimicrobial susceptibility tests were performed for 15 patients (Table 3). For patients for whom FON received more than one strain, the results of these tests were compared. Notably, for patient 4, the first and third isolated strains were susceptible to the trimethoprim–sulfamethoxazole association (cotrimoxazole), whereas the second (isolated 1 month before the third one) was resistant. For patient 9, the first strain was susceptible to imipenem, whereas the two others, isolated, respectively, 3 and 5 months later, were resistant. For patients 8, 10, 12 and 14, the different strains did not present any relevant differences.

Table 3 Antimicrobial susceptibility
Full size table

Twelve patients received an antimicrobial treatment against Nocardia immediately after the first Nocardia isolation, including six of the seven patients presenting an exacerbation, whereas two others (patients 8 and 9) were treated 1 and 3 years later, respectively. Patient 7 was treated by itraconazole for aspergillosis infection and patient 16 did not receive any specific treatment. The main first-line antimicrobial drug was oral cotrimoxazole in monotherapy (3 patients) or in association with ceftazidime, amoxicillin–clavulanic acid or ciprofloxacin and tobramycin (3 patients). For two patients (patients 6 and 8), cotrimoxazole was used as second-line treatment. The duration of cotrimoxazole treatment varied between 15 days and 3 months. The other antimicrobial drugs used were linezolid, amoxicillin–clavulanic acid, penems, cephalosporins, aminoglycosides and fluoroquinolones.

Among the six patients with exacerbations and immediately treated specifically for a Nocardia infection (as explained before, patient 7 was treated for aspergillosis), a significant improvement of respiratory symptoms and spirometry was noted for 4 patients: patients 5 and 14 for who Nocardia was isolated with other pathogens and patients 10 and 12 for who no other pathogen than Nocardia was identified. In the same way, this clinical improvement was also noted for the two patients (patients 8 and 9) treated later, i.e., when a deterioration of the lung function was observed.

Discussion

Due to the increasing life expectancy of CF patients, as well as the use of specific procedures for CF sputum analysis and the development of new methods for bacterial identification, many uncommon bacteria are identified in CF patient sputa [3]. As for several of these opportunistic pathogens, the clinical impact of the Nocardia species in CF remains unclear, notably because of the small number of reported cases. We describe a retrospective French multicentric study with 16 cases of CF patients harboring Nocardia sp. To our knowledge, our study reports, together with one from Thorn et al. [10], the largest number of Nocardia CF cases described in the literature (Table 4).

Table 4 Nocardia cases in patients with CF described in the literature
Full size table

Patients included in our study were young adults or children except for two cases. This is coherent with the life expectancy of CF patients and with the literature: Only 7 out of the 37 previously described patients (18.9 %) were older than 25 years (Table 4). However, these results are unexpected considering general data on nocardiosis. Indeed, Nocardia infections are known to be rare in the general pediatric population, and advanced age is known to be a risk factor for nocardiosis. Therefore, if Nocardia species isolated in CF patient lungs are linked to infections, these results could suggest that CF must be considered as an additional risk factor beside the age. If not, the high level of Nocardia species’ identification in young CF patients may be due to the high number of bacterial sputa analyses that are performed for these child than to the others leading to an increase in the identification of their colonizations.

Concerning the clinical status at first Nocardia isolation, nine patients presented no exacerbation of their pulmonary status suggesting a colonization of the lungs by Nocardia (Table 1), as already proposed by some authors [5, 20]. However, 5 of these 9 patients had notable systemic or respiratory symptoms. For these 5 patients as well as for the 6 with an acute exacerbation (total of patients = 11), the clinical status could be explained by the presence of Nocardia spp. as a real pathogen. This is particularly true when a clinical improvement was observed after specific Nocardia treatment or when Nocardia is the only microorganism isolated from sputa. However, in case of multimicrobial sputa, it is difficult to know which pathogen exacerbates the lung symptoms, and the role of each bacterium in the lung pathophysiology remains uncertain. Due to their known virulence in CF, the major pathogens P. aeruginosa and S. aureus must be treated in priority. However, as described by Aravantagi et al. [11], some patients receiving adapted treatment against major pathogens presented lung deterioration, whereas the clinical improvement was obvious with Nocardia therapy. Therefore, when Nocardia is the sole pathogen isolated in the lung or when the first-line antibiotics fail to restore the lung functions, diagnosis of nocardiosis must be proposed. Moreover, even if Nocardia is not always linked to pulmonary nocardiosis, its presence could enhance lung damages induced by other bacteria [21]. It should be noticed that even if Beucher et al. [7] suggested that the association Nocardia/Pseudomonas seems to be correlated to increased symptoms, in our study, only four sputa contained P. aeruginosa and only two of them were associated with acute exacerbations.

There are currently no guidelines for when (immediately or not) and how to treat the Nocardia infections, and this lack of recommendations might explain the discrepancies among the treatments observed in the literature and within our study [10, 22, 23]. Concerning our cohort, most of the patients were treated immediately after isolation of Nocardia. For two patients, the therapy was performed a long time after the isolation when the persistence of Nocardia colonization was confirmed by the multiple positive samples (no spontaneous eradication was observed) and when the clinical status of these patients got worse. Very short treatment durations were observed (only 2 weeks with amoxicillin–clavulanic acid for patient 4) reflecting the doubt of the clinician to treat and to treat for a long time, especially when no symptoms were noticed (Tables 2, 3). Indeed, antibiotic therapy against Nocardia is often associated with adverse effects [10]. As an example, cotrimoxazole could induce rash, bone marrow suppression and urinary lithiasis [23]. Moreover, CF patients receive multiple treatments and the addition of supplemental treatments should be kept to the strict minimum. However, as suggested by Khon et al., short-term treatments might expose patients to nocardiosis relapses [11]. Several patients presented repeated sputum cultures with Nocardia species even after the instauration of an appropriate antimicrobial treatment and a clinical improvement. This suggests that the Nocardia eradication could take time as earlier indicated and that a prolonged therapy must be necessary in addition to a long-term follow-up of sputum culture. Surprisingly, even with this treatment and the improvement of the symptoms, patient 9 is still colonized by Nocardia. In regard to the literature, treatments from 6 to 12 months are recommended for serious lung infections or in patients with strong immunosuppression. If symptoms are few, treatment could be reduced to 1–3 months [23]. If intravenous therapy is initially used, it could be switched to oral therapy after the clinical improvement [23]. Currently, the use of cotrimoxazole as first-line treatment is reported in the literature as in our study (Table 4) [23]. However, few resistant strains are described, as those of patients 3 and 9 (N. farcinica and N. wallacei, respectively) [24]. Surprisingly, in the case 3, despite this resistance, the patient presented a good evolution with a long-term cotrimoxazole-based treatment. This drug could be used in association with another antibiotic such as an aminoglycoside or a cephalosporin based on the severity of the infection and the Nocardia antimicrobial susceptibility pattern. However, there are some Nocardia species that are resistant to these agents. For example, N. transvalensis and N. wallacei are usually resistant to aminoglycoside antibiotics [25, 26], and this was confirmed by our patient 9. In addition, N. farcinica is most frequently resistant to cephalosporin antibiotics [26], as was the case for patients 2, 3 and 14 (Tables 3 and 4). Carbapenem, minocycline and linezolid have also been shown to be efficient against Nocardia spp. and were also used in our study [8, 22].

Various Nocardia species have been described in CF: N. asteroides complex (20 cases), N. farcinica (6 cases), N. transvalensis (3 cases), N. elegans (1 case), N. asiatica (1 case), N. cyriacigeorgica (1 case) and N. otitidiscaviarum (2 cases) (Table 4).

The current molecular methods of identification have led to the separation of N. asteroides complex mainly into two different species: N. cyriacigeorgica and N. abscessus. Previously published case reports concerning N. asteroides complex used other methods, and therefore, the exact incidence of these two species in these reports remains unknown. If some of the N. asteroides strains described in the literature were in fact N. cyriacigeorgica, our results would be consistent with those published before. Moreover, most of the species found in these patients have already been identified as the etiologic agents most commonly found in France according to the FON [N. farcinica (26 %), N. nova (20 %), N. abscessus (18 %) and N. cyriacigeorgica (12 %)]. Finally, this study is also the first one to describe the species N. wallacei and N. veterana in CF. Due to the long-term infection with N. wallacei, it could be interesting to investigate the adaptability of this pathogen in CF lungs. Moreover, a coinfection case was described in patient 4. This coinfection highlights the need to correctly differentiate the Nocardia morphotypes in the culture in order to give a correct diagnosis and a more appropriate treatment. Indeed, species may present different antibiotic susceptibility patterns.

For several patients, we could observe the evolution of the susceptibility of the isolated strain. In most cases, no modifications in these susceptibility patterns were observed, except for two cases where resistance appeared. Notably, for patient 9 who was first treated by imipenem, a resistance to this drug appeared and the antimicrobial therapy was adapted thereafter.

Pathogenic Nocardia species can grow in different environments [27]. Therefore, activities which imply a direct and repeated exposure with soil (agricultural activities, gardening, etc.) must be performed with the appropriate protection in order to avoid direct inoculations and the contact with aerosols, or simply avoided in the case of high-risk population (chronic obstructive pulmonary disease, bronchiectasis, long-term corticosteroid therapy, organ transplant, cancer, HIV, etc.) [28]. Such recommendations might be extended to the CF population or at least, since it could be constraining for children, to the immunocompromised CF patients.

According to the potential implication of Nocardia spp. in CF lung pathologies, we suggest the following therapeutic strategy for patient colonized by Nocardia spp. (Fig. 1). (1) When bronchopulmonary symptoms are present or increased and/or the PFT results show decreased lung function and/or the imaging shows opacity/nodules, antibiotics should first target and treat usual pathogens isolated in patients’ sputum. In these cases, Nocardia treatment can be postponed. Thereafter, if the clinical status remains unchanged and if the Nocardia is still present, patients should be treated specifically for Nocardia. If major pathogens are absent, Nocardia treatment must be done immediately. (2) If the patient has no symptoms, good lung function and no changes in chest X-ray since the precedent visit, the treatment could be delayed while maintaining a close clinical and microbiological monitoring. To make possible this targeted survey of the Nocardia colonization in lungs, colonized untreated patients and patients with Nocardia history should be clearly identified by the clinicians and the bacteriologists. If a change occurs in the clinical status of the patient, the treatment must be initiated if no other pathogen is incriminated. Similarly, a prolonged isolation of Nocardia in sputa must be taken into account and potentially treated despite stable clinical status. Finally, patients with particular clinical status (lung transplantation or strong immunosuppressive regimen or may be repeated corticoid treatment) must be treated because such conditions are known to induce nocardiosis [29].

Fig. 1
figure 1

Proposed guidelines for the clinical management of Nocardia infections in CF patients

Full size image

For all the CF patients, Nocardia species should be carefully searched for in sputum even if the patient has received, as in our study, a previous cotrimoxazole treatment. Indeed, several studies showed that this antibiotic does not protect from nocardiosis in immunocompromised patients particularly with malignancies or HIV infection [30, 31]. Analyzing another type of sample could be discussed since as indicated by Khon and Conrad [12], a bronchoalveolar lavage allowed the isolation of a Nocardia strain that was missed in sputum. Samples must be inoculated on rich media such as blood or chocolate agar plates or on selective media such as BCYE (buffered charcoal yeast extract medium) and observed at least 7 days at 37 °C and, ideally, at 30 °C [32]. Some Nocardia spp. with slow growth (N. abscessus, N. nova, N. elegans, N. wallacei…) could be searched for by incubating the culture for at least 20 days [8]. Molecular or proteomic techniques such as matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry technology could help in the nocardial identification after cultivation and sometimes a detection of Nocardia directly from the human clinical samples using specific molecular tools to help in the diagnosis [33, 34]. A follow-up of the antimicrobial susceptibility tests should be realized as soon as Nocardia is isolated, particularly in case of lack of clinical improvement under treatment, because some Nocardia species can be multidrug resistant.

When necessary, targeted Nocardia therapy (like cotrimoxazole) has to be initiated and should be continued based on the severity of the lung disease and the evolution of the antimicrobial susceptibility. Treatment duration less than 1 month must be proscribed because of its inefficiency and of the risk of nocardiosis relapse.

Conclusion

With the improving survival of patients with CF, the microbial spectrum described in the pulmonary system continues to evolve, implying a potential rise in the incidence of Nocardia and the need to improving our knowledge on their pathogenicity.

In this series of case reports, we have brought some evidence to consider Nocardia spp. as potential pathogens for certain CF patients and showed that the decision of antibiotic treatment at first isolation remains uncertain, while it should be considered for patients having signs of exacerbation. We therefore proposed guidelines to help clinicians in their therapeutic decisions.