Case of coexisting Nocardia cyriacigeorgica and Aspergillus fumigatus lung infection with metastatic disease of the central nervous system

Discussion

Our patient presented with haemoptysis and cough. He was found to have bilateral pneumonia. Pseudomonas stutzeri, N. cyriacigeorgica and A. fumigatus were cultured from his sputum and BAL samples. All three organisms are commonly found in the soil, further supporting our hypothesis that he acquired the infection a few weeks prior to onset of illness while repairing the engine of a farming truck that was loaded with decaying compost (he did not wear a face mask or respirator). Given his neutropenia and CD4 lymphopenia at the time of admission, it is likely that the leucopenia may have been ongoing for a while, even during the time of the above exposure. With CD4 counts of only 106×10⁹cells/L, his cell-mediated immune system was likely significantly compromised, predisposing him to Aspergillus and Nocardia. He then was started on hydroxychloroquine and prednisone for suspected lupus, which further precipitated the dissemination of these infections to the central nervous system (CNS). Lupus flare, with its associated cytopenia, and the immunosuppressive medications to treat lupus were likely the major contributors for this life-threatening opportunistic infection with disseminated Nocardia and aspergillosis.

Nocardia and Aspergillus are both found in soil and cause opportunistic infection in immunocompromised patients. Nocardia are partially acid-fast, slow-growing filamentous bacteria.2 N. cyriacigeorgica was first described in 2001 by Yassin et al, and since then it has been described as a human pathogen.3 A. fumigatus on the other hand is a fungus known for more than 100 years. Co-infection due to these pathogens in one patient is a rare finding. As both can cause lung and CNS involvement, histopathological diagnosis is often needed for definitive diagnosis. Thus a brain tissue biopsy is usually performed, either through a stereotactic brain aspiration or an open brain biopsy of the abscesses.4 5 However, to avoid possible neurological complications from these invasive procedures, we did not proceed with stereotactic aspiration of the brain abscess. Instead, because of the overall clinical picture, the predilection for Nocardia to involve the brain in up to 44% of cases,6 and the elevated Aspergillus antigen in CSF fluid, we made a presumptive diagnosis that the brain abscesses were likely from the haematogenous spread of Nocardia and Aspergillus from the lung infection. As he received antibiotic treatment for both nocardiosis and aspergillosis, his brain abscesses resolved on repeat MRI imaging after 19 months of antibiotic therapy, further supporting the diagnosis.

In 1984 Casale et al reported a case of co-infection with Nocardia and Aspergillus in a patient with CGD.7 In CGD patients, simultaneous infection due to catalase-producing organisms can be seen. DHR testing was therefore performed on our patient, and we recommend ruling out CGD in such cases. Other known risk factors for pulmonary nocardiosis and aspergillosis are pre-existing lung conditions, iatrogenic immunosuppression and impaired cell-mediated immunity. Our patient initially presented with malar rash, fever and neutropenia 2 weeks prior to his infection. A diagnosis of SLE was made with serological testing: ANA positive >1:2560, pANCA positive 1:1280, anti dsDNA positive >200, anti-histone antibody positive and low C3/C4. It is likely that both neutropenia and lymphopenia associated with lupus flare predisposed him to these opportunistic infections. Other immunological evaluations, including bone marrow biopsy, HIV, flow cytometry of serum and bone marrow and DHR testing, were all unremarkable.

With recent advances in methods of pathogen identification by molecular techniques newer strains are being identified. N. cyriacigeorgica has mostly been described in immunocompromised patients. Interestingly, Rivera et al in 2017 reported a case of paramediastinal abscess in an immunocompetent patient with no known lung disease.8 There have been several case reports of co-infection with Nocardia and Aspergillus to date, but coexisting N. cyriacigeorgica and Aspergillus infection is rarely reported.9–19 This may be partly because N. cyriacigeorgica is a newly identified strain of Nocardia. N. cyriacigeorgica resides in the soil-like A. fumigatus, and in our patient, exposure to rotten corn and soil lead to the infection. Presentation of such co-infection may vary. In contrast to our patient and another case by Gomez et al, where presentation was acute onset cough and haemoptysis, Wu et al reported a case where symptoms were ongoing for 12 months.20 21 Trastoy et al reported the third case of such co-infection with ongoing symptoms for a few weeks in a patient with lymphoma on chemotherapy.22 In 1990 Carter et al reported a case of renal transplant being treated for Aspergillus infection, where the diagnosis of co-existing Nocardia infection was made on autopsy.23 Increased vigilance to diagnose opportunistic co-infections in immunocompromised patients has frequently been emphasised by several authors as these cases have poor outcome, and diagnosis can be frequently missed.

Due to the rarity of such co-infection, there is no current consensus on the initial treatment choice or the duration of therapy. TMP-SMX is most commonly used in treatment of nocardiosis. Most Nocardia sp. demonstrate susceptibility to amikacin, carbapenems and linezolid as well.24 Limited data exists for specific drug susceptibility of N. cyriacigeorgica. Prior to the advent of molecular taxonomy, it was initially considered a part of Nocardia asteroides complex intravenous based on phenotypic characteristics. The N. asteroides group is usually resistant to ampicillin, amoxicillin-clavulanic acid, clarithromycin and ciprofloxacin, and shows susceptibility to ceftriaxone, amikacin, linezolid and imipenem. Though based on prior case reports these antimicrobials may be used in clinical settings, we strongly recommend susceptibility testing for culture isolates. As in our patient, N. cyriacigeorgica is commonly resistant to ceftriaxone, cefepime, ciprofloxacin and clarithromycin. Combination therapy is often used in severe, life-threatening nocardiosis and in cases with dissemination to the CNS. From prior literature, we know that CNS Nocardia infections are linked to worse outcomes with mortality rate as high as 47.8% in CNS nocardiosis.25 Because of the high mortality, we would advocate for the use of two commonly active agents for treatment of severe infection until susceptibility data for Nocardia is available. As in our patient, he was successfully treated with 6 weeks of combination therapy of intravenous imipenem plus TMP-SMX with subsequent de-escalation to oral TMP-SMX monotherapy. In this case we started with intravenous TMP-SMX during the initial treatment phase for 3 weeks, and then de-escalated to oral TMP-SMX when the patient clinically improved. We also checked serum sulfonamide levels (with a goal range of 100–150 µg/mL) to help adjust the dose of TMP-SMX. As for the duration of treatment, most experts recommend a minimum of 12 months for CNS Nocardia infection.26 In our case, we treated him until improvement was noted on follow-up chest X-ray and the brain abscesses fully resolved on MRI imaging (figure 2A,B).

Figure 2

(A) Chest X-ray PA view on presentation showing multiple foci of consolidation in bilateral lung most prominent in right lower lobe (marked by star) and follow-up chest X-ray with resolution of consolidation. (B) MRI of the brain showing rim-enhancing lesions within the brain (marked right arrows) and follow-up MRI showing resolution after medical treatment.

With Aspergillus co-infection, the patient was appropriately started on voriconazole as the first-line agent of choice. It is recommended for serum voriconazole level to be above 1.0 µg/mL for treatment of systemic infection. Our patient required high dose voriconazole at 400 mg two times per day to maintain a drug level of 1.1 µg/mL. This led to an elevated fluoride level and subsequently musculoskeletal symptoms suggestive of voriconazole-induced periostitis. Voriconazole-induced periostitis has been reported in literature. Wang et al in 2009 reported five cases of voriconazole-induced diffuse pain and attributed it to medication-induced fluorosis.27 Mechanism remains unclear, fluoride content of the drug is thought to cause accumulation of fluoride, but interestingly, Allen et al in in-vitro study showed that voriconazole causes increase in expression of vascular endothelial growth factor (VEGF) and platelet-derived growth factor (PDGF) by osteoblasts, unrelated to fluoride level.28 29 Skeletal fluorosis and periostitis has usually been reported as a side effect after long-term use of medication. Barajas et al in a retrospective study showed that 31/32 patients on voriconazole with musculoskeletal symptoms had elevated fluoride levels as well. The median time to fluoride measurement was 128 days after voriconazole initiation (range was 28–692 days after initiation).30 In our patient, the symptoms were noted 3 months after initiation of voriconazole treatment, and they were associated with a significantly elevated fluoride level. His fluoride level went down and normalised 3 weeks after the switch to posaconazole.