Changing Course

A 74-year-old man with end-stage renal disease (ESRD) on dialysis and cardiac sarcoidosis on chronic prednisone 20 mg daily presents to hospital with 2–3 weeks of fevers, confusion, and weakness. He is found to have a large left upper lobe cavitary mass with scattered pulmonary consolidations (Figure 1). During further work-up for possible malignancy, he is found as well to have multiple hypermetabolic subpleural and omental implants and diffuse soft tissue and musculoskeletal deposits involving the abdominal wall and lower extremities noted on a positron emitted tomography—computed tomography scan (PET-CT) (Figure 2). The cavitary chest lesion is also hypermetabolic on nuclear imaging. A magnetic resonance image (MRI) of the brain and spine show multiple rim-enhancing lesions as large as 1.6 × 1.1 cm with foci of suspected ventriculitis as well as an intra-medullary rim-enhancing lesion suspicious for abscess at the level of T5 (Figure 3). CT chest without contrast showing large cavitary mass in left upper lobe and scattered pulmonary consolidations. Abbreviation: CT, computed tomography. PET CT showing hypermetabolic pulmonary cavitary mass and disseminated visceral and soft tissue hypermetabolic nodules. Abbreviations: CT, computed tomography; PET, positron emitted tomography. (A) Contrast enhanced T1 weighted brain MRI showing largest rim-enhancing lesion measuring 1.6 × 1.1 cm, (B) STIR sequence from spine MRI showing rim-enhancing lesion suspicious for abscess at T5. Abbreviations: MRI, magnetic resonance imaging; STIR, short tau inversion recovery. This is an immunocompromised patient with a cavitary lung lesion, soft tissue deposits, and enhancing central nervous system (CNS) lesions. This constellation of findings is most suggestive of a disseminated infection or multifocal malignant process. I am most concerned about a disseminated infection, because coexistent cavitary lung lesions and brain lesions in an immunocompromised host are suggestive of the well-described “lung-brain syndrome.” This is a disease pattern that prioritizes a select group of pathogens acquired from the respiratory tract; these pathogens have a propensity to disseminate to the CNS, with or without evidence of disease in other sites. This includes opportunistic fungal organisms (molds, dimorphic fungi, and Cryptococcus species), mycobacteria, as well as bacterial etiologies including nocardiosis and actinomycosis. It is less likely that these findings are caused by the patient's underlying sarcoidosis since he is on active treatment with corticosteroids and mass-like and cavitary lesions would be unusual. Although it is tempting to search for a single unifying diagnosis following the principle of Occam's razor, Hickam's dictum suggests that a patient—particularly an immunocompromised one—can have multiple diagnoses simultaneously. It would thus be prudent to initiate a broad diagnostic workup in this patient including baseline laboratory investigations, blood cultures, serum fungal markers, a lumbar puncture, and a bronchoscopy with bronchoalveolar lavage (BAL) from the cavitary lesion. I would also discuss with my surgical colleagues if a musculoskeletal or stereotactic brain biopsy could safely be obtained if required. 48 hours following admission, 2 sets of peripheral blood cultures are positive for beaded gram-positive rods with rudimentary branching. The isolate is subsequently identified as Nocardia farcinica by MALDI-TOF. These blood culture results support the presumptive diagnosis of disseminated nocardiosis with pulmonary, CNS, and likely intra-abdominal and soft tissue involvement. Given that disseminated disease, particularly with the features of this case such as pulmonary cavities and space-occupying CNS lesions, is a well-described manifestation of nocardiosis, it would be reasonable to prioritize this as a unifying diagnosis and treat the patient accordingly—that said, should the patient not respond to initial therapy, additional secondary diagnoses should be reconsidered. Nocardia farcinica is a virulent species that is notable for its antimicrobial susceptibility profile, including elevated rates of resistance to third-generation cephalosporins, macrolides, and aminoglycosides (except amikacin), among other agents. Considering the severity of illness and risk of resistance, I would favor an empiric regimen consisting of 3 different agents with adequate CNS penetration (given the presence of infection in the brain and spinal cord) pending the results of antimicrobial susceptibility testing. Trimethoprim-sulfamethoxazole (TMP-SMX) should be considered as a first-line agent in all cases of nocardiosis, and I would also include it in this case. The dose will need to be adjusted for renal function and daily laboratory testing will be required to monitor for toxicity, particularly myelosuppression and hyperkalemia. The risk of adverse events is higher in patients with advanced renal failure, but it remains the most effective drug for nocardiosis overall, so the risk-benefit evaluation in this severe case is in favor of its use. My 2 other preferred antimicrobials would be linezolid (LZD) and moxifloxacin due to their CNS penetration and anticipated in vitro activity against most strains of N. farcinica. These agents are relatively nontoxic over short periods, although thrombocytopenia is likely to occur if linezolid is continued for longer than 2 weeks, especially in combination with TMP-SMX. In other words, I favor these agents because they are expected to be active, will reach the desired anatomic space of infection, and have a relatively low risk of toxicity. Amikacin would also be an option and is commonly used in the treatment of severe nocardiosis, but I am less inclined to use it in this patient with advanced renal disease given the increased risk of toxicity (particularly ototoxicity). It also has suboptimal CNS penetration. Other empiric 3-drug combinations would be acceptable, but regardless of the combination, I would emphasize the importance of using TMP-SMX, which is broadly considered a back-bone drug that is reliably active against Nocardia species. The patient was started on TMP-SMX 5 mg/kg IV daily, imipenem 250 mg IV BID, and amikacin 10 mg/kg after dialysis with cautious monitoring of peaks and troughs, although amikacin is quickly changed to linezolid 600 mg PO BID due to high trough levels and potential for toxicity. He quickly defervesces and his functional status begins to improve. His immunosuppression is maintained at prednisone 20 mg daily and Nocardia antimicrobial susceptibility test results become available: susceptible to amikacin, LZD, TMP-SMX, intermediate to imipenem, quinolones, tetracyclines and resistant to ceftriaxone. The isolate's antimicrobial susceptibility profile confirms elevated minimal inhibitory concentration (MIC) values to imipenem and ceftriaxone, in keeping with current epidemiological data. Although it would be prudent to continue at least 2 antimicrobial agents including TMP-SMX, these susceptibility testing results would likely allow us to safely discontinue imipenem, pending an appropriate clinical and radiologic response. In terms of next steps for follow-up, I would repeat blood cultures to document sterility, monitor for adverse events from the antimicrobials (including bone marrow suppression), and perform repeat imaging of the lesion in approximately 2 weeks to confirm that the patient is responding appropriately to therapy. I would also reassess the patient's needs for ongoing immunosuppression for the treatment of their cardiac sarcoidosis. Decreasing iatrogenic immunosuppression, if safe from the point of view of the underlying disease, is a cornerstone of managing opportunistic infections in the immunocompromised host. A bronchoalveolar lavage specimen and sputum culture are also both positive for Nocardia farcinica. Multiple sets of follow-up blood cultures are negative. Further invasive sampling is not obtained. On follow-up imaging 2 weeks after starting treatment, the cavitary lung lesion is slightly improved, the diffuse soft tissue implants are globally stable to improved, but CNS imaging has worsened. All the brain abscesses have increased in size by up to 5 mm, and several new lesions are seen (Figure 4), as well as doubling of size of the intramedullary T5 abscess and worsening overall ventriculitis. T1 weighted sequence from repeat MRI showing multiple new and worsening rim enhancing brain lesions. Abbreviation: MRI, magnetic resonance imaging. The follow-up imaging is consistent with worsening CNS disease, despite the initiation of appropriate antimicrobial therapy. Treatment failure in this scenario may be due to diagnostic error (ie, the CNS lesions are in fact not due to Nocardia) or therapeutic factors (ie, current therapy for nocardiosis is inadequate or suboptimal). Given the compelling data supporting the working diagnosis of CNS nocardiosis (including the blood and respiratory tract culture results, the characteristic CNS imaging findings, and unequivocal response to therapy for nocardiosis at non-CNS sites), it seems more likely that the progressive CNS imaging findings are due to pharmacokinetic and/or immunologic factors unique to this anatomic compartment. An immune reconstitution type reaction is a consideration, although it is less likely because the lesions have improved elsewhere; his immunosuppression has not changed, and such reactions are not clearly described in Nocardia infections. Each possibility has management implications, and I lean toward intensifying therapy (in terms of dose, frequency, and route of administration), reducing immunosuppression and consulting neurosurgery for a biopsy of the worsening lesions, knowing that the patient's prognosis would be guarded if we do not quickly identify and address the underlying process. No brain or spinal lesions are deemed safe for drainage or aspiration, and a multidisciplinary decision is made to intensify treatment for suspected multisystem nocardiosis while closely monitoring imaging and his clinical status to reevaluate diagnostic procedures soon as needed. His antimicrobial doses are increased, and the route of administration is changed to IV for all agents to eliminate any pharmacokinetic or pharmacodynamic confounders. His TMP-SMX is increased from 5 to 10 mg/kg IV daily, imipenem increased from 250 to 500 mg IV BID, and linezolid changed from 600 mg PO to 600 mg IV BID. In addition, IV amikacin is restarted. The patient's prednisone is decreased to 10 mg daily. He continues to do well clinically without fevers or neurologic symptoms. Fortunately, follow-up imaging 2 weeks later shows clear improvement in all cerebral and spinal lesions, as well as ongoing improvement in chest imaging. However, the patient has developed thrombocytopenia, with a nadir platelet count of 64 000 platelets/µL. The clinical and radiographic evolution is suggestive of an appropriate response to antimicrobial therapy. However, the decrease in platelet count is suggestive of an adverse event from his antimicrobial therapy and is likely attributed to linezolid. The patient has several risk factors for linezolid-associated thrombocytopenia, including: (i) duration of therapy beyond 2 weeks, (ii) renal impairment, and (iii) concomitant usage of other myelosuppressive agents, including TMP-SMX. Although 64 000 platelets/µL is not a dangerous level, it is likely that the platelet count will further decrease over time if no changes are made to the patient's regimen. At this time, I would consider monitoring linezolid drug levels if available and reducing the patient's dose of TMP-SMX. The patient's platelet count continues to decrease, and he continues to improve clinically. Linezolid drug levels are not readily available. The patient's TMP-SMX is decreased to 7.5 mg/kg IV daily, and his linezolid is stopped. He is continued on IV amikacin and imipenem as well and is discharged with follow-up imaging. Three months later, he remains clinically quite well, his interval imaging is much improved at all levels, and he has been switched to oral TMP-SMX monotherapy with plans to continue close follow-up while on this regimen. At 6 months follow-up, he continues to do well. In this case, the discussant quickly identified the “lung-brain syndrome,” a well-established illness script in infectious diseases. This narrowed the differential diagnosis to a limited set of infectious etiologies and led him to a diagnosis where one had not been made previously. Disseminated nocardiosis is an uncommon but severe illness in immunocompromised hosts. There are limited clinical trials to guide the management of this disease, which is complex and requires multiple drugs over prolonged periods [1]. Management reasoning (MR), a concept previously covered in this series, describes the process of clinical decision-making about different aspects of patient management [2]. A central component of MR is the development of management scripts, or schemas developed by the clinicians that help guide the management of certain clinical syndromes [3]. Management scripts encompass a variety of elements of patient management (Figure 5), and both evolve over the course of care of a given patient with time and accumulated experience for a given condition. For example, the discussant in this case described his management script for disseminated nocardiosis in an immunocompromised host, which included initial broad-spectrum, multidrug therapy (later adjusted based on susceptibilities), repeat imaging, and reduction in immunosuppression. An example of a management script adapted to this case, which organizes investigations and treatment decisions at different time points in a case [3]. The management script evolves to reflect changes in the patient's status as a function of time and based on previous decisions. Abbreviations: CNS, central nervous system; CT, computed tomography; MRI, magnetic resonance imaging. Another important component of management scripts is how to monitor response to treatment and adjust therapy if patients are not improving. This is especially important in cases where uncertainty exists regarding the optimal management of a given infection and if there is a moderate to high risk of treatment failure. In this case, the discussant and treating clinicians were faced with apparent worsening of the brain lesions on what was felt to be appropriate therapy. Given the nature of the diseases they treat, this is not an unfamiliar scenario for infectious diseases clinicians. They must therefore have a robust approach to treatment failure, considering variables ranging from an incorrect diagnosis (such as a noninfectious mimic, like cancer) or an inappropriate treatment choice (such as pharmacokinetic or pharmacodynamic failure, or a drug that is inactive against the identified pathogen) (Table 1). Possible Causes of Treatment Failure in Infectious Diseases Abbreviations: AST, antimicrobial susceptibility testing; CMV, cytomegalovirus; CRRT, continuous renal replacement therapy; ECMO, extracorporeal membrane oxygenation; IV, intravenous; TNF, tumor necrosis factor. While troubleshooting unexpected treatment failure, infectious diseases clinicians need to also simultaneously adjust their management scripts based on clinical urgency. For example, a stable patient with non-cavitary pulmonary tuberculosis (TB) who has liver injury due to antimicrobials could safely have their treatment interrupted to re-introduce antibiotics in a stepwise manner. A patient with miliary TB or TB meningitis in a similar situation would likely need to start a completely different regimen without pausing treatment. In our case, because the diagnosis was life-threatening, the discussant felt compelled to address multiple possible causes of treatment failure at once (adding additional drugs, changing the route of administration to IV, increasing doses and optimizing immune response). Overall, this case highlights a difficult-to-manage clinical scenario that was made even more challenging due to unexpected clinical worsening while on treatment. Infectious diseases clinicians must be both nimble and thoughtful in adapting to unexpected developments and toxicities, changing course when necessary, so that they can offer patients the best care while doing no harm. Financial support. No specific source of funding was received to complete this work.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.