Recent-onset systemic lupus erythematosus complicated by acute respiratory failure
CASE PRESENTATION
Chief symptoms
A 38-year-old Hispanic woman with recently diagnosed systemic lupus erythematosus (SLE) presented with fever, cough, and dyspnea.
History of the present illness
The patient was in a normal state of health until 2 months prior, when she developed symptoms of fatigue, fevers, anorexia, diffuse myalgias, sore throat, and a nonproductive cough. During an evaluation in an urgent care clinic, a chest radiograph identified right lower lobe opacities suggestive of community-acquired pneumonia, prompting a 10-day course of doxycycline. Her fevers, fatigue, and cough did not improve with antibiotics, and she additionally developed inflammatory arthritis in her hands, wrists, elbows, shoulders, and knees.
One week after completion of her antibiotic course, the patient was hospitalized for worsening joint pain, and on physical examination she was tachycardic (heart rate 125 beats per minute) with bibasilar inspiratory crackles and synovitis of multiple small and large joints, including her bilateral proximal interphalangeal joints and elbows. Testing for muscle strength was limited by joint pain. There was no rash.
Significant laboratory results included leukopenia (white blood cell [WBC] count 3.0 × 109/liter), decreased absolute lymphocyte count (ALC; 0.6 × 109/liter), thrombocytopenia (platelets 103 × 109/liter), transaminitis (aspartate aminotransferase [AST] 1,122 units/liter and alanine aminotransferase [ALT] 675 units/liter), elevated creatine kinase (CK; 639 units/liter), elevated aldolase (31.4 units/liter), and elevated inflammatory markers (erythrocyte sedimentation rate [ESR] 29 mm/hour and C-reactive protein (CRP) level 3.6 mg/dl) (Table 1). She had a positive antinuclear antibody (ANA; 1:160 titer, fine speckled pattern), positive antibody to SSA/Ro, and a slight decrease in C3 (84 mg/dl, normal range 90–190). She was negative for antibodies to double-stranded DNA, SSB/La, Sm, RNP, and antiphospholipids, including anticardiolipin and β2-glycoprotein I, as well as myositis-associated and specific antisynthetase antibodies. She was negative for viral and autoimmune hepatitis panels. She was negative for human immunodeficiency virus (HIV; enzyme-linked immunosorbent assay for HIV types 1 and 2 antibodies). Additional evaluation was notable for the presence of cytoplasmic antineutrophil cytoplasmic antibodies (1:320), but negative for antibodies to proteinase 3 and myeloperoxidase. A basic metabolic panel and urinalysis were unremarkable, and urine toxicology, including testing for cocaine, was negative. A thoracic high-resolution computed tomography (HRCT) scan revealed findings of interstitial lung disease (ILD) suggestive of a pattern of nonspecific interstitial pneumonia (NSIP).
| Day 0 (diagnosis of SLE) | Day 38 (case presentation) | Normal value | |
|---|---|---|---|
| WBC count, × 109/liter | 3.0 | 6.1 | 3.5–10.5 |
| Hematocrit, % | 30.3 | 31 | 36–45 |
| Platelets, × 109/liter | 103 | 153 | 150–450 |
| ALC, × 109/liter | 0.6 | 0.6 | 1–4.8 |
| Creatinine, mg/dl | 0.6 | 0.6 | 0.4–1.2 |
| AST, units/liter | 1,122 | 195 | 0–47 |
| ALT, units/liter | 675 | 264 | 0–47 |
| ESR, mm/hour | 29 | 34 | 0–20 |
| CRP level, mg/dl | 3.6 | 2.0 | 0–1.0 |
| LDH, units/liter | – | 295 | 98–192 |
| Urinalysis | Negative | Negative | Negative |
| Urine toxicology | – | Negative | Negative |
| HIV | Negative | – | Negative |
| ANA | 1:160, speckled | – | Negative |
| Anti-dsDNA | Negative | – | Negative |
| Anti-SSA/Ro | Positive | – | Negative |
| Anti-SSB/La | Negative | – | Negative |
| Anti-Sm | Negative | – | Negative |
| Anti-RNP | Negative | – | Negative |
| C3, mg/dl | 84 | 95 | 90–190 |
| C4, mg/dl | 39.8 | 31.9 | 13.5–50 |
| CK, units/liter | 639 | 30 | 24–195 |
| Aldolase, units/liter | 31.4 | – | 1.5–8.1 |
| cANCA | 1:320 | – | Negative |
| pANCA | Negative | – | Negative |
| Anti-PR3 | Negative | – | Negative |
| Anti-MPO | Negative | – | Negative |
| Myositis antibodies† | Negative | – | Negative |
- * SLE = systemic lupus erythematosus; WBC = white blood cell; ALC = absolute lymphocyte count; AST = aspartate aminotransferase; ALT = alanine aminotransferase; ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; LDH = lactate dehydrogenase; HIV = human immunodeficiency virus; ANA = antinuclear antibody; anti-dsDNA = anti–double-stranded DNA; CK = creatine kinase; cANCA = cytoplasmic antineutrophil cytoplasmic antibody; pANCA = perinuclear ANCA; anti-PR3 = anti–proteinase 3; anti-MPO = antimyeloperoxidase.
- † Myositis antibodies included Jo-1, Mi-2, PL-7, PL-12, EJ, OJ, signal recognition particle, Ku, PM-Scl, and U2 small nuclear RNP.
In light of her constellation of symptoms, examination findings, and serologic data, she was diagnosed with SLE and started on 1 mg/kg of prednisone, with dramatic improvement in her arthritis, myositis, transaminitis, and cough. She was discharged home 3 days later on prednisone 30 mg twice daily.
One month later, the patient returned to the emergency department with fevers, worsening dyspnea, and productive cough. She denied chest pain, rash, edema, nausea, vomiting, or diarrhea. She reported mild myalgias and arthralgias. She was febrile (38.2°C), tachycardic (heart rate 133 beats per minute), tachypneic (respiratory rate 33 breaths per minute), and hypoxic (85% oxygen saturation on room air), leading to intensive care unit (ICU) admission.
Medical history
Her medical history was unremarkable other than recent- onset SLE.
Social and family history
The patient immigrated to Colorado from Guatemala 7 years prior. She had no history of tobacco, alcohol, or illicit drug use. She had no recent travel or sick contacts. She had no known family history of autoimmune disease.
Medications
Her medications included prednisone 20 mg twice daily for the past 2 weeks, having been tapered from 30 mg twice daily that was initiated 1 month prior.
Review of systems
The patient reported anorexia, fatigue, arthralgias, and myalgias. She denied hemoptysis, sore throat, oral ulcers, headache, vision changes, abdominal pain, or dysuria.
Physical examination
The patient appeared ill and in acute respiratory distress. She was febrile, tachycardic, tachypneic, and hypoxic. She had diffuse inspiratory crackles on respiratory examination. There was no evidence of rash, synovitis, abdominal tenderness, or cardiac murmur.
Laboratory and radiologic evaluation
Laboratory data revealed an overall improvement from previous laboratory values (Table 1), including WBC count 6.1 × 109/liter, platelet count 153 × 109/liter, AST 195 units/liter, ALT 264 units/liter, C3 95 mg/dl, CK 30 units/liter, and CRP level 2.0 mg/dl. She had a stable ALC (0.6 × 109/liter) and ESR (34 mm/hour) and elevated lactate dehydrogenase (295 units/liter, normal range 98–192). She had a normal urinalysis and basic metabolic panel. Intravenous (IV) contrasted CT imaging of her chest was notable for bilateral peripheral ground-glass opacities, fibrotic changes with traction bronchiectasis most prominent in the middle and lower lung fields, right lower lobe consolidation, and no evidence of pulmonary embolism (Figure 1).
Computed tomography image of the chest showing ground-glass opacifications and consolidation in the right lower lung.
CASE SUMMARY
The patient is a 38-year-old previously healthy Hispanic woman with recently diagnosed SLE characterized by fevers, fatigue, polyarthritis, myositis, ANA, antibodies to SSA/Ro, leukopenia, lymphopenia, thrombocytopenia, transaminitis, and HRCT evidence of ILD. She had been treated with high-dose corticosteroids for the past month, and she presented to our institution with fever and acute respiratory distress with findings of diffuse ILD and right lower lobe consolidation on CT.
DIFFERENTIAL DIAGNOSIS
This patient represents a diagnostic challenge that rheumatologists often encounter in the management of patients with SLE who have multiorgan involvement and who are receiving immunosuppression: are these new symptoms a result of infection or active inflammation from SLE? The initial focus for an SLE patient with fever and acute respiratory distress should be to evaluate for an underlying infection; however, the differential diagnosis in this setting is broad and requires comprehensive and multidisciplinary evaluation (Table 2).
| Bacterial infection |
| Staphylococcus species (most commonly Staphylococcus aureus) |
| Streptococcus pneumoniae |
| Klebsiella |
| Pseudomonas aeruginosa |
| Escherichia coli |
| Acinetobacter species |
| Stenotrophomonas maltophilia |
| Nocardia asteroides |
| Fungal infection |
| Aspergillus species |
| Candida species |
| Pneumocystis jiroveci |
| Histoplasma capsulatum |
| Viral infection |
| Epstein-Barr virus |
| Cytomegalovirus |
| Coxsackievirus |
| Influenza |
| Mycobacterial infection |
| Mycobacterium tuberculosis |
| Nontuberculous mycobacteria (most commonly Mycobacterium avium) |
| Inflammatory/autoimmune |
| Pleuritis/pleural effusion |
| Lupus pneumonitis |
| Diffuse alveolar hemorrhage |
| Acute exacerbation of interstitial lung disease |
| Overlap connective tissue disease (e.g., antisynthetase syndrome) |
| Hemophagocytic syndrome |
| Immune reconstitution inflammatory syndrome |
| Aspiration-associated lung injury |
| Malignancy |
| Primary lung cancer |
| Metastatic cancer of the lung |
| Hematologic |
| Pulmonary embolism |
| Drug toxicity |
| Nitrofurantoin |
| Amiodarone |
| Methotrexate |
| Cyclophosphamide |
Bacterial sepsis
Infection is responsible for 22–25% of deaths in patients with SLE and is often attributed to medications used in the treatment of SLE (1, 2). Therefore, bacterial infection was the first consideration in the differential diagnosis for this patient and included routine bacterial infections (Staphylococcus and Streptococcus) as well as hospital-acquired infections (including methicillin-resistant Staphylococcus aureus, vancomycin-resistant Enterococcus, and Pseudomonas aeruginosa). The patient's presentation with fever, tachycardia, and hypoxia met the criteria for systemic inflammatory response syndrome; however, no bacterial infection was identified on physical examination or blood, urine, or bronchial fluid cultures.
Opportunistic infection
In addition to bacterial infection, the patient was at risk for opportunistic infection, including Pneumocystis jiroveci, Aspergillus fumigatus, Histoplasma capsulatum, Mycobacterium tuberculosis, Mycobacterium avium, and Nocardia species, given her exposure to high-dose corticosteroids. Pulmonary opportunistic infection is a common cause of morbidity and mortality in immunocompromised patients without HIV, and although dose and duration of corticosteroid use have been associated with an increased risk for opportunistic infection, no prospective studies have been performed to establish a specific risk for pulmonary opportunistic infections in corticosteroid-induced immunosuppression (3).
Pulmonary embolus
Acute and/or chronic pulmonary embolus was an early consideration in this patient given her new-onset hypoxia in the setting of underlying SLE. Although she had no history of prior thrombosis or miscarriage, antiphospholipid antibodies are present in approximately 28–40% of SLE patients and carry an increased risk for arterial and venous thrombosis (4, 5). While <40% of SLE patients with antiphospholipid antibodies develop thrombosis (2), pulmonary embolus should be considered in SLE patients with hypoxia. The patient was evaluated with serum testing for antiphospholipid antibodies and IV-contrasted CT imaging of her chest, and both were negative, making the diagnosis of pulmonary embolus an unlikely cause of her respiratory symptoms.
Acute lupus pneumonitis
Given the patient's multiorgan inflammation attributable to SLE, acute lupus pneumonitis was considered; however, acute lupus pneumonitis is an uncommon manifestation of SLE (6). It is typically a diagnosis of exclusion that is considered as the cause of hypoxia in SLE only after infectious and other etiologies have been excluded. In this patient, a number of alternative etiologies remained as more likely possibilities, particularly infection, aspiration, or chronic ILD.
Diffuse alveolar hemorrhage (DAH)
Although DAH is an uncommon manifestation of SLE (2–5%), it carries significant mortality. For that reason, DAH should be promptly evaluated with bronchoscopy in SLE patients with cough and alveolar infiltrates, especially in the setting of hemoptysis or increasing anemia. In SLE, DAH can be related to underlying pulmonary capillaritis or bland pulmonary hemorrhage, and it is most often associated with increased SLE disease activity, including active lupus nephritis (7). DAH can mimic bacterial pneumonia, cardiogenic and noncardiogenic pulmonary edema, or pulmonary embolus; therefore, bronchoscopy with bronchoalveolar lavage (BAL) was performed in this patient not only to evaluate for infection, but also to exclude DAH.
Acute exacerbation of interstitial pneumonia
The differential for the patient's worsening respiratory status also included worsening of her previously identified ILD on CT imaging. Acute exacerbations of idiopathic pulmonary fibrosis (IPF) are well described, in which lung histopathology is characterized by diffuse alveolar damage superimposed on the background ILD. Diagnostic criteria for acute exacerbation of IPF include rapid worsening of respiratory symptoms, acute hypoxemia, and new bilateral ground-glass opacities and/or consolidation on CT imaging in the absence of an alternative etiology. While criteria for acute exacerbations in connective tissue disease (CTD)–associated ILD, including NSIP, are not well established, 1 retrospective study estimated an incidence of 1.25% per year in patients with CTD-associated ILD (8), which is less than the incidence observed in IPF. Although the etiology of these acute exacerbations is unknown and the mortality is high (up to 83%), the majority of cases were not associated with a flare of the underlying CTD. In this patient, an acute exacerbation of ILD was considered less likely given the relatively focal, dense area of right lower lobe consolidation that favored organizing pneumonia either due to aspiration or infectious pneumonia.
CTD overlap syndrome
SLE can occur in association with other distinct CTDs, including an overlap syndrome with SLE and polymyositis. Since the patient had elevated muscle enzymes during her initial presentation, an overlap syndrome with polymyositis was considered. The prevalence of ILD in autoimmune inflammatory myositis ranges from 20–80%, depending on the imaging methods utilized and the presence of antisynthetase antibodies (9, 10). In this case, the patient was less likely to have a polymyositis overlap given other more likely etiologies in addition to testing negative for myositis-associated and specific antisynthetase autoantibodies.
Immune reconstitution inflammatory syndrome (IRIS)
IRIS is a condition primarily associated with initiation of highly active antiretroviral therapy (HAART) in patients with HIV and involves an imbalanced inflammatory response in which the immune system paradoxically causes self-tissue damage while attempting to eradicate an infectious organism. IRIS has been described in a variety of infections in patients without HIV in the setting of withdrawal of chronic immunosuppressant medications, including case reports of Cryptococcus-associated IRIS upon withdrawal of chronic corticosteroids in a patient with myasthenia gravis and tuberculosis-associated IRIS upon withdrawal of infliximab in patients with rheumatoid arthritis and Crohn's disease (11, 12). Our patient had been tapered from 60 mg to 40 mg of prednisone; therefore, IRIS was considered in the differential of her respiratory symptoms.
Drug-induced pneumonitis
Drug-induced pneumonitis should be considered in a patient with SLE and worsening respiratory status. The most common immunosuppressive medications used to manage SLE that are associated with pneumonitis include cyclophosphamide and methotrexate (13). Cyclophosphamide-associated lung disease can manifest as acute pneumonitis within the first 6 months of therapy or chronic fibrotic lung disease, whereas methotrexate pneumonitis can occur at any dose at any time. Of additional consideration in SLE patients, case reports of pneumonitis associated with azathioprine, mycophenolate mofetil, statins, amiodarone, nitrofurantoin, aspirin, nonsteroidal antiinflammatory drugs, leflunomide, and rituximab have all been described (14, 15).
Aspiration-associated lung injury
In the setting of fever, acute respiratory distress, and right lower lobe lung consolidation, aspiration-associated lung injury should also be considered. SLE may be associated with risk of aspiration, particularly in certain clinical scenarios often encountered in SLE patients that predispose to aspiration, including gastric reflux, vomiting, neurologic impairment, seizures, stroke, and narcotic use. Our patient was not objectively evaluated for aspiration, but aspiration-associated lung injury is an important consideration in any CTD patient with pulmonary interstitial infiltrates.
HOSPITAL COURSE
The patient was admitted to the ICU, was started on broad-spectrum antibiotics, and underwent urgent bronchoscopy with BAL. The BAL demonstrated 273 nucleated cells/ 10 ml (6% neutrophils, 12% lymphocytes, and 82% monocytes) and 217 red blood cells/10 ml. BAL bacterial cultures were negative, but direct fluorescent antibody staining was positive for P jiroveci. Subsequently, she was diagnosed with P jiroveci pneumonia (PCP), broad-spectrum antibiotics were discontinued, and she was started on oral trimethoprim/sulfamethoxazole (TMP/SMX). Additionally, prednisone was increased to 40 mg twice daily and she received supplemental oxygen support via nasal cannula at 4 liters/minute.
Initially, she showed clinical improvement; however, 3 days later she developed increasing oxygen requirements, dyspnea, and tachypnea. Given her worsening respiratory status, she was reevaluated with repeat bronchoscopy with BAL and CT chest imaging, which were both negative for new infection or DAH, but CT did show increasing bilateral consolidation and ground-glass opacities (Figure 2). Despite treatment, she continued to deteriorate and was intubated for hypoxic respiratory failure. She continued to have high ventilator requirements and ultimately required high-frequency oscillatory ventilation. She showed no clinical improvement with these interventions and died of cardiopulmonary failure 14 days after admission.
Computed tomography image of the chest showing increased consolidation and ground-glass opacities in the lower lung fields.
Autopsy results
The patient's family consented for a restricted autopsy, and a remarkable pathology is shown in Figures 3A–C. Autopsy findings included severely consolidated lungs (right lung 1,010 gm, left lung 1,090 gm, expected weight <350 gm) with areas of interstitial fibrosis and bilateral serosanguineous pleural effusions. Lung tissue also included areas of diffuse patchy intraalveolar hemorrhage that can be seen following high-frequency oscillatory ventilation. There was no evidence of capillaritis, but there was acute and organizing diffuse alveolar damage with extensive hyaline membranes, consistent with a clinical diagnosis of acute respiratory distress syndrome and organizing pneumonia. Grocott's methenamine–silver staining of the lung was negative for P jiroveci, a finding that likely occurred due to her prior antimicrobial therapy.
Gross and microscopic lung pathology. A, Marked diffuse consolidation of the pulmonary parenchyma with patchy areas of alveolar hemorrhage. B, Acute and organizing diffuse alveolar damage with hyaline membranes consistent with acute respiratory distress syndrome. C, Diffuse interstitial fibrosis.
Examination of the kidneys was limited to hematoxylin and eosin staining; however, no signs of lupus nephritis were observed. The patient demonstrated hepatomegaly (2,140 gm, 1,300–1,450 gm expected) with mild macrovesicular steatosis and a peritoneal effusion. Sections of muscle showed type II muscle fiber atrophy, but no evidence of an inflammatory muscle disease. Type II atrophy can be seen in corticosteroid myopathy and disuse atrophy, either or both of which could have been possible in this patient.
FINAL DIAGNOSIS
Pneumocystis jiroveci pneumonia leading to acute respiratory distress syndrome.
DISCUSSION
In this report, we describe a patient with recent-onset SLE treated with high-dose prednisone (>40 mg/day) for 1 month who subsequently developed acute hypoxic respiratory failure. Infection with PCP was demonstrated on BAL, and the patient died of PCP-related acute respiratory distress syndrome unresponsive to antimicrobial or corticosteroid treatment.
PCP infection: background and risk factors
PCP is the pneumonia caused by P jiroveci in immunocompromised individuals and can lead to life-threatening respiratory failure. Previously named Pneumocystis carinii, this fungus is asymptomatically present in the lungs of up to 65% of adults (16) and only leads to clinical disease and tissue damage in the setting of cell-mediated immune dysfunction, predominantly CD4 T lymphocytes.
In patients with underlying CTD, the overall incidence of PCP infection is approximately 1–2%, with rates of up to 6% described in granulomatosis with polyangiitis (Wegener's) (17). Despite this relatively low incidence, it is essential to recognize that in individuals without HIV who develop PCP, including those with underlying CTD, mortality rates reach a staggering 39–59%, compared to only 10–15% in patients with HIV who develop PCP (18). In addition, PCP in individuals without HIV often presents with a more fulminant course developing rapidly over the course of a few days, often leading to mechanical ventilation and ICU-level care. Of interest, the quantitative fungal burden of P jiroveci in the lungs of individuals with PCP infection was found to be significantly lower in patients without HIV compared to those with HIV, supporting a hypothesis that the higher mortality rates and increased severity of disease may be the result of an immune-driven inflammatory process and not a pathogen-driven mechanism of tissue damage (19).
Several risk factors for PCP have been described in individuals with CTD, including corticosteroid use, which is one of the most well described risk factors in this population. Both increased dose and duration have been associated with an elevated risk (3, 20-22). In a retrospective study by Yale and Limper specifically reviewing PCP cases in patients with CTD, the median corticosteroid duration and dose of prednisone were 16 weeks and 40 mg daily, respectively (20). Unlike in our patient, PCP is rarely described in patients with CTD receiving corticosteroids for <4 weeks, and experts have recommend PCP prophylaxis in any patient expected to receive ≥20 mg of prednisone daily for >4 weeks (23). A caveat is that these recommendations are primarily based on retrospective analysis from patients at a tertiary referral cancer center and were likely receiving chemotherapy in addition to corticosteroids, making it difficult to extrapolate these data accurately to CTD patients. Although no prospective studies evaluating PCP risk in patients with CTD have been reported, it is of interest to consider the management of patients with giant cell arteritis in which routine treatment includes high-dose prednisone alone for >1 month. PCP prophylaxis is not routinely prescribed in patients with giant cell arteritis; however, PCP is rarely reported, suggesting more than corticosteroids alone may be necessary to develop PCP infection in patients with underlying CTD.
Additionally, PCP has been associated with other immunosuppressant medications, including cyclophosphamide, methotrexate, azathioprine, cyclosporine, tumor necrosis factor α (TNFα) inhibitors, and rituximab, although the most important risk factor seems to be corticosteroids in combination with these other immunosuppressing medications (24-26). Cyclophosphamide is often used to treat severe organ-threatening flares of systemic autoimmune disease, including severe SLE, and it is known to increase the potential risk for infections, including PCP. In 1 retrospective study of 100 SLE patients treated with cyclophosphamide, the incidence of PCP was elevated at 3% (27); however, it is not surprising given the severity of illness warranting cyclophosphamide that all of these patients were concurrently receiving corticosteroids. One study that included patients with CTD where the majority (73%) was not receiving concurrent corticosteroids is the Scleroderma Lung Study, in which 79 patients with systemic sclerosis–associated ILD were treated with cyclophosphamide (28). Interestingly, with 1 year of cyclophosphamide treatment, there were no reported cases of PCP, suggesting that cyclophosphamide alone, similar to corticosteroids alone, may not be a strong enough risk factor in isolation to develop PCP.
Beyond immunosuppressing medications, other risk factors for PCP encountered in CTD patients include lymphopenia and underlying lung disease. Decreased ALC and CD4 count are associated with an increased risk for PCP; however, specific lymphocyte levels that may necessitate PCP prophylaxis have not been determined in CTD patients. CD4 T lymphocytes are essential to control PCP infection, and PCP prophylaxis in patients with HIV is recommended when CD4 counts decrease to <200 cells/μl, although this guideline does not extend to patients with CTDs. Furthermore, while 1 study found that 91% of PCP cases without HIV had a CD4 count <300 cells/μl, this cutoff level would also include up to 46% of individuals receiving corticosteroids, the majority of whom will not develop PCP (29). It may be that ALC is a better marker for PCP risk, since several studies have identified low ALC with an increased risk for PCP infection in patients with underlying CTD (21, 30-32). For example, 1 case–control study found an ALC <350 cells/mm3 in 4 (67%) of 6 patients with SLE that developed PCP compared to only 1 (5%) of 20 SLE controls that did not develop PCP (22). Further research is needed to determine if CD4 or ALC monitoring can predict PCP risk in patients with CTD.
The presence of underlying lung disease, particularly ILD, also has been strongly associated with the risk of PCP in patients with certain CTDs and warrants strong consideration for PCP prophylaxis. This association may be related to the increased colonization of P jiroveci demonstrated in studies of individuals with other chronic lung diseases (33), or it may be related to the immunosuppressive medications used in the treatment of CTD-associated ILD. Regardless, along with multiple case reports identifying PCP infection in the setting of underlying ILD (26, 34), a retrospective study of 75 hospitalized patients with underlying SLE or inflammatory myositis by Kadoya et al found that 7 (100%) of 7 PCP cases had evidence of underlying ILD compared to 6 (8.8%) of 68 without PCP receiving similar steroid doses (31). Further strengthening this association is the incidence of PCP in granulomatosis with polyangiitis (Wegener's), a disease commonly associated with underlying lung disease and one that has a PCP incidence rate consistently exceeding all other CTDs. Finally, a retrospective analysis of 26 Japanese patients with rheumatoid arthritis diagnosed with acute-onset ILD receiving anti-TNF therapy found 24 of 26 patients to have infection with PCP (35). Prospective studies are needed to further characterize PCP risk in patients with underlying CTD. In our patient, her underlying ILD as well as corticosteroid use likely contributed to her risk for development of PCP infection.
PCP infection: when to initiate prophylaxis in patients with CTD
Unlike in patients with HIV, there are no consensus guidelines at present to guide decisions regarding the use of PCP prophylaxis specific to the CTD population. As a result, uncertainty remains regarding the degree to which various risk factors contribute to the development of PCP as well as which risk factors should warrant PCP prophylaxis. In clinical practice, rheumatologists vary considerably in their decisions to administer PCP prophylaxis. In addition to the finding that only 70% of rheumatologists prescribe PCP prophylaxis, a recent survey of 727 rheumatologists reported great variability in the factors influencing these decisions (36). Only 69% of rheumatologists that reported prescribing PCP prophylaxis were influenced by the type of immunosuppressing treatment, whereas 9% were influenced by the underlying diagnosis and 8% by the medication dose. Furthermore, in those that prescribed PCP prophylaxis, 41% were not influenced by specific prednisone dose and only 17% reported contemplating prophylaxis in patients receiving prednisone doses exceeding 20 mg daily, a remarkably low proportion considering the increased risk of PCP with corticosteroids. These discrepancies in a patient population with extensive disease-related mortality and multiple prophylactic options available to prevent infection highlight a need for consensus guidelines specific to patients with underlying CTD.
As mentioned above, the overall incidence of PCP in patients with CTD is quite low; therefore, it is not appropriate to start prophylaxis in all individuals. Yet, despite this low incidence, the mortality rates are very high, highlighting the critical need for guidelines and appropriate use of PCP prophylaxis in certain CTD patients. Each of the known PCP risk factors discussed above does not seem to impart enough risk when present individually in CTD patients to warrant prophylaxis; however, a combination of risk factors appears to markedly enhance the risk for PCP. Therefore, we suggest there should be strong consideration for PCP prophylaxis in any patient with underlying CTD meeting ≥2 of 4 risk factors listed in Table 3, since they carry the highest risk for the development of PCP.
| Corticosteroids ≥20 mg for >4 weeks |
| Current use of ≥2 DMARDs, including biologic agents |
| Absolute lymphocyte count <350 cells/mm3 |
| Underlying parenchymal lung disease |
- * DMARDS = disease-modifying antirheumatic drugs.
PCP prophylaxis: management options
Each patient requiring PCP prophylaxis must be individually evaluated to weigh risks and benefits for each regimen, since they all pose certain risks in CTD patients.
TMP/SMX is considered the first-line agent for PCP prophylaxis; however, sulfonamide allergic reactions are reported in up to 31% of SLE patients, including rash, worsening SLE, and cytopenias (37). Also, TMP/SMX in combination with methotrexate has been associated with fatal pancytopenia (38, 39); however, all of these reports involved high therapeutic doses of TMP/SMX. There have been no cases associated with the lower TMP/SMX doses used for PCP prophylaxis (TMP 160 mg 3 times weekly), and although caution is warranted, TMP/SMX at prophylactic doses is not contraindicated in patients receiving methotrexate. Its dose is 1 double-strength tablet (800/160 mg) 3 times a week and the cost is $9.00/week.
Dapsone is another option for PCP prophylaxis, but it can precipitate hemolytic anemia in patients with and without G6PDH deficiency. Its dose is 100 mg daily and the cost is $1.43/day.
Another effective option is atovaquone, but it has prohibitive cost. Its dose is 1,500 mg daily and the cost is $53.50/day.
Finally, aerosolized pentamidine is another option for PCP prophylaxis, but it should be noted that it has decreased efficacy in prevention of PCP (40). Its dose is 300 mg inhaled every 4 weeks and the cost is $179.10/month (medication plus administration).
Active PCP infection: management in patients with underlying CTD
We have examined PCP risk factors and options for prophylaxis in patients with CTD. It is also valuable to consider the treatment of PCP infection in this population. TMP/SMX is the most effective treatment for PCP, and both oral and IV formulations are indicated (41-43). Because CTD patients with PCP infection have more severe respiratory disease and are more likely to require intubation (18), it is important to recognize that in the setting of severe PCP infection and significant hypoxia, infectious disease experts recommend IV administration of TMP/SMX. Of note, in the case presented herein, TMP/SMX was not available in the IV formulation due to a national medication shortage, and the patient was treated with oral TMP/SMX as an alternative. The Institute for Safe Medication Practices released a medication safety alert in late 2010 regarding drug shortages and their association with adverse patient outcomes, and in this alert, the shortage of IV TMP/SMX was associated with refractory cases of PCP and delayed treatment (44). In addition to adverse patient outcomes, drug shortages cost hospitals substantial time and money and indicate a barrier to medically necessary treatments that must be reduced in the future to ensure delivery of optimal medical care (45).
Further consideration in the management of PCP infection must take into account the proposed underlying pathophysiology of lung damage. As mentioned previously, the higher mortality rates and increased severity of disease in non-HIV PCP infections are likely related to an immune-mediated inflammatory process (19). As such, it is hypothesized that with strengthening of the immune system as corticosteroids are tapered, an immune reconstitution–like phenomenon may occur, resulting in increased inflammatory lung damage, decreased lung compliance, and impaired gas exchange. IRIS is well described in patients with HIV initiated on HAART, including cases of immune reconstitution in the setting of PCP (46), and may also play a role in non-HIV PCP infections. In the setting of IRIS in HIV, it is thought to represent a surge in lymphocyte count and proliferation leading to a T cell–driven dysregulated inflammatory response and subsequent worsening of disease. Studies of mice deplete of T and B cells infected with PCP show normal lung function and oxygenation, and it is not until they are “immune reconstituted” with functional spleen immune cells that they demonstrate impaired oxygenation in addition to decreased lung surfactant activity and impaired gas exchange (47). Given this, we suggest not decreasing corticosteroids in the setting of active PCP infection. In addition, monitoring for increasing CD4 counts during PCP infection in patients without HIV may be helpful in increasing awareness for the possibility of IRIS as an entity complicating PCP treatment (48).
Although it is not clear if the same pathophysiologic mechanisms of PCP-related organ damage in HIV apply to SLE, it is worth considering IRIS during the treatment of PCP in patients with CTD such as the patient presented herein. Our patient had histologic evidence of diffuse alveolar damage and extensive hyaline membranes in addition to interstitial fibrosis supportive of treated PCP infection with subsequent marked lung inflammatory damage. This raises the question of whether one should actually treat with an increased dose of corticosteroids in the management of PCP infection in CTD patients.
In HIV patients with severe PCP, the initiation of high-dose corticosteroids as an adjunct treatment has shown a mortality benefit (49). The recommended approach includes prednisone 40 mg twice daily for 5 days followed by 40 mg daily for 5 days followed by 20 mg daily for 11 days, but it is unclear if this benefit extrapolates to patients without HIV. Notably, 2 small retrospective studies evaluating this question found no mortality benefit with high-dose steroids in patients without HIV with severe PCP. Delclaux et al evaluated 31 cases of non-HIV PCP and found no significant difference in mortality comparing patients treated with ≥60 mg of prednisone daily (9 [39%] of 23) to no corticosteroids (4 [50%] of 8) (50); the steroid group included those on and off corticosteroids at baseline. Similarly, Pareja et al found no significant mortality benefit in 30 non-HIV PCP cases with 16 receiving ≥60 mg of prednisone daily (all increased from baseline) and 14 receiving ≤30 mg of prednisone daily (36% versus 44%); however, this study did demonstrate that subjects receiving high-dose steroids had a significantly decreased mean time of mechanical ventilation (6.3 versus 20.8 days) and mean length of ICU admission (8.5 versus 15.8 days) (51). Certainly, prospective studies are needed to discern whether high-dose corticosteroids are of benefit in CTD patients with PCP, but they do not appear to worsen mortality and may even decrease time requiring ventilation and ICU-level management.
Conclusion
In conclusion, PCP is an uncommon but devastating complication encountered by patients with underlying CTD. PCP prophylaxis is recommended in those at increased risk, and although specific guidelines do not exist in patients with underlying CTD, strong consideration for PCP prophylaxis should be given to patients with ≥2 of the following: 1) long-term moderate-dose corticosteroids, 2) multiple simultaneous immunosuppressive medications, 3) severe lymphopenia, and 4) underlying parenchymal lung disease. Maintaining or increasing corticosteroid dose may be of benefit in the treatment of PCP infection in patients with underlying CTD, and IV TMP/SMX is recommended in patients with severe PCP infection.
AUTHOR CONTRIBUTIONS
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Demoruelle had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Demoruelle, Kahr, Deane, Fischer, West.
Acquisition of data. Demoruelle, Kahr, Verilhac, Deane, West.
Analysis and interpretation of data. Demoruelle, Kahr, Verilhac, Deane, Fischer, West.
