Pulmonary-renal syndromes: Experience from an Indian Intensive Care Unit.

BACKGROUND: The etiology of patients presenting with pulmonary-renal syndrome (PRS) to Intensive Care Units (ICUs) in India is not previously reported. AIMS: The aim was to describe the prevalence, etiology, clinical manifestations, and outcomes of PRS in an Indian ICU and identify variables that differentiate immunologic causes of PRS from tropical syndromes presenting with PRS.
MATERIALS AND METHODS: We conducted a prospective observational study of all patients presenting with PRS over 1-year. Clinical characteristics of patients with "definite PRS" were compared with those with "PRS mimics".
RESULTS: We saw 27 patients with "provisional PRS" over the said duration; this included 13 patients with "definite PRS" and 14 with "PRS mimics". The clinical symptoms were similar, but patients with PRS were younger and presented with longer symptom duration. Ninety-two percent of the PRS cohort required mechanical ventilation, 77% required vasopressors and 61.5% required dialysis within 48 h of ICU admission. The etiologic diagnosis of PRS was made after ICU admission in 61.5%. Systemic lupus erythrematosus (54%) was the most common diagnosis. A combination of biopsy and serology was needed in the majority (69%, 9/13). Pulse methylprednisolone (92%) and cyclophosphamide (61.5%) was the most common protocol employed. Patients with PRS had more alveolar hemorrhage, hypoxemia and higher mortality (69%) when compared to "PRS mimics".
CONCLUSION: The spectrum of PRS is different in the tropics and tropical syndromes presenting with PRS are not uncommon. Multicentric studies are needed to further characterize the burden, etiology, treatment protocols, and outcomes of PRS in India.

Introduction

“Pulmonary-renal syndrome” (PRS) refers to the co-occurrence of rapidly progressive glomerulonephritis and alveolar hemorrhage.[1] Although any clinical condition with co-occurrence of renal failure and acute respiratory failure can be termed as PRS, this term is reserved for pulmonary infiltrates due to capillaritis and crescentic glomerulonephritis.[2] The most common causes of PRS include anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV), anti-glomerular basement membrane (anti-GBM) disease, and systemic lupus erythematosus (SLE). These three diseases groups contribute to more than 80% of patients with PRS in the west.[34] The syndromic presentation may be similar in all these etiologies; however, there are differences in pathogenesis, diagnostic tests required, histopathological features, treatment protocols, and prognosis.[5] In the tropics, several fulminant infections presenting with alveolar hemorrhage and renal failure may mimic PRS[6] and may be more prevalent than vasculitis presenting with PRS. A high index of suspicion is required to identify PRS early and distinguish it from “PRS mimics” because of treatment implications.[7] We conducted the present study to describe the prevalence, etiology, spectrum, and outcomes of PRS in an Indian Intensive Care Unit (ICU) and identify clinical characteristics at presentation that may help distinguish patients subsequently diagnosed with “definite PRS” from “PRS mimics.”

Materials and Methods

The study was a prospective observational study between January 2014 and December 2014 in a large tertiary care hospital in South India. All patients aged >18 years with a medical illness requiring intensive care were admitted to the medical ICU/High Dependency Unit (HDU) of our hospital. At ICU/HDU admission, clinical status examination, blood chemistry, blood gases, urine routine examination and chest radiography were performed in all patients. The measure of the severity of illness at ICU admission was performed by the Acute Physiological and Chronic Health Evaluation II (APACHE II) scoring system within 6 h of ICU admission. Choice of antibiotics, enteral nutrition protocol, blood sugar monitoring and glycemic protocol, decision on dialysis and administration of blood products, choice of fluid and vasopressors were determined by written ICU protocols.

Study subjects

Case series

Patients were diagnosed with “PRS” provisionally at ICU admission if they fulfilled all the below; acute illness (≤4 weeks) with evidence of renal and lung involvement at ICU admission, “active urinary sediment” on urine examination and pulmonary involvement consistent with diffuse alveolar hemorrhage (DAH) or vasculitis. “Active urinary sediments” was defined by the presence of at least one of the below; albuminuria (≥3+ on dipstick testing, >300 mg on 24-h collection or a urine protein/creatinine ratio >45 mg/mol), red blood cells (RBCs) or casts (RBCs or leukocyte casts) in urine microscopy examination with or without azotemia.[8] DAH was defined by the presence of at least two of the following; chest infiltrates consistent with DAH, hemoglobin ≤11 g/dL, and hemoptysis. In the absence of hemoptysis or a definite drop in hemoglobin, bronchoalveolar lavage fluid (BALF) showing grossly bloody returns or increasing bloody aliquots or hemosiderin-laden macrophages ≥20% was used as criteria for DAH.[9] All patients with “provisional PRS” who did not have a clear diagnosis at presentation underwent an echocardiogram, ultrasound of the kidneys, and appropriate bacterial cultures; the need for peripheral smears and antigen (histidine-rich protein 2) testing for malaria, IgM enzyme-linked immunosorbent assay (ELISA) for leptospirosis, nested PCR for scrub typhus or other appropriate tests was decided by the treating intensivist. Patients initially labeled as “provisional PRS,” but with no definite evidence of DAH or had a nonimmune etiology for PRS on subsequent evaluation were labeled as “PRS mimics” [Figure 1]. In patients with “definite PRS,” evaluation for the underlying etiology of PRS was initially performed with a panel of anti-nuclear antibody by indirect immunofluorescence (IIF), antinuclear cytoplasmic antibodies (ANCA) by IIF, anti-GBM antibodies by ELISA and complement levels (C3, C4). Appropriate guided-biopsies were performed from involved target organs (kidneys and/or lungs) to further characterize the individual diagnosis of PRS.[1011] Treatment was initiated when the syndromic diagnosis of “definite PRS” was made with pulse methylprednisolone 1 g daily for 3 days, followed by 1 mg/kg/day enteral prednisolone and IV cyclophosphamide 750 mg/m2 in the absence of infection or recent major gastrointestinal bleeding. Pantoprazole was administered to all patients during steroid therapy. The dose of cyclophosphamide was modified for an estimated eGFR <10 mL/min to 75% of the estimated dose. Six hundred mg mesna IV in three divided doses was administered in all patients along with cyclophosphamide. Rituximab was not used because of unavailability. Plasmapheresis (hemonitics PCS® 2 centrifugal pump) was initiated with 1.5 plasma volume exchanges (60 mL/kg) in the presence of definite DAH or serum creatinine >5.6 mg/dL and “definite PRS;” fresh frozen plasma (FFP) was used as replacement. Four percent albumin was also used when available. The ICU treating team in concurrence with the nephrology team initiated hemodialysis. Hemodialysis was performed by intermittent hemodialysis over 4 h or sustained low-efficiency dialysis over 9–12 h, when hemodynamically unstable. The decision on the number of sessions of plasmapheresis and hemodialysis and subsequent immunosuppression was decided in consensus with the rheumatology and nephrology teams. Bleeding was managed with random donor platelets or single donor platelets, FFPs, cryoprecipitate and packed red cell transfusions (RBCs) as guided by the results of platelet count, hemoglobin, prothrombin time, activated partial thromboplastin time, and fibrinogen levels. Prothrombin concentrates or activated factor VIIa were not used due to unavailability. Asymptomatic platelet counts ≤20,000/ml were treated prophylactically to a count greater than this level. Packed RBCs were transfused if ongoing major bleeding or if hemoglobin was ≤7 g/dL to a level greater than this.

Figure 1

Composite image of the chest radiograph (left) and computed tomography (CT) of the chest (right) of patient 1 with showing bilateral lower lobe consolidation with corresponding asymmetric ground glass opacities and crazy-paving on CT. A clinical diagnosis of alveolar hemorrhage (DAH) was made

Outcome measures

Data were abstracted prospectively in a predefined data collection form [Appendix 1] and a descriptive analysis was performed. The age, gender, symptoms and their duration, APACHE II scores at ICU admission, time to diagnosis after ICU admission, treatment administered and outcomes were compared between patients with “definite PRS” and “PRS mimics.”

Ethics

The study was approved by the scientific society (JSASC) and Research Ethics Committee of our Hospital. All data abstracted was anonymized, thus ensuring confidentiality and patient privacy.

Statistics

Statistical analyses were performed using SPSS version 14 (IBM Corporation, United States). Continuous variables were described in a descriptive fashion (mean ± standard deviation, median, interquartile range), and discrete variables were described as frequency proportions. Comparisons for continuous variables were performed using the Mann–Whitney test and proportions using the Chi-square test or Fisher's exact test where appropriate. Statistical significance was assessed at the 2-sided P ≤ 0.05 level.

Results

We saw 27 patients with “provisional PRS” over the said duration; this included 13 patients with a subsequent diagnosis of “definite PRS” and 14 patients with “PRS mimics” [Table 1 and Figure 2]. The “definite PRS” cohort accounted for 5% (13/254) of ICU/HDU admissions with a diagnosis of renal failure requiring dialysis and 0.9% (13/1325) of the total ICU/HDU admissions during this period.

Table 1

Summary of clinical findings of patients with immunologic etiology of PRS in our ICU (n = 13)

Figure 2

Study flow-chart

The clinical characteristics of the individual patients are described in Table 1. Two of these patients (case 9 and 12, both with IgA vasculitis) have been previously described. Patients presenting with PRS were 20–30 years and SLE was the most common etiology of PRS (Table 2, 54% of all patients). Breathlessness and leg swelling were the most common complaints; hemoptysis at presentation was seen in 46% (6/13) only. Patients were symptomatic for a median of 2 (8) weeks before seeking healthcare at any center and were admitted to our ICU after a median of another 4 (5.5) weeks evaluation at several other centers. Patients were admitted with advanced organ dysfunction to the ICU, with a mean APACHE II score of 19.8 ± 7.6 at admission. Mechanical ventilation was required in 92% (12/13) for hypoxemic respiratory failure due to DAH, 77% (10/13) required vasopressors, and 61.5% (8/13) required dialysis within 48 h of ICU admission. The etiologic diagnosis of the PRS was made after ICU admission in 8/13 (61.5%) of patients, with a median delay of 4 (5.5) days. In those with a diagnosis prior to ICU admission, a diagnosis of SLE (4/5, 100%) and IgA nephropathy (1/5) had been made at a median of 9 (12) months prior to the current symptoms. Fifteen percent (2/13) had evidence of nephrotic syndrome at presentation. Recognition of alveolar hemorrhage was mostly clinical, with 77% (10/13) having a drop in hemoglobin and 9/13 (69%) having infiltrates consistent with alveolar hemorrhage [Figures 1 and 3]. Only 3/13 (23%) of our cohort needed broncho-alveolar lavage for the diagnosis of alveolar hemorrhage and BALF showed increasing hemorrhagic returns [Figure 3] in all. The etiologic diagnosis of PRS was made using a combination of biopsy (mostly renal, Figures 4-6) and serology in most of the patients (69%, 9/13). Percutaneous or surgical lung biopsy was performed in 15% (2/13) of the “definite PRS” cohort only [Figure 5, right]. Pulse methylprednisolone followed by enteral steroids (12/13, 92%) and pulse cyclophosphamide (8/13, 61.5%) was the most common protocol employed. In those who did not receive cyclophosphamide, a fulminant fatal course (4/5, 80%) or an active infection (1/5, 20%) prevented initiation of this protocol. Plasmapheresis was administered in 46% (6/13), for a median of 4 (4.5) sessions per patient. PRS was associated with very high mortality (9/13, 69%). Mortality was biphasic with an early peak at 48 h due to fulminant DAH and later (median day 8) due to infection aggravated by immunosuppression. Morbidity in survivors was also high; chronic kidney disease in survivors was seen in 75% (3/4) and brain abscess in another patient (1/4). Two of the survivors are dialysis-dependent at 3 months after discharge (50%, 2/4).

Table 2

Summary of clinical findings of patients with immunologic cause of PRS in our ICU (n = 13)

Figure 3

Composite image of the chest-computed tomography (left) of patient 10 with granulomatosis with polyangiitis showing bilateral lower lobe air-space nodules with ground glass opacities. A thick-walled cavity in the right upper lobe was also present (not shown). Broncho-alveolar lavage showed increasingly hemorrhagic returns diagnostic of alveolar hemorrhage. Surgical biopsy confirmed granulomatosis and polyangiitis

Figure 4

Composite image of percutaneous renal biopsy specimen (left, H and E ×400) of patient 4 with systemic lupus erythrematosus showing endocapillary proliferation with basement membrane thickening and duplication. Immunofluorescence (right, FITC stain, monoclonal antibody for IgG, DAKO, USA, ×20) showed full-house pattern (also positive for IgM, C3, C1q, κ and λ [not shown])

Figure 6

Composite image of percutaneous renal biopsy specimen (left, H and E ×400) of patient 3 with microscopic polyangiitis (MPA) showing a cellular crescent and fibrinoid necrosis within the glomerulus and afferent arteriole. The photomicrograph on the right (H and E ×400, patient 8 with MPA) shows perivascular inflammatory infiltrate with few palisading macrophages and a small focus of fibrinoid necrosis. Immunofluorescence of the kidney specimens showed no staining with IgG, IgM, C3, C1q, κ and λ (pauci-immune, not shown)

Figure 5

Composite image of percutaneous renal biopsy specimen (left, H and E×400) of patient 2 with systemic lupus erythematosus showing endocapillary proliferation and cellular crescent and postmortem lung biopsy specimen showing alveolar septa expanded by inflammatory infiltrate and diffuse alveolar hemorrhage (right, H and E, ×400). Immunofluorescence of the kidney specimen showed full-house pattern (not shown)

Patients subsequently diagnosed to “PRS mimics” were older, had a high-grade fever and shorter duration of symptoms when compared to patients with PRS. Clinical symptoms and severity of illness at presentation, however, were similar in both groups [Table 3]. A drop in hemoglobin (odds ratio [OR] 20, P = 0.01) and worse hypoxemia was seen more often in the PRS group and this was associated with a greater need for transfusions and vasopressors (OR 6, P = 0.03). The diagnosis of PRS was associated with higher mortality (OR 5.6, P = 0.04) when compared to syndromes mimicking PRS.

Table 3

Comparison of clinical findings between patients with PRS and PRS “mimics”

Discussion

Small vessel vasculitis is the most common cause of PRS worldwide and this syndrome represents its polar fulminant presentation.[312] AAV and Goodpasture's syndrome are rare disorders; an incidence of 10 cases/million and 1 case/million, respectively, has been reported in Caucasians.[13] While small series of systemic vasculitic syndromes have been reported from India,[1415] prevalence in the Indian population remains unclear. SLE is a commonly recognized autoimmune disease in our population.[16] SLE presenting with alveolar hemorrhage is rare, is reported in 1–5.4% of patients[17] and often occurs in association with lupus nephritis.[1819] There is no systematic data on patients presenting with PRS from India; a search in PubMed using the terms “vasculitis” or “PRS” and “India” supplemented by an IndMed search using the term “vasculitis” turned up 1334 references with representative cases of every reported vasculitis syndrome and several tropical infections mimicking vasculitis. Large tertiary hospitals in India report a diagnosis of about 10–15 patients of systemic vasculitis each year.[15]

The exact etiology or trigger for small-vessel vasculitis remains unknown. The pathogenesis involves in-situ microangiopathy due to a Type II hypersensitivity, immune-complex mediated damage, and autoantibody driven cell-mediated damage; PRS can be classified as Type I (GPS), Type II (SLE), and Type III (AAV) based on this pathogenetic model. The nomenclature and our understanding of the individual vasculitis syndromes are evolving.[1015]

A wide variety of diseases can present with PRS [Table 4a].[2] Our cohort was different from prior published series of PRS in several aspects; a large proportion were patients in whom the diagnosis was first made in the ICU, presentation was late with severe hypoxemia and need for multi-organ support and a large proportion of patients were diagnosed with SLE. The diagnosis is usually made prior to ICU admission in the majority of patients[20], though upto 50% may be undiagnosed.[21] The larger contribution of SLE in our series might reflect a referral bias or a relative larger burden of SLE compared to primary vasculitic syndromes in South India.

Table 4a

Causes of PRS

Management of severe vasculitis in tropical ICUs poses several peculiar challenges; patients may present late without a prior diagnosis due to poor recognition of vasculitic syndromes in primary care, unavailability or longer turnaround time of specialized investigations like ANCA, difficulty or unavailability of expertise in performing guided biopsies in critically ill patients for diagnosis and lack of skilled personnel to perform specialized therapies like plasmapheresis. These add to the usual clinical dilemmas caused by the protean manifestations [Table 4b] that the vasculitis syndromes [Table 5] can present with.[2122] Finally, diagnosis may be delayed due to the similarity of the clinical syndrome to several tropical infections presenting with pulmonary and renal involvement [Table 4b]. Though the clinical findings were very similar, a diagnostic dilemma necessitating invasive surgical/percutaneous lung biopsies and delay in empiric treatment occurred in only 18.5% (5/27, Figure 7) of our cohort of “provisional PRS” patients. The diagnoses made in the “PRS mimics” cohort included pulmonary complications of immunosuppression for crescentic glomerulonephritis as well as leptospirosis, scrub typhus, and paraquat poisoning; these are all well-recognized causes of PRS in the tropics.

Table 4b

Characteristic clinical findings in PRS

Table 5

Features that help distinguish the individual PRS

Figure 7

Composite image of percutaneous lung biopsy specimen (left, H and E ×200) from a patient with crescentic glomerulonephritis and cavitating nodules showing alveoli filled with neutrophils and collection of hyphal forms. High-power view (right, Gomori-Methanamine Stain ×400) confirms thin septate fungal hyphae with acute angle branching, strongly suggestive of aspergillus species

We observed a high mortality in patients with PRS admitted to the ICU in our series. The large proportion of patients with DAH related to SLE, high APACHE II scores and the need for vasopressors and dialysis at the presentation could explain the observed poor outcome. The need for catecholamine support and renal replacement therapy is independently associated with mortality.[23] Apart from the early mortality due to fulminant DAH, infections aggravated by immunosuppression also worsened outcomes. Though mortality for PRSs is improving,[2425] survival rates as low as 20–50% remain common.[23242526]

The strengths of the study are the prospective and uniform nature of data collection and real-world experience with PRS in the tropics. The limitations include the single-center experience, limited numbers, inability to perform lung biopsy in a large majority, and lack of information on disease-specific activity scores. Lung biopsy is seldom performed in sick patients with alveolar hemorrhage undergoing mechanical ventilation and adds no information when the diagnosis of DAH is clear and disease activity is evident. Activity scores correlate only with long-term outcomes and ICU-specific severity scores have been associated with short-term outcomes,[24] and this was our practice at the time of the study.

Conclusions

The spectrum of PRS is different in the tropics and tropical syndromes presenting with PRS are at least as common as small-vessel vasculitis. SLE was the most common etiology of PRS in our cohort. Patients with PRS were younger, had a longer duration of symptoms, and had a greater drop of hemoglobin and hypoxemia with higher mortality when compared to patients with “PRS mimics.” Multicentric studies are needed to further characterize the burden, etiology, treatment protocols, and outcomes of PRS in India.

Financial support and sponsorship

Nil.

Conflict of Interest

There are no conflict of interest.