Keywords

Case Presentation

A 57 year old man with history of idiopathic panniculitis on long-term steroid treatment presented tothe hospital with a 1 month history of dyspnea, malaise, nausea and vomiting. He was hypotensive and tachycardic with an acute kidney injury. Prior to this presentation, he had been hospitalized for 2 weeks at a nearby hospital with similar symptoms and had been found to have multifocal ground glass opacities on CT of the chest. His condition after admission improved with fluid resuscitation, stress-dose steroids, and treatment for presumed sepsis. Infectious and rheumatologic workups were negative. He was noted to have bilateral pulmonary infiltrates (Fig. 73.1a, b). One week after discharge, however, he presented again with systemic inflammatory response syndrome (SIRS). A bronchoscopy was performed which revealed only acute and chronic organizing pneumonia, with no evidence of infection. A bone marrow biopsy was performed, which demonstrated a hypercellular marrow. Pulse dose steroids were initiated for organizing pneumonia and he improved significantly over the next week with continuation of empiric antimicrobials and tapering of steroids. After one week of this regimen, however, he again decompensated with hypoxia and hypotension. Worsening infiltrates were noted on his chest radiograph (Fig. 73.1c). Ferritin was elevated at 7652 ng/mL, and sIL-2R was elevated at 3432 U/mL. Trigylcerides were normal and fibrinogen was elevated. Given his unremarkable bone marrow biopsy 7 days before and elevated fibrinogen, the possibility of hemophagocytic lymphohistiocytosis (HLH) was dismissed as his hyperferritinemia and elevated sIL-2R were attributed to history of blood transfusion and occult infection. He continued to improve with empiric antimicrobials and steroids, and was discharged from the hospital 5 days after admission to the ICU to complete a course of levofloxacin on 40 mg of prednisone daily (Fig. 73.1d).

Fig. 73.1
figure 1

Chest radiograph on admission (a) demonstrated bilateral infiltrates, which corresponded to bilateral ground glass opacities and interstitial thickening on a CT performed the same day (b). After 1 month, pulmonary infiltrates were consistently more severe, despite multiple courses of antibiotics and diuresis (c). After treatment with etoposide and dexamethasone, infiltrates resolved by the time of discharge (d)

He presented to the hospital for the third time 2 days after discharge with hypotension and hypoxia requiring endotracheal intubation and vasopressors, and antibiotics and steroids were restarted. He was admitted to the ICU and his pulmonary status rapidly improved. After 3 days, however, he again decompensated with hypotension and an elevated lactate. He received intravenous fluids and intravenous steroids. Given the unclear association between changes in his antimicrobial regimen, steroid dosing, and clinical condition, his steroid dose was decreased from 40 mg prednisone daily by 10 mg each day. A repeat bone marrow biopsy was performed, which showed rare hemophagocytic macrophages but no evidence of malignancy. The next day, in the setting of reducing his steroid dose, he again had an episode of hypoxia and hypotension. Ferritin was measured at that time and was 9148 ng/mL, increased from 4074 ng/mL the day before (Fig. 73.2). sIL2-R was measured and was elevated at 3494 U/mL. A diagnosis of HLH was made, and he was started on chemotherapy with etoposide and high dose dexamethasone. Eight weeks had elapsed since his initial presentation.

Fig. 73.2
figure 2

Degree of hyperferritinemia varied significantly over the patient’s hospital course, with marked elevations at the time of diagnosis (arrow) and relapse of HLH

Question

How is hemophagocytic lymphohistiocytosis treated?

Answer

The patient was initiated on etoposide and dexamethasone without cyclosporine per the HLH-94 protocol. He rapidly improved and was moved to the hematology/oncology ward. His course was complicated by recurrent fevers and rising ferritin with the addition of filgrastim to his regimen after 3 weeks of therapy. This was discontinued and he improved uniformly from then on. After 1 month he was transferred to the physical medicine service for rehabilitation, and was subsequently discharged to home. He continued to do well 2 months after his diagnosis, and was preparing for allogeneic bone marrow transplant at that time.

HLH is a group of disorders characterized by aberrant immune activation leading to uncontrolled inflammation and, in many cases, critical illness. The underlying causes of HLH have been divided into primary causes, which encompass genetic defects in immune system regulation and often present in the neonatal period or early childhood, and secondary causes, which may be related to autoimmune disease, viral infection, or malignancy [1, 2]. However, there is growing appreciation that many adolescent and adult patients develop HLH due to the interaction of an underlying genetic defect with an environmental trigger, such as viral infection with Epstein-Barr virus or cytomegalovirus [3].

Presentation

In adults, HLH often presents with dramatic, but nonspecific, features of including fever, organomegaly, cytopenias, kidney injury, disseminated intravascular coagulation and altered mental status [4, 5]. As illustrated in the case above, in our experience rapid fluctuating hemodynamic instability and hypoxic respiratory failure are also common features resulting from the abnormal inflammatory response. These presenting symptoms and laboratory findings overlap significantly with severe sepsis. It is not uncommon for adult patients with HLH to have recently been hospitalized for presumed sepsis, and HLH is entertained as a diagnosis after one or more unexplained relapses. The most important step in recognizing HLH in the critically ill patient, therefore, is to consider it in the differential diagnosis alongside more common causes of shock and respiratory failure.

Diagnosis

While primary HLH is a genetic disorder and may be diagnosed by molecular testing with a suggestive family history and clinical presentation, secondary HLH is a clinical syndrome with diverse causes. Diagnostic criteria for HLH are currently based on the selection criteria for the HLH-2004 trial, which enrolled pediatric patients (Table 73.1, [6]).

Table 73.1 Diagnostic criteria for HLH derived from the HLH-2004 study

The presence of fever, organomegaly, cytopenia, and hypertriglyceridemia or hypofibrinogenemia, while nonspecific, is widely accepted as helpful criteria in defining the syndrome. The presence of hemophagocytosis, despite lending HLH its name, is also nonspecific [7]. The search for appropriate biomarkers, then, remains an area of active and important investigation, as detailed below. In practice, the most critical information leading to the diagnosis of HLH is often derived from following the patient over several days, observing the response, or lack thereof, to treatment for sepsis or other common conditions, and the relationship of biomarkers to the patient’s evolving clinical condition.

Treatment

The largest trial of HLH treatment are the HLH-94 and HLH-2004 trials, which investigated the treatment of HLH with etoposide, dexamethasone, and cyclosporine in children and adolescents [6, 8]. Etoposide and dexamethasone have become the mainstay of treatment in adult patients with HLH. Early consultation with a hematologist and center experienced in the treatment of HLH is critical, both because immunomodulatory therapy for HLH is an area of active investigation, and because bone marrow transplantation is considered the curative therapy for patient with a known or suspected genetic cause of HLH.

Early therapies for HLH may be limited by end-organ damage present at the time of diagnosis, especially liver and kidney injury. As highlighted in the case presented here, treatment with high dose steroids may provide adequate temporizing therapy, allowing the patient to stabilize and ultimately receive definitive therapy.

Evidence Contour

Prognosis

As the case above highlights, adult patients in the ICU are often diagnosed with HLH after several episodes of apparent improvement followed by worsening, with periods of critical illness.

There are no randomized controlled trials of HLH treatment. The most comprehensively, longitudinally studied cohort of patients with HLH was enrolled in the HLH-1994 study, which exclusively studied children under age 16 [9]. In this cohort, cumulative 5 year mortality was 46 %, with 25 % mortality within 8 weeks of diagnosis. Survival was 66 % among those patients who underwent HSCT. HLH in adults is far more likely to be secondary than HLH in the pediatric population, though late presentations of primary HLH are possible [3]. While survival to discharge in many adult cohorts mirrors the pediatric population, long-term mortality in secondary HLH may be more reflective of underlying malignancy [10, 11].

Short-term mortality from HLH depends strongly on local treatment resources and supportive care, and varies widely among single-center studies from 50 to nearly 80 % prior to discharge [4, 12]. Across centers, however, HLH related to malignancy is found to have the highest mortality, followed by autoimmune and then infectious causes. These differences may be particularly striking when advanced supportive care is available. In a North American tertiary care center, patients with HLH secondary to causes other than malignancy had a median survival of 48 months, compared to less than 2 months in cases associated with malignancy [13]. Severity of illness at the time of diagnosis of HLH also appears to bear on outcome despite treatment, with thrombocytopenia, marked hyperferritinemia, and shock at ICU admission associated with mortality [13, 14].

Biomarker Testing

As highlighted in the case presentation, diagnosis of HLH is often difficult and delayed. Multiple conditions share at least several of the diagnostic criteria of HLH. Therefore, specific biomarkers for HLH would likely shorten time to diagnosis and therapy. Bone marrow hemophagocytosis, while lending its name to HLH, is neither sensitive nor specific for the disorder [7]. NK cell activity and sIL-2R levels are part of the HLH diagnostic criteria, but these laboratory studies are available in only a few reference laboratories. Ideally, biomarker testing would rely on commonly and rapidly available assays, such as those for determining ferritin levels.

In the pediatric population, hyperferritinemia (>10,000 μg/L) has a greater than 90 % specificity for HLH [15]. Furthermore, ferritin levels also reflect disease course and activity, and in children a marked decrease in ferritin (>96 %) portends decreased short term mortality during HLH treatment [16]. In critically ill adults, however, the increased prevalence of conditions which elevate ferritin, such as infection, malignancy, autoimmune disease, liver injury and chronic blood transfusion make hyperferritinemia a nonspecific finding for HLH [17, 18]. In addition, particular causes of secondary HLH may be associated with variable biomarker patterns. For example, HLH due to lymphoma may present with a high ratio of sIL-2R to ferritin compared to other causes [19]. Given that HLH is a disorder of immune dysregulation, efforts have been made to differentiate HLH from infection based on cytokine patterns in the pediatric population [20]. Whether this approach will be fruitful in adults requires further study.