Cytomorphologic Features Found in Cerebrospinal Fluid Specimens of Hemophagocytic Lymphohistiocytosis Patients

Abstract

Objectives

Central nervous system involvement is present in 70% of patients with hemophagocytic lymphohistiocytosis (CNS-HLH). CNS-HLH is defined by neurologic deficits, neuroimaging abnormalities, or positive cerebrospinal fluid (CSF) findings. The CSF cytomorphologic spectrum of CNS-HLH, however, has not been well investigated.

Methods

A retrospective review was performed on 64 CSF specimens from pediatric and adult patients with HLH. Ten patients had clinicoradiologic evidence of CNS involvement.

Results

We identified five CSF cytomorphologic patterns: (1) hemophagocytosis, (2) vacuolated macrophages without evidence of hemophagocytosis, (3) monocytes and/or nonvacuolated macrophages, (4) acellular specimens, and (5) bloody specimens. Patterns 1 and 2 were common in CNS-HLH and rare in patients without CNS involvement. The CSF cytomorphologic patterns did not correlate well with WBC counts or protein concentration.

Conclusions

Our study offers a comprehensive view of the cytomorphologic features seen in CSF specimens from patients with HLH.

Key Points.

• Five cytomorphologic patterns are identified in cerebrospinal fluid (CSF) from patients with hemophagocytic lymphohistiocytosis (HLH). Pattern 1 (hemophagocytosis) and pattern 2 (vacuolated macrophages) are commonly observed in patients with known central nervous system (CNS) involvement.

• The cytomorphologic patterns of CSF do not correlate well with WBC counts or protein concentration, parameters that have been previously used to diagnose CNS-HLH but are poor indicators of CNS involvement.

• In the appropriate clinicoradiologic context, cytopathologic features of CSF may be associated with an increased risk of CNS-HLH.

Hemophagocytic lymphohistiocytosis is (HLH) is a rare, life-threatening disease of extreme pathologic immune activation, particularly of cytotoxic T cells and antigen-presenting cells. Clinical manifestations include fever, hepatosplenomegaly, lymphadenopathy, rashes, and neurologic abnormalities.¹

HLH can be further classified as primary (genetic) and secondary (acquired). Primary HLH is typically seen in young children with defined genetic abnormalities affecting the lysosomal function of T cells and natural killer cells. Secondary HLH occurs in older individuals as a secondary event of infection or malignancy.^2,3 The main histopathologic feature of HLH is the diffuse infiltration of lymphocytes and macrophages into visceral organs, lymph nodes, bone marrow, and the central nervous system (CNS).² Studies have demonstrated that 40% to 70% of both adult and pediatric patients with HLH display CNS involvement. CNS-HLH is not a disease in and of itself but part of a systemic immune response. CNS involvement by HLH demonstrates a wide range of neurologic signs and/or symptoms (eg, seizures, mental status changes, meningismus, and focal neurologic signs).^4,5 This leads to a broad differential diagnosis, including other more common inflammatory diseases of the brain, such as acute disseminated encephalomyelitis (ADEM), vasculitis, multiple sclerosis, and encephalitis.⁶

So far, there is no consensus on the standard treatment of CNS-HLH, but the common regimens are similar to those used in systemic HLH, which include systemic corticosteroid combined with etoposide and cyclosporine A. Intrathecal corticosteroid and methotrexate will be added in patients with poor response or as part of the initial treatment for children. Antithymocyte globulin (ATG) and hematopoietic stem cell transplantation (HSCT) have also been attempted to treat CNS-HLH.⁶ Therapy must be initiated without inappropriate delay to prevent long-term neurologic sequelae and poor outcome.^5,7-10 The ability to recognize CNS involvement by HLH at an early stage is therefore highly critical.

While the definition of CNS disease in HLH has yet to be standardized, evaluation tends to involve (1) the presence of neurologic signs/symptoms, (2) neuroimaging abnormalities, and (3) evaluation of cerebrospinal fluid (CSF).^5,9,10 The modality of choice for evaluating CNS involvement in HLH is brain magnetic resonance imaging (MRI). Multifocal and bilateral abnormalities on T2-weighted imaging in general are the most common finding in both adult and pediatric populations. The finding of symmetric brain involvement on imaging is also helpful in supporting a diagnosis of HLH over other inflammatory entities such as ADEM.^6,9 However, a significant proportion of patients have been reported to have normal cerebral MRI scans even when presenting with both neurologic symptoms and abnormal CSF findings.^9,10 It has been theorized that this is because the neurologic symptoms are secondary to neurotoxic cytokines secreted by inflammatory cells prior to diffuse infiltration of the brain parenchyma by T lymphocytes and macrophages.^6,10

Lumbar puncture is also one of the important diagnostic tools to diagnose CNS involvement by HLH and is routinely performed on patients if there are no contraindications present. The cytomorphologic spectrum of findings in CSF involved by HLH, however, has not been well investigated. Our study, therefore, aimed to characterize the cytomorphologic spectrum of CSF involvement by HLH.

Materials and Methods

We performed a retrospective cytomorphologic study on 64 CSF specimens from 25 patients diagnosed with HLH by clinical criteria¹ at three institutions (Children’s National Health System, Washington, DC; University of Maryland Medical Center, Baltimore, MD; and Johns Hopkins Hospital, Baltimore, MD). Most specimens were supported by histopathologic confirmation of hemophagocytosis in tissue sections. The study acquired approval from the institutional review board of individual institutions.

All CSF specimens were obtained by lumbar puncture. Papanicolaou stain was performed in conjunction with either modified Giemsa or Diff-Quik stain. Pertinent clinical and histologic findings were reviewed. Student t test was performed for the statistical analysis.

Results

Our cohort consisted of 15 (60%) pediatric and 10 (40%) adult patients diagnosed with HLH with possible CNS involvement. The age of pediatric patients ranged from 6 weeks to 14 years (median, 2 years). The male to female ratio was 0.88. Among the adult cohort, the age ranged from 29 to 69 years (median, 49.5 years). The male to female ratio was 1. Two (8%) patients were believed to have developed HLH secondary to viral infection (one Epstein-Barr virus and one cytomegalovirus), two (8%) secondary to hematologic malignancy (one T-cell-rich B-cell lymphoma and one chronic lymphocytic leukemia), and one (4%) secondary to ehrlichiosis and/or immunosuppression for Crohn disease. The secondary HLH cases were all adult patients.

Seven (46.7%) of the pediatric cases and three (30.0%) of the adult cases were reported to have CNS involvement determined by the presence of neurologic signs/symptoms and/or neuroimaging abnormalities. A significant portion of the patients, especially children, had received repeated (>1) lumbar punctures: the average number of lumbar punctures was 2.6 (3.3 in pediatric and 1.5 in adult patients). Several patients had received bone marrow, liver, or brain biopsy, which showed infiltration of tissue by abundant histiocytic and occasional lymphocytic infiltrates Image 1.

Image 1.

H&E images of hemophagocytic lymphohistiocytosis involving the liver (A), brain (C), and bone marrow (E). Note the obvious presence of foamy macrophages (A, C) accompanied by lymphocytic inflammation. Macrophages are further highlighted by CD163 (B) or CD68 (D, F) immunostain. (×400)

We identified five distinct CSF cytomorphologic patterns within this CNS-HLH cohort: (1) hemophagocytosis, defined by the presence of internalized cells; (2) the presence of vacuolated macrophages without evidence of hemophagocytosis; (3) the presence of monocytes and/or nonvacuolated macrophages; (4) acellular specimens; and (5) bloody specimens Image 2.

Image 2.

A, B, Cerebrospinal fluid (CSF) pattern 1: hemophagocytosis with an internalized lymphocyte. C, Hemophagocytosis with an internalized neutrophil. D, Hemophagocytosis with an RBC. Note that the hemophagocytic macrophages demonstrate vacuolated cytoplasm. E, CSF pattern 2: vacuolated macrophages. Note that in patterns 1 and 2, the nuclei are eccentric in macrophages. F, CSF pattern 3: monocyte or nonvacuolated macrophage. (Modified Giemsa stain, ×400)

Of the 64 CSF specimens, 12 (18.8%) were pattern 1, 12 (18.8%) were pattern 2, 32 (50.0%) were pattern 3, 6 (9.4%) were pattern 4, and 2 (3.1%) were pattern 5. Among the pediatric cohort, 12 (24.5%) were pattern 1, 9 (18.4%) were pattern 2, 24 (49.0%) were pattern 3, 3 (6.1%) were pattern 4, and 1 (2.0%) was pattern 5. Among the adult cohort, 0 (0%) were pattern 1, 3 (20.0%) were pattern 2, 8 (53.3%) were pattern 3, 3 (20.0%) were pattern 4, and 1 (6.7%) was pattern 5 Table 1 and Table 2.

Table 1.

Pediatric Patients With Suspected HLH and Subsequent Pathologic Findings^a

Case No.	Age	Sex	CNS-HLHb	Specimen	CSF WBC, Cells/µL	CSF Protein, mg/dL	CSF Cytology Pattern
1	4 y	M	+	CSF (LP) #1	12	134	3
				CSF (LP) #2	3	NA	1
				CSF (LP) #3	2	48	2
				CSF (LP) #4	2	33	3
				CSF (LP) #5	96	NA	1
				CSF (LP) #6	2	NA	3
				CSF (LP) #7	3	NA	3
				CSF (LP) #8	2	NA	3
				CSF (LP) #9	2	52	3
				CSF (LP) #10	7	103	3
				CSF (LP) #11	0	39	1
				CSF (LP) #12	1	97	3
2	2 y	M	–	CSF (LP) #1	3	30	3
3	10 y	M	+	CSF (LP) #1	0	42	1
4	7 w	M	+	CSF (LP) #1	4	25	1
			+	CSF (LP) #2	0	NA	3
				CSF (LP) #3	0	347	4
				CSF (LP) #4	0	29	3
5	1 y	F	+	CSF (LP) #1	6	29	1
				CSF (LP) #2	1	31	3
6	7 w	M	+	CSF (LP) #1	18	92	1
				CSF (LP) #2	1	NA	2
				CSF (LP) #3	1	NA	3
				CSF (LP) #4	1	25	1
				CSF (LP) #5	0	22	3
				CSF (LP) #6	0	26	2
				CSF (LP) #7	1	24	2
				CSF (LP) #8	NA	NA	4
7	8 y	F	+	CSF (LP) #1	3	58	1
				CSF (LP) #2	1	40	3
				CSF (LP) #3	1	NA	2
				CSF (LP) #4	0	22	1
				CSF (LP) #5	1	22	2
				CSF (LP) #6	0	18	3
				CSF (LP) #7	0	75	1
				CSF (LP) #8	1	40	3
				CSF (LP) #9	1	18	3
				CSF (LP) #10	0	NA	4
8	5 d	F	+	CSF (LP) #1	1	127	1
9	6 w	F	–	CSF (LP) #1	4	86	3
				CSF (LP) #2	0	23	3
10	3 y	M	–	CSF (LP) #1	6	35	3
11	4 y	F	–	CSF (LP) #1	1	22	3
				CSF (LP) #2	1	NA	3
19	13 y	F	–	CSF (LP) #1	1	40	3
22	10 mo	F	–	CSF (LP) #1	325	3.1	5
23	14 y	F	–	CSF (LP) #1	6	46	2
				CSF (LP) #2	2	37	2
25	15 mo	M	–	CSF (LP) #1	0	37	2

CSF, cerebrospinal fluid; CNS, central nervous system; HLH, hemophagocytic lymphohistiocytosis; LP, lumbar puncture; NA, not available; +, positive; –, negative.

^aPattern 1, hemophagocytosis; pattern 2, vacuolated macrophages without evidence of hemophagocytosis; pattern 3, monocytes and/or nonvacuolated macrophages; pattern 4, acellular specimens; and pattern 5, bloody specimens.

^bCNS-HLH status was determined by neuroimaging.

Table 2.

Adult Patients With Suspected HLH and Subsequent CSF Cytologic Findings

Case No.	Age, y	Sex	CNS-HLHa	Specimen	CSF WBC, Cells/µL	CSF Protein, mg/dL	CSF Cytology Pattern
12	63	M	–	CSF (LP) #1	1	70	2
13	62	M	–	CSF (LP) #1	1	47	3
14	36	F	–	CSF (LP) #1	1	37	2
15	36	F	–	CSF (LP) #1	1	49	4
16	50	F	–	CSF (LP) #1	4	53	3
17	29	M	–	CSF (LP) #1	1	33	3
18	69	M	–	CSF (LP) #1	7	50	3
20	41	F	+	CSF(LP) #1	0	28	3
				CSF (LP) #2	0	25	4
21	53	M	+	CSF (LP) #1	2	48	2
				CSF (LP) #2	1	44	3
				CSF (LP) #3	1	31	3
				CSF (LP) #4	2	54	5
				CSF (LP) #5	0	28	4
24	49	F	+	CSF (LP) #1	NA	NA	3

CSF, cerebrospinal fluid; CNS, central nervous system; HLH, hemophagocytic lymphohistiocytosis; LP, lumbar puncture; NA, not available; +, positive; –, negative.

^aPattern 1, hemophagocytosis; pattern 2, vacuolated macrophages without evidence of hemophagocytosis; pattern 3, monocytes and/or nonvacuolated macrophages; pattern 4, acellular specimens; and pattern 5, bloody specimens.

^bCNS-HLH status was determined by neuroimaging.

Of note, vacuolated macrophages (pattern 2) and pleocytosis in the form of monocytes and/or nonvacuolated macrophages (pattern 3) are nonspecific findings that are frequently seen in patients with HLH without clinical or radiologic evidence of CNS involvement Table 3. Excluding these two patterns, the overall average rate of pattern 1 (hemophagocytosis) in any given individual’s CSF was 35.5% in CNS-HLH and 0% in HLH without CNS involvement. In children, the average rate of pattern 1 CSF was 50.7% from CNS-HLH and 0% from HLH without CNS involvement. In adults, however, hemophagocytosis was completely absent in the CSF, even in those with clinical or radiologic evidence of CNS involvement. The calculation took account of all CSF, including pre- and posttreatment samples, although subsequent lumbar punctures after the second one appeared less important for diagnosis, given that they were mostly posttreatment.

Table 3.

Cytopathologic Findings of CSF in Adult and Pediatric Patients With HLH With Suspected CNS Involvement Stratified by Cytomorphologic Pattern^a

Case No.	CNS-HLHb	No. of LPs	Pattern 1	Pattern 2	Pattern 3	Pattern 4	Pattern 5
Pediatric
1	+	12	3 (25.0)	1 (8.3)	8 (66.7)	0 (0)	0 (0)
2	–	1	0 (0)	0 (0)	1 (100)	0 (0)	0 (0)
3	+	1	1 (100)	0 (0)	0 (0)	0 (0)	0 (0)
4	+	4	1 (25.0)	0 (0)	2 (50.0)	1 (25.0)	0 (0)
5	+	2	1 (50.0)	0 (0)	1 (50.0)	0 (0)	0 (0)
6	+	8	2 (25.0)	3 (37.5)	2 (25.0)	1 (12.5)	0 (0)
7	+	10	3 (30.0)	2 (20.0)	4 (40.0)	1 (10.0)	0 (0)
8	+	1	1 (100)	0 (0)	0 (0)	0 (0)	0 (0)
9	–	2	0 (0)	0 (0)	2 (100)	0 (0)	0 (0)
10	–	1	0 (0)	0 (0)	1 (100)	0 (0)	0 (0)
11	–	2	0 (0)	0 (0)	2 (100)	0 (0)	0 (0)
19	–	1	0 (0)	0 (0)	1 (100)	0 (0)	0 (0)
22	–	1	0 (0)	0 (0)	0 (0)	0 (0)	1 (100)
23	–	2	0 (0)	2 (100)	0 (0)	0 (0)	0 (0)
25	–	1	0 (0)	1 (100)	0 (0)	0 (0)	0 (0)
Average, %		3.3	23.7	17.7	48.8	3.2	6.7
Adult
12	–	1	0 (0)	1 (100)	0 (0)	0 (0)	0 (0)
13	–	1	0 (0)	0 (0)	1 (100)	0 (0)	0 (0)
14	–	1	0 (0)	1 (100)	0 (0)	0 (0)	0 (0)
15	–	1	0 (0)	0 (0)	0 (0)	1 (100)	0 (0)
16	–	1	0 (0)	0 (0)	1 (100)	0 (0)	0 (0)
17	–	1	0 (0)	0 (0)	1 (100)	0 (0)	0 (0)
18	–	1	0 (0)	0 (0)	1 (100)	0 (0)	0 (0)
20	+	2	0 (0)	0 (0)	1 (50.0)	1 (50.0)	0 (0)
21	+	5	0 (0)	1 (20.0)	2 (40.0)	1 (20.0)	1 (20.0)
24	+	1	0 (0)	0 (0)	1 (100)	0 (0)	0 (0)
Average, %		1.5	0	22.0	59.0	17.0	2.0
Overall average, %		2.6	14.2	19.4	52.9	8.7	4.8

CNS, central nervous system; HLH, hemophagocytic lymphohistiocytosis; LP, lumbar puncture; +, positive; –, negative.

^aThe table shows the distribution of pattern 1 to 5 specimens associated with a given patient. Values are presented as number (%) unless otherwise indicated. Pattern 1, hemophagocytosis; pattern 2, vacuolated macrophages without evidence of hemophagocytosis; pattern 3, monocytes and/or nonvacuolated macrophages; pattern 4, acellular specimens; and pattern 5, bloody specimens.

^bCNS-HLH status was determined by neuroimaging.

Furthermore, we analyzed the relationship between cytomorphologic patterns and WBC count and protein concentration in CSF (Tables 1 and 2; Figure 1). Although specimens with pattern 1 demonstrated a higher WBC count (11.0 ± 27.2 cells/μL) than the other patterns (pattern 2, 1.5 ± 1.6 cells/μL; pattern 3, 2.2 ± 2.7 cells/μL; pattern 4, 0.2 ± 0.4 cells/μL; and pattern 5, 2.0 ± 0 cells/μL), the difference was not statistically significant due to a wide variation of the data values in the pattern 1 group (Figure 1A). In terms of the protein concentration, pattern 4 had the highest value (112.3 ± 156.9 mg/dL), but again, the difference was not statistically significant compared with the other patterns (pattern 1, 53.4 ± 34.8 mg/dL; pattern 2, 39.5 ± 14.4 mg/dL; pattern 3, 45.6 ± 29.2 mg/dL; and pattern 5, 54.0 ± 0 mg/dL; Figure 1C). In addition, we compared the WBC count and protein concentration between cases with and without CNS involvement. Cases with CNS involvement did not show statistically higher WBC count or protein concentration than cases without CNS involvement (WBC count, 3.3 ± 5.1 vs 2.4 ± 1.9 cells/μL; protein concentration, 55.6 ± 37.8 vs 43.4 ± 16.5 mg/dL). The result suggests that CSF WBC counts or protein concentration may not be reliable parameters to diagnose CNS-HLH.

Figure 1.

Cerebrospinal fluid (CSF) WBC count (A) and protein concentration (C) associated with each cytologic pattern, as well as WBC count (B) and protein concentration (D) of CSF from patients with hemophagocytic lymphohistiocytosis with or without central nervous system involvement (Student t test; NS, nonsignificant; one case of traumatic tapping was excluded from the analysis).

Discussion

Abnormal CSF findings are seen in a significant proportion of HLH cases. These abnormalities include elevated cell count, pleocytosis, elevated protein levels, and hemophagocytosis.^6,10 While examination for hemophagocytosis is part of the routine workup of CSF in patients with suspected HLH, most literature to date has focused on the laboratory values. One of the larger prospective HLH studies, HLH-94, reported the presence of hemophagocytosis in the CSF in 50 (32%) of 158 pediatric patients,¹¹ and Deiva et al¹⁰ reported 11 (24%) of 46 children to have hemophagocytosis. An older and smaller study by Komp et al¹² also reported a similar frequency of hemophagocytosis in the CSF in 33% (3/9) of children with hemophagocytic syndromes. The frequency of hemophagocytosis found in the CSF in adult HLH has not been reported. In contrast, hemophagocytosis has been described in up to 91% of brain biopsy specimens, most commonly located in the meninges.⁴ Brain biopsy is not a common procedure for diagnosing CNS-HLH but is occasionally performed in patients who have positive brain imaging findings but did not receive a bone marrow biopsy.¹³ However, the invasive nature of a brain biopsy makes it a less than ideal diagnostic modality for CNS-HLH.

Compared with brain biopsy, cytomorphologic examination of CSF is a much less invasive procedure to evaluate for CSF involvement by HLH. To our knowledge, this is the first study to characterize the cytomorphologic findings in CSF specimens from patients with HLH beyond the presence of hemophagocytosis. The definitional criteria of an abnormal CSF are inconsistent among the limited studies of HLH. Horne et al⁵ reported 52% of pediatric patients with HLH to have an abnormal CSF by elevated protein or cell count but did not provide cutoff values. Jovanovic et al¹⁴ defined CNS involvement as elevated CSF protein (>350 mg/L) or cell count (>5 cells/μL). Kim et al⁸ classified patients to have CNS involvement based on the presence of neurologic symptoms and elevated WBC counts in the CSF, regardless of the presence of protein in the CSF. Of note, our data suggest that neither WBC counts nor protein concentration alone is a reliable indicator of CNS involvement by CSF, and therefore, CSF cytomorphology may offer a more accurate way to diagnose CNS-HLH.

In this study, we observed the frequent presence of hemophagocytosis in CSF from patients with CNS-HLH, but only in children. The finding of vacuolated macrophages in CSF, although nonspecific, can appear occasionally in both pediatric and adult patients with HLH with CNS involvement. The differential diagnosis for vacuolated macrophages within the CSF is broad, as monocytic activation is frequently found in association with both inflammatory and noninflammatory CNS processes. The macrophages may also be subdivided into different classes by the vacuolar content, such as lipophages, erythrophages, siderophages, and lymphophages.^15,16 Lipophages, as demonstrated by Sudan black or oil red O staining, in particular are associated with destruction of CNS tissue.^15,17 The presence of vacuolated macrophages has been most studied in association with multiple sclerosis and has been found more frequently in cases of long disease duration.¹⁵ Vacuolated macrophages in the CSF have also been reported in cases of meningitis,¹⁸ intraventricular hemorrhage,¹⁹ Lyme disease,^20,21 schizophrenia,²² and after myelogram.²³ One study by Chester et al²⁴ identified fat-laden macrophages in pediatric patients with a variety of conditions, including bacterial and viral meningitis, leukemia, tumors, and hydrocephalus.

In children with CNS disease of HLH, the average rate of hemophagocytosis in CSF was 50.7%, including posttreatment specimens. These findings suggest that cytopathologic evaluation of CSF may be particularly useful in identifying CNS involvement in pediatric cases of HLH. CSF cytology in adult patients with HLH, in comparison, is of less diagnostic value, given that none of the cases with known CNS involvement had findings of hemophagocytosis. As most adult HLH is believed to be an acquired disease, most commonly due to malignancy or inflammatory/infectious disorders, the variation in findings may be secondary to treatment of previous or coexisting conditions. In addition, while clinical features are similar in children and adults, there is debate as to whether HLH is the same disease between the two populations and whether different diagnostic criteria are warranted.²⁵ The discrepancy in CSF findings may then also represent an underlying difference between adult and pediatric HLH and its clinical course.

Diagnosing CNS-HLH requires a multimodal approach, as the clinical, laboratory, and radiologic findings all include broad differential diagnoses. Horne et al⁶ recommend regarding the presence of neurologic symptoms and/or signs or any abnormality in the CSF or brain MRI compatible with an inflammatory process to be consistent with a diagnosis of CNS-HLH and that therapy should be started in all HLH cases with neurologic symptoms even if a lumbar puncture or MRI has not been obtained. Studies have reported abnormal CSFs demonstrating pleocytosis and/or increased protein in 16% to 76% of HLH cases.^5,6,10,11,26 One study by Howells et al²⁷ reported increased CSF neopterin in 2 (50%) of 4 pediatric patients with hemophagocytic disorders, while another study by Janka²⁸ reported increased CSF neopterin in 6 (9%) of 65 patients. Deiva et al¹⁰ reported an abnormal MRI in 15 (33%) of 46 children with primary HLH, 13 (45%) of 29 in children with neurologic symptoms at disease onset, and 2 (12%) of 17 in children without neurologic symptoms. Fitzgerald et al²⁹ reported abnormal radiologic findings in 16 (88%) of 18 patients, although this included both MRI and computed tomography examinations. To our knowledge, no studies definitively examining the sensitivity or specificity of any one modality in diagnosing CNS-HLH have been published. Based on our data, CSF cytology appears comparable and possibly superior to other modalities for the evaluation of CNS-HLH, especially in pediatric patients.

In this study, we have found hemophagocytosis to be a common and specific CSF finding in pediatric HLH with CNS involvement. Other common but nonspecific findings include vacuolated macrophages and pleocytosis. Although further studies are required to validate these data, it appears that the microscopic examination of CSF can, along with other clinicoradiologic findings, contribute to a diagnosis of CNS-HLH.

This work was supported by National Institutes of Health grant K08NS102468 to Dr Ho.