Correlation of Hemophagocytosis with Clinical Criteria of Hemophagocytic Lymphohistiocytosis and Recommendations for Screening Bone Marrow Samples in Adult Patients

Introduction

Hemophagocytic lymphohistiocytosis (HLH) is a rare condition resulting from a dysregulated inflammatory response. It can prove difficult to diagnose and portends a poor prognosis. Bone marrow (BM) biopsy is an easily accessible test that is often used to identify the presence of hemophagocytosis and assess for underlying malignancy. Currently there are no evidence-based guidelines on the reporting of hemophagocytosis on BM biopsy and no reports of a correlation between hemophagocytosis with the clinical diagnostic criteria for HLH.

We therefore aimed to assess if the amount of hemophagocytosis identified in the BM biopsy correlates with HLH-2004 criteria. Secondary aims were to evaluate inter-observer variability in reporting hemophagocytosis, and to formulate recommendations for screening in BM specimens.

Method

A retrospective review of bone marrow biopsies from adult patients under investigation for HLH was undertaken independently by 2 hematopathologists who were blinded to the original biopsy report. Relevant clinical and laboratory data was extracted from medical records.

The average number of actively hemophagocytic cells in each slide prepared from BM aspirates were quantified into 0, 1, 2-4 and ≥5. On trephine samples, hemophagocytosis was reported as either ‘present’ or ‘absent’, with the assistance of the CD68 immunohistochemical stain. Cases with discordance pertaining to the degree of hemophagocytosis were reviewed by both assessors to reach a consensus.

Results

Sixty-two specimens from 59 patients were available for assessment. An underlying hematological condition was identified in 34 cases (58%). The most common underlying hematological condition was lymphoma, found in 15 cases (25%).

There was a significant association between the amount of hemophagocytosis identified on the aspirate samples and the number of HLH-2004 criteria met (p<0.05). In patients where hemophagocytosis was present (n=31), there was a significant correlation between the amount of hemophagocytosis and ferritin levels (p<0.05). Interobserver variability was present in 63% of cases.

Based on our review, we make the following recommendations for reporting of hemophagocytosis in the BM samples:

• Count only macrophages ingesting intact hemopoietic cells.
• Quantify the average number of active histiocytes per aspirate slide.
• Count histiocytes away from particles where the cellular outline is clear.
• Avoid counting conglomerates of histiocytes where the cellular margins are indistinct
• On the aspirate specimen, assess for hemophagocytosis on both the trail and squash preparations.
• Delineating hemophagocytosis on trephine samples is difficult without the use of a CD68 immunohistochemical stain.

Interestingly, a study by Ho et al found no association between the BM histologic findings and the probability of hemophagocytosis (Ho et al, American Journal of Clinical Pathology, 2014). This difference highlights the need for standardised reporting of BM specimens.

Conclusion

Our findings indicate that the amount of hemophagocytosis present on BM samples correlates with the number of HLH-2004 criteria met. We found marked interobserver variability which we anticipate can be rectified with our recommendations on the reporting of hemophagocytosis.