Article Text

Concise report
The multifaceted clinical presentations and manifestations of Erdheim–Chester disease: comprehensive review of the literature and of 10 new cases
  1. Giulio Cavalli1,
  2. Barbara Guglielmi1,
  3. Alvise Berti1,
  4. Corrado Campochiaro1,
  5. Maria Grazia Sabbadini1,2,
  6. Lorenzo Dagna1,2
  1. 1Vita-Salute San Raffaele University, Milan, Italy
  2. 2The Unit of Medicine and Clinical Immunology, San Raffaele Scientific Institute, Milan, Italy
  1. Correspondence to Professor Lorenzo Dagna, Unit of Medicine and Clinical Immunology, Vita-Salute San Raffaele University, San Raffaele Scientific Institute, Via Olgettina, 60, Milano I-20132, Italy; lorenzo.dagna{at}unisr.it

Abstract

Objectives Erdheim–Chester disease (ECD) is a rare inflammatory disorder characterised by organ infiltration by non-Langerhans’ histiocytes. Although rare, ECD is clearly an overlooked diagnosis. No data specifically addressing the most frequent presentations of ECD at the time of onset in a large cohort of patients are currently available.

Methods We reviewed all the published cases in the English literature of histologically-confirmed ECD. We excluded reports in which data regarding onset and diagnosis were not univocal, as well as repeated reports of the same case(s). We also included in the analysis 10 new unpublished patients from our cohort. We analysed the disease presentation with particular regard to the manifestations that induced patients to seek medical attention and their subsequent evolution.

Results In the cumulative cohort of 259 cases, ECD predominantly presented with skeletal symptoms, diabetes insipidus, neurological and constitutional symptoms. Diabetes insipidus and constitutional symptoms, if not present at onset, seemed to only seldom develop. There were differences in ECD presentation and course among different age groups of patients.

Conclusions Physicians should be aware of the extraordinarily heterogeneous clinical presentations and manifestations of ECD in order to include ECD in the differential diagnosis of several conditions.

  • Epidemiology
  • Inflammation
  • Autoimmune Diseases

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Erdheim–Chester disease (ECD) is a rare, systemic inflammatory disease of unknown aetiology characterised by multi-organ infiltration by CD68-positive, CD1a-/S100-negative lipid-laden macrophages.1 The pathogenesis of ECD is largely obscure. We and others identified a cytokine–chemokine network resembling the ‘cytokine storm’ described in Langerhans’ cell histiocytosis which may be responsible for recruitment and activation of histiocytes in ECD lesions.2–5 Although ECD is undoubtedly rare, it is arguably a largely overlooked diagnosis. Thus, data available on the overall incidence of this disease are not completely reliable.1 ,6 The disease virtually always involves the long bones.1 In about half of the cases ECD may also feature extra-skeletal involvement determining a significant worsening in prognosis, particularly when specific organs are involved, such as the central nervous system, heart and lung.1 ,6 ,7

Since ECD has protean clinical presentations and subsequent manifestations may develop after several years,6 a timely diagnosis is puzzling. Here, we describe the clinical presentations and clinical evolution of ECD, with specific regard to the manifestations that induced patients to seek medical attention.

Patients and methods

We searched the English-language literature indexed in PubMed using the keywords ‘Erdheim-Chester’ or ‘Erdheim-Chester disease’ and published up to June 2012. To ensure medical accuracy, two independent medical doctors reviewed the full texts of all the retrieved by our literature search. A third doctor rechecked the analysis for consistency. We only included in our analysis histologically-confirmed cases and excluded reports in which data regarding onset and diagnosis were not univocal, as well as repeated reports of the same case(s). We also included in our analysis 10 unpublished patients with ECD currently followed at our Institution. Statistical analyses were performed using GraphPad Prism V.5.0b for Macintosh (GraphPad Software, San Diego, California, USA).

Results and discussion

ECD is a still largely overlooked disease, mostly affecting middle-aged men. Its clinical spectrum at presentation is broad, since the infiltration of different organs leads to the development of protean manifestations. Current literature identified bone, central nervous system and retroperitoneal as the most common localisations of ECD.1 ,6 Still, the clinical picture at the time of onset may often be substantially different and seldom specific for ECD. Thus, the diagnosis is particularly challenging at onset. Furthermore, as successive manifestations can occur over years, the clinical picture may be a slowly forming mosaic.

In order to characterise the clinical manifestations of ECD at the time of the presentation, we performed a thorough review of the published literature and included 10 unpublished patients from our cohort, with particular regard to those symptoms that first induced patients to seek medical attention.

We identified 341 manuscripts describing patients diagnosed with ECD published in the English language up to June 2012. As previously specified, in order to avoid discrepancies, we only considered histologically-proven ECD cases, and excluded reports in which data regarding onset and diagnosis were not univocal, as well as repeated reports of the same case(s). We eventually found 249 patients fulfilling the search criteria (see online supplementary table S1 for bibliographic references of individual cases and motivations for each excluded paper). In the cumulative group of 259 patients (including 10 unpublished patients from our cohort (see online supplementary table S2)), most (60%) were male subjects. The median age at onset was 49 years (range, 4–87); the median age at diagnosis was 53 years (range, 4–87). The median diagnostic delay was 1 year (range, 0–34). In 19% of patients, ECD developed below the age of 40 years, in 71% between 40 and 70, and in 10% developed after 70. The frequencies of symptoms at disease onset in the whole study population and in age-stratified subgroups are summarised in table 1 and figure 1.

Table 1

Clinical presentation of a cohort of 259 histologically-proven Erdheim–Chester disease patients

Figure 1

Manifestations of Erdheim–Chester disease (ECD) at onset and over the course of the disease. The frequency of different clinical presentations of ECD was analysed in a cumulative cohort of 259 patients. (A) Depicts the cumulative frequency of the different clinical disease manifestations at the time of onset (black boxes) and over the subsequent course of the disease (white boxes). The most frequent presentations of ECD included skeletal and neurological manifestations, diabetes insipidus, and constitutional symptoms. (B) Represents the initial clinical manifestations of ECD in patients divided in different groups according to their age at disease onset (<40 years (n=49); 40–70 years (n=184); >70 years (n=26)). In patients aged less than 40 years, ECD onset was most frequently characterised by neurological and skeletal manifestations and diabetes insipidus, whereas it never featured retroperitoneal involvement. In patients aged 40–70 years, other manifestations such as constitutional symptoms, retroperitoneal, pulmonary and cutaneous involvements started to emerge. In patients who were over 70 years at disease onset, pulmonary and cardiovascular involvements also were not infrequent presentations. DI, diabetes insipidus.

We found that ECD presented more frequently with skeletal symptoms, diabetes insipidus, neurological and constitutional symptoms. There were no significant differences between presentations and subsequent manifestations of ECD between male and female patients (data not shown).

Bone involvement, either clinically silent or symptomatic, was virtually present in all patients with ECD.1 Typical radiological findings (symmetric diaphyseal osteosclerosis or symmetric uptake of long bone extremities at bone scintigraphy) are almost pathognomonic of the disease, and represent a cornerstone in the assessment of the diagnosis.6 ,8–12 Over the course of the disease, bone involvement became symptomatic in about 50% of patients.1 In half of these cases, this was the first manifestation of ECD, although it rarely prompted the disease-revealing radiological investigations.

Neurological involvement, recently indicated as an independent predictor of death in ECD patients,7 emerged from our observations as a prominent feature of ECD, occurring in 51% of the patients over the course of the disease and in 23% at disease onset.

According to our data, diabetes insipidus was the third most frequent manifestation of ECD at onset and the most overlooked and misdiagnosed presentation (median diagnostic delay, 5 years; range, 0–34). Intriguingly, although diabetes insipidus was a frequent presenting feature of ECD, patients without diabetes insipidus only rarely (3%) developed it during follow-up.

Constitutional symptoms represented the initial complaint of ECD patients in 20% of cases, whereas their frequency appeared to decline in the overall course of the disease, although they could persist in a relapsing-remitting fashion, often accompanying new manifestations of ECD.

We found that retroperitoneal and pulmonary involvement were moderately frequent presenting manifestations of ECD.13 ,14 Although retroperitoneal involvement was one of the most frequent features of ECD, occurring in more than 30% of patients in the overall course of the disease, it was not particularly represented at presentation (14%), possibly because frequently asymptomatic and thus not noticed. Also, this manifestation was quite frequently misdiagnosed as idiopathic retroperitoneal fibrosis. Pulmonary involvement, causing lung fibrosis, associates with a poor prognosis.14 Dyspnoea and cough account for the large majority of respiratory symptoms in ECD. Pulmonary involvement induced the patients to seek for medical attention only in a minority of cases (12%).

Together with pulmonary and neurological involvement, cardiovascular ECD associates with a poor prognosis.15–17 It most typically manifests as pericardial effusion,15 16 circumferential thickening of the aorta (‘coated aorta’)18 or right atrial wall infiltration,19 which can result in cardiac tamponade.20 Despite a significant overall frequency (it occurred in 22% of ECD patients over the course of the disease), cardiac involvement was uncommon at presentation, accounting for less that 7% of initial complaints.

We then stratified the patients into different subgroups, according to their age at disease onset and analysed if the reciprocal frequencies of each clinical presentation could be different in specific age groups (figure 1B). In patients aged less than 40 years, ECD onset was most frequently characterised by neurological manifestations, bone pain, diabetes insipidus and constitutional symptoms, whereas it only rarely featured pulmonary, cardiac or retroperitoneal involvement. In patients aged between 40 and 70, the most common manifestations at onset were bone pain, diabetes insipidus, neurological and constitutional symptoms, and retroperitoneal involvement. Also, pulmonary and cardiac manifestations started to emerge as relatively frequent. Similarly, patients aged over 70 are more likely to present with bone pain, neurological symptoms, diabetes insipidus and constitutional symptoms. Intriguingly, in the elderly, the incidence of cardiac manifestations reached its peak. There were no differences in disease presentation between male and female patients in the age-stratified subgroups (data not shown).

In conclusion, ECD typically manifests with bone pain, diabetes insipidus, neurological and/or constitutional symptoms, regardless of the age of presentation. However, there seem to be differences in ECD presentation and course in different age-groups. Physicians should be aware of this rare but overlooked clinical entity and consider it in the differential diagnosis of many different clinical manifestations, in particular when multi-organ involvement is present (figure 2).

Figure 2

Typical clinical, imaging and histological findings in patients with Erdheim–Chester disease (ECD). ECD typically manifests with bone pain, diabetes insipidus, neurological and/or constitutional symptoms, but virtually every organ may be involved by the disease. Bone x-ray typically document bilateral and symmetric cortical osteosclerosis of the diaphyseal and metaphyseal regions in the long bones, in particular in the lower limbs (A). 99mTc bone scan classically shows a symmetrical and abnormally increased uptake of the radiotracer in the distal ends of the long bones of the lower (and sometimes also upper) limbs (B). This finding is highly suggestive of ECD. Imaging studies may also typically disclose mediastinal histiocyte infiltration around the heart, thoracic aorta and great vessels ((C), contrast enhanced CT scan, arrowheads), and retroperitoneal infiltration, often determining hydronephrosis, accompanied by the classical findings of ‘coated aorta’ and ‘hairy kidney’ ((D), contrast enhanced CT scan, arrowheads). Considering the rarity of ECD and its protean manifestation, pathological studies on a surgical biopsy of a lesion are necessary to confirm the diagnosis. Histology shows a xanthogranulomatous infiltrate composed by foamy histiocytes accompanied by fibrosis ((E), H&E, 200×). Immunohistochemical studies reveal that the infiltrating histiocytes are positive for CD68 ((F), CD68 immunostaining, 200×), but negative for the dendritic cell markers CD1a ((G), CD1a immunostaining, 200×) and S100 ((H), S100 immunostaining, 200×).

Acknowledgments

We are indebted to Dr Claudio Doglioni for the histological images.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • Handling editor Tore K Kvien

  • Contributors LD and MGS designed the research; BG, AB and CC did the literature search, collected and analysed the data on patients; GC and LD performed statistical analysis; GC, GS and LD interpreted data; GC and LD drafted the manuscript. All authors had full access to the original data reviewed in the manuscript. LD takes responsibility for the integrity of the data and the accuracy of the data analysis. All the authors have seen and approved the final version of the manuscript.

  • Funding This work was supported in part by a research grant from the ECD Global Alliance (De Ridder, LA) and from the Italian Ministry of Health (Rome, Italy - GR-2009-1594586) to Dr Dagna.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.