Peripheral eosinophilia in children with transient synovitis of the hip: 7-year experience from a single centre in New Zealand

Introduction

Hip pain with limping is a common presentation in childhood [1]. The most common diagnosis in young children is transient synovitis of the hip (TSH), a benign and self-limiting condition that is diagnosed once other more sinister diagnoses have been excluded. TSH is sometimes confirmed by the presence of a hip effusion on radiological imaging. Whilst an exact cause is yet to be elucidated, an antecedent viral illness or trivial trauma is commonly reported [2]. Patients often present with hip pain (which may refer to the back or ipsilateral knee), decreased range of movement and/or refusal to weight-bear [2].

Epidemiological data on the incidence of TSH are generally lacking. One study from the Netherlands revealed an incidence of 76.2 cases per 100,000 person-years [1], while a recent English study found an annual incidence of 25.1 cases per 100,000 children aged 14 years or less [3]. The peak age of incidence appears to be between 4 and 10 years, with boys being more than twice as likely to develop TSH as girls [1–3].

Eosinophils represent a subgroup of granular leukocytes, whose function primarily involves combating helminthic infections and mounting eosinophil-mediated inflammatory responses [4]. Whereas eosinopaenia may be caused by either pyogenic infections or steroid excess (endogenous or exogenous), many causes have been documented for eosinophilia [5].

Eosinophilia is generally defined as an eosinophil count exceeding 450–500 cells/µL [5]. The most common cause of marked eosinophilia worldwide is infections (almost exclusively by helminths) [5]. Other causes of eosinophilia, albeit often low-grade, include medications [6], atopic and allergic diseases, malignancies (especially haematological), rheumatological diseases and the hypereosinophilic syndromes [5].

In our clinical practice, we observed eosinophilia in children presenting with irritable hip more commonly than would otherwise be expected. However, there is a paucity of published literature on the subject. Therefore, we sought to explore this observation; the aims of this study were to assess the prevalence of eosinophilia in children with TSH, and to evaluate the clinical outcomes of this sub-group of patients.

Methods

Results

Management

Most of the patients (n = 91) had one or more imaging studies of the affected hip on admission. The most common modality was X-ray, followed by ultrasound scan and magnetic resonance imaging (see Fig. 1). The most common abnormality was small or moderate joint effusions. There was no difference in the rate of any radiological abnormality in eosinophilic (16.7 %) versus non-eosinophilic (16.7 %) children (p = 1.00). None of the hip joints was aspirated. All patients were managed conservatively with observation and as-required analgesia.

OPEN IN VIEWER

Fig. 1

Types of radiological investigations used. Some patients had more than one modality of imaging. MRI magnetic resonance imaging, USS ultrasound scan

Discussion

To the authors’ knowledge, this is the first study to explore the relationship between TSH and eosinophilia. We found a sizeable minority (15.6 %) of children with TSH to have eosinophilia. This is much higher than the reported rate of 1.0–1.5 % incidental eosinophilia found in the general population [7]. To explore whether this observed eosinophilia was not circumstantial, we compared TSH children with age- and sex-matched children who had blood samples taken as part of their acute admission investigations. TSH children had significantly higher eosinophil counts than controls (380 vs. 303 cells/μL, p = 0.049).

TSH children in our sample did not differ in their clinical presentation or outcome from TSH patients without eosinophilia. However, they appeared to have a shorter median hospital stay (16.3 vs. 21.5 h, p = 0.058). While the difference in hospital stay was not statistically significant, a correlation between peripheral eosinophilia and length of hospital stay of TSH patients is possible. Whether this correlation is clinically meaningful remains to be answered.

Whilst the cause of the observed association between TSH and eosinophilia is not clearly understood, we suggest several hypotheses. A likely explanation is that a sub-group of patients with TSH, which is thought to be a reactive condition to an unknown primary trigger, may be susceptible to exhibiting eosinophilia as a reaction to the same trigger. The pathophysiology of eosinophilia in rheumatology patients is yet to be elucidated [5], although a reactive process secondary to increased circulating cytokines (especially interleukin-5) appears feasible [4].

An alternative explanation is that the observed eosinophilia may have been iatrogenic; NSAIDs (a common choice for paediatric analgesia) are a known cause of eosinophilia [6]. This hypothesis is unlikely, however, since most blood samples would have been obtained on admission prior to any medication administration. Almost 8 % of our sample (split roughly equally between TSH children with eosinophilia 3.8 %, and children without 4.2 %) were given medications by parents prior to presenting; this was, however, almost always paracetamol, a non-NSAID analgesic. Moreover, other paediatric inpatients, who are also commonly given NSAIDs, do not exhibit this eosinophilic reaction nearly as frequently as 15 %. Finally, the possibility that eosinophilic transient synovitis accounted for the observed association had to be entertained. This also is unlikely given the exceeding rarity of this clinical entity in which synovial fluid (rather than peripheral) eosinophilia is the pathological hallmark [8].

The demographics of our sample are in keeping with those reported elsewhere [1, 3]. Most of the children in our sample were between 3 and 7 years, and males were more than twice as likely as females to present with TSH. Children with eosinophilia tended, albeit non-significantly, to have shorter hospital stay. The abnormal result (i.e. eosinophilia) may have been an impetus for the admitting doctor to prescribe more or stronger analgesia [9, 10], leading therefore to an apparent significant improvement in pain and earlier discharge. This, however, remains a conjecture. In the absence of quantitative data on TSH severity or standardised pain scores, it is difficult to ascertain whether an association exists between eosinophilia and a milder course of TSH.

A few limitations inherent in the design of our study ought to be mentioned. The results presented remain a single centre's experience evaluated in retrospect. The relationship detected between eosinophilia and TSH remains, at best, an association, since causality cannot be established with this type of study. This study, however, remains the first study to report this association. Future research should focus on larger TSH samples with longer follow-up to ascertain their disease course.