Diabetes insipidus in Langerhans cell histiocytosis: When is treatment indicated?

Diabetes insipidus (DI) is the most common central nervous system (CNS)/endocrine manifestation of Langerhans cell histiocytosis (LCH). Earlier studies have shown the frequency of DI in LCH patients to be between 25% and 50%, while in more recent studies DI has been observed in only 7–20% [1]. The frequent use of chemotherapy in multisystem (MS) patients may partially explain this lower incidence. DI can occur months or years before the diagnosis of LCH, although it usually develops within 5 years after diagnosis. A water deprivation test is mandatory to confirm the diagnosis, and to discriminate between partial and complete DI. Risk factors for developing DI include MS disease and craniofacial bone lesions particularly involving the frontal bones, orbits, middle ears, and mastoids. The risk is also significantly increased when LCH is active for a prolonged period of time and with disease reactivations [1]. In a study by Pollono et al. [2], 209 LCH reactivation episodes occurred within 24 months from diagnosis, 15 of them presented as DI. More recently, an international retrospective analysis showed that DI occurred in 23 out of 124 patients with MS LCH before first reactivation, and in 11 patients at first reactivation. Seventeen patients out of 90 who at first reactivation still had no DI developed DI later (12 within the context of a subsequent disseminated reactivation and 5 as isolated DI) [3]. On histopathology, active LCH granulomas with S100þ and CD1aþ histiocytes and mixed with CD68þmacrophages have been reported in biopsies of hypothalamic/pituitary space-occupying lesions [4]. Patients with LCH and DI have a high risk of progression to anterior pituitary dysfunction, with a 54% 10-year risk of growth hormone insufficiency (GHI) [5] and a lesser, but significant risk of deficiency of the other anterior pituitary hormones. In a study by Nanduri et al. [6], GHI occurred in 21 out of 144 patients with MS LCH, at a median interval of 3.5 years after diagnosis. Thirteen of these patients had a final height with a median standard deviation (SDS) of 1.2. DI was found in 49 of the 144 patients and occurred within 5 years of diagnosis, occasionally as the presenting symptom of LCH. Furthermore, patients with DI may develop even more serious CNS LCH and neurodegenerative (ND) disease. Grois et al. [7] found that 76% of patients with DI developed radiologic ND changes when followed with brain MRI for 5 years. In this study, 46 of the 59 patients with DI had MS disease, while 13 had single system disease; 37 of 59 (75%) with DI had craniofacial bone lesions on MRIs during their disease course. At least 25% of patients with radiologic ND will develop clinical neurodegeneration, associated with potentially devastating long-term neurocognitive deficits [8]. While anterior pituitary and CNS LCH may occasionally be seen in patients without DI, the majority of these patients develop DI as the first manifestation of endocrine/CNS disease. The question, therefore, arises whether prevention of DI can prevent the development of these permanent consequences. The present state of knowledge does not allow for a definitive answer; however, there is at least suggestive evidence that this may be possible. For example, retrospective evidence suggests that prolonged systemic therapy can reduce LCH reactivations; since multiple reactivations increase the risk of DI, it follows that prevention of LCH reactivations may decrease the incidence of DI. In the DAL-HX studies which utilized five drugs and 12 months of chemotherapy for patients with MS LCH, the reactivation rate was 27% compared to a reactivation rate of 53% and 61%, respectively, with 6 months of therapy given on the less intensive LCH-I and LCH-II protocols [9]. In addition, more prolonged and intensive therapy appeared to reduce the incidence of DI to 19% in patients with MS LCH [10], as opposed to 36% rate when a more conservative approach was adopted [11]. However, this is a retrospective comparison of trials done at different time points and the results may suffer from the biases associated with any retrospective comparison. The most recent LCH-III protocol was, in part, designed to compare in a prospective randomized fashion whether 6 or 12 months of low dose chemotherapy are most effective in reducing reactivations in patients with MS LCH. The results are currently being analyzed. Further long-term follow-up will be necessary to evaluate the incidence of DI and adverse late effects in the two arms. Recommendations for management of new onset established DI include desmopressin (DDAVP) to control DI symptoms; in addition, regular neurological evaluation and whole brain MRI scans for early detection of active CNS-LCH and neurodegeneration. Patients with DI and/or hypothalamo-pituitary lesions on MRI should have early assessment of anterior pituitary function, as early replacement with GH may improve final height [6]. Apart from this, the optimal management of patients with LCH presenting with isolated DI remains controversial. The histology of hypothalamic/pituitary space-occupying lesions in LCH patients with DI suggests that DI is a manifestation of active LCH [4]. Established DI is usually not reversible [1], but a number of cases have been reported in which the DI remitted, spontaneously [12], with treatment such as 2-chlorodeoxyadenosine (2-CDA) [13] and etoposide [14], or after hypothalamic–pituitary radiation therapy (HPRT) [15,16]. An early retrospective study of LCH-DI patients which included 30 who received HPRT and 17 who did not, reported a clinical response (5 complete and 1 partial) in 6/28 (21%) of radiated patients; 5/19 (26%) of those who completed HPRT in