Studies were included when patients were at least 18 years old and diagnosis was histologically confirmed. We screened cohort studies, case–control studies, case series, case reports and letters that clearly reported the outcome of the respective systemic treatments. As we focused on systematic therapies, papers dealing with topical treatments were excluded. In addition, some articles were removed due to duplicate information. Studies were checked for eligibility by the first author, and then results were reviewed by the last author.

A review by Miguel et al. helped to identify relevant cases from 1980 to 2014. Only patients who had received systemic therapy were included. As a second step, we searched PubMed, Medline and Web of Science databases using the queries “necrobiotic xanthogranuloma and therapy” until 2021. Following the database search, studies were compiled into a single list with all duplicates removed. Further exclusion criteria were studies with aggregated data, an unclear diagnosis, only topical treatment mentioned, no proper description of treatment, or response to treatment not mentioned.

The primary outcome was the reported response to systemic treatment in the papers. These were classified as “complete response”, “partial response”, “stable disease” or “progressive disease”. The response to therapy was evaluated by reviewing each patient’s medical record (as reported). Complete response to treatment was used for the absence of all detectable NXG lesions and stable hematological symptoms. Partial response was defined as a decrease in the size or number of NXG lesions and an improvement of the hematological symptoms. Stable disease was defined as no change in the size or number of the NXG lesions and stable hematological symptoms. Progressive disease was defined as an increase in the size or number of the NXG lesions or worsening of the hematological condition. In mixed response scenarios (reduction in size or regression of individual lesions with simultaneous appearance of new lesions), we rated as “progressive disease”. The sole response of cutaneous lesions with simultaneous progression of the hematological condition, or vice versa, were also rated as “progressive disease”.

The review by Miguel et al. helped to identify 101 patients [1–3, 14–59]. The additional literature search yielded 45 records. After removal of duplicates, 39 papers were subject to fulltext-review. 13 records were excluded: 6 did not discuss systemic treatment of NXG, a further 2 did not report any treatment, another study provided ambiguous information on treatment, 3 studies discussed an alternative diagnosis to NXG and another study failed to mention the response to treatment. A total of 26 studies were included based on the above-mentioned criteria. These 26 articles present the therapy options and the course of therapy of 69 patients [4, 60–84]. 5 institutional patients (University Medical Center Regensburg) were included (Table 1, see Additional file 1). We were thus able to assess the outcome of systemic therapies in 79 studies and 175 patients (Fig. 1, representative institutional case in Fig. 2).Table 1

Clinical data of institutional case reports: For details, see Additional file 1

Most of the studies were case reports and some were case series and the sample size of all studies was small. Since a scale for severity of NXG does not exist, clinical response is difficult to classify. In our systematic review, clinical response is essentially based on the individual report of each author. This makes a comparative statements difficult, which is a limitation of the study. All studies were uncontrolled, leading to a high risk of confounding. Due to the lack of randomization, the risk of selection bias was high. Risk of reporting bias was high due to lack of blinding. It is difficult to comment on publication bias, however, as the main part of evidence is from case reports, the question arises whether ineffective therapies have been published in the same way as effective ones.

The most common association between NXG and hematologic disorders has been with plasma cell dyscrasias. Of the 175 patients, 95 patients had paraproteineinemia (55%). The most common subtype was IgG-kappa. 19 of 175 (11%) patients had a malignant condition: Multiple myeloma, in 12 patients (7%), was the most common type. However, Hodgkin lymphoma and chronic lymphatic leukemia (CLL) were also observed. The overall perecentage of patients with simultaneous paraproteinemia and/or a malignant condition was 65% (114 patients).

Different treatments have been used for NXG with a wide variety of responses, such corticosteroids, IVIG, lenalidomide, cyclophosphamide, chlorambucil, thalidomide, melphalan, infliximab/rituximab, cladribine, bortezomib, vincristine, interferon alpha-2a, dapsone, ibarubicin, adalimumab, etretinate, cyclosporine, mycophenolate-mofetil, clofazimine, minocycline, doxycycline, acitretin, azathioprine and combined therapies (FCR, RCVP, vincristine/melphalane/cyclophosphamide/prednisolone).

The effect of treatments administered are presented in Fig. 3. Corticosteroids were the most frequently used treatment for NXG. Corticosteroids were used in 45 cases. Complete response occured in 5 patients (11%), and partial response in 9 patients (20%), stable disease was achieved in 16 patients (36%) and progressive disease was observed in 15 patients (33%). The use of IVIG turned out to be the most effective therapy. IVIG were used in 26 patients. Complete response was achieved in 7 patients (27%) and partial response in 14 patients (54%). Two patients exhibited stable disease (8%) and three patients did not respond to the therapy (11%). Another sufficient therapy option was the use of lenalidomide in combination or without corticosteroids. Complete response was observed in 4 patients (18%) and partial response in 7 patients (32%). Six patients (27%) achieved partial response and no response was noticed in five patients (23%).Fig. 3

Efficacy of systemic therapies for necrobiotic xanthogranuloma. The numbers denote the cases with respective treatments

The overall response was improvement (complete response and partial response) in 128 patients (73%) and stable disease in 25 patients. 22 (13%) patients showed progress of disease.

Since patient data were collected from individual case reports, follow-up data were only occasionally available. The duration of response was set to be at least the timespan reported in the case reports in case patients were either lost to follow-up or no other information was given.

Of the 26 patients treated with IVIG, follow-up data were available for 8 patients. The median duration of response (2 patients with complete response and 6 patients with partial response) for the 8 patients was 12 months (range 6–48 months, mean 15.75 months). Furthermore, we wanted to illustrate the follow-up data of the second promising therapy, the use of corticosteroids. Of the 45 patients treated with corticosteroids, follow-up data were available for 10 patients. The median duration of response (4 patients with complete response, 5 patients with partial response and 1 patients with stable disease) for the 10 patients was 12 months (range 2–24 months, mean 11.9 months).

Additional file 1. Supplementary information.