Use of glycosylated ferritin assay to aid the diagnosis of adult-onset Still’s disease: a local laboratory experience in Hong Kong

Abstract

Adult-onset Still’s disease (AOSD) is a multi-systemic inXammatory disorder aVecting all ethnic groups. Even though there are several published diagnostic criteria, AOSD remains diYcult to diagnose clinically due to the overlapping features with infective, neoplastic and other rheumatological conditions [1, 2]. The problem is further complicated by the fact that treatments for AOSD, including steroid and immunosuppressive agents, can worsen the clinical course of the above conditions. Regarding laboratory markers, a markedly increased serum ferritin concentration is a common diagnostic marker among AOSD patients in active disease [3–5]. Generally speaking, serum ferritin concentration above 2,000 g/L is considered to be consistent with AOSD. Fautrel et al. [6] reported that a Wvefold increase in serum ferritin over upper limit of normal had 80% sensitivity and 41% speciWcity for AOSD. As marked elevation of serum ferritin concentration can also be seen in other disorders, the low speciWcity limits its diagnostic use. The percentage of glycosylated ferritin (GF), an isoform of ferritin, over total ferritin oVers additional diagnostic value. In healthy participants, 50–80% of ferritin is glycosylated. In inXammatory diseases, the glycosylated portion drops to 20–50%, while in AOSD, it is often ·20%. Using ·20% as the cut-oV, the sensitivity and speciWcity of percentage GF for AOSD were 80 and 66%, respectively. The combination of a percentage GF · 20% with ferritin over Wve times of normal produced a sensitivity of 43% and speciWcity of 93% [6]. We decided to develop an assay for the measurement of serum GF. Based on method described by Vignes et al. [7], GF could be easily separated from non-glycosylated ferritin by dividing the target serum sample into two aliquots, adding concanavalin A (Con-A) Sepharose 4B to one aliquot and only Sepharose 4B to the other aliquot and incubating them in a water bath for 2 h at room temperature. GF bound to Con-A, whereas non-glycosylated ferritin did not. The two aliquots were then centrifuged at 3,000 rpm for 15 min. The non-glycosylated ferritin unbound to Con-A was recovered in the supernatant and measured by immunometric assay. Accordingly, the non-glycosylated ferritin and total ferritin concentrations could be determined in the aliquots with Con-A Sepharose 4B and Sepharose 4B, respectively. The percentage GF was calculated from the diVerence between total ferritin and non-glycosylated ferritin divided by total ferritin. We had retrospectively analysed archived serum samples taken on admission from 5 newly diagnosed AOSD patients referred to our laboratory, in whom diagnosis was made clinically according to the Yamaguchi’s criteria. These 5 patients had serum ferritin concentrations of 8,047–41,479 g/L and percentage GF of 1.9–16.2%. The combined results were compatible with the diagnosis of AOSD. We had also randomly selected 15 patient samples with ferritin level over 2,000 g/L. These were patients with chronic anaemia requiring regular transfusion and patients with fulminant sepsis. Their serum ferritin concentrations ranged from 2,092 to 13,773 g/L and percentage GF varies from 28.6 to 80.6%. J. S.-S. Kwok (&) · M. H.-M. Chan Department of Chemical Pathology, The Chinese University of Hong Kong, Hong Kong SAR, China e-mail: jeffreykwok@cuhk.edu.hk