Clinical Care/Education/Nutrition/Psychosocial Research ORIGINAL ARTICLE

Evaluation of Serum 1,5 Anhydroglucitol Levels as a Clinical Test to Differentiate Subtypes of

1,2 2,4 APARNA PAL, BM BCH RHIANNON KLYNE formation about prognosis and guides 1,2,3 1,2 ANDREW J. FARMER, DM JILLY P. GREW, SRN investigation of family members. Despite 1,2 1,2,5 CHRISTINA DUDLEY, RGN MARK I. MCCARTHY, MD 3 1,2 these clear advantages, individuals with MARY P. SELWOOD, MSC ANNA L. GLOYN, DPHIL 1,2 1,2 MODY are frequently misdiagnosed as BERYL A. BARROW, RGN KATHARINE R. OWEN, MD having either type 1 or or do not have confirmatory molecular test- ing performed even when MODY is OBJECTIVE — Assignment of the correct molecular diagnosis in diabetes is necessary for suspected. informed decisions regarding treatment and prognosis. Better clinical markers would facilitate Although HNF1A-MODY and MODY discrimination and prioritization for genetic testing between diabetes subtypes. Serum 1,5 an- hydroglucitol (1,5AG) levels were reported to differentiate maturity-onset diabetes of the young due to GCK mutations (GCK-MODY) due to HNF1A mutations (HNF1A-MODY) from type 2 diabetes, but this requires further have distinct phenotypes (1,3), differen- validation. We evaluated serum 1,5AG in a range of diabetes subtypes as an adjunct for defining tiating these from each other and from diabetes etiology. common forms of diabetes can be chal- lenging in clinical practice. Molecular ge- RESEARCH DESIGN AND METHODS — 1,5AG was measured in U.K. subjects with: netic testing, if positive, is definitive but is HNF1A-MODY (n ϭ 23), MODY due to mutations (GCK-MODY, n ϭ 23), type 1 ϭ ϭ currently too expensive for indiscriminate diabetes (n 29), latent autoimmune diabetes in adults (LADA, n 42), and type 2 diabetes use. Therefore, there is a need for novel (n ϭ 206). Receiver operating characteristic curve analysis was performed to assess discrimina- tive accuracy of 1,5AG for diabetes etiology. biochemical screening tools to identify and direct efficient genetic analysis in RESULTS — Mean (SD range) 1,5AG levels were: GCK-MODY 13.06 ␮g/ml (5.74–29.74), those for whom a probable monogenic di- HNF1A-MODY 4.23 ␮g/ml (2.12–8.44), 3.09 ␮g/ml (1.45–6.57), LADA 3.46 agnosis of diabetes exists. Ideally such a ␮g/ml (1.42–8.45), and type 2 diabetes 5.43 (2.12–13.23). Levels in GCK-MODY were higher test would be highly specific for a MODY Ϫ than in other groups (P Ͻ 10 4 vs. each group). HNF1A-MODY subjects showed no difference subtype and would allow differentiation in unadjusted 1,5AG levels from type 2 diabetes, type 1 diabetes, and LADA. Adjusting for A1C ϭ between type 1 and type 2 diabetes. revealed a difference between HNF1A-MODY and type 2 diabetes (P 0.001). The discrimi- A recent report suggests that mea- native accuracy of unadjusted 1,5AG levels was 0.79 for GCK-MODY versus type 2 diabetes and 0.86 for GCK-MODY versus HNF1A-MODY but was only 0.60 for HNF1A-MODY versus type surement of serum 1,5 anhydroglucitol 2 diabetes. (1,5AG) may represent such a test, at least to discriminate HNF1A-MODY from type CONCLUSIONS — In our dataset, serum 1,5AG performed well in discriminating GCK- 2 diabetes (4). 1,5AG is a metabolically MODY from other diabetes subtypes, particularly HNF1A-MODY. Measurement of 1,5AG levels inactive that reaches could inform decisions regarding MODY diagnostic testing. steady state between ingestion and uri- nary excretion with near complete renal Diabetes Care 33:252–257, 2010 reabsorption at a specific - mannose active transporter (5,6). Due to n estimated 2% of diabetes in Eu- factor 1-␣ (HNF1A) and glucokinase structural similarity, competi- rope is caused by monogenic disor- (GCK) (1). Making the correct molecular tively inhibits this reabsorption, such that A ders of the ␤-cell (maturity-onset diagnosis allows individualization of in times of significant glycosuria, 1,5AG is diabetes of the young [MODY]) (1). The treatment, for example the use of low- excreted in the urine and consequently two most common types of MODY in dose sulfonylurea as a first line in MODY serum levels fall (7). Thus, poor glycemic clinical practice are caused by mutations due to HNF1A mutations (HNF1A- control is associated with low serum in the genes encoding hepatocyte nuclear MODY) (2). It also conveys important in- 1,5AG levels (8). A low renal threshold for ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● glucose also results in a serum 1,5AG level lower than expected (9). As HNF1A From the 1Diabetes Research Laboratories, Oxford Centre for Diabetes, and Metabolism (OCDEM), University of Oxford, Oxford, U.K.; the 2Oxford National Institute of Health Research, Bio- mutations are characterized by low renal medical Research Centre, Churchill Hospital, Oxford, U.K.; the 3Department of Primary Care Medicine, glucose threshold (10) due to decreased University of Oxford, Oxford, U.K.; the 4Diabetes Trials Unit, OCDEM, University of Oxford, Oxford, expression of the high-affinity low- 5 U.K.; and the Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, U.K. capacity glucose co-transporter 2 Corresponding author: Katharine R. Owen, [email protected]. Received 8 July 2009 and accepted 13 November 2009. Published ahead of print at http://care. (SGLT2) (11), it was hypothesized that diabetesjournals.org on 23 November. DOI: 10.2337/dc09-1246. 1,5AG levels could be a biomarker for © 2010 by the American Diabetes Association. Readers may use this article as long as the work is properly HNF1A-MODY. An initial report of se- cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons. rum 1,5AG levels in Polish subjects (4) org/licenses/by-nc-nd/3.0/ for details. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby found that mean 1,5AG levels were 50% marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. lower in patients with HNF1A-MODY

252 DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 care.diabetesjournals.org Pal and Associates compared with those with type 2 diabetes Clinical details, anthropometry, and Upon first examination of the unad- matched for glycemic control. fasting blood samples were collected for justed data, we found no difference in We sought to evaluate this hypothesis all subjects (Table 1). The study was ap- mean 1,5AG levels between subjects with in a larger independent sample set with a proved by the Oxfordshire Local Research HNF1A-MODY and type 2 diabetes (P Ͼ wider range of diabetes subtypes and to Ethics Committee, and all subjects gave 0.05). There was also no difference in assess the performance of serum 1,5AG informed consent. 1,5AG levels between HNF1A-MODY levels as an adjunctive test in identifying 1,5AG was measured using an enzy- and either of the autoimmune groups subtypes of diabetes. matic colorimetric assay (GlycoMark, (P Ͼ 0.05). As previously reported (16), GlycoMark, NY) (13). Intra-assay coeffi- we found 1,5AG levels were higher in RESEARCH DESIGN AND cient of variance (CV) was 0.46%, and those with type 2 diabetes compared with METHODS — Subjects were collected inter-assay CV ranged from 1.74 to those in the autoimmune groups (type 2 in Oxford, U.K. The MODY samples com- 2.37%. GADA was measured by a radio- vs. type 1 diabetes, P ϭ 0.011, type 2 prise subjects with a confirmed mutation immunoassay using 35S-labeled full- diabetes vs. LADA, P ϭ 0.015). After con- in either HNF1A (n ϭ 23 from 12 fami- length GAD65, and results were expressed trolling for A1C, adjusted 1,5AG level was lies) or GCK (n ϭ 23 from 10 families). in WHO units per milliliter derived from a lower in HNF1A-MODY than in type 2 Median family size was 2.5 members standard curve calibrated from interna- diabetic subjects (P ϭ 0.001). Adjusting (range 1–6), and half the families com- tional reference material (National Institute for A1C did not alter the relationships be- prised only one individual. Nineteen of for Biological Standards and Control code tween 1,5AG levels in the other diabetic the HNF1A-MODY subjects had diabetes, 97/550). Samples were considered positive subgroups. Reanalysis adjusting for one had IGT, and three were normogly- if they had levels above 14 WHO units/ml shared family membership in the MODY cemic. Oral (97.5th percentile of healthy school chil- cases and excluding the non-European (OGTT) data from the time of sampling dren) (14). individuals had no effect on these find- were available for all nondiabetic Values for age of diagnosis, duration ings (data not shown). Duration of diabe- subjects. of diabetes, BMI, creatinine, A1C, fasting tes also did not have a significant effect on The remaining subjects were from the plasma glucose (FPG) and 1,5AG level 1,5AG levels, but in the type 2 diabetic Young Diabetes in Oxford (YDX) study, were not normally distributed and were subjects we observed a progressively comprising subjects diagnosed with dia- log10 transformed. Geometric mean and lower 1,5AG level (and higher A1C) with betes Յ45 years of age recruited from ei- SD range were calculated. ANOVA was escalating treatment requirement. This ther primary (n ϭ 82) or secondary (n ϭ calculated across the groups. For 1,5AG was not seen in the MODY subgroups. 198) care. Within the group are cases of levels, pairwise comparisons (using T test The three panels in Fig. 1 show scatter classical type 1 diabetes (n ϭ 29), latent with Bonferroni correction for multiple plots of unadjusted 1,5AG levels plotted autoimmune diabetes in adults (LADA, testing) were also calculated between the against A1C. These illustrate the consid- n ϭ 42), and type 2 diabetes (n ϭ 209). different diabetes subtypes. We then ex- erable overlap between HNF1A-MODY Type 1 diabetes was defined as permanent amined the effect of correcting 1,5AG lev- and common forms of diabetes at all val- treatment since diagnosis with ad- els for A1C, and, in the type 2 diabetes ues of A1C. This suggests that, in our ditional evidence of severe ␤-cell dys- and MODY subgroups, for the effects of dataset at least, 1,5AG levels will not be a function (C-peptide undetectable or treatment modality. As 1,5AG levels can very useful clinical indicator of HNF1A- homeostasis model assessment of ␤-cell be lowered in chronic renal failure (15), MODY. We investigated this question fur- function [HOMA %B] Ͻ10%), positive subjects with serum creatinine Ͼ150 ther by constructing ROC curves. For GAD antibodies (Ͼ14 World Health Or- ␮mol/l were excluded from the analysis HNF1A-MODY and type 2 diabetes, the ganization [WHO] units/ml), or both. (one subject with LADA and nine with area under the curve (AUC) was 0.60 (Fig. LADA was defined as diabetes with posi- type 2 diabetes). 2B), confirming that in our set of subjects tive GAD antibodies but no requirement Receiver operating characteristic unadjusted 1,5AG levels are poorly dis- for insulin treatment within 3 months of (ROC) curve analysis was performed to criminative between HNF1A-MODY and diagnosis. Those not requiring perma- assess the discriminative accuracy of type 2 diabetes. We then repeated the nent insulin treatment at diagnosis with 1,5AG with regard to diabetes etiology. ROC curve analysis using 1,5AG levels negative antibodies were classified as hav- The performance of 1,5AG level as a di- adjusted for A1C. This improved the AUC ing type 2 diabetes. Subjects in the type 2 agnostic discriminator was compared to 0.75. diabetic group did not meet current clin- with A1C and FPG. All statistical analysis Our most striking finding was that ical criteria for MODY diagnostic testing was performed in SPSS version 16, and subjects with GCK-MODY had a higher (12) or had been tested and were negative P Ͻ 0.05 was assumed to be significant. 1,5AG level than any of the other groups for mutations in HNF1A/HNF4A (n ϭ 9) (P Յ 0.0003 for all pairwise comparisons, or GCK (n ϭ 4). Briefly, clinical criteria RESULTS — Table 1 shows the char- both uncorrected and corrected for A1C). for HNF1A-MODY testing was young on- acteristics of the subjects and the results Figure 1 illustrates that the GCK-MODY set (Ͻ25 years) of familial non–insulin of the biochemical investigations. Geo- cases show good separation from other dependent diabetes and for GCK testing metric mean (SD range) 1,5AG levels kinds of diabetes. ROC curve analysis to was young onset of mild fasting hypergly- were: GCK-MODY 13.06 ␮g/ml (5.74– examine the discriminative accuracy of cemia (5.5–8 mmol/l). 29.74), HNF1A-MODY 4.23 ␮g/ml 1,5AG level for GCK-MODY (from type 2 Two of the LADA subjects and 31 of (2.12–8.44), type 1 diabetes 3.09 ␮g/ml diabetes) gave an AUC of 0.79 in our sub- the type 2 subjects were of non-European (1.45–6.57), LADA 3.46 ␮g/ml (1.42– jects (Fig. 2A). Similarly for GCK-MODY ethnicity (14 Asian, 11 Black, 1 Chinese, 8.45), and type 2 diabetes 5.43 ␮g/ml versus HNF1A diabetic subjects, the AUC and 7 mixed or other). (2.12–13.23). was 0.86 (Fig. 2C). Both of these estimates care.diabetesjournals.org DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 253 Serum 1,5AG levels in subtypes of diabetes

were improved by adjusting for A1C to values of 0.94 and 0.96, respectively. 0.08 0.003 0.07 0.09 We calculated threshold values of Ͻ 0.001 Ͻ 0.001 Ͻ 0.001 Ͻ 0.001 (ANOVA) 1,5AG designed to reflect maximum sen- P sitivity and specificity in our dataset; for GCK-MODY versus type 2 diabetes, a 1,5AG Ͼ11 ␮g/ml gave a sensitivity and specificity of 75% for identifying the ␮ g/ml for male patients and HNF1A 0.70 GCK-MODY cases, while 70% of the type 4 Ϯ 25 NA 2 diabetes cases fall below this cut off. For GCK-MODY versus HNF1A-MODY, Ͼ ␮

mutation carriers 1,5AG 7.5 g/ml gave sensitivity of Nondiabetic 86% and specificity of 84% for identifying GCK-MODY, while 89% of HNF1A- MODY diabetic subjects had 1,5AG levels below this.

3.53 19.83 In clinical practice, A1C Ͻ8% and Ϯ

24 46 FPG 5.5–8.5 mmol/l are often used as biochemical discriminators to identify

GCK-MODY those individuals with apparent mono- genic diabetes most likely to have GCK- MODY (12). In our dataset, these criteria did not perform as well as 1,5AG. Al- though they gave similar sensitivity to the

2.50 15.84 above for identifying GCK-MODY cases, Ϯ 19 38 specificities were much lower at 42, 58,

MODY and 66% for A1C, FPG, and the combined criteria, respectively. Diabetic HNF1A- We examined how serum 1,5AG could perform as a preselection for cases for GCK diagnostic sequencing. This was done by estimating the proportion of pos-

3.28 5.25 itive tests for GCK-MODY that would be 61 Ϯ

209 identified by setting different prevalences of GCK-MODY in the baseline sample set

Type 2 diabetes using the threshold levels described above. When the type 2 diabetes and GCK-MODY groups from this study are combined, 25% of cases in this combined group with a 1,5AG value Ͼ11 ␮g/ml value refers to ANOVA across the diabetic groups. Manufacturer’s quoted reference range for 1,5AG: 10.7–32.0 2.38 7.58

P have a GCK mutation. This would be a Ϯ 42 69 similar rate of positive test pick-up to that achieved using standard clinical criteria by the U.K. diagnostic testing center (17). However, the prevalence of GCK-MODY in this combined group is 10%, which is

SD adjusted for A1C. somewhat higher than we would expect, Ϯ

1.73 4.99 even in a group referred for genetic inves- Ϯ

29 55 tigation. Recalculating based on a more realistic prevalence of GCK-MODY repre-

4.09 senting 5% of the diabetes cases would 7.9 (7.1 Ϫ 8.8) 7.9Ϫ 9.1) (6.9 7.8 (6.5 Ϫ 9.3) 7.2 (6.2 Ϫ 8.3) 6.8 (5.6 Ϫ 8.3) 5.2 (4.9 Ϫ 5.6) Type 1 diabetes LADA 3.09 (1.45 Ϫ 6.57) 3.46 (1.42 Ϫ 8.45) 5.43 (2.12 Ϫ 13.23) 4.23 (2.12 Ϫ 8.44) 13.06 (5.74 Ϫ 29.74) 18.03 (10.48 Ϫ 31.0) 26.9 (23.1 Ϫ 31.4) 27.2 (22.4 Ϫ 33) 33 (26.9 Ϫ 40.4) 25.2 Ϫ 30.3) (21.1 27.7 (22 Ϫ 34.8) 22.8 (20.2 Ϫ 25.8) 0.26 halve the positive test rate to 12.5%, which would probably still be an accept- able pick-up rate given the changes in management and prognosis that result from rediagnosing type 1 or type 2 diabe- tes as a GCK mutation. This is also com-

) parable with many other diagnostic 2

␮ g/ml for female patients. *Age at sampling. genetic tests in the U.K. Other models that add fasting glucose and A1C thresh- olds to the calculation do not improve Adjusted 1,5AG 1,5AG ( ␮ g/ml) Fasting glucose (mmol/l) 9.5 (5.3 Ϫ 17.0) 8.4 (5.6 Ϫ 12.7) 8.2Ϫ 11.9) (5.7 8.3 (5.9 Ϫ 11.5) 7.7 (5.8 Ϫ 10.3) 5.3 (4.3 Ϫ 6.4) A1C (%) BMI (kg/m Serum creatinine ( ␮ mol/l) 94 (83 Ϫ 106) 92 (76 Ϫ 110) 96 (73 Ϫ 125) 84 (75 Ϫ 95) 87 (75 Ϫ 101) 82 (78 Ϫ 86) Diabetes duration (years) 17.8 (9.4 Ϫ 33.6) 10.3 (3.2 Ϫ 33.2) 11.7 (4.5 Ϫ 30.3) 15.1 (6.1 Ϫ 37.2) 12.3 (6.5 Ϫ 23.5) Diagnosis age (years) 16.7 (7.2 Ϫ 38.5) 32.9 Ϫ 43.3) (25 36.2 (29.1 Ϫ 45.1) 21.5 (13.5 Ϫ 29.1) 21.5 (11.6 Ϫ 40.0) 30.0*Ϫ 46.6) (19.3 n % Male Table 1— Characteristics of the subjects studied Data are geometric means (SD range) and means 6.8–29.3 these rates in our dataset.

254 DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 care.diabetesjournals.org Pal and Associates

Figure 1—Scatter plots of serum 1,5AG levels (␮g/ml) versus A1C (%) for the different subtypes of diabetes. For clarity the data points are plotted on three panels with a filled symbol to emphasize a different diabetic subtype in each panel: lilac diamonds, subjects with autoimmune diabetes (type 1 ϩ LADA combined); blue circles, type 2 diabetes; orange squares, HNF1A-MODY; and green squares, GCK-MODY. A: Distribution of autoimmune diabetes. B: Type 2 diabetes. C: Both MODY subtypes. Three subjects with A1C Ͼ12.5% are not shown for increased clarity of the figure but were included in the analysis.

CONCLUSIONS — Identifying mo- for A1C. This is necessary because the higher A1C, as we have limited data on nogenic forms of diabetes and assigning lower A1C in our HNF1A-MODY group A1C Ͼ9%. A further limitation for a role the subtypes correctly currently depends (7.2 vs. 7.8%, P ϭ 0.04) has the effect of of 1,5AG in a diagnostic strategy to detect on recognizing a clinical phenotype and diminishing the difference in unadjusted HNF1A-MODY is that we found no dif- arranging confirmatory molecular testing. 1,5AG levels (the previous study was well ference in 1,5AG levels between HNF1A- Additional biochemical tests that aid pri- matched at baseline for A1C). Adjusting MODY and either form of autoimmune oritization of cases for genetic testing for A1C increased the discriminative ac- diabetes. would have great clinical utility. curacy of 1,5AG to identify HNF1A- The most striking finding in our study 1,5AG is an attractive candidate MODY from type 2 diabetes with the AUC was the higher 1,5AG levels in subjects marker for HNF1A-MODY because it uti- of the ROC curve rising from 0.60 to 0.75. with GCK-MODY compared with all lizes the known characteristic of low renal This is still of rather limited clinical utility other groups. This is likely to be ex- threshold for glucose seen in HNF1A mu- (AUC of Ն0.8 representing a useful test) plained by the known modest postchal- tations. We confirm the previous finding and would require further validation to lenge glucose increment seen in those (4) that 1,5AG levels are lower in HNF1A- design a suitable model that includes with GCK mutations (18). Postprandial MODY than type 2 diabetes, but this dif- A1C. Ideally this validation would in- glucose levels rarely rise high enough to ference is only apparent after adjustment clude more HNF1A-MODY cases with cause glycosuria, resulting in levels of

Figure 2—ROC curves illustrating discriminative capacity of unadjusted 1,5AG to distinguish between diabetes subgroups. A: GCK-MODY and type 2 diabetes. B: HNF1A-MODY and type 2 diabetes. C: GCK-MODY and HNF1A-MODY. care.diabetesjournals.org DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 255 Serum 1,5AG levels in subtypes of diabetes

1,5AG that are mainly within the normal consuming, difficult to interpret in those References range. Our ROC curve analyses result in on treatment, and has large day-to-day 1. Owen K, Hattersley AT. Maturity-onset promising estimates for the ability of variability (21). A 1,5AG measurement diabetes of the young: from clinical de- 1,5AG to discriminate GCK-MODY from would in theory be a useful, more cost- scription to molecular genetic character- both type 2 diabetes and HNF1A-MODY. effective, and practical alternative to an ization. Best Pract Res Clin Endocrinol Though 1,5AG levels are likely to be OGTT, as it reflects postprandial glucose Metab 2001;15:309–323 influenced by recent dietary intake, we excursion from a single nonfasting blood 2. Pearson ER, Starkey BJ, Powell RJ, Gribble made no attempt to adjust for this. The sample. OGTT data were not available on FM, Clark PM, Hattersley AT. Genetic cause of hyperglycaemia and response to ROC curve results are reassuring that this our GCK-MODY cases, so we were not test will perform well in routine clinical treatment in diabetes. Lancet 2003;362: able to directly assess the correlation with 1275–1281 practice where prior dietary information 2-h glucose. However, a relationship be- is not likely to be available. 3. Murphy R, Ellard S, Hattersley AT. Clini- tween 1,5AG levels and OGTT has been cal implications of a molecular genetic There are some limitations to extrap- examined previously; in subjects with classification of monogenic beta-cell dia- olating this finding to a general type 2 IGT 1,5AG levels were strongly correlated betes. Nat Clin Pract Endocrinol Metab diabetes clinic population: The subjects (r ϭϪ0.8) with 2-h glucose levels, and 2008;4:200–213 in this study were selected for age of di- this was greater than the correlation seen 4. Skupien J, Gorczynska-Kosiorz S, Klupa agnosis Յ45 years, where the pretest T, Wanic K, Button EA, Sieradzki J, Ma- probability of possessing a GCK mutation with FPG and the correlation between lecki MT. Clinical application of 1,5-an- is likely to be higher than in an unselected A1C and either fasting or 2-h values (22). hydroglucitol measurements in patients group of patients with type 2 diabetes Similarly, 1,5AG levels showed good cor- with hepatocyte nuclear factor-1alpha (GCK mutations in fasting relation with postprandial continuous maturity-onset diabetes of the young. Di- fall from a prevalence of 40% in children glucose monitoring system readings (23) abetes Care 2008;31:1496–1501 [19] to ϳ1% in adults diagnosed over 50 and 2-h postprandial capillary measure- 5. Yamanouchi T, Shinohara T, Ogata N, Ͻ ments (24). This supports the use of Tachibana Y, Akaoka I, Miyashita H. years of age [20]). In the 45 years of age Common reabsorption system of 1,5-an- 1,5AG as a surrogate for postchallenge range the likely prevalence of GCK muta- hydro-D-glucitol, fructose, and mannose tions is difficult to estimate but probably glucose and merits further validation in in rat renal tubule. Biochim Biophys Acta in the range of 3–5%. GCK-MODY. 1996;1291:89–95 We found that duration of diabetes did In conclusion, we suggest future re- 6. Yamanouchi T, Tachibana Y, Akanuma H, not have a significant effect on 1,5AG level; search should focus on the role of the Minoda S, Shinohara T, Moromizato H, however, the type 2 diabetic patients on diet 1,5AG level as a tool to differentiate Miyashita H, Akaoka I. Origin and dis- treatment had a higher 1,5AG than those on MODY subtypes in those already sus- posal of 1,5-anhydroglucitol, a major oral hyperglycemic agents or insulin. This pected of having a monogenic form of polyol in the human body. Am J Physiol 1992;263:E268–273 suggests that postprandial glucose excur- diabetes. sion is not normalized by treatment of dia- 7. Akanuma Y, Morita M, Fukuzawa N, Yamanouchi T, Akanuma H. Urinary betes. Therefore, we would predict that excretion of 1,5-anhydro-D-glucitol ac- 1,5AG levels might be less useful in dis- Acknowledgments— This study was sup- companying glucose excretion in diabetic criminating GCK-MODY from those ported by the National Institute for Health patients. Diabetologia 1988;31:831–835 with type 2 diabetes who are well- Research (NIHR) Biomedical Research Cen- 8. McGill JB, Cole TG, Nowatzke W, controlled on diet treatment. tre, Oxford; Diabetes U.K.; the EUFP6 inte- Houghton S, Ammirati EB, Gautille T, There was very little overlap between grated project MolPAGE (LSHG-512066); Sarno MJ, U.S. trial of the GlycoMark as- the 1,5AG levels in our GCK-MODY and the European Community FP7 program say. Circulating 1,5-anhydroglucitol lev- HNF1A-MODY groups. This discrimina- CEED3 (HEALTH-F2-2008-223211); and els in adult patients with diabetes reflect tive performance benefits from the fact the Oxford Hospitals Charitable Fund. This longitudinal changes of glycemia: a U.S. that two major characteristics of these study is also supported by the NIHR Thames trial of the GlycoMark assay. Diabetes subtypes of MODY (low renal threshold Valley Diabetes Local Research Network, Care 2004;27:1859–1865 9. Kilpatrick ES, Keevilt BG, Richmond KL, part of the U.K. Clinical Research Network. and low postchallenge increment) have Newland P, Addison GM. Plasma 1,5-an- A.P. is a Medical Research Council (MRC)- opposite effects on 1,5AG levels and sug- hydroglucitol concentrations are influ- funded Clinical Training Research Fellow, gests 1,5AG analysis might have the most enced by variations in the renal threshold A.J.F. is supported by the NIHR School of potential as a discriminative test between for glucose. Diabet Med 1999;16:496– these two MODY subtypes. Currently pa- Primary Care Research, A.L.G. is an MRC 499 tients suspected of having MODY are fre- New Investigator (grant ref. 81696), and 10. Menzel R, Kaisaki PJ, Rjasanowski I, Hei- quently selected for GCK rather than K.R.O. is an NIHR-funded Clinician nke P, Kerner W, Menzel S. A low renal HNF1A mutation testing on the basis of Scientist. threshold for glucose in diabetic patients No potential conflicts of interest relevant to the characteristic pattern seen on OGTTs: with a mutation in the hepatocyte nuclear this article were reported. factor-1alpha (HNF-1alpha) gene. Diabet In GCK-MODY a mild fasting hyperglyce- This study was presented in abstract form at mia with a modest 2-h increment (90th Med 1998;15:816–820 Ͻ the Diabetes UK Annual Professional Confer- 11. Pontoglio M, Prie´D, Cheret C, Doyen A, centile 4.6 mmol/l) is observed (12), ence, Glasgow, Scotland, 11–13 March 2009, Leroy C, Froguel P, Velho G, Yaniv M, while in HNF1A-MODY FPG may be nor- and at the European Association for the Study Friedlander G. HNF1alpha controls renal mal but with a high 2-h postchallenge of Diabetes 45th annual meeting, Vienna, Aus- glucose reabsorption in mouse and man. level (mean Ͼ5 mmol/l) (18). tria, 29 September–2 October 2009. EMBO Rep 2000;1:359–365 Although an OGTT is quoted as the We are grateful to all the patients and fam- 12. Ellard S, Bellanne´-Chantelot C, Hattersley gold standard investigation (12), it is time ilies involved in our research. AT, European Molecular Genetics Quality

256 DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 care.diabetesjournals.org Pal and Associates

Network (EMQN) MODY group. Best a combination test of hemoglobin A1c and 21. Selvin E, Crainiceanu CM, Brancati FL, practice guidelines for the molecular ge- 1,5-anhydroglucitol. Metabolism 1992;41: Coresh J. Short-term variability in mea- netic diagnosis of maturity-onset diabetes 862–867 sures of glycemia and implications for the of the young. Diabetologia 2008;51:546– 17. Institute of Biomedical Sciences PMS, Ex- classification of diabetes. Arch Intern Med 553 eter. Genetic types of diabetes including 2007;167:1545–1551 13. Nowatzke W, Sarno MJ, Birch NC, Stickle maturity-onset diabetes of the young 22. Yamanouchi T, Inoue T, Ogata E, Kashi- DF, Eden T, Cole TG. Evaluation of an (MODY) [website]. Available from http:// wabara A, Ogata N, Sekino N, Yoshimura assay for serum 1,5-anhydroglucitol (Gly- www.diabetesgenes.org/. Accessed 1 July T, Ichiyanagi K, Kawasaki T. Post-load coMark) and determination of reference 2009 glucose measurements in oral glucose tol- intervals on the Hitachi 917 analyzer. Clin 18. Stride A, Vaxillaire M, Tuomi T, Barbetti erance tests correlate well with 1,5-anhy- Chim Acta 2004;350:201–209 F, Njølstad PR, Hansen T, Costa A, Con- droglucitol, an indicator of overall 14. Bingley PJ, Bonifacio E, Williams AJ, get I, Pedersen O, Søvik O, Lorini R, glycaemic state, in subjects with impaired Genovese S, Bottazzo GF, Gale EA. Pre- Groop L, Froguel P, Hattersley AT. The glucose tolerance. Clin Sci (Lond) 2001; diction of IDDM in the general popula- genetic abnormality in the deter- 101:227–233 tion: strategies based on combinations of mines the response to an oral glucose 23. Dungan KM, Buse JB, Largay J, Kelly MM, autoantibody markers. Diabetes 1997;46: load. Diabetologia 2002;45:427–435 Button EA, Kato S, Wittlin S. 1,5-anhy- 1701–1710 19. Feigerlova´ E, Pruhova´ S, Dittertova´L, droglucitol and postprandial hyperglyce- 15. Shimizu H, Shouzu A, Nishikawa M, Lebl J, Pinterova´ D, Kolostova´ K, Cerna´ mia as measured by continuous glucose Omoto S, Hayakawa T, Miyake Y, M, Pedersen O, Hansen T. Aetiological monitoring system in moderately con- Yonemoto T, Inada M. Serum concentra- heterogeneity of asymptomatic hypergly- trolled patients with diabetes. Diabetes tion and renal handling of 1,5-anhydro- caemia in children and adolescents. Eur Care 2006;29:1214–1219 D-glucitol in patients with chronic renal J Pediatr 2006;165:446–452 24. Stettler C, Stahl M, Allemann S, Diem P, failure. Ann Clin Biochem 1999;36:749– 20. Gloyn AL, van de Bunt M, Stratton IM, Schmidlin K, Zwahlen M, Riesen W, 754 Lonie L, Tucker L, Ellard S, Holman RR. Keller U, Christ E. Association of 1,5-an- 16. Yamanouchi T, Moromizato H, Shinohara Prevalence of GCK mutations in individ- hydroglucitol and 2-h postprandial blood T, Minoda S, Miyashita H, Akaoka I. Esti- uals screened for fasting hyperglycaemia. glucose in type 2 diabetic patients. Diabe- mation of plasma glucose fluctuation with Diabetologia 2009;52:172–174 tes Care 2008;31:1534–1535

care.diabetesjournals.org DIABETES CARE, VOLUME 33, NUMBER 2, FEBRUARY 2010 257