Subgroups of adult-onset diabetes: a prospective follow-up study of progression of insulin resistance and deficiency and association with liver steatosis and fibrosis

OBJECTIVETo assess changes in insulin secretion and sensitivity in subgroups of adult-onset diabetes (derived from cluster analysis of GADA-positivity, BMI, age, HbA1c, HOMA2-B and HOMA2-IR), the stability of the subgroup assignment, and the association between steatotic liver disease (SLD), liver fibrosis (LF) and tissue insulin resistance. RESEARCH DESIGN AND METHODSParticipants registered in a regional diabetes register within three years from diagnosis, were restudied 2-10 years later with fasting laboratory tests and a questionnaire (n=547). A subset (n=194) participated in a clinical investigation including estimation of insulin secretion and sensitivity with i.v. glucagon-insulin tolerance test, and SLD/LF with elastography. RESULTSDifferences between subgroups attenuated during the follow-up. Obesity and age-related subgroups (MOD & MARD) were most stable with >76% assigned to the same cluster at registration and follow-up, whereas that hold for <25% in the groups characterized by insulin deficiency or resistance (SIDD & SIRD). Despite stable BMI, HOMA2-B and HOMA2-IR decreased significantly during follow-up but remained highest in the SIRD group. Prevalence of SLD/LF was highest in SIRD (70%/42%) and MOD (60%/41%). The differences between subgroups were mainly explained by BMI. However, LF was associated (OR [95% CI]) with insulin resistance in liver (3.90 [2.05-7.42]; p<0.001), adipose tissue (1.14 [1.06-1.22]; p<0.001]) and muscle (0.33 [0.17-0.64]; p<0.001), independently of BMI or SLD. CONCLUSIONSCluster assignment should be performed near diagnosis, as the more severe phenotypic features attenuate with time (or treatment). To assess the risk of LF in individuals with diabetes, the degree of insulin resistance should be evaluated, not only BMI. TWITTER SUMMARYDiabetes consists of heterogenous disease phenotypes, and disease stratification could improve targeting of interventions. Previously, we suggested dissecting adult-onset diabetes into five cluster-based subgroups using six easily accessible variables (GAD-antibodies, BMI, age, HbA1c, HOMA2-B and HOMA2-IR) measured at or near diagnosis. In this study, we assessed changes in insulin secretion and sensitivity, stability of subgroup assignment and association of steatotic liver disease (SLD) and liver fibrosis (LF) with tissue insulin resistance 2-10 years after diagnosis. Differences between subgroups attenuated during follow-up. In the obesity and age-related subgroups (MOD & MARD) >76% were assigned to the original cluster when classified at follow-up, whereas this was true for <25% of the groups characterized by insulin deficiency or resistance (SIDD & SIRD). HOMA2-B and HOMA2-IR decreased significantly but were still highest in the SIRD group at the follow-up visit. SLD/LF were most common in SIRD and MOD, with differences compared to other subgroups mainly explained by BMI. LF was also associated with insulin resistance in liver, adipose tissue and muscle, independently of BMI or SLD. Cluster assignment performed near diagnosis is thus preferred, as phenotypic features attenuate with time. Insulin resistance should be evaluated when assessing the risk of LF in individuals with diabetes. ARTICLE HIGHLIGHTSa) We investigated cluster-stability, liver health and changes in insulin resistance and insulin deficiency in cluster-based diabetes subgroups b) Are there differences between subgroups? Does insulin resistance associate with liver steatosis and fibrosis in subgroups of diabetes? c) We observed a major shift towards the more common obesity and age-related clusters after the follow-up. Liver fibrosis was mainly associated with insulin-resistant and obesity-related subgroups and, independently of BMI or liver steatosis, with insulin resistance of different tissues. d) The phenotypic features of the subgroups attenuate with time and treatment, indicating the need for clustering to be performed at diagnosis. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=109 SRC="FIGDIR/small/25340818v1_ufig1.gif" ALT="Figure 1"> View larger version (35K): org.highwire.dtl.DTLVardef@1e608c2org.highwire.dtl.DTLVardef@1491e75org.highwire.dtl.DTLVardef@16ea258org.highwire.dtl.DTLVardef@6455ae_HPS_FORMAT_FIGEXP M_FIG C_FIG