Den första studie som har jämfört två vanliga behandlingsalternativ för LADA i en studiepopulation av acceptabel storlek. Resultaten har tidigare också sammanfattats i Läkartidningen under ”Nya rön”.

Diabetes Obes Metab.
 2019 Oct;21(10):2219-2227. doi: 10.1111/dom.13797. Epub 2019 Jun 19.

Investigating optimal β-cell-preserving treatment in latent autoimmune diabetes in adults: Results from a 21-month randomized trial.

Hals IK1,2Fiskvik Fleiner H3Reimers N1Astor MC4Filipsson K5Ma Z6Grill V2Björklund A6.

Author information


Department of Endocrinology, St Olavs Hospital, Trondheim University Hospital, Trondheim, Norway.


Department of Clinical and Molecular Medicine, Faculty of Medicine and Health Sciences, Norwegian University of Science and Technology (NTNU), Trondheim, Norway.


Department of Clinical Pharmacology, St Olavs Hospital, Trondheim University Hospital, Trondheim, Norway.


Department of Medicine, Haukeland University Hospital, Bergen, Norway.


Department of Clinical Sciences, Lund University, Malmö, Sweden.


Department of Molecular Medicine and Surgery, Endocrine and Diabetes Unit, Karolinska University Hospital, Stockholm, Sweden.



To compare outcomes of glucagon-stimulated C-peptide tests (GSCTs) in people with latent autoimmune diabetes in adults (LADA) after a 21-month intervention with either insulin or the dipeptidyl peptidase-4 inhibitor sitagliptin.


We included 64 glutamic acid decarboxylase (GAD) antibody-positive individuals, who were diagnosed with diabetes <3 years before the study, aged 30 to 70 years, and without clinical need for insulin treatment. We stratified participants by age and body mass index (BMI) and evaluated β-cell function by GSCT after a 48-hour temporary withdrawal of study medication.


Age at randomization (mean 53 years), BMI (mean 27 kg/m2 ) and metabolic markers were similar between treatment arms. Glycated haemoglobin concentrations during intervention did not differ between arms. Fasting C-peptide concentrations after the intervention were similar, as were stimulated C-peptide levels (0.82 ± 0.63 nmol/L after insulin, 0.82 ± 0.46 nmol/L after sitagliptin; nonsignificant). Autoimmunity in the study population (estimated from GAD antibody titres and positivity/no positivity for zinc transporter 8 and islet antigen 2 antibodies) affected the evolution of the GSCT results significantly, which deteriorated in participants with high but not in those with low autoimmunity. Adjustment using analysis of covariance for the degree of autoimmunity did not alter the findings of no difference between treatment arms.


β-cell function after intervention was similar in patients with insulin- and sitagliptin-treated LADA, regardless of the strength of autoimmunity. Further, participants with low levels of GAD antibodies did not experience progressive deterioration of β-cell function over a 21-month period. Taken together, these findings could be useful for clinicians' choices of treatment in people with LADA

Inskickat till red 191127 av Valdemar Grill


Läs hela artikeln som pdf utan lösenord

From the article


Latent autoimmune diabetes in adults (LADA) is usually diagnosed according to the following criteria: onset of diabetes above the age of 30 years; presence of β‐cell directed antibodies, and mostly glutamic acid decarboxylase (GAD) antibodies; and no clinical need for insulin during the first 6 months after the diagnosis of diabetes. Using these criteria, LADA is a common form of diabetes, at least in populations in Europe.1 People with LADA make up one‐tenth of the total population with diabetes in many countries, and the condition may be more common than insulin‐requiring type 1 diabetes2; treatment of LADA is therefore an important clinical issue.

Deficiency of β cells progresses faster in many people with LADA than in those with type 2 diabetes, presumably because of ongoing autoimmune assault in LADA.3 The deficiency leads to insulin dependence, which occurs on average earlier in LADA than in type 2 diabetes.4 β‐cell deficiency to the point of insulin dependence is associated with poor metabolic control and diabetic complications,5-7 which may be worse in LADA than in type 2 diabetes.8 There is thus an urgent need for a therapy that retards β‐cell demise; however, randomized studies on treatment for LADA are scarce and provide insufficient evidence with which to decide on the optimal treatment.9, 10

In particular, there is ongoing debate regarding whether to treat LADA similarly to type 2 diabetes, with per‐oral agents or with insulin (ie, before insulin is clinically needed). Evidence favouring early insulin treatment comes to some extent from pre‐clinical data,11 but mainly from a Japanese randomized study which found that insulin treatment retarded the progression of β‐cell insufficiency as compared to antidiabetic sulphonylureas (SUs).12 That study has been criticized, however, for the use of an SU as a comparator, since SU has been shown in the long term to hasten β‐cell deficiency in type 2 diabetes13 and to exert β‐cell toxicity in vitro.14 A Cochrane review9 concluded that randomized treatment studies without SUs in LADA are needed, and especially studies rigorously designed to assess a possible benefit of early insulin treatment.9

Against this background we designed a randomized study comparing the impact of early insulin treatment with a dipeptidyl peptidase‐4 (DPP‐4) inhibitor, sitagliptin, which prolongs the effect of the endogenous incretin hormones glucagon‐like peptide‐1 and gastric inhibitory polypeptide, thereby stimulating insulin secretion. Sitagliptin was chosen as comparator because it is a currently favoured drug for the treatment of type 2 diabetes, and, in contrast to SUs, has no recognized deteriorating effects on β cells in type 2 diabetes and has even been reported to exert beneficial effects in people with LADA.15

In Norway and Sweden, LADA populations are mostly overweight or obese, with ensuing insulin resistance. We therefore designed arms of the study to be add‐ons to treatment with metformin, which is an insulin‐sensitizer. We restricted recruitment to people who had neither near‐optimal glucose control nor very poor glucose control, necessitating more intense pharmacological treatment. In this way we sought to avoid either over‐ or undertreating participants according to commonly accepted goals of treatment. After randomization we aimed for the metabolic control to be similar between treatment arms, thereby minimizing any influence of “glucotoxicity” on measures of β‐cell function.


In the present study we investigated β‐cell function in people with LADA over time in relation to two clinically relevant alternatives of treatment. Focusing on β‐cell function, the aim of the present study design was, in contrast to many other clinical trials,17 to achieve similar metabolic control in the two arms of the study. A sizeable difference in metabolic control would raise the question of whether the degree of metabolic control per se could have influenced measures of β‐cell function. The goal of comparable (and acceptable) metabolic control between arms of the study was achieved; a “glucotoxicity” effect should therefore not be relevant when interpreting the results. Further, in order to exclude the impact of ambient exposure to insulin or sitagliptin we performed tests of β‐cell function after a 48‐hour temporary withdrawal of sitagliptin and insulin.

Metformin is currently the first choice of treatment in adult‐onset, non‐insulin‐requiring and usually overweight people with diabetes. It was therefore natural to design a protocol that included add‐ons of study medications to metformin. Beneficial effects of metformin on β‐cell function have been reported in clinical studies18; however, these could probably be explained by ameliorated glucose control which would counteract “glucotoxicity.” In any case a putative β‐cell effect would probably be equal for the insulin and sitagliptin arms of the present study, therefore, the use of metformin should not confound our results.

Sitagliptin and other DPP‐4 inhibitors have become frequent add‐on treatments in non‐insulin‐dependent persons who are not optimally controlled on metformin. Clinical interest in DPP‐4 inhibitors as a treatment for LADA has arisen from reports of a possible effect on autoimmunity19 and beneficial effects of sitagliptin on β‐cell function when added to insulin in people with LADA.15 In addition, an observational study of another DPP‐4 inhibitor, saxagliptin, indicated a favourable effect on β‐cell function,19 as did other small studies.17 In clinical practice, however, LADA treatment usually entails a choice between insulin on one hand and a per‐oral antidiabetic drug on the other. Except for the Japanese study12 no direct comparison has, to our knowledge, been carried out between insulin versus a per‐oral drug in a randomized study. This lack of knowledge motivated the present study.

We chose the results of GSCTs as the primary endpoint. Such tests have been validated over many decades as reflective of β‐cell function in people with type 2 as well as type 1 diabetes.20, 21 The tests are relatively easy to perform, an advantage in a multicentre study in which the capacities for testing varied among the centres. The tests were uniformly performed at normal or only slightly hyperglycaemic glucose levels; hence, a confounding influence of hypoglycaemia or marked hyperglycaemia on our measurements can be excluded. We acknowledge that additional testing of intravenous glucose and/or responses to a test meal would have added details of interest on the time dynamics of β‐cell performance in our participants.

Both treatment arms were well accepted by the participants, and side effects were infrequent. In addition, the dropout rate was low. The weight loss during treatment with sitagliptin (on average 3.4 kg) was surprising as the effects of DPP‐4 inhibitors are reportedly weight‐neutral.22 Participants in the study were mostly overweight and were recommended appropriate dieting and exercise; this may have influenced the evolution of body weight. The reduction in body weight could have helped to uphold an acceptable level of HbA1c in the sitagliptin‐treated participants; however, we note that HOMA2‐IR, a measure of insulin resistance, was not affected by treatment in either the sitagliptin or the insulin arm.

Demise of β‐cells in people with LADA is well recognized, but the extent and the time scale of demise is variable.4, 23 This heterogeneity is well illustrated in the present study. Hence, high titres of GAD antibodies, in conjunction with the presence of ZnT8 and/or IA‐2 antibodies, were associated with a marked and significant decrease in C‐peptide values, and an increased ratio of proinsulin: C‐peptide, the latter indicating β‐cell stress.24 Importantly, adjusting for the impact of autoimmunity did not alter the findings of no difference in β‐cell function between the insulin and the sitagliptin arms. The present study also fails to suggest in other respects (ie, similar development in antibody titres) an influence of study medication on the process of autoimmunity.

The recruitment criteria excluded both those with optimal and those with markedly deranged metabolic control. This restricts the study population in terms of its representativeness of the general population with LADA, but makes the study population more clinically relevant. Epidemiological studies in people with LADA indicate similarities to the present study population; age at onset of LADA and the degree of excess body weight were similar to findings from the large population‐based health study in Nord‐Trøndelag (HUNT) in Norway (which potentially includes all adults in the Nord‐Trøndelag area).25 A gender difference in GAD antibody titres (higher in women) reported in epidemiological studies1, 26was also observed in the present study population (results not shown). The percentage of participants with LADA who were positive for more than one antibody was higher than in the HUNT epidemiological study.27 This difference could relate to time between diagnosis and measurements. Measurements in the present study population were usually performed <1 year after diagnosis, whereas in epidemiological studies, such as the HUNT prevalence studies, time from diagnosis could be decades. A successive decrease in antibodies occurs with time27 and, with regard to antibodies against IA‐2 and ZnT8, this was also observed in the present study; thus the prevalence of more than one antibody decreases with the duration of LADA.

We did not observe differences in β‐cell function between treatments during or after 21 months of intervention (a possible exception pertains to stimulated insulin data; however, as this was not accompanied by effects on C‐peptide levels, this could be secondary to changes in insulin extraction over the liver). The question arises of whether our finding of no difference was attributable to a type 2 statistical error. We think that this is unlikely because there was no borderline significance, P = .1 being closest for C‐peptide values. Another concern could be that the intervention did not register differences between treatments because endpoints were not much affected in the whole study population; however, we registered a clear decline in C‐peptide levels during the intervention in those participants who displayed strong markers of autoimmunity, thereby demonstrating that significant deterioration of β‐cell function occurred in many participants during the timeframe of the study.

No deterioration of β‐cell function was apparent in participants with low GAD antibody titres and absence of other antibodies. This finding is consistent with a recent report.28 It implies that such patients are for many years at low risk of becoming insulin‐dependent. Consequently, they could be treated similarly to those with type 2 diabetes and should not need insulin as a safety measure against a swift change to insulin dependence. Further studies are needed to establish a threshold in titres for no deterioration and to test for deterioration/no deterioration beyond the present time of observation.

Strengths of the present study include the comparatively large study population with LADA, especially when viewed in relation to the rather strict inclusion and exclusion criteria. Furthermore, the study population was well characterized and underwent a follow‐up of β‐cell‐related variables that included a 48‐hour temporary omission of study medication before the validated and standardized tests were undertaken. In addition, the dropout rate was low. Limitations of the study include the heterogeneity of autoimmune activity, which characterizes a typical LADA population. Also, we recognize that levels of GAD antibodies from identical samples may differ among laboratories29; the absolute levels reported should be viewed in that context.

In summary, we observed, to our knowledge for the first time, that β‐cell function was similarly affected in insulin‐ and sitagliptin‐treated individuals with LADA, regardless of the strength of autoimmunity. The results also imply that people with low levels of GAD antibodies and absence of other antibodies do not experience progressive deterioration of β‐cell function over a 21‐month timeframe. Taken together, the present findings could help clinicians' choices of treatment in people with LADA.

www red DiabetologNytt