The aim of this study was to perform a systematic review and meta-analysis on insulin resistance in adult patients with type 1 diabetes mellitus compared to healthy controls, assessed by hyperinsulinemic euglycemic clamp studies.
Design and methods
We conducted a systematic search of publications using PubMed, EMBASE, Web of Science and COCHRANE Library. Hyperinsulinemic euglycemic clamp studies comparing adult patients with type 1 diabetes mellitus to healthy controls were eligible. Primary outcome measures were pooled mean differences of insulin sensitivity of endogenous glucose production (EGP), of glucose uptake and of lipolysis. We estimated mean (standardized) differences and 95% CIs using random effects meta-analysis.
We included 38 publications in this meta-analysis. The weighed mean differences in EGP during hyperinsulinemia between patients and controls was 0.88 (95% CI: 0.47, 1.29) in the basal state and 0.52 (95% CI: 0.09, 0.95) in insulin stimulated conditions, indicating decreased hepatic insulin sensitivity in patients. Insulin sensitivity of glucose uptake was either reported as M value (M), glucose infusion rate (GIR), glucose disposal rate (GDR) or metabolic clearance rate (MCR). Weighed mean differences were similar for M −3.98 (95% CI: −4.68, −3.29) and GIR −4.61 (95% CI: −5.86, −3.53). Weighed mean difference for GDR was −2.43 (95% CI: −3.03, −1.83) and −3.29 (95% CI: −5.37, −1.22) for MCR, indicating decreased peripheral insulin sensitivity in patients. Insulin mediated inhibition of lipolysis was decreased in patients, reflected by increased non-esterified fatty acid levels.
Insulin resistance is a prominent feature of patients with type 1 diabetes mellitus and involves hepatic, peripheral and adipose tissues.
Although type 1 diabetes mellitus is characterized by absolute insulin deficiency, type 1 diabetes is also associated with insulin resistance (1, 2). Insulin resistance is defined as a decreased biological response to a certain concentration of insulin. Impaired insulin action has been described in both poorly and adequately controlled patients with type 1 diabetes mellitus (3, 4), although the degree of insulin resistance varies substantially between patients (2, 5). In patients with type 1 diabetes mellitus, insulin resistance is an important risk factor for development of micro-and macrovascular complications (5, 6). Cardiovascular disease is still the leading cause of mortality in these patients (7). Therefore, it is important to understand the relation between insulin resistance and type 1 diabetes mellitus.
The hyperinsulinemic euglycemic clamp method, first described by DeFronzo et al. (8), is considered the reference standard for measurement of basal and insulin-stimulated insulin sensitivity under a variety of circumstances. Using isotope tracers, tissue specific insulin sensitivity can be quantified (9). Briefly, insulin is infused intravenously at a constant rate in a subject after an overnight fast. Glucose is ’clamped’ at a predetermined level (euglycemic clamp glucose ∼5 mmol/l or isoglycemic at basal glucose levels) by titrating a variable glucose infusion rate (GIR). As a consequence of hyperinsulinemia, glucose uptake in skeletal muscle and adipose tissue is increased, whereas lipolysis and EGP are suppressed. Under steady state conditions, the rate of glucose infusion equals the rate of glucose uptake by peripheral tissues. The degree of insulin resistance is inversely related to the amount of glucose necessary to maintain euglycemia. EGP can be calculated using isotope tracers of glucose and insulin resistance is reflected by increased EGP during a certain degree of hyperinsulinemia. Increased lipolysis, as a consequence of insulin resistance, results in increased levels of non-esterified fatty acids (NEFA). In summary, insulin resistance is characterized by decreased glucose uptake and increased EGP and NEFA levels.
Our aim was to perform a systematic review and meta-analysis on insulin resistance in adult patients with type 1 diabetes mellitus compared to healthy controls, assessed by hyperinsulinemic clamp studies.
Data sources and searches
In collaboration with a trained librarian, a search string was composed. The following databases were searched from their inception to December 1, 2013: PubMed, EMBASE, Web of Science, and COCHRANE Library. The search strategy consisted of the combination of three concepts:
Glucose clamp technique.
Type 1 diabetes.
For these three items, relevant keyword variations were used, which included keyword variations in the controlled vocabularies of the various databases, and also free text word variations. The search strategy was optimized for all consulted databases, taking into account differences of controlled vocabularies as well as differences of database-specific technical variations. The reference lists of all potentially relevant articles were screened for additional publications. The search was restricted to the English language (see Supplementary Table 1, see section on supplementary data given at the end of this article for details of the literature search).
Experimental studies using the hyperinsulinemic, iso- or euglycemic clamp technique comparing adult patients with type 1 diabetes mellitus and healthy controls were eligible.
We excluded studies in patients with renal and/or pancreatic transplantations, since immune suppressive drugs are known to induce insulin resistance. Studies performed in children and adolescents were beyond the scope of this review.
Data extraction and risk of bias assessment
Data extraction was performed independently by two investigators (ED and JAR) using a standard data-extraction sheet. In case of disagreement, consensus was reached after discussion with a third reviewer (OD).
Insulin sensitivity parameters were extracted as provided in the included studies. Information at the study level was extracted on i) patient characteristics (age, gender, mean duration of diabetes, BMI, insulin treatment, mean daily insulin use, HbA1c values), ii) clamp characteristics (including type of glucose clamp technique, use of stable isotopes), iii) type of outcome measure (EGP, glucose disposal rate (GDR, M), GIR, metabolic clearance rate (MCR) and levels of non-esterified free fatty acids (NEFA)).
In case authors compared one control group to two groups of patients with type 1 diabetes mellitus (i.e. poor vs well regulated diabetes), data of this control group were used twice in this meta-analysis. If necessary, values were recalculated to international system units, i.e. mmol/l for glucose and pmol/l for insulin.
Assessment of risk of bias was based on study components that could potentially bias the association between insulin resistance and type 1 diabetes mellitus. The following components were considered: i) adequacy of the clamp protocol. The clamp protocol was considered adequate in case predetermined fixed insulin dose and predetermined steady state glucose levels were described and sampling of blood glucose levels was performed at brief intervals throughout the experiment. Furthermore, the duration of insulin infusion and the clamp should be described. ii) Adequacy of steady state in insulin-stimulated phase of glucose clamp study. Calculations of insulin sensitivity parameters are based on the assumption that in steady state, glucose infusion equals glucose uptake. Studies were considered to have a low risk of bias in case duration of steady state, steady state insulin and glucose levels were reported and stable during steady state in both patients and controls. Studies that did not report steady state values for glucose and insulin were scored as unclear. iii) Inclusion of patients with poorly regulated diabetes, defined as an HbA1c level >9%. Chronic hyperglycemia is known to increase insulin resistance (10) and, therefore, insulin resistance could be overestimated in these patients. iv) Presence of impaired glucose tolerance in healthy controls. Studies were considered to have a low risk of bias if healthy controls were reported to have a normal glucose tolerance, tested prior to the study.
Data synthesis and analysis
The primary outcome measures in this meta-analysis were pooled mean differences of EGP, peripheral glucose uptake and NEFA levels during hyperinsulinemic euglycemic clamp conditions in patients with type 1 diabetes compared to healthy controls. Parameters for peripheral glucose uptake could be reported as GDR, glucose metabolized (M value, M), GIR or MCR. Therefore, we used standardized mean outcome measures for peripheral glucose uptake to allow comparison between studies. Values for EGP, GDR, M and GIR were reported in mg/kg per min and recalculated if necessary when reported in other units. MCR was reported in ml/kg per min and NEFA levels in μmol/l. We performed random effects meta-analysis using (standardized) mean differences and 95% CIs. In case of <5 studies a fixed effect analysis was performed, as in such a circumstance the between study variance cannot be estimated reliably. We did not check formally for funnel plot asymmetry as all our included studies were of similar sample size. Statistical analyses were performed with STATA Statistical Software (Statacorp, College Station, TX, USA) version 12.1.
The initial search included 300 publications after removal of duplicates and ultimately 64 articles were retrieved for detailed assessment (Fig. 1). Publications were excluded for the following reasons: no type 1 diabetes mellitus (n=3), age of participants was below 18 years (n=7), absence of a healthy control group (n=5), use of a clamp technique, not comparable to the iso-or euglycemic clamp method (n=3), no quantitative outcome measures reported (n=7) and one publication was excluded because participants had newly onset type 1 diabetes mellitus, in whom endogenous insulin production might be partly preserved. Finally, 38 publications were included in this meta-analysis, of which five publications (11, 12, 13, 14, 15) consisted of two and one publication (16) of three studies.
Study characteristics of included studies are presented in Table 1. In total, 38 publications were included, of which 14 publications used stable isotope tracer infusions for assessment of tissue specific insulin sensitivity (1, 3, 4, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45). Mean age of included patients ranged from 22 to 50 years and from 23 to 64 years in healthy controls. BMI ranges in patients (21–28 kg/m2) and controls (22–29 kg/m2) were similar. Most clamp studies used the hyperinsulinemic euglycemic clamp method, and only two publications performed a hyperinsulinemic isoglycemic clamp. The insulin infusion rates and steady state insulin levels differed considerably between studies, and ranged between 20 and 120 mU/m2 per min for insulin infusion rates and between 160 and 8972 pmol/l for steady state insulin levels. Both patients with short and long duration of disease were included and a large variation in diabetes regulation was present between studies.
Study characteristics of the included publications. Range of means was reported of included studies.
|Mean age (years)||22–50||23–64|
|Mean BMI (kg/m2)||21.0–28.3||21.5–28.9|
|Mean duration of diabetes (years)||1.2–38.9||NA|
|Mean daily insulin dose (U/day)||34–61||NA|
|Mean HbA1c (%)||5.6–14.2||4.8–6.6|
|Insulin infusion rates clamp (mU/m2 per min)||20–120||20–120|
|Mean glucose levels steady state (mmol/l)||4.0–9.6||4.0–5.5|
|Mean insulin levels steady state (pmol/l)||167–8972||160–7320|
Risk of bias
Risk of bias assessment is summarized in Table 2. Most publications (33 out of 38) reported an adequate clamp protocol. Steady state values and durations were reported in 32 publications. Twelve publications included patients with a mean HbA1c values >9%. Twenty-four publications did not report on impaired glucose tolerance in healthy controls. Overall, the risk of bias in the included publications was considered to be limited.
Potential sources of bias.
|References||Journal||Isotopes||Steady state||Clamp protocol||Controls without diabetes||Poorly regulated|
|(17)||Journal of Clinical Endocrinology and Metabolism||No||Yes||No||Yes||Yes|
|(18)||Journal of Clinical Endocrinology and Metabolism||Yes||Yes||Yes||NR||No|
|(21)||Journal of Clinical Endocrinology and Metabolism||No||NR||Yes||NR||No|
|(22)||Diabetes/Metabolism Research and Reviews||Yes||Yes||Yes||NR||No|
|(24)||Nephrology, Dialysis, Transplantation||No||Yes||Yes||NR||No|
|(25)||Journal of Clinical Endocrinology and Metabolism||No||Yes||Yes||Yes||No|
|(3)||Journal of Internal Medicine||No||Yes||Yes||NR||No|
|(28)||Hormone and Metabolic Research||Yes||Yes||Yes||Yes||No|
|(33)||Journal of Clinical Endocrinology and Metabolism||No||Yes||Yes||NR||No|
|(34)||Diabetes Research and Clinical Practice||No||Yes||Yes||Yes||Yes|
|(35)||American Journal of Physiology||No||Yes||Yes||Yes||No|
|(13)||Journal of Clinical Endocrinology and Metabolism||No||Yes||Yes||NR||No|
|(37)||American Journal of Physiology||Yes||Yes||Yes||Yes||No|
|(42)||Journal of Clinical Investigation||No||Yes||Yes||Yes||No|
|(43)||Journal of Clinical Endocrinology and Metabolism||Yes||Yes||Yes||NR||No|
|(15)||Journal of Clinical Endocrinology and Metabolism||Yes||Yes||Yes||NR||Yes|
|(15)||Journal of Clinical Endocrinology and Metabolism||No||Yes||Yes||Yes||Yes|
NR, not reported.
Endogenous glucose production
From studies using stable isotopes, ten studies reported data obtained during hyperinsulinemia on basal EGP and five studies on clamp EGP (Figures 2 and 3). In the basal state, mean EGP in diabetes patients varied from 2.0 to 4.5 mg/kg per min vs 1.6 to 3.0 mg/kg per min in controls. The weighed mean difference in basal EGP between patients and controls was 0.88 (95% CI: 0.47, 1.29), indicating less suppression of EGP in diabetic patients compared to controls. Under steady state hyperinsulinemic conditions, mean reported EGP values ranged from 0.0 to 1.2 mg/kg per min in diabetes patients and from 0.0 to 0.6 mg/kg per min in controls, with a weighed mean difference of 0.52 (95% CI: 0.09, 0.95), compatible with less suppression of EGP under insulin-stimulated conditions. This means that EGP during hyperinsulinemia was higher in patients compared to controls, indicating decreased hepatic insulin sensitivity in type 1 diabetes mellitus.
Peripheral glucose uptake
M value, GIR, GDR and MCR during hyperinsulinemic clamp conditions are parameters of insulin-mediated glucose uptake (Fig. 4). Most clamp studies reported M values (n=14) or GIR (n=15). Mean M values ranged from 3.9 to 9.9 mg/kg per min for diabetic patients and from 7.0 to 14.0 for healthy controls. Mean GIR in patients varied between 2.4 and 8.9 mg/kg per min and in controls between 4.2 and 15.6 mg/kg per min. Weighed mean difference for M was −3.98 (95% CI: −4.68, −3.29), in accordance with the weighed mean difference in GIR (−4.61 (95% CI: −5.86, −3.53)). The weighed mean difference for GDR was −2.43 (95% CI: −3.03, −1.83) and weighed mean difference for MCR was −3.53 (95% CI: −4.06, −2.99). All values indicate decreased peripheral glucose uptake under insulin stimulated conditions, indicating decreased peripheral insulin sensitivity in patients with type 1 diabetes compared to healthy controls.
Lipolysis and levels of NEFA
Eleven studies reported data on basal NEFA levels and nine studies reported NEFA levels during clamp conditions (Figures 5 and 6). In the basal state, mean NEFA levels ranged from 473 to 1220 μmol/l in patients and 370 to 860 μmol/l in healthy controls. NEFA levels during clamp conditions ranged from 20 to 380 μmol/l in patients and 10 to 390 μmol/l in controls. Weighed mean difference for basal NEFA levels between patients and controls was 134 (95% CI: 7, 261) and for clamp NEFA levels 24 (95% CI: 1, 47). This reflects decreased inhibition of lipolysis in patients with type 1 diabetes mellitus, and the values obtained during standardized hyperinsulinemic conditions indicate decreased insulin sensitivity of lipolysis.
In this meta-analysis, we compared insulin resistance between patients with type 1 diabetes mellitus and healthy controls using data from hyperinsulinemic euglycemic clamp studies. The present meta-analysis shows that insulin resistance is a prominent feature of patients with type 1 diabetes mellitus which involves hepatic, peripheral and adipose tissues. This association is consistent and present in both poorly and well regulated patients.
Major strengths of this meta-analysis are the relatively large number of included publications and the large number of included participants, 522 patients and 698 controls. A limitation of this meta-analysis is the potential for confounding. For instance, in patients with type 1 diabetes recent hyperglycemia or hypoglycemia preceding the clamp study could have negatively influenced clamp outcome measures on insulin sensitivity (46, 47). Furthermore, the use of certain drugs is associated with decreased insulin sensitivity. For example, β blockers and calcium channel antagonist are known to alter autonomic nervous system activity and decrease insulin sensitivity (48). In this meta-analysis, some publications excluded subjects who were using these drugs, whereas other publications did not report on the use of medication. Nonetheless, although we cannot completely exclude confounding, the findings from the included clamp studies are all very similar and consistently indicate decreased insulin sensitivity in multiple tissues in patients with type 1 diabetes mellitus. Clamp protocols differed substantially between included publications, although all studies refer to the same original study design by DeFronzo et al. (8). For instance, the insulin infusion rates and duration of the clamp procedures varied considerably, which resulted in a wide range of steady state insulin levels in the included clamp studies. Therefore, differences in insulin sensitivities found between clamp studies can be partly explained by different steady state levels of insulin. To allow comparison between studies, we only included publications directly comparing patients with type 1 diabetes mellitus with healthy controls.
In both clinical practice and clinical research, little attention is paid to insulin resistance as a common feature of type 1 diabetes mellitus. In contrast to studies in patients with type 2 diabetes mellitus, data on the effect of lifestyle factors on insulin sensitivity in type 1 diabetes mellitus are sparse (49). Although previous studies have assessed the relation between insulin resistance and development of cardiovascular complications, only a few studies are available on improving insulin resistance in type 1 diabetes mellitus. For example, Rosenfalck et al. (39) reported that a low fat diet during 3 months improved peripheral insulin sensitivity by 30% in patients with type 1 diabetes. A twelve week exercise program in adolescents with type 1 diabetes mellitus improved insulin sensitivity by 23%, although this was not reflected in HbA1c levels (50). However, these studies looked at short-term outcome measures and were not designed to investigate the relation between improvement of insulin sensitivity in type 1 diabetes mellitus and prevalence of cardiovascular complications. Conversely, insulin resistance can be readily increased in patients with type 1 diabetes by only a single night of partial sleep deprivation (51). These intervention studies indicate that insulin resistance is not a fixed pathophysiological condition in patients with type 1 diabetes.
The pathophysiological basis of insulin resistance in multiple tissues in patients with type 1 diabetes is most likely to be complex. Different concepts have been introduced to explain the presence of insulin resistance in subgroups of patients with type 1 diabetes. The term ‘double diabetes’ was introduced for patients with type 1 diabetes, with a family history of type 2 diabetes who have a higher risk to develop components of type 2 diabetes at some age (52). This notion of ‘double diabetes’ should be differentiated from the accelerator concept, which proposes that autoimmune diabetes is triggered by factors like BMI and insulin resistance. However, neither of these two concepts fully explains the observations of the current meta-analysis, since the included studies were not merely obtained in subgroups of patients with type 1 diabetes.
Other factors that may be involved are the nonpulsatile pharmacokinetic profiles of exogenous insulin compared to normal endogenous insulin secretion and the posthepatic vs prehepatic route of insulin delivery in patients vs healthy subjects. Intermittent insulin exposure in patients with type 1 diabetes mellitus resulted in a 40% reduction in insulin delivery, compared to continuous insulin exposure (53). Chronic peripheral hyperinsulinemia could alter expression and activity of the insulin receptor and insulin signaling pathways in peripheral tissues (54). Hyperglycemia also contributes to the observed insulin resistance in patients with type 1 diabetes mellitus and intensive insulin treatment has shown to improve insulin sensitivity (11, 15).
Insulin resistance seems to reflect a general phenomenon in patients with type 1 diabetes mellitus. Nonetheless, it is possible that certain conditions in subgroups with type 1 diabetes including genetic factors and life style may exacerbate insulin resistance in patients with type 1 diabetes.
In conclusion, type 1 diabetes mellitus is characterized by both insulin deficiency and insulin resistance in multiple metabolic pathways. Therefore, studies are warranted to assess if interventions aimed at reducing insulin resistance provide clinical benefits in patients with type 1 diabetes mellitus.
This is linked to the online version of the paper at http://dx.doi.org/10.1530/EJE-14-0911.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Author contribution statement
E Donga and J A Romijn conceived this study, O M Dekkers and E Donga analyzed the data, and E Donga, O M Dekkers, Eleonora P M Corssmit and J A Romijn contributed to the collection of data. All authors critically reviewed various drafts of the manuscript, and all authors approved the final version. E Donga and J A Romijn are responsible for the integrity of the work as a whole.
KacerovskyMBrehmAChmelikMSchmidAISzendroediJKacerovsky-BieleszGNowotnyPLettnerAWolztMJonesJG. Impaired insulin stimulation of muscular ATP production in patients with type 1 diabetes. Journal of Internal Medicine2011269189–199. (doi:10.1111/j.1365-2796.2010.02298.x).
OrchardTJOlsonJCErbeyJRWilliamsKForrestKYSmithlineKLEllisDBeckerDJ. Insulin resistance-related factors, but not glycemia, predict coronary artery disease in type 1 diabetes: 10-year follow-up data from the Pittsburgh Epidemiology of Diabetes Complications Study. Diabetes Care2003261374–1379. (doi:10.2337/diacare.26.5.1374).
BravardALefaiEMeugnierEPesentiSDisseEVouillarmetJPerettiNRabasa-LhoretRLavilleMVidalH. FTO is increased in muscle during type 2 diabetes, and its overexpression in myotubes alters insulin signaling, enhances lipogenesis and ROS production, and induces mitochondrial dysfunction. Diabetes201160258–268. (doi:10.2337/db10-0281).
Brown TL Sippl RM & Snell-Bergeon JK. 2012 Healthy pre-monopausal women with type 1 diabetes have increased peripheral insulin resistance compared to non diabetic women. Diabetes Conference: 72nd Scientific Sessions of the American Diabetes Association Philadelphia: June 2012
Del PratoSNosadiniRTiengoATessariPAvogaroATrevisanRValerioAMuggeoMCobelliCToffoloG. Insulin-mediated glucose disposal in type I diabetes: evidence for insulin resistance. Journal of Clinical Endocrinology and Metabolism198357904–910. (doi:10.1210/jcem-57-5-904).
GreenfieldJRSamarasKChisholmDJ. Insulin resistance, intra-abdominal fat, cardiovascular risk factors, and androgens in healthy young women with type 1 diabetes mellitus. Journal of Clinical Endocrinology and Metabolism2002871036–1040. (doi:10.1210/jcem.87.3.8324).
MakimattilaSVirkamakiAGroopPHCockcroftJUtriainenTFageruddJYki-JarvinenH. Chronic hyperglycemia impairs endothelial function and insulin sensitivity via different mechanisms in insulin-dependent diabetes mellitus. Circulation941996a1276–1282. (doi:10.1161/01.CIR.94.6.1276).
PerseghinGLattuadaGDannaMSereniLPMaffiPDeCFBattezzatiASecchiADelMALuziL. Insulin resistance, intramyocellular lipid content, and plasma adiponectin in patients with type 1 diabetes. American Journal of Physiology. Endocrinology and Metabolism2003285E1174–E1181. (doi:10.1152/ajpendo.00279.2003).
SchauerIESnell-BergeonJKBergmanBCMaahsDMKretowskiAEckelRHRewersM. Insulin resistance, defective insulin-mediated fatty acid suppression, and coronary artery calcification in subjects with and without type 1 diabetes: the CACTI study. Diabetes201160306–314. (doi:10.2337/db10-0328).
TessariPNosadiniRTrevisanRDe KreutzenbergSVInchiostroSDunerEBioloGMarescottiMCTiengoACrepaldiG. Defective suppression by insulin of leucine-carbon appearance and oxidation in type 1, insulin-dependent diabetes mellitus. Evidence for insulin resistance involving glucose and amino acid metabolism. Journal of Clinical Investigation1986771797–1804. (doi:10.1172/JCI112504).
TrevisanRNosadiniRAvogaroALippeGDunerEFiorettoPDeanaRTessariPTiengoAVelussiM. Type I diabetes is characterized by insulin resistance not only with regard to glucose, but also to lipid and amino acid metabolism. Journal of Clinical Endocrinology and Metabolism1986621155–1162. (doi:10.1210/jcem-62-6-1155).
Yki-JarvinenHTaskinenMRKiviluotoTHildenHHelveEKoivistoVANikkilaEA. Site of insulin resistance in type 1 diabetes: insulin-mediated glucose disposal in vivo in relation to insulin binding and action in adipocytes in vitro. Journal of Clinical Endocrinology and Metabolism591984b1183–1192. (doi:10.1210/jcem-59-6-1183).
SandovalDAGuyDLRichardsonMAErtlACDavisSN. Acute, same-day effects of antecedent exercise on counterregulatory responses to subsequent hypoglycemia in type 1 diabetes mellitus. American Journal of Physiology. Endocrinology and Metabolism2006290E1331–E1338. (doi:10.1152/ajpendo.00283.2005).