Review of the Literature
Kelly et al performed a systematic review to evaluate the association between elevated plasma insulin levels and the traits associated with metabolic syndrome.9 The metabolic traits examined included hypertension, obesity, certain cancers, polycystic ovarian syndrome (PCOS), sleep apnea, atherosclerosis, dyslipidemia, cardiovascular disease (CVD), nonalcoholic fatty liver disease, renal failure, and T2D. The researchers selected 58 articles in which insulin level was measured and evaluated them for associations between hyperinsulinism and the selected metabolic conditions. They documented 63 instances of hyperinsulinism being found in association with the abovementioned conditions and noted only 3 instances of normal insulin levels correlating with hypertension, PCOS, and certain cancers. They concluded that the consistent links between elevated insulin level and 11 disorders associated with metabolic syndrome indicated the potential pathogenic impact of elevated insulin level on a wide range of body tissues.
Pataky et al performed a cross-sectional longitudinal study involving 1211 adult participants to identify associations between fasting insulin level, insulin sensitivity, and lifestyle parameters with cardiometabolic risk factors.10 Among the cardiometabolic risk factors examined were blood glucose level, low-density lipoprotein (LDL), blood pressure, triglyceride level, and waist circumference. Follow-up was concluded at the 3-year mark, and multiple regression analysis was utilized to identify independent contributors to cardiometabolic risk. The researchers determined that adiponectin, tobacco use, BMI, and elevated fasting insulin level were all independently associated with the number of risk factors. Changes in fasting insulin and BMI were also found to positively correlate with the number of cardiometabolic risk factors present. The researchers concluded that fasting hyperinsulinemia was associated with cardiometabolic changes independent of hyperglycemia and insulin resistance.
Lunger et al evaluated methods for identifying insulin resistance in women with PCOS.11 Their retrospective review focused on the medical records of 147 women meeting the criteria for a diagnosis of PCOS and evaluated the impact particular variables had on insulin levels and insulin resistance indices. The efficacy of fasting insulin assessment in identifying insulin resistance was determined by calculating specificity, sensitivity, and the predictive values at various levels. Women identified as being insulin resistant via 3-hour OGTT were found to have significantly higher fasting insulin levels. Receiver operating characteristics (ROC) analysis found fasting insulin to be an effective measure of insulin resistance. The researchers determined that setting cut-off points for fasting insulin at <7 μU/mL and >13 μU/mL could accurately identify nearly two-thirds of the women with insulin resistance.
Welsh et al explored the relationship between early markers of insulin resistance with T2D and CVD.12 Using data previously collected in the Prospective Study of Pravastatin in the Elderly at Risk (PROSPER) trial cohort (ISRCTN Registry Number: ISRCTN40976937), they examined the connection between the homeostatic model assessment of insulin resistance (HOMA-IR), CVD, all-cause mortality, and T2D. Fasting insulin levels or HOMA-IR data were available on 4742 elderly individuals. At the 3.2-year follow-up, 283 participants had received a diagnosis of T2D. Statistical analysis of these data indicated a strong association between T2D and fasting insulin level and the HOMA-IR. Conversely, the researchers found no correlation between HOMA-IR and fasting insulin with CVD-related deaths in this population.
Ghasemi et al followed 4942 Iranian adults in an effort to determine cut-off points for measures of insulin sensitivity, insulin resistance, beta-cell function, and fasting insulin for identification of T2D risk.13 Cut-off points for predicting T2D were determined using ROC curve analysis to estimate the specificity and sensitivity of each laboratory assessment. The researchers identified the cut-off points using the Youden index and setting the sensitivity to 75% in an effort to capture high-risk individuals and avoid misclassification. They identified fasting insulin and HOMA-IR as independent predictors of T2D. The researchers elaborated on these findings by identifying optimal cut-off points (75% sensitivity) for HOMA1-IR at 1.86 for women and 1.34 for men, HOMA2-IR at 1.01 for women and 0.75 for men, and fasting insulin at 7.51 μU/mL in women and 5.48 μU/mL in men.
Saravia et al conducted a study involving 3200 male participants to evaluate the connection between fasting insulin levels and HbA1c with metabolic syndrome.14 Both HbA1c and fasting insulin levels were found to correlate strongly with metabolic syndrome and its associated traits (eg, BMI, cholesterol levels, blood pressure, waist circumference, and blood glucose). However, plasma insulin levels showed stronger associations than HbA1c to all the associated traits of metabolic syndrome except for hyperglycemia, “… and there were significant differences between the low- and mid-range of insulin for all criteria except high blood pressure.”14 The researchers concluded that elevated insulin levels appeared to identify certain cardiometabolic changes earlier than both HbA1c and measures of glucose.
Derakhshan et al investigated the relationship between beta-cell function, fasting insulin levels, and insulin resistance and progression to T2D.15 Participants included 1532 male and 2221 female Iranian adults with no apparent glycemic impairment. They found the annual incidence rate of diabetes in this population to be 3.73/1000 person-years in the 9.2-year follow up. HOMA-IR and hyperinsulinemia were found to be associated with progression to T2D, impaired glucose tolerance, and impaired fasting glucose in both genders. However, the researchers were unable to link beta-cell dysfunction to the incidence of diabetes and surmised that this could be explained by the natural progression of diabetes. The researchers concluded that underlying pathophysiology indicates there is “…a compensatory increase in insulin secretion before the development of glucose abnormalities.”15
Yang et al assessed variances in insulin resistance in their study involving 80 adult participants.16 Individuals with glycemic dysfunction were categorized into 4 groups: normal glucose tolerance (NGT), hyperinsulinemia with NGT (HINS), impaired glucose tolerance (IGT), and newly diagnosed T2D. The researchers collected baseline data, performed OGTT, calculated simple indices of insulin resistance, and used the hyperinsulinemic-euglycemic clamp technique to measure accurate levels of insulin resistance. They determined that insulin sensitivity was lower in the HINS, IGT, and T2D groups but not statistically significantly different from one another. This finding indicates that insulin sensitivity is already impaired in the HINS state long before glucose tolerance becomes impaired. The researchers employed stepwise regression analysis to identify waistline measurement and fasting insulin level as independent predictors of insulin resistance. These findings support the suggestion that the progression to T2D includes a hyperinsulinemic stage and a prediabetic stage.
Ruijgrok et al analyzed baseline data from 1349 participants from the city of Hoorn, in the Netherlands.17 Their study examined the links between fasting insulin, HbA1c, FPG, the HOMA-IR, and 2-hour plasma glucose with progression to T2D. Baseline data were collected on the aforementioned laboratory assessments as well as BMI, family history, social history, blood pressure, exercise habits, and diet. The outcomes measure was identified as the incidence of T2D utilizing the 2011 WHO diagnostic criteria. At the 6.4-year follow-up, 11.2% of the participants had been diagnosed with T2D. Logistic regression models were utilized to calculate the probability of developing T2D for each quintile of fasting insulin, HbA1c, FPG, HOMA-IR, and 2-hour plasma glucose. The researchers determined that there were strong nonlinear associations between FPG and HbA1c with T2D. As the FPG and HbA1c increased, the incidence of T2D increased more sharply. However, the correlations between insulin resistance and fasting insulin with T2D were found to be more linear. Ultimately, the researchers determined that after adjusting for metabolic and lifestyle variables, fasting insulin was not statistically significantly associated with the incidence of T2D.
Pennings et al investigated the relationship between fasting insulin levels and weight gain.18 Their study included 3840 participants identified from the National Health and Nutrition Examination Survey (NHANES). Participants were divided into 6 groups based on fasting insulin level and level of glycemic impairment. The researchers identified 5.48 µU/mL as the cut-off point for normal fasting insulin. This study identified that participants with normal fasting insulin levels gained an average of 1.40 lb in the 10-year follow-up period while individuals with elevated fasting insulin gained an average of 11.12 lb. Additionally, the researchers found “…BMI, race, smoking status, fasting glucose, weight change, and HOMA-IR” to be significantly associated with increases in the fasting insulin level.18 They calculated that each 1-μU/mL increase in fasting insulin resulted in 0.52 lb of added weight. Remarkably, most of the weight gain occurred before glycemic impairment was measurable, which would implicate hyperinsulinism, not hyperglycemia, as an instrumental factor in weight gain.