Renal Denervation for Resistant Hypertension Fails to Improve Insulin Resistance as Assessed by Hyperinsulinemic-Euglycemic Step Clamp

More than 50% of patients with essential hypertension are considered to have insulin resistance (IR) (1). The link between these two disorders is partly unknown, but it has been hypothesized that increased sympathetic nerve activity (SNA) plays an important role (2). Sympathetic activation results in peripheral vasoconstriction through the release of norepinephrine, which acts upon vascular muscle adrenoreceptors. A subsequent reduction in skeletal muscle blood flow leads to reduced whole-body glucose utilization (3). Consequently, increased SNA may reduce insulin sensitivity (IS) through direct effects on the regulation of glucose uptake by skeletal muscle (4,5).

Targeting both sympathetic tone and IR might be of relevance to reduce the cardiovascular risk associated with hypertension. Renal denervation (RDN) has been shown to lower SNA by the ablation of both afferent and efferent nerves (6). The first study showing improved IS after RDN used only surrogate indexes of whole-body IR (7). To investigate the possible effects of RDN on IR in more detail, we applied a two-step hyperinsulinemic-euglycemic clamp (HEC) with glucose tracer infusion and labeled glucose infusate to separately assess glucose turnover, as well as hepatic and peripheral IS before and after RDN. Patients were also subjected to an oral glucose tolerance test (OGTT).

Twenty-three patients with treatment-resistant hypertension were included in the study and underwent RDN. Resistant hypertension was defined as office systolic blood pressure (BP) of >140 mmHg despite regular intake of maximally tolerated doses of four or more antihypertensive drugs, including a diuretic. In addition, subjects had to have an average daytime systolic BP of >135 mmHg, as measured by ambulatory BP monitoring after an investigator witnessed the intake of their antihypertensive drugs. Patients with known diabetes or an HbA_1c level of ≥6.5% were excluded. Details about the selection criteria have been published previously (8) (clinical trial reg. no. NCT01630928, clinicaltrials.gov).

The included patients gave their written, informed consent. The study was conducted in accordance with the protocol; applicable regulatory requirements; and the ethical principles of the Declaration of Helsinki and Title 45, U.S. Code of Federal Regulations, Part 46, Protection of Human Subjects, Revised 13 November 2001, effective 13 December 2001. The Regional Committee for Medical and Health Research Ethics as well as the Data Protection Officer at University Hospital of North Norway gave their approval.

All measurements were obtained within the last 4 weeks before and 6 months after RDN. Metabolic syndrome was diagnosed according to the International Diabetes Federation criteria from 2006 (9). Venous blood samples were drawn after an overnight fast (12 h). Levels of insulin during the OGTT and C-peptide were measured by ELISA (C-peptide, EIA-1293; insulin, EIA-2935; AH Diagnostics, Aarhus, Denmark). Insulin (endogenous and lispro) during HEC was measured with radioimmunoassay methods (HI-14K; Millipore, Billerica, MA) at the Hormone Laboratory, Oslo University Hospital (Aker, Norway). HEC was performed after a 12-h fast, as previously described (10). After the drawing of fasting blood samples, a primed (3 mg/kg/5 min), continuous (2.4 mg/kg/h) infusion of D-[6,6-²H₂] glucose was given for the measurement of basal glucose turnover. Tracer infusion was continued, and a primed (127 mU/m²/min for 10 min), continuous (13 mU/m²/min) infusion of human insulin (insulin lispro) was commenced after 150 min. Glucose (200 mg/mL) with D-[6,6-²H₂] glucose added at a 1.25 atom percent enrichment was variably infused to maintain normoglycemia (5 mmol/L). The second step clamp began after 270 min, when insulin infusion was increased to 40 mU/m²/min, and continued for 120 min. Sampling, chemical analysis, and the determination of tracer enrichment were performed as previously described (11) using liquid chromatography mass spectroscopy.

Whole-body IS was expressed as the glucose infusion rate (GIR) (in milligrams per kilogram per minute) during the last 40 min of each step of the clamp (steady state). The IS index (ISI) was calculated as the mean GIR divided by the mean insulin concentration at each step. Endogenous glucose release (EGR) and whole-body glucose disposal (WGD) were calculated using modified versions of Steele’s equations (12,13). The following calibrated infusion pumps were applied: care fusion Alaris Guardrails (BD, San Diego, CA) syringe pumps were used for insulin, and infusions of D-[6,6-²H₂] glucose and a tracer-enriched glucose solution were performed using Alaris Medsystem III (BD).

A standard (82.5 g of glucose monohydrate) OGTT was performed, with plasma samples obtained at 0, 30, 60, 90, and 120 min after the glucose load. Postload glucose and insulin responses were calculated as incremental area units during the 2-h sampling time, and were expressed as the area under the curve (AUC) for glucose and insulin (14).

HOMA-IR was calculated as follows: HOMA-IR = (glucose t0 [mmol/L] × insulin t0[μIU/mL]/22.5) (15). Simple Index Assessing IS OGTT (SIisOGTT) and Quantitative IS Check Index (QUICKI) were calculated using the following formulas: 1/(log [Σ glucose t 0–30–90–120] [mmol/L] + log [Σ insulin t 0–30–90–120] [μIU/mL]) and 1/(log [fasting glucose (mg/dL)]) + log (fasting insulin [μIU/mL]), respectively (16,17).

A 20% change in basal EGR (∼0.4 mg/kg/min) was considered to be clinically relevant. With an α-level of 0.05 and a power of 80%, 20–25 patients were needed to demonstrate a 20% difference in basal EGR before and after intervention (18).

Frequency distribution was checked for all variables. Data were reported as the mean ± SD if normally distributed, and median (minimum–maximum) if distribution was skewed. Accordingly, paired t tests and Wilcoxon signed rank tests were used for the comparison of variables between baseline and follow-up. Significance was accepted at P < 0.05. All analyses were performed with the SPSS statistical package version 22 (IBM, Armonk, NY).

Baseline characteristics are displayed in Table 1. The mean age was 53 ± 8 years. Twenty-one of 23 patients had central obesity, and 18 patients had metabolic syndrome (9) at baseline. Three patients were excluded from the clamp because of technical problems encountered during the procedure. The mean systolic and diastolic office BP and ambulatory BP fell significantly after RDN treatment despite a significant reduction in the number of prescribed drugs, as previously reported (8). Fifteen patients had normal fasting glycemia, 8 patients had impaired fasting glycemia, and 17 patients had impaired glucose tolerance. The OGTT-derived AUC for glucose and insulin remained unchanged after RDN (Table 1). High insulin and C-peptide concentrations were seen at baseline and remained unchanged after 6 months (Table 1). Accordingly, the indirect indices of IR, QUICKI, SIisOGTT, and HOMA-IR were high at baseline and did not improve after RDN (Table 1).

Fasting and steady-state plasma C-peptide and insulin levels during the clamp remained unaltered after RDN (Table 2). Basal EGR and WGD measured by glucose tracer infusion did not change significantly after RDN (2.12 ± 0.36 vs. 2.15 ± 0.41 mg/kg/min [P = 0.34], and 2.20 ± 0.36 vs. 2.14 ± 0.40 mg/kg/min [P = 0.35], respectively). During the two-step HEC, no significant changes in GIR and ISI were seen, indicating unaltered whole-body IS (Table 2). The suppression of EGR decreased from 0.9 ± 0.4 to 0.8 ± 0.4 mg/kg/min (P = 0.02) during low-dose insulin infusion, but remained unchanged during high-dose insulin infusion. The increase in WGD during high-dose insulin infusion was modest and remained unaltered at follow-up (Table 2). No improvement in IS was observed in a subanalysis of nine patients with extensive systolic ambulatory BP reduction (>10 mmHg) after RDN.

To our knowledge, this is the first study in which peripheral and hepatic IS have been separately assessed by gold standard methods before and 6 months after RDN in a group of patients with IR and treatment-resistant hypertension. Despite a significant reduction in BP, we found no improvement in IS, as measured by two-step HEC with infusion of a glucose tracer and a labeled glucose infusate. OGTT-based ISIs gave the same findings.

Because IS and SNA are closely associated, the previously reported amelioration of glucose metabolism after RDN in this group of patients has been followed with great interest (7). Most authors have calculated changes in IR according to HOMA-IR, a simple surrogate index that is most sensitive for changes in hepatic IR (15) and does not actually measure IS accurately on an individual basis.

The currently used two-step HEC with infusion of glucose tracer and a labeled glucose infusate remains the only reliable noninvasive method to separately assess hepatic and peripheral IS. During step clamp 2, insulin was infused at a rate that leads to the physiological hyperinsulinemia seen after a meal and that effectively increases WGD in individuals with normal IS values. The modest increase in WGD with increments in insulin infusion, both at baseline and at follow-up, supports the notion that our group of treatment-resistant hypertensive patients also experienced severe peripheral IR, which persisted after RDN. A substantial residual endogenous insulin release was present even during the highest insulin infusion rate in our study population, as indicated by the lack of decline in C-peptide levels during clamping. Interestingly, it has previously been shown that insulin inhibition of insulin secretion during HEC, as measured by the decrease in C-peptide levels, is correlated with IS (19).

Tracer dilution during HEC further makes it possible to assess the ability of insulin to suppress EGR. In order to optimize the detection of changes in hepatic IS, the insulin infusion rate during step clamp 1 was chosen to achieve a level of circulating insulin near to the median effective dose at the steep part of the sigmoidal-shaped hepatic insulin dose-response curve. Most of our patients had impaired fasting glucose levels and hyperinsulinemia, as well as high HOMA-IR index, indicating reduced hepatic IS. A smaller reduction in EGR during the low-dose insulin clamp at follow-up was statistically significant, but its clinical relevance is doubtful. A trend toward increased GIR during clamp 2 after RDN was probably related to the slightly higher serum insulin concentrations during clamping at follow-up. Accordingly, the ISI, which corrects for the prevailing insulin concentrations, did not change after RDN.

Importantly, and similar to the findings of Mahfoud et al. (7), patients in the current study experienced a significant reduction in office BP after RDN. However, in contrast to our study, 15 of 16 patients with type 2 diabetes mellitus in the study of Mahfoud et al. (7) used metformin, and adherence to the prescribed drugs was not rigorously assessed. Thus, one cannot exclude that improved patient compliance and a more rigorous metformin intake after RDN could have explained the findings in that study. In our study, HbA_1c level was significantly reduced after RDN, but the magnitude of this decrease was small and probably without clinical relevance.

Although the first RDN studies were promising, the sham-controlled HTN-3 Trial revealed no significant treatment effect of RDN on BP. Furthermore, in the DREAMS Study (20), which included only patients with metabolic syndrome, no effect on glucose metabolism was found 6 and 12 months after RDN, and neither HOMA-IR nor SIisOGTT values were changed after 6 months. Almost all patients in our study had metabolic syndrome with central obesity at baseline, and, although we did not investigate sympathetic drive directly, there are reasons to believe that increased SNA was prevalent at baseline in this group. In contrast to the DREAMS Study, in which almost drug-naive participants were included and measurements were performed during drug-free intervals, all our patients used four or more antihypertensive drug classes and were examined after witnessed drug intake. Despite these obvious differences in study populations and examination methods, no effect of RDN on IR was revealed in either study. However, even with witnessed intake and a significant reduction in the number of antihypertensive drugs used (8), we cannot exclude the possibility that improved compliance could have contributed to the fall in BP.

Although the recommended number of bilateral ablations was performed (8), we cannot exclude the possibility that the failure of RDN treatment to improve IR in our study might also be due to either insufficient ablation or true recurrence, possibly due to nerve regeneration or the sprouting of nerve fibers from the injured sympathetic chain. Indeed, previously reported improvements in IS measured by HOMA, OGTT (7), or HEC (21) were registered 3 months after RDN, whereas follow-up performed 6 months after RDN in the DREAMS Study and the current study had negative results with regard to these metabolic parameters.

The lack of a control group and the inclusion of only Caucasian patients are both important limitations of our data.

The current study shows that IS did not improve 6 months after successful RDN treatment in a highly selected group of treatment-resistant hypertensive patients with IR.

Acknowledgments. The authors thank all of the participating patients and the study staff. The authors also thank the late Professor Ingrid Toft (University Hospital of North Norway and UiT The Arctic University of Norway), who also conceived this study, without whom none of this work would have been possible.

Funding. All authors are funded by governmental nonprofit organizations in Norway and The Norwegian Diabetes Association. A.K.M. and M.D.S. are funded by The North Norwegian Health authorities. P.F.G., M.D.S., and T.K.S. are funded by UiT The Arctic University of Norway. The study was supported by grants from The North Norwegian Health authorities and an unrestricted grant from Medtronic.

The funders had no access to the study data and had no role in the design, conduct, or reporting of the study.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. A.K.M. and P.F.G. conceived and conducted the study, researched and analyzed the data, and reviewed and edited the manuscript. M.D.S., T.G.J., and T.K.S. conceived and conducted the study, analyzed the data, and reviewed and edited the manuscript. O.M.F. analyzed clamp samples and reviewed the manuscript. T.K.S. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in poster form at the European Society of Hypertension 2016 Annual Meeting, Paris, France, 10–13 June 2016.