Salsalate-induced Changes in Lipid/Lipoprotein/Apoprotein Concentrations in Obese/Overweight, Insulin Resistant, Nondiabetic Individuals

Abstract

Background and Objective

Although salsalate administration consistently lowers plasma triglyceride (TG) concentrations in patients with type 2 diabetes, prediabetes, and/or insulin resistance, changes in low-density lipoprotein cholesterol (LDL-C) concentrations have been inconsistent; varying from no change to a significant increase. To evaluate the clinical relevance of this discordance in more detail we directly measured LDL-C, as well as obtained a comprehensive assessment of changes in lipid/lipoprotein/apoprotein concentrations associated with salsalate use in insulin resistant individuals, overweight or obese, but without diabetes, using Vertical Auto Profile (VAP) method.

Methods

A single-blind, randomized, placebo-controlled study was performed in volunteers who were overweight or obese, without diabetes, and insulin resistant on the basis of their steady-state plasma glucose concentration during an insulin suppression test. Participants were randomized 2:1 to receive salsalate 3.5 grams/day (n=27) or placebo (n=14) for 4 weeks. Comprehensive lipid/lipoprotein/apoprotein analysis by VAP was obtained after an overnight fast, before and after study intervention.

Results

There was no change in directly measured LDL-C concentration in salsalate-treated individuals. However, salsalate administration was associated with various changes considered to decrease atherogenicity; including decreases in TG and total very-low-density lipoprotein cholesterol (VLDL-C) concentrations, a shift from small denser LDL lipoproteins towards larger, more buoyant LDL particles, decreases in VLDL₁₊₂-C and LDL₄-C, and non-significant decreases in non-high-density lipoprotein cholesterol and apolipoprotein B. No significant changes occurred in the placebo-treated group.

Conclusion

Atherogenicity of the lipid/lipoprotein/apoprotein profile of insulin resistant individuals who were overweight or obese improved significantly in association with salsalate treatment. The clinical importance of this finding awaits further study.

Introduction

There seems to be general agreement that salsalate administration lowers fasting plasma glucose concentration in patients with type 2 diabetes (1–3), as well as in nondiabetic individuals with prediabetes and/or resistance to insulin-mediated glucose disposal (4, 5). However, discordance exists concerning the associated change in low-density lipoprotein cholesterol (LDL-C) concentration. Goldfine, et al (1, 3) reported that LDL-C concentration significantly increased in salsalate-treated patients with type 2 diabetes (T2DM), whereas LDL-C concentration was essentially identical following salsalate treatment of patients with T2DM studied by Faghihimani and associates (2). A similar disparity in the changes in LDL-C associated with salsalate treatment was observed in persons without diabetes, with Goldfine, et al (4) noting an increase in LDL-C in persons with prediabetes, whereas Kim and colleagues could not discern any change in LDL-C concentration in salsalate-treated, insulin resistant individuals without diabetes (5). Although these studies differed in many respects, they all used the Friedewald equation (6) to provide an indirect estimate of LDL-C concentration. In this context, results (7) from an analysis of approximately 1.3 million samples demonstrated that the Friedewald equation consistently underestimated LDL-C concentration as compared to a direct measurement, and this seemed relevant to the uncertainty surrounding the effect of salsalate on LDL-C concentration. Of particular interest was evidence that differences between Friedewald-estimated and directly measured LDL-C concentrations were particularly a problem if TG concentration was ≥ 150 mg/dL (7). In light of this information, it seemed important to obtain direct measurements of LDL-C concentrations, not an estimate, in salsalate-treated individuals.

An additional reason to obtain a more detailed assessment of the effect of salsalate administration on lipid metabolism stems from the fact that in addition to an improvement in glycemia, salsalate treatment has also been associated with decreases in triglyceride (TG) concentration (1–5). Although hypertriglyceridemia may not be an independent risk factor for coronary heart disease (CHD), there is considerable evidence that a number of changes in lipoprotein metabolism associated with an elevated TG concentration contribute to CHD risk (8). Also, there is a paucity of information regarding the effect of salsalate on non-high-density lipoprotein cholesterol (non-HDL-C) and apolipoprotein B which are known markers of cardiovascular risk (9–11).

For the aforementioned reasons, we submitted plasma samples from a recently completely study (5) of salsalate administration to insulin resistant individuals without diabetes for analysis by the Vertical Auto Profile (VAP) method (12, 13). This particular method was chosen because it provides a direct measure of LDL-C concentration, and was used in the study comparing Friedewald-estimated vs. directly measured LDL-C concentrations referred to above (7). In addition, it also directly measures cholesterol concentrations in high-density (HDL), low-density (LDL), very-low-density (VLDL), intermediate-density (IDL) lipoproteins, and lipoprotein (a) (Lp (a)) and various subclasses (HDL₂-C, HDL₃-C, LDL₁-C, LDL₂-C, LDL₃-C, LDL₄-C, IDL₁-C, IDL₂-C, VLDL₁₊₂-C, VLDL₃-C) as well as measures LDL diameter, and calculates concentrations of apolipoproteins (apo) B and A-1. Thus, in addition to providing the first directly measured changes in LDL-C concentration associated with salsalate treatment, we sought to gather information as to the improvement, if any, in the atherogenic lipid/lipoprotein/apoprotein profile of salsalate-treated persons.

Materials and Methods

Study Design and Patient Population

The experimental data represent the findings from a single-blind, randomized, placebo-controlled study, with the goal of defining the metabolic effects of salsalate on volunteers who were overweight or obese, who did not have diabetes, but were insulin resistant (5). Participants were randomized 2:1 to receive salsalate 3.5 grams/day or placebo in two divided doses. The Stanford Institutional Review Board approved the protocol, and informed consent was obtained from all volunteers. Volunteers were aged 30–75 years, with overweight/obesity (BMI 25–40 kg/m²), nondiabetes (fasting glucose <126 mg/dL (7 mmol/L)) taking no glucose-lowering drugs, and without cardiac, hepatic, or renal disease, and classified as being insulin resistant with a steady-state plasma glucose (SSPG) concentration of >150 mg/dL (8.3 mmol/L) based on results of the insulin suppression test as described previously (14–17). Briefly, after an overnight fast, an intravenous catheter was placed in each of the subjects’ arms. One arm was used for the administration of a 180-minute infusion of octreotide (0.27 μg per m²/min), insulin (32 mU per m²/min) and glucose (267 mg per m²/min); the other arm was used for collecting blood samples. Blood was drawn at 10-minute intervals from 150 to 180 minutes of the infusion to determine the SSPG and steady-state plasma insulin (SSPI) concentrations. Since endogenous insulin is suppressed by the octreotide, SSPI concentrations are similar in all subjects. Consequently, the SSPG concentration provides a direct measure of the ability of insulin to mediate disposal of an infused glucose load; therefore, the higher the SSPG concentration, the more insulin resistant the individual. It should be noted that insulin-mediated glucose disposal as determined by the IST is closely correlated with those obtained with the euglycemic, hyperinsulinemic clamp technique (15, 17).

Participants were block randomized 2:1 (salsalate: placebo) by sex and BMI (<30, ≥30 kg/m²), and blinded to treatment assignment for the duration of the study. Salsalate was administered at a dose of 3.5 grams (0.5 grams/pill), divided into two daily doses, and continued for 4 weeks. Placebo-treated subjects took the same number of pills. Participants were seen weekly to monitor for symptoms and instructed to maintain medication, activity regimen, and current weight. The salsalate dose was decreased to 3.0 mg/day in 4 individuals due to persistent tinnitus.

Experimental Measurements

Plasma samples were obtained after an overnight fast at baseline and 4 weeks later for measurement of plasma glucose and lipid/lipoprotein/apoprotein concentrations. Glucose was determined by the oxidase method (Analyzer 2; Beckman, Brea, CA, USA), and samples, maintained at −80° C, were submitted to Atherotech Inc. (Birmingham, AL, USA) for VAP analysis, which uses an inverted rate zonal, single vertical spin, density gradient ultracentrifugation. This approach provides direct measurement of cholesterol concentrations of all five lipoprotein classes: HDL, LDL, VLDL, IDL, and Lp (a) and various subclasses (HDL₂-C, HDL₃-C, LDL₁-C, LDL₂-C, LDL₃-C, LDL₄-C, IDL₁-C, IDL₂-C, VLDL₁₊₂-C, VLDL₃-C) (12).

Statistical Analysis

Statistical analyses were performed using SAS (version 9.4; SAS Institute, Cary, NC, USA). Data are reported as mean ± SD or percent where applicable, unless otherwise specified. Mean differences within each of the two treatment groups (follow-up vs. baseline) were assessed by paired t-tests. Differences between salsalate and placebo groups were assessed using an independent t-test. For categorical variables, univariate analyses were performed using chi-squared tests. Statistical significance was defined as a p-value <0.05 and all testing was two-tailed.

Results

Forty-one participants were initially randomized (27 salsalate and 14 placebo) to salsalate vs. placebo. One participant assigned to salsalate dropped out due to tinnitus and heartburn symptoms, and another participant assigned to placebo dropped out for personal reasons. Therefore, 39 participants completed the 4-week treatment period and were included in the data analysis.

Table 1 compares baseline demographic and metabolic characteristics of the two experimental groups, and indicates that the only difference between them was a higher systolic blood pressure of marginal significance (p=0.05) in the salsalate-treated group. All subjects were overweight/obese, with values of body mass index elevated to a comparable degree in both groups. Both fasting plasma glucose concentration and the prevalence of prediabetes were increased somewhat in the salsalate-treated group, but neither of these differences was statistically significant. Finally, SSPG concentration, the measure of insulin resistance, was increased to a comparable degree in the two groups.

Table 1.

Baseline demographic characteristics: Salsalate vs Placebo

Variable	Salsalate (n=26)	Placebo (n= 13)	P value
Age (yrs)	55 ± 10	54 ± 10	0.95
Sex (% male)	46	31	0.49
Non-Hispanic White (%)	62	54	0.74
BMI (kg/m2)	33.0 ± 3.8	33.3 ± 2.9	0.81
WC (cm)	108 ± 11	108 ± 8	0.91
SBP (mmHg)	131 ± 13	122 ± 12	0.05
DBP (mmHg)	80 ± 8	76 ± 9	0.14
NGT/PreDM (%)	19/81	46/54	0.13
FPG (mmol/l)	6.0 ± 0.5	5.8 ± 0.6	0.18
SSPG (mmol/l)	11.9 ± 2.5	11.6 ± 1.6	0.63

BMI, body mass index; DBP, diastolic blood pressure; FPG, fasting plasma glucose; NGT, normal glucose tolerance; PreDM, pre-diabetes; SSPG, steady-state plasma glucose; SBP, systolic blood pressure; WC, waist circumference.

Table 2 compares lipid, lipoprotein and apoprotein concentrations as quantified by VAP analysis before and 4-weeks after administration of salsalate or placebo. Of direct relevance to the goal of this study was that directly measured total LDL-C concentrations were essentially identical before and after 4 weeks of salsalate treatment. However, there were a number of changes in salsalate-treated persons worthy of notice. Both TG and VLDL-C concentrations decreased significantly, associated with a shift away from small, dense LDL particles (pattern B) towards larger, more buoyant LDL particles (pattern A and AB), and decreases in VLDL₁₊₂-C and LDL₄-C (the smallest, most dense and atherogenic LDL subfraction). In addition, concentrations of non-HDL-C decreased by 3.8% and apoB by 5.0%, although in neither case did these changes reach the conventional level of statistical significance, i.e., p=0.25 and, 0.07 respectively.

Table 2.

Changes (mean ± SD) in lipid and lipoprotein concentrations as measured by VAP in individuals receiving salsalate or placebo

	Salsalate (n = 26)				Placebo (n = 13)				Salsalate vs. Placebo
	Baseline	4-weeks	Within Group Difference (95%CI)	P-value	Baseline	4-weeks	Within Group Difference (95%CI)	P-value	P-value
Direct-Measured Cholesterol
TC (mmol/L)	5.14 ± 0.90	4.99 ± 0.78	−0.15 (−0.42, 0.13)	0.29	5.96 ± 1.31	5.95 ± 1.55	−0.004 (−0.23,0.23)	0.97	0.41
Total LDL-C (mmol/L)	3.20 ± 0.76	3.13 ± 0.67	−0.06 (−0.31, 0.18)	0.60	3.90 ± 1.10	3.86 ± 1.26	−0.04 (−0.23, 0.14)	0.62	0.89
LDLr-C (mmol/L)	2.62 ± 0.65	2.55 ± 0.59	−0.07 (−0.28, 0.14)	0.52	3.15 ± 0.94	3.14 ± 1.10	−0.01 (−0.18, 0.16)	0.90	0.67
LDL1-C (mmol/L)	0.51 ± 0.18	0.55 ± 0.23	0.04 (−0.04, 0.12)	0.29	0.67 ± 0.29	0.63 ± 0.28	−0.04 (−0.12, 0.4)	0.30	0.20
LDL2-C (mmol/L)	0.64 ± 0.43	0.81 ± 0.34	0.17 (0.04, 0.30)	0.01	0.68 ± 0.33	0.59 ± 0.39	−0.09 (−0.26, 0.07)	0.24	0.02
LDL3-C (mmol/L)	1.14 ± 0.39	1.03 ± 0.42	−0.11 (−0.26, 0.04)	0.14	1.36 ± 0.43	1.47 ± 0.66	0.11 (−0.05, 0.27)	0.16	0.07
LDL4-C (mmol/L)	0.34 ± 0.30	0.17 ± 0.17	−0.17 (−0.26, −0.08)	0.0009	0.44 ± 0.44	0.45 ± 0.33	0.01 (−0.18, 0.21)	0.88	0.05
Total HDL-C (mmol/L)	1.22 ± 0.32	1.22 ± 0.28	0.002 (−0.05, 0.05)	0.93	1.21 ± 0.40	1.21 ± 0.32	0.002 (−0.09, 0.09)	0.96	1.00
HDL2-C (mmol/L)	0.27 ± 0.11	0.27 ± 0.11	0.004 (−0.02, 0.03)	0.72	0.31 ± 0.15	0.29 ± 0.12	−0.01 (−0.05, 0.02)	0.43	0.37
HDL3-C (mmol/L)	0.95 ± 0.22	0.95 ± 0.19	0.002 (−0.04, 0.04)	0.92	0.90 ± 0.25	0.91 ± 0.20	0.01 (−0.05, 0.07)	0.77	0.86
Total VLDL-C (mmol/L)	0.72 ± 0.23	0.63 ± 0.24	−0.08 (−0.14, −0.03)	0.004	0.85 ± 0.28	0.88 ± 0.33	0.03 (−0.08, 0.15)	0.52	0.03
VLDL1+2-C (mmol/L)	0.32 ± 0.14	0.27 ± 0.13	−0.05 (−0.08, −0.03)	0.0002	0.39 ± 0.18	0.41 ± 0.18	0.02 (−0.04, 0.08)	0.54	0.01
VLDL3-C (mmol/L)	0.40 ± 0.11	0.36 ± 0.11	−0.03 (−0.06, 0.001)	0.06	0.46 ± 0.10	0.47 ± 0.17	0.01 (−0.05, 0.08)	0.65	0.15
Total IDL-C (mmol/L)	0.45 ± 0.17	0.44 ± 0.18	−0.01 (−0.07, 0.05)	0.71	0.58 ± 0.24	0.58 ± 0.27	−0.004 (−0.07, 0.06)	0.90	0.89
IDL1-C (mmol/L)	0.14 ± 0.06	0.13 ± 0.06	−0.01 (−0.04, 0.01)	0.22	0.18 ± 0.07	0.19 ± 0.11	0.01 (−0.02, 0.04)	0.43	0.18
IDL2-C (mmol/L)	0.31 ± 0.12	0.31 ± 0.13	0.002 (−0.04, 0.04)	0.90	0.41 ± 0.17	0.39 ± 0.18	−0.02 (−0.07, 0.03)	0.36	0.46
Other CVD Risk Factors
TG (mmol/L)	1.82 ± 0.82	1.30 ± 0.53	−0.52 (−0.71, −0.34)	<0.0001	2.07 ± 0.64	2.09 ± 0.62	0.02 (−0.30, 0.35)	0.87	0.002
Non-HDL-C (mmol/L)	3.92 ± 0.83	3.77 ± 0.78	−0.15 (−0.41, 0.11)	0.25	4.74 ± 1.29	4.74 ± 1.49	<0.0001 (−0.18, 0.18)	1.0	0.33
Lp(a)-C (mmol/L)	0.18 ± 0.10	0.21 ± 0.09	0.02 (−0.0004, 0.05)	0.05	0.23 ± 0.16	0.20 ± 0.17	−0.03 (−0.06, 0.001)	0.06	0.009
LDL Density (% Pattern B)	42	15	n/a	0.02	38	62	n/a	0.08	n/a
apoB (mmol/L)	1.01 ± 0.20	0.96 ± 0.18	−0.05 (−0.11, 0.004)	0.07	1.21 ± 0.30	1.21 ± 0.33	0.005 (−0.04, 0.05)	0.82	0.10
apoA-1 (mmol/L)	1.45 ± 0.22	1.44 ± 0.20	−0.02 (−0.05, 0.02)	0.41	1.45 ± 0.26	1.46 ± 0.23	0.01 (−0.05, 0.07)	0.72	0.44
apoB/A1 ratio	0.71 ± 0.17	0.68 ± 0.16	−0.03 (−0.07, 0.01)	0.15	0.86 ± 0.24	0.84 ± 0.21	−0.02 (−0.07, 0.04)	0.49	0.71

CI, confidence interval; TC, total cholesterol; LDL-C, low-density lipoprotein cholesterol; LDLr-C, low-density lipoprotein cholesterol as the sum of LDL_1+2+3+4-C sufractions; IDL-C, intermediate-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; VLDL-C, very-low-density lipoprotein cholesterol; TG, triglyceride; Lp(a)-C, lipoprotein (a) cholesterol; apoB, apolipoprotein B; apoA-1; apolipoprotein A-1.

There were no significant changes in any of the lipid/lipoprotein/apoprotein concentrations in the placebo-treated group, and all of the significant within group changes observed in the salsalate-treated individuals resulted in statistically significant differences between the salsalate and placebo-treated groups.

Conclusions

Salsalate lowers plasma glucose and TG concentrations in both insulin resistant persons and patients with type 2 diabetes (1–5), raising the possibility that its use might be of benefit in several clinical syndromes. However, reports of increases in LDL-C concentration in some (1, 3, 4), but not all (2, 5), studies of the metabolic effects of salsalate raise some concern as to its overall metabolic impact. A possible explanation for these discordant findings is that LDL-C concentration in these studies was not a direct measurement, but an estimate, derived from the Friedewald equation (6). In an effort to help clarify this situation, we used the VAP analytic method to directly quantify LDL-C concentration (12, 13). In addition, since previous studies consistently demonstrated that TG concentrations decrease in salsalate-treated persons, we thought that obtaining a complete lipid/lipoprotein/apoprotein profile with VAP methodology should also help evaluate the potential clinical benefit of salsalate treatment.

The central main outcome of our study is that there was no change in plasma LDL-C concentration associated with salsalate administration, when quantified directly, and specifically by the VAP technique. In addition to the lack of an adverse effect on LDL-C concentrations, the VAP results showed that the improvement in lipid/lipoprotein profile in salsalate-treated patients is not limited to a decrease in TG concentration, but includes decreases in LDL₄-C and the proportion of small, dense LDL particles (Pattern B), as well as concentration of total VLDL-C and VLDL₁₊₂-C. Finally, there was also a non-significant trend in the decrease of non-HDL-C and apoB which are both strong markers of cardiovascular risk (9–11). Thus, in the absence of an untoward effect of salsalate on LDL-C concentration, and the overall improvement in the atherogenic lipid/lipoprotein profile of salsalate-treated persons with insulin resistance, the possibility that salsalate might have therapeutic potential should not be abandoned.

Several population-based studies have demonstrated that small, dense LDL-particles predict CHD (18–20), and a recent study (21) has documented “the association of small, dense LDL with greater coronary atherosclerosis progression.” This latter study (21) is of interest in that four methods were used to analyze lipoprotein particle subfractions, including VAP, leading the authors to conclude that “consistency of the relationships of coronary atherosclerosis progression with measures of small LDL by four independent methods supports their use for assessing the pathophysiologic effects of these particles, and for managing risk for coronary artery disease.” Moreover, the Expert Panel of the National Lipid Association has emphasized the superiority of non-HDL-C and apoB over LDL-C in predicting cardiovascular disease risk, which further supports consideration of the use of salsalate in persons at high cardio-metabolic risk as examined in our study population of individuals who are overweight or obese and classified as insulin resistant (9–11).

In conclusion, although we studied relatively few individuals, and the treatment period lasted only one month, we think our findings are interesting for the following reasons. To begin with, the salsalate-treated group was at greatly increased risk of cardio-metabolic disease based on being overweight/obese, insulin resistant, with 81% already meriting the classification of prediabetes. In addition, to the best of our knowledge, this represents the first time that a direct and extensive analysis has been made of the changes in the lipid/lipoprotein/apoprotein profile associated with salsalate administration. Therefore, it now appears that in persons at high cardio-metabolic risk, salsalate is able to not only lower plasma glucose concentration, but also to decrease the atherogenicity of circulating lipid/lipoproteins in the absence of any increase in LDL-C concentration. Whether salsalate will eventually emerge as a useful therapeutic agent remains to be seen, but our findings suggest that: 1) it might eventually achieve this status; and 2) there is no reason not to pursue the possibility.

Highlights.

• Population studied was overweight/obese, nondiabetic, insulin resistant individuals

• Directly measured LDL-C concentration did not change in salsalate-treated subjects

• Salsalate was associated with changes considered to decrease atherogenicity

• TG and VLDL-C levels fell, along with a shift to larger, more buoyant LDL particles