Potassium Magnesium Citrate is Superior to Potassium Chloride in Reversing Metabolic Side Effects of Chlorthalidone

Abstract

Background:

Thiazide diuretics (TD) are the first-line treatment of hypertension (HTN) because of its consistent benefit in lowering BP and cardiovascular risk. TD is also known to cause excess risk of diabetes mellitus, which may limit long-term use. Although potassium (K) depletion was thought to be the main mechanism of TD-induced hyperglycemia, TD also triggers magnesium (Mg) depletion. However, the role of Mg supplementation in modulating metabolic side-effects of TD has not been investigated. Therefore, we aim to determine the effect of KMgCitrate on fasting plasma glucose (FPG) and liver fat by magnetic resonance imaging (MRI) during TD therapy.

Methods:

Accordingly, we conducted a double-blinded RCT in 60 nondiabetic HTN patients to compare the effects of KCl vs. KMgCitrate during chlorthalidone (CTD) treatment. Each patient received CTD alone for 3-weeks before randomization. Primary endpoint was the change in FPG after 16-weeks of KCl or KMgCit supplementation from CTD alone.

Results:

Mean age of subjects was 59±11 years (30% Black participants). CTD alone induced a significant rise in FPG, and a significant fall in serum K, serum Mg, and 24-h urinary citrate excretion (all p<0.05). KMgCit attenuated the rise in FPG by 7.9 mg/dL vs. KCl (p<0.05), which was not observed with KCl. There were no significant differences in liver fat between the two groups.

Conclusion:

KMgCit is superior to KCl, the common form of K supplement used in clinical practice, in preventing TD-induced hyperglycemia. This action may improve tolerability and cardiovascular safety in hypertensive patients treated with this drug class.

Graphical Abstract

Introduction

Thiazide diuretics (TD), particularly chlorthalidone (CTD), are considered the first-line treatment of hypertension (HTN) by many guidelines^1,2 because of its superiority in reducing BP compared to other classes of antihypertensive drugs. TD is known to cause a number of metabolic side effects such as hypokalemia, dyslipidemia, and insulin resistance, raising safety concerns over lifetime use^3,4. Mechanisms underlying TD-induced insulin resistance are poorly understood. Hypokalemia was thought to be the main mechanism of TD-induced hyperglycemia, possibly via impairing pancreatic insulin release and peripheral insulin sensitivity⁵. Another mechanism proposed is the TD-induced activation of renin-angiotensin-aldosterone-system (RAAS), which can cause an inflammatory state in addition to renal K wasting⁶. However, potassium (K) supplementation alone has not consistently prevented TD-induced hyperglycemia, despite maintaining normal serum K concentration^4,7. In addition to reducing serum K, TD also triggers magnesium (Mg) depletion, which is preventable by Mg supplementation alone or in combination with K^8,9. Clinical implications of Mg depletion associated with TD use is unknown, but numerous epidemiological studies have demonstrated a protective role of dietary Mg against the metabolic syndrome^10,11. Accordingly, we conducted a randomized clinical trial to determine the effect of KMgCit in modulating metabolic side effects of TD in hypertensive patients.

Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request. We conducted studies in 60 subjects with treated or untreated HTN with office systolic BP 130–159 or diastolic BP 80–99 mmHg (NCT02665117) after obtaining written informed consent. The study was approved by the Institutional Review Board of the University of Texas Southwestern Medical Center. Key exclusion criteria include concomitant cardiovascular diseases, chronic kidney diseases, or compelling indication for treatment with angiotensin converting enzyme inhibitor (ACEI), beta blockers (BB), or angiotensin receptor blocker (ARB), type-2 diabetes mellitus (T2DM), unprovoked hypokalemia (serum K < 3.5 mmol/L) or hyperkalemia (serum K > 5.3 mmol/L), or requirement for any K supplement on a regular basis for any reason, such as primary aldosteronism. Complete exclusion criteria are provided in the Supplemental Materials.

The study included four consecutive phases (Figure 1). Phase 1 (3–12 weeks) was a screening period during which patients continued their individual background therapy. During phase 2, the patients were asked to discontinue their antihypertensive medications for 3 weeks and were seen weekly in the research clinic for BP monitoring⁴. Phase 3 was the active-treatment period in which patients received open label chlorthalidone (CTD, 25 mg daily) for 3 weeks. Participants who developed hypokalemia (serum K of < 3.5 mmol/L) were able to enter phase 4 after 2 weeks of CTD therapy without delay for safety reasons. During phase 4, each subject was randomized to receive either KCl powder containing 40 mEq KCl /day (n=31) vs. KMgCit (IND 116,208) powder containing 40 mEq K, 20 mEq Mg and 74 mEq citrate/day (n=29) in a double-blinded manner. Additional details regarding each study phase, biochemical^12–14, hormonal¹⁵, muscle free Mg concentration^16,17, side effects⁹, and hepatic fat measurements¹⁸, are provided in the Supplemental Materials.

Figure 1.

Study visits and interventions.

Note: total of 39 participants underwent MRI for measures of muscle Mg and hepatic fat (20 from KCl group, and 19 from KMgCit group, respectively).

Statistical Methods

It was determined that 60 participants (randomized 1:1 to each study arm) would provide power of at least 0.80 at α=0.05 to detect differences in FPG (primary outcome) of 6.5 mg/dL with SD of delta of 8.5 between the two study arms for the 4-month response. This sample size would also provide power to detect a net mean difference in HOMA-IR of 0.59 (assuming SD of 0.8), muscle Mg of 0.06 mM (assuming SD of 0.08), and hepatic fat of 2.9% (assuming SD of 4.0). Although it was expected that serum K would be similar in both arms, the sample size of 60 yielded power to detect the mean difference in mean serum of K of at least 0.22 mEq/L (assuming SD of 0.3). To account for a dropout rate of 20%, we had planned to enroll 72 participants for this study. During the COVID-19 pandemic between 2020–2021, the research MRI scans were all suspended and the requirement for MRI was discontinued. Participants were allowed to continue study intervention and collection of data for primary endpoint (FPG) and other secondary endpoints.

All statistical analysis was conducted using SAS 9.4 (SAS Institute Inc., Cary, NC). To assess the extent of CTD-induced metabolic side effects, a linear mixed model was used to determine changes in plasma and urinary levels of fasting glucose (FPG), insulin, HOMA-IR, potassium (K), Mg, sodium (Na), calcium (Ca), phosphate (Pi), renin, aldosterone, parathyroid hormone (PTH), Klotho, fibroblast growth factor-23 (FGF-23), high-sensitivity C-reactive protein (hs-CRP), and 8-isoprostane (8-IsoP) after CTD alone for 3 weeks (during phase 3) from washout period in phase 2. Similarly, a linear mixed model was used to compare the change during KCl vs. KMgCit supplementation at the end of phase 4 from CTD alone in phase 3 of plasma and urinary levels of FPG, insulin HOMA-IR, K, Mg, Na, Ca, Pi, renin, aldosterone, PTH, Klotho, FGF-23, hs-CRP, and 8-IsoP. The model included treatment group effect (KCl vs. KMgCit), visit, and group-by-visit interaction with the corresponding baseline plasma and urinary levels prior to randomization (at the end of phase 3 or CTD alone) as covariates. If the model assumptions were not met, rank transformation was applied. The primary endpoint was FPG. The primary efficacy endpoint was summarized by previous treated vs. untreated HTN. For continuous secondary efficacy variables, the same analysis model was used. Significance (α-level) was defined as p <0.05. No multiplicity adjustment was applied to the analysis of all variables.

Results

Overall, 109 individuals were screened, and 81 participants met inclusion/exclusion criteria. The most common reason for exclusion prior to randomization was abnormal laboratory test results or medical history (18% of all screened patients). Among the 81 participants who met inclusion/exclusion criteria, 21 dropped out prior to randomization. The reasons for dropout included 7 new medical problems unrelated to study interventions (n=7), side effects to CTD (n=4), relocation/scheduling issue (n=6), incidental findings detected during MRI scan (n=2), claustrophobia (n=1), and non-compliance to CTD (n=1). The remaining 60 subjects were randomized to KCl or KMgCit (figure 1) and all completed 16-week study intervention. Due to the suspension of research MRI scans during COVID-19, muscle mg and liver data were only available in 39 participants (20 from KCl group and 19 from KMgCit group). Characteristics of participants during the screening period (phase 1) are shown in Table 1.

Table 1.

Subject Characteristics at Screening

Variables	KCl (n=31)	KMgCit (n=29)
Age (year)	58.1 ± 12.8	60.7 ± 9.3
Sex (M/F)	20/11	15/14
Black Participants (%)	22.58%	34.48%
Body Mass Index, kg/m2	29.0 ± 5.7	28.8 ± 4.6
Systolic Blood Pressure, mmHg	127.9 ± 13.0	128.9 ± 10.6
Diastolic Blood Pressure, mmHg	80.2 ± 7.1	77.4 ± 8.8
Heart Rate, bpm	66.6 ± 12.6	63.6 ± 8.9
Fasting Plasma Glucose, mg/dL	92.6 ± 7.9	93.6 ± 9.5
Creatinine Clearance, mL/min	123.4 ± 42.8	104.2 ± 33.4
Serum Creatinine, mg/dL	0.84 ± 0.14	0.89 ± 0.16
Serum Sodium, mmol/L	139.4 ± 2.23	138.9 ± 2.04
Serum Potassium, mmol/L	4.19 ± 0.27	4.2 ± 0.36
Serum Calcium, mg/dL	9.51 ± 0.45	9.59 ± 0.28
Serum Magnesium, mg/dL	2.08 ± 0.11	2.07 ± 0.18
Serum Phosphorus, mg/dL	3.42 ± 0.48	3.65 ± 0.50

Responses to CTD alone

We found that, prior to randomization, CTD alone induced a significant reduction in office BP, body weight, serum K, Na, 24-hour urinary Ca, 24-hour urinary uric acid, and 24-hour urinary citrate excretion; and a significant increase in levels of heart rate, urine pH, FPG, insulin, HOMA-IR, total cholesterol, LDL-C, serum Ca, klotho, renin, and aldosterone levels (Table 2 and Figure 2). Serum Mg was also reduced by CTD (p = 0.0506, Figure 2). These changes were not different among participants in two groups (all interaction p-value > 0.05, Table 2). CTD had no effect on levels of HDL-C, triglyceride, plasma 8-IsoP, 24-hour urinary Na, or 24-hour urinary K, and 24-hour urinary Mg excretion.

Table 2.

Biochemical and BP Response to CTD Alone

Variables	KCl (n=31)		KMgCit (n=29)		P - value
	Baseline (washout)	Post-CTD alone	Baseline (washout)	Post-CTD alone	CTD Factor	Group	Interaction *
Body weight, kg	86.0 ± 20.8	85.2 ± 20.5	81.9 ± 14.3	80.7 ± 14.2	<0.01	0.35	0.12
SBP, mmHg	142.5 ± 18.3	132.0 ± 12.9	144.8 ± 15.3	129.0 ± 14.0	<0.01	0.92	0.18
DBP, mmHg	85.8 ± 8.8	82.5 ± 7.1	84.9 ± 6.5	81.2 ± 8.0	<0.01	0.54	0.84
Heart Rate, bpm	66.5 ± 12.2	68.5 ± 11.1	65.2 ± 10.2	69.7 ± 10.9	<0.01	0.98	0.14
Glucose, mg/dL	92.1 ± 10.6	95.1 ± 12.6	94.0 ± 10.9	96.1 ± 9.7	0.04	0.56	0.71
Insulin, mU/L	12.6 ± 7.4	14.6 ± 8.6	14.8 ± 9.2	16.2 ± 9.5	<0.01	0.41	0.56
HOMA-IR	1.6 ± 1.0	1.9 ± 1.1	1.9 ± 1.2	2.4 ± 1.9	<0.01	0.37	0.76
Total Chol, ng/ml	176.9 ± 39.6	180.5 ± 39.7	180.3 ± 36.7	187.0 ± 40.3	0.01	0.58	0.35
HDL-C, mg/dL	54.1 ± 17.0	52.9 ± 16.1	52.4 ± 16.6	54.6 ± 17.4	0.61	0.98	0.04
LDL-C, mg/dL	102.5 ± 32.5	106.6 ± 35.1	106.5 ± 28.9	109.3 ± 30.7	0.03	0.62	0.99
TG, mg/dL	101.8 ± 49.6	113.7 ± 73.2	113.8 ± 49.1	125.5 ± 45.3	0.07	0.38	0.94
iPTH, pg/mL	51.9 ± 23.4	53.8 ± 52.0	60.6 ± 35.3	60.0 ± 63.4	0.05	0.47	0.91
Klotho, pmol/L	19.1 ± 10.7	24.3 ± 13.3	19.6 ± 10.5	20.6 ± 13.0	0.01	0.61	0.11
FGF-23, pg/mL	247.5 ± 953.2	691.4 ± 2916.3	92.10 ± 111.6	114.3 ± 191.9	0.42	0.72	0.55
Renin, pg/mL	374.3 ± 209.0	652.2 ± 444.7	332.1 ± 216.0	714.2 ± 416.8	<0.01	0.91	0.26
Aldo, pg/mL	189.1 ± 91.0	221.9 ± 92.8	159.9 ± 74.8	227.8 ± 114.2	<0.01	0.56	0.18
Serum Measures
Na, mmol/L	139.2 ± 2.08	136.9 ± 2.42	138.8 ± 2.27	137.1 ± 2.34	<0.01	0.78	0.51
K, mmol/L	4.23 ± 0.26	3.69 ± 0.31	4.19 ± 0.21	3.59 ± 0.31	<0.01	0.24	0.49
Ca, mg/dL	9.30 ± 0.38	9.57 ± 0.38	9.37 ± 0.34	9.61 ± 0.37	<0.01	0.55	0.68
Mg, mg/dL	2.05 ± 0.10	1.99 ± 0.17	2.00 ± 0.13	1.98 ± 0.21	0.05	0.39	0.41
Pi, mg/dL	3.43 ± 0.38	3.35 ± 0.42	3.33 ± 0.51	3.45 ± 0.42	0.79	0.97	0.16
CO2, mmol/L	24.4 ± 2.0	26.5 ± 2.3	24.6 ± 2.3	26.3 ± 2.6	<0.01	0.91	0.40
8-isoP, pg/mL	577.6 ± 530.0	593.8 ± 625.1	596.3 ± 537.3	577.3 ± 543.7	0.82	0.52	0.98
hs-CRP, ng/mL	5.49 ± 11.70	7.54 ± 15.24	5.74 ± 8.29	4.87 ± 5.65	0.14	0.48	0.45
Urinary Measures
Na, mmol/24-h	129.6 ± 61.4	134.2 ± 61.8	140.6 ± 66.2	134.9 ± 50.3	0.94	0.66	0.52
K, mmol/24-h	54.1 ± 20.6	55.4 ± 23.5	47.4 ± 15.0	51.1 ± 20.7	0.25	0.25	0.57
Ca, mg/24-h	169.6 ± 109.0	119.6 ± 90.9	164.3 ± 86.5	105.0 ± 102.8	<0.01	0.67	0.62
Mg, mg/24-h	95.6 ± 42.6	100.2 ± 48.4	90.0 ± 31.9	96.6 ± 38.3	0.20	0.64	0.81
Pi, mg/24-h	827.6 ± 380.3	751.7 ± 323.4	799.6 ± 364.9	739.9 ± 338.5	0.11	0.81	0.85
Citrate, mg/24-h	562.8 ± 216.1	427.3 ± 299.5	626.1 ± 304.3	405.8 ± 206.3	<0.01	0.73	0.14
NAE, mEq/24-h	52.1 ± 36.7	53.2 ± 30.7	53.3 ± 33.3	52.7 ± 32.3	0.96	0.96	0.86
8-isoP, ng/mg Cr	1.14 ± 0.64	1.05 ± 0.37	1.29 ± 1.16	1.26 ± 0.87	0.50	0.35	0.75
pH	6.0 ± 0.5	6.1 ± 0.5	5.8 ± 0.4	6.1 ± 0.5	0.02	0.65	0.18
Uric Acid, mg/24-h	587.0 ± 268.7	533.4 ± 233.7	506.2 ± 167.6	475.8 ± 176.9	0.04	0.19	0.57

^*Interaction between CTD effect and group (KMgCit vs KCl).

Aldo; aldosterone, Ca; calcium, DBP; diastolic blood pressure, FGF-23; fibroblast growth factor 23, HDL-C; high density lipoprotein cholesterol, hs-CRP; high-sensitivity C-reactive protein, iPTH; intact parathyroid hormone, K; potassium, LDL-C; low density lipoprotein cholesterol, Mg; magnesium, Na; sodium, NAE; net acid excretion, Pi; phosphorous, SBP; systolic blood pressure, TG; triglycerides, 8-isoP; 8-isoprostanes.

Figure 2.

Summary data showing mean and 95% confidence intervals.

Fasting serum A) glucose, B) insulin, C) potassium, and D) magnesium in the KCl vs. KMgCit group at baseline and in response to CTD alone prior to randomization. * P < 0.05 post-CTD alone vs Baseline.

Responses to KCl vs. KMgCit

As expected, 24-h urinary citrate excretion was significantly increased while 24-h NAE was significantly reduced in the KMgCit group than the KCl group (all p < 0.05, Table 3). Urine pH was also significantly increased in the KMgCit group then the KCl group (interaction p-value < 0.01, Table 3). Moreover, 24-h urinary Mg and 24-h urinary uric acid significantly increased in both KCl and KMgCit group, however there was no significant interaction (p < 0.05 for time, interaction p-value = 0.10, Table 3). Serum K was also increased significantly in both KCl and KMgCit group, but the magnitude of increase was greater in the KMgCit than KCl group at week 16 visit (p < 0.01 for time, interaction p value = 0.04, Table 3 and Figure 3). There was no significant change in serum Mg, or muscle tissue Mg by MRI, between the 2 groups (Table 3 and Figure 3, and Figure S1, respectively). KMgCit significantly reduced FPG by 5.6 mg/dL compared to pre-randomization values, and by 7.9 mg/dL compared to KCl group at week 16 (p < 0.05, Figure 3), while KCl had no significant effect on FPG. Fasting plasma insulin, HOMA-IR, Na, Ca, Pi, renin, aldosterone, PTH, Klotho, FGF-23, hs-CRP, and 8-IsoP, were not altered by KMgCit or KCl (Table 3). Further, hepatic fat was not altered by KMgCit or KCl (Figure S1). Similarly, SBP, DBP, and HR were not altered by KMgCit or KCl (Table 3). Response to KMgCit vs KCl among untreated hypertensive patients (n=9), and previously treated hypertensive patients (n=51), are shown in Table S1 and S2, respectively. Overall, FPG response to KMgCit vs KCl was not modified by treatment status prior to study enrollment (interaction p-value = 0.9).

Table 3.

Biochemical and BP Response to KCl vs. KMgCit from CTD Alone

Variables	KCl (n=31)		KMgCit (n=29)		P - value
	8 weeks	16 weeks	8 weeks	16 weeks	Time	Interaction *
ΔBody weight, kg	−0.1 ± 2.0	0.0 ± 2.7	0.0 ± 1.6	0.0 ± 1.4	0.96	0.99
ΔSBP, mmHg	−5.3 ± 11.8†	−9.4 ± 11.5†	−4.8 ± 14.8†	−5.6 ± 11.9†	<0.01	0.38
ΔDBP, mmHg	−3.7 ± 6.7†	−5.4 ± 7.1†	−3.3 ± 8.0†	−4.9 ± 7.3†	<0.01	0.96
ΔHeart Rate, bpm	0.9 ± 7.2	0.9 ± 5.5	0.1 ± 7.4	1.3 ± 10.1	0.58	0.79
ΔGlucose, mg/dL	0.8 ± 11.9	2.3 ± 9.9	−2.3 ± 9.2	−5.6 ± 13.3†‡	0.53	0.03
ΔInsulin, mU/L	1.2 ± 4.1	2.8 ± 7.7	2.1 ± 9.5	1.6 ± 9.3	0.68	0.85
ΔHOMA-IR	0.2 ± 0.5	0.4 ± 1.0†	0.2 ± 1.1	0.2 ± 1.1	0.08	0.41
ΔTotal Chol,	1.1 ± 24.5	2.8 ± 20.9	11.8 ± 25.7†	4.9 ± 22.5	0.10	0.14
ΔHDL-C, mg/dL	1.7 ± 6.3	0.9 ± 8.7	3.0 ± 6.6†	3.8 ± 8.5†	0.02	0.39
ΔLDL-C, mg/dL	2.2 ± 19.5	3.4 ± 14.5	8.8 ± 23.4†	2.7 ± 21.2	0.10	0.26
ΔTG, mg/dL	−12.9 ± 50.8	−6.0 ± 71.6	4.3 ± 46.0	−8.1 ± 49.9	0.62	0.19
ΔiPTH, pg/mL	−0.3 ± 49.3	−7.4 ± 47.0	−10.0 ± 53.1	−4.0 ± 28.4	0.57	0.38
ΔKlotho, pmol/L	−1.7 ± 10.6	−1.6 ± 10.10	0.6 ± 8.5	−0.9 ± 7.1	0.59	0.50
ΔFGF-23, pg/mL	−3.4 ± 74.4	13.6 ± 104.2	19.2 ± 98.4	38.1 ± 123.3	0.19	0.68
ΔRenin, pg/mL	−37.1 ± 208.3	12.4 ± 202.6	1.3 ± 175.4	1.5 ± 237.2	0.57	0.56
ΔAldo, pg/mL	50.3 ± 105.8†	46.0 ±	14.5 ± 121.5	27.1 ± 123.1	0.03	0.44
Serum Measures
ΔNa, mmol/L	0.55 ± 1.82	1.06 ± 2.93	−0.04 ± 2.30	0.29 ± 2.39	0.12	0.46
ΔK, mmol/L	0.21 ± 0.49†	0.00 ± 0.28	0.12 ± 0.38	0.18 ± 0.34†	<0.01	0.04
ΔCa, mg/dL	−0.10 ± 0.26†	−0.05 ± 0.34	−0.01 ± 0.29	0.00 ± 0.25	0.39	0.30
ΔMg, mg/dL	0.00 ± 0.14	0.04 ± 0.19	0.09 ± 0.21†‡	0.08 ± 0.17†	0.04	0.09
ΔPi, mg/dL	−0.13 ± 0.35†	0.05 ± 0.35	−0.11 ± 0.38	−0.05 ± 0.48	0.03	0.94
ΔCO2, mmol/L	−0.4 ± 2.3†	−0.5 ± 2.6	0.8 ± 2.1‡	0.1 ± 2.4	0.42	0.10
Δ8-isoP, pg/mL	−140 ± 491.1	−81.8 ± 756.0	64 ± 616.5	23.2 ± 509.6	0.36	0.68
Δhs-CRP, ng/mL	−1.23 ± 6.85	1.63 ± 11.82	−0.34 ± 5.05	0.68 ± 4.42	0.10	0.99
Urinary Measures
ΔNa, mmol/24-h	16.1 ± 75.0	15.2 ± 58.8	14.5 ± 70.4	28.5 ± 65.5†	0.06	0.66
ΔK, mmol/24-h	31.4 ± 32.7†	25.7 ± 28.1†	35.6 ± 31.9†	31.0 ± 29.1†	<0.01	0.80
ΔCa, mg/24-h	22.6 ± 69.5†	26.8 ± 67.0†	10.4 ± 64.9	22.1 ± 63.6	0.02	0.70
ΔMg, mg/24-h	20.3 ± 37.9†	12.7 ± 36.5	34.1 ± 58.7†	40.3 ± 46.6†‡	<0.01	0.10
ΔPi, mg/24-h	58.5 ± 298.5	87.9 ± 273.5	10.0 ± 279.2	58.7 ± 425.6	0.26	0.82
ΔCitrate, mg/24-h	98.0 ± 204.1†	32.5 ± 204.9	260.6 ± 270.4†‡	266.2 ± 287.0†‡	<0.01	<0.01
ΔNAE, mEq/24-h	13.7 ± 60.8	15.2 ± 34.0	−38.3 ± 34.2†‡	−25.6 ± 56.5†‡	0.15	<0.01
Δ8-isoP, ng/mg Cr	0.33 ± 1.5	0.09 ± 0.48	0.12 ± 0.63	0.21 ± 0.67	0.22	0.43
ΔpH	0.01 ± 0.53	0.03 ± 0.53	0.48 ± 0.53†‡	0.42 ± 0.59†‡	<0.01	<0.01
ΔUric Acid, mg/24-h	60.5 ± 168.0	69.2 ± 195.0	51.7 ± 192.8	92.2 ± 176.1†	<0.01	0.80

^*Interaction between group (KMgCit vs KCl) by time.

^†P < 0.05 vs week 0

^‡P < 0.05 vs KCl.

Aldo; aldosterone, Ca; calcium, DBP; diastolic blood pressure, FGF-23; fibroblast growth factor 23, HDL-C; high density lipoprotein cholesterol, hs-CRP; high-sensitivity C-reactive protein, iPTH; intact parathyroid hormone, K; potassium, LDL-C; low density lipoprotein cholesterol, Mg; magnesium, Na; sodium, NAE; net acid excretion, Pi; phosphorous, SBP; systolic blood pressure, TG; triglycerides, 8-isoP; 8-isoprostanes.

Figure 3.

Summary data showing mean change (Δ) and 95% confidence intervals post-randomization. A) fasting plasma glucose, B) serum potassium, C) serum magnesium, D) 24-hour urinary K exertion, E) 24-hour urinary Mg excretion, and F) 24-hour urinary citrate excretion in the KCl vs. KMgCit group during follow-up. * P < 0.05 vs week 0, † P < 0.05 vs KCl.

Discussion

There are three major findings of our study. First, Mg supplementation in the form of KMgCit is more effective in reversing detrimental effect of CTD, a thiazide-like diuretic, on plasma glucose than KCl, which is the typical form of K supplementation used in clinical practice. Second, the beneficial effects of KMgCit were not accounted for by modulation in RAAS, which was activated by thiazide diuretics. Third, the improvement in plasma glucose was not accompanied by alteration in oxidative stress marker, liver fat, or tissue Mg in the skeletal muscle.

Thiazide diuretics (TD) are known to induce a variety of metabolic disturbances, including electrolyte abnormalities, elevated plasma glucose, increased total cholesterol levels, and escalated risk of diabetes mellitus ^19–21. Although hypokalemia is proposed to be the main mechanisms underlying a lot of its metabolic side effects, K supplementation in the form of KCl has not been shown to prevent CTD induced elevation in FPG ^4,7. In addition to hypokalemia, TD induce a significant reduction in serum Mg which was evident within 2–3 weeks of administration⁸. Previous studies have also shown that short term administration of KMgCit is equally effective as KCl in attenuating hydrochlorothiazide (HCTZ)-induced hypokalemia^22,23. However, the duration of KMgCit in previous studies was limited to only 3 weeks^22,23, which was not sufficient to allow detection of its potential effect on the glucose metabolism. In the present study, CTD induced a small but significant reduction in serum Mg and K within 3 weeks of administration similar to effects of HCTZ. Furthermore, we found that 16-week of KMgCit is superior to KCl in attenuating CTD-induced hyperglycemia.

Mechanisms underlying benefits of KMgCit in preventing CTD-induced hyperglycemia is unknown. In addition to hypokalemia, TD triggers a sustained activation of RAAS ²⁴ and of the sympathetic nervous system⁴ in hypertensive patients, which is not explained by hypokalemia alone^4,7. Sympathetic activation has been implicated in the pathogenesis of insulin resistance, both by skeletal muscle blood flow-dependent²⁵, and flow-independent mechanisms²⁶, while RAAS activation is also known to inhibit insulin signaling in the adipocytes and skeletal muscle^27,28, resulting in impaired insulin-mediated glucose uptake²⁸. Although we did not assess sympathetic nerve activity directly, KMgCit did not attenuate the rise in circulating renin and aldosterone levels, or the small but significant rise in heart rate induced by CTD in our study, suggesting the effect of Mg and/or citrate may be direct.

Several studies also suggested a protective role of Mg intake on the metabolic syndrome. In rats subjected to a Mg-deficient diet, hypomagnesemia had a deleterious effect on glucose metabolism by impairing both insulin secretion and signaling²⁹. In normal volunteers, exposure to low Mg diet for 4 weeks resulted in impaired intravenous glucose tolerance test³⁰. Mg supplementation was also shown to reduce and delay the postprandial rise in serum non-esterified fatty acid and triglyceride levels after a fat load in normal volunteers³¹. In a meta-analysis from 6 cross-sectional studies, including a total of 24,473 individuals, an inverse association between Mg intake and the risk of metabolic syndrome was observed¹⁰. Similarly, citrate supplementation in the form of K citrate (KCit) was shown to reduce marker of insulin resistance (HOMA-IR) in prediabetic subjects, which was not observed with KCl administration³². Several populational studies have shown an association between dietary acid load and insulin resistance as well as the risk of T2DM³³. In the NHANES surveys, low serum bicarbonate and high anion gap were independently associated with increased plasma fasting insulin and high triglyceride (TG) levels ³⁴.

A previous trial in hypertensive patients with heart failure with preserved ejection fraction showed that low sodium plus Dietary Approaches to Stop Hypertension (DASH) diet, which is rich in K, Mg, and alkali from the high fruits and vegetables content, reduced urinary F2-isoprostanes by 31%, suggesting decreased oxidative stress³⁵. Our previous trial in patients with Pre- or Stage-I HTN also showed that KMgCit is more effective than KCit, KCl, and placebo, in reducing oxidative stress as evidenced by a significant decline in urinary 8-isoprostanes¹². However, none of hypertensive subjects in previous study received diuretic treatment and fasting plasma glucose and insulin were not measured during the study. In the current study, however, we did not find alterations in serum or urinary 8-isoprostane levels with KMgCit during CTD treatment.

Our study is limited by the short duration of the study and inability to separate effects of citrate from Mg on the metabolic outcomes. The increase in serum Mg after KMgCit was relatively modest but nevertheless identical to change observed in our previous study¹². The lack of statistical difference in serum Mg between KMgCit vs. KCl group (p = 0.09) is likely to be due to small sample size. Citrate supplementation may further promote metabolic alkalosis induced by chlorthalidone alone and may limit the use of KMgCit in patients with preexisting systemic alkalosis. The glucose lowering effect of KMgCit was observed only among patients treated with CTD in our study and may not be extrapolated to population without diuretic therapy. This is an important distinction as previous randomized clinical trials have not shown a consistent reduction in plasma glucose after Mg supplementation in type 2 DM without diuretic therapy^36–41. Despite this limitation, our study has significant implications as TD are the second most prescribed class of antihypertensive in the US ⁴². Despite the increasing popularity, the excess risk of T2DM associated with TD use may fuel the epidemic of T2DM in the U.S.^43,44, and limit its potential cardiovascular benefit when given as a life-long antihypertensive agent. Our current study suggests a novel effect of KMgCit supplementation in addressing not only electrolyte abnormalities more completely, but also in preventing adverse metabolic effects without increased gastrointestinal side-effects. Future studies are needed to determine if KMgCit is superior to KCl in improving cardiovascular outcomes in hypertensive patients treated with thiazide diuretics.

Perspectives

In summary, we identify a novel action of potassium-magnesium-citrate (KMgCit) supplementation in preventing chlorthalidone (CTD)-induced rise in fasting plasma glucose in hypertensive patients, which was not observed with KCl supplementation. These findings suggest that KMgCit may offer a superior alternative to KCl, typically used in clinical practice, in maintaining serum potassium during thiazide-diuretic treatment while avoiding of thiazide-induced hyperglycemia.

Novelty and Relevance:

What is new?

• We identify a novel action of potassium-magnesium-citrate (KMgCit) supplementation in preventing chlorthalidone (CTD)-induced rise in fasting plasma glucose in hypertensive patients, which was not observed with KCl supplementation.

• The beneficial effects of KMgCit were not explained by modulation in the renin-angiotensin-activation system, oxidative stress markers, liver fat, or tissue Mg in the skeletal muscle, suggesting a potential direct effect of Mg and/or citrate.

What Is Relevant?

• KMgCit may offer a superior alternative to KCl in maintaining serum potassium during thiazide-diuretic treatment while avoiding of thiazide-induced hyperglycemia; this may alleviate the rising epidemic of type-2 diabetes mellitus.

Clinical/Pathophysiological Implications?

• Future studies are needed to determine the mechanism of KMgCit action and if the benefit of KMgCit will translate into improved cardiovascular outcomes during thiazide-diuretic treatment.

Nonstandard Abbreviations

TD

Thiazide diuretics

CTD

Chlorthalidone

HTN

Hypertension

RAAS

Renin-Angiotensin-Aldosterone-System

KCl

Potassium Chloride

KMgCit

Potassium Magnesium Citrate