Elevated circulating growth differentiation factor 15 is related to decreased heart rate variability in chronic kidney disease patients

Figures & data

Table 1. Characteristics of the Study Population (N = 355).

	Mean ± SD / Number (%)
Age, years	57.4 ± 14.2
BMI, kg/m^2	25.8 ± 4.0
Male	224 (63.1%)
Current Smoker	93 (26.2%)
Diabetes	158 (44.5%)
Cardiovascular disease	82 (23.1%)
Etiology
Diabetic kidney disease	113 (31.8%)
Hypertensive nephropathy	37 (10.4%)
Glomerulonephritis	121 (34.1%)
Others	37 (10.4%)
Undetermined	47 (13.2%)
Statins	152 (42.8%)
Epo/ iron supplements	91(25.6%)
Systolic Blood pressure, mm Hg	137.0 ± 20.4
Diastolic Blood pressure, mm Hg	84.5 ± 11.5
Heart rate, bpm	74.5 ± 12.7
Hemoglobin, g/L	117.5 ± 24.4
CRP, mg/L	0.9(0.5—3.2)
Total cholesterol, mmol/L	4.7 ± 2.1
Triglycerides, mmol/L	2.1 ± 1.7
HDL cholesterol, mmol/L	1.1 ± 0.5
LDL cholesterol, mmol/L	3.0 ± 1.6
Calcium, mmol/L	2.1 ± 0.2
Phosphorus, mmol/L	1.2 ± 0.3
Urine PCR, mg/g	150.0 (40.5—358.5)
eGFR, mL/min/1.73m^2	50.1 ± 33.2
cfPWV, m/s	10.4 ± 3.4
RHI	2.0 ± 0.6

BMI: body mass index; EPO: erythropoietin; CRP: C-reactive protein; HDL-C: high-density lipoprotein cholesterol; LDL-C: low-density lipoprotein cholesterol; PCR: protein to creatinine ratio; eGFR: estimated glomerular filtration rate; cfPWV: carotid–radial pulse wave velocity; RHI: Reactive hyperemia index.

Table 2. GDF15 level in all patients and subgroups.

	Mean ± SD, pg/mL	p
All (n = 339)	1394.7 ± 610.1
Diabetes mellitus		<.001
No (n = 190)	1254.65 ± 616.24
Yes (n = 149)	1573.28 ± 554.84
Previous history of CVD		.03
No (n = 262)	1360.04 ± 634.05
Yes (n = 77)	1512.64 ± 506.51
CKD stages		<.001
CKD stage 1 + 2 group (n = 119)	1019.3 ± 532.7
CKD stage 3 group (n = 106)	1250.8 ± 414.6
CKD stage 4 + 5 group (n = 114)	1920.4 ± 458.9
uPCR stages		<.001
<30mg/g (n = 65)	1051.5 ± 449.0
30mg/g–300mg/g (n = 172)	1329.2 ± 609.3
>300mg/g (n = 97)	1725.1 ± 535.4

Figure 1. Correlation between GDF15 and eGFR and urine protein-to-creatinine ratio (n = 355) (A) GDF15 is significantly correlated with eGFR. (B) GDF15 is negatively significantly correlated with uPCR. Data are displayed as scatterplots. Abbreviations: eGFR estimated glomerular filtration rate, PCR protein-to-creatinine ratio. uPCR was logarithm transformed in correlation analysis.

Table 3. Association between GDF15 and cardiovascular evaluation results.

	r	p
Systolic blood pressure	0.303	<.001
Diastolic blood pressure	–0.062	.253
cfPWV*	0.023	<.001
LnRHI	0.074	.183
SDNN*	–0.216	<.001
RMSSD*	–0.116	.036
Triangular Index*	–0.291	<.001
LF/HF*	–0.173	.002

cfPWV: carotid–radial pulse wave velocity; RHI: Reactive hyperemia index; SDNN: standard deviation of the normal-to-normal (NN) interval; RMSSD: square root of the mean of the sum of the squares of differences between adjacent NN intervals; LF/HF: low frequency and high frequency ratio.

*These parameters were logarithm transformed in regression analysis due to their skewed distribution. Because of the closely association between PWV and SBP, the association of GDF15 with PWV was analyzed after adjustment for blood pressure.

Table 4. Associations between cardiovascular evaluation results and GDF15.

	Plasma GDF15
	Model-1			Model-2
	ß	95% confidence interval	p	ß	95% confidence interval	p
CfPWV*	0.024	0.012 − 0.035	<.001	0.007	–0.007 to 0.021	.3
SDNN*	–0.058	–0.088 to −0.028	<.001	–0.044	–0.083 to −0.055	.026
RMSSD*	–0.053	–0.087 to −0.018	.003	–0.05	–0.095 to −0.005	.029
Triangular Index*	–0.053	–0.075 to −0.032	<.001	–0.038	–0.066 to −0.01	.008
LF/HF*	–0.032	–0.07 − 0.006	.099	0.006	–0.044 to 0.056	.801

cfPWV, carotid–radial pulse wave velocity; SDNN, standard deviation of the normal-to-normal (NN) interval; RMSSD, square root of the mean of the sum of the squares of differences between adjacent NN intervals; LF/HF, the ratio of low frequency to high frequency components. Model-1: adjust for age, sex, body mass index. Model-2: adjust for age, sex, body mass index, smoker, diabetes mellitus, previous history of cardiovascular disease, eGFR, CRP, LDL, SBP.

* These parameters were logarithm transformed in regression analysis due to their skewed distribution.