The association between peripheral eosinophil count and chronic kidney disease: evidence from NHANES 1999–2018

Figures & data

Table 1. Evidence Supporting the association between PEC and CKD.

Publication	Location	Inclusion criteria	Sample size	Main findings involving PEC and CKD
Hattori K et al. 2023 [6]	Japan	Patients with an eGFR of 10–60 mL/min/1.73 m^2	2877	Patients in the highest PEC quartile were more likely to have lower eGFR and higher proteinuria. A higher baseline PEC was significantly associated with a higher rate of renal replacement therapy initiation (PEC, >289 cells/uL vs. <289 cells/uL, HR, 1.83, 95%CI 1.33 − 2.51).
Kielar D et al. 2023 [7]	America	Adult (≥18 years) patients with CKD stage 3 or 4	383708	High PEC (≥300 cells/µL) was associated with CKD progression. Those with PEC ≥300 cells/µL had a relatively lower eGFR, higher all-cause mortality, and reduced time to CKD progression than those with <300 cells/µL.
Pongdee T et al. 2022 [8]	America	Individuals aged 30 years or older	44035	CKD prevalence was higher in the population with high eosinophils (210 - 350cells/uL) than in those with low eosinophils (<90cells/uL): 5.5% vs. 3.7%.
Yen CH et al. 2021 [9]	China	Age older than 30 years and absence of pregnancy.	2128	PEC was not significantly higher in CKD patients than non-CKD population (p = 0.06), and PEC was not significantly correlated with eGFR.
Kang DH et al. 2020 [10]	America	Patients with ESKD who initiated dialysis	107506	Eosinophilia (>350 cells/μL) was observed in 23.4% of patients with ESKD. There was a gradual increase in PEC over time after hemodialysis initiation.
Yang J et al. 2020 [11]	China	Patients receiving peritoneal dialysis	1432	After peritoneal dialysis initiation, PEC and percentage of patients with eosinophilia were significantly increased compared with baseline.
Tariq A et al. 2020 [12]	America	Patients who underwent native or transplant kidney biopsies	178	Peripheral eosinophilia acounted for 37%. Eosinophilia was associated with higher risk of ESKD (OR, 15.9; 95%CI, 1.9, 134.7); Peripheral eosinophilia had a significant linear association with kidney tissue eosinophils.
Hildebrand S et al. 2016 [13]	UK	Patients receiving hemodialysis	510	Persistent eosinophilia were identified as 4.7%.
Nakhjavani M et al. 2015 [14]	Iran	Patients with type 2 diabetes and healthy controls.	376	Diabetic patients with CKD had higher PEC compared with both diabetic patients without CKD and healthy controls. PEC was negatively associated with eGFR.
Ishii R et al. 2016 [15]	Japan	Cardiac inpatients with CKD	1022	Patients with higher CKD stages had a greater prevalence of the highest PEC quartile or eosinophilia (>500cells/uL). the highest PEC quartile was positively associated with higher risk of CKD 4 - 5 (OR, 1.99, 95%CI, 1.20–3.31).
Agarwal R et al. 2011 [16]	America	Patients with CKD and healthy controls	420	Patients with CKD had more eosinophils compared to controls. Eosinophil (%) increase > 3% was associated with higher risk of all-cause mortality and ESKD among people with CKD (HR 1.58, 95%CI, 1.13 − 2.22).
Diskin CJ et al. 2011 [17]	America	Patients referred to the nephrology service	1339	The percentage of eosinophils was significantly higher in patients with proteinuria and hematuria and lowest in patients with transplants.
Fukui M et al. 2009 [18]	Japan	Patients with type 2 diabetes	783	PEC was independent determinants of albuminuria in men, but not in women.

CKD: chronic kidney disease; ESKD: end-stage kidney disease; HR: hazard ratio; OR: odds ratio; CI: confidence interval; PEC: peripheral eosinophil counts.

Figure 1. Flowchart of participants selection. CKD: chronic kidney disease; eGFR-EPI: estimated glomerular filtration rate based on the EPI formula; UACR: urinary albumin to creatinine ratio. Weighted numbers are the representative national estimates, after incorporating sample weights, strata, and primary sampling units in accordance with NHANES recommendations.

Table 2. Baseline characteristics of participants with chronic kidney disease stratified by quartiles of eosinophil counts.

		PEC (cells/uL)
Characteristic	Overall	Quartile 1 (< 109)	Quartile 2 (110–176)	Quartile 3 (177–268)	Quartile 4 (> 269)	p value
Patients, no.	9224	2313	2283	2278	2350
PEC, cells/uL	176 (109, 268)	78 (57, 95)	143 (127, 159)	217 (194, 242)	362 (307, 470)	<0.001
Age, years	64 (48, 76)	60 (42, 74)	64 (48, 76)	65 (49, 76)	66 (52, 76)	<0.001
Sex, female no. (%)	4844 (52.52%)	1381 (59.71%)	1244 (54.49%)	1198 (52.59%)	1021 (43.45%)	<0.001
Race/ethnicity						<0.001
Non-Hispanic White	4290 (46.51%)	942 (40.73%)	1057 (46.30%)	1093 (47.98%)	1198 (50.98%)
Non-Hispanic Black	2306 (25.00%)	730 (31.56%)	579 (25.36%)	527 (23.14%)	470 (20.01%)
Mexican American	1367 (14.82%)	342 (14.78%)	347 (15.20%)	352 (15.45%)	326 (13.87%)
Other Hispanic	619 (6.71%)	144 (6.23%)	160 (7.01%)	141 (6.19%)	174 (7.40%)
Others	642 (6.96%)	155 (6.70%)	140 (6.13%)	165 (7.24%)	182 (7.74%)
eGFR, ml/min/1.73m^2	64.35 (51.86, 95.96)	77.58 (54.38, 102.09)	65.46 (52.82, 95.86)	60.79 (50.28, 94.62)	59.16 (49.71, 91.63)	<0.001
CKD stages						<0.001
Stage 1	2490 (26.99%)	743 (32.12%)	640 (28.03%)	577 (25.33%)	530 (22.55%)
Stage 2	2107 (22.84%)	525 (22.70%)	520 (22.77%)	523 (22.96%)	539 (22.94%)
Stage 3	4226 (45.82%)	990 (42.80%)	1,041 (45.60%)	1,075 (47.19%)	1,120 (47.66%)
Stage 4	345 (3.74%)	47 (2.03%)	70 (3.07%)	91 (3.99%)	137 (5.83%)
Stage 5	56 (0.61%)	8 (0.35%)	12 (0.53%)	12 (0.53%)	24 (1.02%)
UACR, mg/g	40.64 (12.72, 90.47)	40.39 (16.60, 83.12)	40.00 (12.38, 83.70)	40.95 (11.79, 91.02)	41.24 (11.70, 104.02)	0.723
WBC,10^3 cells/uL	7.20 (5.90, 8.60)	6.50 (5.40, 7.90)	7.00 (5.80, 8.30)	7.30 (6.10, 8.60)	8.00 (6.60, 9.60)	<0.001
BMI, kg/m^2	28.70 (24.71, 33.78)	27.30 (23.42, 32.05)	28.74 (24.70, 33.43)	29.32 (25.40, 34.70)	29.50 (25.40, 34.60)	<0.001
Hypertension, no. (%)	6309 (68.40%)	1449 (62.65%)	1549 (67.85%)	1624 (71.29%)	1687 (71.79%)	<0.001
Diabetes, no. (%)	2820 (30.57%)	568 (24.56%)	698 (30.57%)	726 (31.87%)	828 (35.23%)	<0.001
Asthma, no. (%)	1231 (13.35%)	254 (10.98%)	262 (11.48%)	331 (14.53%)	384 (16.34%)	<0.001
Smoking status						<0.001
Current smoker	1527 (16.55%)	317 (13.71%)	374 (16.38%)	372 (16.33%)	464 (19.75%)
Former smoker	2987 (32.38%)	664 (28.71%)	708 (31.01%)	761 (33.41%)	854 (36.34%)
Never smoker	4710 (51.07%)	1332 (57.58%)	1201 (52.61%)	1145 (50.26%)	1032 (43.91%)
NHANES cycle, no. (%)						0.350
1999–2000	662 (7.18%)	170 (7.35%)	177 (7.75%)	156 (6.85%)	159 (6.77%)
2001–2002	886 (9.61%)	234 (10.12%)	223 (9.77%)	224 (9.83%)	205 (8.72%)
2003–2004	898 (9.74%)	218 (9.42%)	221 (9.68%)	239 (10.49%)	220 (9.36%)
2005–2006	912 (9.89%)	223 (9.64%)	236 (10.34%)	216 (9.48%)	237 (10.09%)
2007–2008	1043 (11.31%)	269 (11.63%)	238 (10.42%)	262 (11.50%)	274 (11.66%)
2009–2010	986 (10.66%)	229 (9.90%)	248 (10.86%)	260 (11.41%)	249 (10.60%)
2011–2012	916 (9.93%)	228 (9.86%)	261 (11.43%)	214 (9.39%)	213 (9.06%)
2013–2014	994 (10.78%)	278 (12.02%)	211 (9.26%)	245 (10.76%)	260 (11.06%)
2015–2016	922 (10.00%)	218 (9.42%)	209 (9.15%)	244 (10.71%)	251 (10.68%)
2017–2018	1005 (10.90%)	246 (10.64%)	259 (11.34%)	218 (9.58%)	282 (12.00%)

BMI: body mass index; CKD: chronic kidney disease; eGFR: estimated glomerular filtration rate; PEC: peripheral eosinophil count; UACR: urinary albumin to creatinine ratio; WBC: white blood cells; NHANES: National Health and Nutrition Examination Survey. Continuous variables with non-normal distribution are shown as medians (25th, 75th percentiles), categorical variables are described as numbers (percentages).

Table 3. Associations between eosinophil counts and UACR and eGFR.

	Model 1	p value	Model 2	p value	Model 3	p value
eGFR
β per 100 cells/uL increase in PEC (95% CI)	−1.56 (−2.02, −1.11)	<0.001	−0.83 (−1.16, −0.49)	<0.001	−0.71 (−1.04, −0.37)	<0.001
Quartile 1 (< 109 cells/uL)	Ref.		Ref.		Ref.
Quartile 2 (110–176 cells/uL)	−4.45 (−6.72, −2.19)	<0.001	−1.14 (−2.58, 0.28)	0.115	−1.14 (−2.34, 0.05)	0.061
Quartile 3 (177–268 cells/uL)	−6.35 (−8.31, −4.39)	<0.001	−2.63 (−4.00, −1.25)	<0.001	−2.09 (−3.23, −0.95)	<0.001
Quartile 4 (> 269 cells/uL)	−9.83 (−11.91, −7.74)	<0.001	−4.65 (−6.16, −3.14)	<0.001	−3.83 (−5.18, −2.47)	<0.001
p for trend	< 0.001		< 0.001		< 0.001
UACR
β per 100 cells/uL increase in PEC (95% CI)	17.58 (9.14, 26.03)	<0.001	17.82 (9.16, 26.49)	<0.001	10.21 (1.37, 19.06)	0.024
Quartile 1 (< 109 cells/uL)	Ref.		Ref.		Ref.
Quartile 2 (110–176 cells/uL)	30.97 (−1.87, 63.83)	0.064	37.65 (4.50, 70.79)	0.026	27.26 (-7.84, 62.36)	0.127
Quartile 3 (177–268 cells/uL)	38.67 (2.44, 74.91)	0.036	44.31 (8.96, 79.65)	0.014	25.69 (-8.62, 60.00)	0.141
Quartile 4 (> 269 cells/uL)	100.90 (60.78, 141.03)	<0.001	107.28 (67.25, 147.31)	< 0.001	62.72 (25.46, 99.97)	0.001
p for trend	< 0.001		< 0.001		0.002

CI: confidence interval; eGFR: estimated glomerular filtration rate; UACR: urinary albumin to creatinine ratio. Model 1 was unadjusted model; Model 2 was adjusted for age (continuous), sex (male or female), and race/ethnicity. Model 3 was further adjusted for BMI (<25, 25 ∼ 29.9, and 30 kg/m²), WBC (continuous), smoking behavior (former, current, or never smoker), hypertension (yes or no), diabetes (yes or no), asthma (yes or no), UACR (≥ 30 or <30 mg/g) or eGFR (≥ 60 or <60mL/min/1.73m²), and NHANES cycles (1999–2000, 2001–2002, 2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012, 2013–2014, 2015–2016, 2017–2018).

Figure 2. Restricted cubic spline curve for the association between eosinophils, albuminuria and estimated glomerular filtration rate. eGFR: estimated glomerular filtration rate; UACR: urinary albumin to creatinine ratio. A, eosinophils and eGFR in model 3; B, eosinophils and UACR in model 3; C, subgroup analysis for eosinophils and eGFR stratified by smoking status in model 3; D, subgroup analysis for eosinophils and UACR stratified by hypertension in model 3. Model 3 was adjusted for age (continuous), sex (male or female), race/ethnicity, BMI (<25, 25 ∼ 29.9, and 30 kg/m²), WBC (continuous), smoking behavior (former, current, or never smoker), hypertension (yes or no), diabetes (yes or no), asthma (yes or no), and UACR (≥ 30 or <30 mg/g) or eGFR (≥ 60 or <60mL/min/1.73m²), and NHANES cycles (1999–2000, 2001–2002, 2003–2004, 2005-2006, 2007-2008, 2009-2010, 2011–2012, 2013–2014, 2015-2016, 2017-2018).

Figure 3. Subgroup analysis between eosinophils and albuminuria and estimated glomerular filtration rate in fully adjusted model. eGFR, estimated glomerular filtration rate; UACR, urinary albumin to creatinine ratio. Fully adjusted model was adjusted for age (continuous), sex (male or female), race/ethnicity, BMI (<25, 25 ∼ 29.9, and 30 kg/m²), WBC (continuous), smoking behavior (former, current, or never smoker), hypertension (yes or no), diabetes (yes or no), asthma (yes or no), and UACR (≥ 30 or <30 mg/g) or eGFR (≥ 60 or <60mL/min/1.73m²), and NHANES cycles (1999–2000, 2001–2002, 2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012, 2013–2014, 2015–2016, 2017–2018).

Table 4. Sensitivity analysis.

	Model 1	p value	Model 2	p value	Model 3	p value
Analysis after excluding extreme 5% of PEC [β per 100cells/uL increase in PEC (95% CI)]
eGFR	−3.03 (−3.87, −2.21)	< 0.001	−1.72 (−2.34, −1.10)	< 0.001	−1.46 (−2.00, −0.93)	< 0.001
UACR	36.58 (19.19, 53.97)	< 0.001	38.98 (21.84, 56.12)	< 0.001	23.28 (7.81, 38.75)	0.003
Analysis using PEC percentage [β per 1% increase in PEC percentage (95% CI)]
eGFR	−1.61 (−2.02, −1.20)	< 0.001	−0.61 (−0.89, −0.32)	< 0.001	−0.46 (−0.71, −0.21)	< 0.001
UACR	4.88 (−0.34, 10.09)	0.066	5.04 (−0.26, 10.34)	0.062	3.82 (−1.84, 9.48)	0.184

CI: confidence interval; eGFR: estimated glomerular filtration rate; PEC: peripheral eosinophil count; UACR: urinary albumin to creatinine ratio. Model 1 was unadjusted model; Model 2 was adjusted for age (continuous), sex (male or female), and race/ethnicity. Model 3 was further adjusted for BMI (<25, 25 ∼ 29.9, and 30 kg/m²), WBC (continuous), smoking behavior (former, current, or never smoker), hypertension (yes or no), diabetes (yes or no), asthma (yes or no), and UACR (≥ 30 or <30 mg/g) or eGFR (≥ 60 or <60mL/min/1.73m²), and NHANES cycles (1999–2000, 2001–2002, 2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012, 2013–2014, 2015–2016, 2017–2018).

Hattori K, Sakaguchi Y, Oka T, et al. Interstitial eosinophilic aggregates and kidney outcome in patients with CKD. Clin J Am Soc Nephrol. 2023;18(12):1563–1572. doi: 10.2215/CJN.0000000000000277.

 PubMed Web of Science ®Google Scholar

Kielar D, Jones AM, Wang X, et al. Association between elevated blood eosinophils and chronic kidney disease progression: analyses of a large United States electronic health records database. Int J Nephrol Renovasc Dis. 2023;16:269–280. doi: 10.2147/IJNRD.S431375.

 PubMedGoogle Scholar

Pongdee T, Manemann SM, Decker PA, et al. Rethinking blood eosinophil counts: epidemiology, associated chronic diseases, and increased risks of cardiovascular disease. J Allergy Clin Immunol Glob. 2022;1(4):233–240. doi: 10.1016/j.jacig.2022.09.001.

 PubMedGoogle Scholar

Yen CH, Wu IW, Lee CC, et al. The prognostic value of peripheral total and differential leukocyte count in renal progression: a community-based study. PLOS One. 2021;16(10):e0258210. doi: 10.1371/journal.pone.0258210.

 PubMed Web of Science ®Google Scholar

Kang DH, Lee Y, Kleine CE, et al. Eosinophil count and mortality risk in incident hemodialysis patients. Nephrol Dial Transplant. 2020;35(6):1032–1042. doi: 10.1093/ndt/gfz296.

 PubMed Web of Science ®Google Scholar

Yang J, Fan J, Fan L, et al. Higher eosinophils predict death-censored technique failure in peritoneal dialysis patients. Int Arch Allergy Immunol. 2020;181(10):765–773. doi: 10.1159/000509085.

 PubMed Web of Science ®Google Scholar

Tariq A, Okamato K, Tariq A, et al. Eosinophilia and risk of incident end stage kidney disease. BMC Nephrol. 2020;21(1):14. doi: 10.1186/s12882-020-1685-3.

 PubMedGoogle Scholar

Hildebrand S, Corbett R, Duncan N, et al. Increased prevalence of eosinophilia in a hemodialysis population: longitudinal and case control studies. Hemodial Int. 2016;20(3):414–420. doi: 10.1111/hdi.12395.

 PubMed Web of Science ®Google Scholar

Nakhjavani M, Aghajani Nargesi A, Salabati M, et al. Changes in leukocyte subpopulations with decline in glomerular filtration rate in patients with type 2 diabetes. Acta Med Iran. 2015l;53(7):425–431.

 PubMedGoogle Scholar

Ishii R, Fujita S, Kizawa S, et al. Association between ­absolute blood eosinophil count and CKD stages among cardiac patients. Heart Vessels. 2016;31(2):198–205. doi: 10.1007/s00380-014-0590-8.

 PubMed Web of Science ®Google Scholar

Agarwal R, Light RP. Patterns and prognostic value of total and differential leukocyte count in chronic kidney disease. Clin J Am Soc Nephrol. 2011;6(6):1393–1399. doi: 10.2215/CJN.10521110.

 PubMed Web of Science ®Google Scholar

Diskin CJ, Stokes TJ, Dansby LM, et al. The prevalence and meaning of eosinophilia in renal diseases on a ­nephrology consultation service. Nephrol Dial Transplant. 2011;26(8):2549–2558. doi: 10.1093/ndt/gfq745.

 PubMed Web of Science ®Google Scholar

Fukui M, Tanaka M, Hamaguchi M, et al. Eosinophil count is positively correlated with albumin excretion rate in men with type 2 diabetes. Clin J Am Soc Nephrol. 2009;4(11):1761–1765. doi: 10.2215/CJN.03330509.

 PubMed Web of Science ®Google Scholar

Data availability statement

The data used this study are openly available in NHANES at https://www.cdc.gov/nchs/nhanes/.