A nomogram prediction model for mild cognitive impairment in non-dialysis outpatient patients with chronic kidney disease

Figures & data

Figure 1. Flowchart recruitment process of the study population.

Table 1. Clinical characteristics of non-dialysis outpatient patients with CKD with or without MCI.

Variables	Total (n = 416)	MCI group (n = 210)	Non-MCI group (n = 206)	Statistic	p Value
Age/years				χ^2 =172.67	<.001
≥65	208(50.0)	172(81.9)	36(17.5)
<65	208(50.0)	38(11.1)	170(82.5)
Sex				χ^2=7.521	.006
Male	204(49.0)	89(42.4)	115(54.8)
Female	212(51.0)	121(57.6)	91(44.2)
Educational level middle school and below				χ^2=71.254	<.001
	310(74.5)	194(92.3)	116(56.3)
High school/secondary school and above	106(25.5)	16(7.7)	90(43.7)
Occupation status				χ^2=146.813	<.001
Retired/unemployed	269(64.7)	194(92.3)	75(36.4)
Employed	147(35.3)	16(7.7)	131(63.6)
Spouse				χ^2=9.940	.002
Yes	357(85.8)	169(80.5)	188(91.3)
No	59(14.2)	41(19.5)	18(8.7)
CKD stage				χ^2=14.844	.005
1	37(8.9)	9(4.3)	28(13.6)
2	60(14.4)	27(12.9)	33(16.0)
3	150(36.1)	76(36.2)	74(35.9)
4	80(19.2)	45(21.4)	35(17.0)
5	89(21.4)	53(25.2)	36(17.5)
Exercise				χ^2=68.149	<.001
Never/occasionally	302(72.6)	190(90.5)	112(54.4)
Frequently	114(27.4)	20(9.5)	94(45.6)
Hypertension				χ^2=23.167	<.001
Yes	308(74.0)	177(84.3)	131(63.6)
No	108(26.0)	33(15.7)	75(36.4)
Diabetes				χ^2=42.258	<.001
Yes	147(35.3)	107(51.0)	40(19.4)
No	269(64.7)	103(49.0)	166(80.6)
Coronary heart disease				χ^2=13.487	<.001
Yes	66(84.1)	47(22.4)	19(9.2)
No	350(15.9)	163(77.6)	187(90.8)
Meat and vegetable collocation				χ^2=32.246	<.001
Balance	346(83.2)	153(72.9)	193(93.7)
Unbalance	70(16.8)	57(27.1)	13(6.3)
Reading books and newspapers				χ^2=23.380	<.001
Yes	41(9.9)	6(2.9)	35(83.0)
No	375(90.1)	204(97.1)	171(17.0)
Chess and cards				χ^2=37.855	<.001
Yes	234(56.3)	87(41.4)	147(71.4)
No	182(43.8)	123(58.6)	59(28.6)
Use of smartphone				χ^2=134.751	<.001
Yes	290(69.7)	92(43.8)	198(96.1)
No	126(30.3)	118(56.2)	8(3.9)
Primary disease				χ^2=55.734	<.001
Chronic glomerulonephritis	170(40.9)	55(26.2)	115(55.8)
Hypertensive nephropathy	83(20.0)	48(22.9)	35(17.0)
Diabetic nephropathy	95(22.8)	74(35.2)	21(10.2)
Polycystic kidney	11(2.6)	7(3.3)	4(1.9)
Lupus nephritis	2(0.5)	0(0.0)	2(1.0)
Others	55(13.2)	26(12.4)	29(14.1)
Depression				χ^2=24.328	<.001
Yes	83(20.0)	62(29.5)	21(10.2)
No	333(80.0)	148(70.5)	185(89.8)
Sleep disorders				χ^2=59.646	<.001
Yes	184(44.2)	132(62.9)	52(25.2)
No	232(55.8)	78(37.1)	154(74.8)
Total protein (g/L)	69.55(64.18, 74.98)	69.01(64.00, 74.80)	70.15(64.25, 75.45)	Z=−0.527	.598
Albumin (g/L)	40.87(36.40, 44.65)	39.50(35.60, 44.05)	42.65(37.70, 45.10)	Z=−3.241	.001
Triglyceride (mmol/L)	1.60(1.17, 2.47)	1.56(1.16, 2.47)	1.69(1.17, 2.47)	Z=−0.830	.407
Total cholesterol (mmol/L)	4.63(3.82, 5.36)	4.50(3.57, 5.26)	4.81(4.05, 5.61)	Z=−3.276	.001
HDL (mmol/L)	1.33(1.09, 1.78)	1.33(1.10, 1.75)	1.340(1.08, 1.80)	Z=−3.940	.694
LDL (mmol/L)	2.62(1.99, 3.23)	2.41(1.93, 2.96)	2.74(2.20, 3.41)	Z=−3.230	.001
eGFR (ml/min/1.73m^2)	38.51(17.24, 57.53)	31.68(14.53, 50.84)	43.73(21.39, 70.08)	Z=−3.437	.001
Urea (mmol/L)	11.41(7.39, 18.95)	12.79(9.15, 21.24)	9.89(6.06,16.00)	Z=−4.199	<.001
Uric acid (µmol/L)	393.95(323.00, 465.75)	391.80(309.50, 458.30)	402.50(332.50, 482.40)	Z=−1.644	.100
Cystatin C (mg/L)	2.30(1.51, 4.00)	2.55(1.71, 4.16)	2.02(1.25, 3.48)	Z= −3.309	.001
Potassium (mmol/L)	4.34(3.96, 4.81)	4.37(3.93, 4.80)	4.34(4.01, 4.83)	Z=−0.380	.704
Calcium (mmol/L)	2.270(2.13, 2.41)	2.26(2.10, 2.41)	2.30(2.15, 2.40)	Z=−1.552	.121
Sodium (mmol/L)	141.00(139.00, 143.40)	140.90(138.30, 143.40)	141.30(139.30, 143.40)	Z=−1.301	.193
Phosphorus (mmol/L)	1.21(1.02, 1.38)	1.18(1.02, 1.35)	1.23(1.05, 1.42)	Z=−1.174	.240
WBC (10^9/L)	6.95(5.56, 8.32)	6.81(5.31, 8.21)	7.10(5.76, 8.59)	Z=−1.806	.071
Hemoglobin (g/L)	117.54 ± 24.70	109.64 ± 23.29	125.59 ± 23.51	t = 6.954	<.001

Note: n: number of patients; HDL: high-density lipoprotein; LDL: low-density lipoprotein; eGFR: estimation of glomerular filtration rate; WBC: white blood cells.

Table 2. Comparison of total MoCA score and scores of each dimension in non-dialysis outpatient patients with CKD [M (P25, P75)].

Items	MCI group (n = 210)	Non-MCI group (n = 206)	Z value	p Value
Visuospatial/executive	2(1,3)	4(4,5)	−15.647	<.001
Naming	3(2,3)	3(3,3)	−7.369	<.001
Attention	4(3,5)	6(6,6)	−14.973	<.001
Language	2(1,2)	3(2,3)	−13.062	<.001
Abstraction	0(0,1)	2(1,2)	−13.129	<.001
Delayed recall	2(2,3)	4(4,5)	−13.885	<.001
Orientation	2(2,3)	4(4,5)	−12.674	<.001
Total score	19(15,23)	28(27,29)	−17.556	<.001

Note: n: number of patients.

Table 3. Univariate binary logistic regression analyses of MCI in non-dialysis outpatient patients with CKD.

Variables	β	SE	Wald χ^2	OR	95%CI	p Value
Age (years)						<.001
<65 years	–	–	–	1.00 (ref)	–
≥65 years	3.026	0.256	142.529	21.374	12.929–35.36
Sex						.006
Male	–	–	–	1.00 (ref)	–
Female	0.541	0.198	7.475	1.718	1.166–2.532
Spouse						.002
No	–	–	–	1.00 (ref)	–
Yes	−0.930	0.302	9.480	0.395	0.218–0.713
Educational level						<.001
Junior high school and below	–	–	–	1.00 (ref)	–
Senior high school / technical	−2.241	0.117	57.495	0.106	0.060–0.190
Secondary school and above
Occupation status						<.001
Unemployed/retired	–	–	–	1.00 (ref)	–
Employed	−3.053	0.298	105.174	0.047	0.026–0.085
Exercise						<.001
Never / occasionally	–	–	–	1.00 (ref)	–
Often	−2.076	0.274	57.599	0.125	0.073–0.214
Sleep disorders						<.001
No	–	–	–	1.00 (ref)	–
Yes	1.612	0.215	56.329	5.012	3.290–7.635
Depression						<.001
No	–	–	–	1.00 (ref)	–
Yes	0.136	0.267	4.094	3.690	2.151–6.333
Hypertension						<.001
No	–	–	–	1.00 (ref)	–
Yes	1.122	0.239	22.114	3.071	1.924–4.901
Diabetes						<.001
No	–	–	–	1.00 (ref)	–
Yes	1.461	0.224	42.636	4.311	2.780–6.685
Coronary heart disease						<.001
No	–	–	–	1.00 (ref)	–
Yes	1.043	0.292	12.741	2.838	1.601–5.032
Meat and vegetable collocation						<.001
Unbalance	–	–	–	1.00 (ref)	–
Balance	−1.710	0.326	27.549	0.108	0.095–0.342
Reading books and newspapers						<.001
No	–	–	–	1.00 (ref)	–
Yes	−1.940	0.454	18.272	0.144	0.059–0.350
Chess and cards						<.001
No	–	–	–	1.00 (ref)	–
Yes	−1.259	0.208	36.553	0.284	0.189–0.427
Use of smartphone						<.001
No	–	–	–	1.00 (ref)	–
Yes	−3.485	0.365	80.092	0.032	0.014–0.067
Albumin (g/L)	−0.029	0.014	4.308	0.971	0.945–0.998	.038
Total cholesterol (mmol/L)	−0.238	0.075	10.095	0.788	0.681–0.913	.001
LDL (mmol/L)	−0.236	0.093	6.495	0.790	0.685–0.947	.011
Urea (mmol/L)	0.024	0.010	5.713	1.024	1.004–1.044	.017
Cystatin C (mg/L)	−0.001	0.003	0.322	0.999	0.993–1.004	.570
Hemoglobin (g/L)	−0.30	0.005	39.115	0.971	0.962–0.980	<.001
eGFR (ml/min/1.73m^2)	−0.008	0.003	6.094	0.992	0.986–0.998	.014
CKD stage
1	–	–	13.724	1.00 (ref)	–	.008
2	0.0934	0.463	4.076	2.545	1.028–6.305	.043
3	1.162	0.417	7.778	3.195	1.412–7.228	.005
4	1.386	0.445	9.725	4.000	1.674–9.560	.002
5	1.552	0.440	11.969	4.580	1.934–10.847	.003
Primary disease
Chronic glomerulonephritis	–	–	49.174	1.00 (ref)	–	<.001
Hypertensive nephropathy	1.053	0.276	14.548	2.868	1.669–4.297	<.001
Diabetic nephropathy	1.997	0.297	45.320	7.368	4.119–13.179	<.001
Polycystic kidney	1.297	0.648	4.009	3.659	1.028–13.027	.045
Lupus nephritis	−20.465	28,420.722	<0.001	0.000	0.000–	.999
Others	0.628	0.316	3.956	1.875	1.009–3.482	.47

Notes: ref: reference; LDL: low-density lipoprotein; eGFR: estimation of glomerular filtration rate. Independent variable assignment: age: < 65 years old =0, ≥65 years old = 1; gender: male =0, female = 1; spouse: none =0, yes = 1; education level: junior high school and below =0, senior high school/technical secondary school and above = 1; occupational status: retired or unemployed =0, employed = 1; exercise situation: never or occasionally =0, often = 1; sleep disorders: no =0, yes = 1; depression: no =0, yes = 1; hypertension: no =0, yes = 1; diabetes: no =0, yes = 1; coronary heart disease: no =0, yes =1; meat and vegetable combinations: unbalance =0, balance = 1; reading books and newspapers: no =0, yes = 1; chess and cards: no =0, yes = 1; use of smartphone: no =0, yes = 1; CKD stages: 1 = 0,2 = 1, 3 = 2,4 = 3,5 = 4; primary diseases: chronic glomerulonephritis =0, hypertensive nephropathy =1, diabetic nephropathy =2, polycystic kidney disease =3, lupus nephritis =4, others =5. The regression analysis takes ‘0’ as the reference.

Table 4. multivariate binary logistic regression analysis of MCI in non-dialysis outpatient patients with CKD.

Variables	β	SE	Wald χ^2	OR	95%CI	p Value
Age (years)						.004
<65 years	–	–	–	1.00 (ref)	–
≥65 years	1.177	0.410	8.257	3.244	1.454–7.240
Sex						.901
Male	–	–	–	1.00 (ref)	–
Female	−0.045	0.363	0.016	0.956	0.469–1.948
Spouse						.991
No	–	–	–	1.00 (ref)	–
Yes	0.007	0.599	0.000	1.007	0.311–3.254
Educational level						.007
Junior high school and below	–	–	–	1.00 (ref)	–
Senior high school / technical	−1.211	0.446	7.365	0.298	0.124–0.714
Secondary school and above
Occupation status						<.001
Unemployed/retired	–	–	–	1.00 (ref)	–
Employed	−1.684	0.434	15.020	0.186	0.079–0.435
Exercise						.157
Never / occasionally	–	–	–	1.00 (ref)	–
Often	−0.603	0.426	2.006	0.547	0.238–1.260
Sleep disorders						.004
No	–	–	–	1.00 (ref)	–
Yes	1.050	0.366	8.255	2.859	1.396–5.853
Depression						.378
No	–	–	–	1.00 (ref)	–
Yes	−0.467	0.530	0.778	0.627	0.222–1.770
Hypertension						.401
No	–	–	–	1.00 (ref)	–
Yes	−0.360	0.429	0.704	0.698	0.301–1.617
Diabetes						.514
No	–	–	–	1.00 (ref)	–
Yes	0.292	0.447	0.427	1.339	0.558–3.216
Coronary heart disease						.669
No	–	–	–	1.00 (ref)	–
Yes	0.197	0.292	0.183	1.218	1.601–5.032
Meat and vegetable collocation						.203
Unbalance	–	–	–	1.00 (ref)	–
Balance	−0.688	0.540	1.622	0.502	0.174–1.449
Reading books and newspapers						<.001
No	–	–	–	1.00 (ref)	–
Yes	−1.128	0.717	2.473	0.324	0.059–0.350
Chess and cards						.559
No	–	–	–	1.00 (ref)	–
Yes	0.212	0.362	0.342	1.236	0.608–2.514
Use of smartphone						<.001
No	–	–	–	1.00 (ref)	–
Yes	−2.358	0.525	20.210	0.095	0.034–0.264
Albumin (g/L)	0.000	0.003	0.001	1.000	0.994–1.006	.975
Total cholesterol (mmol/L)	−0.283	0.307	0.849	0.754	0.413–1.375	.357
LDL (mmol/L)	0.147	0.385	0.145	1.158	0.544–2.465	.703
Urea (mmol/L)	0.001	0.003	0.073	1.001	0.995–1.006	.787
Hemoglobin (g/L)	−0.020	0.009	5.013	0.980	0.964–0.998	.025
CKD stage
1	–	–	5.404	1.00 (ref)	–	.248
2	0.855	0.818	1.092	2.351	0.473–11.685	.296
3	0.289	0.735	0.154	1.334	0.316–5.631	.694
4	−0.544	0.800	0.462	0.581	0.121–2.787	.497
5	−0.114	0.817	0.019	0.892	0.180–4.422	.889
Primary disease
Chronic glomerulonephritis	–	–	0.877	1.00 (ref)	–	.972
Hypertensive nephropathy	−0.358	0.471	0.578	0.699	0.278–1.760	.447
Diabetic nephropathy	−0.228	0.569	0.160	0.797	0.261–2.428	.689
Polycystic kidney	0.212	1.159	0.033	1.236	0.128 – 11.980	.855
Lupus nephritis	−18.217	28,245.945	0.000	0.000	0.000–	.999
Other	0.107	0.577	0.034	1.113	0.359–3.446	.853
Constant	5.749	1.770	10.547	313.943	–	.001

Notes: ref: reference; LDL: low-density lipoprotein. The independent variables are assigned the same values as in the univariate binary logistic regression analysis.

Figure 2. Nomogram risk prediction of MCI in non-dialysis outpatient patients with CKD.

Figure 3. Web calculator based on dynamic nomogram of MCI in non-dialysis outpatient patients with CKD.

Table 5. The final multivariate binary logistic regression analysis.

Variables	β	SE	Wald χ2	OR	95%CI	p Value
Age						.002
<65 years	–	–	–	1.00 (ref)	–
≥65 years	1.079	0.347	9.668	2.941	1.490–5.805
Educational level						<.001
Junior high school and below	–	–	–	1.00 (ref)	–
Senior high school / technical	−1.405	0.403	12.130	0.245	0.111–0.541
Secondary school and above
Occupational status						<.001
Unemployed/retired				1.00 (ref)
Employed	−1.590	0.382	17.362	0.204	0.097–0.431
Sleep disorders						.001
No	–	–	–	1.00 (ref)	–
Yes	0.977	0.305	10.275	2.658	1.462–4.831
Use of smartphone						<.001
No	–	–	–	1.00 (ref)	–
Yes	−2.348	0.470	24.935	0.096	0.038–0.240
Hemoglobin	−0.017	0.007	6.385	0.984	0.971–0.996	.012
Constant	3.601	1.001	12.939	36.625	–	<.001

Notes: ref: reference. Independent variable assignment: age: < 65 years old =0, ≥65 years old = 1; –education level: junior high school and below =0, senior high school/technical secondary school and above = 1; occupational status: retired or unemployed =0, employed = 1; sleep disorders: no =0, yes = 1; use of smartphone: no =0, yes = 1; the regression analysis takes ‘0’ as the reference.

Figure 4. Restricted cubic spline diagram of HB and MCI risk.

Figure 5. The ROC curve analysis.

Figure 6. The calibration plot.

Figure 7. The DCA analysis.

Table 6. Predictive performance of the model.

AUC (95% CI)	Sensitivity	Specificity	Accuracy	PPV	NPV	F1-score	LR+	LR−
0.926(0.902, 0.951)	0.767	0.932	0.849	0.920	0.797	0.836	11.5	0.255

Note: AUC: area under the receiver operating characteristic curve; NPV: negative predictive value; PPV: positive predictive value; LR+: positive likelihood ratio; LR−: negative likelihood ratio.