Developing Non-Laboratory Cardiovascular Risk Assessment Charts and Validating Laboratory and Non-Laboratory-Based Models

Razieh Hassannejad; Marjan Mansourian; Hamidreza Marateb; Mohammad Reza Mohebian; Thomas Andrew Gaziano; Rodney T Jackson; Emanuele Di Angelantonio; Nizal Sarrafzadegan

doi:10.5334/gh.890

Developing Non-Laboratory Cardiovascular Risk Assessment Charts and Validating Laboratory and Non-Laboratory-Based Models

Volume 16 (2021): Issue 1

By: Razieh Hassannejad, Marjan Mansourian, Hamidreza Marateb, Mohammad Reza Mohebian, Thomas Andrew Gaziano, Rodney T Jackson, Emanuele Di Angelantonio and Nizal Sarrafzadegan

Open Access

|Sep 2021

Figures & Tables

Table 1

Non- Laboratory-based and laboratory-based models for predicting cardiovascular disease outcomes, Isfahan Cohort Study, 2001–2011.

Risk factors	Estimate	Hazard Ratio	95% CI*
Non-laboratory-based (SPARS)
Age at baseline per 10 years	0.38227	1.466	1.371–1.566
Male	0.32309	1.381	1.076–1.774
WHR
Female: <0.85, male: <1
Female: 0.85-0.90, male: 1–1.05	0.10478	1.11	0.867–1.422
Female: 0.90–0.95, male: 1.05–1.10	0.17986	1.197	0.898–1.595
Female: >=0.95, male: >=1.10	0.2944	1.342	1.012–1.780
SBP (mm Hg)
<120^a
120–139	0.47524	1.608	1.316–1.966
140–159	0.79045	2.204	1.743–2.788
>=160	1.13051	3.097	2.405–3.989
Self-reported Diabetes	0.73001	2.075	1.697–2.537
Smoking	0.2656	1.304	1.079–1.577
Harrell’s C (95% CI):0.73 (0.71–0.74) Nam-D’Agostino χ²:11.01 (p = 0.27).
Laboratory-based (PARS)^‡
Age at baseline per 10 years	0.376	1.456	1.361–1.558
Male	0.28957	1.335	1.111–1.508
TC (mg/dl)
<150^a	–	–
150–200	0.20759	1.231	0.879–1.723
200–250	0.34201	1.408	1.013–1.957
250–300	0.45316	1.573	1.113–2.225
>300	0.54847	1.731	1.172–2.556
SBP (mm Hg)
<120^a	–	–
120–139	0.45643	1.578	1.291–1.929
140–159	0.73697	2.09	1.651–2.644
>=160	1.0467	2.848	2.207–3.676
Diabetes	0.63041	1.878	1.570–2.247
High WHR^b	0.26989	1.31	1.072–1.601
FH of CVD	0.40182	1.495	1.116–2.002
Smoking	0.28974	1.336	1.104–1.617
Harrell’s C (95% CI): 0.73 (0.71–0.75) Nam-D’Agostino χ²: 10.82 (p = 0.29)

[i] Abbreviations: CI, confidence interval; CVD, cardiovascular Disease; TC, total cholesterol; SBP, systolic blood pressure; FH, family history; WHR, waist to hip ratio.

^a Reference category.

^b WHR ≥ 0.95 in men and ≥ 0.8 in women was considered as a high WHR.

^* 95% CI for HR.

^‡ PARS risk model, published previously by Sarrafzadegan et al. [22].

SPARS chart for prediction of 10-year risk of fatal and non-fatal cardiovascular disease in ICS population, 2001–2011.

Patients correctly classified as a high or low risk at various cutoff levels of 10-year risk of cardiovascular disease.

Table 2

True positive and true negative based on PARS and SPARS models across different risk thresholds.

Thresholds	PARS		SPARS
Thresholds	True positive	True negative	True positive	True negative
5%	50.1% (353/705)	81.3% (3844/4727)	66.2% (467/705)	68.1 (3217/4727)
10%	17.6% (124/705)	96.5% (4561/4727)	31.6% (223/705)	90.4% (4273/4727)
15%	5.2% (37/705)	98.9% (4676/4727)	13.9% (98/705)	96.7% (4573/4727)

Table 3

Baseline Risk Factors, Person-Years of Follow-up, and CVD events in the ICS and TLGS Cohorts.

	ICS*						TLGS
	Men (N = 2648)			Women (N = 2784)			Men (N = 2306)			Women (N = 2976)
	Participants WithRisk Factor, %	Person-years of follow-up	CVD events	Participants WithRisk Factor, %	Person-years of follow-up	CVD events	Participants WithRisk Factor, %	Person-years of follow-up	CVD events	Participants WithRisk Factor, %	Person-years of follow-up	CVD events
Total		23931	379		25522	326		23812.9	464		32169.9	360
TC (mg/dl)
<150	11.4	2826.4	27	7.5	1917	15	5.3	1294.2	17	3	991.4	2
150–200	35	8507.1	112	30.4	7938.5	71	34.4	8375.3	125	26.9	8999	39
200–250	34.1	8144.3	131	36.4	9206.3	123	42.2	10136.2	193	40.9	13156.3	148
250–300	14.2	3253.8	79	18.6	4710.25	77	15.1	3369.5	104	22.3	7036.2	112
>300	5.2	1198.9	30	7.2	1749.8	40	3	637.6	25	6.9	1986.9	59
SBP(mm/Hg)
<120	44.4	11132.3	96	44.9	11652	67	48	11988.3	141	45.3	15165	85
120–139	37.1	8741.5	148	33.7	8746.2	116	33.9	8038	162	33.5	10835.4	105
140–159	12	2683	76	13.4	3347.7	78	12.6	2699.7	102	14.5	4337.4	105
>=160	6.6	1374.1	59	8	1776.1	65	5.6	1086.9	59	6.7	1832.1	65
Diabetes
Laboratory-based	9.3	2059.8	82	12.6	3040.7	87	15.4	3009.7	142	17.8	5022.5	155
Self-reporting	5.8	1289.8	51	8.6	1984.4	69	10.7	2085.3	96	12.5	3480.3	123
Smoker	41.6	9809.1	159	3.3	805.3	18	27.6	6487.9	120	3.8	1188.6	14
FH of CVD	5	1205.7	22	5.7	1415.8	27	15.6	3618.3	83	19.9	6188.4	101
WHR
1	86.3	20719.7	299	13.8	3673.1	25	79.1	19236.2	306	39.8	13312.5	61
2	9.3	2236.6	49	16.4	4103.2	44	15.8	3557.3	107	21.6	7010.7	77
3	3.3	714.1	25	24.4	6272.7	71	4.2	841.3	41	20.8	6590.4	97
4	1.1	260.5	6	45.4	11472.9	186	0.9	178.1	10	17.8	5256.2	125

[i] Abbreviations: ICS, Isfahan Cohort Study; TLGS, Tehran Lipid and Glucose Study; CVD, cardiovascular Disease; TC, total cholesterol; SBP, systolic blood pressure; FH, family history; WHR, waist to hip ratio.

* Data for the ICS risk factors are from Sarrafzadegan et al. [22].

Table 4

Relative risk and performance comparisons of ICS and TLGS based on non-laboratory SPARS risk function.

Risk factors	ICS		TLGS
Risk factors	Estimate	Hazard Ratio(95% CI)	Estimate	Hazard Ratio(95% CI)
Age	0.03823	1.039 (1.032–1.046)	0.04494	1.046 (1.039–1.053)
Male^†	0.32309	1.38 (1.076–1.774)	0.76677	2.153 (1.801–2.573)
WHR
1			–	–
2	0.10478	1.11 (0.867–1.422)	0.35549	1.427 (1.188–1.713)
3^†	0.17986	1.197 (0.898–1.595)	0.55846	1.748 (1.399–2.184)
4^†	0.2944	1.342 (1.012–1.780)	0.69904	2.012 (1.577–2.567)
SBP (mm/Hg)
<120^a			–	–
120-139	0.47524	1.608 (1.316–1.966)	0.21399	1.239 (1.029–1.490)
140-159	0.79045	2.204 (1.743–2.788)	0.70242	2.019 (1.643–2.481)
>=160	1.13051	3.097 (2.405–3.989)	0.84719	2.333 (1.825–2.982)
History of Diabetes	0.73001	2.075 (1.697–2.537)	0.88382	2.42 (2.064–2.837)
Smoking	0.2656	1.304 (1.079–1.577)	0.47038	1.601 (1.308–1.959)
	Harrell’s C (95% CI): 0.73 (0.71–0.74) Nam–D’Agostino χ²: 11.01 (p = 0.27)		Harrell’s C (95% CI): 0.77 (0.75–0.78) Nam–D’Agostino: 29.89 (p = 0.001)

[i] Abbreviations: ICS, Isfahan Cohort Study; TLGS, Tehran Lipid and Glucose Study; CVD, cardiovascular Disease; SBP, systolic blood pressure; WHR, waist to hip ratio.

^† Hazard ratio in the ICS is significantly different from that in the TLGS (P-value < 0.05).

Table 5

Relative risk and performance comparisons of ICS and TLGS based on laboratory-based PARS risk function.

Risk factors	ICS^*		TLGS
Risk factors	Estimate	Hazard Ratio(95% CI)	Estimate	Hazard Ratio(95% CI)
Age	0.03759	1.038 (1.031–1.045)	0.04592	1.047 (1.04–1.054)
Male ^†	0.28957	1.335 (1.111–1.508)	0.71764	2.05 (1.748–2.403)
TC (mg/dl)
<150^a	–	–	–	–
150–200	0.20759	1.231 (0.879–1.723)	0.00956	1.01 (0.627–1.625)
200–250	0.34201	1.408 (1.013–1.957)	0.41855	1.52 (0.956–2.417)
250–300	0.45316	1.573 (1.113–2.225)	0.55737	1.746 (1.087–2.804)
>300	0.54847	1.731 (1.172–2.556)	0.95743	2.605 (1.571–4.321)
SBP (mm/Hg)
<120^a	–	–	–	–
120–139^†	0.45643	1.578 (1.291–1.929)	0.15221	1.164 (0.967–1.402)
140–159	0.73697	2.09 (1.651–2.644)	0.60783	1.836 (1.493–2.259)
>=160	1.0467	2.848 (2.207–3.676)	0.74208	2.1 (1.643–2.684)
Diabetes	0.63041	1.878 (1.570–2.247)	0.79142	2.207 (1.899–2.564)
High WHR	0.26989	1.31 (1.072–1.601)	0.31902	1.376 (1.154–1.640)
FH of CVD	0.40182	1.495 (1.116–2.002)	0.38066	1.463 (1.240–1.727)
Smoking	0.28974	1.336 (1.104–1.617)	0.49554	1.641 (1.341–2.010)
	Harrell’s C (95% CI): 0.73 (0.71–0.75) Nam–D’Agostino χ²: 10.82 (p = 0.29)		Harrell’s C (95% CI): 0.78 (0.76–0.79) Nam–D’Agostino χ²: 28.57 (p = 0.001)

* Data for the ICS risk factors are from Sarrafzadegan et al. [22].

^† Hazard ratio in the ICS is significantly different from that in the TLGS (P-value < 0.05).

Actual and predicted cardiovascular disease events by deciles of risk for the laboratory and non-laboratory-based model in ICS and TLGS.

References

Authors

Metrics

Articles in this issue

DOI: https://doi.org/10.5334/gh.890 | Journal eISSN: 2211-8179

Journal RSS Feed

Language: English

Submitted on: Jul 28, 2020

Accepted on: Aug 20, 2021

Published on: Sep 2, 2021

Published by: Ubiquity Press

In partnership with: Paradigm Publishing Services

Publication frequency: 1 issue per year

Keywords:

Cardiovascular Disease,

risk assessment,

Laboratory-based model,

Non-laboratory-based model,

Isfahan Cohort Study

© 2021 Razieh Hassannejad, Marjan Mansourian, Hamidreza Marateb, Mohammad Reza Mohebian, Thomas Andrew Gaziano, Rodney T Jackson, Emanuele Di Angelantonio, Nizal Sarrafzadegan, published by Ubiquity Press
This work is licensed under the Creative Commons Attribution 4.0 License.

Volume 16 (2021): Issue 1

Developing Non-Laboratory Cardiovascular Risk Assessment Charts and Validating Laboratory and Non-Laboratory-Based Models

Figures & Tables

Table 1

Figure 1

Figure 2

Table 2

Table 3

Table 4

Table 5

Figure 3

Paradigm

My account