Variations in how medical researchers report variables in risk scores or models to predict prognosis of patients after percutaneous coronary intervention: a retrospective analysis of published articles

doi:10.21203/rs.3.rs-4939290/v1

Download PDF

Research Article

Variations in how medical researchers report variables in risk scores or models to predict prognosis of patients after percutaneous coronary intervention: a retrospective analysis of published articles

https://doi.org/10.21203/rs.3.rs-4939290/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background: The use of risk variables in prognostic risk scores/models to evaluate patients after percutaneous coronary intervention (PCI) has been a controversial topic in medical literature. We therefore analyzed variations in risk scores/model variables to assess the prognosis of patients after percutaneous coronary intervention (PCI) in detail.

Methods: Articles were included from inception to December 2023 in PubMed/MEDLINE database using a combination of key words "Risk score" or "Risk model" AND "Percutaneous coronary intervention" (n=822). All English-language articles involving risk scores or models for assessment of patient prognosis after PCI were retained (n=183). We collected information on the extracted risk scores/models for patients after PCI (n=138) from the included articles and analyzed the variation variables in the relevant risk scores/models in detail.

Results: Among the risk scores/models, age, kidney function index, ACS presentation, diabetes, LVEF, culprit coronary artery, heart failure, SBP, heart rate, and sex were the top ten variables used. There were statistically significant differences in the use of variables such as kidney function index (χ²=6.995, P=0.008), ACS presentation (χ²=9.611, P=0.002), culprit coronary artery (χ²=3.937, P=0.047), SBP (χ²=10.556, P=0.001), heart rate (χ²=10.704, P=0.001), and ST-segment deviation (χ²=11.489, P=0.001) between Caucasian participants (n=74) and non-Caucasian participants (n=58).

Conclusions: in the risk scores/models for prognostic assessment after PCI. When constructing scores/models, the variable selection should fully consider the ethnic background of the study population.

PCI

scores

models

variable

prognosis

Percutaneous coronary intervention (PCI) is one of the most common types of coronary heart disease (CHD) worldwide.¹ At present, and significant progress has been made in the development of related technologies and devices. Despite this, there is still a significant increase in major adverse cardiac and cerebrovascular events (MACCE) after PCI.^2–5 Further improving the prognosis of patients after PCI remains a major challenge for clinicians. Therefore, it is necessary to evaluate the prognostic risk of patients after PCI and to take appropriate measures to guide clinical treatment.

Several risk scores/models can be used to predict the short- and/or long-term prognosis after PCI, such as the GRACE risk score,^5,6 TIMI risk score,⁷ SYNTAX score,⁸ EuroSCORE,^8,9 Mayo Clinic Risk Score,¹⁰ etc. Nevertheless, clinicians and researchers face the basic question of selecting variables when developing scores/models. Are the variables in one score or model appropriate for another patient population? What are the generalities and particularities of the variable characteristics in these risk scores/models?

Researchers parse diversity differently with different expectations and requirements for variables in scores or models. To clarify these issues, we conducted a detailed survey of articles related to the prognostic assessment of risk scores/models for patients after PCI in various professional journals, paying particular attention to the application of variables in relevant scores/models under different scenarios, which is a novel contribution to the literature.

2.1 Search strategy and selection criteria

We performed a retrospective analysis of articles published between inception and December 2023 in the PubMed/MEDLINE database. Articles were retrieved using a combination of the keywords "Risk score" or "Risk model" AND "Percutaneous coronary intervention.”

All English language articles involving risk scores or model development, validation, and/or comparison for the assessment of patient prognosis after PCI were included. The type of CHD was not specified, and included stable angina pectoris, unstable angina pectoris, non-ST-segment elevation myocardial infarction (NSTEMI), and ST-segment elevation myocardial infarction (STEMI). Prognostic outcomes were defined as 1) unplanned readmission, 2) mortality, 3) major adverse cardiac events (MACE), or 4) MACCE. The duration of follow-up was not limited, including short-term (in-hospital, 30-day, or 6 months) and long-term (≥ 1 year). The exclusion criteria were as follows: 1) duplicate literature; 2) abstracts, editorial comments, letters, corrections, expert consensus or guidelines, meta-analyses, and reviews; 3) the study population was mixed with patients without PCI; and 4) irrelevant to the study purpose.

2.2 Data analysis

Two researchers independently performed the literature search, screened the full-text articles, selected the risk scores or models, and extracted variables and other data. There was a good concordance coefficient between researchers (kappa index = 0.91). Discrepancies were discussed independently with the corresponding author until a consensus was reached.

We extracted risk scores or models from the included articles. For each risk score/model, we analyzed what date about age, sex/gender, smoking, body mass index (BMI), heart rate, systolic blood pressure (SBP), left ventricular ejection fraction (LVEF), Killip class, kidney function index (serum creatinine, estimated glomerular filtration rate (eGFR), chronic kidney disease (CKD), acute kidney injury (AKI), kidney insufficiency, and dialysis), acute coronary syndrome (ACS, including acute myocardial infarction (AMI), unstable angina) presentation, culprit coronary artery (including left main (LM) coronary artery, multivessel disease, and/or chronic total occlusion (CTO)), hypertension, chronic lung disease (CLD, including chronic obstructive pulmonary disease (COPD)), diabetes, heart failure (HF), peripheral vascular disease (PVD), cerebrovascular disease (including TIA), prior PCI, prior coronary artery bypass grafting (CABG), cardiogenic shock, cardiac arrest, TIMI flow grade, elevated cardiac markers (myocardial enzyme and/or cardiac troponin I/T), Brain natriuretic peptide (BNP)/N-terminal proBNP (NT-proBNP), and/or ST-segment deviation. Each article was coded to indicate whether the variables were reported. We also collected information on the year of publication, journal, country, or region of the study population, and names of risk scores/models for each article.

We calculated the percentage of articles that reported or did not report the names of the risk scores/models, the proportions of variables used in the prognostic risk scores/models of patients after PCI based on race (Caucasian participants or not), and the percentage of time each variable (age, sex, sex, smoking, BMI, heart rate, SBP, LVEF, Killip class, kidney function index, ACS presentation, culprit coronary artery, hypertension, CLD, diabetes, HF, PVD, CVD, prior PCI, prior CABG, cardiogenic shock, cardiac arrest, TIMI flow grade, elevated cardiac markers, BNP/NT-proBNP, and/or ST-segment deviation) were used. We also counted the number of scores/models constructed based on geography and the countries in which study participants were located. Chi-square tests for each index comparison were performed using SPSS Statistics (version 26.0; IBM, Armonk, NY, USA).

The initial literature search yielded a total of 822 articles. After removing 52 duplicates and 13 non-full-text articles, 757 titles and abstracts were screened. After a careful review, 574 articles were excluded. Finally, 183 articles were included in the analysis. The literature search process and the reasons for exclusion are shown in Fig. 1.

All included articles were published after 2000. The distribution of the publication year of the included articles is shown in Fig. 2. Over the past two decades, the largest number of related articles were published in 2021 (24 articles, 13.1%). Of the 183 included articles, 75 (41.0%) were registry studies, and the study population covered more than 26 countries and regions (Fig. 3). Among them, the largest number of related articles were published with the Chinese study population (44 articles, 24.0%), followed by articles with the USA study population (39 articles, 21.3%), and the Japanese study population (14 articles, 7.7%). The study populations of the UK, Italy, and Turkey all published 9 articles, the Netherlands published 8 articles, and South Korea published 6 articles. Moreover, except for seven international study populations involving multiple countries, other countries and regions published fewer than five articles each.

The top 10 journals that published articles on prognostic risk scores/models for patients after PCI were Catheterization and cardiovascular interventions (14 articles), The American journal of cardiology (13 articles), International Journal of Cardiology (11 articles), Frontiers in Cardiovascular Medicine (8 articles), Journal of the American College of Cardiology (JACC, 8 articles), JACC: Cardiovascular Interventions (8 articles), American Heart Journal (8 articles), BMC Cardiovascular Disorders (7 articles), Circulation: Cardiovascular Interventions (7 articles), and PLoS One (5 articles). Other journals published fewer than 5 articles each. Among the 183 selected articles, 138 risk scores/models were constructed, validated, modified, and/or compared and 29.0% (40/138) were unnamed (Table 1).

Table 1

Prognostic risk scores/model for study populations after PCI in different countries or regions
Country of study populations	Number of articles	Examples of risk scores or models involved
China	44	ACEF score, Blood Routine Test Parameters Score, Bio-C-CACS risk score, CHA2DS2-VASc Score, EPICOR score, modified GRACE risk score, HAMIOT score, Mehran risk score, MELD-XI, NERS score, PARIS thrombotic risk score, PMI score, SCAMI-NSTEMI score, SYNTAX score, SYNTAX score II, triglyceride-glucose index (TyG index), Unnamed-Risk score, Unnamed-Risk Model
USA	39	CHIP score, Ischemic Risk Model, Framingham Risk Score, Intermountain Risk Scores, LASH score, Mayo Clinic Risk Score, New Mayo Clinic Risk Score, NCDR risk score, PAMI risk score, New York State multivariate risk score, PROGRESS-CTO MACE scores, STEMI shock index, CADILLAC Risk Score, NCDR CathPCI beside risk score, TIMI risk score, Clinical Outcomes Assessment Program model, Unnamed-Risk score, Unnamed-Risk Model
Japan	14	Modified ACEF score, Japan-CTO extension score, Geriatric Nutritional Risk Index
Italy	9	ACEF score, EuroSCORE, Global Risk Classification, Mehran CIN-Risk score, PARIS thrombotic risk score, Unnamed-Risk score, Unnamed-Risk Model
Turkey	9	Modified ACEF score, CHA2DS2-VASc score, ATRIA risk score, PRECISE-DAPT Score, Intermountain Risk Score, SYNTAX score, SYNTAX score II, TIMI risk score, Zwolle risk score
UK	9	ATI score, Mayo Clinic Risk Score, New York PTCA score, PPCI risk score, Unnamed-Risk score, Unnamed-Risk Model
Netherlands	8	Multimarker score, Unnamed-Risk score, Unnamed-Risk Mode, Logistic Clinical SYNTAX score, Clinical SYNTAX score, modified Zwolle risk score
International*	7	ACUITY-PCI risk score, ABC-ACS ischemia score, GRACE risk score, GRACE risk score 2.0, Updated logistic clinical SYNTAX score, STSm score, EuroScore, the logistic Euroscore, Mayo Clinic risk score, the New York State risk score, Unnamed-Risk Model
South Korea	6	Mehran risk score, MVD risk model, TIMI risk score, modified TIMI risk score, KAMIR Score, KorMI Score,
Germany	4	EuroHeart score, EuroSCORE, clinical SYNTAX score, logistic EuroSCORE, Unnamed–Risk score
Poland	4	ANIN risk score, modified TIMI risk score, Parsonnet risk scores, SYNTAX Score
Spain	4	GRACE score, Zwolle risk score, Unnamed-Risk score
Australia	3	Zwolle risk score, CADILLAC risk score, TIMI risk score, and GRACE risk score; Unnamed-Risk Model
Brazil	3	DESIRE Score, Toronto score, Unnamed-Risk score
Israel	3	TRS2°P risk score, CathPCI Registry Risk Score, Unnamed-Risk model
Canada	2	British Columbia PCI risk score, Toronto score
Pakistan	2	TIMI risk score
Portugal	2	NCDR risk score, Mayo Clinic Risk Score, Zwolle risk score
Serbia	2	Echocardiographic Parameters, SYNTAX score, GRACE score, TIMI risk score, Zwolle risk score, CADILLAC risk score, PAMI risk scores
Belgrade	1	RISK-PCI score
France	1	ORBI risk score
Greece	1	ACEF score, EuroSCORE, EuroSCORE II, SYNTAX score, Global Risk Score, and Clinical Sxscore
Hungary	1	ALPHA score
Indonesia	1	New Mayo Clinic Risk Score
Mexico	1	TIMI risk score
New Zealand	1	GRACE score
Switzerland	1	MI SYNTAX score
Thailand	1	GRACE score
*The study population included participants from three or more countries.

The selection of variables varied across all 138 risk scores/models (Fig. 4). Most of them (84.8%, 117/138) used age as a variable, 55.8% (77/138) used kidney function index as a variable, 41.3% (57/138) used ACS presentation, and 35.5% (49/138) used diabetes. LVEF was 32.6% (45/138), culprit coronary artery was 31.9% (44/138), HF 24.6% (34/138), and SBP 23.9% (33/138). The next variables in order of application frequency from high to low were heart rate (31 times), sex/gender (29 times), cerebrovascular disease (27 times), Killip class (25 times), ST-segment deviation (25 times), PVD (23 times), hypertension (22 times), prior CABG (19 times), cardiogenic shock (18 times), smoking status (17 times), CLD (16 times), elevated cardiac markers (15 times), cardiac arrest (14 times), TIMI flow grade (13 times), body mass index (BMI, 12 times), prior PCI (11 times), and BNP/NT-proBNP (10 times). The remaining variables appeared less than 10 times.

Of the 138 risk scores/models, six studies were based on international populations, involving multiple countries and ethnicities. The remaining 132 scores/models were divided into two groups according to the study population: Caucasian participant group (n = 74) and non-Caucasian participant group (n = 58). Variables such as age (89.2% vs. 81.0%), kidney function index (84.1% vs. 31.0%), ACS presentation (44.6% vs. 19.0%), diabetes (39.2% vs. 29.3%), LVEF (39.2% vs. 25.9%), culprit coronary artery (39.2% vs. 29.3%), sex/gender (24.3% vs. 17.2%), CVD (17.6% vs. 17.2%), PVD (18.9% vs. 12.1%), hypertension (18.9% vs. 12.1%), cardiogenic shock (17.6% vs. 6.9%), smoking (12.2% vs. 12.1%), chronic lung disease (14.9% vs. 6.9%), cardiac arrest (10.8% vs. 8.6%), TIMI flow grade (10.8% vs. 5.2%), BMI (9.5% vs. 8.6%), and prior PCI (8.1% vs. 5.2%) were more commonly used in the Caucasian group. Variables such as HF (32.8% vs. 20.3%), SBP (37.9% vs. 13.5%), heart rate (36.2% vs. 16.2%), Killip class (20.7% vs. 14.9%), ST-segment deviation (30.0% vs. 8.1%), prior CABG (19.0% vs. 10.8%), elevated cardiac markers (12.1% vs. 5.4%), and the BNP/NT-proBNP ratio (6.9% vs. 6.8%) were more commonly used in the non-Caucasian group (Table 2). Statistically significant differences were observed between the two groups in the use of variables such as kidney function index (χ² = 6.995, P = 0.008), ACS presentation (χ² = 9.611, P = 0.002), culprit coronary artery (χ² = 3.937, P = 0.047), SBP (χ² = 10.556, P = 0.001), heart rate (χ² = 10.704, P = 0.001), and ST-segment deviation (χ² = 11.489, P = 0.001).

Table 2

Proportion of variables used in the prognostic risk score/model of patients after PCI based on Caucasian participants or not*.
Variable	Caucasian participants (n = 74)	No-Caucasian participants (n = 58)	χ²	P-value
Age, n (%)	66 (89.2)	47 (81.0)	1.755	0.185
Renal function index, n (%)	40 (54.1)	18 (31.0)	6.995	0.008
ACS presentation, n (%)	33 (44.6)	11 (19.0)	9.611	0.002
Diabetes, n (%)	29 (39.2)	17 (29.3)	1.398	0.237
LVEF, n (%)	29 (39.2)	15 (25.9)	2.599	0.107
Culprit coronary artery, n (%)	30 (40.5)	14 (24.1)	3.937	0.047
Heart failure, n (%)	15 (20.3)	19 (32.8)	2.652	0.103
SBP, n (%)	10 (13.5)	22 (37.9)	10.556	0.001
Heart rate, n (%)	9 (12.2)	21 (36.2)	10.704	0.001
Sex/gender, n (%)	18 (24.3)	10 (17.2)	0.976	0.323
Cerebrovascular disease (TIA), n (%)	13 (17.6)	10 (17.2)	0.002	0.961
Killip class, n (%)	11 (14.9)	12 (20.7)	0.767	0.381
ST-segment deviation, n (%)	6 (8.1)	18 (31.0)	11.489	0.001
Peripheral vascular disease, n (%)	14 (18.9)	7 (12.1)	1.140	0.286
Hypertension, n (%)	14 (18.9)	7 (12.1)	1.140	0.286
Prior coronary artery bypass grafting (CABG), n (%)	8 (10.8)	11 (19.0)	1.755	0.185
Cardiogenic shock, n (%)	13 (17.6)	4 (6.9)	3.300	0.069
Smoking, n (%)	9 (12.2)	7 (12.1)	< 0.001	0.987
Chronic lung disease, n (%)	11 (14.9)	4 (6.9)	2.050	1.152
Elevated cardiac markers, n (%)	4 (5.4)	7 (12.1)	1.118	0.290
Cardiac arrest, n (%)	8 (10.8)	5 (8.6)	0.176	0.675
TIMI flow grade, n (%)	8 (10.8)	3 (5.2)	0.716	0.398
BMI, n (%)	7 (9.5)	5 (8.6)	0.028	0.868
Prior PCI, n (%)	6 (8.1)	3 (5.2)	0.100	0.752
BNP/NT-proBNP, n (%)	5 (6.8)	4 (6.9)	< 0.001	1.000
* Six risk scores/models studies based upon international populations were excluded for involving multiple ethnicities.

Age, kidney function index, ACS presentation, diabetes, LVEF, culprit coronary artery, heart rate, sex, cerebrovascular disease, Killip class, ST-segment deviation, PVD, hypertension, prior CABG, cardiogenic shock, smoking status, CLD, elevated cardiac markers, cardiac arrest, TIMI flow grade, BMI, prior PCI, and BNP/NT-proBNP were the most widely applied variables in the risk prediction scores/models for patients after PCI. Moreover, our results show that variables are often used variously by researchers and participants, which suggests that ethnicity or geographical considerations influence how these variables are used. These results complement and expand on recent guidelines ^1,11 that prognostic risk assessment may contribute to further precise secondary prevention after PCI.

Even after PCI, patients with CHD require precise risk assessment and secondary prevention, which can significantly improve patient outcomes. Guidelines recommend risk scores/models used to identify risk stratification to guide therapeutic decision-making and help physicians improve the prognosis of patients.^1,11,12 Several scoring systems have been developed to evaluate the prognosis of patients with CHD within the last two decades,^{5–10,13−16} but few have been unanimously recognized as being suitable specifically for prognosis assessment of patients after PCI. Risk scores/models use covariates (each variable is assigned relative weights) to estimate the probability that a certain outcome (e.g., death or readmission) will occur in the future.^17–19 Therefore, variables are the basic elements of risk scores/models, and different variables may play different roles in different risk scores/models. Hence, it is necessary to conduct a detailed analysis of the variables contained in the risk scores/models to identify suitable candidates for further study and discussion.

In risk prediction scores/models, some variables, such as demographics, personal history, comorbidities, physical examination, and test results, have relatively stable predictive performance and universal applicability.^18,19 Our results showed that age, sex, sex, BMI, diabetes, hypertension, smoking, and SBP are representative of these variables (Table 2). We found that age, as a well-known prognostic predictor, was treated as a continuous variable (e.g., per year, per 10 years, or 30 years) in some scores/models,^5,6,10,13 whereas in other scores/models, it was applied as a categorical variable (e.g., > 40 years, > 50 years, > 60 years, > 65 years, > 70 years, > 75 years, or > 80 years).^7,9,20–26 Similarly, SBP and some scores/models regard per mmHg, per 10 mmHg, or per 20 mmHg as a continuous variable,^5,6,27,28 while more scores/models used it as a categorical variable based on < 100 mmHg or not.^7,29–31 At present, there is no uniform regulation on whether it is better to define such variables as continuous or categorical predictors, and researchers tend to favor their own needs for building models (e.g., increasing the discrimination power), which will cause a certain degree of selection bias. However, the guidelines ^1,11,32 and several well-known risk scores ^7,9,33 seem to favor age and SBP as categorical variables to assess risk and assist in the treatment of patients with CHD, which is also preferred by our panel to construct a score/model for patient prognosis assessment after PCI.

Diabetes, hypertension, and smoking have long been established as risk factors for CHD and used as predictors to assess patient outcomes after PCI.^11,22,24,34 Similar to CHD, CVD, and PVD are both arterial diseases, and the pathogenesis is both atherosclerosis, which leads to vascular stenosis or blockage and, in turn, causes ischemia of the target organ.³⁵ Therefore, CVD and PVD are also often included as variables in the risk scores/models of patients.^36,37 CLD (including COPD) is also a common concomitant disease of CHD.³⁸ CLD is closely related to the prognosis of patients with CHD, and is often selected as an important variable in the prognostic scoring system of PCI patients.^24,25,37,39 Perhaps it would be a good idea to construct a scoring system where all risk variables are comorbidities.

Regarding sex in predicting the prognosis of patients after PCI, there is still a matter of controversy among different scores/models. In general, if other factors are not considered, women with CHD have worse prognosis than men. However, there is now growing evidence that differences in adverse outcomes for women after PCI in different situations, compared to men, either no longer present or are significantly reduced after adjusting for confounding factors.⁴⁰ Specifically, some scores/models took female as the risk variable of the prognosis score after PCI (e.g. ATRIA risk score, EuroSCORE, EuroSCORE II, SYNTAX score II, and STS score) ^2,9,41,42, while others took male as the risk variable of the prognosis score after PCI (e.g. CHIP score, COAP risk model, and Framingham Risk Score).^24,34,43 Sex/gender, in any case, is a very important variable in the risk assessment scores/models for patients after PCI, and whether men or women are more at risk depends on the situation.

Obesity is an independent risk factor for CHD onset and development. Interestingly, a large body of literature has shown that overweight and obese patients with CHD have a better prognosis than the lean patients, which is known as the "obesity paradox".⁴⁴ For example, in models such as KORMI score,⁴⁵ MVD score,⁴⁶ and TIMI score,⁷ body weight/BMI less than a certain cut-off value is considered to be a risk factor for poor prognosis of patients after PCI, while in models such as NCDR CathPCI risk score ³⁶ and Logistic Clinical SYNTAX score,⁴⁷ BMI higher than a certain cutoff value is a predictor of poor prognosis in patients after PCI. In addition, some researchers have suggested that either a higher or lower BMI is a risk variable for long-term mortality after PCI,³⁷ with a U-shaped association. We are in favor of the last view; however, the conclusion needs to be verified by more high-quality, multicenter, and large-sample risk score/model studies with long-term follow-up.

Some indicators related to cardiac function, such as variables in the risk scores/models, also influence the prognosis of patients after PCI. Cardiogenic condition is a serious condition. In Mayo Clinic Risk Score,¹⁰ NERS score II,⁴⁸ NCDR CathPCI risk score,³⁶ PPCI risk score,⁴¹ and Toronto score,⁴⁹ cardiogenic shock was used to predict the in-hospital and 30-day mortality of AMI patients after PCI (patients with AMI are more prone to cardiogenic shock). More risk scores/models are needed to explore the impact of cardiogenic shock on long-term prognosis after PCI. As we all know, LVEF and Killip class are the criteria used to evaluate cardiac function, BNP and NT-proBNP are important biomarker for diagnosis and evaluation of HF in many clinical settings, and their level represents whether HF has occurred and the severity of HF.^50,51 For patients with CHD, the level of cardiac function indicates the clinical consequences of coronary ischemia. Therefore, cardiac function-related indicators can be used as indispensable variables in the construction of prognosis scores for post-PCI patients.

Additionally, prior PCI,^36,52,53 prior CABG,^52,53 ACS presentation,⁵⁴ ST-segment deviation,^5–7,55 elevated cardiac markers,^5–7 culprit coronary artery,^{10,20,24,48,56} and TIMI flow grade ^20,54,57,58 are directly related to an increased risk of poor prognosis after PCI. Studies have shown that previous revascularization (PCI or CABG) is associated with adverse cardiac events of patients with repeat PCI.^59,60 ACS presentation, ST-segment deviation, and elevated cardiac markers represent an unstable state of CHD, even AMI, which means greater and more variable risk.⁶¹ As variables, they occupy an important place in the prognostic scoring model of patients after PCI. The culprit coronary artery and TIMI flow grade are two factors that are directly related to PCI. Left main, multi-vessel, and chronic occlusive diseases are more complex and high-risk stratified coronary vascular lesions. They not only pose risks to PCI itself but also to the prognosis after PCI.^{10,20,24,48,56,62} The TIMI flow grade indicates whether the coronary blood supply is adequate, and pro- or post-procedural TIMI flow grade can affect the clinical outcome of PCI patients.^{20,54,57–59} Therefore, this type of risk variable is essential for creating a score/model to predict the outcome of patients after PCI.

Another variable worth discussing when constructing prognostic scores for patients after PCI is kidney function index. The kidney function index, measured by serum creatinine level, includes indicators such as AKI,⁶³ CKD,^22,64 eGFR,^48,64 kidney insufficiency,^65,66 and dialysis,^24,26 excluding contrast-induced nephropathy, and has different degrees of influence on the clinical outcomes of patients with stable CHD or ACS after PCI. The heart and kidney are both important organs in the human body, and their simultaneous onset of the two organs can affect each other, resulting in a series of pathophysiological changes known as cardiorenal syndrome, which increases the incidence of adverse prognosis. For AKI or CKD, mild kidney insufficiency, or end-stage kidney disease requiring dialysis, the serum creatinine level was the final evaluation criterion. Similarly, serum creatinine level was used as either a categorical variable ^{22,24,48,64–66} or a continuous variable ^8,9,13,20,28 in the risk score/model after PCI.

However, another issue is worth noting: The proportion of the above variables applied differed in the risk scores/models built for different ethnic groups (Table 2). In other words, many variables differ between Caucasians and non-Caucasians; some variables may be more suitable for Caucasians (e.g., kidney function index and ACS presentation), whereas other variables may be more suitable for non-Caucasians (e.g., SBP and heart rate). Even for the same variables, the cut-off values differed across ethnic groups. For example, Asians generally have lower BMIs than Caucasians.⁶⁷ Therefore, when constructing relevant scores/models, the selection of variables should also fully consider the ethnic background of the study population.

The limitations of this study include the fact that it was based on a literature analysis of previous studies. Although we thoroughly analyzed the high-frequency application variables in the scores/models of prognostic evaluation after PCI, it is still necessary to verify whether these variables can construct a high-performance score/model to predict the prognosis of patients after PCI in a real cohort. Additionally, applying keywords to filter articles may result in missing scores/models.

In the risk scores/models for prognostic assessment after PCI, the variables applied vary. When constructing scores/models, the variable selection should fully consider the ethnic background of the study population. Analyzing articles by variables in risk scores/models may improve the predictive power of related scores/models after PCI in follow-up studies. Clinicians and researchers need to critically consider the variables selected to construct a “super” risk model or score that can be widely used in the prognostic assessment of patients after PCI.

Clinical trial number

Not applicable

Ethical approval

Not applicable

Consent to participate

Not applicable

Competing interests

The authors have no conflicts of interest to disclose.

Availability of data and materials

Data sharing is not applicable to this article, as no new data were created or analyzed in this study.

Funding

Hebei Key Laboratory of Cardiac Injury Repair Mechanism Study (2024YDYY-KF01)

Authors’ contributions

H.L.Z. designed the study, conducted the research and drafted the manuscript. J. S. edited the manuscript. H.L.Z. and J.S. performed literature searches. G. Q. Q. and M. Q. Z. conducted statistical analyses and interpreted the data. G.L. revised the manuscript accordingly. All authors have read and approved the final manuscript.

Acknowledgements

The authors would like to thank all researchers who built the original scores/models used in this study. We thank the editor and all peer reviewers for their opinions and suggestions.

Writing Committee Members, Lawton JS, Tamis-Holland JE, Bangalore S, Bates ER, Beckie TM, Bischoff JM, et al. 2021 ACC/AHA/SCAI Guideline for Coronary Artery Revascularization: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. J Am Coll Cardiol. 2022;79(2):e21–9.
Castro-Dominguez YS, Wang Y, Minges KE, McNamara RL, Spertus JA, Dehmer GJ, et al. Predicting In-Hospital Mortality in Patients Undergoing Percutaneous Coronary Intervention. J Am Coll Cardiol. 2021;78(3):216–29.
Xu W, Tu H, Xiong X, Peng Y, Cheng T. Predicting the Risk of Unplanned Readmission at 30 Days After PCI: Development and Validation of a New Predictive Nomogram. Clin Interv Aging. 2022;17:1013–23.
Doll JA, O'Donnell CI, Plomondon ME, Waldo SW. Contemporary Clinical and Coronary Anatomic Risk Model for 30-Day Mortality After Percutaneous Coronary Intervention. Circ Cardiovasc Interv. 2021;14(12):e010863.
Bai XF, Zhang YP, Zhou J, Wu Y, Li RF, Sun LZ, et al. Combination of the CYP2C19 metabolizer and the GRACE risk score better predicts the long-term major adverse cardiac events in acute coronary syndrome undergoing percutaneous coronary intervention. Thromb Res. 2018;170:142–7.
Raposeiras-Roubín S, Abu-Assi E, Cabanas-Grandío P, Agra-Bermejo RM, Gestal-Romarí S, Pereira-López E, et al. Walking beyond the GRACE (Global Registry of Acute Coronary Events) model in the death risk stratification during hospitalization in patients with acute coronary syndrome: what do the AR-G (ACTION [Acute Coronary Treatment and Intervention Outcomes Network] Registry and GWTG [Get With the Guidelines] Database), NCDR (National Cardiovascular Data Registry), and EuroHeart Risk Scores Provide? JACC Cardiovasc Interv. 2012;5(11):1117–25.
Furnaz S, Karim M, Ashraf T, Ali S, Shahid I, Ali S, et al. Performance of the TIMI risk score in predicting mortality after primary percutaneous coronary intervention in elderly women: Results from a developing country. PLoS ONE. 2019;14(7):e0220289.
Sinning JM, Asdonk T, Erlhöfer C, Vasa-Nicotera M, Grube E, Nickenig G, et al. Combination of angiographic and clinical characteristics for the prediction of clinical outcomes in elderly patients undergoing multivessel PCI. Clin Res Cardiol. 2013;102(12):865–73.
Romagnoli E, Burzotta F, Trani C, Siviglia M, Biondi-Zoccai GG, Niccoli G, et al. EuroSCORE as predictor of in-hospital mortality after percutaneous coronary intervention. Heart. 2009;95(1):43–8.
Singh M, Lennon RJ, Holmes DR Jr, Bell MR, Rihal CS. Correlates of procedural complications and a simple integer risk score for percutaneous coronary intervention. J Am Coll Cardiol. 2002;40(3):387–93.
Byrne RA, Rossello X, Coughlan JJ, Barbato E, Berry C, Chieffo A, et al. 2023 ESC Guidelines for the management of acute coronary syndromes. Eur Heart J Acute Cardiovasc Care. 2024;13(1):55–161.
Virani SS, Newby LK, Arnold SV, Bittner V, Brewer LC, Demeter SH, et al. 2023 AHA/ACC/ACCP/ASPC/NLA/PCNA Guideline for the Management of Patients With Chronic Coronary Disease: A Report of the American Heart Association/American College of Cardiology Joint Committee on Clinical Practice Guidelines. Circulation. 2023;148(9):e9–119.
Biondi-Zoccai G, Romagnoli E, Castagno D, Sheiban I, De Servi S, Tamburino C, et al. Simplifying clinical risk prediction for percutaneous coronary intervention of bifurcation lesions: the case for the ACEF (age, creatinine, ejection fraction) score. EuroIntervention. 2012;8(3):359–67.
Henriques JP, Claessen BE, Dangas GD, Kirtane AJ, Popma JJ, Massaro JM, et al. Performance of currently available risk models in a cohort of mechanically supported high-risk percutaneous coronary intervention–From the PROTECT II randomized trial. Int J Cardiol. 2015;189:272–8.
Xhepa E, Tada T, Kufner S, Ndrepepa G, Byrne RA, Kreutzer J, et al. Long-term prognostic value of risk scores after drug-eluting stent implantation for unprotected left main coronary artery: A pooled analysis of the ISAR-LEFT-MAIN and ISAR-LEFT-MAIN 2 randomized clinical trials. Catheter Cardiovasc Interv. 2017;89(1):1–10.
Ono M, Kawashima H, Hara H, Gamal A, Wang R, Gao C, et al. External validation of the GRACE risk score 2.0 in the contemporary all-comers GLOBAL LEADERS trial. Catheter Cardiovasc Interv. 2021;98(4):E513–22.
Zhao HL, Gao XL, Liu YH, Li SL, Zhang Q, Shan WC, et al. Validation and derivation of short-term prognostic risk score in acute decompensated heart failure in China. BMC Cardiovasc Disord. 2022;22(1):307.
Moons KG, Kengne AP, Woodward M, Royston P, Vergouwe Y, Altman DG, et al. Risk prediction models: I. Development, internal validation, and assessing the incremental value of a new (bio)marker. Heart. 2012;98(9):683–90.
Moons KG, Kengne AP, Grobbee DE, Royston P, Vergouwe Y, Altman DG, et al. Risk prediction models: II. External validation, model updating, and impact assessment. Heart. 2012;98(9):691–8.
Montalto C, Kotronias RA, Marin F, Terentes-Printzios D, Shanmuganathan M, Emfietzoglou M, et al. Pre-procedural ATI score (age-thrombus burden-index of microcirculatory resistance) predicts long-term clinical outcomes in patients with ST elevation myocardial infarction treated with primary percutaneous coronary intervention. Int J Cardiol. 2021;339:1–6.
Zhao X, Liu C, Zhou P, Sheng Z, Li J, Zhou J, et al. Estimation of Major Adverse Cardiovascular Events in Patients With Myocardial Infarction Undergoing Primary Percutaneous Coronary Intervention: A Risk Prediction Score Model From a Derivation and Validation Study. Front Cardiovasc Med. 2020;7:603621.
Sato A, Hoshi T, Kakefuda Y, Harunari T, Watabe H, Hiraya D, et al. Effect of the Mehran risk score for the prediction of clinical outcomes after percutaneous coronary intervention. J Cardiol. 2015;66(5):417–22.
Minges KE, Herrin J, Fiorilli PN, Curtis JP. Development and validation of a simple risk score to predict 30-day readmission after percutaneous coronary intervention in a cohort of medicare patients. Catheter Cardiovasc Interv. 2017;89(6):955–63.
Brener SJ, Cunn GJ, Desai PH, Faroqui M, Ha LD, Handa G, et al. A Novel Risk Score to Predict One-Year Mortality in Patients Undergoing Complex High-Risk Indicated Percutaneous Coronary Intervention (CHIP-PCI). J Invasive Cardiol. 2021;33(4):E253–8.
Hannan EL, Zhong Y, Cozzens K, Ling FSK, Jacobs AK, King SB 3rd, et al. New York Risk Model and Simplified Risk Score for In-Hospital/30-Day Mortality for Percutaneous Coronary Intervention. Am J Cardiol. 2023;206:23–30.
Spirito A, Sharma A, Cao D, Sartori S, Zhang Z, Nicolas J, et al. New Criteria to Identify Patients at Higher Risk for Cardiovascular Complications After Percutaneous Coronary Intervention. Am J Cardiol. 2023;189:22–30.
Hizoh I, Gulyas Z, Domokos D, Banhegyi G, Majoros Z, Major L, et al. A novel risk model including vascular access site for predicting 30-day mortality after primary PCI: The ALPHA score. Cardiovasc Revasc Med. 2017;18(1):33–9.
Li L, Zhang X, Wang Y, Yu X, Jia H, Hou J, et al. A Novel Risk Score to Predict In-Hospital Mortality in Patients With Acute Myocardial Infarction: Results From a Prospective Observational Cohort. Front Cardiovasc Med. 2022;9:840485.
Addala S, Grines CL, Dixon SR, Stone GW, Boura JA, Ochoa AB, et al. Predicting mortality in patients with ST-elevation myocardial infarction treated with primary percutaneous coronary intervention (PAMI risk score). Am J Cardiol. 2004;93(5):629–32.
He PC, Duan CY, Liu YH, Wei XB, Lin SG. N-terminal pro-brain natriuretic peptide improves the C-ACS risk score prediction of clinical outcomes in patients with ST-elevation myocardial infarction. BMC Cardiovasc Disord. 2016;16(1):255.
Millo L, McKenzie A, De la Paz A, Zhou C, Yeung M, Stouffer GA. Usefulness of a Novel Risk Score to Predict In-Hospital Mortality in Patients ≥ 60 Years of Age with ST Elevation Myocardial Infarction. Am J Cardiol. 2021;154:1–6.
Ibanez B, James S, Agewall S, Antunes MJ, Bucciarelli-Ducci C, Bueno H, Caforio ALP, et al. 2017 ESC Guidelines for the management of acute myocardial infarction in patients presenting with ST-segment elevation: The Task Force for the management of acute myocardial infarction in patients presenting with ST-segment elevation of the European Society of Cardiology (ESC). Eur Heart J. 2018;39(2):119–77.
Capodanno D, Dipasqua F, Marcantoni C, Ministeri M, Zanoli L, Rastelli S, et al. EuroSCORE II versus additive and logistic EuroSCORE in patients undergoing percutaneous coronary intervention. Am J Cardiol. 2013;112(3):323–9.
Sara JD, Lennon RJ, Gulati R, Singh M, Holmes DR Jr, Lerman LO, et al. Utility of the Framingham Risk Score in predicting secondary events in patients following percutaneous coronary intervention: A time-trend analysis. Am Heart J. 2016;172:115–28.
Alagarsamy KN, Mathan S, Yan W, Rafieerad A, Sekaran S, Manego H, et al. Carbon nanomaterials for cardiovascular theranostics: Promises and challenges. Bioact Mater. 2021;6(8):2261–80.
Peterson ED, Dai D, DeLong ER, Brennan JM, Singh M, Rao SV, et al. Contemporary mortality risk prediction for percutaneous coronary intervention: results from 588,398 procedures in the National Cardiovascular Data Registry. J Am Coll Cardiol. 2010;55(18):1923–32.
Wu C, Camacho FT, King SB 3rd, Walford G, Holmes DR Jr, Stamato NJ, et al. Risk stratification for long-term mortality after percutaneous coronary intervention. Circ Cardiovasc Interv. 2014;7(1):80–7.
Roversi S, Roversi P, Spadafora G, Rossi R, Fabbri LM. Coronary artery disease concomitant with chronic obstructive pulmonary disease. Eur J Clin Invest. 2014;44(1):93–102.
Song Y, Gao Z, Tang X, Ma Y, Jiang P, Xu J, et al. Usefulness of the SYNTAX score II to validate 2-year outcomes in patients with complex coronary artery disease undergoing percutaneous coronary intervention: A large single-center study. Catheter Cardiovasc Interv. 2018;92(1):40–7.
Khamis RY, Ammari T, Mikhail GW. Gender differences in coronary heart disease. Heart. 2016;102(14):1142–9.
Çetinkal G, Koçaş C, Balaban Koçaş B, Arslan Ş, Abacı O, Karaca OŞ, et al. Comparative performance of AnTicoagulation and Risk factors In Atrial fibrillation and Global Registry of Acute Coronary Events risk scores in predicting long-term adverse events in patients with acute myocardial infarction. Anatol J Cardiol. 2018;20(2):77–84.
Andrews M, Iqbal J, Wall JJ, Teare D, El-Omar M, Fath-Ordoubadi F, et al. Development and Validation of a Novel Risk Score for Primary Percutaneous Coronary Intervention for ST-Elevation Myocardial Infarction. Cardiovasc Revasc Med. 2019;20(11):980–4.
Wu Y, Jin R, Grunkemeier GL. Validating the Clinical Outcomes Assessment Program risk model for percutaneous coronary intervention. Am Heart J. 2006;151(6):1276–80.
Tutor AW, Lavie CJ, Kachur S, Milani RV, Ventura HO. Updates on obesity and the obesity paradox in cardiovascular diseases. Prog Cardiovasc Dis. 2023;78:2–10.
Song PS, Ryu DR, Kim MJ, Jeon KH, Choi RK, Park JS, et al. Risk Scoring System to Assess Outcomes in Patients Treated with Contemporary Guideline-Adherent Optimal Therapies after Acute Myocardial Infarction. Korean Circ J. 2018;48(6):492–504.
Ryu KS, Bae JW, Jeong MH, Cho MC, Ryu KH, and other Korea Acute Myocardial Infarction Registry Investigators. Risk Scoring System for Prognosis Estimation of Multivessel Disease Among Patients with ST-Segment Elevation Myocardial Infarction. Int Heart J. 2019;60(3):708–14.
Farooq V, Vergouwe Y, Généreux P, Bourantas CV, Palmerini T, Caixeta A, et al. Prediction of 1-year mortality in patients with acute coronary syndromes undergoing percutaneous coronary intervention: validation of the logistic clinical SYNTAX (Synergy Between Percutaneous Coronary Interventions With Taxus and Cardiac Surgery) score. JACC Cardiovasc Interv. 2013;6(7):737–45.
Chen SL, Han YL, Zhang YJ, Ye F, Liu HW, Zhang JJ, et al. The anatomic- and clinical-based NERS (new risk stratification) score II to predict clinical outcomes after stenting unprotected left main coronary artery disease: results from a multicenter, prospective, registry study. JACC Cardiovasc Interv. 2013;6(12):1233–41.
Chowdhary S, Ivanov J, Mackie K, Seidelin PH, Dzavík V. The Toronto score for in-hospital mortality after percutaneous coronary interventions. Am Heart J. 2009;157(1):156–63.
Bozkurt B, Coats AJS, Tsutsui H, Abdelhamid CM, Adamopoulos S, Albert N, et al. Universal definition and classification of heart failure: a report of the Heart Failure Society of America, Heart Failure Association of the European Society of Cardiology, Japanese Heart Failure Society and Writing Committee of the Universal Definition of Heart Failure: Endorsed by the Canadian Heart Failure Society, Heart Failure Association of India, Cardiac Society of Australia and New Zealand, and Chinese Heart Failure Association. Eur J Heart Fail. 2021;23(3):352–80.
Del Buono MG, Montone RA, Rinaldi R, Gurgoglione FL, Meucci MC, Camilli M, et al. Clinical predictors and prognostic role of high Killip class in patients with a first episode of anterior ST-segment elevation acute myocardial infarction. J Cardiovasc Med (Hagerstown). 2021;22(7):530–8.
Gallone G, D'Ascenzo F, Conrotto F, Costa F, Capodanno D, Muscoli S, et al. Accuracy of the PARIS score and PCI complexity to predict ischemic events in patients treated with very thin stents in unprotected left main or coronary bifurcations. Catheter Cardiovasc Interv. 2021;97(2):E227–36.
Endo A, Kawamura A, Miyata H, Noma S, Suzuki M, Koyama T, et al. Angiographic Lesion Complexity Score and In-Hospital Outcomes after Percutaneous Coronary Intervention. PLoS ONE. 2015;10(6):e0127217.
Brener SJ, Leon MB, Serruys PW, Smits PC, von Birgelen C, Mehran R, et al. Derivation and external validation of a novel risk score for prediction of 30-day mortality after percutaneous coronary intervention. EuroIntervention. 2019;15(6):e551–7.
Garg S, Sarno G, Garcia-Garcia HM, Girasis C, Wykrzykowska J, Dawkins KD, et al. A new tool for the risk stratification of patients with complex coronary artery disease: the Clinical SYNTAX Score. Circ Cardiovasc Interv. 2010;3(4):317–26.
de Mulder M, Gitt A, van Domburg R, Hochadel M, Seabra-Gomes R, Serruys PW, et al. EuroHeart score for the evaluation of in-hospital mortality in patients undergoing percutaneous coronary intervention. Eur Heart J. 2011;32(11):1398–408.
Auffret V, Cottin Y, Leurent G, Gilard M, Beer JC, Zabalawi A, et al. Predicting the development of in-hospital cardiogenic shock in patients with ST-segment elevation myocardial infarction treated by primary percutaneous coronary intervention: the ORBI risk score. Eur Heart J. 2018;39(22):2090–102.
Lim TW, Karim TS, Fernando M, Haydar J, Lightowler R, Yip B, et al. Utility of Zwolle Risk Score in Guiding Low-Risk STEMI Discharge. Heart Lung Circ. 2021;30(4):489–95.
Al Suwaidi J, Velianou JL, Berger PB, Mathew V, Garratt KN, Reeder GS, et al. Primary percutaneous coronary interventions in patients with acute myocardial infarction and prior coronary artery bypass grafting. Am Heart J. 2001;142(3):452–9.
Arjomand H, Willerson JT, Holmes DR Jr, Bamlet WR, Surabhi SK, Roukoz B, et al. Outcome of patients with prior percutaneous revascularization undergoing repeat coronary intervention (from the PRESTO Trial). Am J Cardiol. 2005;96(6):741–6.
Kotecha T, Rakhit RD. Acute coronary syndromes. Clin Med (Lond). 2016;16(Suppl 6):s43–8.
De Marzo V, D'amario D, Galli M, Vergallo R, Porto I. High-risk percutaneous coronary intervention: how to define it today? Minerva Cardioangiol. 2018;66(5):576–93.
Cosentino N, Resta ML, Somaschini A, Campodonico J, D'Aleo G, Di Stefano G, et al. ST-Segment Elevation Acute Myocardial Infarction Complicated by Cardiogenic Shock: Early Predictors of Very Long-Term Mortality. J Clin Med. 2021;10(11):2237.
Shoji S, Kohsaka S, Kumamaru H, Nishimura S, Ishii H, Amano T, et al. Risk prediction models in patients undergoing percutaneous coronary intervention: A collaborative analysis from a Japanese administrative dataset and nationwide academic procedure registry. Int J Cardiol. 2023;370:90–7.
Polańska-Skrzypczyk M, Karcz M, Rużyłło W, Witkowski A. Bedside prediction of 9-year mortality after ST–segment elevation myocardial infarction treated with primary percutaneous coronary intervention. Kardiol Pol. 2019;77(7–8):703–9.
Madan P, Elayda MA, Lee VV, Wilson JM. Predicting major adverse cardiac events after percutaneous coronary intervention: the Texas Heart Institute risk score. Am Heart J. 2008;155(6):1068–74.
Park MH, Kwak SH, Kim KJ, Go MJ, Lee HJ, Kim KS, et al. Identification of a genetic locus on chromosome 4q34-35 for type 2 diabetes with overweight. Exp Mol Med. 2013;45(2):e7.

No competing interests reported.

Download PDF

Editorial decision: Revision requested
04 Sep, 2024
Editor assigned by journal
03 Sep, 2024
Submission checks completed at journal
03 Sep, 2024
First submitted to journal
19 Aug, 2024

You are reading this latest preprint version

Variations in how medical researchers report variables in risk scores or models to predict prognosis of patients after percutaneous coronary intervention: a retrospective analysis of published articles

Status:

Version 1

Abstract

Figures

1. Background

2. Methods

2.1 Search strategy and selection criteria

2.2 Data analysis

3. Results

4. Discussion

5. Limitations

6. Conclusions

Declarations

References

Additional Declarations

Status:

Version 1