The Osaka Prognostic Score as a Novel Predictor of Contrast-Induced Nephropathy and In-Hospital Mortality in STEMI Patients Undergoing Primary Percutaneous Coronary Intervention

Abstract

Background

The Osaka Prognostic Score (OPS), derived from C-reactive protein, serum albumin, and total lymphocyte count, reflects systemic inflammation and nutritional status. This study evaluated the prognostic value of the OPS for predicting contrast-induced nephropathy (CIN) and in-hospital all-cause mortality in patients with ST-elevation myocardial infarction (STEMI) undergoing primary percutaneous coronary intervention (pPCI).

Methods

In this retrospective study, 2,769 consecutive STEMI patients treated with pPCI were analyzed. OPS was calculated on admission (range 0-3). Multivariable regression, receiver operating characteristic, and spline analyses were used to evaluate predictors and optimal cutoff values.

Results

CIN occurred in 188 patients (6.8%), and 58 patients (2.1%) died during hospitalization. The mean OPS was significantly higher in patients who developed CIN compared with those who did not (1.86 ± 0.82 vs. 0.73 ± 0.70, p < 0.001). OPS independently predicted CIN (odds ratio: 6.817, 95% confidence interval: 5.157-9.012). For in-hospital mortality, OPS remained an independent predictor across all parsimonious multivariable Cox models, with adjusted hazard ratios ranging from 1.045 to 1.211 (all p < 0.05). An OPS cutoff value of ≥ 1.5 optimally predicted CIN (area under the curve [AUC]: 0.83) and in-hospital mortality (AUC: 0.73). Patients with an OPS of 3 had markedly higher rates of CIN (71.4%) and mortality (15.9%) than those with an OPS of 0 (0.7% and 0.5%, respectively; both p < 0.001). Kaplan-Meier analysis demonstrated a stepwise reduction in in-hospital survival with increasing OPS (log-rank p < 0.001).

Conclusions

The OPS is a simple and effective tool for the early prediction of CIN and in-hospital mortality in STEMI patients undergoing pPCI, and may facilitate early risk stratification and targeted management.

Abbreviations

ACS, Acute coronary syndrome

AUC, Area under the curve

CI, Confidence interval

CIN, Contrast-induced nephropathy

CRF, Chronic renal failure

CRP, C-reactive protein

DM, Diabetes mellitus

eGFR, Estimated glomerular filtration rate

GRACE, Global Registry of Acute Coronary Events

HF, Heart failure

HRs, Hazard ratios

LVEF, Left ventricular ejection fraction

OPS, Osaka Prognostic Score

OR, Odds ratio

PCI, Percutaneous coronary intervention

pPCI, Primary percutaneous coronary intervention

RCS, Restricted cubic splines

ROC, Receiver operating characteristic

STEMI, ST-elevation myocardial infarction

TIMI, Thrombolysis in myocardial infarction

TLC, Total lymphocyte count

INTRODUCTION

Coronary artery disease remains the leading cause of death worldwide, with ST-elevation myocardial infarction (STEMI) representing its most critical manifestation.¹ Despite significant improvements in outcomes through rapid reperfusion therapy, particularly primary percutaneous coronary intervention (pPCI) — the current gold standard in acute STEMI management — patients remain at risk for serious in-hospital complications.^2,3 Among these complications, contrast-induced nephropathy (CIN), an acute kidney injury triggered by iodinated contrast media, is one of the most frequent and clinically significant.³ CIN develops in approximately 10% of STEMI patients undergoing percutaneous coronary intervention (PCI) and is associated with prolonged hospitalization, increased morbidity, and higher in-hospital mortality.⁴ Moreover, early mortality after STEMI remains substantial despite optimal reperfusion and medical therapy, with a recent meta-analysis reporting a rate of nearly 10%.⁵ Identifying high-risk patients at admission is therefore essential for optimizing outcomes.

Early risk stratification forms the cornerstone of acute cardiovascular care. Widely used scores such as Thrombolysis in Myocardial Infarction (TIMI) and Global Registry of Acute Coronary Events (GRACE) guide clinical decision-making by integrating clinical and laboratory variables to estimate mortality risk, whereas the Mehran score specifically predicts CIN following PCI.^6,7 However, conventional models may overlook systemic factors such as inflammation and nutritional status, which significantly influence outcomes. Recently, prognostic tools combining inflammatory and nutritional markers have gained attention.^8,9 The Osaka Prognostic Score (OPS), calculated from C-reactive protein (CRP), serum albumin, and total lymphocyte count (TLC), reflects both systemic inflammation and nutritional reserve.^10-12 Each component of the OPS reflects key clinical aspects of inflammation and nutrition; however, its overall predictive role in STEMI remains unexplored. Therefore, the aim of this study was to evaluate whether admission OPS independently predicts CIN and in-hospital all-cause mortality in STEMI patients undergoing pPCI.

METHODS

Study design and population

This retrospective observational study included 2,769 patients diagnosed with STEMI who underwent PCI between December 2022 and January 2025. All patients were treated at a high-volume tertiary care center with a 24/7 catheterization laboratory. Patients were eligible for inclusion if they were 18 years of age or older, had a confirmed diagnosis of STEMI based on electrocardiographic findings, and received emergency PCI with iodinated contrast media. The exclusion criteria were the presence of sepsis, chronic infections, chronic inflammatory conditions (such as autoimmune or connective tissue diseases), regular use of anti-inflammatory medications, severe liver disease or active hepatitis, end-stage renal disease requiring dialysis, and a history of malignancy.

All patients were followed throughout their hospital stay, from admission to discharge or in-hospital death, whichever occurred first. Since all participants presented with STEMI, they all underwent emergency coronary angiography and PCI within a maximum of 120 minutes from first medical contact, in accordance with international guidelines. None of the patients received fibrinolytic therapy prior to PCI. All patients received guideline-directed medical therapy for STEMI, as recommended by the European Society of Cardiology.² Heart failure (HF) status at admission was assessed using the Killip classification system, and patients with clinical signs of HF received optimal, guideline-recommended medical therapy for HF throughout their hospitalization.¹³

This study was conducted in accordance with the principles outlined in the Declaration of Helsinki and all applicable ethical standards. Ethical approval was obtained from the Clinical Research Ethics Committee of Ankara Etlik City Hospital (approval date and number: March 12, 2025 / AEŞH-BADEK-2025-0324). Due to the retrospective nature of the study, the ethics committee waived the requirement for informed consent.

Clinical data collection and definitions

All clinical, laboratory, and procedural data were extracted from the institutional electronic medical record system and the digital coronary angiography database. Baseline demographic characteristics, Killip class, cardiovascular risk factors, laboratory results, and procedural details were obtained retrospectively. Diabetes mellitus (DM) and hypertension were defined according to the most recent guideline criteria. DM was defined as fasting plasma glucose ≥ 126 mg/dL, random plasma glucose ≥ 200 mg/dL, HbA1c ≥ 6.5%, or ongoing treatment with antidiabetic medication. Hypertension was defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, or current antihypertensive therapy. Left ventricular ejection fraction (LVEF) was calculated via transthoracic echocardiography using the modified Simpson’s biplane method during the index hospitalization. Failed PCI was defined as the inability to achieve complete vessel patency in the culprit coronary lesion. The estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation. Chronic renal failure (CRF) was defined as an eGFR of ≤ 60 mL/min/1.73 m².

The OPS was calculated using three biomarkers: CRP, serum albumin, and TLC. Each variable was scored as follows:⁹

• CRP ≤ 10.0 mg/L = 0 points, > 10.0 mg/L = 1 point

• Albumin ≥ 3.5 g/dL = 0 points, < 3.5 g/dL = 1 point

• TLC ≥ 1,600/μL = 0 points, < 1,600/μL = 1 point

The cumulative score ranged from 0 to 3, and the patients were stratified into four groups accordingly.

CIN was defined as an absolute increase in serum creatinine of ≥ 0.5 mg/dL or a relative increase of ≥ 25% from baseline within 48 hours following exposure to iodinated contrast.¹⁴ Serum creatinine was measured at baseline and 48 hours post-procedure as part of routine clinical care.

Statistical analysis

Continuous variables were expressed as mean ± standard deviation or median (range), depending on the distribution assessed by the Kolmogorov-Smirnov test. Categorical variables were reported as frequencies (percentages). Between-group comparisons for continuous variables were performed using the Student’s t-test or Mann-Whitney U test, as appropriate. The chi-square test or Fisher’s exact test was used for categorical variables. The patients were first divided into two groups according to the presence or absence of CIN. Variables associated with CIN were analyzed using univariable logistic regression. Parameters with p values < 0.10 were included in a multivariable logistic regression model to identify independent predictors of CIN. Multicollinearity among candidate variables was evaluated using variance inflation factor analysis, and variables representing overlapping clinical domains were not included simultaneously in the multivariable model. The predictive performance and optimal OPS cutoff value for CIN were evaluated using receiver operating characteristic (ROC) curve analysis, and the area under the curve (AUC) was calculated. The optimal cutoff values of the OPS for CIN and in-hospital mortality were determined using the Youden index derived from the ROC curve analysis. Subsequently, the entire population was divided into survivor and non-survivor groups to analyze in-hospital all-cause mortality. Univariable Cox proportional hazards regression was performed, followed by multivariable Cox regression using variables with p < 0.10 in univariable analysis. Given the limited number of in-hospital deaths, multivariable Cox regression analyses were performed using several parsimonious models to avoid overfitting and to assess the robustness of the association between the OPS and mortality across different clinical scenarios. Hazard ratios (HRs) and 95% confidence intervals (CIs) were reported. The best OPS cutoff value for predicting in-hospital mortality was also determined by ROC analysis. Kaplan-Meier survival curves were generated based on this threshold and compared using the log-rank test. To explore the potential nonlinear association between the OPS and clinical outcomes, logistic regression models with restricted cubic splines (RCS) (3 knots) were constructed. Statistical significance of nonlinearity was assessed using likelihood ratio tests. A bootstrap resampling technique with 100 iterations was applied to generate 95% CIs for the estimated probabilities. All patients were classified into four groups based on OPS score (0, 1, 2, and 3), and intergroup comparisons of baseline characteristics and clinical outcomes were performed. When comparing continuous variables across these four groups, one-way analysis of variance was used for normally distributed variables, and the Kruskal-Wallis test was used for non-normally distributed variables. For categorical variables, the chi-square test or Fisher’s exact test was used as appropriate. Kaplan-Meier analysis was used to compare survival among the four OPS categories. All statistical analyses were performed using SPSS software (version 29.0; IBM Corp., Armonk, NY, USA). A two-tailed p value of < 0.05 was considered statistically significant.

RESULTS

A total of 2,769 STEMI patients who underwent pPCI were retrospectively analyzed. The median age of the study cohort was 57 years, and 16.8% were female. The prevalence of CIN was 6.8% (n = 188), and the overall in-hospital mortality rate was 2.1% (n = 58). The median length of hospital stay for the entire cohort was 3 days (range, 1-12 days).

OPS and CIN

As shown in Table 1, the patients who developed CIN were significantly older (mean age: 65.1 ± 12.5 vs. 56.3 ± 11.3 years, p < 0.001) and had a greater prevalence of cardiovascular risk factors, including hypertension, DM, and CRF. The prevalence of female patients was significantly higher in the group that developed CIN compared to the group that did not. In addition, the patients who developed CIN more frequently presented with Killip class III-IV than those without CIN (19.1 vs. 11.0%, p < 0.001). They also presented more frequently with impaired LVEF (41.7 ± 13.1 vs. 48.8 ± 9.7%, p < 0.001) and worse baseline renal function, as indicated by a lower eGFR value (71 vs. 99 mL/min/1.73 m², p < 0.001). CRP levels were significantly higher in the CIN group, while hemoglobin, TLC, and serum albumin levels were lower. Notably, the mean OPS score was markedly higher among the patients who developed CIN (1.86 ± 0.82 vs. 0.73 ± 0.70, p < 0.001), and in-hospital mortality was also substantially higher in this group (13.3 vs. 1.3%, p < 0.001). The rates of regular pre-admission use of acetylsalicylic acid, anticoagulants, beta-blockers, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, and statins were similar between the two groups. In terms of periprocedural characteristics, the amount of contrast agent used during the procedure was comparable between the two groups; however, the incidence of failed PCI (13.8 vs. 5.1%, p < 0.001) and the length of hospital stay (7.1 ± 2.0 vs. 3.3 ± 0.9 days, p < 0.001) were significantly higher in the patients who developed CIN.

Table 1. Baseline clinical, laboratory, and procedural characteristics of patients stratified by the development of CIN.

	Patients without CIN (n = 2581)	Patients with CIN (n = 188)	p value
Age, years	56.3 ± 11.3	65.1 ± 12.5	< 0.001
Female sex, n (%)	418 (16.2)	47 (25.0)	0.002
Killip class III-IV, n (%)	283 (11.0)	36 (19.1)	< 0.001
Comorbidities
Diabetes mellitus, n (%)	552 (21.4)	70 (37.2)	< 0.001
Hypertension, n (%)	703 (27.2)	95 (50.5)	< 0.001
Smoking, n (%)	1024 (39.7)	84 (44.7)	0.175
Chronic renal failure, n (%)	27 (1.0)	16 (8.5)	< 0.001
Coronary artery disease, n (%)	373 (14.4)	35 (18.6)	0.119
Atrial fibrillation, n (%)	32 (1.2)	4 (2.1)	0.299
Laboratory data
LVEF, %	48.8 ± 9.7	41.7 ± 13.1	< 0.001
Hemoglobin, g/dL	13.7 ± 1.7	13.0 ± 2.1	< 0.001
Platelets, ×103/mm3	231 (60-778)	228 (60-626)	0.572
Lymphocytes, ×103/mm3	1.7 (0.2-7.3)	1.5 (0.2-7.5)	0.003
C-reactive protein, mg/L	7.0 (0-21)	11.5 (0-39)	< 0.001
eGFR, mL/min/1.73 m2	99 (18-196)	71 (55-89)	< 0.001
Serum creatinine, mg/dL	0.8 (0.2-3.3)	1.1 (0.3-4.2)	< 0.001
Serum sodium, mmol/L	136.5 ± 4.7	135.7 ± 4.4	0.251
Serum albumin, g/dL	3.8 (0.7-4.9)	3.1 (1.4-4.4)	< 0.001
Triglycerides, mg/dL	138 (68-201)	141 (70-471)	0.862
HDL-cholesterol, mg/dL	38.0 ± 10.1	39.0 ± 12.4	0.326
LDL-cholesterol, mg/dL	109 (81-405)	115 (65-264)	0.251
Osaka Prognostic Score	0.73 ± 0.70	1.86 ± 0.82	< 0.001
Drugs used before hospital admission
Acetylsalicylic acid, n (%)	387 (15.0)	34 (18.1)	0.254
Anticoagulant, n(%)	29 (1.1)	3 (1.6)	0.558
Beta-blocker, n (%)	868 (33.6)	71 (37.8)	0.248
ACEi or ARB, n (%)	1049 (40.6)	67 (35.6)	0.177
Statin, n (%)	1134 (43.9)	71 (37.8)	0.100
Periprocedural features
Contrast agent amount, mL	200 (20-700)	200 (100-600)	0.201
Failed PCI, n (%)	135 (5.1)0	26 (13.8)	< 0.001
Length of hospitalization, days	3.3 ± 0.9	7.1 ± 2.0	< 0.001
In-hospital mortality, n (%)	33 (1.3)	25 (13.3)	< 0.001

Values are presented as mean ± standard deviation or median (range) for continuous variables and as number (%) for categorical variables.

ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; CIN, contrast-induced nephropathy; eGFR, estimated glomerular filtration rate; HDL, high-density lipoprotein; LDL, low-density lipoprotein; LVEF, left ventricle ejection fraction; PCI, percutaneous coronary intervention.

Univariable analysis identified that multiple variables were significantly associated with CIN, including age, female sex, Killip class III-IV, DM, hypertension, CRF, LVEF, hemoglobin, TLC, CRP, eGFR, serum creatinine, serum albumin, OPS, and failed PCI. In multivariable logistic regression, the independent predictors of CIN were elevated OPS score (odds ratio [OR]: 6.817, 95% CI: 5.157-9.012, p < 0.001), advanced age (OR: 1.054 per 1 year, 95% CI: 1.035-1.073, p < 0.001), hypertension (OR: 2.007, 95% CI: 1.352-2.981, p = 0.001), CRF (OR: 1.238, 95% CI: 1.071-1.431, p < 0.001), and reduced LVEF (OR: 0.970 per %, p = 0.001) (Table 2).

Table 2. Univariable and multivariable logistic regression analyses for predictors of contrast-induced nephropathy.

	Univariable regression	Multivariable regression
	p value	OR (95% CI)	p value
Age	< 0.001	1.054 (1.035-1.073)	< 0.001
Female sex	0.002	1.121 (0.677-1.855)	0.657
Killip class III-IV *	< 0.001	-	-
Diabetes mellitus	< 0.001	1.294 (0.850-1.970)	0.230
Hypertension	< 0.001	2.007 (1.352-2.981)	0.001
Chronic renal failure †	< 0.001	1.238 (1.071-1.431)	< 0.001
LVEF *	0.001	0.970 (0.953-0.987)	< 0.001
Hemoglobin	< 0.001	0.969 (0.863-1.089)	0.602
Lymphocytes #	0.005	-	-
C-reactive protein #	< 0.001	-	-
eGFR †	< 0.001	-	-
Serum creatinine †	0.001	-	-
Serum albumin #	< 0.001	-	-
Statin usage	0.102	-	-
Osaka Prognostic Score #	< 0.001	6.817 (5.157-9.012)	< 0.001
Failed PCI	0.001	1.031 (0.989-1.065)	0.056

Odds ratios (OR) and 95% confidence intervals (CI) are presented for each variable.

Multicollinearity was evaluated using the variance inflation factor analysis.

* Only LVEF was included instead of the Killip class. ^# Only Osaka Prognostic Score (OPS) was included instead of its individual components (C-reactive protein, serum albumin, and lymphocyte count). ^† Chronic renal failure was included instead of serum creatinine and eGFR in the multivariable model.

Abbreviations are same as in Table 1.

ROC curve analysis confirmed the predictive performance of the OPS for CIN, yielding an AUC of 0.830 (95% CI: 0.790-0.856, p < 0.001). Using a cutoff value of ≥ 1.5, the OPS demonstrated a sensitivity of 66.5% and specificity of 86.6%, indicating excellent discriminative ability for CIN (Figure 1A). An RCS model further illustrated a progressive and nonlinear increase in CIN probability with increasing OPS scores (p < 0.001) (Figure 1B).

Figure 1.

Predictive performance of the Osaka Prognostic Score (OPS) for contrast-induced nephropathy (CIN). (A) Receiver operating characteristic (ROC) curve demonstrating the discriminative ability of OPS for CIN (area under the curve [AUC] = 0.83); (B) Association between OPS and the predicted probability of CIN estimated using a logistic regression model with restricted cubic splines (3 knots). The solid line represents the adjusted predicted probability, and the shaded area indicates the 95% confidence interval (CI). A significant nonlinear association was observed (likelihood ratio test, p < 0.001).

OPS and in-hospital mortality

Baseline characteristics of the patients stratified by in-hospital survival status are shown in Table 3. The patients who died during hospitalization were older (mean age: 67.3 ± 12.9 vs. 56.7 ± 11.5 years, p < 0.001) and had higher rates of DM and CRF. The proportion of female patients among those who died was higher than that among the survivors (31 vs. 16.5%, p = 0.003). In addition, advanced HF at presentation (Killip class III-IV) was markedly more frequent among the non-survivors compared with the survivors (51.7 vs. 10.7%, p < 0.001). Compared to the survivors, the non-survivors had significantly lower LVEF (34.4 ± 12.2 vs. 48.6 ± 9.9%, p < 0.001), eGFR, hemoglobin and serum albumin, and higher CRP levels. The mean OPS score in the non-survivors was nearly double that of the survivors (1.55 ± 0.88 vs. 0.79 ± 0.75, p < 0.001). When evaluating periprocedural characteristics, the amount of contrast agent used during the procedure and the length of hospital stay were similar between the two groups; however, the rates of CIN and failed PCI were significantly higher in the non-survivors (both p < 0.001).

Table 3. Clinical, laboratory, and procedural characteristics of patients stratified by in-hospital mortality status.

	Alive (n = 2711)	Dead (n = 58)	p value
Age, years	56.7 ± 11.5	67.3 ± 12.9	< 0.001
Female sex, n (%)	447 (16.5)	18 (31.0)	0.003
Killip class III-IV, n (%)	289 (10.7)	30 (51.7)	< 0.001
Comorbidities
Diabetes mellitus, n (%)	590 (21.8)	32 (55.2)	< 0.001
Hypertension, n (%)	775 (28.6)	23 (39.7)	0.066
Smoking, n (%)	1085 (40.0)	23 (39.6)	0.748
Chronic renal failure, n (%)	36 (1.3)	7 (12.1)	< 0.001
Coronary artery disease, n (%)	401 (14.8)	7 (12.1)	0.563
Atrial fibrillation, n (%)	34 (1.2)	2 (3.4)	0.144
Laboratory data
LVEF, %	48.6 ± 9.9	34.4 ± 12.2	< 0.001
eGFR, mL/min/1.73 m2	102 (11-377)	60 (18-111)	< 0.001
Hemoglobin, g/dL	13.6 ± 1.8	13.1 ± 1.9	0.013
Platelets, ×103/mm3	230 (60-778)	237 (60-442)	0.505
Lymphocytes, ×103/mm3	1.7 (0.2-10.3)	1.60 (0.6-9.2)	0.295
C-reactive protein, mg/L	7 (0-21)	11 (0-41)	< 0.001
Serum creatinine, mg/dL	0.8 (0.1-3.3)	1.2 (0.2-4.2)	< 0.001
Serum sodium, mmol/L	136.5 ± 4.7	135.5 ± 5.9	0.123
Serum albumin, g/dL	3.8 (0.7-4.9)	3.4 (1.4-4.3)	< 0.001
Triglycerides, mg/dL	138 (66-332)	120 (60-285)	0.302
HDL-cholesterol, mg/dL	38.1 ± 10.2	34.3 ± 12.8	0.103
LDL-cholesterol, mg/dL	110 (85-405)	89 (68-240)	0.211
Osaka Prognostic Score	0.79 ± 0.75	1.55 ± 0.88	< 0.001
Drugs used before hospital admission
Acetylsalicylic acid, n (%)	410 (15.1)	11 (18.9)	0.420
Beta-blocker, n (%)	920 (33.9)	19 (32.7)	0.557
ACEi or ARB, n (%)	1093 (40.0)	23 (39.0)	0.706
Statin, n (%)	1181 (43.5)	24 (41.3)	0.743
Periprocedural features
Contrast agent amount, mL	200 (30-700)	200 (20-400)	0.794
Contrast-induced nephropathy, n (%)	163 (6.0)	25 (43.1)	< 0.001
Failed PCI, n (%)	142 (5.2)	19 (32.8)	< 0.001
Length of hospitalization, days	3.8 ± 2.3	3.8 ± 2.0	0.897

Values are presented as mean ± standard deviation or median (range) for continuous variables and as number (%) for categorical variables.

Abbreviations are same as in Table 1.

As shown in Table 4, univariable Cox proportional hazards analysis identified several clinical and laboratory variables that were significantly associated with in-hospital all-cause mortality. Advanced age, female sex, Killip class III-IV, DM, CRF, failed PCI, and the occurrence of CIN were all significant clinical predictors of mortality. In addition, lower LVEF, reduced eGFR, lower hemoglobin levels, and elevated CRP levels were significantly associated with increased in-hospital mortality. The OPS also demonstrated a strong univariable association with in-hospital mortality. As detailed in Table 5, multivariable Cox regression analyses were performed using several parsimonious models to account for the limited number of in-hospital mortality events. Across all models incorporating different combinations of clinical and laboratory variables, the OPS remained independently associated with in-hospital all-cause mortality. Specifically, after adjustment for age, sex, hemodynamic status, renal function, procedural failure, and CIN, higher OPS values consistently conferred an increased risk of in-hospital mortality, with adjusted HRs ranging from 1.05 to 1.21 across models (all p < 0.05).

Table 4. Univariable Cox regression analyses for associated variables of in-hospital all-cause mortality.

	Covariates	p value
Clinical variables	Age	< 0.001
	Female sex	0.004
	Killip class III-IV	< 0.001
	Diabetes mellitus	< 0.001
	Hypertension	0.101
	Chronic renal failure	< 0.001
	Failed PCI	< 0.001
	Contrast-induced nephropathy	< 0.001
Laboratory variables	LVEF	< 0.001
	eGFR	< 0.001
	Hemoglobin	0.011
	C-reactive protein *	< 0.001
	Serum creatinine	< 0.001
	Serum albumin *	< 0.001
	Osaka Prognostic Score *	< 0.001

Multivariable regression models include variables found to be significant in univariable analysis.

* Only Osaka Prognostic Score (OPS) was included instead of its individual components (C-reactive protein and serum albumin) in the multivariable models.

Abbreviations are same as in Table 1.

Table 5. Multivariable Cox regression models including combined clinical and laboratory parameters together with the Osaka Prognostic Score (OPS) for predicting in-hospital mortality.

Models	Covariates	Adjusted hazard ratio for OPS (95% confidence interval)	p value
Clinical + laboratory parameters (Model A)	Age	1.211 (1.113-1.398)	0.001
	Female sex
	Killip III-IV
	eGFR
	Osaka Prognostic Score
Clinical + laboratory parameters (Model B)	Age	1.167 (1.097-1.258)	0.002
	Female sex
	Diabetes mellitus
	LVEF
	Osaka Prognostic Score
Clinical + laboratory parameters (Model C)	Age	1.098 (1.055-1.151)	0.010
	Killip III-IV
	Chronic renal failure
	Failed PCI
	Osaka Prognostic Score
Clinical + laboratory parameters (Model D)	Killip III-IV	1.182 (1.113-1.245)	0.001
	Diabetes mellitus
	Contrast-induced nephropathy
	Hemoglobin
	Osaka Prognostic Score
Clinical + laboratory parameters (Model E)	Diabetes mellitus	1.045 (1.024-1.102)	0.021
	Failed PCI
	Contrast-induced nephropathy
	LVEF
	Osaka Prognostic Score

Abbreviations are same as in Table 1.

The ROC analysis confirmed that the OPS had good discriminatory power for predicting all-cause in-hospital mortality, with an AUC of 0.730 (95% CI: 0.665-0.793, p < 0.001). Using a cutoff value of ≥ 1.5, the OPS demonstrated a sensitivity of 46.6% and specificity of 83.6% for in-hospital mortality (Figure 2A). Moreover, RCS analysis revealed a continuous, nonlinear increase in predicted mortality risk with increasing OPS score (p < 0.001), with an accentuated slope beyond an OPS of 2 (Figure 2B).

Figure 2.

Predictive performance of the Osaka Prognostic Score (OPS) for all-cause in-hospital mortality. (A) Receiver operating characteristic (ROC) curve showing the discriminative ability of OPS for in-hospital mortality (area under the curve [AUC] = 0.73); (B) Relationship between OPS and predicted in-hospital mortality risk modeled using logistic regression with restricted cubic splines (3 knots). The solid line denotes the adjusted predicted probability, and the shaded area represents the 95% confidence interval (CI). The nonlinear association between OPS and mortality risk was statistically significant (likelihood ratio test, p < 0.001).

Patient stratification according to OPS

Baseline clinical, laboratory, and procedural characteristics of the patients categorized by OPS group (0, 1, 2, and 3) are detailed in Table 6. Progressive increases in age, prevalence of comorbidities (hypertension and CRF), Killip class III-IV, and inflammatory burden (increased CRP and decreased TLC) were observed with a higher OPS (all p < 0.05). LVEF values showed a stepwise decline across increasing OPS groups. Serum albumin levels were inversely correlated with the OPS, while serum creatinine levels were significantly elevated. Similarly, the incidence of CIN and all-cause in-hospital mortality showed a graded increase across OPS groups (CIN: 0.7% in OPS 0 vs. 71.4% in OPS 3; mortality: 0.5% in OPS 0 vs. 15.9% in OPS 3, p < 0.001 for both). In addition, the need for hemodialysis and the rate of failed PCI increased significantly with higher OPS.

Table 6. Baseline characteristics of patients categorized by Osaka Prognostic Scores (OPS) (0–3).

	OPS = 0 (n = 1071)	OPS = 1 (n = 1226)	OPS = 2 (n = 409)	OPS = 3 (n = 63)	p value
Age, years	54.1 ± 10.9a	57.8 ± 11.4b	60.0 ± 12.2c	65.5 ± 12.6d	< 0.001
Female sex, n (%)	171 (16.0)	208 (17.0)	74 (18.1)	12 (19.0)	0.735
Killip class III-IV, n (%)	70 (6.5)a	122 (10.0)a	102 (24.9)b	25 (39.7)c	< 0.001
Diabetes mellitus, n (%)	219 (20.4)	287 (23.4)	98 (24.0)	18 (28.6)	0.155
Hypertension, n (%)	261 (24.3)a	385 (31.4)b	123 (30.1)a,b	29 (46.0)b	< 0.001
Smoking, n (%)	481 (44.9)a	448 (36.5)b	160 (39.1)a,b	19 (30.2)a,b	< 0.001
Chronic renal failure, n (%)	11 (1.0)a	15 (1.2)a	14 (3.4)b	3 (4.8)a,b	0.001
LVEF, %	50.1 ± 9.3a	47.9 ± 10.1b	45.8 ± 10.7c	39.5 ± 12.7d	< 0.001
Hemoglobin, g/dL	14.0 ± 1.7a	13.5 ± 1.6b	13.3 ± 1.9c	12.6 ± 1.9d	< 0.001
Platelets, ×103/mm3	246.9 ± 71.7a	236.8 ± 71.6b	230.9 ± 67.3b	229.8 ± 74.1a,b	< 0.001
Lymphocytes, ×103/mm3	2.3 (0.6-7.3)a	1.4 (0.2-6.8)b	1.3 (0.2-5.0)c	1.1 (0.2-1.5)c	< 0.001
C-reactive protein, mg/L	6 (0-10)a	7 (0-16)b	8 (0-27)c	16 (11-39)d	< 0.001
Serum creatinine, mg/dL	0.8 ± 0.2a	0.8 ± 0.2b	0.9 ± 0.3c	1.1 ± 0.4d	< 0.001
Serum sodium, mmol/L	137.0 ± 3.4a	136.2 ± 5.8b	136.2 ± 3.6b	135.2 ± 4.4b	< 0.001
Serum albumin, g/dL	3.8 ± 0.2a	3.6 ± 0.4b	3.2 ± 0.4c	2.9 ± 0.4d	< 0.001
Triglycerides, mg/dL	108 (44-276)a	127 (61-303)a	133 (82-392)a	149 (93-475)b	< 0.001
HDL-cholesterol, mg/dL	37.5 ± 9.7	38.5 ± 10.3	38.2 ± 10.9	39.5 ± 12.7	0.131
LDL-cholesterol, mg /dL	103 (64-238)a	109 (77-267)a	106 (81-301)a	115 (96-405)b	< 0.001
Contrast-induced nephropathy, n (%)	8 (0.7)a	55 (4.5)b	80 (19.6)c	45 (71.4)d	< 0.001
Hemodialysis, n (%)	0 (0)a	8 (0.7)b	9 (2.2)c	7 (11.1)d	< 0.001
In-hospital mortality, n (%)	5 (0.5)a	26 (2.1)b	17 (4.2)b	10 (15.9)c	< 0.001
Failed PCI, n (%)	49 (4.6)a	64 (5.2)a	39 (9.5)b	9 (14.3)b	< 0.001

Values are presented as mean ± standard deviation or median (range) for continuous variables and as number (%) for categorical variables.

Superscripts (a, b, c, d) indicate statistically significant differences between specific OPS subgroups.

Abbreviations are same as in Table 1.

Survival analysis

Kaplan-Meier survival curves were generated to evaluate in-hospital outcomes according to OPS group. When dichotomized at a cutoff value of 1.5, the patients with a higher OPS (≥ 1.5) had significantly worse in-hospital survival compared to those with an OPS < 1.5 (Figure 3A, log-rank p < 0.001). Stratification into four distinct OPS groups (0, 1, 2, 3) showed a clear stepwise gradient in survival, with increasing mortality across groups (Figure 3B, log-rank p < 0.001). Notably, the patients with an OPS of 3 had the steepest decline in survival probability.

Figure 3.

Kaplan-Meier analysis of in-hospital survival according to the Osaka Prognostic Score (OPS). (A) Survival curves stratified by OPS category using a predefined cutoff value (< 1.5 vs. ≥ 1.5); (B) Survival curves stratified according to individual OPS categories (OPS 0, 1, 2, and 3). Patients with higher OPS values showed significantly lower in-hospital survival in both analyses (log-rank test, p < 0.001).

DISCUSSION

The results of this study demonstrated that the OPS, a composite of CRP, albumin, and TLC, was an independent predictor of both CIN and in-hospital all-cause mortality in STEMI patients undergoing pPCI. While the prognostic value of each individual component of the OPS has been previously reported in cardiovascular settings, this study is the first to evaluate the composite OPS as a unified index in this population and to demonstrate its dual predictive utility for both renal and early mortality outcomes (Central Illustration).

Central Illustration.

Graphical summary illustrating the relationship between the Osaka Prognostic Score (OPS) and clinical outcomes in ST-elevation myocardial infarction (STEMI) patients treated with primary percutaneous coronary intervention (pPCI). OPS — based on serum C-reactive protein (CRP), albumin, and total lymphocyte count — was identified as an independent predictor of both contrast-induced nephropathy (CIN) and all-cause in-hospital mortality. CI, confidence interval; HRs, hazard ratios; OR, odds ratio.

Inflammation, nutritional status, and immune response are increasingly recognized as key contributors to both acute kidney injury and adverse cardiovascular events.^15-18 CRP is a well-established marker of systemic inflammation and is associated with increased infarct size, endothelial dysfunction, and higher mortality in STEMI patients.¹⁵ Hypoalbuminemia is traditionally considered a nutritional marker but also reflects systemic inflammation, and it has been linked with both CIN and cardiovascular mortality.^16,17 Lymphopenia, a surrogate of immune suppression, has been associated with a poor prognosis in patients with acute coronary syndrome (ACS) and may indicate an impaired response to stress or inflammation.¹⁸ While previous studies have evaluated these biomarkers individually, combining them into a single score, such as the OPS, provides an integrative assessment of the host response. A few prior studies have explored combinations of these parameters. For example, the CRP-to-albumin ratio has been associated with mortality in patients with ACS,¹⁹ and the systemic immune-inflammation index, which includes neutrophil, platelet, and TLC, has also shown prognostic value in patients with STEMI.²⁰ However, to our knowledge, none of these indices have been shown to predict both CIN and mortality simultaneously, nor have they been explicitly validated in a high-risk STEMI cohort undergoing emergent PCI. The OPS, through its balanced integration of inflammatory and nutritional markers, may offer a more comprehensive and clinically applicable risk-stratification tool.

Only one previous study has demonstrated an association between the OPS and CIN. The retrospective study by Özbeyaz and Algül²¹ included all patients with ACS undergoing PCI, and found an independent association between OPS and CIN. Although the patient cohort in their study differs from ours, our analysis, which includes a larger, more comprehensive sample, supports their findings. Established risk models for CIN such as the Mehran risk score primarily incorporate clinical and procedural variables, including contrast volume, intraaortic balloon pump use, and HF status.⁷ While these models have demonstrated robust predictive performance, they require multiple inputs that may not be readily available at admission. In contrast, the OPS is derived exclusively from routine laboratory parameters obtained at presentation, and may therefore serve as a complementary early risk assessment tool prior to coronary angiography. The association between OPS and CIN observed in our study is biologically plausible. Systemic inflammation promotes endothelial dysfunction, oxidative stress, and microvascular injury — key mechanisms in CIN development.²² Inflammatory biomarkers such as CRP and leukocytosis have been repeatedly linked to CIN risk in STEMI and PCI populations.^23,24 Similarly, hypoalbuminemia may reflect both poor baseline health and a proinflammatory milieu, reducing renal perfusion and resilience to contrast exposure.²⁵ The inclusion of TLC adds an immune function component, which may also play a role in CIN pathogenesis, although this aspect has been less studied.²⁶

Regarding mortality, previous research has linked each component of the OPS to both short- and long-term outcomes in STEMI. For example, several studies reported that both low albumin and TLC independently predicted mortality following acute myocardial infarction.^17,27 In addition, elevated CRP levels have been consistently associated with a worse prognosis, including reinfarction, HF, and death.²⁸ Nevertheless, the circulating levels of these biomarkers are modulated by numerous metabolic factors, and their prognostic implications may vary across diverse patient populations. Current risk stratification tools for mortality in ACS, including STEMI, such as the GRACE and TIMI risk scores, predominantly rely on clinical variables, hemodynamic status, and electrocardiographic findings.^29,30 Although these scores have been widely validated, their performance may be limited by the need for multiple clinical inputs and their relative insensitivity to acute inflammatory and nutritional states. In this context, inflammation-based indices have gained increasing attention as adjunctive prognostic tools. However, most of these indices have been evaluated solely with regard to mortality or cardiovascular outcomes, without simultaneous assessment of renal complications. Hence, using these parameters in isolation for outcome prediction may not consistently yield accurate results in clinical settings, whereas combining these three markers into a simplified categorical score, such as the OPS, enables efficient, real-time stratification in clinical settings. In this study, the OPS also effectively discriminated between patient risk across the full range of scores (0-3), with a consistent stepwise increase in CIN, mortality, need for hemodialysis, and failed PCI in the higher-score groups. These findings emphasize its potential utility in real-world triage, and the potential to identify high-risk patients on admission who may benefit from intensified monitoring, nephroprotective strategies, and multidisciplinary care. Notably, our study design ensured uniform PCI timing and optimal medical therapy across all patients, thereby reinforcing the robustness of the OPS as a predictive biomarker, independent of procedural variability. Another strength of our findings is the consistency across multiple statistical approaches including logistic and Cox regression models, ROC curves, RCS regression, and survival analyses, all of which demonstrated that higher OPS values corresponded to greater clinical risk. Notably, the nonlinearity of this relationship, as visualized by spline models, underscores the importance of even modest increases in OPS.

From a clinical perspective, the OPS is appealing for its simplicity. It relies exclusively on three widely available laboratory parameters without the need for complex calculations or subjective clinical input. This practicality enhances its utility, especially in high-volume, resource-constrained emergency settings. Moreover, because the OPS reflects a patient’s overall systemic condition, it may be useful not only for risk prediction but also for identifying candidates for early interventions, such as nephroprotection, closer hemodynamic monitoring, or modulation of nutritional and inflammatory factors. Moreover, the OPS offers practical advantages in the setting of acute STEMI. Unlike traditional risk scores, it relies solely on three routine, readily available laboratory markers, requires no subjective inputs, and integrates multiple physiologic domains. Its simplicity and objectivity make it highly suitable for use in emergency and resource-limited settings. Given its associations with both CIN and early mortality, the OPS may serve as a dual-purpose tool, aiding in decision-making for both renal protection and overall risk stratification.

Study limitations

Despite its strengths, this study has several limitations. First, its retrospective and single-center design may introduce selection bias and limit the generalizability of the findings to other populations or healthcare settings. Second, we did not perform a head-to-head comparison between the OPS and established risk models such as the Mehran score⁷ to predict CIN or the GRACE and TIMI scores^29,30 to predict mortality; future prospective studies should address these comparisons. Third, although the OPS components are routinely measured, their levels may still vary due to acute stress responses, hemodynamic instability, or other metabolic factors present during STEMI, which could influence risk estimation. Fourth, the timing of biomarker measurements was limited to admission values, and serial changes during hospitalization were not evaluated, which may provide additional prognostic insight. Fifth, as all participants were treated at a tertiary center with uniform PCI protocols, the applicability of the results to different clinical settings or lower-resource institutions remains uncertain. Finally, the follow-up period was confined to in-hospital outcomes, and long-term prognostic implications of the OPS after discharge were not assessed.

CONCLUSIONS

In this study of STEMI patients undergoing pPCI, the OPS emerged as a novel, independent predictor of both CIN and all-cause in-hospital mortality. The composite nature of the OPS in integrating inflammation, nutritional status, and immune response allows for a comprehensive assessment of systemic vulnerability, making it a practical and powerful tool in acute cardiovascular care. Incorporation of the OPS into early risk stratification protocols may facilitate timely clinical decision-making and the development of personalized care strategies. Further prospective and multicenter investigations are warranted to validate these findings and to explore the long-term prognostic utility of OPS in broader ACS populations.

DECLARATION OF CONFLICT OF INTEREST

All the authors declare no conflict of interest.

FUNDING

None.