The role of the modified Glasgow prognostic score in predicting postoperative mortality following open-heart surgery

Purpose

The Modified Glasgow Prognostic Score is an inflammation-based index utilizing C-reactive protein and albumin levels. It has been investigated in various diseases such as cancer, heart failure, and myocardial infarction. Open heart surgery, particularly with cardiopulmonary bypass, induces an acute-phase inflammatory response that is associated with postoperative complications and mortality. This study aimed to evaluate the prognostic value of preoperative and early postoperative Modified Glasgow Prognostic Score in predicting early outcomes in patients undergoing open heart surgery, with the goal of improving patient stratification and surgical decision-making.

Methods

A retrospective review of medical records was conducted for 456 patients who underwent open heart surgery between January 1, 2022, and October 1, 2024. All procedures were elective cases. Patients undergoing emergency operations, including those with acute type A aortic dissection, were excluded from the study. Data were collected on demographics, laboratory results, procedure duration, ejection fraction, and presence of multi-vessel disease. All patients included in the study underwent open heart surgery with the use of cardiopulmonary bypass (CPB). Patients who underwent off-pump coronary artery bypass grafting (OPCAB) or other procedures not requiring CPB were excluded. Both preoperative and postoperative Modified Glasgow Prognostic Score were analyzed to assess their association with early postoperative outcomes.

Results

Preoperative mGPS was found to be significantly associated with early mortality (p < 0.05). A one-unit increase in Modified Glasgow Prognostic Score was associated with a 2.85-fold increase in mortality risk (95% CI: 1.85–4.41). However, no significant relationship was observed between postoperative Modified Glasgow Prognostic Score and mortality (p > 0.05).

Conclusions

Preoperative Modified Glasgow Prognostic Score serves as a valuable prognostic marker for early mortality in open heart surgery patients. It provides independent predictive value and can enhance preoperative risk assessment, ultimately improving patient outcomes and guiding surgical management.

Key points

• a. What is known about the topic?

The modified Glasgow Prognostic Score (mGPS) is widely recognized as an inflammation-based prognostic scoring system, primarily used in cancer patients. It combines CRP and serum albumin levels to predict clinical outcomes, particularly in assessing systemic inflammation and its role in disease prognosis. Cardiopulmonary Bypass (CPB) used during open-heart surgeries is known to trigger a systemic inflammatory response, which can lead to postoperative complications and increased mortality. While mGPS has been validated in various settings, its utility in predicting postoperative mortality in cardiac surgery patients, especially those undergoing CPB, has not been extensively studied.

• b. What does this study add?

This study provides evidence that preoperative mGPS is significantly associated with early mortality in adult patients undergoing open-heart surgery with CPB. It suggests that mGPS can serve as a valuable preoperative screening tool for identifying patients at higher risk of mortality. Additionally, this study highlights mGPS as an independent predictor of poor clinical outcomes in the short term, emphasizing its potential role in risk assessment and the management of postoperative complications in cardiac surgery. The findings underscore the need for larger, multi-center studies to validate mGPS as a practical tool in the clinical management of cardiac surgery patients.

Introduction

Cardiovascular disease (CVD) remains the leading cause of death and disability worldwide, accounting for approximately 17.5 million deaths annually [1]. While advancements in medical and surgical techniques have improved outcomes, many patients with coronary artery disease and valvular heart disease still require surgical intervention [2]. Given the high morbidity and mortality associated with open-heart surgery, accurate preoperative risk stratification remains crucial—especially for high-risk populations.

Traditional risk factors such as age, heart failure, and renal function are commonly used to estimate postoperative outcomes, but these parameters may not fully capture the patient’s inflammatory status or nutritional reserve. Inflammatory responses induced by cardiopulmonary bypass can significantly affect recovery and outcomes. Therefore, identifying preoperative biomarkers that reflect systemic inflammation may improve the prediction of postoperative mortality and enhance surgical decision-making.

The Modified Glasgow Prognostic Score (mGPS), originally developed for oncologic patients, is an inflammation-based scoring system derived from serum C-reactive protein (CRP) and albumin levels [3–5]. It assigns a score of 0 if CRP ≤ 10 mg/L and albumin ≥ 35 g/L; 1 if CRP > 10 mg/L and albumin ≥ 35 g/L; and 2 if CRP > 10 mg/L and albumin < 35 g/L.⁶ The scoring system is based on serum CRP and albumin concentrations. A detailed summary of the mGPS scoring system is provided in Table 1 for clarity.

Table 1.

Components and scoring of the modified Glasgow prognostic score (mGPS)

CRP Level (mg/L)	Albumin Level (g/L)	mGPS Score
≤ 10	≥ 35	0
> 10	≥ 35	1
> 10	< 35	2

CRP: C-reactive protein; mGPS: Modified Glasgow Prognostic Score

The mGPS reflects systemic inflammation and nutritional status. Higher scores are associated with worse prognosis

This scoring system captures the severity of systemic inflammation and its prognostic implications. Combining CRP and albumin—both acute-phase proteins influenced by proinflammatory cytokines—provides a reliable reflection of the patient’s systemic response [6].

The prognostic value of mGPS has been validated in various conditions, including non-small cell lung cancer [7] and other inflammatory diseases [8]. Its simplicity, reproducibility, and cost-effectiveness make it an attractive tool for broader clinical use. However, its role in cardiovascular surgery remains underexplored.

This study aimed to evaluate the prognostic significance of preoperative and early postoperative mGPS in predicting short-term mortality in patients undergoing open-heart surgery. By assessing the utility of this inflammation-based index, we sought to improve risk stratification, guide clinical decision-making, and optimize postoperative management.

Materials and methods

Data collection

The reliability of these datasets was ensured through consistent data verification and standardization protocols across all involved departments.

• Study design: This retrospective, observational cohort study was conducted at Mersin University Faculty of Medicine Education and Research Hospital, a tertiary academic center providing cardiovascular surgery services. The study included consecutive patients who underwent open-heart surgery between January 1, 2022, and October 1, 2024. During this period, an annual average of 300 open-heart surgeries were performed at our facility, leading to an estimated total of 900 procedures over the study timeframe. A nested case-control design was employed within the cohort. Assuming an odds ratio of 1.5 for factors such as mGPS associated with mortality, a confidence interval width of 25% was accepted, resulting in a sample size of 456 patients. This study design enabled the identification of factors associated with postoperative mortality and facilitated comparisons between high- and low-risk patient groups. The number of deceased patients was matched to non-deceased patients at a 1:4 ratio. The mortality data presented in this study reflect all-cause mortality, including deaths due to cardiac and non-cardiac causes such as infection and multi-organ failure.

• Data collection: Data on patient demographics, laboratory test results, procedure duration, left ventricular ejection fraction (EF), and the presence of multi-vessel disease were reviewed. Additionally, intraoperative variables such as cardiopulmonary bypass (CPB) duration and aortic cross-clamp time were recorded, given their potential influence on postoperative inflammatory responses, including CRP levels. EF values were assessed preoperatively based on transthoracic echocardiography performed within one week before surgery. The Exitus group included patients who experienced mortality within 30 days postoperatively. Patients who survived beyond this period were classified under the Alive group. Demographic and clinical variables were carefully reviewed to ensure the accuracy and reliability of the data used for prognostic assessments. Venous blood samples were collected preoperatively upon admission and daily postoperatively in ethylenediaminetetraacetic acid (EDTA) tubes. Blood samples for biochemical analysis were collected at hospital admission (preoperative), and on postoperative days 0, 1, and 2 to monitor early postoperative trends and assess potential complications. Complete blood count (CBC) and biochemical measurements were performed at multiple time points using an automated hematology analyzer. These measurements allowed for the identification of trends and potential deviations in key biomarkers that could affect patient outcomes.

Data analysis

• Statistical analysis: Continuous variables were presented as mean ± standard deviation and minimum-maximum values, while median (25-75% interquartile range) was used for ordinal variables. Frequencies and percentages were used to describe categorical variables. Multivariate analysis was performed to identify independent predictors of mortality, adjusting for potential confounders such as age, comorbidities, and surgical parameters. Mortality-related comparisons of continuous variables were conducted with Student’s t-test, while the Mann-Whitney U test was used for ordinal mGPS scores. Paired t-tests were used for repeated continuous measurements, and the Wilcoxon test was employed for ordinal variables. Chi-square tests were applied to assess relationships between mortality status and categorical variables. For factors thought to be associated with mortality, odds ratios with 95% confidence intervals were calculated. This comprehensive approach was aimed at ensuring robust statistical power and minimizing the risk of Type I and Type II errors. Statistical significance was set at p < 0.05.

• Software: Statistical analyses were conducted using IBM SPSS 21 and MedCalc statistical software packages.

Results

A total of 456 diagnosed patients were included in the study. Their baseline characteristics and clinical data are presented in Table 2. The distribution of surgical procedures among these 456 patients is as follows: Coronary artery bypass grafting (CABG) 65% (290 patients), Valve surgery (isolated or combined) 20% (90 patients), Aortic surgery 10% (45 patients), Other complex procedures 5% (21 patients). All patients underwent elective procedures. None of the aortic surgery cases included patients with acute type A aortic dissection. This procedural distribution highlights the diverse surgical interventions performed at our facility and their potential impact on postoperative outcomes.

Table 2.

Distribution of Socio-Demographic characteristics in patients undergoing Open-Heart surgery in cardiovascular surgery (n = 456)

Identifying features	Mean ± SD	Median(Min-Mak)
Age (year)	66.5 ± 11.2	68(28–88)
	Count (n)	Percentage (%)
Gender
Male	310	68
Female	146	32
DM
No	169	37.1
Yes	287	62.9
HT
No	272	59,6
Yes	184	40,4
Mortality
Alive	355	77,9
Exitus	101	22,1
		Median (Min-Maks.)
EF	50.8 ± 8.1	52(27–66)
PREOP		Median (Min-Maks.)
Creatinine(mg/dL)	0.86 ± 0.62	1.05(0.50–10.2)
Ure (mg/dL)	37.63 ± 17.5	38.5(18.2-120.7)
Neu(103mcl)	6.05 ± 1.75	5.50(1.10–16.3)
Lym(103mcl)	2.18 ± 0.91	2.05(0.30–6.2)
Plt(103mcl)	245.5 ± 69.2	240(85–525)
Crp(mg/l)	26.7 ± 20.3	10.5(0.50–420.0)
Albumın(mg/l)	36.9 ± 4.22	37.2(23.8–45.9)
Preop mGPS	0.66 ± 0.43	0(0–2)
	Mean ± SD	Median (Min-Maks.)
CPB Duration (minutes)	125.37 ± 34.01	114.7(35–267)
Cross-clamp Duration (minutes)	74.76 ± 29.11	65.6(20–158)

SD: Standard Deviation, p-value: Student’s t-test was used for continuous variables, and the Chi-Square test was used for categorical variables. CRP: C reactive protein, PLT: Platelets, NEU: Neutrophil, LYM: lymphocyte

According to Table 2, the study included 456 patients who underwent open-heart surgery. The median age was 68 years (range: 28–88), and 68% of the participants were male. Diabetes mellitus was present in 62.9% of the patients, and 40.4% had hypertension. The overall 30-day mortality rate was 22.1%. The median preoperative ejection fraction (EF) was 52% (range: 27–66).

Preoperative laboratory findings revealed a median creatinine level of 1.05 mg/dL (range: 0.50–10.2), a median urea level of 38.5 mg/dL (range: 18.2–120.7), and a median CRP level of 10.5 mg/L (range: 0.50–420.0). The median albumin level was 37.2 g/L (range: 23.8–45.9). The median neutrophil and lymphocyte counts were 5.50 × 10³/mcL and 2.05 × 10³/mcL, respectively, while the median platelet count was 240 × 10³/mcL.

The median duration of cardiopulmonary bypass (CPB) was 114.7 min (range: 35–267), and the median aortic cross-clamp time was 65.6 min (range: 20–158).

The preoperative mGPS ranged from 0 to 2, with a median score of 0. Among these parameters, preoperative urea, CRP, albumin, and mGPS were found to be significantly associated with mortality (p < 0.05). No statistically significant association was observed between preoperative creatinine, neutrophil, lymphocyte, or platelet counts and mortality (p > 0.05).

According to Table 3, several parameters showed statistically significant differences between the Alive and Exitus groups. Preoperative ejection fraction (EF) was significantly lower in the Exitus group (49.7 ± 9.2) compared to the Alive group (54.3 ± 7.1) (p = 0.007). Diabetes mellitus was significantly more prevalent in the Exitus group (p < 0.001). Gender was also significantly associated with mortality (p = 0.025), with a higher proportion of males in the Alive group.

Table 3.

Evaluation of differences and associations in Socio-Demographic and biochemical measurements by mortality status (n = 456)

		Alive (n = 355)	Exitus (n = 101)
Features		Mean ± SD	Mean ± SD	p-value
Age (year)		62.9 ± 10.2	65.8 ± 12.4	0.48*
EF		54.3 ± 7.1	49.7 ± 9.2	0.007*
Pre-Creatinine(mg/dL)		0.92 ± 0.53	1.38 ± 0.41	0.18*
Post-Creatinine(mg/dL)		0.92 ± 0.53	1.45 ± 0.63	< 0.001*
p value**		0.003	< 0.001
Pre-Ure (mg/dL)		37.9 ± 15.4	44.3 ± 14.9	0.015*
Post-Ure (mg/dL)		39.4 ± 12.8	55.2 ± 22.5	< 0.001*
p value**		< 0.001	< 0.001
Pre-NEU(103mcL)		5.42 ± 2.35	5.81 ± 2.51	0.68*
Post- NEU(103mcL)		9.85 ± 3.75	13.02 ± 5.42	< 0.001*
p value**		< 0.001	< 0.001
Pre-LYM(103mcL)		2.08 ± 0.72	2.21 ± 1.18	0.41*
Post-LYM(103mcL)		2.12 ± 0.45	2.49 ± 0.99	0.019*
p value**		< 0.001	< 0.001
Pre-PLT(103mcL)		245.3 ± 68.4	230.6 ± 73.2	0.57*
Post-PLT(103mcL)		152.1 ± 46.9	140.3 ± 69.7	0.09*
p value**		< 0.001	< 0.001
Pre-CRP(mg/L)		18.1 ± 16.8	27.8 ± 21.1	0.33*
Post-CRP(mg/L)		148.1 ± 55.2	136.9 ± 52.1	0.26*
p value**		< 0.001	< 0.001
Pre-Albumin(mg/L)		39.2 ± 3.85	34.8 ± 5.35	0.002*
Post-Albumin(mg/L)		29.1 ± 12.2	24.2 ± 5.1	0.018*
p value**		< 0.0001	< 0.0001
		Median(Min-Mak)	Median(Min-Mak)
Preop mGPS		0(0–1)	1(0–2)	< 0.001*****
Postop mGPS		2(2–2)	2(2–2)	0.48*****
p value****		< 0.001	0.001
		n(%)	n(%)
Gender	Male	254(71.5)	56(55.4)	0.025***
	Female	101(28.5)	45(44.6)
DM+		223(77,7)	64(22,3)	< 0.001***
HT+		145(78,8)	39(21,2)	0.13***

*Student’s t test, **Paired t test, ***Ki-Kare test****Wilcoxon test, *****Mann-Whitney U test, ( p < 0.05 anlamlılık), Student’s t-test was used for continuous variables, the paired t-test for preoperative and postoperative comparisons, the Chi-Square test for categorical variables, the Mann-Whitney U test for ordinal variables, and the Wilcoxon test for preoperative and postoperative ordinal comparisons. A p-value < 0.05 was considered statistically significant

Biochemical parameters revealed that preoperative urea (p = 0.015), postoperative creatinine (p < 0.001), postoperative urea (p < 0.001), and postoperative neutrophil count (p < 0.001) were significantly higher in the Exitus group. Postoperative lymphocyte count (p = 0.019) and postoperative albumin (p = 0.018) also differed significantly between the groups. Preoperative and postoperative albumin levels, as well as preoperative and postoperative mGPS, were significantly associated with mortality (p < 0.05).

There was no statistically significant difference between the Alive and Exitus groups regarding preoperative creatinine, preoperative CRP, preoperative neutrophils, preoperative lymphocytes, preoperative platelets, postoperative CRP, and postoperative platelets (p > 0.05). Hypertension did not show a significant association with mortality (p = 0.13). Postoperative mGPS did not differ significantly between groups (p = 0.48).

According to Table 4, logistic regression analysis revealed that ejection fraction (EF) was significantly associated with mortality, with an odds ratio of 0.91 (95% CI: 0.88–0.94; p = 0.001). Male gender was also significantly associated with increased mortality risk (OR: 2.05; 95% CI: 1.25–3.45; p = 0.012). The presence of diabetes mellitus was significantly related to mortality (OR: 3.15; 95% CI: 2.02–5.10; p < 0.001).

Table 4.

Evaluation of the relationship between mortality, age, gender, and chronic Diseases(n = 456)

Variables	Odds ratio	95% CI	p-value
Age	1.15	1.01–1.09	0.42
Ejection Fraction (EF)	0.91	0.88–0.94	0.001
Gender (Risk: Male)	2.05	1.25–3.45	0.012
Diabetes Mellitus (DM) (Risk:1)	3.15	2.02–5.10	< 0.001
Hypertension (HT) (Risk:1)	1.72	1.05–2.85	0.35

Age and hypertension did not show statistically significant associations with mortality (p = 0.42 and p = 0.35, respectively).

According to Table 5, preoperative urea levels were significantly associated with mortality (OR: 1.08; 95% CI: 1.01–1.06; p = 0.014). Preoperative albumin levels were also significantly related to mortality (OR: 0.75; 95% CI: 0.70–0.84; p < 0.001). In addition, preoperative mGPS showed a significant association with mortality (OR: 2.85; 95% CI: 1.85–4.41; p < 0.001).

Table 5.

Evaluation of the relationship between preoperative biochemical parameters and mortality (n = 456)

Variables	Odds ratio	95% CI	p-value
Pre-Creatinine(mg/dL)	1.35	0.95–1.77	0.22
Pre-Ure (mg/dL)	1.08	1.01–1.06	0.014
Pre-NEU(103mcL)	1.10	0.96–1.24	0.45
Pre-LYM(103mcL)	1.30	0.88–1.79	0.19
Pre-PLT(103mcL)	0.93	0.91–1.02	0.48
Pre-CRP(mg/L)	1.05	0.97–1.02	0.11
Pre-Albumin(mg/L)	-0.75	0.70–0.84	< 0.001
Pre-mGPS	2.85	1.85–4.41	< 0.001

No statistically significant associations were found for preoperative creatinine, neutrophil count, lymphocyte count, platelet count, or CRP levels (p > 0.05).

According to Table 6, several postoperative biochemical parameters were significantly associated with mortality. These included postoperative creatinine (OR: 3.1; 95% CI: 1.80–5.2; p < 0.001), urea (OR: 1.12; 95% CI: 1.06–1.14; p < 0.001), neutrophil count (OR: 1.28; 95% CI: 1.15–1.39; p < 0.001), lymphocyte count (OR: 2.1; 95% CI: 1.45–3.05; p = 0.001), platelet count (OR: 0.90; 95% CI: 0.88–0.94; p = 0.012), and albumin levels (OR: 0.80; 95% CI: 0.60–0.73; p < 0.001).

Table 6.

Evaluation of the relationship between postoperative biochemical parameters and mortality (n = 456)

Variables	Odds ratio	95% CI	p-value
Post-Creatinine(mg/dL)	3.1	1.80–5.2	< 0.001
Post-Ure (mg/dL)	1.12	1.06–1.14	< 0.001
Post- NEU(103mcL)	1.28	1.15–1.39	< 0.001
Post-LYM(103mcL)	2.1	1.45–3.05	0.001
Post-PLT(103mcL)	0.90	0.88–0.94	0.012
Post-CRP(mg/L)	1.08	1.02–1.13	0.18
Post-Albumin(mg/L)	-0.80	0.60–0.73	< 0.001
Post-mGPS	1.01	0.01–1.02	0.99

No statistically significant association was found between mortality and postoperative CRP (p = 0.18) or postoperative mGPS (p = 0.99).

Discussion

In this single-center retrospective study, it was determined that preoperative mGPS is a good predictor of mortality after open-heart surgery. mGPS has proven to be a robust indicator for predicting mortality risk, and its integration into clinical practice could markedly enhance risk stratification in cardiac surgeries.

Cardiac surgeries performed under cardiopulmonary bypass CPB frequently trigger systemic inflammatory responses that contribute to postoperative morbidity and prolonged hospitalization [3]. Despite the growing use of off-pump CABG and endovascular approaches, CPB remains essential for many procedures. Although the safety profile of CPB has improved, controlling inflammation after cardiac surgery is still inadequate [9]. Kirklin’s early work in 1980 introduced the concept of a systemic inflammatory response caused by CPB, coining the term “post-perfusion syndrome,” which is now considered a subset of systemic inflammatory response syndrome [10, 11].

Our findings confirm the significance of systemic inflammation by showing a strong association between preoperative inflammatory biomarkers and postoperative mortality. In particular, elevated CRP levels were significantly associated with increased mortality, in line with previous studies. For instance, Agda Mezzomo et al. reported that high preoperative CRP levels were linked to readmission and cardiovascular mortality [12]. Likewise, CRP remains one of the most sensitive markers in acute myocardial infarction for detecting necrosis and inflammatory activity [13]. Eissa et al. also demonstrated the predictive role of CRP in in-hospital mortality in patients undergoing cardiac surgery [4].

Albumin plays a dual role as a negative acute-phase reactant and a nutritional marker. During systemic inflammation, albumin synthesis decreases while catabolism increases, resulting in lower serum levels [5]. These mechanisms explain the frequent association of hypoalbuminemia with adverse outcomes after surgery. Our findings confirm that low preoperative albumin levels are significantly associated with higher mortality, which supports prior literature showing similar results [14–19].  A study indicated that among patients with preoperative albumin levels below 30 g/L, mortality was approximately 36.2% and the need for reoperation due to bleeding exceeded 32%[20]. Preoperative albumin < 30 g/L has also been linked to prolonged ICU and hospital stays and increased mortality [21]. These data underscore albumin’s importance beyond nutrition, reinforcing its utility as an inflammatory biomarker [22].

Based on these two biomarkers—CRP and albumin—the modified Glasgow Prognostic Score (mGPS) was initially developed in oncology but has gained attention in cardiac and vascular surgery. It offers a simple, reproducible, and cost-effective approach for stratifying inflammatory risk [6, 23, 24]. The mGPS is calculated as follows: CRP ≤ 10 mg/L and albumin ≥ 35 g/L scores 0; CRP > 10 mg/L and albumin ≥ 35 g/L scores 1; CRP > 10 mg/L and albumin < 35 g/L scores 2 [6, 25].

Although the mGPS provides a simple and cost-effective method for assessing inflammatory burden, it does not incorporate surgical or clinical variables such as age, ejection fraction, comorbidities, or procedure type. In contrast, established surgical risk models like the Society of Thoracic Surgeons (STS) score and the European System for Cardiac Operative Risk Evaluation (EUROScore) offer a comprehensive approach by integrating a wide range of perioperative factors. However, these scores are more complex to calculate and do not focus primarily on inflammatory status. The integration of mGPS with established models such as STS or EUROScore may potentially enhance risk stratification by providing both inflammatory and procedural risk perspectives. Future studies are warranted to investigate the additive value of combining mGPS with conventional scoring systems.

The prognostic performance of mGPS has been validated in non-cancer populations as well. Anna Cho et al. found that in patients with heart failure, an elevated mGPS was associated with a 3.0 to 3.5 times increased risk of death over three years, independent of age and NT-proBNP levels [26]. Our findings are in line with this, showing that each unit increase in preoperative mGPS is associated with a 2.85-fold increase in mortality. Similarly, J. Li et al. observed that patients with higher mGPS scores undergoing percutaneous coronary intervention (PCI) had significantly worse long-term mortality outcomes [6]. Jia et al. and others extended this evidence to STEMI and AMI populations, demonstrating that mGPS and GPS were independent predictors of in-hospital and long-term mortality [27–29].

Another explanation for the prognostic utility of mGPS lies in the inflammatory triggers inherent in cardiac surgery, such as ischemia, CPB, and surgical trauma [30]. Although mGPS does not replace a clinical diagnosis, it reflects systemic inflammatory stress and may identify vulnerable patients preoperatively. Our results showed that postoperative mGPS values remained around 2 in survivors due to expected high CRP and low albumin levels in the early postoperative period. However, in non-survivors, mGPS values were elevated preoperatively, despite normal WBC, neutrophil, and CRP levels—suggesting that mGPS may capture subtle systemic inflammation not otherwise apparent in isolated lab values.

In addition to inflammation-based markers such as CRP and albumin, myocardial damage biomarkers—particularly high-sensitivity Troponin T (hs-TnT) and N-terminal pro-B-type natriuretic peptide (NT-proBNP)—have also been identified as valuable predictors of postoperative complications in cardiac surgery. A recent prospective study by Duchnowski et al. (2024) involving patients undergoing valve surgery demonstrated that elevated preoperative NT-proBNP and postoperative hs-TnT levels were independently associated with postcardiotomy cardiogenic shock requiring mechanical circulatory support (MCS). These biomarkers reflect myocardial stress and injury, providing additional insights into postoperative risk, especially in hemodynamically vulnerable patients [31]. Although not evaluated in our cohort, future studies integrating these markers alongside mGPS may further improve early postoperative risk stratification.

Moreover, our data suggest that mGPS might be particularly valuable in detecting patients who are not flagged by conventional risk scores but still have poor postoperative outcomes. In our cohort, some patients with high preoperative mGPS had no overt inflammatory or infectious conditions, suggesting the score’s sensitivity in capturing subclinical risk.

Given that surgical complexity is a significant contributor to postoperative morbidity and mortality, the inclusion of procedural details enhances the overall interpretation of our findings. Future research with larger cohorts may further elucidate the impact of different surgical techniques on Modified Glasgow Prognostic Score (mGPS)-related risk stratification.

In our series, CABG accounted for approximately 65% of the procedures, followed by valve surgeries (20%), aortic surgeries (10%), and complex hybrid operations (5%). The role of surgical invasiveness and its potential interaction with mGPS is an area warranting further study. For instance, inflammation and nutritional status may have greater predictive power in high-risk, complex operations compared to isolated CABG.

Our analysis also revealed that mGPS was more predictive preoperatively than postoperatively. This may be due to the universal rise in CRP and drop in albumin following surgery, leading to an expected elevation in postoperative mGPS scores in most patients. Thus, postoperative mGPS may lose its discriminatory power in this setting.

In conclusion, this study demonstrated that preoperative mGPS is an independent and significant predictor of mortality in patients undergoing open-heart surgery. The ease of measuring CRP and albumin, and the established scoring structure of mGPS, make it an attractive adjunct to standard preoperative assessments. While postoperative inflammatory markers also rise, preoperative values provide the greatest prognostic utility. Combining mGPS with additional parameters—such as cardiac biomarkers and established risk scores—may further enhance patient stratification, reduce perioperative risk, and guide clinical decision-making.

Limitations of the study

This study has several potential limitations. Although the analysis included a relatively large patient group, the single-center observational retrospective design inherently carries some constraints. This design limited significant subgroup analyses, particularly in high preoperative risk groups such as patients with heart failure, those with a high EuroSCORE, and patients undergoing emergency or reoperation. Additionally, while the total sample size of 456 patients provides preliminary insights, it may not be sufficient to draw definitive conclusions about all subgroups. The absence of enriched representation from high-risk preoperative categories—such as patients with severely reduced ejection fraction, ongoing myocardial ischemia, or urgent surgical indications—may limit the applicability of the findings to broader or more complex surgical populations. Furthermore, the single-center nature of the study introduces potential institutional bias and may restrict the generalizability of the results across different settings or surgical teams. Therefore, larger-scale studies involving these high-risk patient groups are needed.

Additionally, this study was unable to elucidate the underlying pathophysiology between the preoperative modified Glasgow Prognostic Score and early complications in adult cardiac surgery. Surgical mortality rates can vary significantly depending on the type of cardiac procedure performed, as different interventions carry distinct levels of complexity and associated risks. Therefore, the generalizability of our findings may be limited when applied to different surgical subgroups. We aimed to minimize bias by using multivariate logistic regression analysis to account for variables that could affect early clinical outcomes in cardiac surgeries performed with CPB (Cardiopulmonary Bypass). The study also sought to reduce inter-center variability by analyzing patients who were operated on using the same techniques by the same experienced surgical team.

Moreover, we only assessed overall mortality without an in-depth analysis of the specific causes of death. Furthermore, the study did not include a detailed analysis of the specific causes of death or individual postoperative complications. Mortality was recorded as all-cause mortality without classification into cardiac, infectious, or other systemic etiologies. This limits the ability to distinguish between different clinical pathways leading to poor outcomes. Future studies incorporating cause-specific mortality and complication profiling may provide deeper insights into the mechanisms driving early postoperative mortality. To confirm whether mGPS is a simple and effective marker in clinical practice and to determine whether it influences clinical outcomes, larger-scale randomized controlled trials are necessary.

Conclusions

In conclusion, mGPS may be a useful preoperative screening tool for predicting early mortality in adult patients undergoing open-heart surgery and has also been found to be an independent determinant of poor clinical outcomes. In patients undergoing open-heart surgery with CPB, preoperative mGPS was significantly associated with mortality, showing a meaningful relationship with short-term mortality. Incorporating mGPS in preoperative assessments not only has the potential to predict early mortality but also serves as a guide for tailored interventions that could mitigate postoperative complications. This highlights its valuable role in risk assessment, particularly in managing complications that contribute to increased postoperative mortality.

Author contributions

Burak Toprak: Made substantial contributions to the study design, manuscript writing, conceptualization, and execution of the research, contributed to data collection, performed data analysis, substantially revised the manuscript, and critically interpreted the results. Abdulkadir Bilgiç: Contributed to the conceptualization phase of the research and the study design.

Funding

No financing available.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Ethical approval for the study was obtained from the Mersin University Ethics Committee with the decision numbered 2024/982 and dated 16/10/2024. Committee.

Consent for publication

Written informed consent for publication was obtained from all patients or their legal guardians where applicable.

Competing interests

The authors declare no competing interests.

Declaration of Helsinki

The study and the writing of the article were prepared in accordance with the Declaration of Helsinki.

Statement on the use of Artificial Intelligence

No artificial intelligence application was used.