The merits of the prognostic nutritional index and neutrophil-to-lymphocyte ratio in predicting outcomes of coronary artery bypass grafting: a retrospective cohort study

Abstract

Background

Coronary artery bypass grafting (CABG) carries a risk for serious adverse events. The Prognostic Nutritional Index (PNI) and Neutrophil-to-Lymphocyte Ratio (NLR), are simple measures of nutritional and inflammatory statuses, have been proposed as a potential predictor of postoperative outcomes. This study aimed to assess the predictive performance of preoperative PNI and NLR and their dynamic changes for post-CABG outcomes.

Methods and Materials

The study investigated the medical records of 327 patients who underwent CABG between February 20, 2020, and February 19, 2025.

Results

The overall complication and mortality rates were 15%, and 3.36% respectively. The preoperative PNI decreased significantly (P1 < 0.001) at 24-h postoperative and at the time of discharge. At the time of discharge, the mean PNI value increased significantly (P2 < 0.001) compared to that determined at 24-h postoperative. The magnitude of ∆PNI at discharge was significantly smaller than that at 24-h postoperatively (P3 < 0.001). Multivariate binary logistic regression identified preoperative PNI as an independent predictor of 1-month mortality, while ΔPNI at 24 h was independently associated with 1-month mortality.

Conclusion and recommendations

PNI is a simple and inexpensive marker that may serve as a potential predictor of short-term outcomes after CABG. Lower preoperative PNI and greater early postoperative PNI decline were associated with adverse short-term outcomes after CABG, whereas preoperative NLR did not retain independent significance after multivariable adjustment; however, further prospective multicenter studies are required to confirm these findings.

Introduction

Coronary artery bypass grafting (CABG) is the most performed major open-heart operation, with an increasing global frequency, especially in developed countries [1]. Despite advancements in clinical practice, CABG is still a complex and high-risk operation [2].

CABG carries a significant risk for serious adverse events, bleeding requiring further surgery, perioperative mortality and major adverse cardiac and cerebrovascular events [3], cardiac arrhythmias, and atrial fibrillation (AF) [4], and deep sternal wound infection [5], which are the most common perioperative complications, and each has an approximate risk of 1–2% [6].

Early consultation and swift diagnosis of these conditions are crucial for optimizing the patient’s hemodynamics and preventing early peri-CABG mortality [7]. However, preoperative risk prediction in patients scheduled for CABG is vital to provide an estimated risk of early postoperative complications and death [8]. Furthermore, risk factors and the duration of exposure to these risks can help improve existing assessment tools and create tailored preventive strategies [9].

Inflammation and malnutrition are related to adverse clinical outcomes in patients with coronary artery disease (CAD) [10]. Malnutrition is an independent risk factor for worse prognosis of CAD, with increased all-cause mortality and major adverse cardiovascular events by 72% and 47%, especially in patients with stable CAD [11].

Anthropometric measures, biochemistry-based nutrition status, and blood inflammatory markers constitute various indices, which are touted as early-warning measures with relevant preventive interventions for patients at risk of malnutrition and as significant indicators of disease prognosis [12].

The Prognostic Nutritional Index (PNI) is a simple composite score derived from peripheral lymphocyte counts and albumin levels [13]. PNI has been widely used as a nutritional metric in patients undergoing cardiac surgery because of its ability to incorporate both nutritional and inflammatory statuses. The predictability of preoperative and postoperative PNI and its dynamic changes for CABG outcomes require being evaluated [14].

Complementing these indices, the neutrophil-to-lymphocyte ratio (NLR) has emerged as an established inflammatory biomarker and an independent cardiovascular risk predictor [15]. NLR reflects the delicate balance between neutrophil-mediated innate inflammation and lymphocyte-mediated adaptive immune regulation. Elevated preoperative NLR has been strongly associated with an increased risk of postoperative complications -such as atrial fibrillation and acute kidney injury- as well as increased 30-day and long-term mortality following CABG [16]. By capturing the interplay between non-specific inflammatory responses and physiological stress, NLR provides a cost-effective and readily accessible tool for personalized risk stratification in cardiac surgery [17].

The study aimed to assess the performance of preoperative PNI and NLR and their dynamic changes secondary to surgical intervention in predicting the outcomes of patients planned to undergo CABG.

Methodology

Study design

A retrospective cohort study was conducted evaluating medical records of patients who underwent elective CABG surgeries.

Study area and period

The study evaluated CABG patients’ medical records over the preceding five years from February 20, 2020, to February 19, 2025, in Department of Cardiothoracic surgery, Faculty of Medicine, Benha University, Egypt.

The study protocol was submitted to the Research Ethics Committee (REC), Faculty of Medicine, Benha University, Egypt, on 6 November 2025, when the protocol reference number RC 6-11-2025 was assigned. Final ethical approval was granted in March 2026.

Population and Eligibility criteria

All records were explored to ensure fulfillment of the inclusion requirements and were free of exclusion criteria for CABG under the conditions of the current analysis. These exclusion criteria included other cardiac surgeries, the presence of malignancy, systemic inflammatory disorders, active infection, autoimmune disorders, maintenance on immunosuppressants for any indication, liver diseases, and patients who received exogenous albumin infusion during the perioperative period were also excluded, as exogenous albumin directly elevates serum albumin levels and would artificially inflate postoperative PNI values, thereby confounding the assessment of nutritional status. Patients with incomplete baseline records for eligibility and preoperative PNI and NLR calculation were excluded from the primary cohort.

Sample size, technique and procedure

The minimum sample size was estimated using Epi Info (Epidemiological Information Package), version 7.2.5.0, using a single-proportion approach. Based on the postoperative poor-outcome frequency of 31.3% reported by Alagha et al. (2023), with an absolute margin of error of 5% and a confidence level of 80%, the estimated minimum sample size was approximately 141 patients. As this was a retrospective cohort study, all eligible patient records available during the predefined study period were screened to maximize the available sample and improve the precision of the estimates. Of 358 records screened, 31 were excluded according to the predefined eligibility criteria or because of incomplete baseline data, leaving 327 patients for the final analysis.

Data collection tool and procedure

Age, sex, body weight and height, the presence of medical disorders, whether controlled or not, smoking status, and the preoperative grade according to the American Society of Anesthesiologists (ASA) were determined. Cardiac disease-related data, including the preoperative functional grade according to the New York Heart Association (NYHA), ejection fraction (EF), number of vessels to be revascularized, previous PCI, and its findings were retrieved. Comorbidities are defined as: diabetes (IDDM or NIDDM by treatment regimen), hypertension (systolic BP ≥ 140 mmHg or diastolic ≥ 90 mmHg, or antihypertensive medication use), smoking (non-smoker, ex-smoker ≥ 12 months cessation, or current), prior cardiac interventions (previous PCI or CABG), and joint disorders (any musculoskeletal condition that may affect rehabilitation). The hypertension prevalence of 29.7% reflects formally documented diagnoses at the time of referral; patients with controlled hypertension not formally re-documented at referral may not have been captured. The risk for mortality according to the collected data was calculated using the European System for Cardiac Operative Risk Evaluation (EuroSCORE) II. Operative data and findings included the durations of ischemia due to aortic clamping, cardiopulmonary bypass (CPB), and surgery. Also, the data concerning the amount of intraoperative blood loss, the frequency of transfusion of blood or its products, the incidence of intraoperative complications and mortality were obtained. Postoperative data included the duration of ICU and hospital stays, the frequency and severity of complications, and mortalities encountered during the first 24 h and one month after surgery were determined. The reported morbidity rate represents the proportion of patients who experienced at least one complication during the follow-up period. Although individual patients may have experienced more than one complication, each patient was counted only once for the overall morbidity rate. For the statistical analyses presented in Table 4, patient-level binary classification was used, with patients categorized as having at least one event versus having no event, rather than using event-level counting. Accordingly, the reported 15% overall morbidity rate represents the proportion of the 327-patient cohort who experienced at least one adverse event during the entire follow-up period. Acute kidney injury (AKI) was defined and classified according to the KDIGO (Kidney Disease: Improving Global Outcomes) criteria based on serum creatinine rise and urine output. Postoperative atrial fibrillation (POAF) was defined as any new-onset atrial fibrillation requiring treatment occurring within 30 days of surgery. Stroke was defined as a new neurological deficit persisting beyond 24 h, confirmed clinically. Prolonged mechanical ventilation was defined as ventilator dependence exceeding 48 h postoperatively. Wound infection was classified as superficial or deep according to the CDC (Centers for Disease Control and Prevention) surgical site infection criteria. Respiratory infection was defined as new consolidation on chest imaging with clinical signs of pneumonia requiring antibiotic treatment.

Table 4.

Univariate comparison of clinical variables between patients with and without perioperative complications and mortality following CABG surgery (n = 327). Mann-Whitney U test for continuous variables; chi-square test for categorical variables:

Variable	Events Median [Q1–Q3] or %	No Events Median [Q1–Q3] or %	P-value	Sig.
Intraoperative Morbidity (n = 6 events / 321 no-events)
Age (years)	67 [61–68.5]	59 [56–65]	0.034	*
BMI (kg/m²)	35.75 [33.8–36.29]	29.76 [27.61–31.91]	0.010	**
Preoperative PNI	41.98 [39.69–43.62]	56.35 [51–60.55]	< 0.001	***
Preoperative NLR	2.98 [2.85–3]	1.36 [1.09–2.08]	< 0.001	***
CPB time (min)	67.5 [56.25–75]	75 [65–85]	0.033	*
Operative time (min)	210 [199–210]	190 [170–210]	0.002	**
Intraoperative Mortality (n = 2 events / 325 no-events)
Age (years)	78 [77.5–78.5]	59 [56–66]	0.015	*
EuroSCORE II	3.94 [3.94–3.94]	1.07 [0.93–2.26]	0.021	*
Preoperative PNI	41 [38.65–43.35]	56.25 [50.45–60.3]	0.026	*
Preoperative NLR	2.91 [2.89–2.92]	1.36 [1.11–2.09]	0.020	*
Ischemia time (min)	75 [71.00–75]	55 [50–65]	0.021	*
Hospital Postoperative Morbidity (n = 22 events / 305 no-events)
Age (years)	67.00 [63.50–68.75]	59 [56–65]	< 0.001	***
BMI (kg/m²)	31.76 [29.13–33.97]	29.76 [27.46–31.91]	0.021	*
Preoperative PNI	45.55 [44.20–47.17]	56.80 [53.10–61.10]	< 0.001	***
Preoperative NLR	2.05 [1.60–2.57]	1.34 [1.09–2.08]	0.003	**
Hospital Postoperative Mortality (n = 4 events / 323 no-events)
Age (years)	68 [65–71.25]	59 [56–65.5]	0.031	*
BMI (kg/m²)	36.22 [35.36–36.49]	29.76 [27.54–31.91]	0.002	**
EuroSCORE II	3.19 [2.66–3.38]	1.07 [0.93–2.26]	0.036	*
Preoperative PNI	43.2 [41.52–44.7]	56.25 [50.67–60.42]	0.002	**
Preoperative NLR	2.99 [2.92–3]	1.36 [1.10–2.09]	< 0.001	***
Operative time (min)	210 [190–210]	190 [170–210]	0.010	*
1-Month Postoperative Morbidity (n = 49 events / 278 no-events)
Preoperative PNI	54.45 [49.40–58.6]	56.58 [51.11–60.89]	0.085	NS
Female gender	44.9%	26.6%	0.015	*
1-Month Postoperative Mortality (n = 11 events / 316 no-events)
Age (years)	65 [64–67.50]	59 [56 − 65.25]	0.023	*
BMI (kg/m²)	33.51 [30.81–34.06]	29.76 [27.46–31.91]	0.009	**
EuroSCORE II	2.26 [2–3.19]	1.07 [0.93–2.26]	0.015	*
Preoperative PNI	46.65 [44.38–54.8]	56.45 [50.96–60.76]	0.001	**
Preoperative NLR	1.97 [1.46–2.43]	1.36 [1.09–2.09]	0.054	NS
Ischemia time (min)	65 [60–72.5]	55 [50 − 65]	0.003	**
CPB time (min)	90 [77.5–90]	75 [65 − 85]	0.044	*
Operative time (min)	210 [200 − 210]	190 [170 − 210]	0.008	**
ΔPNI at 24 h (%)	−27.21 [− 34.89–−23.85]	−16.31 [− 22.56–−11.89]	< 0.001	***
ΔNLR at 24 h (%)	+ 25.61 [22.05–29.47]	+ 18.59 [14.61–22.58]	0.002	**
Female gender	63.6%	28.2%	0.028	*

Data presented as median [Q1–Q3] for continuous variables and percentage for categorical variables. P-values derived from Mann-Whitney U test (continuous) and chi-square test (categorical)

PNI Prognostic nutritional index, NLR: Neutrophil-to-lymphocyte ratio, BMI Body mass index, CPB Cardiopulmonary bypass, EuroSCORE, European system for cardiac operative risk evaluation II, ΔPNI/ΔNLR Percentage change at 24h postoperative [(post−pre)/pre×100], negative ΔPNI PNI declined, positive ΔNLR NLR increased

*P <0.05; **P <0.01; ***P <0.001; NS: not significant

The sampling protocol entailed collecting peripheral blood samples preoperatively, 24 h after surgery, and at the time of discharge. Investigations included a complete blood count, including differential leucocytic counts and estimation of serum albumin and creatinine levels.

The following evaluation tools were also collected:

1. Body mass index (BMI): was calculated as weight (kg) divided by height (m2), was graded after the guidelines of the National Institute of Health.

2. Neutrophil-to-lymphocyte ratio (NLR): was calculated as the ratio between the neutrophil and lymphocyte counts measured in peripheral blood [17].

3. Prognostic Nutritional Index (PNI): was calculated as the estimated serum albumin (g/L) plus the result of multiplying the total peripheral blood lymphocyte count (103/µL) by five [18].

4. ΔPNI: was calculated as [(postoperative PNI − preoperative PNI) / preoperative PNI] × 100 (i.e., ΔPNI = [(postPNI − prePNI) / prePNI] × 100, where a negative value indicates PNI declined after surgery and a more negative value represents greater nutritional deterioration. Similarly, ΔNLR = [(postoperative NLR − preoperative NLR) / preoperative NLR] × 100, where a positive value indicates NLR increased after surgery.

Study endpoints

Primary endpoint

Composite postoperative morbidity within 30 days after CABG.

Secondary endpoints

Intraoperative mortality, in-hospital mortality, 30-day mortality, individual postoperative complications, ICU length of stay, hospital length of stay, and readmission.

Predictor variables of interest

Preoperative PNI, preoperative NLR, ΔPNI, and ΔNLR.

Data management and statistical analysis

The data normality was assessed using the Kolmogorov-Smirnov test. The data are presented as mean±standard deviation (SD) for normally distributed continuous variables, median [interquartile range, IQR] for non-normally distributed variables, and numbers and percentages for categorical variables. Comparisons between patients were performed using t-test and Mann-Whitney U test for continuous variables when appropriate and chi-square test for categorical variables. Multivariate binary logistic regression analysis was subsequently performed to identify independent predictors of postoperative morbidity and mortality. Results are reported as odds ratios (OR) with 95% confidence intervals (CI) and P-values. The Receiver Operating Characteristic (ROC) curve analysis was used to assess the discriminative ability of PNI and NLR for morbidity and mortality outcomes, as judged by the area under the curve (AUC). For each outcome, the dependent variable in all analyses was a patient-level binary outcome: each patient was coded 1 if they experienced at least one complication or death within the specified period, and 0 if they experienced none. This patient-level binary coding was used for all univariate comparisons, logistic regression analyses, and ROC analyses. The total number of complication events may exceed the number of uniquely affected patients because some patients experienced more than one complication; both unique patient counts and total event counts are reported separately throughout. Given the limited number of mortality events (n = 11), all mortality-related analyses including multivariate models are presented as exploratory and hypothesis-generating only. Statistical analysis was conducted using IBM SPSS Statistics for Windows, Version 28.0 (IBM Corp., Armonk, NY, USA, 2021). A P-value of less than 0.05 was considered statistically significant.

Results

A total of 358 patient records were screened. Thirty-one were excluded based on the predefined exclusion criteria or because of incomplete baseline data, leaving 327 patients for final analysis. Postoperative biomarker measurements were available for 325 patients at 24 h and 319 patients at discharge; analyses at these time points were performed using the available measurements. The recruitment demographics and general medical data, as well as the preoperative data, are presented in (Table 1).

Table 1.

Preoperative data (n = 327)

Data			Findings (n = 327)
Demographics	Age (years)	< 50	37 (11.3%)
		50–59	126 (38.5%)
		60–69	147 (45%)
		70–79	17 (5.2%)
		Mean (± SD)	60.2 ± 7.35
	Gender: Male: Female		231: 96
	Body mass index (kg/m2)	< 25	42 (12.8%)
		25–30	122 (37.4%)
		30–35	144 (44%)
		> 35	19 (5.8%)
		Mean (± SD)	29.6 ± 3.7
Smoking	Non: Ex: Current		106:153:68
Preoperative medical data	Number of morbidities: 0:1:2:3:4		18:78:118: 93:20
	Type of morbidities	Insulin-dependent diabetes mellitus	28 (8.6%)
		Non-insulin-dependent diabetes mellitus	63 (19.3%)
		Hypertension	97 (29.7%)
		Joint disorders	110 (33.6%)
		Others	11 (3.4%)
	Previous non-cardiac surgeries		85 (26%)
	Previous cardiac interventions		97 (29.7%)
	NYHA functional grade: I:II: III: IV		47: 59:125:96
	Ejection fraction	< 40	63 (19.3%)
		40–49	91 (27.8%)
		50–59	139 (42.5%)
		60–69	34 (10.4%)
		Mean (SD)	49.2 ± 9.3
	ASA grade: I:II: III		78: 214:35
	EuroSCORE		1.07 [0.93–2.26] *
Preoperative lab data	Hemoglobin concentration (g/dl)		11.65 ± 0.73
	Serum albumin (g/l)		39.56 ± 4.34
	Serum creatinine (mg/dl)		0.82 ± 0.27
	Neutrophil count (1000/cc)		5.43 ± 0.68
	Lymphocyte count (1000/cc)		3.33 ± 0.35

The data were shown as means and standard deviations, numbers, percentages, ratios, and *median and interquartile range NYHA New york heart association, ASA American society of anesthesiologists, EuroSCORE the European System for Cardiac Operative Risk Evaluation II

All patients had elective CABG through sternotomy. The surgical procedure was performed with CPB for 258 patients (78.9%) and only 69 patients (21.1%) were operated off-pump. The number of patients who required single-vessel grafting was 27.5%, while 237 patients (72.5%) required grafting of two or three vessels. The mean amount of intraoperative blood loss was 710 ± 174.2 ml. The amount of intraoperative blood loss was positively correlated with the number of grafted vessels (r = 0.457, P < 0.001) and was greater in female patients. Blood or blood products transfusion was required in 38.5% of patients. The frequency of intraoperative complications was 1.83% and all were related to bleeding. Two patients died intraoperatively following severe bleeding; both had been receiving anticoagulant and antiplatelet therapies preoperatively.

During hospital stay, 305 patients had an uneventful recovery, while 22 patients developed postoperative complications. Bleeding necessitated re-exploration in the operative room occurred in one patient, who subsequently died. Two patients developed postoperative atrial fibrillation that was managed successfully. Another two patients developed deteriorated renal function and required renal replacement therapy. Infections occurred in three patients: two developed sternal wound infections, while one developed a respiratory infection that failed to respond to treatment, and the patient subsequently died. Five patients required prolonged mechanical ventilation; two could not be successfully weaned from mechanical ventilation and subsequently died due to concomitant respiratory and wound infections. The remaining nine patients experienced other postoperative complications that were managed conservatively during the hospital stay.

During the month after surgery, 30 patients developed postoperative morbidities for a total rate of 9.35%. Twenty patients (6.23%) were re-admitted and managed as hospital cases, while ten patients (3.12%) were managed as outpatient cases. Three cases were re-admitted for the development of myocardial ischemia and required redo CABG; two died. Two patients developed stroke and another two showed features consistent with acute kidney injury (AKI). One of those with AKI could not compensate and died. Seven patients developed sternal wound infection. Three patients with deep infections and three with superficial infections were readmitted. The seventh patient presented with a severe deep infection had poorly controlled diabetes mellitus and died. Three patients with respiratory infection were readmitted; two responded to treatment but one died. Two patients with sternotomy incision sepsis and one with leg incision sepsis were readmitted, received appropriate treatment and were discharged without evidence of infection. Ten patients with mild sepsis were managed as outpatient cases and responded to conservative treatment. The overall 1-month postoperative morbidity rate was 15.0% (49/327), which includes the 30 patients with specifically documented major complications and an additional 19 patients with milder adverse events not requiring hospital admission. The overall mortality rate was 3.36% (11/327) (Fig. 1; Table 2).

Fig. 1.

Patient outcome flow diagram following coronary artery bypass grafting surgery (n = 327)

Table 2.

Operative and postoperative data:

Data				Findings
Intraoperative data	Type of surgery	On-pump		258 (78.9%)
		Off-pump		69 (21.1%)
	Number of grafted vessels: One: Two: Three			90:182:55
	Aortic clamping time (min)			57.1 ± 10.1
	CPB time (min)			76.1 ± 12.5
	Total operative time (min)			188 ± 24.1
	Intraoperative events	Amount of blood loss (ml)		710 ± 174.2
		Frequency of transfusion of blood or blood products		126 (38.5%)
		Complications		6 (1.83%)
		Mortality		2 (0.62%)
Postoperative data	During hospital stay	Duration of ICU stay (h)		61 ± 20.8
		Amount of chest tube drainage (cc)		1045 ± 145.9
		Morbidities	No	305(93.3%)
			Bleeding	1 (0.31%)
			POAF	2 (0.62%)
			AKI	2 (0.62%)
			Infection	3 (0.92%)
			Prolonged ventilation	5 (1.5%)
		Mortality		4 (1.23%)
		Length of hospital stay (days)		9.6 ± 2
	A month after surgery	Morbidities required hospital re-admission	Myocardial infarction	3 (0.93%)
			Stroke	2 (0.62%)
			AKI	2 (0.62%)
			Respiratory infection	3 (0.93%)
			Deep incisional SSI — sternal site	3 (0.93%)
			Superficial incisional SSI — sternal site.	4 (1.25%)
			Sternal incisional SSI with septic features — sternotomy site	2 (0.62%)
			Incisional SSI — leg harvest site with septic features	1 (0.31%)
			Total re-admission rate	20 (6.23%)
		Morbidities didn’t require hospital re-admission	Superficial incisional SSI — sternal site (managed outpatient)	4 (1.25%)
			Superficial incisional SSI — sternotomy site (managed outpatient)	3 (0.93%)
			Superficial incisional SSI — leg harvest site	3 (0.93%)
		Morbidity rate		30 (9.35%)
		Mortality rate		5 (1.56%)
Overall outcome rate	Total overall morbidity rate			49 (15%)
	Total overall mortality rate			11 (3.36%)

ICU Intensive care unit, POAF Postoperative atrial fibrillation, AKI, Acute kidney injury, PNI Prognostic nutritional index, NLR Neutrophil-to-lymphocyte ratio, SSI  Surgical Site Infection

Preoperative PNI decreased significantly (P1 < 0.001) at 24-h postoperative and at discharge. At discharge, mean PNI increased significantly (P2 < 0.001) compared to 24-h postoperative values. The magnitude of ∆PNI at discharge was significantly smaller than that at 24-h postoperatively (P3 < 0.001). Conversely, NLR was markedly elevated at 24 h after surgery (P1 < 0.001) and at discharge (P1 = 0.0007) compared to preoperative values. At discharge, NLR was significantly lower (P2 = 0.035) than at 24-h postoperative. ∆NLR at discharge decreased significantly (P3 < 0.001) versus that at 24-h postoperative (Table 3).

Table 3.

Preoperative PNI and NLR values of the studied patients:

Data	PNI	NLR
Preoperative (n = 327)	56.10 ± 7.6	1.63 ± 0.62
24-h postoperative (n = 325)	45.62 ± 7.4	1.91 ± 0.68
P1 value	< 0.001	< 0.001
Percentage change at 24-h postoperative (%)	-18.95 ± 6.15	18.33 ± 8.6
At discharge (n = 319)	51.62 ± 8.13	1.8 ± 0.65
P1 value	< 0.001	< 0.001
P2 value	< 0.001	0.035
Percentage change at discharge (%)	-8.54 ± 7.6	11.14 ± 6.9
P3 value	< 0.001	< 0.001

Differences between paired measurements were compared using the paired t-test. P1 represents the comparison with preoperative values; P2 represents the comparison between 24-hour postoperative and discharge values; and P3 represents the comparison between the percentage changes at 24 h and at discharge. P < 0.05 was considered statistically significant

PNI, Prognostic nutritional index, NLR Neutrophil-to-Lymphocyte ratio

The study demonstrated several significant associations between preoperative, demographic, and operative variables and postoperative outcomes. Intraoperative morbidity was significantly associated with older age (P = 0.034), higher BMI (P = 0.010), lower preoperative PNI (P < 0.001), higher preoperative NLR (P < 0.001), shorter CPB time (P = 0.033), and longer operative time (P = 0.002). Intraoperative mortality was significantly associated with older age (P = 0.015), higher EuroSCORE II (P = 0.021), lower preoperative PNI (P = 0.026), higher preoperative NLR (P = 0.020) and longer ischemia time (P = 0.021).

Hospital postoperative morbidity was significantly associated with older age (P < 0.001), higher BMI (P = 0.021), lower preoperative PNI (P < 0.001), and higher preoperative NLR (P = 0.003). Hospital postoperative mortality was significantly associated with older age (P = 0.031), higher BMI (P = 0.002), higher EuroSCORE II (P = 0.036), lower preoperative PNI (P = 0.002), higher preoperative NLR (P < 0.001), and longer operative time (P = 0.010).

For 1-month postoperative morbidity, preoperative PNI was not significantly different between groups (P = 0.085), whereas female sex was significantly associated with morbidity (P = 0.015). For 1-month postoperative mortality, significant associations were observed with age (P = 0.023), BMI (P = 0.009), EuroSCORE II (P = 0.015), lower preoperative PNI (P = 0.001), longer ischemia time (P = 0.003), longer CPB time (P = 0.044), longer operative time (P = 0.008), greater decline in PNI at 24 h (P < 0.001), greater increase in NLR at 24 h (P = 0.002), and female sex (P = 0.028).

Overall, lower preoperative PNI showed consistent associations with several adverse perioperative outcomes, whereas the associations observed for preoperative NLR were less consistent across endpoints.

All analyses used patient-level binary outcomes: The following numbers of unique patients experienced each outcome: intraoperative morbidity (6 patients), intraoperative mortality (2 patients), hospital postoperative morbidity (22 patients), hospital postoperative mortality (4 patients), 1-month postoperative morbidity (49 patients), and 1-month postoperative mortality (11 patients). As some patients experienced more than one complication, total complication events may exceed the number of uniquely affected patients (Table 4).

Data presented as median [Q1–Q3] for continuous variables and percentage for categorical variables. P-values derived from Mann-Whitney U test (continuous) and chi-square test (categorical). PNI: Prognostic Nutritional Index; NLR: Neutrophil-to-Lymphocyte Ratio; BMI: Body mass index; CPB: Cardiopulmonary bypass; EuroSCORE: European System for Cardiac Operative Risk Evaluation II; ΔPNI/ΔNLR: percentage change at 24 h postoperative [(post − pre)/pre×100]; negative ΔPNI = PNI declined; positive ΔNLR = NLR increased. *P < 0.05; **P < 0.01; ***P < 0.001; NS: not significant.

The ROC curve analysis suggested a potential predictive ability of preoperative PNI and NLR for intraoperative morbidities. High specificity values were observed within this cohort, with AUC values indicating apparent discriminative performance (Table 5; Fig. 2a & b). However, these findings should be interpreted with caution given the relatively low number of mortality events, which may affect the stability and generalizability of these estimates.

Table 5.

ROC curve analysis of patients at increased risk of intraoperative and postoperative morbidity and/or mortality

	Complications				Mortalities
Time	Variable	AUC	P	95%CI	Variables	AUC	P	95%CI
IO	Pre. PNI	0.818	0.003	0.222, 1	Age	0.988†	0.012	0.962, 1
	Pre. NLR	0.850	< 0.001	0.687, 1	Ischemia time	0.985	0.015	0, 1
					Pre. PNI	0.966	0.028	0.954, 1
24-h PO	Age	0.728	0.0004	0.601, 0.855	Age	0.813	0.008	0.583, 1
	Female gender	0.562	0.055	0.454, 0.670	Female gender	0.731	0.065	0.485, 0.977
	BMI	0.647	0.033	0.511, 0.784	BMI	0.801	0.018	0.551, 1
	EuroSCORE II	0.584	0.073	0.441, 0.727	EuroSCORE II	0.949	< 0.001	0.879, 1
	Pre. PNI	0.907	0.026	0.540, 1	Pre. PNI	0.962	< 0.001	0, 1
	Pre. NLR	0.690	0.003	0.569, 0.812	Pre. NLR	0.984†	< 0.001	0.956, 1
	Operative time	0.585	0.069	0.448, 0.721	Operative time	0.867	0.012	0.842, 0.891
1-m PO	Age	0.633	0.001	0.552, 0.715	Age	0.701	0.015	0.537, 0.863
	Females	0.591	0.017	0.516, 0.666	EuroSCORE II	0.730	0.008	0.558, 0.901
	BMI	0.633	0.001	0.552, 0.715	Pre. PNI	0.728	0.011	0.401, 1
	Pre. PNI	0.687*	0.019*	0.530, 0.843*	Ischemia time	0.785	< 0.001	0.658, 0.874
	Pre. NLR	0.557	0.047	0.458, 0.655	Operative time	0.766	0.002	0.578, 0.848
	Percentage change in PNI (%)	0.633	0.003	0.543, 0.723	Percentage change in PNI (%)	0.868	< 0.001	0.745, 0.992
	Percentage change in NLR (%)	0.642	0.001	0.556, 0.727
	Operative time	0.596	0.015	0.518, 0.674

IO Intraoperative, PO Postoperative, Pre Preoperative, AUC Area under curve, CI Confidence interval, SE, PNI Prognostic nutritional index, NLR Neutrophil-to-Lymphocyte ratio, BMI Body mass index, EuroSCORE The european system for cardiac operative risk evaluation II. )

†AUC values ≥ 0.980 are likely artificially inflated due to the very small event count and should not be used for clinical decision-making without external validation in larger cohorts

Fig. 2.

a ROC curve for preoperative predictors of IO morbidity. b ROC curve for preoperative predictors of IO mortality. c ROC curve for the predictors of 24-h complication rate. d ROC curve for the predictors of 24-h mortality. e ROC curve for predictors of one-month complication rate. f ROC curve for predictors of one-month mortality rate

Regarding intraoperative mortality, no significant differences were observed between the AUCs of age, preoperative PNI, and ischemia time. These variables demonstrated high specificity within this dataset; however, the observed performance metrics should be interpreted cautiously due to potential overestimation related to the limited number of events.

ROC curve analysis suggested that age, BMI, preoperative NLR, and PNI may be associated with the risk of developing complications during the first 24 h after surgery. Similarly, lower preoperative PNI, higher preoperative NLR, and elevated EuroSCORE II were associated with an increased risk of mortality within the first 24 h postoperatively.

Older age, higher BMI, and longer operative time also showed potential associations with early postoperative mortality (Table 5; Fig. 2c & d). However, these findings should be interpreted with caution given the limited number of mortality events, which may affect the robustness of these associations.

The ROC curve analysis suggested that older age, higher BMI, and greater percentage changes in NLR (∆NLR) and PNI (∆PNI) may be associated with an increased risk of complications within one month after surgery. Female gender, higher preoperative NLR, and lower PNI also showed potential associations, although with weaker statistical significance. Additionally, greater ∆PNI, longer ischemia and operative times, and higher preoperative EuroSCORE II were associated with an increased risk of mortality within one month postoperatively. Preoperative PNI and older age also demonstrated possible associations with mortality, though with lower statistical strength (Table 5; Fig. 2e & f).

However, these findings should be interpreted with caution due to the limited number of mortality events, which may affect the stability and generalizability of the observed associations.

Multivariate binary logistic regression confirmed preoperative PNI (OR = 0.818, 95%CI: 0.723–0.925, P = 0.003) as an independent predictor of 1-month postoperative mortality. ΔPNI at 24 h (OR = 1.132, 95%CI: 1.057–1.212, P < 0.001) was independently associated with 1-month postoperative mortality after adjustment for age, EuroSCORE II, preoperative NLR, and ischemia time. For 1-month morbidity, female gender was the only significant independent predictor (OR = 2.434, 95%CI: 1.222–4.848, P = 0.023); preoperative PNI and NLR did not retain independent significance after adjustment.

ROC curve analysis demonstrated modest discriminative performance for 1-month morbidity. Among the evaluated biomarkers, ΔPNI at 24 h showed the highest AUC (AUC = 0.691, 95% CI: 0.6–0.78, P < 0.001), followed by preoperative PNI (AUC = 0.687, 95% CI: 0.53–0.843, P = 0.019) and preoperative NLR (AUC = 0.634, 95% CI: 0.53–0.738, P = 0.011). ΔNLR at 24 h did not demonstrate significant discrimination for morbidity (P = 0.128). For 1-month mortality, ΔPNI at 24 h showed an AUC of 0.707 (95% CI: 0.504–0.911, P = 0.045), whereas ΔNLR at 24 h was not statistically significant (P = 0.301) (Table 6; Fig. 3a & b).

Table 6.

Multivariate Binary Logistic Regression And ROC Curve Analyses for 1-Month Postoperative Morbidity and Mortality

Binary Logistic Regression		OR	95% CI		p
1-Month Postoperative Mortality
Preoperative PNI		0.818	0.723 – 0.925		0.003
ΔPNI at 24 h		1.132	1.057 – 1.212		<0.001
1-Month Postoperative Morbidity
Female gender		2.434	1.222 – 4.848		0.023

ROC Curve Analysis		AUC (SE)	P	95%CI	Sn.	Sp.	PPV	NPV	Accuracy
Total morbidity	Pre. PNI	0.687 (0.08)	0.019	0.53, 0.843	40	77.2	28.9	84.7	71
	Pre. NLR	0.634 (0.06)	0.011	0.53, 0.738	29.1	77.6	23.1	82.5	69.4
	∆PNI (%) 24-h	0.691 (0.05)	<0.001	0.6, 0.78	87.3	79.4	49.6	96.4	80.7
	∆NLR (%) 24-h	0.581 (0.05)	0.128	0.477, 0.686	29.1	75	21.2	82	67.3
Total mortality	∆PNI (%) 24-h	0.707 (0.1)	0.045	0.504, 0.911	69.2	76.8	40.8	91.5	76.5
	∆NLR (%) 24-h	0.614 (0.11)	0.301	0.389, 0.83	61.5	75.8	8.8	98.1	75.2

Pre Preoperative, AUC Area under curve, SE Standard error, Sn Sensitivity; Sp Specificity, PPV Positive predictive value, NPV Negative predictive value, PNI Prognostic nutritional index, NLR Neutrophil-to-Lymphocyte ratio, CI Confidence interval

Fig. 3.

a ROC curve for predictors of the total complication rate until the month after surgery. b ROC curve for predictors of the overall mortality rate until the month after surgery

Discussion

The present study demonstrated that lower preoperative PNI was significantly associated with increased perioperative morbidity and early postoperative mortality after CABG. Moreover, preoperative PNI remained an independent predictor of 1-month mortality after adjustment for established clinical variables. These findings are consistent with previous studies demonstrating that reduced PNI is associated with Major Adverse Cardiovascular Events (MACEs) and mortality in patients with coronary artery disease (CAD), particularly those undergoing coronary revascularization, whereas higher PNI scores are associated with better clinical outcomes [19–21]. Similarly, Ling et al. [22] reported that lower PNI was associated with poorer outcomes among patients with acute myocardial infarction undergoing Percutaneous Coronary Intervention (PCI). Demirci et al. [23] further demonstrated that reduced preoperative PNI independently predicted long-term mortality after emergency CABG in patients with ST-elevation myocardial infarction and identified a PNI cutoff value of 44.9 for mortality prediction. Likewise, Peng et al. [24] found that patients with acute myocardial infarction and low PNI had a 2.55-fold higher risk of major adverse cardiovascular events following coronary intervention, while Sun et al. [25] reported significantly better long-term survival among CABG patients with normal preoperative PNI. Collectively, these findings support the concept that impaired nutritional status adversely influences postoperative recovery and highlight 8 the potential value of routine preoperative nutritional assessment in CABG candidates.

Beyond its association with adverse cardiovascular outcomes, lower preoperative PNI in the present study was also significantly associated with important baseline comorbidities, particularly diabetes mellitus. This observation is consistent with the findings of Ling et al. [22], who reported that poor nutritional status frequently coexisted with metabolic disorders among patients with cardiovascular disease. Since diabetes is a well-recognized determinant of postoperative complications following CABG, the interaction between impaired nutritional status and metabolic dysfunction may partially explain the poorer outcomes observed in patients with lower PNI. These findings further support the hypothesis that PNI reflects overall physiological reserve rather than nutritional status alone.

The prognostic utility of inflammatory and nutritional biomarkers observed in the present study is further supported by previous investigations evaluating similar indices in CABG patients. Serhatlioglu et al. [26] reported significantly higher serum C-reactive protein, NLR, and neutrophil-to-albumin ratio among patients who developed new-onset atrial fibrillation after CABG compared with those who remained free of this complication, emphasizing the contribution of perioperative inflammation to adverse postoperative outcomes. Likewise, Yücel et al. [28] evaluated the CALLY Index, a composite score incorporating lymphocyte count, serum albumin, and C-reactive protein, and demonstrated its usefulness as a simple and practical tool for preoperative risk assessment, allowing identification of high-risk patients and potentially improving surgical outcomes. These findings further support the growing role of composite nutritional and inflammatory biomarkers in perioperative risk stratification after CABG.

Further supporting the potential value of composite nutritional and inflammatory indices, Erkan et al. [29] evaluated the Hemoglobin–Albumin–Lymphocyte–Platelet (HALP) score in patients undergoing elective CABG and reported that lower postoperative HALP scores were associated with postoperative adverse events. Their findings are consistent with the present study, in which lower preoperative PNI was associated with several adverse perioperative outcomes and greater postoperative changes in PNI were associated with 1-month mortality. Since both HALP and PNI incorporate serum albumin and lymphocyte count, these observations suggest that albumin-containing composite indices may reflect clinically relevant nutritional and inflammatory alterations in CABG patients. However, HALP additionally incorporates hemoglobin and platelet count and therefore represents a broader hematological and inflammatory profile than PNI.

The present study also demonstrated that several demographic, clinical, and operative variables, including advanced age, female sex, higher body mass index, reduced left ventricular ejection fraction, prolonged aortic cross-clamp time, and longer operative duration, were significantly associated with adverse perioperative outcomes. Similar observations have been reported previously. Demirci et al. [23] identified age, reduced ejection fraction, impaired renal function, higher Killip class, and graft selection LIMA-LAD as important determinants of long-term mortality following emergency CABG. Likewise, Das et al. [30] reported significantly higher postoperative complications and mortality among patients aged 65 years or older and identified advanced age as an independent mortality predictor. Wang et al. [31] demonstrated that reduced modified body mass index was associated with increased postoperative complications and in-hospital mortality, particularly among women and patients with multiple comorbidities. Furthermore, Liu et al. [32] developed a preoperative prediction model incorporating advanced age, hypoalbuminemia, and elevated serum creatinine for early mortality after surgical revascularization, while Duman et al. [33] reported that the visceral adiposity index could predict postoperative respiratory complications after CABG. Together, these findings emphasize that nutritional and inflammatory biomarkers should be interpreted alongside established clinical risk factors to achieve more comprehensive perioperative risk assessment.

NLR also demonstrated important prognostic value in the present study. Although elevated preoperative NLR was significantly associated with adverse perioperative outcomes in univariate analysis, it did not retain independent significance after multivariable adjustment, suggesting that its predictive value may partly reflect its association with other established clinical risk factors. Nevertheless, this finding is consistent with previous studies reporting that elevated NLR is associated with postoperative atrial fibrillation, acute kidney injury, major adverse cardiovascular events, and mortality after CABG. Engin et al. 27 identified lymphocyte count together with visceral adiposity index as independent predictors of postoperative atrial fibrillation. Raeisi et al. [34] reported significantly higher NLR among patients who developed postoperative acute kidney injury, while Garganeeva et al. [35] demonstrated significant associations between NLR and mortality in patients with chronic heart failure undergoing CABG. Similarly, Atasoy and Guven [36] identified NLR as an independent predictor of Post-Operative Atrial Fibrillation (POAF) after off-pump CABG, whereas Ceylan et al. [37] reported that elevated NLR independently predicted major adverse cardiovascular events. Moreover, a recent systematic review confirmed that both preoperative and postoperative NLR were significantly associated with POAF [38], and Söylemez et al. [39] demonstrated that inflammatory biomarkers remained important predictors of mortality following CABG. Taken together, these findings support the role of systemic inflammation in determining postoperative outcomes while suggesting that NLR may be most informative when interpreted in conjunction with other clinical variables.

An important observation of the current study was that the magnitude of early postoperative PNI decline (ΔPNI) was associated with adverse short-term postoperative outcomes. Patients experiencing greater postoperative PNI decline had poorer short-term outcomes, suggesting that dynamic perioperative changes in nutritional status may have potential prognostic relevance. This observation agrees with Hayasaka etal. [40], who demonstrated that the postoperative-to-preoperative PNI ratio was a more sensitive predictor of mortality in lung cancer surgery than postoperative PNI alone. Similarly, Tatsuta et al.[41] identified perioperative PNI changes as independent predictors of overall and recurrence-free survival in colorectal cancer, while Miura et al.[42] reported that decreased postoperative PNI independently predicted poor survival following head and neck cancer surgery. Comparable findings have also been reported after cardiac surgery, where Li et al. [43] showed that postoperative nutritional deterioration following aortic valve surgery was associated with unfavorable clinical outcomes, and Özmen et al. [44] demonstrated significant correlations between preoperative PNI and operative complexity, including cardiopulmonary bypass duration, aortic cross-clamp time, and Intensive Care Unit (ICU) stay. More recently, Toprak and Bilgiç [45] confirmed that both preoperative and postoperative PNI values independently predicted mortality after CABG, emphasizing that serial perioperative assessment of nutritional status may improve risk stratification and facilitate timely nutritional and perioperative interventions.

Toprak et al.[46] evaluated the prognostic significance of PNI in patients undergoing CABG and demonstrated an association between perioperative PNI and in-hospital mortality. Their findings are consistent with the present study, in which lower preoperative PNI was associated with adverse postoperative outcomes and greater ΔPNI at 24 h was independently associated with 1-month mortality. In another recent CABG study,Toprak et al.[47] investigated time-dependent changes in NLR and other systemic inflammatory indices during cardiopulmonary bypass and reported significant associations between perioperative inflammatory dynamics and in-hospital mortality. Similarly, in our study, higher preoperative NLR and perioperative changes in NLR were associated with adverse outcomes, although preoperative NLR did not retain independent significance after multivariable adjustment.

Study limitations

While this study offers important insights, several limitations should be considered. The retrospective design introduces potential selection bias, reliance on routinely collected and potentially non-standardized clinical records, and possible residual confounding. The exclusion of patients with incomplete records may have introduced selection bias, as these patients could differ systematically from those included in the final analysis. This may limit the generalizability of the findings. Furthermore, as a single-center retrospective study, the study population may not fully represent the broader population of patients undergoing CABG. The relatively short follow-up period restricts the assessment of long-term outcomes. Additionally, the combination of on-pump (78.9%) and off-pump (21.1%) CABG patients represents a potential confounding factor, as cardiopulmonary bypass directly affects inflammatory markers, albumin concentration, and postoperative biomarker changes through hemodilution, fluid shifts, and the systemic inflammatory response. A sensitivity analysis stratified by surgical technique was not performed, which may limit the interpretation of dynamic biomarker changes. Future studies should stratify analyses by CPB exposure. Furthermore, postoperative changes in PNI and NLR may partly reflect hemodilution, transfusion (38.5% of patients received blood products), and the acute stress response rather than nutritional deterioration exclusively. The temporal relationship between biomarker measurements and outcomes also warrants consideration: postoperative ΔPNI and ΔNLR at 24 h cannot precede intraoperative events and are therefore presented as early postoperative prognostic associations with subsequent outcomes only. Discharge biomarker values are additionally subject to survivor bias, as only patients who survived to discharge contributed these measurements. In addition, the use of composite morbidity without detailed stratification (e.g., by severity) may limit clinical interpretation. Critically, with only 11 mortality events, all mortality-related analyses — including univariate comparisons, multivariate models, ROC analyses, and reported AUC values — are inherently unstable and should be regarded as hypothesis-generating only. Some AUC values approaching 1.0 (e.g., 0.988 for age and 0.984 for preoperative NLR for mortality) may be artificially inflated due to the small event count and represent overfitted estimates that are unlikely to replicate externally. AUC values derived from such small event counts are known to be susceptible to artificial inflation and may not be replicated in external cohorts; accordingly, the mortality AUC values reported in Tables 5 and 6 should be interpreted with particular caution. Finally, there is a risk of over- or underestimation of the associations between PNI and NLR and postoperative outcomes. These findings should be considered preliminary evidence pending validation in larger, prospective multicenter studies.

Conclusion and recommendations

PNI is a simple, inexpensive, and biologically plausible score derived from routine blood tests that may provide prognostic information regarding short-term outcomes after CABG. Lower preoperative PNI was associated with adverse perioperative outcomes after CABG, while preoperative PNI and early postoperative PNI change were independently associated with 1-month mortality. Elevated preoperative NLR was associated with adverse outcomes in univariate analysis but did not retain independent significance after multivariable adjustment. These findings suggest that perioperative nutritional and inflammatory markers may provide additional information regarding short-term outcomes after CABG. However, given the retrospective single-center design and the limited number of mortality events, the findings should be considered exploratory and require validation in larger prospective multicenter studies.

Abbreviations

CABG

Coronary artery bypass grafting

PNI

Prognostic Nutritional Index

NLR

Neutrophil-to-Lymphocyte Ratio

ROC

Receiver Operating Characteristic

AUC

Area Under the Curve

MACE

Major Adverse Cardiovascular Events

EF

Ejection Fraction

ICU

Intensive Care Unit

BMI

Body Mass Index

Authors' contributions

• Mohammed Elgazzar: Conceptualization, Validation, Data curation & Supervision.• Basma Hani; Methodology, Validation, Formal analysis & Writing—review and editing.• Abdelhamed Ali; Conceptualization, Soft-ware & Visualization.• Mahmoud Elshafie; Soft-ware, Investigation & Project administration.• Ashraf Wahdan; Validation, Resources, Writing—original draft preparation & Funding acquisition.• All authors have read and agreed to the published version of the manuscript.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Data availability

The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethical approval and consent to participate

The study protocol was submitted to the Research Ethics Committee (REC), Faculty of Medicine, Benha University, Egypt, on 6 November 2025, when the protocol reference number RC 6-11-2025 was assigned. Final ethical approval was granted in March 2026. CABG procedures were conducted in accordance with the Declaration of Helsinki. Given the retrospective nature of the study and the use of de-identified medical record data, the ethics committee waived the requirement for informed consent. The privacy of all research participants was safeguarded, and a high level of confidentiality was maintained.

Consent for publication

Not applicable. This study did not include any individual patient data, images, videos, or other identifiable personal information requiring consent for publication.

Competing interests

The authors declare no competing interests.