ABSTRACT
Objective
To evaluate whether the fibrinogen-to-albumin ratio (FAR), measured mainly in the third trimester, is associated with and can predict preeclampsia (PE) and small-for-gestational-age (SGA) neonates in pregnant women.
Methods
In this retrospective study involving 30 patients (PE group) and 330 controls, the FAR was calculated, and its role in predicting PE and SGA was investigated. Additionally, gestational age at delivery, birth weight, percentile values, hemogram parameters, and biochemical variables were compared between the groups.
Results
FAR did not demonstrate sufficient diagnostic power to predict PE [receiver operating characteristic, area under the curve (AUC)= 0.527]; however, decreased serum albumin level was significantly associated with increased risk of PE [odds ratio (OR)= 0.768]. Albumin was identified as an independent predictor of PE (multivariate OR= 0.768, 95% confidence interval: 0.662-0.891, p < 0.001). Furthermore, no significant association was found between FAR and birth weight, SGA, or birth weight percentile values.
Conclusion
FAR alone does not possess sufficient diagnostic power (AUC= 0.527). Albumin emerged as a consistent predictor across all analyses (univariate OR= 0.768; multivariate OR= 0.768). Urea and alanine aminotransferase were also identified as independent predictors in the multivariate model. No significant association was detected between FAR and birth weight or SGA.
INTRODUCTION
Preeclampsia (PE), affecting approximately 5-10% of pregnant women, is a pregnancy complication characterized by hypertension and proteinuria occurring after the 20th week of gestation.1, 2 PE, a multifactorial disorder characterized by hypertension and organ dysfunction, is associated with maternal and neonatal mortality and morbidity worldwide.2
An increasing incidence of PE has been observed in the United States in recent years. This trend has been attributed to a higher prevalence of known risk factors such as obesity and advanced maternal age.3
Although the etiology of PE has not been fully elucidated, clinical and pathological studies indicate that systemic inflammatory response, immune dysfunction, abnormalities in the coagulation and fibrinolytic systems, and aberrant placentation play central roles in the pathogenesis of the disease.2 Moreover, a more pronounced hypercoagulable state has been reported in preeclamptic pregnancies compared with normotensive pregnancies, and alterations in the coagulation system are considered to contribute to the disease process.4
Early prediction and recognition of PE are among the primary objectives of obstetric care, as appropriate clinical surveillance and timely intervention can prevent serious maternal and perinatal complications. To this end, various biomarkers and imaging parameters, including soluble fms-like tyrosine kinase-1, placental growth factor, and uterine artery Doppler indices, have been extensively investigated for their potential in predicting PE.2
In recent years, a growing number of studies have evaluated the clinical utility of the fibrinogen-to-albumin ratio (FAR), derived from routine laboratory parameters and reflecting systemic inflammation and coagulation activity, in predicting the presence and severity of PE. However, the diagnostic and prognostic value of FAR for PE has not yet been conclusively established.
Fibrinogen (Factor I) is one of the major proteins in plasma, present in high concentrations, and plays a central role in the coagulation cascade. As the principal structural component of clot formation, fibrinogen is an indispensable element of the coagulation process; effective clot formation cannot occur in the absence of fibrinogen, even when other coagulation factors are present at normal levels.5 Synthesized in hepatic parenchymal cells, fibrinogen is a hexameric glycoprotein with a molecular weight of approximately 340 kDa. Its normal plasma concentration ranges from 200 to 400 mg/dL, with a half-life of approximately 3 to 5 days.6
Albumin, another important plasma protein synthesized in the liver, possesses antioxidant, anti-inflammatory, and tissue-repair-supporting properties. Albumin, the most abundant protein in plasma, tends to decrease in conditions such as chronic inflammation, malnutrition, and liver disease.7-9 During pregnancy, serum albumin levels decline due to physiological plasma volume expansion and hemodilution; mild hypoalbuminemia is frequently regarded as a normal finding.9
The FAR, calculated from serum fibrinogen and albumin levels, has attracted attention as a novel composite biomarker that simultaneously reflects inflammatory and coagulation processes. FAR has been shown to be useful in predicting clinical outcomes in various cardiovascular diseases and malignancies.10, 11 In obstetrics, its prognostic value has been investigated in threatened miscarriage, recurrent pregnancy loss, and several obstetric complications, and promising results have been reported.12
During inflammatory processes, FAR increases as a consequence of rising fibrinogen levels and declining albumin levels. Consequently, FAR is regarded as a composite indicator of the degree of systemic inflammation. Recent studies have suggested that FAR may be an important marker for evaluating numerous inflammation-associated conditions and predicting clinical outcomes.10, 11
MATERIALS AND METHODS
This retrospective cohort study was conducted in the Department of Obstetrics and Gynecology at Erzincan Binali Yıldırım University Mengücek Gazi Training and Research Hospital. Pregnant women with a gestational age ≥20 weeks who were followed at our clinic between January 2023 and June 2024 and who had serum fibrinogen and albumin levels measured were included in the study.
Ethical approval was obtained from the Erzincan Binali Yıldırım University Non-Interventional Clinical Research Ethics Committee (decision no.: 485828, date: 29.09.2025). All procedures were conducted in accordance with the principles of the Declaration of Helsinki. The study was retrospective; therefore, informed consent was not required, as only anonymized data from routine clinical records and patient files were analyzed and no intervention was performed on the patients.
The study population comprised 360 pregnant women, classified into the PE group (n=30) and the control group (n=330). PE was diagnosed based on the clinical assessment recorded at admission, according to the diagnostic criteria of the Ministry of Health of the Republic of Türkiye, including admission blood pressure (≥140/90 mmHg). Because individual blood pressure values could not be retrieved retrospectively from the electronic records, severity subclassification could not be performed.
Exclusion criteria included platelet (PLT) count <100,000/mm3 for other than PE (either pre-existing or occurring during pregnancy), nephropathy, elevated liver enzymes attributable to causes other than PE, active bleeding, and intrahepatic cholestasis of pregnancy.
The study included pregnant women aged 18 years or older who were hospitalized for the following indications: decreased fetal movement, threatened preterm labor, trauma, hypertension, membrane rupture, uterine contractions, or other obstetric reasons. Demographic characteristics, gestational age at delivery, mode of delivery, birth weight, and neonatal percentile values were obtained from patients’ files.
Advanced maternal age was defined as ≥35 years. Small for gestational age (SGA) was defined as a birth weight below the 10th percentile for gestational age. The FAR was calculated by dividing serum fibrinogen (mg/dL) by serum albumin (g/L) and reported as a unitless ratio. Levels of fibrinogen, albumin, hemoglobin, PLT, alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatinine, urea, and calcium were obtained from routine laboratory assessments performed at the time of hospitalization or approximately 3-4 weeks before delivery.
The primary analysis compared FAR and other laboratory parameters between the PE and control groups. The secondary analysis evaluated the association between FAR and birth weight, gestational age at delivery, and neonatal percentile values across the entire study population.
Statistical Analysis
The normality of the data was assessed using the Shapiro-Wilk test. Normally distributed continuous variables are presented as mean ± standard deviation and compared between groups using the independent samples t-test. Non-normally distributed continuous variables were expressed as medians [interquartile range (Q1-Q3)] and analyzed using the Mann-Whitney U test. Categorical variables were presented as frequencies and percentages and were compared using Fisher’s Exact Test.
The diagnostic performance of FAR for predicting PE was evaluated using receiver operating characteristic (ROC) curve analysis. The 95% confidence intervals (CIs) for the area under the curve (AUC) were calculated using the bootstrap method with 2,000 iterations. The optimal cut-off point was determined based on the Youden Index. The DeLong test was used for comparisons between ROC curves. Associations between variables were assessed using Spearman’s correlation analysis. Multivariate logistic regression analysis was performed to identify independent predictors of PE. In the multivariate model, FAR was included a priori as the marker of interest, and albumin, urea, and ALT were included based on their univariate association with PE. AST was not entered because of its strong correlation with ALT (Spearman r= 0.609) to avoid collinearity. Model fit was assessed using the Nagelkerke R-squared and the Hosmer-Lemeshow test. A two-sided P value of <0.05 was considered statistically significant. ROC analyses, bootstrap CIs (2,000 iterations), DeLong comparisons were performed using Python; figures were generated using Python; the remaining analyses were performed using IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp., Armonk, NY, USA).
RESULTS
Comparison of Clinical and Laboratory Characteristics
The clinical, demographic, and laboratory characteristics of all 360 pregnant women included in the study are presented in Table 1. The group-specific distributions of key laboratory parameters are illustrated in Figure 1 using box plots.
Serum albumin levels were significantly lower in the PE group compared with the control group (31.67 ± 4.19 g/L vs. 33.66 ± 2.27 g/L; P < 0.001). This difference is also reflected in Figure 1, where box plots demonstrate a marked separation between the two groups.
FAR did not differ significantly between the two groups (control: 10.98 ± 3.01; PE: 11.51 ± 3.54; P = 0.365; Table 1). The expected increase in fibrinogen was limited in this sample.
Urea was significantly elevated in the PE group [15.00 (12.00-18.00) vs. 17.50 (14.25-21.75) mg/dL; P = 0.008; Table 1), consistent with other studies in the literature.13 Gestational age at delivery was significantly shorter in the PE group [38.00 (37.00-39.00) vs. 37.00 (36.50-38.00) weeks; P = 0.008). ALT, AST, PLT count, creatinine, calcium, and birth weight did not differ significantly between the groups (all p > 0.05; Table 1).
Correlation Analysis Between FAR and Clinical Parameters
Spearman correlation results between FAR and eight clinical parameters are presented in Table 2. A statistically significant negative correlation was identified between FAR and AST (r = -0.170; P = 0.001). No significant correlations were found with other parameters (PLT, urea, ALT, calcium, creatinine, birth weight, or percentile).
ROC Analysis of FAR and Combined Models
The standalone diagnostic accuracy of FAR and that of the combined models are presented in Table 3, and the corresponding ROC curves are shown in Figure 2.
The AUC value for FAR alone was 0.527 (95% CI: 0.420-0.633). That the CI encompasses the 0.500 reference value indicates that FAR does not significantly exceed the chance level (Table 3). The optimal threshold determined by the Youden index was 12.60, yielding a sensitivity of 43.3%, a specificity of 69.1%, and an negative predictive value of 93.1%.
The ROC performance of models combining FAR with six parameters is also presented in Table 3. None of the FAR-combined models provided a statistically significant improvement over FAR alone.
Univariate and Multivariate Logistic Regression
Univariate and multivariate logistic regression results are presented in Table 4, and comparative forest plots are presented in Figure 3. In univariate analysis, albumin was the strongest predictor [odds ratio (OR)= 0.768; 95% CI: 0.670-0.880; P < 0.001; Table 4]. An OR below 1 indicates that each 1 g/L increase in albumin level is associated with a 23.2% reduction in the odds of PE. Urea (OR= 1.107; P = 0.004), ALT (OR= 1.030; P = 0.005), and AST (OR= 1.019; P = 0.008) also demonstrated significant associations. FAR did not yield a significant result (OR= 1.058; P = 0.364). In the multivariate model, albumin (OR= 0.768; P < 0.001), urea (OR= 1.086; P = 0.038), and ALT (OR= 1.027; P = 0.035) were identified as independent predictors (Table 4). FAR lost its independent predictive significance (OR= 1.017; P = 0.790). The model showed an acceptable fit (Nagelkerke R-squared =0.178; Hosmer-Lemeshow P = 0.085).
Association of FAR with Birth Weight and SGA
No significant correlation was found between FAR and birth weight among the 326 patients with available birth weight data (Spearman r = 0.046; P = 0.408; n=326). FAR did not differ significantly between 32 SGA and 296 non-SGA infants (median: 11.95 vs. 11.04; P = 0.118). The ROC AUC for FAR in predicting SGA was 0.584, which is close to chance level. The correlation between FAR and birth weight percentile was also non-significant (r = -0.029; P = 0.603). These results are summarized in Table 5.
Association Between Maternal Age, FAR and SGA
No statistically significant correlation was identified between age and FAR across all patients (Spearman r = -0.007; P = 0.894; n=360; Table 6). This finding is supported by the scatter plot, in which the nearly horizontal regression line indicates that age has no deterministic effect on FAR.
No significant correlation was found in either the control group (r = -0.023; P = 0.683) or the PE group (r = 0.102; P = 0.592) when the groups were analyzed separately (Table 6). When the advanced maternal age (≥35 years, n=63 in the whole sample) and normal-age (<35 years, n=297) groups were compared, no significant difference in FAR values was observed [advanced-age median: 10.84 (9.64-12.83) vs. normal-age median: 11.05 (8.76-13.05); P = 0.492], indicating that FAR is independent of maternal age.
Regarding SGA, no statistically significant difference was found in maternal age when comparing mothers of 32 SGA infants and 296 non-SGA infants (26.0 vs. 28.0 years; P = 0.149; Table 6). Age was not an independent predictor of SGA (OR= 0.954; 95% CI: 0.891-1.023; P = 0.185). ROC analysis showed that age had no clinical utility in predicting SGA (AUC= 0.423; an AUC below 0.5 indicates an inverse association). No significant correlation was found between age and birth weight percentile (r = 0.041; P = 0.457), and SGA prevalence did not differ between the advanced age and normal age groups (3.6% vs. 11.0%; Fisher P = 0.134).
DISCUSSION
In this study, the diagnostic value of FAR for predicting PE was examined using rigorous statistical methods, and the effect of maternal age on FAR and SGA was evaluated in additional analyses.
FAR alone did not demonstrate sufficient diagnostic power (AUC= 0.527; Table 3). The limited performance of FAR is primarily attributable to the absence of a significant difference in fibrinogen levels between the two groups.
Albumin emerged as a consistent predictor across all analyses (univariate OR= 0.768, multivariate OR= 0.768; Table 4). Urea and ALT were also identified as independent predictors in the multivariate model. No significant association was found between FAR and birth weight or SGA (Table 2, Figure 3).
In a retrospective cohort study encompassing 3,249 pregnancies, patients were classified into control, mild, and severe PE groups. The investigators reported that FAR was independently associated with PE and could be a useful biomarker for clinical risk stratification. FAR could also contribute to the planning of delivery management and perinatal care, although the need for larger, prospective, multicenter studies to validate these findings was emphasized.3 One study reported that FAR demonstrated superior diagnostic performance for predicting PE severity compared with other routine laboratory indicators. Because fibrinogen and albumin levels are readily obtained as part of routine biochemical workup, FAR may serve as a practical and cost-effective biomarker in clinical practice.1 Although FAR levels were slightly higher in the PE group in our study, the difference was not statistically significant (P = 0.365), and discriminatory power in ROC analysis limited. These discrepancies may be attributable to differences in patient populations, sample sizes, and study designs.
Studies evaluating FAR for the early prediction of PE, its disease severity, and its association with adverse perinatal outcomes have reported FAR as a promising biomarker. Given its low cost, ease of access, and strong predictive potential, it has been proposed that FAR could be integrated into first-trimester screening protocols alongside existing biomarkers.2
In contrast, our study assessed FAR during the second and third trimesters and investigated the association between FAR measured at more advanced gestational weeks and PE. This represents a different timing compared with previous studies, which predominantly focus on early pregnancy. While our findings suggest that FAR may be associated with PE even at advanced gestational ages, they demonstrate that it does not exhibit strong standalone diagnostic performance. This discrepancy may be explained by temporal variability in FAR throughout pregnancy and by variation in the pathophysiology of PE across trimesters.
In a study evaluating the diagnostic performance of FAR for predicting PE, which comprised 131 PE cases and 131 control cases, the combination of second-trimester FAR and PLT distribution width (PDW) demonstrated better diagnostic performance than FAR alone. Nevertheless, the investigators emphasized the need for prospective studies to validate these findings.14 In our study, PDW was not included in the analysis and only FAR was evaluated. The possibility that adding PDW to the model might improve diagnostic performance is consistent with the literature, but this could not be confirmed with our current data and remains an area for investigation in future multivariate prospective studies.
Serum albumin level in cases of PE has been reported to reflect disease severity and to serve as a useful biomarker for determining the timing of delivery.15 Studies examining the association between serum albumin levels and PE have reported that hypoalbuminemia is associated with increased vascular permeability, systemic inflammation, and endothelial dysfunction; consequently, low albumin levels may correlate with the presence and severity of PE. It has been further emphasized that albumin, a negative acute-phase reactant, decreases during the inflammatory response and may have clinical implications, particularly in hypertensive disorders of pregnancy.16, 17 In our study, serum albumin levels were significantly lower in the PE group compared with the control group, and albumin was identified as the strongest independent predictor of PE in multivariate analysis (OR= 0.768; P < 0.001). These results are consistent with findings reported in the literature and support the notion that albumin is not merely a nutritional indicator but also a potent biomarker reflecting the systemic inflammatory burden and endothelial damage in PE.
In a study evaluating the role of first-trimester Hemoglobin, Albumin, Lymphocyte, PLT (HALP) score and FAR in predicting PE, FAR levels were significantly higher in both severe and non-severe PE cases than in healthy pregnant women.17 In our study, FAR values were also numerically higher in the PE group than in the control group, although the difference was not statistically significant. However, FAR alone did not demonstrate sufficient discriminatory power in ROC analysis, suggesting that while FAR may be associated with PE, it may not serve as a robust standalone diagnostic marker.
Significantly elevated serum urea levels have been reported in preeclamptic pregnancies compared with normotensive pregnancies.18 Similarly, in our study, serum urea levels were significantly higher in the PE group than the control group. This finding is consistent with impaired renal function and reduced glomerular filtration rate in PE, which lead to accumulation of nitrogenous waste products; this interpretation is supported by the literature.
In a study evaluating 62 fetal growth restriction (FGR) cases and 62 control cases, first-trimester C-reactive protein, fibrinogen, and FAR levels were significantly higher in the FGR group.12 In our study, however, no statistically significant association was found between SGA and FAR. This discrepancy may be related to the small number of SGA cases (n=32) and the study’s overall limited sample size. Larger multicenter prospective studies are needed to elucidate more clearly the relationship between FAR and SGA.
FAR has been reported as a prognostic marker in hepatocellular carcinoma, where its prognostic value varied according to AST level.10 Because both fibrinogen and albumin are synthesized in the liver, hepatic dysfunction can affect both proteins and, thereby, the FAR. In our study, the FAR-AST correlation was weak (r = -0.170). Although statistically significant, it was of limited clinical relevance and was not in the same direction as the strong positive correlation reported in hepatocellular carcinoma.
Because the number of PE cases in our study was limited and the majority were classified as non-severe, patients were evaluated as an overall PE group. In this context, FAR levels in non-severe PE cases were similar to those in the control group.
Furthermore, in our study, FAR and other laboratory parameters were evaluated primarily during the third trimester, approximately one month prior to delivery, whereas many studies have measured these parameters during the first and second trimesters. This difference in timing may be an important factor explaining the heterogeneity observed across study results.
Study Limitations
The imbalance between the PE group (n=30) and the control group (n=330) limited the statistical power for FAR analyses. The retrospective study design and the absence of data on the precise timing of measurements represent major limitations. The SGA group (n=32) should be interpreted with caution, given the small sample size. Severity classification for PE could not be directly applied because blood pressure data were unavailable; most patients in the PE group had non-severe disease. Because of the small number of events (30 events; approximately 7.5 events per variable), the multivariate model is at risk of overfitting, and the AUCs for the combined model were derived in-sample and are therefore optimistic. The creatinine OR, with its very wide CI, and the ALT discrepancy between the Mann-Whitney and regression analyses reflect the narrow measurement range, possible outliers, and the small number of events, and both should be interpreted with caution. Severe PE cases below 34 weeks of gestation were referred for delivery to a tertiary university hospital with a neonatal intensive care unit, according to the regional referral chain; therefore, gestational age at delivery was unavailable for 11 and birth weight for 12 of the 30 patients in the PE group, resulting in missing delivery data. In addition, the control group included women hospitalized for various obstetric indications (e.g., preterm premature rupture of membranes, trauma, threatened preterm labor), which may themselves affect fibrinogen and albumin levels and thus FAR.
CONCLUSION
The FAR demonstrated limited diagnostic performance as a standalone marker for predicting PE, with an AUC of 0.527 on ROC analysis (Table 3). In contrast, albumin was identified as the strongest independent predictor of PE in all analyses (multivariate OR= 0.768; P < 0.001; Table 4), a finding that is consistent with the existing literature.15
No statistically significant association was found between FAR and birth weight, SGA, maternal age, or birth weight percentile (Tables 2, 5 and 6). Moreover, advanced maternal age (≥35 years) had no significant effect on either FAR or SGA (Table 6).
FAR measurements in our study were predominantly obtained during the third trimester or within 3-4 weeks before delivery. This constitutes an important difference from studies in the literature that rely on first- and second-trimester FAR measurements. In light of these findings, it is necessary to compare the clinical value of FAR in predicting PE across pregnancy trimesters. Prospective, multicenter studies comparing second-trimester FAR measurements with peripartum data are particularly recommended.
