The inverted U-shaped association between blood fibrinogen and rehospitalization risk in patients with heart failure

Fibrinogen, a biomarker of thrombosis and inflammation, is related to a high risk for cardiovascular diseases. However, studies on the prognostic value of blood fibrinogen concentrations for heart failure (HF) patients are few and controversial. We performed a retrospective analysis among acute or deteriorating chronic HF patients admitted to a hospital in Sichuan, China, between 2016 and 2019, integrating electronic health care records and external outcome data (N = 1532). During 6 months of follow-up, 579 HF patients were readmitted within 6 months, and 46 of them died. Surprisingly, we found an inverted U-shaped association of blood fibrinogen levels with risk of readmission within 6 months but not with risk of death within 6 months. It was found that HF patients had the highest risk for readmission within 6 months after reaching the turning point for blood fibrinogen (2.4 g/L). In HF patients with low fibrinogen levels < 2.4 g/L, elevated fibrinogen concentrations were still significantly associated with a higher risk for readmission within 6 months [OR = 2.3, 95% CI (1.2, 4.6); P = 0.014] after controlling for relevant covariates. There was no significant association between blood fibrinogen and readmission within 6 months [(OR = 1.0, 95% CI (0.9, 1.1); P = 0.675] in HF patients with high fibrinogen (> 2.4 g/L). The effect difference for the two subgroups was significant (P = 0.014). However, we did not observe any association between blood fibrinogen and death within 6 months stratified by the turning point, and the effect difference for the stratification was not significant (P = 0.380). We observed an inverted U-shaped association between blood fibrinogen and rehospitalization risk in HF patients for the first time. Additionally, our results did not support that elevated blood fibrinogen was related to increased death risk after discharge.

stroke had a significantly increased mean plasma fibrinogen concentration (> 4 g/L) 10 .A significant association between blood fibrinogen concentrations and the presence of ischemic lesions on cerebral computed tomography was observed: patients with a fibrinogen concentration > 3.41 g/L showed a 3.29-fold increased risk of ischemic lesions 10 .Moreover, a prospective study containing 14,916 men aged 40-84 years in the Physicians' Health Study evaluated the relationship between the plasma concentration of fibrinogen and myocardial infarction (MI) risk 11 , reporting that a high blood fibrinogen concentration (≥ 343 mg/dL) had a twofold increase in MI risk (RR = 2.09; 95% CI 1.15-3.78)compared with those with fibrinogen below 343 mg/dL, independent of other CVD risk factors 10 .Moreover, a previous meta-analysis with 1773 men who were free of HF or cardiac arrhythmias reported 131 sudden cardiac deaths within 22 years of follow-up, thus suggesting that blood fibrinogen concentrations were positively, log-linearly and independently associated with an increased risk of sudden cardiac death 12 .However, there are currently few or unclear explorations for the association of blood fibrinogen concentration with prognosis in HF patients.Only two clinical studies have primarily evaluated the effect of blood fibrinogen on the prognosis of HF, and their results are controversial and have obvious limitations.One of those studies reported that high fibrinogen levels (≥ 284 mg/dL) independently predicted 90-day mortality in critically ill patients with acute exacerbation of chronic HF, but those authors emphasized that their findings need to be further validated by large prospective studies with longer follow-up times 13 .Another study also only analyzed 120 patients with chronic HF and did not observe that elevated fibrinogen levels were associated with a high risk for all-cause mortality during a follow-up period of 2 years 14 .
In this study, our objective was to further evaluate the value of blood fibrinogen for predicting prognosis (risk for readmission and all-cause mortality within 6 months) with a large sample size of elderly HF patients (N = 1532) who were admitted to a hospital in Sichuan, China between 2016 and 2019.

HF patients
Our study subjects comprised a representative sample from a large number of Chinese HF individuals (N = 2008) who had been admitted to Zigong Fourth People's Hospital with HF between December 2016 and June 2019 15,16 .These patients were hospitalized due to acute or worsening chronic HF, and electronic health care records and external outcome data were integrated.Therefore, this study was designed as a cohort study to investigate risk factors for related outcomes (death risk and readmission risk) 15,16 .Of a total of 2008 HF patients selected into the dataset, subsequent hospital readmission and mortality risk were collected from a mandatory follow-up visit at 28 days, 3 months and 6 months 15,16 .The follow-up visit was replaced by a telephone if the HF patients were not able to visit the clinical center.According to the European Society of Cardiology (ESC) criteria 17 , HF was defined as follows: (1) the presence of symptoms and/or signs of HF; (2) elevated levels of BNPs (BNP > 35 pg/ mL and/or NT-proBNP > 125 pg/mL); and (3) objective evidence of other cardiac functional and structural alterations underlying HF.In case of uncertainty, a stress test or invasively measuring elevated left ventricle (LV) filling pressure may have been needed to confirm the diagnosis 17 .For the present analyses, after excluding HF patients with missing information on blood fibrinogen, relevant covariates and outcomes (N = 476), a final cohort of 1532 individuals with nonmissing information was enrolled in our analysis.Figure 1

Statistical analysis
We used descriptive analyses to summarize the admission characteristics for these HF patients.First, crosssectional associations of blood fibrinogen levels with readmission within 6 months and death within 6 months were evaluated by using calculated logistic regression (adjusted for age and sex in Model 1; adjusted for age, sex, BMI, MAP, type of heart failure and NYHA cardiac function classification in Model 2).All analyses were performed using odds ratios (ORs) with 95% confidence intervals (CIs).To further characterize the shapes of associations, a smooth curve was used to confirm real-world relationships between blood fibrinogen levels and readmission within 6 months and death within 6 months with ORs and 95% CIs.As the association showed a nonlinear shape, a turning point for blood fibrinogen (2.4 g/L) was detected in which HF patients had the highest risk for readmission within 6 months.Therefore, we conducted hierarchical analysis through the turning point (HF patients with fibrinogen < 2.4 g/L and with fibrinogen > 2.4 g/L).Finally, subgroup analyses were performed to examine statistical evidence of effect modification by clinically relevant characteristics, such as comorbidities including congestive HF, MI, cerebrovascular disease, diabetes, PVD, moderate-to-severe CKD, dementia, COPD and liver disease.Empower version 4.1 was used for analysis, and P ≤ 0.05 was considered to indicate statistical significance.

Ethics approval
The research ethics committee of Zigong Fourth People's Hospital (Approval Number: 2020-010) approved the study, and informed consent was waived by the research ethics committee of Zigong Fourth People's Hospital due to the retrospective design that all methods were performed in accordance with the Declaration of Helsinki guidelines.

Admission characteristics in patients with HF
The admission characteristics for a total of 1532 HF patients who met the inclusion criteria are described in Table 1.The number of male individuals among these HF patients was 637 (41.58%), and the number of individuals > 60 years old was 1401 (91.45%).The median blood fibrinogen level in all HF patients was 3.05 (2.51-3.76)g/L.Among these patients, the number for NYHA cardiac function classification from II to IV in these patients was 265 (17.30%), 800 (52.22%) and 467 (30.48%), respectively.Diabetes (23.56%) and moderateto-severe CKD (23.43%) were the most common comorbidities.Because 1426 (93.08%) of them had congestive HF, 47.72% of HF patients were admitted to the emergency room with worsening HF.Importantly, during 6 months of follow-up, 579 (37.79%) of these HF patients were readmitted within 6 months, and 46 (3.00%) died within 6 months.Then, we used logistic regression to explore the association of blood fibrinogen level with risk for readmission within 6 months and risk for death within 6 months in Table 2.Not surprisingly, we did not observe significant associations between blood fibrinogen and readmission within 6 months [OR = 1.03, 95% CI (0.93, 1.13); P = 0.6174] or death within 6 months [OR = 1.23, 95% CI (0.96, 1.57); P = 0.1065] in Model 1 when only covariates including age and sex were adjusted for.After other covariates, including BMI, MAP, type of heart failure, NYHA cardiac function classification and creatinine, were further added into the above model, their associations of blood fibrinogen with readmission within 6 months [OR = 1.04, 95% CI (0.94, 1.15); P = 0.4116] and death within 6 months [OR = 1.16, 95% CI (0.89, 1.51); P = 0.2713] remained meaningless.

Curvilinear association between blood fibrinogen and the outcomes within 6 months in HF patients
We further explored whether there was a curvilinear association between blood fibrinogen and adverse outcomes within 6 months (readmission and death).Unexpectedly, we observed an inverted U-shaped relationship between blood fibrinogen and risk for readmission within 6 months (Fig. 2A, P = 0.163) rather than death within 6 months (Fig. 2B, P = 0.272) after age, sex, BMI, MAP, type of heart failure, NYHA cardiac function classification and creatinine were adjusted.It was found that HF patients had the highest risk for readmission within 6 months after reaching the turning point for blood fibrinogen (2.4 g/L).This is an interesting and even strange discovery because the normal value of fibrinogen is 1.5-3.5 g/L.Therefore, grouping analysis through the turning point was performed (the low group with fibrinogen < 2.4 g/L and high group with fibrinogen > 2.4 g/L), as shown in Table 3.In HF patients with low fibrinogen < 2.4 g/L, we found that elevated fibrinogen concentrations of 1 g/L were still significantly associated with an increased risk of 130% for readmission within 6 months [OR = 2.3,   NYHA cardiac function classification and creatinine.Interestingly, there was no significant association between fibrinogen and readmission within 6 months [OR = 1.0, 95% CI (0.9, 1.1); P = 0.675] in HF patients with high fibrinogen > 2.4 g/L.The log likelihood ratio tests for the two subgroups was significant (P = 0.014).However, we still did not observe any significant associations between blood fibrinogen and death within 6 months stratified by a turning point for blood fibrinogen (2.4 g/L), and the effect difference for the stratification was not significant (P = 0.380).In addition, we also conducted stratified analysis by smooth curve fitting and found that the results were similar to the above (Fig. 3A and 3B).

Subgroup analysis of composite diseases for readmission risk within 6 months in HF patients
Subgroup analysis was conducted to evaluate the association between blood fibrinogen and risk for readmission and death within 6 months in different subgroups by composite diseases (Table 4).Significant differences in stratification for the association between fibrinogen and readmission within 6 months were not observed in MI, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, COPD, diabetes, moderate-to-severe CKD and liver disease respectively (all P > 0.05).Predictably, these subgroup analyses also showed no association between blood fibrinogen and death risk within 6 months (data not shown).

Discussion
Our analysis using an electronic health database comprising patients with HF from a hospital in Sichuan, China, between 2016 and 2019 provided several new findings that have not been reported previously.
(1) Indeed, blood fibrinogen was not strongly correlated with the risk for death within 6 months in HF patients after controlling for some traditional and nontraditional risk factors.( 2) We additionally found a nonlinear (inverted U-shaped) association of baseline fibrinogen with the risk for readmission within 6 months.It was found that HF patients had the highest risk for readmission within 6 months after reaching the turning point for blood fibrinogen (2.4 g/L).
A large amount of evidence has confirmed that increased blood fibrinogen levels contribute to CVD risk and worsen prognosis via platelet aggregation, plasma viscosity and fibrin formation in various populations.For instance, Danesh et al. conducted a meta-analysis of 31 prospective studies, indicating that each 1 g/L increase in blood fibrinogen concentration could contribute to an increased risk of 142% for CAD, 106% for stroke and 176% for other vascular deaths 18 .These significant relationships were independent of age, sex and other cardiovascular risk factors 18 .Another meta-analysis of 52 prospective studies performed by Kaptoge et al. involving 246,669 individuals without previously diagnosed CVD examined the prognostic effects of blood fibrinogen levels on CVD risk 19 .These researchers reported that elevated blood fibrinogen levels were related to an increased risk of 15% for CVD and drew a conclusion that assessing blood fibrinogen levels may be a very favorable indicator for detecting an intermediate CVD risk 19 .In addition, there is evidence also supporting that blood fibrinogen may be a valuable biomarker for subclinical atherosclerosis.In the Coronary Artery Risk Development in Young Adults study including 1396 participants aged 25-37 who were assessed for carotid intimal/medial thickness (CIMT) and coronary artery calcification (CAC) 20 , Green et al. found that the prevalence values of CAC with increasing quartiles of fibrinogen were 14.4%, 15.2%, 20.0% and 29.1%, and a similar trend was observed for CIMT.They concluded that increased fibrinogen concentration in these individuals was independently associated www.nature.com/scientificreports/with subclinical CVDs 20 .Although these clinical investigations have populations with various CVDs, their conclusions were almost similar that high fibrinogen is closely related to poor prognosis, which is partly consistent with the results of our study.We found that HF patients with blood fibrinogen at the turning point (2.4 g/L) had the highest risk for readmission within 6 months, and elevated baseline fibrinogen independently increased the risk of readmission by 130% within 6 months in HF patients with fibrinogen < 2.4 g/L with each 1 g/L increase, while there was no significant association in these patients with high fibrinogen > 2.4 g/L.This is an interesting and even strange discovery because the normal value of fibrinogen in blood is 1.5-3.5 g/L, with a half-life of 3-5 days in generally healthy people [21][22][23] .In fact, many risk factors and health statuses could increase blood fibrinogen concentrations, including older age, female sex, Black race, smoking and drinking, hypertension, obesity, diabetes, metabolic syndrome and others.Elderly patients with HF are often accompanied by these risk factors, placing them in a hypercoagulable state and increasing their rehospitalization risk.Another potential explanation is that fibrinogen is also an acute phase protein, as its biosynthesis increases during the inflammatory process [24][25][26] .An increased systemic inflammatory response may contribute to a high risk for cardiovascular events requiring readmission, along with elevated blood fibrinogen concentrations (< 2.4 g/L).In fact, few studies have reported the prognostic value of blood fibrinogen in predicting death risk in acute or exacerbated states of chronic HF.One recent study included a total of 554 HF patients from the Medical Information Mart for Intensive Care III (MIMIC III) database, and 90-day mortality was defined as the primary outcome 13 .They found that in the unadjusted Cox model, the 90-day mortality hazard ratio (HR) with 95% confidence intervals of the high fibrinogen level was 3.33 (95% CI 2.15-5.15)compared with the low fibrinogen level (< 284 mg/dL).The HR of the high fibrinogen level for 90-day mortality in multivariable Cox models changed slightly.They concluded that high fibrinogen levels (≥ 284 mg/dL) independently predict mortality in critically ill patients with acute exacerbation of chronic HF 13 .In contrast, our conclusion did not support the close correlation between elevated fibrinogen levels and a high risk for all-cause mortality, even though we observed a significant turning point at 2.4 g/L for blood fibrinogen.This difference may come from the fact that our outcome variable was all-cause mortality rather than disease-specific mortality.In another prospective cohort study of 1773 men aged 42-61 years who were free of HF or cardiac arrhythmias and whose plasma fibrinogen was measured at baseline, fibrinogen was positively and independently associated with the risk of sudden cardiac death during a 22-year follow-up 12 ; however, those individuals mainly came from non-HF populations with younger ages from 42 to 61 years old, so their results are not comparable with those of our study.Moreover, our study data mainly came from patients discharged from the cardiology department after treatment, but their data came from patients in various ICUs (MIMIC III).Therefore, the difference in population selection may be an important reason for the difference in conclusions.Importantly, consistent with our findings, Chin et al. 14 analyzed 120 patients with chronic and stable HF and did not observe that elevated fibrinogen levels were associated with a high risk for all-cause mortality.Similarly, our HF subjects were included in this analysis after receiving standard treatment for HF with a relatively stable condition.The only difference is that the follow-up time for our study was 6 months, while their follow-up time was 2 years.
The major advantage of our study was that a large sample of HF patients were enrolled in the study population from elderly patients admitted to a hospital in Sichuan, China, where there were no losses during the follow-up of 6 months and clinical variables were allowed for adequate adjustment.It is also the first exploration showing an inverted U-shaped association between blood fibrinogen and risk for readmission 6 months in a Chinese population with HF, which provided more clinical evidence for previous controversial studies concerning HF prognosis.This may also provide a new candidate biomarker for the prognosis of HF patients in clinical practice that "blood fibrinogen = 2.4 g/L" among these patients had the highest risk for readmission within 6 months.Naturally, some limitations should be noted in this study.First, we could not control for the effects of unmeasured or unknown confounders for our findings, even though traditional CVD risk factors were controlled for.Second, fibrinogen in our study was only measured once at admission.Additional measurements of fibrinogen levels at discharge and/or follow-up may reliably determine the relationship of fibrinogen with prognosis.Third, the dataset was collected from a single center, so models developed using the data may not be generalizable.Fourth, the dataset did not offer time series data throughout the hospitalization period.For example, administration times for medications are not available.Thus, drug variables were not included in our analysis.Finally, we cannot clearly explain why there was no association between blood fibrinogen and all-cause mortality risk in the Chinese population, which is contrary to the results in previous investigations with the American population 13 .It is also unknown whether there are racial differences in this relationship.This should encourage more similar research to provide more evidence for our conclusion in different races.

Conclusion
In summary, our findings suggested a curvilinear association (inverted U-shaped association) between blood fibrinogen and rehospitalization risk.The "blood fibrinogen = 2.4 g/L" among these patients had the highest risk for readmission within 6 months.This might also provide new evidence and perspective on the prognosis of blood fibrinogen for HF patients in clinical practice.certainly, more clinical evidence were needed to validate our conclusions.

Figure 1 .
Figure 1.Flow chart of final HF patients for analysis.

Figure 2 .
Figure 2. Curve analysis for associations between blood fibrinogen and risk for readmission and death within 6 months in HF patients.

Figure 3 .
Figure 3. Stratified analysis for associations between blood fibrinogen and risk for readmission and death within 6 months in HF patients.

Table 2 .
The association between blood fibrinogen and outcomes within 6 months in HF patients.Model 1: Adjusted for age and gender.Model 2: Adjusted for age, gender, BMI, MAP, type of heart failure, NYHA cardiac function classification and creatinine.HF heart failure, MAP mean arterial pressure, BMI body mass index, NYHA New York Heart Association.

Table 3 .
The turning point analysis for association between blood fibrinogen and outcomes within 6 months in HF patients.Adjusted for age, gender, BMI, MAP, type of heart failure, NYHA cardiac function classification and creatinine.HF heart failure, MAP mean arterial pressure, BMI body mass index, NYHA New York Heart Association.Turning point

Table 4 .
The stratified analysis for association between blood fibrinogen and readmission within 6 months in HF patients.Adjusted for age, gender, BMI, MAP, type of heart failure, NYHA cardiac function classification and creatinine.HF: heart failure; COPD: chronic obstructive pulmonary disease; CKD: chronic kidney disease; Moderate to severe CKD: chronic kidney disease with glomerular filtration rate < 60 ml/min; MAP: mean arterial pressure; NYHA: New York Heart Association.