Red Cell Distribution Width as a Predictor of Survival in Patients with Hepatocellular Carcinoma

Background and Objectives. Hepatocellular carcinoma (HCC) and the intrahepatic biliary tract cancers are estimated to rank sixth for incidence among solid cancers worldwide, and third for mortality rates. A critical issue remains the need for accurate biomarkers for risk stratification and overall prognosis. The aim of this study was to investigate the ability of a biomarker of heterogeneity of the size of red blood cells, the red cell distribution width (RDW), to predict survival in patients with HCC. Materials and Methods. A consecutive series of patients with a histologic diagnosis of HCC were included into this study irrespective of their age, stage of the disease, and treatment administered, and followed-up for a period of three years. Demographic, anthropometric [age, sex, body mass index (BMI)], and clinical data (Charlson Comorbidity Index, Child–Pugh score, etc.), along with laboratory tests were retrieved from clinical records. Results. One-hundred and four patients were included in this study. Among them, 54 (69%) were deceased at the end of the follow-up. Higher RDW values, but not other hematological and biochemical parameters, were significantly associated with mortality in both univariate and multivariate analysis. The optimal RDW cut-off value identified with the Youden test for survival was 14.7%, with 65% sensitivity and 74% specificity (AUC  =  0.718, 95% CI 0.622–0.802, p  <  0.001). Kaplan–Meier survival curves showed significantly lower survival with higher RDW values (HR = 3.5204; 95% CI 1.9680–6.2975, p < 0.0001) with a mean survival of 30.9 ± 9.67 months for patients with RDW ≤ 14.7% and 22.3 ± 11.4 months for patients with RDW > 14.7%. Conclusions. The results of our study showed that RDW can perform better than other blood-based biomarkers in independently predicting prognosis in patients with HCC.


Introduction
Hepatocellular carcinoma (HCC), together with the intrahepatic biliary tract cancers, is estimated to rank sixth for incidence among solid cancers worldwide, and third for mortality rates, with more than 900,000 new cases and more than 830,000 deaths in 2020, according to the Global Cancer Observatory (GLOBOCAN) data [1].Albeit the prevalence of HCC is highest in Eastern Asia and sub-Saharan Africa, where chronic hepatitis B infection is endemic and the food is often contaminated with the mycotoxin aflatoxin B1, at least one quarter of the cases is idiopathic; recent cancer epidemiology reports have shown an increasing incidence of liver cancer in Western countries in the recent two decades, especially in women [2,3].The close link between incidence and mortality rates reflects the lack of effective screening strategies, difficulties in early diagnosis, and poor outcomes with current therapeutic options, especially when diagnosed in advanced stage and in patients with compromised liver function and general clinical conditions.Patients with advanced stage disease are generally treated with multidisciplinary therapies, including surgical, locoregional, systemic, and supportive treatment options encompassing several specialties comprising surgery, radiology, oncology, hepatology, and palliative care.A critical issue, however, remains the need for accurate risk stratification in order to select the best treatment option and predict clinical outcomes.Circulating biomarkers such as bilirubin and albumin are useful for the classification of the underlying cirrhosis, and alpha fetoprotein is used in combination with other clinical and imaging factors for disease stage, while numerous other biomarkers are currently under evaluation [4].

Materials and Methods
A consecutive series of patients with a histologic diagnosis of HCC attending the Department of Medicine, Surgery and Pharmacy of the University of Sassari, Italy, between 2009 and 2017, were included in this study irrespective of their age, stage of the disease, and treatment.Patients were followed-up for three years, without any censoring.Demographic and anthropometrical data [age, sex, body mass index (BMI)], as well as data on comorbidities and clinical status (Charlson Comorbidity Index, Child-Pugh score, etc.), along with laboratory tests were retrieved from patients' clinical files and saved in a digital database.In addition, data regarding single (medical, surgical, or ablative) or combined treatments were collected.Pediatric patients, as well as those with no complete demographic, pathological, and clinical data were excluded.All patients were informed about the aims and the procedures of this study and provided written consent.This study was performed in accordance with the principles of the declaration of Helsinki and was approved by the ethical committee of the University Hospital of Cagliari (A.O.U Cagliari, Protocol number PG/2019/4493).
Results were reported as means and standard deviations (SDs) or medians and interquartile ranges (IQRs).Individual variables' distribution was classified using the Shapiro-Wilk test.Statistical differences were investigated with the unpaired Student's t-test or the Mann-Whitney rank-sum test, according to variable types.Univariate and multivariate logistic regression analysis was performed to assess correlations between individual variables and three-year mortality (primary end-point).Receiver operating characteristics (ROC) curve tests were carried out to identify the best cut-off values, maximizing sensitivity and specificity.Sensitivity and specificity were calculated using the optimal ROC curve value according to the Youden Index.For survival analysis, time zero was set as the time of diagnosis.Survival probability was investigated using the Kaplan-Meyer approach and the log-rank test, with death being the end point.Cox proportional hazards regression was carried out for both univariate and multivariate analyses.Hazard ratios were calculated from Cox analysis.Multivariate analyses were adjusted for those confounders that showed associations with p-values of <0.1 in univariate analysis.Statistical analyses were performed using MedCalc for Windows, version 20.014 64 bit (MedCalc Software, Ostend, Belgium).

Discussion
RDW is a simple and inexpensive parameter which reflects the degree of heterogeneity of erythrocyte volume (known as anisocytosis) and accelerated red cell turnover; it is traditionally used in laboratory hematology for the differential diagnosis of anemias [48].Anisocytosis, defined as the presence of red blood cells (RBCs) with a broad heterogeneity of size and volume in peripheral blood, can be caused by numerous clinical conditions, including congenital erythrocyte disorders (i.e., β-thalassemia, sickle cell disease, hereditary spherocytosis), anemia (e.g., due to iron, folate, or vitamin B deficiencies), blood transfusions, some forms of hemolytic anemias, oxidative stress, inflammation, impaired renal function, and cancer [48].RDW is calculated by dividing the standard deviation (SD) of erythrocyte volumes for the mean corpuscular volume (MCV) (i.e., RDW = SD/MCV).Although the result can be expressed either in absolute values (i.e., RDW-SD) or as a percentage (i.e., RDW-%), the latter is more widely used in routine laboratory practice and was the one adopted in this study [49].
RDW alterations in cancer may be due to various events, including the direct stimulation of red cell turnover by the tumors themselves, the immune responses to their dissemination, or to a state of chronic occult bleeding and anemia.Fancellu et al. recently described significant RDW differences between right-sided and left-sided colorectal cancers [31].They hypothesized that the right colon, especially the caecum, is significantly larger than the left colon, and has less anatomical relations with neighboring organs in comparison to the rectum.This allows right-sided tumors to grow and remain clinically silent for a long time and explains why these lesions affect older patients and have an exophytic, ulcerated aspect with a higher stage at diagnosis.Older studies estimated that the time interval between cancer onset and clinical manifestations varies between 4.8 and 5.8 years [50].During growth, the lesions become ulcerated and bleed.Chronic occult bleeding causes iron loss and thus chronic microcytic anemia.Bleeding is common in other long-lasting slowly growing solid tumors, especially those of the gastrointestinal, respiratory, and genitourinary tract, and this may be the reason for the differences in RDW values.In addition, central necrosis with intratumoral red cell consumption is common in solid parenchymal tumors, such as those of the liver and the intrahepatic biliary tract.
A few studies have investigated the role of the RDW in predicting prognosis in HCC.The first study was published in 2015 by Smirne et al. [51].The authors retrospectively analyzed a training cohort of 208 patients with HCC and an independently prospectively collected validated cohort of 106 patients with HCC.In both cohorts, median survival time was significantly lower in patients with RDW ≥14.6% at the time of diagnosis.The median survival in the training cohort was 1026 days in the low RDW group (RDW ≤14.6%) vs. 282 days in the high RDW group (HR = 0.43; 95% CI: 0.31-0.60,p < 0.0001).The patients HCC were then classified in four quartiles according to their RDW values, and the median survival decreased progressively with each increasing RDW quartile.In the validation cohort, the median survival was 868 days in patients with RDW <14.6%, whereas it was 340 days in patients with RDW >14.6%.In the training cohort, the survival rates at 1, 2, and 3 years were 79%, 57%, and 42% in patients with RDW <14.6% compared to 48%, 29%, and 18% in patients with RDW >14.6%, respectively.In both the training and validation cohorts, an increased RDW value was independently associated with higher mortality.Notably, the cut-off level identified by the authors (14.7%) is identical to the one identified in our study (14.7%).
Hu et al., in a study comprising 298 patients (including 81 HCC cases) and 66 healthy controls, showed that the RDW was associated with worse hospital outcomes (AUC 0.76 (95% CI 0.67-0.84)[43].RDW values above 15.15% were independently associated with poor hospital outcomes after adjustment for serum bilirubin, platelet count, prothrombin time, albumin, and age, with an odds ratio of 13.29 (95% CI 1.67-105.68).In 2016, Zhou et al. investigated the relationship between the preoperative serum RDW value and the clinic and pathologic characteristics of 106 patients with HCC who underwent curative resection [30].Two groups were identified based on a cut-off level that was very similar to the one described by Smirne et al., as well as in our study: high RDW (>14.5%, n = 28) and low RDW (≤14.5%, n = 78).The patients with preoperative high levels of RDW had significantly worse survival than those with low RDW (p < 0.05).According to multivariate analysis, high RDW (HR = 1.89, p = 0.002), TNM stage (HR = 1.70, p = 0.019), tumor size (HR = 1.33, p = 0.045), tumor number (HR = 1.42, p = 0.027), and vascular invasion (HR = 1.64, p = 0.009) were independent prognostic factors of overall survival in this study.The same year, Wei et al. retrospectively reviewed the medical records of 110 treatment-naive patients with HCC and 68 controls and found that the RDW was significantly increased in patients with HCC and correlated with the liver function tests [52].No correlation between RDW and the stage of the disease was found in this study.
Subsequently, Howell et al. investigated the prognostic role of the RDW in patients with HCC treated with sorafenib [53].A total number of 442 subjects were enrolled with a median follow-up time of 7.1 months.The baseline RDW was found to be significantly associated with survival (HR 1.23, 95% CI: 1.12-1.29).Interestingly, the authors found that the AUC markedly increased from 0.669 to 0.787 after the RDW, and no treatmentrelated diarrhea was considered as a factor when determining the CLIP score (Cancer of Liver Italian Program Score) for predicting 12-month survival.In a retrospective study encompassing 442 patients with HCC, Jing et al. reported that the RDW in patients with HCC with a tumor size larger than 10 cm was higher than that of patients with HCC with a tumor size smaller than 3 cm, 3-5 cm, and 5-10 cm (14.77 ± 2.35%, 15.27 ± 2.65%, 15.32 ± 2.40% vs. 15.97 ± 2.39%, p < 0.001); furthermore, the RDW level significantly increased with worsening Child-Pugh grades and BCLC stages [54].Finally, they found that patients with HCC with RDW values below 14.15% (HR 0.530, 95 CI 0.395-0.710;p < 0.001) had lower mortality rates.
Golriz et al. investigated the role of preoperative, six-and twelve-month RDW values in predicting recurrence after curative resection for HCC in 395 patients [55].The authors reported that recurrence-free survival (RFS) was significantly higher among patients with low RDW at 6 and 12 months, postoperatively (p < 0.001 and p = 0.028).In addition, RDW values higher than 16.15% at 6 months (HR: 2.047, p < 0.001) and higher than 15.85% at 12 (HR: 3.105, p < 0.002) months after liver resection were independent predictors of relapse-free survival.More recently, Tan et al. conducted a retrospective study comprising 745 patients with hepatitis B(HBV)-related HCC, 253 patients with chronic hepatitis B (CHB), as well as 256 healthy individuals, which included as a control group to enable a comparison to be made regarding hematological parameters and RDW values [56].Potential risk factors for long-term all-cause mortality in patients with HBV-related HCC were predicted using multivariate Cox regression.The RDW of patients with HBV-related HCC was significantly higher when compared to CHB and healthy controls.In the former, splenomegaly, liver cirrhosis, larger tumor diameter, multiple tumor number, portal vein tumor thrombus, and lymphatic or distant metastasis were significantly increased, and a greater Child-Pugh grade and BCLC stage was correlated with a higher RDW.Furthermore, multivariate Cox regression analysis identified RDW as an independent risk factor for predicting long-term all-cause mortality in patients with HBV-related HCC.Finally, the authors generated a nomogram incorporating RDW with other clinical parameters and tested it in a training and a validation cohort; their nomogram showed high accuracy in predicting the survival and prognosis of patients with HBV-related HCC, suggesting its potential utility in clinical practice.
Despite a consistent methodological and treatment heterogeneity among these studies, they all showed that RDW has a potential predictive role in HCC prognosis.In our study, we included patients treated with different approaches.In addition, the cut-off level for a good predictive performance (AUC 0.718) found in our study was very similar to those described in other studies.Notably, the RDW maintained its predictive ability, even after we adjusted for numerous clinical factors, and performed better than other blood-based biomarkers.Although the mechanisms underpinning the superior performance of the RDW remain elusive, it has been hypothesized that a state of chronic inflammation and oxidative stress, typically associated with HCC, could explain the accelerated red blood cell turnover and anisocytosis [57].
Our study has some limitations, particularly regarding its retrospective design and the relatively small number of patients, which did not allow for consistent subgroup analyses to be performed in accordance to the treatment received by the patients.In addition, the data regarding the dimensions of the tumors and details on the underlying etiological causes were not analyzed.It must also be underlined that most of the patients included in this study had a low BCLC stage and Child-Pugh score, which may have negatively impacted the predictive capacity of the RDW.Further studies should include a proportion of patients with advanced-stage HCCs.On the other hand, our study shows the potential of RDW to independently predict prognosis in patients with HCC, regardless of the underlying etiology and treatment, and the ability to perform better than other biomarkers.

Conclusions
Our results show the potential of RDW to independently predict prognosis in patients with HCC, regardless of the treatment, and the ability to perform better than other biomarkers.Considering that RDW is a low-cost, easy-to-perform interpretable biomarker, its potential clinical application in this context is attractive.Nevertheless, such an application needs to be further evaluated in future prospective, randomized trials.
Institutional Review Board Statement: This study was conducted in accordance with the Declaration of Helsinki and was approved by the ethical committee of the University Hospital of Cagliari (A.O.U Cagliari, protocol number PG/2019/4493, 28/03/2019).
Informed Consent Statement: Informed consent was obtained from all subjects involved in this study.

Figure 1 .
Figure 1.Receiver operating characteristics (ROC) curves for RDW as predictor of survival.

Medicina 2024 , 13 Figure 2 .
Figure 2. Kaplan-Meier curves for three-year survival probability of HCC in patients with different RDW levels.

Figure 2 .
Figure 2. Kaplan-Meier curves for three-year survival probability of HCC in patients with different RDW levels.

Table 1 .
Demographic, clinical, and laboratory data of the whole population and at three yearssurvivor and non-survivor sub-groups.

Table 2 .
Univariate logistic regression analysis assessing association between demographic, clinical, and laboratory data with three-year mortality.

Table 3 .
Multivariate logistic regression analysis assessing association between studied variables and three-year mortality.

Table 4 .
Univariate COX regression analysis assessing association between studied variables and survival.

Table 5 .
Multivariate COX regression analysis assessing association between studied variables and survival.