We collected 102 patient records for cirrhosis in the study. Male patients were predominantly represented, accounting for 61.7% (sex ratio: 1.6). The average age was 54.7 ± 16.3 years, ranging from 17 to 86 years. Table 1 shows the distribution of the population according to age and sex.

Hepatic encephalopathy was observed in 62.7% of patients, with grade I predominating (38.3%). Ascites was present in 82.4% of patients, mainly grade III (45.1%).

Esophageal varices (EV) were found in 52% of patients (n = 53). Grade III EVs accounted for 35.2% of cases, followed by grade II (14.8%) and grade I (2%). Gastrointestinal bleeding was observed in 15.7% of cases (n = 16).

Biologically, patients had low prothrombin levels (TP) in 76.4% (n = 63), hypoalbuminemia in 68.6% (n = 68), and hyperbilirubinemia in 46.1% of cases (n = 47).

The main causes of cirrhosis were HBV (36.3%), HCV (30.4%), alcohol (21.5%), and diabetes (5.9%).

The majority of patients, 52% (n = 53), were in stage C of the Child-Pugh classification, followed by Child-Pugh B score in 45% of cases (n = 46). The frequency of anemia was 75.5% (n = 77).

The average age of patients with anemia was 53.9 ± 17.5 years (range: 17 - 86 years) compared to 56.8 ± 12.3 years (range: 29 - 78 years) for patients without anemia. The frequency of anemia increased with age but was more severe among younger subjects, with no statistically significant difference. Male patients were relatively more affected (57.2%). However, anemia was more severe in female patients, with an average Hb of 7.9 ± 1.8 g/dL compared to 8.2 ± 1.8 g/dL in male patients, with no significant difference ( Table 1 ).

In most cases (52%), it was a moderate, normocytic, normochromic anemia (mean Hb: 8.1 g/dL; range: 3.8 - 10.9 g/dL; mean MCV: 80.4 fL; mean MCHC: 27.8 pg). Table 2 shows the distribution of patients by anemia characteristics.

Diagnostically, hemoglobin concentration decreased with the presence of gastrointestinal bleeding, the stage of hepatic encephalopathy, and the abundance of ascites. Anemia was significantly associated with decreased TP and albumin levels. It did not vary according to the grade of EV. Table 3 presents these results.

Anemia was more frequent in post-hepatic cirrhosis due to HBV and HCV, with 67.5% (n = 52/77), and was more severe in patients with HBV, though without significant difference. Table 4 shows the distribution of patients with anemia by etiology.

Others: HVB + HVC (n = 4), HVB + Alcohol (n = 1), HVB + diabetes (n = 1).

Unlike patients without anemia, the average hemoglobin concentration significantly decreased with the Child-Pugh score in those with anemia (p = 0.002). Figure 1 illustrates the distribution of cirrhotic patients with anemia according to the Child-Pugh score.

Information on meal composition and frequency was not mentioned in the records. Additionally, no iron profile was performed for any patient.

Our study aimed to determine the prevalence of anemia among cirrhotic patients hospitalized in the GEMI department at the University Hospital of Brazzaville, to identify its various characteristics, and to examine its relationship with different epidemiological, diagnostic, and prognostic parameters to improve patient care. In this context, the discussion focused primarily on two aspects.

In this study, we assessed the role of anemia in the complications of cirrhosis and investigated its association with various patient profiles. Patients were selected based on data gathered from medical records and operational definitions we chose for anemia and other variables.

The main challenge was accessing medical records post-hospitalization. During the study period, 234 cases of cirrhosis were admitted to the Gastroenterology Department. Due to the high cost of certain tests and poor record-keeping, several files were unusable, and many patients were discharged against medical advice. These factors, along with a high rate of loss to follow-up, introduced selection bias, explaining the low sample size in our study. This also applied to reticulocyte counts, initially planned for our study to explore normochromic normocytic anemias, and other tests like erythrocyte ferritin levels, soluble transferrin receptor, and inflammatory profiles (ESR, CRP, fibrinogen) for microcytic anemias. Dietary factors that could have contributed to the etiological research were not studied because they were not mentioned in the medical records.While nutritional surveys may be performed for certain patients with specific malnutrition issues, they are not done systematically. Similarly, iron profile tests and other explorations may be conducted for persistent anemias, but in cirrhotic patients, these are generally considered only for severe and persistent anemia in the absence of bleeding. In the context of this retrospective survey, information on dietary factors was not available.

The frequency of anemia in cirrhosis varies across studies. Our results align with those of Nalcoulma et al. in Burkina Faso, who reported a 74.5% rate in 2007 [5] . In 2011, Sawadogo in Burkina Faso and Doumbia in Mali reported rates of 54.9% and 44.9%, respectively [10] [11] . These differences could be attributed to various factors, including the causes of anemia, the severity of liver disease, and the operational definitions used in each study, as in the study by Boivin et al., where the criterion chosen was the isotopic measurement of globular volume [12] . Maruyama et al. found 70%, defining normal Hb levels at 13 g/dl for men and 12 g/dl for women [6] . Eichner and Hilman reported 60%, considering anemia as a hematocrit below 38% in alcoholic patients [7] . Despite these differences, it can be concluded that the prevalence of anemia is high in cirrhosis.

The male predominance among study patients may explain the higher anemia frequency in this population. The severity of anemia in women, not described in other studies, could be linked to additional mechanisms not reported here, such as blood loss due to menstruation.

Regarding the increased frequency of anemia with age, our findings align with the literature, as tissue tolerance is lower in older individuals. Conversely, the severity of anemia in younger patients could be explained by its rapid onset or physical activity. Indeed, these conditions might account for the earlier and more severe presentation of anemia in younger patients compared to older ones [13] .

The diagnostic approach to anemia in cirrhosis is not different from general practice. However, two factors complicate it: first, usual diagnostic tools lack specificity in cirrhosis; second, there is a degradation of some diagnostic markers due to the disease (ferritin, MCV, reticulocytes, haptoglobin, etc.) [8] [14] [15] .

The approach to diagnosing anemia in cirrhosis is similar to that for other conditions but is complicated by the disease’s multifactorial and often concurrent nature and the lack of specificity of usual biological tests [8] . In cirrhosis, several diagnostic markers deteriorate, including reticulocyte counts, giving a false impression of central anemia, increased ferritin levels, reduced hepatic synthesis of transferrin associated with increased transferrin saturation, and decreased plasma haptoglobin levels [14] - [18] . Similarly, the dilution and splenic sequestration phenomena associated with cirrhosis may produce or worsen false anemia. However, an etiological search is essential as there is a correlation between liver disease severity, assessed by the Child-Pugh score, and anemia severity [6] [19] . The pathophysiology, informed by knowledge of red blood cells and erythroblast physiology, enables a logical diagnostic approach.

The predominant normochromic normocytic character found in our study is consistent with Nacoulma et al. [5] . Normocytic anemia in cirrhotic patients can have various causes, including hemorrhagic anemia, dilutional (false) anemia, hemolytic anemia, and marrow insufficiency. The term “normocytic” here should be understood in a broad sense, encompassing slightly macrocytic anemias, especially after ruling out vitamin deficiencies, since macrocytosis may be simply an expression of the alcohol-related macrocytosis often seen in cirrhotics [9] [12] [13] [16] . In cirrhotics, a normocytic anemia, especially if it presents acutely or with poor clinical tolerance, should raise suspicion of gastrointestinal bleeding, even if reticulocyte counts are not significantly elevated [8] . However, normocytic anemia in cirrhosis is often false anemia due to hemodilution, linked to plasma volume expansion [12] [20] [21] . This dilution can be exacerbated by overhydration, as seen in alcohol withdrawal. This contributes to what is commonly referred to as the “simple and common anemia of cirrhosis” [22] . In the absence of gastrointestinal bleeding, red cell morphological abnormalities, signs of hemolysis, or iron/vitamin deficiency, this diagnosis may be retained for moderate anemia, with hemoglobin levels above 10 g/dl in women and 11 g/dl in men.

The origin of normocytic anemia may also be hemolytic. Hemolysis is a major determinant of anemia in cirrhosis, as shown by studies measuring erythrocyte half-life isotopically in patients with shortened erythrocyte lifespans, indicating pathological hemolysis [20] . Hemolytic anemia, in its simplest form, reflects hypersplenism. Splenic sequestration and the impact of splenomegaly contribute to this often-latent hemolysis.

Bone marrow failure may be a cause of normocytic anemia seen in most patients. Bone marrow failure is almost constant in cirrhotic anemia due to the direct toxicity of alcohol or its metabolites, or medications. It may also be a consequence of partial iron or vitamin deficiencies (nutritional anemias). This marrow insufficiency, more often relative than absolute, becomes a cofactor of anemia in cases of hemolysis or mild bleeding [12] [22] . Anemia results because inefficient erythropoiesis fails to compensate for losses. For over a decade, erythropoietin secretion in cirrhotics has been a research focus, though results are highly controversial. Some studies report higher EPO levels in cirrhotics than in non-anemic controls, while others find lower EPO levels in cirrhotics than in anemic controls [23] [24] . A blood smear would have allowed us to specify possible marrow involvement after ruling out other causes [8] .

Microcytic anemias are also frequently observed in cirrhosis, often as microcytic hypochromic anemias, as seen in our study and in previous ones [22] [25] [26] . The most common cause of these anemias is iron deficiency, which must be distinguished from inflammatory anemia, especially if there are clinical and biological signs of inflammation (elevated ESR, CRP, low transferrin). Iron deficiency anemia is a significant factor in cirrhotic anemias, raising questions about the role of portal hypertension; portal hypertension gastropathy (telangiectasias, lesions similar to angiodysplasia, and colonic varices) is common in cirrhotic patients and may cause slow gastrointestinal bleeding [26] . However, Calès et al. found no significant relationship between large esophageal varices and hemoglobin levels after adjusting for the Child-Pugh score. Similarly, there is no significant relationship between serum iron and these endoscopic signs of portal hypertension [25] . Erythrocyte ferritin measurement, a marker less affected by liver disease than serum ferritin, would have enabled diagnosis [27] .

Inflammatory microcytic anemia may also occur in cirrhosis, although it is less frequently discussed. Its impact is likely underestimated and should be more accurately assessed, especially considering the current focus on inflammatory processes in alcoholic cirrhosis [22] .

Normocytic hypochromic and microcytic normochromic ane-mias are less frequent. Nalcoulma et al. and Doumbia observed these in 11.7% and 10% of cases, respectively [5] [11] , which is close to our cumulative frequency of 13% (10/77). These unusual hematologic features, difficult to categorize, may result from multiple anemia-inducing factors, such as malnutrition due to vitamin and trace element deficiencies and intestinal parasitic infections.

Regarding macrocytic anemia, Nalcoulma et al. and Doumbia reported it in 25% and 5% of cases, respectively [5] [11] . Our results are lower than those of Dénié et al. in France, who found a rate of 48% [4] . This type of anemia, commonly seen in alcoholic cirrhosis, is much more prevalent in Europe, where alcoholism is the main etiology of cirrhosis. Chronic alcoholism often leads to macrocytosis with increased gamma-glutamyl transferase levels [5] . Macrocytosis regression following alcohol cessation supports the diagnosis [6] . Folate or vitamin B12 deficiencies are the other main causes of macrocytic anemia, especially in cases of moderate macrocytosis, with a mean cell volume between 95 and 105 fl [15] .

New biological tools provide relevant insights for diagnosing anemia in cirrhosis, particularly of alcoholic origin. For instance, analyzing erythrocyte morphology on blood smears shows promise for diagnosing hemolytic anemias [6] [7] [15] . The protein profile, particularly the albumin–transferrin dissociation, soluble transferrin receptor levels, orosomucoid–haptoglobin dissociation, and iron reserves measured on Perls-stained bone marrow smears, are uncommon in clinical practice but could contribute to characterizing and investigating the etiology of anemia in cirrhotic patients, especially in alcoholic cirrhosis [28] - [30] .

Data regarding anemia’s impact on hepatic encephalopathy align with literature findings that gastrointestinal hemorrhage complicated by anemia is a risk factor for the occurrence and worsening of hepatic encephalopathy [31] [32] .

Aside from platelet counts, the prothrombin level is one of the most commonly prescribed tests for detecting hemorrhagic disease, whether acquired or congenital. In cirrhosis (an acquired condition), hepatocellular insufficiency results in a coagulopathy with signs that depend on the severity of liver damage, regardless of its cause. This leads to a decrease in PT, which may be accompanied by anemia [33] .

As for the association between anemia and hypoalbuminemia, it could be linked to malnutrition, a frequent complication of cirrhosis, or to a pseudo-hypoalbuminemia related to hemodilution. The same goes for the relationship between anemia and ascites grade, which could suggest a false anemia due to hemodilution.

The effects of anemia on the splanchnic circulatory changes of cirrhosis are not precisely known; only preliminary findings suggest that anemia may worsen the hyperdynamic circulation characteristic of the disease [34] . In the study by Calès et al., a significant relationship was indeed found between the size of esophageal varices and hemoglobin reduction; however, this relationship disappeared after adjusting for the Child-Pugh score [8] [25] . Conversely, gastric antral vascular ectasia or “watermelon stomach” in its most advanced form is less controversial and is well recognized as a cause of iron deficiency anemia [8] [35] .

The prognostic significance of anemia in cirrhosis varies widely, ranging from a mere epiphenomenon to severe anemia that can threaten life, particularly when caused by gastrointestinal bleeding. Our findings support existing literature, which confirms a correlation between the severity of liver disease, as measured by the Child-Pugh score, and the depth of anemia [4] [11] .