Evaluation of hepatic toxicity in autoimmune hemolytic anemia (AIHA) and Evans syndrome patients: a single-center Egyptian study

Abstract

Introduction

Benign auto-immune illnesses include Evans syndrome (ES) and auto-immune hemolytic anemia (AIHA). Despite being benign in nature, the patients’ livers are burdened by the disease’s chronicity and the accompanying problems beyond the course of treatment. An additional burden stems from HCV infection, of which a significant proportion of Egyptians are positive. The purpose of this study was to identify the hepatotoxicity risks and the variables that influence the prognosis and survival of patients with AIHA/ES. There are 126 AIHA patients in this observational study, which is retrospective. From June 2009 to March 2021, patients visited the Haematology Unit of the Oncology Centre in Egypt. One hundred and sixteen patients have available data.

Results

There was no significant difference between primary and secondary AIHA groups as regards baseline hemoglobin (Hb), bilirubin, LDH, or reticulocyte count. Thirty-four patients (29.31%) had HCV-positive tests and 1 patient (0.9%) had HBV. There was no difference between HCV-positive and negative cases as regards mean Hb concentration, mean platelet, or immune markers (P > 0.05). AIHA patients with HCV-positive showed a significantly higher relapse rate (56%) than HCV-negative patients (32%) (P = 0.034). HCV positivity and low platelet counts at diagnosis were poor predictors for overall survival (OS) (P 0.022 and 0.04, respectively). Median OS was significantly better in patients with no viral hepatitis infection (1101 days, 95% CI 592–2068) than in patients with positive HCV infection (521, 95% CI 326–1325) (P = 0.019).

Conclusions

Azathioprine is the least hepatotoxic in AIHA patients under treatment. Viral hepatitis represents a superadded damage to the liver besides AIHA concerning clinical characteristics and outcomes.

Background

The term “auto-immune hemolytic anemia” (AIHA) refers to elevated erythrocyte turnover caused by immune system processes. One to three cases per 100,000 people is the estimated yearly incidence [1]. An autoimmune disorder known as Evans syndrome (ES) is characterized by the presence of two or more cytopenias, most commonly immune thrombocytopenia (ITP) and AIHA, either with or without immune neutropenia [2]. Based on the presence or absence of related diseases [3], such as systemic lupus erythematosus (SLE), lymphoproliferative disorders (LPD), malignancies, and drug-induced [2, 4], AIHA and Evan’s syndrome are classified as main (idiopathic) or secondary disorders. Immune dysregulation or the production of neo-antigens can be brought on by medications and hematologic treatments such as hematopoietic stem cell transplantation (HSCT) and checkpoint inhibitors (CPI) [5]. The direct antiglobulin test (DAT), which allows the disease to be classified based on the isotype and thermal properties of the autoantibody, is the gold standard for the diagnosis of AIHA [6]. The appearance of reticulocytosis, unconjugated hyperbilirubinemia, decreased haptoglobin, and elevated LDH can all be indicators of hemolytic anemia. Since these signs are more common in liver illnesses, baseline examinations in patients with ES, AIHA, or both should include evaluations of the liver architecture, virology screening, and liver function testing in order to rule out or confirm the presence of hepatic affection. Hemolytic anemia may exhibit false-positive laboratory results in liver cell failure [7].

The etiology of AIHA is extremely intricate. Molecular mimicry, the formation of prohibited clones, polyclonal B cell activation and antibody production, and the disruption of T-lymphocyte homeostasis by antigen-presenting cells (APC) are the primary immunological mechanisms that lead to tolerance breakdown. Increased levels of interleukin (IL)-2, IL-4, IL-6, IL-10, and IL-12 are accompanied by a decrease in interferon-gamma (IFN-γ), which promotes differentiation, T-helper response, and humoral and cellular autoimmune. B cells interact with T-helper cells; extravascular hemolysis occurs when IgG autoantibodies are produced and induce destruction of the red blood cells (RBCs) through antibody-dependent cellular cytotoxicity and phagocytosis, primarily in the spleen; IgM autoantibodies strongly activate the complement system. Transforming growth factor beta (TGF-β) is increased leading to T-helper 17 differentiation and production of IL-17 that decreases T-regulatory levels. Extravascular hemolysis is caused by IgG autoantibodies, which destroy RBCs through antibody-dependent cellular cytotoxicity and phagocytosis, primarily in the spleen. On the other hand, IgM autoantibodies strongly activate the complement system, which destroys RBCs through C3b opsonization and phagocytosis, primarily in the liver. B cells interact with T-helper cells. IgM has the potential to cause terminal complement activation, which could result in intravascular hemolysis until the membrane assault complex forms. The effectiveness of the bone marrow compensatory response determines the degree of anemia [5].

In response to corticosteroids, primary warm-AIHA (w-AIHA) responds well. As a second-line treatment, splenectomy is effective in 70% of primary AIHA patients [8]. Rituximab is increasingly being used as the first-line treatment for cold agglutinin disease (CAD) and as the preferred second-line treatment for steroid-resistant w-AIHA [9]. Additional B cell targeting agents, such as obexelimab and ianalumab, are being investigated to lengthen the duration of response and lessen toxicity. Agents that target plasma cells, such as daratumumab (ant-CD38 monoclonal antibody) and proteasome inhibitors like bortezomib, have demonstrated effectiveness in cases that are severe and refractory [10]. Immunosuppressants or the management of the underlying illness are two possible treatments for secondary AIHA, which is similar to how primary AIHA is treated [11].

The hepatitis virus’s involvement in the start or progression of AIHA has not been extensively studied in reports. On the other hand, reports of hepatitis virus coinfection linked to AIHA are uncommon [12]. AIHA may be an uncommon extrahepatic sign of long-term HCV infection [13, 14]. Hemolytic anemia and acute hepatitis B rarely coexist [15]. In 25–50% of patients, it can cause subclinical hemolysis [16]. W-AIHA has been observed in hepatitis B virus carriers who do not exhibit any symptoms [17, 18]. Many medications, including azathioprine and androgens, which are used to treat AIHA and/or Evans syndrome, have varying degrees of hepatotoxicity and should be used cautiously in patients who have hepatic impairment [7].

The Egyptian Demographic and Health Surveys (EDHS) measured antibody prevalence among the adult population aged 15–59 years at 14.7% in 2009 and at 10.0% in 2015, significantly higher than the global levels estimated at 1.4% by the World Health Organization. Hepatitis C virus (HCV) prevalence in Egypt was estimated at 11.9% in a 2018 meta-analysis. As a result, 6 million people in Egypt have a chronic infection [19]. According to the WHO, HCV infection is a serious public health issue that affects 3% of the world's population and results in 2–3 million new cases annually [20]. According to estimates, 125,000 viremic individuals become infected with HCV annually in Egypt, where the prevalence rate of HCV infection was 872,000 in 2013 (15% of the population) [21].

The objectives of this study were to determine the liver status in patients with Evans syndrome and AIHA, evaluate the impact of various AIHA treatment modalities on liver function, and identify any potential influence of viral hepatitis on the clinical features and prognosis of AIHA.

Methods

From June 2009 to March 2021, 126 AIHA patients who visited the Haematology Unit at the Oncology Centre in Egypt were included in this retrospective observational study. There were 116 patients’ worth of data. This study did not include patients with sickle cell disease, thalassemia, or hereditary spherocytosis, among other hereditary hemolytic disorders. Hepatocellular carcinoma, hypersplenism, and end-stage liver disease (Child score C) patients were not accepted. Patients with primary, secondary, and/or Evans syndrome who were 16 years of age or older and of either gender were eligible for this study.

Information was gathered from the registered medical records (https://srv137.mans.edu.eg/mus/newSystem/index.py) in accordance with the 1975 Helsinki Declaration, which was updated in 2008. The study was approved by the Faculty of Medicine’s Ethical Committee (Code Number: RP.21.11.121).

Patients’ complaints, medication histories for treating AIHA and/or Evans syndrome (corticosteroids, azathioprine, cyclosporine A, rituximab, splenectomy, etc.), and other medications that may cause secondary AIHA were among the information gathered for the presentation. Data from the physical examination were updated to include routine systemic examination for secondary cause detection as well as signs of hemolysis, such as jaundice, evaluation of comorbidities like diabetes, hypertension, or other conditions, and examination of the liver, spleen, and lymphoid regions.

In order to make a diagnosis, laboratory data were obtained at the patient’s initial presentation, at 6 months, and after 6 months of starting therapy, as well as during their final follow-up appointment. Using an automated hematology analyzer, the following laboratory and radiological data were gathered: total leucocytic count, platelet count, and complete blood count (CBC). Total leucocytic, and platelets count using automated hematology analyzer Cell dyn 1700 and Cell dyn emerald hematology analyzer. The measuring unit for hemoglobin (Hb) was gm/dl, red cell count (RBC) was m/µL, total leucocytic count (TLC) was k/µL, and platelet count was k/µL. Laboratory data also included blood smear and reticulocyte count. Lactate dehydrogenase (LDH) (U/l) and Coomb’s (direct and indirect).

Serum albumin (gm/dl), serum total bilirubin (0.3–1 mg/dl), direct and indirect bilirubin, aspartate aminotransferase (AST) (15–37 U/L), and alanine aminotransferase (ALT) (30–65 U/L) were among the tests used to estimate liver function using the Roche Cobas Integra-800 auto-analyzer. INR (International Normalised Ratio) as reported by Japan’s Sysmex autoanalyzer. Virology markers, such as ELISA testing for HIV, HBV, and HCV. When possible, use an immune profile that includes ASMA, ANCA, anti-ds-DNA, and ANA to detect autoimmune hepatitis or SLE. For patients with concomitant thrombocytopenia, a bone marrow examination was performed to rule out lymphoproliferative disorders or underlying cancer. The radiological investigations consisted of a CT scan (of the chest, abdomen, and pelvis) to rule out solid tumors or lymphomas as hepatocellular carcinoma suspected for 2ry AIHA, as well as an abdominal ultrasound to comment on hepatic texture, spleen size, and the presence of ascites.

Treatment data included (according to our institutional protocols):

• a. Standard first-line treatment was corticosteroids in the form of oral prednisone 1 mg/kg/day for at least 3 weeks then followed by gradual tapering or high dose dexamethasone (HDD) 40 mg intravenous (iv) for 4 days, or methylprednisolone (pulse steroid) iv 500–1000 mg/day for 3–5 days.

• b. As a second line, treatment with rituximab (whenever available) at doses of 100 mg/m²/week for 4 consecutive weeks or splenectomy. Oral azathioprine and cyclosporine A (CSA), mycophenolate mofetil (MMF), and others were also used as second or subsequent lines of treatment.

• c. The response criteria were defined as follows [22]:

  • Complete response (CR): normalization of hemoglobin with no evidence of hemolysis (normal bilirubin, LDH, and reticulocytes), and no need for transfusions.

  • Response (R): increase in hemoglobin by > 2 g/dL or normalization of hemoglobin without biochemical resolution of hemolysis; and absence of transfusion.

  • No response (NR): failure to achieve a response.

  • Steroid resistance: failure to obtain a hematologic response within 3 weeks on at least 1 mg/kg prednisolone.

  • Refractory disease failure to respond to at least 3 lines of therapy; in w-AIHA including splenectomy and/or at least one immunosuppressant.

• d. Supportive blood/platelet transfusion once indicated:

Blood transfusion was done for any patient with hemoglobin < 7 g/dl and in cases of hemodynamic instability. While platelet transfusion was ordered once the platelet count ≤ 10,000 and in severe bleeding manifestations.

Statistical analysis

The statistical analysis and the visualization of the data were done using R/R-Studio (version 3.0) program. The data were read from Excel (Office program 2010). For the statistical analysis, different packages were used including “survival”, “condsurv”, “ggsigniff”, “ggfortify”, “Survminer”, “rstatix”, and “corrplot”. For the visualization of the data, mainly the “tidyverse” package was used for the manipulation of the plots. The Shapiro–Wilk test was used to test the normality of the data. The qualitative and descriptive data were illustrated as numbers and percentages using the chi-square test. Quantitative data were analyzed as means, medians, and standard deviations. Wilcoxon test was used for the non-parametric data for the comparison between the two groups. ANOVA test was used for comparison between more than two non-parametric groups. Also, Pearson correlation was used to correlate between different parameters. Kaplan–Meier test was used for the survival analysis. COX regression was used for univariate and multivariate analysis for the survival predictors. The significant P value was considered when less than 0.05.

Results

Baseline laboratory investigations of all patients are illustrated in Table 1 showing female predominance (71%) with a mean age of 43.3 ± 15 years. Most of the cases were primary AIHA (80.17%). Different autoimmune markers were assessed: direct Coomb’s test (DCT) was positive in 99 patients. ASMA and ANCA tests were performed in 9 cases with suspected autoimmune hepatitis, and they were positive in only one patient.

Table 1 Clinical and laboratory characteristics of AIHA and/or ES studied patients n = 116

There was no significant difference between both AIHA groups as regards baseline Hb levels (P = 0.36), platelet counts (P = 0.13), or response to therapy (Fig. 1).

Fig. 1

PRBCs transfusion requirements. Patients who did not need PRBCs transfusion at any time of therapy = 25. Patients who needed PRBCs transfusion at any time of therapy = 87. Unknown = 4

HCV-positive patients showed a mean Hb concentration of 6.82 g/dL, mean platelet count of 176.53 × 10⁹/L. There was no difference between HCV-positive and negative cases as regards mean Hb, mean platelets, LDH, or immune markers including (DCT, ANA, anti-ds DNA) (P > 0.05). While AIHA patients with HCV-positive status showed a significantly higher relapse rate (56%) than HCV-negative patients (32%) (P = 0.034).

Liver function follow-up in HCV-positive vs. HCV-negative cases at different milestones (at 1st presentation and then at follow-up intervals; 3–6 months after treatment, > 6 months from diagnosis, and at the last follow-up visits). There was a significant difference in SGOT level between the 2 studied groups when comparing the diagnosis to the last follow-up (P = 0.0045). A significant difference was also detected regarding the level of serum bilirubin between evaluated milestones.

A comparison of CBC parameters at the presentation and other time points is illustrated in Table 2. This comparison showed significant improvement in Hb levels from the baseline to the last follow-up evaluations rather than platelet or TLC.

Table 2 CBC parameters at different time points

It was found that, the mean SGPT levels showed no significant difference between patients who received or did not receive 2nd line therapy. On the other hand, the mean SGOT level and total bilirubin showed a statistically significant difference between the diagnosis levels and levels at the last follow-up in both groups (P < 0.05).

Median OS was significantly better in patients with no viral hepatitis infection (1101 days) than in patients with positive HCV infection (521 days) (P = 0.019).

In Table 3, patients with Hb levels ≤ 7 g/dL (severe anemia) at presentation had no significant difference in median OS than patients presenting with Hb levels > 7 g/dL (P = 0.8). While patients with platelet counts < 150 × 10⁹/L showed significantly shorter median OS than patients with normal counts (672 days vs. 1239 days)(P = 0.038). There was no significant difference in median OS between corticosteroid responders and those who failed to respond to it (P = 0.99).

Table 3 Kaplan–Meier method in correlations with OS

Tables 4, 5, and 6 show uni- and multivariate analysis.

Table 4 Univariate analysis (COX regression for predictors of OS)

Table 5 Multivariate analysis (COX regression for predictors of survival) (HCV with all the variables) (significant variables only)

Table 6 Multivariate analysis (COX regression for predictors of survival) (platelets with all the variables) (significant variables only)

HCV positivity and low platelet counts at diagnosis were poor predictors for the OS (P 0.022 and 0.04, respectively). Other factors such as age, gender, etiology, Evans syndrome, second-line intake, liver functions, hemoglobin level, Coomb’s test status, LDH level, relapse, and refractoriness status did not impact the OS significantly.

The combinations with a significant impact on OS were platelet count with HCV, Direct Coomb’s test, and second-line intake (P: 0.0048, 0.018, and 0.017 respectively).

Discussion

Benign auto-immune diseases include Evans syndrome and auto-immune hemolytic anemia. Even though these conditions are not cancerous, the fact that they are chronic and have accompanying problems—whether they are brought on by the illness or the therapy—makes them difficult for the patients who have them, particularly the hepatic complications. This study sought to identify the covariates influencing the survival and outcome of patients with AIHA/Evans syndrome, as well as the risks associated with hepatotoxicity, taking into account the additional burden on the liver from HCV infection, which is positive in a significant portion of the Egyptian population.

This is a retrospective study of 116 patients with a median age of 44 years (range, 16–72) which is slightly younger than that recorded in previous studies where the median ages were over 50 years [23, 24] which could be attributed to different sample sizes and older than what reported in a previous Indian study [25].

In this study (Table 1), there is a female predominance (71%) which is in agreement with different studies and reviews [26]. Most of our cases were primary AIHA (~ 80%) with no obvious underlying etiology after performing the workup and ~ 19% of the cases were secondary and 43.5% of them were related to autoimmune disease, e.g., SLE and RA (rheumatoid arthritis); this is corresponding to various previous studies [23, 24, 27]. Malignant hematological diseases were found in 17.4% of patients presenting with hemolysis. Bone marrow examination detected underlying hematological disorders in 4 patients [n = 1 plasma cell dyscrasias, n = 1 hairy cell leukemia (HCL), and n = 2 Myelodysplastic syndrome]. CT scans revealed lymphadenopathy in 2 cases and were diagnosed as non-Hodgkin lymphoma (NHL).

Basma Atef et al. illustrated that the prevalence of AIHA, ITP (idiopathic thrombocytopenic purpura), and Evans syndrome was 6.9%, 3%, and 2% respectively among 101 chronic lymphocytic leukemia (CLL) patients in a previous observational retrospective study performed at our center [28].

Direct Coomb’s test (DCT) was positive in 99 patients with immune cytopenia, and 14 patients had a negative test. A negative DCT test in warm autoimmune hemolytic anemia (w-AIHA) could be misleading and necessitates further evaluation. The most common cause of a negative test in AIHA is technical, other causes include erythrocyte-bound antibodies below the detection limit by standard tests, or the presence of low-affinity IgG, IgA, or IgM antibodies reacting at warm temperatures and monomeric IgM not fixing complement. The test may need to be repeated using anti-IgG and anti-C3d reagents. DCT-positive or DCT-negative w-AIHA have the same clinical criteria and response to therapy [1].

There was no significant difference between 1 ry and 2 ry AIHA groups as regards baseline hemoglobin, bilirubin, LDH, or reticulocyte count [29]. This is consistent with our study. Also, we did not even detect significance between them as regards age, gender, or response to therapy. AIHA 1ry vs. 2ry did not show a difference in response to corticosteroids [30] which is also seen in our results (Table 7).

Table 7 Comparison between different parameters in 1ry and 2ry AIHA

Evans syndrome (ES) is rare (1–9 patients/1000,000 people/year) and most data we get about it are from pediatric retrospective studies so less is understood about adults [31]. The incidence of ES among our studied patients was 26% and 67% of them were females with a mean age of 40 years and the incidence of positive HCV test among them was 26% (7 patients). The ES international retrospective study showed that age, baseline Hb level < 8 g/dL, and occurrence of relapse, infection, and thrombosis were poor prognostic indicators of survival [31]. In our ES cases, we did not find a correlation with age, HCV status, response to 1st line, relapse, or refractoriness to further treatments. ES diagnosis did not affect survival, these differences between our results and the international multicenter study could be attributed to their larger sample size and their focus on ES characteristics.

Viral infections have been observed with AIHA, e.g., human immune deficiency virus (HIV), Epstein-Barr virus (EBV), cytomegalovirus (CMV), Parvovirus B19, and HCV. Viral infections trigger autoimmunity by mimicking molecular pathways between self-antigens and foreign viral antigens. The first International Consensus Meeting recommended testing for HBV, HCV, and HIV at baseline for patients presenting with AIHA and before initiation of therapy, while other viruses like EBV, CMV, and Parvovirus B19 are requested under certain circumstances [4, 22, 32]. Viral hepatitis is capable of inducing extrahepatic immune dysregulation and affection leading to immune cytopenia as AIHA [33]. HCV-associated AIHA has been described before in research as case reports or series [13, 34], while other reports have attributed hemolytic anemia to anti-viral hepatitis drugs [35, 36].

In our study, 34 patients (29.31%) had HCV-positive tests, 1 patient (0.9%) had HBV positive test, and no one was HIV-positive at diagnosis, while virology data were missing in 13 patients. A previous study showed that the incidence of AIHA among Egyptian patients with chronic HCV was 0.7% [37]. In another Egyptian study, AIHA was diagnosed in only 2 (0.95%) HCV patients, and 1 of them was induced by anti-viral therapy [38].

In Table 8, there was no difference between HCV-positive and negative cases as regards mean Hb concentration, mean PLT, or immune markers (P > 0.05). Previously 17 HCV-positive AIHA patients (F/M: 15/2; mean age, 55.8 years) were studied; their mean Hb concentration was 7.1 g/dL (range, 2.8–10.0 g/dL). ANA and anti-ds-DNA represented 41% and 6%, respectively of their immune profile [39].

Table 8 Comparison between HCV-positive and HCV-negative AIHA and/or Evans syndrome patients

Concerning the treatment protocol, all patients received corticosteroids as first-line treatment of choice for 1ry AIHA and/or Evans syndrome and in combination with the definitive therapy for 2ry AIHA patients with treatable causes. The steroid responders were estimated as the patients who achieved complete response (CR) or response (R) [n = 48 patients (41.38%)], and the non-responders (NR) were estimated as the patients who were steroid resistant [n = 65 patients (56.03%)], there were 3 patients with unknown response to steroids as they lost follow-up. Approximately 62.93% of the total included patients received second-line treatments.

To define the correlation between HCV status and the response to steroids, the chi-square test was used revealing that there is a positive correlation between HCV positivity and steroid response but with no statistical significance (P = 0.13) (see Table 8). AIHA patients with HCV-positive status showed a significantly higher relapse rate (56%) than HCV-negative patients (32%) (P = 0.034). On the other hand, AIHA patients with HCV-negative status showed a higher treatment refractoriness or resistance rate (57%) than HCV-positive patients (32%) (P = 0.025). Hepatic impairment after treatment was more pronounced in HCV-positive patients (39%) than HCV-negative patients (10%) (P = 0.002), which indicates a higher burden of HCV on these patients.

The significant difference that was found in SGOT level (rather than SGPT) between the 2 studied groups when comparing the diagnosis to the last follow-up (P < 0.05), as well as, in the level of serum bilirubin between evaluated milestones could be related to the super added effect of HCV on liver condition besides the drug-induced liver injury caused by different lines of treatment especially cyclosporine (see Table 9).

Table 9 Liver functions in HCV-positive vs. HCV-negative cases at different milestones (n = 103)

The liver enzymes and total bilirubin were compared in patients who received 2nd line therapy to patients who did not receive 2nd line. The patients were compared as a total and then sub-classified according to the type of the 2nd line received. The mean SGOT level and bilirubin (rather than SGPT) showed a statistically significant difference between the diagnosis levels and levels at the last follow-up in both groups. This can be attributed to the cholestatic pattern of liver injury induced by these drugs. To dissect the previous data and go into its depth, the enzymes, and total bilirubin were analyzed regarding each 2nd line therapy used.

In patients who received rituximab, there was no significant difference in the mean SGPT or SGOT levels between the diagnosis and any time point in contrast to the patients who did not receive rituximab who showed a significant difference in the mean of the enzymes’ levels between diagnosis and last follow-up. However, the total bilirubin showed a significant difference in the means at diagnosis and the last follow-up in patients who received rituximab. In patients who received azathioprine, SGPT levels did not show any significant differences between diagnosis and different time points. SGOT levels showed significant differences between diagnosis and last follow-up in patients who did not receive azathioprine. Total bilirubin showed significant differences between diagnosis and all the time points in both groups. Serum bilirubin showed a significant difference between the time of diagnosis and all time-point levels in the arm of patients who didn’t receive cyclosporine and only a significant difference between the diagnosis and last follow-up levels in patients who received cyclosporine. The same significant differences were seen when we assessed the patients who received combined therapies as in Table 10.

Table 10 Comparison between patients who received and did not receive different Second-line therapies regarding liver functions (SGPT, SGOT, and total bilirubin)

These results indicate that liver functions are improving better in patients who did not receive 2nd line therapy. SGOT and bilirubin are more sensitive to hemolysis status and drug toxicity. Based on the results, the least hepatotoxic drug that could be considered is azathioprine as the liver functions improved between the diagnosis and the last follow-up in both groups.

According to survival, there was no significant difference in the overall survival (OS) regarding the gender of studied patients; the median OS for male patients was 839 days (95% CI 506–1743) and the median OS for female patients was 1101 days (95% CI 583–1720) (P = 0.86). There was also no significant difference in OS between 1ry or 2ry cases (P = 0.75). However, median OS was significantly better in patients with no viral hepatitis infection (1101 days, 95% CI 592–2068) than in patients with positive HCV infection (521, 95% CI 326–1325) (P = 0.019) (Fig. 2).

Fig. 2

Survival probability between HCV-positive and HCV-negative cases

HCV positivity and low platelet counts at diagnosis were found to be poor predictors for the OS (P 0.022 and 0.04, respectively). Multi-factorial combinations were analyzed by combining them using COX regression. A negative impact on OS with statistical significance was detected between HCV status with gender, HCV status with age, HCV status with platelet counts, and HCV with second-line intake (P 0.04, 0.051, 0.0048, and 0.024, respectively). Other HCV combinations may have a significant protective impact on the OS of AIHA patients such as HCV status with steroid response status, and HCV with relapse or refractoriness status (P 0.048, 0.054, and 0.025 respectively) as shown in Table 5.

We did not find a relation between response to corticosteroids and baseline Hb or platelet levels or HCV status. Abdallah et al. and Barcellini et al. agreed that baseline Hb level had a negative predictive effect on response [40], while in our study there was no impact of baseline Hb concentration on OS (P 0.8), and it was the platelet count < 150 × 10⁹/L that had a negative impact on OS as shown by Kaplan–Meier, univariate and multivariate COX regression analyses.

Limitations facing our study included its retrospective nature, the sample size, and the lack of uniformity during second-line management. We faced difficulties in analyzing relapses and refractoriness, and collecting liver function data during follow-up because there are no uniform guidelines for choosing a drug over the other, we usually face problems with drug availability financially and on the other hand the patient’s affordability and tolerability to treatments.

However, there was an analysis of quite a large number of patients including the section of patients with Evans syndrome. Also, we compared primary and secondary AIHA as well as HCV-positive and HCV-negative patients receiving different lines of treatment.

In conclusion, there is impairment of hepatic status either clinical (more decompensation) or laboratory parameters (SGOT, SGPT, and serum bilirubin) in patients with AIHA and Evans syndrome. The effect of different lines of treatment of AIHA on liver enzymes is variable and it needs meticulous follow-up, especially for serum bilirubin. The least hepatotoxic drug observed was azathioprine. Viral hepatitis infection (especially HCV) represents a superadded damage to the liver besides AIHA concerning clinical characteristics and outcomes. Median OS was significantly better in patients with no viral hepatitis infection (P = 0.019).

Follow-up liver enzymes and serum bilirubin are a must when we shift the treatment from one line to another and once suspected hepatic adverse events. Larger prospective studies are needed to validate the results.