Randomised clinical trial: intravenous vs oral iron for the treatment of anaemia after acute gastrointestinal bleeding

Summary Background Acute gastrointestinal bleeding is prevalent condition and iron deficiency anaemia is a common comorbidity, yet anaemia treatment guidelines for affected patients are lacking. Aim To compare efficacy and safety of intravenous ferric carboxymaltose (FCM) and oral ferrous sulphate (FeSulf) in patients with anaemia secondary to non‐variceal gastrointestinal bleeding Methods A prospective 42‐day study randomised 61 patients with haemoglobin <10 g/dL upon discharge (Day 0) to receive FCM (n = 29; Day 0: 1000 mg, Day 7: 500 or 1000 mg; per label) or FeSulf (n = 32; 325 mg/12 hours for 6 weeks). Outcome measures were assessed on Days 0 (baseline), 7, 21 and 42. The primary outcome was complete response (haemoglobin ≥12 g/dL [women], ≥13 g/dL [men]) after 6 weeks. Results A higher proportion of complete response was observed in the FCM vs the FeSulf group at Days 21 (85.7% vs 45.2%; P = 0.001) and 42 (100% vs 61.3%; P < 0.001). Additionally, the percentage of patients with partial response (haemoglobin increment ≥2 g/dL from baseline) was significantly higher in the FCM vs the FeSulf group (Day 21:100% vs 67.7%; P = 0.001, Day 42:100% vs 74.2%; P = 0.003). At Day 42, normalisation of transferrin saturation to 25% or greater was observed in 76.9% of FCM vs 24.1% of FeSulf‐treated patients (P < 0.001). No patient in the FCM group reported any adverse event vs 10 patients in the FeSulf group. Conclusion FCM provided greater and faster Hb increase and iron repletion, and was better tolerated than FeSulf in patients with iron deficiency anaemia secondary to non‐variceal acute gastrointestinal bleeding.


| INTRODUC TI ON
Acute gastrointestinal bleeding (GIB) is a prevalent condition (90-108 cases per 100 000 incidence) with a substantial economic burden and is a direct cause of substantial mortality (3%-14%). 1,2 Iron deficiency anaemia (IDA) following non-variceal upper GIB is estimated to occur in approximately 50% of all cases. Notably, mortality among patients with upper GIB is mostly related to the decompensation of underlying diseases, in which anaemia can play a key role. 3 The treatment of GIB-associated anaemia, therefore, requires both correction of the underlying cause and adequate repletion of iron stores in case of IDA. 4 Despite the known frequency of GIB and the negative consequences of IDA, guidelines for the effective treatment of GIB-associated anaemia remain poorly defined. 5 Oral administration of ferrous sulphate (FeSulf) is still recognised as the first-line treatment for IDA patients. However, for patients with gastrointestinal disorders, oral iron shows substantial limitations due to insufficient absorption, slow course of action and severe gastrointestinal side effects that can exacerbate existing symptoms. [5][6][7] Due to this intolerance, it is estimated that oral iron treatment is discontinued in up to 50% of IDA patients. 8 For patients with poor tolerance of oral iron, a large iron deficit or need for rapid response to treatment, intravenous (iv) iron administration is the treatment of choice. [9][10][11] Intravenous iron is administered as an iron-carbohydrate complex, and various formulations that can be given at different maximum doses or minimum administration times (mainly depending on the product's stability and premature release of iron) are currently available. 12,13 By bypassing the gut, iv iron circumvents slow absorption associated with oral iron, promotes rapid uptake by the reticuloendothelial system and prevents gastrointestinal inflammation. 14 Additionally, large doses of iv iron can be administered over a short period of time, facilitating a rapid and long-lasting response. Overall, iv iron administration is associated with improved treatment compliance and reduced demand for red blood cell transfusions, which is highly recommended given their scarcity, cost and potential risks. 2,6,[15][16][17] The use of iv iron has been well studied for the treatment of IDA associated with a wide range of underlying digestive pathologies, including inflammatory bowel disease, coeliac disease and colorectal cancer as well as non-digestive haemorrhagic events. 8,[18][19][20][21][22][23][24] However, only few studies have analysed the efficacy of different iron regimens in patients with IDA secondary to acute GIB and even less for patients with lower GIB. One randomised study in patients with upper GIB compared a fixed single dose of iv iron (ferric carboxymaltose, FCM) vs oral iron (FeSulf). While the increase of haemoglobin (Hb) was independent of the iron administration route, the correction of iron deposits was more rapidly achieved by iv FCM than oral FeSulf. 25 A recently published retrospective analysis of patients who were admitted with acute GIB and treated with FCM either alone or combined with red blood cell transfusions suggested effectiveness and safety of FCM in this population. 26 The randomised study presented here compared the efficacy and safety of iv FCM given at weight-and Hb-adjusted doses according to the summary of product characteristics (SmPC) and oral FeSulf in patients with anaemia secondary to non-variceal acute GIB, aiming to improve the evidence base in lieu of established treatment protocols.

| Patients
Clinically stable patients admitted to the University General Hospital of Valencia, Spain, with non-variceal GIB (aged >18 years) and subsequent diagnosis of anaemia secondary to acute GIB (Hb <10 g/dL on the day of hospital discharge, Day 0) who did not meet the exclusion criteria were enrolled into the trial. GIB was defined as any sign or symptom of macroscopic exteriorisation of haemorrhage accompanied by changes in relevant laboratory values (eg decrease of Hb).
Obscure origin GIB was defined as exteriorisation of macroscopic anal haemorrhage of undefined cause after gastroscopy and colonoscopy.
Exclusion criteria were pathologies that could influence anaemia progression (chronic renal failure, liver disease, inflammatory bowel disease, neoplasia, uncompensated thyroid disorders, malabsorption, HIV, haematological disorders, folic acid or vitamin B12 deficiency and previous gastric surgery). Further exclusion criteria were non-iron deficiency anaemia or anaemia of mixed origin, a history of chronic anaemia, treatment with erythropoietin, iron supplements, folic acid or vitamin B12 in the year prior to inclusion and contraindications to treatment with study drugs. In addition, available clinical history and laboratory data of patients covering one year prior to inclusion were reviewed to exclude patients with undiagnosed iron deficiency prior to the GIB that lead to hospital admission.
The criteria for withdrawal from the study were relapse of GIB or requiring transfusion after hospital discharge.

| Study design
The study was designed as a single centre, prospective, unblinded study and was conducted in accordance with the Helsinki Declaration and adherence to Good Clinical Practice guidelines. At each visit, patients were asked for treatment adherence. Patients declaring that they took >90% of tablets were considered adherent.
No patient diaries were used, but patients were asked how many tablets they had left in the box. In case of non-adherence, patients were asked if it was due to adverse effects. Patients who did not show Hb normalisation at the end of the study were treated according to local clinical practice (ie iv FCM for patients not responding to oral FeSulf).

| Outcome measures
The primary outcome measure was complete response rate (percent of patients who reached Hb levels ≥12 g/dL or ≥13 g/dL in women and men respectively) at Day 42. Secondary outcome measures included partial response rate (percentage of patients with Hb increase ≥2 g/dL vs baseline at Day 7, 21 or 42), the rate of iron status normalisation (percentage of patients who achieved TSAT ≥25%; centre-specific cut-off for normal TSAT) and serum ferritin levels.
Adverse events were considered treatment related when they were described in the drugs' SmPCs and chronologically related to the treatment administration (ie the first 24 hours after infusion for FCM and the entire 42-day treatment period for oral iron).

| Statistics
The sample size was calculated on the premise that an increase in Hb of 0.5 g/dL per 100 mg of absorbed iron and a minimum Hb increase of 1.6 g/dL would be clinically relevant. Comparison of both groups utilised a standard deviation of 2, an alpha value of 0.05 and expected power of 80%. Thus, with an estimated 15% loss to followup, each group needed to include 29 patients completing the study.
For the analysis of the categorical variables (complete response and partial response), the response rates were calculated and Fisher exact test or chi-square test was used. For the comparison of the quantitative parametric variables (demographic and analytical descriptive variables), mean values and standard deviation (SD) were calculated followed by application of Student's t test. In case of quantitative non-parametric variables, the Mann-Whitney U test (Wilcoxon Rank Sum Test) was applied.

| Hb response
At the end of the study (Day 42), complete response was achieved by 100% of FCM-treated patients compared with 61.3% of oral   Figure 3.
Mean serum ferritin levels increased rapidly in the FCM group and remained >100 µg/L from Day 7-42 ( Table 2).

| Adherence to treatment
All FCM-treated patients adhered to treatment compared with 84.4% in the FeSulf group (P = 0.03). Non-adherence to oral FeSulf was due to abdominal pain and related dose reduction or forgetting to take the treatment. Notably, two patients in the FeSulf group required FCM rescue medication due to very low Hb levels (8 and 8.9 g/dL) on Day 21.   10,18,20,21,24,29,[31][32][33][34][35][36] The results are also consistent with a meta-analysis that assessed studies using different iv iron preparation to treat IDA in patients with different conditions and concluded that although all iv and oral formulations could correct IDA, FCM was the superior treatment. 37 The rapid increase in serum ferritin as well as TSAT is typical for the rapid uptake and utilisation of intravenously administered iron, bypassing the slow and limited enteral absorption of oral iron. As it is known that iv iron rapidly increases ferritin, this is the reason for high observed ferritin levels and not an inflammatory background as basal values were normal. Furthermore, our study assessed ferritin levels early, which therefore cannot be compared with the assess-  for patients at risk of organ dysfunction or haemodynamic instability. 39 In any case, decisions about the use of transfusion during this period were not influenced by the study because inclusion into the study was decided on the day of hospital discharge.

| D ISCUSS I ON
In our study, treatment with FCM was well tolerated during administration and the reported events were without clinical repercussion. Based on a terminal half-life of 7-12 hours for the clearance of FCM and its average residence time of 11-18 hours, 40 observations of potential drug-related events were only expected in this early period during/after iv administration (Day 0-7). No acute effects were reported in this period and despite frequent assessment of potential adverse events on Days 0, 7, 21 and 42 over the 42-day study period only minor events that were considered unrelated to FCM were observed. Even sensitive assessment of liver function (by the measurement of aspartate transaminase and alanine transaminase) showed only two recorded elevations of alanine transaminase that were below 1.5 times the upper limit of normal and considered not clinically relevant. Conversely, in the FeSulf group, a substantial proportion of gastrointestinal adverse events was detected in the FeSulf group, which even forced the reduction of the FeSulf dose by half in one of the patients. The association of gastrointestinal adverse events with oral iron administration is known from several studies. 19,21,[41][42][43][44] Consistent with the good tolerability of FCM, adherence to treatment was 100%. The adherence observed in the FeSulf group (84%) was lower compared to the FCM group but still high compared to prior published studies. 8,45 However, this high percentage is based on patients' self-declaration when they were asked for adherence during examinations, and therefore, non-adherence may be underreported.
Subjective state of health at the end of the study as assessed by EQ -VAS scale was significantly better in FCM-treated compared with FeSulf-treated patients. To our knowledge, this is the only currently available study reporting quality of life in patients with F I G U R E 5 Results of quality of life questionnaire EQ-5D-3L and EQ-VAS. For EQ-5D-3L dimensions (mobility, self-care, usual activities, anxiety/depression and pain/discomfort) higher percentages correspond to more patients with problems. For EQ-VAS, higher scores correspond to better health status (quality of life  46,47 Although no great differences in the dimensions self-care and usual activities of patients were observed between both treatment groups, there seems to be a higher percentage of patients in the oral FeSulf group who reported problems in mobility, pain-discomfort and anxiety dimen- While cost-effectiveness was not an endpoint of this study, a recently published cohort study of patients with gastrointestinal disease and IDA suggested a potential advantage of iv over oral iron in terms of healthcare utilisation due to lower risk of hospital reattendance and shorter length of stay. 48 Our study showing rapid correction of both anaemia and iron deficiency among FCM-treated patients with Hb levels <10 g/dL can support the identification of a feasible cut-off when to prefer iv over oral iron in patients with gastrointestinal disease.
In conclusion, this study demonstrates that in patients with anaemia after acute gastrointestinal bleeding, iv ferric carboxymaltose offers a faster and more efficient normalisation of haemoglobin and iron status parameters compared to the oral administration of ferrous sulphate. Additionally, the administration of ferric carboxymaltose displays a low profile of adverse events and improved patient quality of life, further contributing to an overall favourable benefit-risk profile.
Overall, the results of this study could provide clinical support for the inclusion of ferric carboxymaltose into treatment schemes and guidelines for patients with anaemia secondary to acute haemorrhage of gastrointestinal origin.

ACK N OWLED G EM ENTS
Declaration of personal interests: Dr. Gisbert has served as a speaker, a consultant and advisory member for or has received research funding from MSD, Abbvie, Hospira, Pfizer, Kern Pharma, Biogen, Takeda,