Several hypotheses may account for how anemia influences on cardiovascular-related mortality. First, anemia augments the imbalance between myocardial oxygen supply and demand by decreasing oxygen-carrying capacity and increasing myocardial oxygen consumption to increase cardiac output. Second, experimental data suggest that ACS patients with anemia may have an impaired capacity for vascular healing [15]. In humans, anemia has consistently been associated with adverse cardiovascular outcomes in the general population. Many prospective cohort studies provided consistent evidence for mortality or adverse cardiovascular outcomes [16]. Data from the cohort study including 3,921 patients showed that anemia defined by World Health Organization (WHO) increased all-cause mortality and death caused by progressive heart failure in patients diagnosed with acute myocardial infarction (AMI) [16], whereas in a study by Valeur and colleagues anemia was an independent risk factor for mortality in ACS patients with heart failure only [17]. Recently, a large database of over 422,855 ACS patients from England and Wales showed the prevalence of anemia in ACS cohort of 28% and anemia as an independent risk factor for mortality [18].

In the case of heart failure, anemia increased risks of all-cause mortality by 21–47% with more than 1 year of follow-up [16]. Lower Hb concentrations were consistently associated with increased mortality without an obvious increase in specific cardiovascular events [14]. In addition, a prospective, observational study showed that reduced Hb was associated with a poorer quality of life in patients with heart failure [19].

Causes of anemia in patients with cancer are multifactorial and can be a direct result of cancer invasion, its treatment (radiation or chemotherapy), and/or CKD [20]. The cancer itself can cause or exacerbate anemia in a number of ways [21]. Anemia from a direct invasion of cancer is due to malignancy invading normal tissues causing blood loss, marrow infiltration inhibiting erythropoiesis, and/or inflammation leading to FID. Myelosuppressive chemotherapy either alone or in combination with radiation commonly causes anemia, which is referred to as CIA. Renal injury can occur from tumor invasion, chemotherapy, and/or age-related decline among cancer patients [20].

Anemia adversely impacts on tumor behavior. The transcription factor, hypoxia-inducible factor (HIF)-1α, responds to hypoxia. Head and neck cancer patients with an increase in tumor HIF-1α survived shorter than patients with non-hypoxic tumor (P=0.008) [22]. A similar result was shown in patients with cervical cancer, for whom a higher mortality rate with hypoxic tumors was seen than with non-hypoxic tumors (P=0.0039) [2324].

The prevalence of anemia among patients with cancer at initial presentation ranges from 30 to 90% from a systematic review of the literature, depending on the type of cancer, the definition of anemia (<9 g/dL vs. <11 g/dL), disease stage, and whether patients have been treated. In this review, patients with anemia had poorer survival and local tumor control than non-anemic counterparts. Quality of life (QoL) appeared to be positively correlated with the Hb level [25]. The European Cancer Anaemia Survey showed that the prevalence of anemia (Hb <12 g/dL) in patients with different cancer types was 39% at enrolment and 67% becoming anemic at least once during the 6-month survey period. Low Hb levels correlated significantly with poor performance status in this survey [26]. Although the number of good-quality studies are small, overall results of these studies suggest that anemia or low Hb is associated with decreased survival or poor QoL [14].

A systematic review including 5 randomized-controlled trials and 13 observational studies concluded that studies consistently showed an association between reduced Hb and increased mortality [27]. Eight prospective cohort studies showed a significant relationship between different Hb concentrations and all-cause or cardiovascular mortality [14]. However, the nature of the relationship between Hb level and mortality in CKD patients remains unclear due to the heterogeneity of patient populations and insufficient power [27]. A trial that enrolled 416 patients with renal anemia (pre-dialysis, hemodialysis, and peritoneal dialysis patients) showed a trend, but not a statistically significant difference, in 52-week survival favoring the greater target Hb group [28]. Although survivals did not differ, the QoL improved in the dialysis patients when Hb was normalized. Of note, this trial enrolled hemodialysis patients without a recent history of CHF or ACS [28]. A trial which included 1,233 hemodialysis patients with CHF or ischemic heart disease did not show survival improvement at a target Hb level of 14 g/dL compared with a target of 10 g/dL [29]. The trial by Foley and colleagues showed improved QoL in the higher Hb group (Hb 10 g/dL vs. 13.5 g/dL) in hemodialysis patients with asymptomatic cardiomyopathy [30]. From the observational studies, there was a consistent trend of decreasing mortality with Hb levels increasing to 10 g/dL but no further significant effect of increasing the Hb level above 11 g/dL. A similar trend was shown in studies using 11 to 12 g/dL as a reference Hb level. Therefore, the relationship may be affected by the preexisting conditions and the evidence is insufficient for generalization of risks and benefits of normalization of Hb level [27].

Given insufficient data from medical patients, especially those with MI, the threshold of RBC transfusion to optimize oxygen delivery and reduce workload to maintain cardiac output remains controversial in clinical practice.

A prospective, parallel-group randomized, pilot trial examined two thresholds (hematocrit (Hct) <30% vs. Hct <24%) for RBC transfusion in 45 patients with MI. The restrictive strategy was associated with a 36% relative decrease in the number of units transfused per patient compared to a 46% decrease in the other group of patients receiving any transfusion. Although the trial was not powered to detect differences in clinical outcomes due to a small number of patients, patients with liberal strategy showed significantly higher adverse outcomes, largely related to heart failure [33]. However, a recent trial did not favor the restrictive transfusion strategy. A randomized controlled trial including 110 patients with ACS and stable angina undergoing a cardiac catheterization showed somewhat different result from the previous pilot trial. Patients with the restrictive strategy of the lower threshold (Hb <8 g/dL) showed higher all-cause mortality, cardiac mortality, and cardiac events than with the liberal strategy of the higher threshold (Hb <10 g/dL). Of note, left ventricular ejection fraction at baseline tended to be lower in patients who were randomized to the restrictive strategy [34].

Other than small prospective parallel randomized trials, a few retrospective observational cohort studies, yet with large samples, seem to provide a clue for the optimal threshold for ACS patients. These studies, albeit retrospective in nature, consistently demonstrated that RBC transfusion at a higher Hb or Hct level is associated with a higher risk of mortality [353637]. A large retrospective study of 78,974 elderly patients with MI showed RBC transfusion at lower Hct (<33%) was associated with lower 30-day mortality while RBC transfusion at Hct >36% adversely influenced on 30-day mortality [37]. The other large cohort study with the enrollment of 39,922 patients with ACS, transfusion was associated with a decreased risk of cardiovascular death when the baseline Hb was <12 g/dL (adjusted odds ratio [OR] 0.42, 95% CI 0.20–0.89) but not when Hb was ≥12 g/dL (adjusted OR 1.42, 95% CI 0.94–2.17) [36]. Data of 44,242 patients with non-ST-segment elevation ACS (NSTE ACS) in 400 hospitals in the United States showed that in patients with a nadir Hct ≤24%, transfusion seemed to lower mortality (Hct ≤24% adjusted OR 0.68, 95% CI 0.45–1.02). In the median range (Hct 24–27%), transfusion did not impact on mortality (adjusted OR 1.01, 95% CI 0.79–1.30). Although rare, those transfused with nadir Hct ≥27% had higher mortality (Hct of 27% to 30% adjusted OR 1.18, 95% CI 0.92–1.50 and Hct ≥30% adjusted OR 3.47, 95% CI 2.30–5.23) [35].

Two meta-analyses assessed the association between blood transfusion and clinical outcomes among patients with MI, in which more than 200,000 patients were included each [3839]. While Chatterjee and colleagues found that transfusion adversely affected all-cause mortality and infarction [38], Garfinkle and colleagues showed that transfusion had beneficial or neutral effects on mortality below 8 g/dL and harmful or neutral effects above 11 g/dL [39].

Given the lack of high-quality data from randomized trials, the American Association of Blood Banks (AABB) chose not to recommend for or against a liberal or restrictive transfusion threshold in patients with ACS [4041]. Patient blood management guideline by the National Blood Authority recommends that RBC transfusion not be advisable in ACS patients with a Hb concentration >10 g/dL. In patients with ACS and a Hb of <8 g/dL, RBC transfusion may be associated with reduced mortality and is likely to be appropriate from the National Blood Authority's guideline. For patients with ACS and a Hb 8 g/dL–10 g/dL, any decision to transfuse should be made with caution and based on careful consideration of the risks and benefits [14].

The impact of transfusion on treatment outcomes or mortality in patients with cancer remains controversial [42]. Most studies were retrospective analyses and evaluated the impact of RBC transfusions on the treatment outcome in patients with locoregional cancer. A favorable influence on locoregional recurrence or survival were observed after RBC transfusions in patients with head and neck cancer and cervical cancer who received surgery, radiotherapy, or chemoradiotherapy [23]. For example, in a study of 56 patients with unresectable esophageal cancer receiving chemoradiotherapy, blood transfusion was associated with an increase in overall survival (OS) (Hazard ratio [HR], 0.26) [43]. A retrospective analysis of 605 patients with cervical cancer treated with radiotherapy showed that patients whose average weekly nadir Hb level during radiotherapy maintained ≥12 g/dL had a significant reduction in both pelvic and distant recurrence. Data suggested that blood transfusion reduced the negative prognostic implication of low Hb [44]. However, a large retrospective cohort study including 504,208 hospitalizations of cancer patients raised a concern of RBC transfusion. Among 504,208 patients, 70,542 patients received at least 1 RBC transfusion. The study reported that the incidence of venous TE (VTE) and arterial TE (ATE) of the transfused patients were 7.2% and 5.2% which were significantly greater than rates of 3.8% and 3.1% of those without RBC transfusion, respectively. RBC transfusion was also associated with an increased risk of in-hospital mortality (OR 1.34). However, the increased TE events and mortality may reflect a bias of more severe anemia and/or more advanced cancer in patients who require transfusions [45]. Therefore, National Comprehensive Cancer Network (NCCN) recommends that a decision to give RBC transfusion should not be made on the basis of whether the Hb level of the patient has reached a certain threshold [42].

There would be a question as to whether the guidelines for patients with solid tumor could be applied to the patients with hematologic malignancies or undergoing hematopoietic stem cell transplantation (HSCT). Two randomized controlled trials of RBC transfusion for patients with hematologic malignancies were conducted to evaluate the threshold for RBC transfusion [4647]. Given the experimental evidence in animal model suggesting the bleeding time to be inversely correlated with the Hct, Webert et al. evaluated whether or not the higher target of Hb level (12 g/dL compared to 8 g/dL) could decrease the risk of bleeding in thrombocytopenic patients [4748]. The study enrolled 29 patients in control group (2 units of RBC transfusion for Hb <8 g/dL) and 31 patients in experimental group (2 units of RBC transfusion for Hb <12 g/dL) who received induction/reinduction for acute leukemia or HLA-matched myeloablative allogeneic HSCT. The proportion of patients with clinically significant bleeding was 75.9% and 71% for the control and experimental groups, respectively, while the experimental group received more RBC transfusions (transfusion/patient-day) than the control group (0.233 vs. 0.151, P=0.003) [47]. A randomized (2:1) study of 89 patients with acute leukemia was conducted to compare restrictive Hb trigger (7 g/dL) to higher Hb trigger (8 g/dL). While the restrictive trigger group was transfused on average 8 units/patient, the higher trigger group received 11.7 units/patient (P=0.0003). There was no significant difference in bleeding episodes or neutropenic fever between the two groups [46]. However, clinical experience has shown that the leukemia and HSCT patient population requires multiple transfusions over the treatment course. Therefore, further studies are warranted to evaluate if it would be possible to reduce transfusions safely in this clinical setting. The recommendations for RBC transfusion in medical patients are summarized in Table 1.

Use of ESAs has consistently been shown to reduce transfusions and increase the Hb level in patients with anemia that arises during or shortly after myelotoxic chemotherapy [52]. The practice guidelines are available from the American Society of Hematology/American Society of Clinical Oncology (ASH/ASCO) last updated in 2010, from the European Organisation for Research and Treatment of Cancer (EORTC) last updated in 2006, and from National Comprehensive Cancer Network (NCCN) [425253].

In an early clinical trial, ESA epoetin alpha clearly increased Hb levels in cancer patients who were receiving non-platinum chemotherapy compared to placebo [54]. An early Cochrane meta-analysis demonstrated that cancer patients treated with ESA had a lower Relative Risk (RR) of receiving a blood transfusion than untreated patients (RR=0.67, 95% CI=0.62 to 0.73). In particular, ESA-treated patients with baseline Hb levels lower than 10 g/dL were more likely to have a hematologic response than untreated patients (RR=3.60, 95% CI=3.07 to 4.23) [55]. However, the U.S. Food and Drug Administration (the U.S. FDA) issued a black box warning in 2007 for the rHuEPO, i.e., warnings of greater mortality, serious cardiovascular and TE events, and tumor progression, based on 4 clinical trials. Among those receiving anti-cancer treatments, a higher mortality rate was observed in patients receiving rHuEPO than in control patients (Erythropoietin to treat head and neck cancer patients with anaemia [ENHANCE study] and Epoetin alfa versus best standard of care in anemic patients with metastatic breast cancer [BEST study]), and in anemic cancer patients not receiving chemotherapy (EPO-CAN-20 non-small cell lung cancer [NSCLC] study and Amgen 103 study) [23]. A meta-analysis showed that there was a higher risk for TE events in patients receiving ESA than in the control group [56]. A recent meta-analysis of survival outcomes from 53 randomized controlled trials showed that ESAs increased on-study mortality (combined hazard ratio [HR] 1.17; 95% CI, 1.06–1.30) and worsened overall survival (combined HR 1.06; 95% CI, 1.00–1.12). Thirty-eight trials enrolled 10,441 patients receiving chemotherapy. The combined HR for on-study mortality was 1.10 (95% CI, 0.98–1.24) and 1.04 (95% CI, 0.97–1.11) for overall survival [57]. Therefore, decision-making to use ESAs for anemic patients with cancer should be individualized. The practice guidelines on the use of ESAs from ASH/ASCO and NCCN are summarized in Table 2.

Iron studies, including serum iron, total iron binding capacity (TIBC), and serum ferritin, should be performed prior to ESA treatment to rule out absolute iron deficiency (ID), which may respond to oral or IV iron monotherapy without ESA. Treatment for iron deficiency is guided by iron status, defined as absolute ID, FID, possible FID, or no ID in NCCN guidelines. For absolute ID (TSAT <20% and ferritin <30 ng/mL), either IV or oral iron products alone (without ESAs) are recommended although IV iron is preferred. If cancer patients have FID (ferritin between 30 ng/mL and 500 ng/mL and TSAT <50%), IV iron supplements with ESAs should be considered. However, IV iron monotherapy without ESAs also can reduce the number of RBC transfusions. Possible FID is defined by a TSAT level <50% and a ferritin level >500 but ≤800 ng/mL. There are insufficient data to support the routine use of IV iron in this setting [42].

Although anemia in patients with CKD is multifactorial in origin, it is primarily associated with relative EPO production deficiency as the glomerular filtration rate (GFR) declines [58]. The introduction of rHuEPO into clinical practice in the 1980s was a major breakthrough in the treatment of the anemia of renal disease. In the early years, rHuEPO was considered beneficial for long-term dialysis patients who were transfusion-dependent. Then, the use of rHuEPO was extended to non-dialysis patients with CKD, and Hb targets increased progressively. Numerous observational studies demonstrated an inverse association between Hb concentration up into the normal range and cardiovascular events, TE events, and death [59]. A new Cochrane Review by Palmer and colleagues evaluated the evidence for pros and cons of using ESAs for anemic adults with CKD from 56 eligible studies involving 15,596 patients [60]. Epoetin alfa (OR 0.18), epoetin beta (OR 0.09), darbepoetin alfa (OR 0.17), and methoxy polyethylene glycol-epoetin beta (OR 0.15) prevented blood transfusions compared to placebo. However, a study examining the risks and benefits of normalizing the Hct in hemodialysis patients with cardiac disease raised concern of the increase in mortality and MI among normal-Hct group (HR 1.3, 95% CI 0.9–1.9) [29]. A trial enrolling CKD patients with diabetes compared achieving Hb level 13 g/dL versus 9 g/dL using ESA. Although there was no difference in death or cardiovascular events in both groups, stroke occurred more often in the Hb 13 g/dL group (HR 1.92, 95% CI 1.38–2.68) [61]. Therefore, when initiating and maintaining ESA therapy, the potential benefits of reducing blood transfusion and anemia-related symptoms should be balanced against the risk of harm in individual patients such as a history of stroke or malignancy [59]. The evidence now suggests that Hb target levels between 10 and 11.5 g/dL should be the aim for patients with CKD, and certainly not levels >13 g/dL [6263]. The U.S. FDA has emphasized an increased risk for patients with a target Hb >11 g/dL. An appropriate target Hb for ESA therapy has not been defined in patients with CKD, but caution is recommended in patients with a Hb >10 g/dL [1459].

In patients with CKD, correction of iron deficiency with oral or IV iron supplementation can improve anemia to some degree. Iron supplementation is widely used in CKD patients to treat iron deficiency, to prevent its development in patients receiving ESA therapy, to raise Hb levels in the presence or absence of ESA therapy, and to reduce ESA doses [59]. A systematic review with a meta-analysis of intravenous (IV) versus oral iron supplementation in patients with CKD (stages III to V) was performed. Thirteen trials were included in this meta-analysis. Compared with oral iron, IV iron significantly increased the Hb level in dialysis patients (weighted mean difference, 0.83 g/dL, 95% CI 0.09–1.57). For patients with CKD either on dialysis or not, there was a small but significant difference in the Hb level favoring the IV iron group (weighted mean difference, 0.31 g/dL, 95% CI, 0.09–0.53). There was no difference in adverse events between those receiving IV and oral iron supplements [64]. A randomized controlled trial comparing IV iron sucrose to oral ferrous sulfate in non-dialysis patients with CKD stages 3 to 5 showed IV iron therapy to be more effective in increasing Hb ≥1 g/dL from baseline than oral iron therapy (44.3% vs. 28.0%, P=0.0344). No serious adverse drug events were seen in patients receiving IV iron sucrose as 200 mg IV over 2 to 5 minutes, although drug-related hypotension occurred in two females (weighing less than 65 kg) after 500 mg were given intravenously over 4 hours. Use of ESA was associated with a higher proportion of patients achieving the primary endpoint (an increase in Hb ≥1 g/dL) in both the IV iron group (EAS use vs. non-use: 53.1% vs 38.3%) and the oral iron group (32.2% vs. 25.5%) [65].

However, the assessment of iron status is difficult. The diagnostic utility of serum ferritin and TSAT test is limited in estimating body iron stores. Even the examination of bone marrow iron stores does not predict erythropoietic responsiveness to iron supplements in patients with CKD with a high degree of accuracy. Kidney Disease: Improving Global Outcomes (KDIGO) Anemia Work Group recommended iron supplements if TSAT is ≤30% and ferritin is ≤500 ng/mL when an increase in Hb concentration without starting ESA is desired or a decrease in ESA dose is desired although the evidence to support a recommendation for specific TSAT and ferritin level is limited. Intravenous iron supplements are preferred, but a 1–3 months trial of oral iron therapy can be an alternative in nondialysis patients with CKD [59].