INTRODUCTION — Radiation therapy (RT) plays a major role in the management of head and neck squamous cell carcinomas. Despite therapeutic and technological advances, some patients will have persistence of irradiated tumor or develop locoregional failure, resulting in significant morbidity and mortality [1]. Radioresistance is a broad term that describes the relative resistance of individual cells, tissues, organs, or entire organisms to the biologic effects of RT [2].
Mechanisms of radioresistance to RT in head and neck cancer and strategies used to overcome this resistance are discussed here. Concurrent chemoradiation is discussed in detail separately. (See "Locally advanced squamous cell carcinoma of the head and neck: Approaches combining chemotherapy and radiation therapy" and "Definitive radiation therapy for head and neck cancer: Dose and fractionation considerations".)
MECHANISMS OF RADIORESISTANCE — Many factors affect the responsiveness of tumors to radiation therapy (RT). Individual patients with tumors of similar size and stage can respond very differently to RT.
Relevant factors are related to the primary tumor (eg, volume, size, grade, human papillomavirus [HPV] status), the patient (eg, hemoglobin levels, smoking status), and biologic factors (eg, hypoxia, expression of DNA repair genes, alterations of many other genes, and proliferation status).
Clinical factors affecting radiation response
Primary tumor characteristics — The complexity and variability of clinical outcomes in head and neck cancer are reflected in the spectrum of T stage descriptors used in the American Joint Committee on Cancer (AJCC)/Union for International Cancer Control (UICC) tumor, node, metastasis (TNM) staging system for the different primary sites [3]. Larger tumors and/or those with more extensive local invasion have higher T classifications, corresponding to a likely higher malignant cell burden and poorer prognosis. Larger tumors are also more likely to harbor biological factors associated with treatment resistance. (See "Overview of the diagnosis and staging of head and neck cancer".)
Radiation-related cell killing is a random event; the potentially lethal DNA double-strand breaks induced by irradiation are distributed randomly throughout the radiation-sensitive cells and will not kill all cells. As tumor burden increases, the probability of lethally damaging all of the cancer cells decreases, even if the total dose is escalated to high levels.
The location of the primary tumor site may influence response to therapy. As an example, oral cavity tumors demonstrate poorer response to RT than laryngeal tumors [4]. Tumor grade may also play a role. Well-differentiated tumors, which retain the ability to accelerate repopulation during RT, are characteristically more radioresistant than their moderately and poorly differentiated counterparts [5]. (See 'Repopulation' below.)
Inherent tumor cell radioresistance — Differences in intrinsic tumor cell radiosensitivity are seen experimentally and may contribute significantly to radiation failure [6,7]. Head and neck cancers associated with the HPV have an improved prognosis relative to their HPV-negative counterparts, most likely due to their enhanced sensitivity to chemotherapy and radiation. (See "Epidemiology, staging, and clinical presentation of human papillomavirus associated head and neck cancer".)
Patient factors
Hemoglobin — Lower pretreatment hemoglobin levels (<14.5 g/dL for males and <13 g/dL for females) are correlated with reduced local control and survival in patients with head and neck cancer [8,9]. However, the relationship between hemoglobin concentration, tumor oxygenation status, and radioresistance is complex. Most anemic patients have tumors that are hypoxic, but the absence of anemia does not ensure that a tumor will be well oxygenated.
Smoking status — Patients with head and neck cancer who continue to smoke during RT have lower rates of response and survival than patients who do not smoke during RT [10].
Biological factors determining radiation response
Hypoxia — The high-energy photons used in RT exert their biologic effects via secondary electrons generated in tissue through a process known as the Compton effect [11]. These secondary electrons induce DNA damage through both direct and indirect actions. With direct action, the electrons interact with and ionize atoms of the DNA helix. With indirect action, the electrons first ionize molecules around the DNA, such as water, producing highly reactive free radical species that then interact with and damage DNA. Unrepaired DNA damage leads to inhibition of cell proliferation, fatal chromosomal aberrations, and cell death. For photon radiation, the indirect mechanism predominates [2].
The presence of oxygen enhances the indirect process by prolonging the life span of the highly reactive free radicals generated by indirect ionization. Additionally, oxygen decreases the ability of cells to repair sublethal DNA damage, reacting with the broken ends of DNA to create stable organic peroxides that are not as readily repaired [12]. Well-oxygenated cells are two and one-half to three times more radiosensitive than hypoxic cells (oxygen partial pressure [pO2] <10 mmHg), a relationship described by the oxygen enhancement ratio (OER): the ratio of RT doses needed in hypoxic versus aerobic conditions to achieve the same biologic effect.
Intratumor hypoxia can be chronic, arising as the tumor outgrows the limits of its blood supply, or acute, when there is a disruption in microregional blood flow [13]. Tumor hypoxia in head and neck cancer correlates with radioresistance and is an adverse prognostic factor [14,15]. In one report, patients with a median tumor pO2 >10 mmHg, compared with those with tumor pO2 levels <10 mmHg, had significantly better two-year local control (82 versus 41 percent), disease-free survival (70 versus 31 percent), and overall survival (80 versus 46 percent) [14].
Other biologic mechanisms also modulate the link between hypoxia and radiosensitivity [12]. Many signaling pathways are impacted by hypoxia, including those involved in angiogenesis, glucose transport, pH regulation, and erythropoiesis [16].
As an example, the hypoxia-inducible factor (HIF) family of transcription factors plays an important role in the cellular response to oxygen homeostasis in vivo. HIF-1 alpha overexpression is seen in over one-half of all head and neck cancers and correlates with poor response to RT [17,18]. Overexpression of the HIF-1-regulated hypoxia-related protein lysyl oxidase is associated with increased metastases, progression, and death in patients with head and neck cancer [19,20]. However, there is also evidence that HIF-1 inhibition may both suppress and enhance the effects of RT [21].
Repopulation — Repopulation is defined as regrowth of tumor cells after the initiation of RT [22]. Accelerated tumor cell repopulation is an important factor contributing to the radioresistance of head and neck cancer [23]. Potential mechanisms include accelerated stem cell division, loss in asymmetrical division, and cell recruitment [24].
Accelerated stem cell division describes the shortening of the cell cycle after the start of RT. Loss in asymmetry refers to the shift in stem cells from asymmetric division, which produces one stem cell and one progenitor cell with limited self-renewal potential, to symmetric division, which gives rise to two identical stem cells. Lastly, quiescent cells in G0 phase may be recruited to re-enter more-active phases of the cell cycle.
Cyclins, cyclin inhibitors, Ki-67, and the epidermal growth factor receptor (EGFR) may all serve as markers of tumor proliferation [16]. Expression of Ki-67 and EGFR has been associated with radioresistance and has been studied to select patients for and predict response to regimens aimed to overcome repopulation, such as accelerated RT or anti-EGFR therapy [22,25,26].
Inherent tumor cell radioresistance — Differences in intrinsic tumor cell radiosensitivity are seen experimentally and may contribute significantly to radiation failure [6,7]. Head and neck cancers associated with HPV have an improved prognosis relative to HPV-negative tumors. There are some data to suggest that the better prognosis of these tumors may be attributable to their enhanced sensitivity to chemotherapy and radiation. (See "Epidemiology, staging, and clinical presentation of human papillomavirus associated head and neck cancer".)
Cancer stem cells — The potential role of cancer stem cells in radioresistance has been demonstrated in an increasing number of solid cancers [27-30]. If not eliminated by RT, stem cells have the capacity to self-renew and differentiate into the heterogeneous cells that constitute a tumor mass [31,32]. Experimental evidence suggests that a higher proportion of cancer stem cells within a tumor correlates with higher radioresistance [33]. Furthermore, cancer stem cells have been shown to possess inherent resistance mechanisms, such as amplified checkpoint activation and DNA damage repair [34]. Extrinsic factors, such as hypoxia, may also impact the radioresistance of cancer stem cells [35].
Other factors — Radioresistance may also occur as a consequence of other biologic factors: amplification of DNA repair genes, increased cellular production of free radical scavengers (eg, glutathione), activation of certain protooncogenes, and stromal interactions [7].
As an example, the excision repair cross-complementation group 1 protein (ERCC1) is a drug-resistance-related DNA repair protein that prevents apoptosis of cancer cells by removing platinum-DNA adducts generated by cisplatin. ERCC1 may also play a role in the repair of radiation-induced DNA damage. High expression of ERCC1 in nasopharyngeal cancer patients has been associated with poor survival after concurrent chemoradiation [36]. In addition, cells with a mutated p53 tumor suppressor gene fail to arrest at the G1/S cell cycle checkpoint following irradiation and are less likely to undergo apoptosis [37]. Similarly, overexpression of survivin, an oncogene of the inhibitor of apoptosis (IAP) family, may contribute to treatment resistance via negative regulation of programmed cell death and enhanced DNA repair [38].
Radioresistance involves a complex interplay between biologic processes. Investigators are applying genomics, proteomics, and gene expression profiling techniques to identify new targets and novel interactions.
STRATEGIES TO OVERCOME RADIORESISTANCE — Multiple strategies have been developed and tested in randomized clinical trials to improve the clinical results obtained with conventional radiation therapy (RT) given alone.
Altered fractionation RT — RT regimens using accelerated fractionation schedules can shorten the overall treatment duration, thereby attempting to minimize tumor repopulation as a cause of treatment failure. Pure hyperfractionation regimens, by treating twice daily with smaller fraction sizes, permit a higher total radiation dose without increasing late toxicity. Late effects on normal tissues are influenced by the individual dose fraction size. Altered fractionation strategies are discussed separately.
The clinical application of altered fractionation schedules is discussed separately, as are the applications of these approaches with concurrent chemotherapy. (See "Definitive radiation therapy for head and neck cancer: Dose and fractionation considerations", section on 'Accelerated fractionation RT' and "Definitive radiation therapy for head and neck cancer: Dose and fractionation considerations", section on 'Hyperfractionation' and "Locally advanced squamous cell carcinoma of the head and neck: Approaches combining chemotherapy and radiation therapy", section on 'Radiation therapy treatment plan'.)
Concurrent systemic therapy
Chemotherapy — Concurrent chemoradiation exploits the additive and synergistic effects that result when chemotherapy and RT are used together. Preclinical studies suggest that many chemotherapeutic agents contribute to improved outcomes through additive effects, via the independent cytotoxic action of the drugs themselves as well as through synergistic effects in which the drugs alter cell radiosensitivity.
Chemotherapy agents with radiosensitizing properties in vitro include cisplatin, fluorouracil, paclitaxel, docetaxel, mitomycin, hydroxyurea, bleomycin, gemcitabine, and etoposide. Many mechanisms have been proposed to explain how the addition of chemotherapy augments the effects of RT [39,40]. These include:
●Making DNA more susceptible to radiation damage and augmenting cell killing
●Interfering with cellular DNA repair after sublethal or potentially lethal damage (cisplatin and gemcitabine)
●Reducing tumor cell repopulation by enhancing the cytotoxic effect (hydroxyurea and fluorouracil)
●Reducing the population of hypoxic tumor cells (mitomycin)
●Exerting cytotoxic effects in the radioresistant S phase (camptothecins)
●Redistribution and accumulation of cells into the more radiosensitive G2 and M phases (paclitaxel and docetaxel)
●Promoting apoptosis
●Inhibiting angiogenesis
Concurrent chemoradiation improves locoregional control and survival compared with RT alone for head and neck cancer patients with unresectable disease and for those in whom a functional organ preservation strategy provides a nonsurgical alternative for cure while maintaining quality of life [1]. To date, concurrent chemoradiation remains the most effective strategy, as attempts to add various other agents to chemoradiation have not consistently resulted in improved outcomes compared with chemoradiation alone. (See "Locally advanced squamous cell carcinoma of the head and neck: Approaches combining chemotherapy and radiation therapy".)
Molecularly targeted agents — Molecularly targeted agents may potentiate the effects of radiation by targeting specific cell signaling pathways involved with radioresistance.
Cetuximab, a monoclonal antibody directed against the epidermal growth factor receptor (EGFR), inhibits tumor proliferation and has been shown to improve locoregional control and overall survival in combination with RT alone. However, in HPV associated oropharyngeal cancer, cetuximab plus chemoradiation has inferior survival outcomes compared with cisplatin-based chemoradiation. These data are discussed separately. (See "Locally advanced squamous cell carcinoma of the head and neck: Approaches combining chemotherapy and radiation therapy", section on 'Cetuximab plus radiation therapy'.)
Several other novel agents including some that are in preclinical or early clinical investigation target key components of various signaling pathways, including those for angiogenesis (vascular endothelial growth factor [VEGF]), DNA repair (poly adenosine diphosphate ribose polymerase [PARP]) and apoptosis, hypoxia (hypoxia-inducible factor 1 [HIF-1]), and proliferation (PI3-kinase/AKT) [41-44].
Checkpoint inhibitor immunotherapy — The programmed cell death receptor 1 (PD-1) inhibitor drugs nivolumab and pembrolizumab are approved for use in patients with unresectable or metastatic squamous head and neck cancer. (See "Treatment of metastatic and recurrent head and neck cancer", section on 'PD-1 inhibitor immunotherapy'.)
Uncontrolled data, especially in melanoma, have suggested a potential synergistic effect with immunotherapy given together with RT, both in the radiation-targeted tumor and at other sites. However, there are no randomized trial data supporting the combination of checkpoint inhibitor immunotherapy and RT in patients with squamous cell carcinoma of the head and neck. For example, a randomized phase III trial (JAVELIN Head and Neck 100) evaluating the addition of the programmed-death ligand 1 (PD-L1) inhibitor avelumab to standard-of-care definitive chemoradiation in this population was terminated because the combination was unlikely to achieve an improvement in progression-free survival [45,46].
OTHER APPROACHES — A number of other approaches have been studied clinically in an effort to overcome the radiation response observed clinically.
Targeting hypoxia — Hypoxia may be an important contributor to radiation resistance. Several different strategies have been aimed at reducing the effects of hypoxia on radiation resistance. None of these has achieved widespread clinical utilization. (See 'Hypoxia' above.)
Hypoxic cell sensitizers — The most extensively studied agents are the hypoxic cell sensitizers or oxygen mimetics, including misonidazole, nimorazole, and other related compounds. These have been studied in numerous randomized trials in patients with head and neck cancer, most extensively in Europe.
A systematic review of the literature and meta-analysis combined results from 18 trials that included almost 3400 patients [47]. The addition of these agents significantly improved locoregional control and decreased disease-specific deaths (odds ratios [ORs] 0.76, 95% CI 0.66-0.88, and 0.74, 95% CI 0.64-0.86, respectively), although the difference in overall survival was not significant (OR 0.87, 95% CI 0.75-1.02, p = 0.08).
Clinical interest in these drugs has been hampered by their marginal efficacy as well as dose-limiting toxicities (eg, delayed peripheral neuropathy with misonidazole).
Hypoxic cell cytotoxins — Tirapazamine is a hypoxic cell cytotoxic that selectively killed hypoxic cells in preclinical experiments and phase I to II studies, and potentiated the effects of radiation therapy (RT) and chemotherapy [48]. However, a large phase III trial failed to demonstrate a survival advantage for the addition of tirapazamine to RT and cisplatin [49]. A second phase III trial was closed early due to excess mortality when tirapazamine was combined with cisplatin plus RT [50].
Hyperbaric oxygen — Hyperbaric oxygen therapy increases the amount of oxygen in the blood and has been studied in a variety of tumors [51]. The amount of dissolved plasma oxygen is 0.3 mL/dL at 1.0 atmosphere; this is increased to 1.5 mL/dL upon administration of 100 percent oxygen and to 6 mL/dL with hyperbaric oxygen delivered at 3.0 atmosphere. Despite this theoretical advantage, studies using hyperbaric oxygen during RT for head and neck cancer have been disappointing. Because of the lack of clear benefit, the difficulty of completing combined RT and hyperbaric oxygen together, and other more effective alternatives, hyperbaric oxygen is generally not used clinically.
A systemic review and meta-analysis that included nine trials with 624 patients found that hyperbaric oxygen significantly improved locoregional control and decreased disease-specific deaths (ORs 0.46, 95% CI 0.33-0.64, and 0.58, 95% CI 0.42-0.81, respectively), although the difference in overall survival was not statistically significant (OR 0.73, 95% CI 0.51-1.05, p = 0.09) [47]. This approach is not used outside of clinical trials.
Functional imaging — Radiographic studies, such as dynamic contrast-enhanced magnetic resonance imaging (MRI) and hypoxia-specific positron emission tomography (PET) scans, can noninvasively and indirectly measure tumor hypoxia [52-54]. Work is ongoing to establish reproducible and relevant imaging correlations that can guide future clinical decision making. With their ability to capture and display regional differences in hypoxia within tumors, these imaging modalities provide potential physical targets for radiation dose escalation via highly conformal RT delivery technologies [55,56]. However, serial imaging has shown day-to-day variability in the spatial distribution of tumor hypoxia [57]. Moving targets can be missed and, thus, may be a source of local failure.
Correction of anemia — Correction of asymptomatic anemia, with either transfusion or the use of erythropoiesis-stimulating agents, has been postulated to decrease tumor hypoxia and reduce radioresistance, and thereby, improve clinical outcomes. However, there is no direct evidence to support these approaches.
Transfusion — Transfusion prior to and during treatment failed to improve outcome in patients with low hemoglobin levels in the Danish Head and Neck Cancer (DAHANCA) 5 study [58]. In this trial, patients were randomly assigned to the hypoxic radiosensitizer nimorazole or to placebo. In addition, 171 patients with low hemoglobin levels (females <13 g/dL and males <14.5 g/dL) were randomly assigned to transfusion or not. Although transfusion increased hemoglobin levels for most patients, receipt of transfusion did not improve locoregional control, disease-specific survival, or overall survival (hazard ratios [HRs] 0.99, 1.07, and 1.10, respectively). Correction of baseline anemia might improve efficacy of RT in the subset of patients with more-severe anemia, but there is limited evidence to support this approach.
Erythropoiesis-stimulating agents — Erythropoiesis-stimulating agents can reduce chemotherapy-induced anemia and the need for transfusions in head and neck cancer patients [59,60]. Retrospective data demonstrated that erythropoiesis-stimulating agents used prior to or during chemoradiation in oral cavity and oropharyngeal tumors corrected anemia and overcame the adverse prognostic effect of low pretreatment hemoglobin levels [61]. However, subsequent prospective trials have failed to confirm a clinical benefit from this approach. In fact, many trials were closed based upon interim analyses due to consistently worse outcomes in patients who received erythropoiesis-stimulating agents during therapy [62-66].
As an example, in one of the largest of these trials, 522 patients with squamous cell carcinoma in the head and neck were randomly assigned to weekly darbepoetin alfa or placebo weekly during RT in the DAHANCA-10 trial [66]. All patients had an initial hemoglobin value less than 14.0 g/dL, and treatment was continued until the hemoglobin value was 15.5 g/dL. The trial was terminated after a planned interim analysis because of the inferiority of the active treatment regimen. The five-year rate of locoregional failure was significantly increased with darbepoetin alpha (47 versus 34 percent, HR 1.53, 95% CI 1.16-2.02). Similar statistically significant results were seen for event-free survival, disease-specific death, and overall survival.
A Cochrane review examined the use of erythropoiesis-stimulating agents with RT or chemoradiation in five trials with a total of 1397 head and neck cancer patients [65]. Of note, the target hemoglobin concentration was higher than recommended in four of the five trials. Significantly worse overall survival was seen in the patients who received an erythropoiesis-stimulating agent (OR 0.73, 95% CI 0.58-0.91, p = 0.005).
Reasons for the adverse outcomes associated with the use of erythropoiesis-stimulating agents during RT are unclear. These trials attempted to correct hemoglobin concentration to supraphysiologic levels, and it has been suggested that an excess of thromboembolic events may have contributed to the worse outcomes [67]. Erythropoiesis-stimulating agents may also promote tumor progression, as erythropoietin receptors are expressed in head and neck cancer cell lines [68]. A follow-up analysis of samples from a phase III trial found that increased expression of the erythropoietin receptor was associated with a poorer local progression-free survival [69].
The issues surrounding the use of erythropoiesis-stimulating agents in cancer patients are discussed in detail separately. (See "Role of erythropoiesis-stimulating agents in the treatment of anemia in patients with cancer".)
Until 2017, the US Food and Drug Administration (FDA) required all erythropoiesis-stimulating agents to be prescribed and used under a risk evaluation and mitigation strategy (REMS) with clear understanding by treating clinicians and patients of the poorer survival, increased risk of treatment failure, and life-threatening side effects associated with their use [70]. While an REMS is no longer required, careful consideration should be given prior to starting these drugs in all patients. Transfusion remains an acceptable option to treat anemia in patients with head and neck cancer undergoing RT.
Radioprotectors — The concomitant administration of a systemic agent that protects normal tissues from the effects of irradiation may reduce treatment-related toxicity, potentially allowing for the use of higher RT doses to overcome radioresistance.
●Amifostine, a thiol-containing compound that may act as a free radical scavenger, has been shown to reduce the risk of xerostomia in head and neck cancer patients treated with RT alone; its role in combination with chemoradiation and the higher doses of radiation used in contemporary treatment regimens remains uncertain. (See "Management and prevention of complications during initial treatment of head and neck cancer", section on 'Amifostine'.)
●Palifermin, a recombinant keratinocyte growth factor, significantly decreased the incidence of severe mucositis in randomized trials in patients with head and neck cancer [71,72]. (See "Management and prevention of complications during initial treatment of head and neck cancer", section on 'Palifermin'.)
Neither palifermin nor amifostine has an established role in conjunction with the use of higher doses of radiation to try to overcome radioresistance and risk of locoregional failure in patients with head and neck cancer.
Hyperthermia — Hyperthermia involves the heating of tissues to supraphysiologic temperatures; its concurrent use with RT may enhance the tumor response to radiation via additive and synergistic mechanisms [73-75].
The independent cytotoxic effects of hyperthermia are complementary to RT. Cells in S phase, the most radioresistant part of the cell cycle, are the most sensitive to hyperthermia [76,77]. Moreover, unlike RT, hypoxic conditions may also make cells more sensitive to hyperthermia [78]. Synergistic radiosensitization by hyperthermia may occur via inhibition of DNA repair pathways [79,80]. Finally, at more moderate temperatures, hyperthermia can improve blood flow, leading to increased perfusion, reoxygenation, and decreased hypoxia [81].
Two randomized trials in locally advanced head and neck squamous cell carcinoma patients showed improved tumor control with the use of hyperthermia concurrent with RT [82,83]. However, these trials were conducted in small numbers of patients using older techniques and dosing regimens, and their clinical relevance is uncertain.
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Head and neck cancer".)
SUMMARY AND RECOMMENDATIONS
●Mechanisms of radioresistance – Multiple clinical and biologic factors contribute to the development of radioresistance, which remains a significant obstacle to optimal locoregional disease control in patients with squamous cell carcinomas of the head and neck. Large tumors, continued smoking, and pretreatment anemia are clinical factors that have been inversely correlated with locoregional control. Tumor hypoxia, repopulation, cancer stem cells, and inherent differences in tumor cell radiosensitivity are biologic factors that have been identified as strongly correlating with radioresistance. In particular, human papilloma virus (HPV) status has emerged as a significant prognostic factor and predictor of response to radiation therapy (RT) in head and neck cancer. (See 'Mechanisms of radioresistance' above.)
●Rationale for concurrent chemoradiation – Concurrent chemoradiation is the standard of care to decrease radioresistance and improve outcomes in patients who are treated with RT for head and neck cancer. Concurrent chemoradiation improves locoregional control and survival compared with RT alone for head and neck cancer patients with unresectable disease and for those in whom a functional organ preservation strategy provides a nonsurgical alternative for cure while maintaining quality of life. (See "Locally advanced squamous cell carcinoma of the head and neck: Approaches combining chemotherapy and radiation therapy".)
•Strategies to further intensify treatment beyond concurrent chemotherapy with conventionally fractionated RT have not demonstrated improved outcomes in unselected patient populations. Recognition of the influence of HPV status in treatment response has led to the design of clinical trials that separately investigate HPV-positive and HPV-negative patient populations. (See 'Strategies to overcome radioresistance' above.)
●Targeting hypoxia – Various approaches to target hypoxia have been evaluated in randomized clinical trials. Although these techniques may improve locoregional control, their clinical impact is less than that of concurrent chemoradiation, and they do not have a major role in the management of head and neck cancer patients. (See 'Targeting hypoxia' above.)
●No role for erythropoietic agents during RT – We recommend against using erythropoietic agents to correct asymptomatic anemia during RT for head and neck cancer (Grade 2B). Although the correction of anemia has a strong theoretical rationale, data do not support the routine use of transfusions or erythropoiesis-stimulating agents in patients with head and neck cancer, and these approaches may actually worsen outcomes. (See 'Erythropoiesis-stimulating agents' above.)
