Patients with cancer have high rates of comorbidities, but hard data are often lacking about the interaction of specific regimens with such health conditions. Experts say more inclusive clinical trials and broader use of risk assessment tools would help improve care.
Hyman Muss, MD
Although patients with cancer have high rates of comorbidities that often correlate with poorer outcomes and complicate treatment decisions, hard data are frequently lacking on the interaction between specific regimens and health conditions.
Oncology experts are increasingly turning their attention to the disconnect between patient populations recruited for the clinical trials that lead to the approval of new drugs and individuals in the real world who are treated with those medicines. A push for better metrics to assess and manage comorbidities is also under way.
The absence of patients with comorbidities from clinical trials is generally attributed to trial sponsors’ concerns about jeopardizing the approval of agents that would prove efficacious in healthier patients as well as about the safety of trial participants.
However, new research findings show that the processes that steer these patients away from drug studies go beyond specific eligibility criteria to a broader cultural failure to integrate people with comorbidities—who represent a majority of patients with cancer—into trials.
Patients with comorbidities were less likely than those without a comorbidity to discuss clinical trials with their doctors or be offered a spot in a trial, with the impact more pronounced as the number of secondary conditions rose, according to self-reported patient data presented at last month’s American Society of Clinical Oncology (ASCO) conference.1 Hypertension, prior cancer, and hearing loss had the greatest impact, even though good hearing is rarely, if ever, a listed eligibility requirement. Patients with at least 1 of those 3 comorbidities were 11% less likely to discuss clinical trials, 18% less likely to be recruited, and 22% less likely to participate.
The study’s findings illustrate the challenges that are faced in efforts to improve access to new drugs for older people and those with comorbidities and to produce more trial data to guide treatment decisions in those populations. ASCO and Friends of Cancer Research have proposed a set of broader eligibility criteria that would boost the number of patients with comorbidities in trials, but “this effort alone is unlikely to substantially increase trial participation rates overall,” the study concluded. Advocates argue that the culture around trial enrollment is beginning to change, but they say it could take years to make a significant impact.
“Clinical trial populations are always healthier and richer than the people in the community,” said Hyman B. Muss, MD, a specialist in geriatric oncology at the Lineberger Comprehensive Cancer Center at the University of North Carolina at Chapel Hill and a Giants of Cancer Care® award winner.
“A lot of the data we get on the Olympic athletes who are going on trials [are] really not applicable to someone who is 75, with bad arthritis, who’s using a cane,” he said in an interview with OncologyLive®. “And yet we treat them the same way, and then they frequently will have horrendous toxicity because those patients are not included in the trials. We’ve been trying to fix it, but it’s been woefully slow.”Studies have found a wide range of comorbidity rates.2 A large study of patients aged 66 years and older with cancer found that 40% had at least 1 of 15 major comorbid conditions, compared with 32% of people without cancer (Figure).3 Twenty-five percent of the patients with cancer had 1 condition, and 15% had 2 or more. Diabetes was most common at 16.0%, followed by chronic obstructive pulmonary disease (COPD) at 15.5%, congestive heart failure at 9.7%, and cerebrovascular disease at 6.0%. Comorbidity rates were about 53% in patients with lung cancer, 41% with colorectal cancer, 32% with breast cancer, and 31% with prostate cancer.Other studies have found greater prevalence of comorbidities. In the study presented at ASCO, 66% of patients with cancer said they had at least 1 comorbidity, led by hypertension at 35%.1 Another study found that 92% of patients had a comorbidity, with an average of 2.7 conditions per person.4 In people both with and without cancer, multiple chronic conditions are more common in older patients, racial minority groups, and patients who are in poverty.5,6
Certain common comorbidities preclude standard treatment approaches. In lung cancer, in which the majority of patients have other serious conditions, those with emphysema and other lung impairments often cannot undergo surgery, said Edward S. Kim, MD, chair of the Department of Solid Tumor Oncology at Levine Cancer Institute, Carolinas HealthCare System, in Charlotte. “Especially if you’re oxygen dependent and that has really reduced your overall activity, that’s tough to overcome. Some of them can’t even receive radiation, and that’s what we use palliatively if they can’t have surgery,” he said.
Some drugs, including doxorubicin (Adriamycin) and trastuzumab (Herceptin), are cardiotoxic and pose greater risk to patients with cancer who have cardiac dysfunction, he noted. Trastuzumab, for example, doubles the risk of cardiomyopathy or heart failure in patients aged 66 to 75 years with breast cancer, and patients with comorbidities have higher rates of discontinuation of therapy.7,8 For some patients, medical management can improve heart function enough to allow use of these drugs but not in all cases and not always quickly enough, Kim said.
Other significant conditions include kidney and liver diseases, which may rule out some agents over concerns that the patient cannot metabolize them normally, he said. Diabetes has a number of impacts; a patient with chronic complications from diabetes faces a doubled risk of developing grade 2-4 peripheral neuropathy after receiving taxane-containing regimens.9 The odds ratio for diabetes with or without chronic complications is 1.67, or about two-thirds greater odds.
“In many instances, it ends up being a dose-limiting toxicity, which is to say patients are so fed up with their neuropathies that they could end up choosing not to continue their chemotherapy, which is critical to their outcomes,” said Joseph Unger, PhD, MS, of the Cancer Prevention Program at Fred Hutchinson Cancer Research Center, University of Washington, Seattle, who was an author on the neuropathy study.
Comorbidities are generally associated with poorer survival. In breast cancer, patterns of survival show that patients with early-stage cancers and comorbidities such as myocardial infarction, peripheral vascular disease, stroke, COPD, and diabetes have outcomes typical of those with malignancies that were 1 stage higher but no comorbidities.10 Patients with more severe comorbidities such as dementia, chronic renal failure, and liver disease had outcomes that shifted 2 stages higher.
For patients with any comorbidity, mortality rates are 1.1 to 5.8 times higher in the years after breast cancer treatment compared with patients without comorbidities, 1.2 to 4.8 times greater for colon cancer, and 1.1 to 1.5 times higher for lung cancer.2
“In certain disease settings, for example, more patients end up dying of noncancer causes than cancer causes. In many instances, it’s because they’re dying of conditions related to cardiovascular risk and they’re experiencing cardiovascular events themselves,” Unger said. “These risks may be overlooked because cancer is the proximal worry. What this research shows is that there are other proximal worries out there, as well, for these cancer patients. They’re not necessarily secondary or as secondary as we might have thought.”Some investigators have attempted to construct mathematical relationships between the number and type of comorbidities and patient outcomes, but Muss said understanding how to treat patients with multiple illnesses requires more granular data, especially in older populations.
“You can have a patient who lists 6 illnesses: They’re diabetic, they’ve got arthritis, and so on. But none of them are very bad, they’re well controlled, and they look great. And then you can have a patient with 1 comorbidity, like atherosclerosis, and they could have 6 stents and some mild heart failure, and that patient can’t shop or walk up a flight of stairs easily. If you were just counting the comorbidities, you would be misled,” Muss said. “The number of comorbidities is not proportional to your functional status.”
Treating such patients requires evaluating and balancing a number of issues in addition to their cancer status, such as the severity of the other conditions and whether to avoid certain standard-of-care treatments or reduce doses. These determinations are made every day, but evidence supporting one choice over another can be spotty, especially for new therapies, because patients with comorbidities rarely participate in clinical trials. Online tools that provide guidance on chemotherapy toxicity are available, but oncologists tend not to use them and depend on their personal judgment instead, Muss said.
“They’re doing it by the eyeball test, and [that is] notoriously inaccurate. They’ll look at a patient and say, ‘Oh, give Mary half the dose, or maybe we shouldn’t treat her,’” he said. “What we’re trying to do in our field is to get quantitative metrics other than ‘She just doesn’t look so good’ to really make much better judgments.”Low trial participation rates are a problem in oncology broadly but are particularly acute for certain subgroups, including older patients11 and people with comorbidities. As Muss’ experience indicates, relevant conditions include both those that are always classified as comorbidities, like COPD, and others such as obesity, mental health issues, and hearing problems that often are not.
“Hearing loss is not typically a formal eligibility [criterion] to be on a trial, but if you think about it conceptually, if a person is having a hard time hearing and communicating, it’s hard for the physician to have a conversation about a trial, and it may be just less likely to come up. That’s a good example of a nonformal way that a comorbidity can get in the way of trial participation,” said Mark Fleury, PhD, of the American Cancer Society Cancer Action Network, who conducted the study of self-reported patient data with Unger and other investigators.
The recommendations issued last fall by ASCO and Friends of Cancer Research urge for greater inclusion of children and of patients with HIV, brain metastases, prior or current malignancy, or renal, liver or cardiac disease.12 Some criteria are still needed to safeguard patient safety, but others have not made sense for decades, Fleury said.
“Many eligibility criteria have just been copied and pasted from previous trials, and sometimes not a lot of reflection has gone into whether or not those are still appropriate,” he said. “If you think back to 25 years ago, being HIV positive could very well be a death sentence because of where we were with drug development at that time. You wouldn’t want to put a [patient with] HIV on a cancer trial, because they may have health complications. Today HIV is extremely well managed and people who are HIV positive can live a normal life span and don’t have any other health issues, but that has persisted over the years.”
Kim, the lead author on the eligibility criteria recommendations, said the proposal to include some patients with organ dysfunction has drawn objections, but he argued that in many cases, minor liver or kidney impairment should not bar participation. He gave the example of pemetrexed (Alimta), which has a label warning against its use in patients whose creatinine clearance (CrCl) level is below 45 mL/min, but is routinely used in those with a CrCl of 40 mL/min.
“If it was restricted to patients with a clearance of 45, there wouldn’t be many people getting it. Ironically, it’s one of the safest chemotherapies you can give. We give it to our more poorperformance-status patients because it’s so easy to give,” he said. “That’s a real-world example of where nobody actually pays attention to that restriction, at a safe range. It does have some renal toxicity, but physicians aren’t excluding people from the drug, even though that’s what the clinical trials do.”
The FDA supports broadening the criteria, and the pharmaceutical companies that sponsor most trials have expressed cautious interest, Fleury and Kim said. The companies remain concerned about endangering patients and jeopardizing drug approvals, but advocates say that options such as separate trial cohorts for patients with comorbidities, with sequestered results, can help make the pharmaceutical industry more comfortable.
“Part of it is creating those trial structures and protocols that can provide some sense of ease and safety for both the patients and the sponsors,” Fleury said. “It’s a little bit of a culture change as well. When you’re talking about a drug development program of hundreds of millions of dollars, it’s understandable that the sponsor might be risk averse in terms of pushing the limits and going with a new population.”
Muss said sponsors could also support expanded trial cohorts after a drug is approved, allowing investigators to accrue patients who are older and have comorbidities. He said new drugs are already given to such patients when they lack other treatment options, but the results are not systemically collected. In some cases, cooperative groups sponsored by the National Cancer Institute are able to conduct these trials, and retrospective real-world studies can be done using data collected by cancer centers, Muss said. Kim said that ASCO’s ongoing multiarm TAPUR study, which has embraced expanded eligibility criteria where possible, shows that patients with comorbidities can be safely included.Muss also urged for much greater use of geriatric assessments that ask patients about their mobility, activity level, social support, and other factors that he says are more practically useful for prognosticating outcomes than straight morbidity counts are. Online tools include the Cancer and Aging Research Group’s calculator for estimating chemotoxicity, at mycarg.org/ SelectQuestionnaire, and the Chemotherapy Risk Assessment Scale for High-Age Patients, or CRASH.13 Kim acknowledged that geriatric assessments are widely available but said they are rarely used by the majority of oncologists who do not specialize in geriatrics because of the amount of time it takes to administer them.
Unger said more data on outcomes in patients with comorbidities are gradually becoming available, thanks in part to expanding trial eligibility criteria but also to initiatives like ASCO’s CancerLinQ, which is collecting real-world clinical data from thousands of oncologists, and to creative uses of other data sources. He noted that the neuropathy and cardiovascular risk studies were made possible by linking Medicare claims data with statistics from the National Cancer Institute—funded SWOG clinical trials program. SWOG is now preparing another study to systematically characterize the phenotype of neuropathy, he said.
“I like to imagine the day when based on this study that we’re developing, a risk calculator for neuropathy will allow clinicians and patients to feed in the most relevant conditions that predict developing neuropathy and figure out what the patient’s risk of neuropathy is and which could help guide treatment,” Unger said. “It’s not only about guiding the immediate treatment choice but anticipating what could happen to the patient down the line.”