December 27, 2017
16 min read

Toxicities associated with immunotherapy warrant ‘very high level of suspicion’

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact

Since 2011, the FDA has approved six immune checkpoint inhibitors designed to reverse the course of multiple cancers.

Those approvals — along with oncologists’ willingness to combine immunotherapies — has led to more durable responses among patients with cancer.

However, many oncologists say the toxicities associated with immunotherapies, especially those used in combination, pose a greater risk to patients than often reported in clinical trials funded by the pharmaceutical companies that develop and distribute these drugs.

If adverse events warrant an ED visit, a patient should consult with his or her oncologist before being incorrectly treated by their primary care physician or internist, according to Igor Puzanov, MD, MSCI, FACP.
If adverse events warrant an ED visit, a patient should consult with his or her oncologist before being incorrectly treated by their primary care physician or internist, according to Igor Puzanov, MD, MSCI, FACP. “We need to spread the knowledge across the oncology spectrum,” he said. “If a clinician doesn’t give steroids for pneumonitis because they think it’s pneumonia, the patient dies.”

Source: Courtesy of Roswell Park Cancer Institute

“A lot of physicians grew up in the era of chemotherapy and targeted therapy, and now immunotherapy comes along and everyone is excited,” Igor Puzanov, MD, MSCI, FACP, director of early phase clinical trials and chief of melanoma at Roswell Park Cancer Institute, told HemOnc Today. “Clinicians want to provide these immunotherapies to their patients but, at the same time, they need to get the help to manage the new toxicities. These drugs are good and they are saving lives, so we must figure it out.”

Toxicities from combination immunotherapy can range from skin rashes, mucositis and diarrhea to colitis, sepsis, hypothyroidism, hyperthyroidism, pneumonitis, myocarditis, arrhythmia, type 1 diabetes and hypophysitis. These can occur within weeks after initiation.

“These side effects can happen and, when they happen, they can be severe and lethal,” Ryan J. Sullivan, MD, assistant professor of medicine at Harvard Medical School and assistant professor of hematology/oncology at Massachusetts General Hospital, told HemOnc Today. “Even when we’re using single-agent PD-1 inhibitors, we need to be vigilant at diagnosing immune-related toxicities and managing them early as opposed to waiting to begin actively pursuing a diagnosis and therapy. The earlier you identify the toxicity, the better the management.”

HemOnc Today spoke with oncologists and researchers about the toxicities associated with immunotherapies; ways to monitor, treat and re-challenge patients affected by them; and whether reduced dosing may help limit adverse events without diminishing the efficacy of immune checkpoint inhibitors.

‘Enormous optimism’

Checkpoint inhibitors targeting cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) and the programmed death-1 receptor (PD-1) and its ligand PD-L1 have become standards of care for an increasing number of indications, including metastatic melanoma, non-small cell lung cancer, renal cell carcinoma, Hodgkin lymphoma, urothelial cell carcinoma and squamous cell carcinoma of the head and neck.


FDA-approved checkpoint inhibitors include ipilimumab (Yervoy, Bristol-Myers Squibb), nivolumab (Opdivo, Bristol-Myers Squibb), pembrolizumab (Keytruda, Merck), atezolizumab (Tecentriq, Genentech), avelumab (Bavencio; EMD Serono, Pfizer) and durvalumab (Imfinzi, AstraZeneca).

“There is enormous optimism surrounding immunotherapy,” Diane Blum, MSW, former CEO of Lymphoma Research Foundation and former executive director of CancerCare Inc., told HemOnc Today. “Diseases that seemed so untreatable are now treatable.

“In the past, patients with metastatic melanoma had no options,” said Blum, who is now senior vice president of the National Executive Service Corps. “These new treatments aren’t just improving outcomes, they’re turning patients around. They have given us miraculous stories. Just look at [former President] Jimmy Carter. That was a miracle, and I think oncologists are inspired by stories like that.”

Diagnosed in 2015 with metastatic melanoma, which began in his liver and spread to his brain, Carter, now 93, has been in remission for more than 2 years. Carter received radiation, surgery and pembrolizumab, which targets PD-1.

Combinations of immunotherapies also have shown great promise.

In a double-blind randomized study published in The New England Journal of Medicine that evaluated the addition of nivolumab to ipilimumab, the objective response rate for the combination therapy far exceeded that of ipilimumab alone (61% vs. 40%) among 109 patients with advanced melanoma who had BRAF wild-type tumors.

Still, these agents, particularly when combined, aren’t without their risks.

In a separate study, also published in The New England Journal of Medicine, Johnson and colleagues reported that 40% of patients treated with advanced melanoma discontinued combination immunotherapy because of adverse events. The study — which used data from Bristol-Myers Squibb’s corporate safety databases — also showed only 18 of 20,594 patients (0.09%) receiving nivolumab alone or with ipilimumab experienced myocarditis — potentially fatal inflammation of the heart muscle. However, incidence of myocarditis was higher among those assigned the combination than nivolumab alone (0.27% vs. 0.06%).

“But, I can tell you, in my practice, I had three patients last year who developed myocarditis out of about 100 treated,” said Puzanov, who also was an author on that study. “That’s still only 3%, but it’s well over 0.27%. If you are one of those three patients, it’s 100% for you.”

Several other combinations of checkpoint inhibitors are under investigation.

A phase 2 study presented at this year’s American Association for Cancer Research Annual Meeting showed pembrolizumab in combination with indoleamine 2,3-dioxygenase inhibitor indoximod (NLG-8189, NewLink Genetics) conferred a 33% ORR. In a phase 1b trial, ipilimumab plus Coxsackievirus A21 (CVA21, Cavatak) conferred a 50% ORR.


The risk-benefit of combining PD-1 and CTLA-4 blockades are expected to be further clarified by data from ongoing phase 3 double-blind trials, such as CheckMate 067.

“If any of those combinations, such as pembrolizumab and epacadostat (Incyte), end up with a positive phase 3 study, it could change the way we manage patients if these are less toxic than ipilimumab-nivolumab,” Sullivan said. “If it’s less toxic and has similar efficacy, that could be very useful.”

Reported vs. real-word rates

According to European Society for Medical Oncology’s 2017 clinical practice guidelines, the most frequent immune-related adverse events can affect the skin, colon, endocrine organs, liver and lungs, and they most often are treated with corticosteroids.

Diarrhea — the toxicity most often associated with anti-CTLA-4 agent ipilimumab — can lead to colitis, and moderate to severe dermatitis. Adverse events associated with ipilimumab occur among 60% to 85% of patients and are mostly grade 1 and grade 2 events, although up to 27% of patients develop grade 3 to grade 4 toxicities, according to the ESMO guidelines.

Fatigue is the most common adverse event associated with anti-PD-1/PD-L1 agents, occurring among 16% to 37% of patients treated with an anti-PD-1 agent and 12% to 24% of patients on an anti-PD-L1. Pneumonitis can occur in up to 10% of patients receiving nivolumab or pembrolizumab.

However, clinical trial-reported adverse events rates — used to form the ESMO guidelines — may be lower than those observed in real-world practice.

According to an independent brief report by Shoushtari and colleagues, published in JAMA Oncology, the combination of ipilimumab and nivolumab led to clinically significant immune-related adverse events among 91% of patients with advanced melanoma. However, if researchers had used the typical reporting methods of clinical trials, the rate of grade 3 to grade 4 adverse events would have been reported as 59%.

Additionally, 72% of patients required systemic steroids, 50% required ED visits and 36% required hospital admission.

Approximately 10% of patients who discontinued therapy for toxic effects developed new late-onset immune-related adverse events that would not have been captured in randomized trials and, thus, may be underreported.

“If oncologists are going to use these combinations, there are a lot more important toxicities than they might think just by reading the clinical trials,” Paul B. Chapman, MD, an oncologist specializing in melanoma at Memorial Sloan Kettering Cancer Center and co-author on that study, told HemOnc Today. “In fact, they’re almost inevitable.”

Paul B. Chapman, MD
Paul B. Chapman

Several oncologists with whom HemOnc Today spoke said they were not surprised by Chapman and colleagues’ finding that half of the patients on treatment required ED visits.


“These are experienced investigators and Paul knows what he’s doing with these agents, as does his group,” Bruce G. Redman, DO, professor of oncology at Michigan Medicine at University of Michigan, told HemOnc Today. “You don’t get those kind of detailed adverse events reported in the drug company studies. He broke it down to show real-world evidence that there are serious toxicities, at least with combinations.”

There also may be a fundamental problem with the toxicity grading system, Chapman said.

“One of the things you learn when you do this long enough is that the grading system we use is arbitrary and not very helpful,” Chapman said, referring to the Common Terminology Criteria for Adverse Events (CTCAE).

For example, Chapman said an immunotherapy-related mild, non-itch rash that covers most of a patient’s body is considered a grade 3 toxicity, whereas a very localized rash “that drives you crazy” is considered a grade 2.

The same inaccuracies occur with the grading system for diarrhea, Chapman added, which is why, in their study, Chapman and colleagues used a different system to grade toxicities.

“We decided to call a serious toxicity anything that required steroids,” he said. “If it needs to be treated, it’s a serious side effect. We found that 91% of patients in this cohort had a clinically significant adverse event that was grade 2 or worse or required steroids. That’s high.”

Chapman noted that grade 2 neuropathy interferes with daily living and would prompt most clinicians to stop immunotherapy.

“Our point is that grading doesn’t mean anything clinically important,” he said.

Recognizing toxicities

Although immunotherapies are far better tolerated than many chemotherapies, oncologists and patients must be keenly aware that even subtle adverse events could require immediate treatment, Sullivan said.

“Oncologists are really good at giving chemotherapy and understanding the timing of when neutropenia is going to happen or when a patient has diarrhea related to targeted therapy,” he said. “We’ve developed strategies to help eliminate or treat those toxicities, but it took a while for the optimization of management of nausea and neutropenia. The optimization management of immune checkpoint inhibitor toxicities is going to be similar.”

A patient with severe diarrhea can be diagnosed with colitis after undergoing a colonoscopy. But, an oncologist might not explore this option if the patient otherwise is well.

To help prevent improper treatment of patients who require hospitalization because of immune-related toxicities, Massachusetts General Hospital recently unveiled a unit with physicians specifically trained to treat immune-related toxicities.


“If you don’t think about it, you may miss something very serious. That’s the challenge,” Sullivan said. “These toxic side effects seem so innocuous — tiredness, headache — but they can represent severe endocrine toxicities. A mild cough may be early pneumonitis. It doesn’t mean you have to treat every one of these patients as if they have a severe immunotherapy complication, but having it at the forefront of the list of potential problems is critical.”

To monitor patients on ipilimumab and nivolumab, most oncologists measure for endocrinopathies by checking cortisol, adrenocorticotropic and thyroid-stimulating hormone levels on a regular basis. Nausea and vomiting could be the first signs of autoimmune hepatitis, which can be detected by high levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) liver enzymes.

According to Society for Immunotherapy of Cancer guidelines, short-term adverse events due to the use of moderate to high-dose corticosteroids — opportunistic infections, sleep disturbance, gastritis and hypertension — should be anticipated. Patients receiving long-term or high-dose corticosteroids are at risk for developing diabetes and osteoporosis and should receive vitamin D and calcium supplementation and, in some cases, antibiotic prophylaxis.

“I really harp on the cardiac toxicities, because they are probably the most fatal of all these conditions,” said Puzanov, who gives his patients weekly troponin tests. “Whether it’s 20%, 30% or 40%, cardiac toxicities are much higher than colitis, pneumonitis or any other ‘itises.’ We need to be aware.”

Even with diligent testing, Redman said many of his patients are hospitalized with adverse events.

“I’ve had patients say after their third dose of therapy that, outside of a little fatigue, they feel fine,” Redman said. “Forty-eight hours after starting the fourth cycle, they call me with diarrhea 8 to 10 times a day. When toxicities hit, they hit.”

According to Redman, few oncologists check for the incidence of diabetes in nondiabetic patients on immunotherapies, although they should.

“The patient may be feeling a little lousy, and their comprehensive metabolic panel comes back showing the blood glucose is 800 for somebody who has no history of diabetes,” he said. “Those occasions are rare, but you need to have a very high level of suspicion to be able to react. There may be something more going on. It could be the first manifestation of adrenal insufficiency.”

Communication with patients is key, Redman said.

“If a patient has serious diarrhea, they shouldn’t think it was something they ate that day,” Redman said. “If they get a cough or feel short of breath, they should call right then and there. They probably didn’t catch something from their wife or husband. It could be pneumonitis, which we need to jump on right away.


“I’ve had patients walk in 1 week after having a normal ALT and AST and have it come back over 1,000, and they only felt tired,” he added. “It’s autoimmune hepatitis.”

It also is imperative that if adverse events warrant an ED visit, a patient should consult with his or her oncologist before being incorrectly treated by their primary care physician or internist, Puzanov said.

“We need to spread the knowledge across the oncology spectrum,” Puzanov said. “If a patient on chemotherapy has a fever and goes to the ED, he or she may be given antibiotics and not steroids. But, if the fever is from immune-associated inflammatory syndrome, antibiotics would not necessarily be the right treatment. He or she might be treated for pneumonia when, in fact, they have pneumonitis. If a clinician doesn’t give steroids for pneumonitis because they think it’s pneumonia, the patient dies.”

Knowing when to re-challenge

Although he called it “a moving target,” Chapman said most oncologists agree the patients who benefit most from combination immunotherapies are those with brain metastases or metastatic mucosal melanoma.

The pressing questions for oncologists with patients experiencing toxic effects from immunotherapy is when to discontinue and, perhaps more importantly, when to resume checkpoint inhibitors. Research has not yet shown whether re-challenging patients with grade 2 or grade 3 adverse events is efficacious.

“People seem to be all over the place — primarily with the question of when to stop,” Chapman said. “If there are toxicities and the patient is responding, you’re kind of stuck, but you have to stop, treat the side effects, and continue to follow the patient as long as the tumor continues to shrink, or at least not grow.”

In addition to ESMO’s clinical practice guidelines, UpToDate provides recommendations on when to re-challenge patients with immune-related toxicities.

A panel of experts compiled similar grading and treatment guidelines for chimeric antigen receptor (CAR) T-cell therapy toxicities, published in Nature Reviews Clinical Oncology in September.

Puzanov served on an expert panel that crafted new guidelines for the Society of Immunotherapy in Cancer, published this month in Journal for ImmunoTherapy of Cancer. ASCO’s new guidelines are expected in early 2018.

In summary, Puzanov recommends:

  • For grade 1 toxicities — Continue immunotherapy with close communication with the patient;
  • For grade 2 toxicities — Hold treatment, initiate steroids and continue on a 4- or 5-week taper, then re-challenge the patient with immunotherapy; and
  • For grade 3 or higher toxicities — Discontinue immunotherapy altogether and initiate steroids.

“If a patient has grade 2 or grade 3 side effects, you have to withhold immunotherapy,” Puzanov said. “But, do not despair. A patient is not going to be shortchanged. If their cancer is meant to respond to the immunotherapy, it will respond.

“That’s important to relay to patients, because you don’t want them not reporting side effects because they are worried their doctor will hold treatment and their cancer will kill them,” he added. “That fear is not rational and should not exist. The immune system, once you kick it — whether there are side effects or not — keeps going.”

Several oncologists told HemOnc Today they grapple with the idea of discontinuing immunotherapy altogether in cases where a cancer begins spreading again.

“If a tumor is shrinking, you’re done [with immunotherapy],” Chapman said. “You don’t have to re-challenge them if you’re winning. If it’s growing, it’s a different story.”

Under those circumstances, it is important to take into account the patient’s preference, Puzanov said. Changing immunotherapies — perhaps from nivolumab to pembrolizumab — is one option, along with introducing or re-introducing chemotherapy.

“That’s when the art of medicine and good communication is important,” Puzanov said. “If a patient tells me, ‘Please don’t re-challenge me. I’m worried about the toxicities,’ I wouldn’t. But if the patient says, ‘Please do whatever it takes,’ that’s different.”

Biomarkers for toxicity

Numerous biomarkers may reflect the pharmacodynamics of checkpoint inhibitors and predict their efficacy — and possibly toxicity — for individual patients. Patients not expected to benefit from therapy based on their biomarkers may thus be spared the toxicities associated with treatment.

However, the relevance of these biomarkers in clinical practice and their potential for routine application remains poorly defined.

Patients with melanoma and a PD-L1 expression less than 1% appear to benefit more from combination therapy than from monotherapy, whereas patients with high expressions of PD-L1 are more often given single-agent therapy, such as pembrolizumab.

“But what about patients with bladder, lung, head and neck, and kidney cancers?” Puzanov said. “PD-L1 in kidney cancer didn’t show a trend either way.”

In a 2016 study published in Annals of Oncology, Manson and colleagues noted PD-L1 biomarkers imperfectly identify patients who are more likely to have toxic adverse events from checkpoint inhibitors. Researchers stressed the need for better knowledge of the mechanisms of action of immunotherapies.

Sullivan said he and other researchers have been “working very hard” to identify biomarkers in patients at highest risk for developing immune-related toxicities, but described such research as in its “early days.”


Sullivan also noted that circulating lymphocytes in subsets of T cells may predict immune-related toxicities.

“These are the type of efforts I see moving forward over the next half decade — studies that stratify for a biomarker to treat or not treat to prevent toxicity,” Sullivan said. “It’s not clear there is one specific biomarker better than others. What does seem clear is, in a subset of patients, an underlying subclinical autoimmune disease gets reactivated during immune checkpoint inhibitor therapy. Just looking for that may help us better understand who’s at risk.”

In a study published this year in Journal of Clinical Oncology, Gowen and colleagues reported that pre-existing autoantibody to antigens enriched in metabolic pathways among patients with metastatic melanoma prior to treatment may predict ipilimumab-induced toxicities. In those patients, antibodies against CTLA-4 were significantly elevated in patients following ipilimumab treatment.

“We know lower doses of ipilimumab are associated with less frequent toxicities, but is it as effective?” said Chapman. “We know that efficacy decreases with very low dosage. There is a current movement to decrease the ipilimumab dose from 3 mg/kg to 1 mg/kg in combination therapy, and there are some data showing that it’s less toxic. We may need to wait to see whether the efficacy is the same.”

Reduced dosing, reduced costs

Because most combination therapies have been approved recently — and others are still in the trial phase — there is a growing debate over whether reduced dosing and treatment duration of immunotherapy might lower toxicities, while also reducing costs.

Redman said one of his longest-term cancer survivors is a man who was diagnosed with metastatic melanoma in 2003. He received three doses of anti-CTLA-4 agent tremelimumab (CP-675,206, MedImmune), experienced severe diarrhea, and was taken off the monoclonal antibody.

“Fourteen years later, he’s still disease free,” Redman said.

In their brief report in JAMA Oncology, Chapman and colleagues noted most patients did not tolerate four cycles of ipilimumab and nivolumab, adding that the prescribed four doses may not be required for clinical benefit.

“When is it safe to stop? That’s a new question for us,” Chapman said. “One doctor will treat for 1 or 2 more years because that’s what the clinical trials did, ignoring the fact it was an arbitrary stopping point because it was a clinical trial.”

In a randomized phase 2 clinical trial presented at the 2017 Genitourinary Cancers Symposium, researchers reported that taking a quarter dose of abiraterone acetate with a low-fat breakfast appeared noninferior to taking a full dose in a fasting state, potentially saving patients up to $6,500 a month.


Chapman said he believes similar clinical trials with immunotherapies would produce similar results.

“Nobody wants to support a trial like that,” Redman said. “The NIH may say, ‘That’s awfully interesting, but our agenda is to take the next step toward a new immunotherapy or a new combination, not go back and try to answer those questions.’”

Redman said he would like to see a trial in which patients receive nivolumab for 1 month, then are randomly assigned to every 4 months vs. every 2 months. He’d also like to see a trial evaluating ipilimumab and nivolumab compared with each agent as monotherapy every 2 weeks.

Trials such as these are the mission of the Value in Cancer Care Consortium, a nonprofit entity aimed at improving the access, affordability and financial sustainability of cancer therapy.

“It’s academia’s responsibility to conduct studies trying to answer these very important questions,” Jonas de Souza, MD, MBA, medical director at Humana and a former assistant professor of medicine at University of Chicago, told HemOnc Today. “Academia and oncologists are starting to look at these issues from a research perspective, and I think there is promising data that we can prove our hypothesis right — that we can potentially give a much lower dose of some drugs for some cancers.”

Until results of those trials are reported, Blum said it is imperative that clinicians inform their patients not only of the toxicities associated with immunotherapies, but the ways in which those toxicities — both physical and financial — can be better managed.

“If a drug could cost less and be more easily tolerated at a reduced dose — but with the same kind of outcome — why wouldn’t you want to share that with a patient?” Blum said. “We need to look at the person in front of us. How old are they? What’s going on in their lives? How far do they need to travel to receive treatments? What comorbidities exist? What will it mean if they have to spend $30,000 or $40,000 of their own money for treatment? That’s what informed, shared decision-making is all about.” – by Chuck Gormley

Click here to read the POINTCOUNTER, “Do enough data exist to properly screen and monitor patients for toxicities associated with immunotherapies?”


Gowen M, et al. J Clin Oncol. 2017;doi:10.1200/JCO.2017.35.15_supple.9559.

Haanen JB, et al. Ann Oncol. 2017;doi:10.1093/annonc/mdx225.

Johnson DB, et al. N Engl J Med. 2016;doi:10.1056/NEJMoa1609214.

Manson G, et al. Ann Oncol. 2016;doi:10.1093/annonc/mdw181.

Neelapu SS, et al. Nat Rev Clin Oncol. 2017;doi:10.1038/nrclinonc.2017.148.

Postow MA, et al. N Engl J Med. 2015;doi:10.1056/NEJMoa1414428.

Shoushtari AN, et al. JAMA Oncol. 2017;doi:10.1001/jamaoncol.2017.2391.

Szmulewitz RZ, et al. Abstract 176. Presented at: Genitourinary Cancers Symposium; Feb. 16-18, 2017; Orlando, Fla.

The following were presented at AACR Annual Meeting; April 1-5, 2017; Washington, DC:

Curti B, et al. Abstract CT114.

Zakharia Y, et al. Abstract CT117.

For more information:

Diane Blum, MSW, can be reached at

Paul B. Chapman, MD, can be reached at

Jonas de Souza, MD, MBA, can be reached at

Igor Puzanov, MD, MSCI, FACP, can be reached at

Bruce G. Redman, DO, can be reached at

Ryan J. Sullivan, MD, can be reached at

Disclosures: Blum, Chapman, de Souza, Puzanov, Redman and Sullivan report no relevant financial disclosures.