Bolte LA, Lee KA, Björk JR, et al. Association of a Mediterranean diet with outcomes for patients treated with immune checkpoint blockade for advanced melanoma. JAMA Oncol. 2023;9(5):705-709.
This study asked whether specific diets are associated with tumor response to immune checkpoint blockade (ICB) in advanced melanoma.
Adherence to a Mediterranean-style diet significantly improved outcomes in melanoma patients treated with checkpoint blockade.
Prospective cohort study
Between 2018 and 2021, dietary and clinical data were prospectively collected from 91 patients receiving treatment with immune checkpoint blockade treatment (ICB) for advanced melanoma in the United Kingdom (UK) and the Netherlands (NL). There were 44 patients in the NL cohort, and their average age was 59.4 years. The 47 in the UK cohort were older with a mean age 66.2 years. In the NL group, half (22) were female. In the UK cohort there were 15 female and 32 male patients. In the combined cohorts, 34 patients were treated with combination ipilimumab and nivolumab, 56 were treated with PD-1/PDL-1 inhibitor monotherapy, and 1 with CTLA-4 therapy.
Eighteen of the patients were taking antibiotics at baseline, and 31 were taking PPIs.
Patients in this combined UK/NL cohort were treated with checkpoint inhibitor agents (anti–programmed cell death 1 and anti–cytotoxic T lymphocyte–associated antigen 4) either as monotherapy or in combination.
Dietary intake was assessed through food frequency questionnaires before treatment.
Study Parameters Assessed
Clinical endpoints were defined as overall response rate (ORR), progression-free survival at 12 months (PFS-12), and immune-related adverse events that were grade 2 or higher. Patients were classified as responders (complete response, partial response, or stable disease) or nonresponders (progressive disease) using the standard Response Evaluation Criteria in Solid Tumors.
Investigators converted data from the food frequency questionnaires into 4 food-based scores used to assess dietary quality:
- Alternate Mediterranean diet score (aMED)
- Original plant-based diet index (oPDI)
- Healthy plant-based diet index (h-PDI)
- Unhealthy plant-based diet index (u-PDI)
Specific food groups and nutrients were also analyzed for association with immune-related adverse events (irAEs).
The Mediterranean diet was positively associated with overall response rate and progression-free survival at 12 months. This positive linear association with the Mediterranean diet was seen in generalized additive models and suggests that this dietary pattern, which is high in whole grains, fish, nuts, fruit, and vegetables, may have significant benefit for these patients.
With the Mediterranean dietary pattern, the probability of overall response rates (ORR) and progression-free survival (PFS-12) were both significantly improved (probability of 0.77 for ORR; P=0.02; probability of 0.74 for PFS-12; P=0.01).
The log odds of being a responder increased by 1.43 for every unit increase in the aMED score.
The authors included a long list of affiliations, with none of them appearing to be a conflict of interest.
This study was funded by the Seerave Foundation.
Practice Implications & Limitations
Although this journal’s primary focus is on natural medicine, it is important for us to discuss the cancer immunotherapy drugs commonly referred to as checkpoint inhibitors or, if used in combination, as an immune checkpoint blockade (ICB). This is because these drug strategies, which a short time ago were considered revolutionary, have been rapidly adopted and are now becoming commonplace: We will all be seeing patients who are undergoing these treatments. It seems that the effects of these drugs are heavily influenced by the gut microbiome and by diet. We, as naturopathic doctors, are already experts in these areas and should counsel our patients on ways to increase the likelihood these new drugs will benefit them. As this study suggests, simple dietary changes may have a significant impact on outcome.
Checkpoint inhibitors are a type of immunotherapy. They block specific proteins that may prevent the immune system from attacking cancer cells. These proteins, known as immune checkpoints, act as regulators of the immune system. Many of them serve as brakes to slow immune responses and are crucial for self-tolerance—that is, preventing the immune system from initiating an autoimmune attack on our own body. Some cancers protect themselves from immune attack by generating proteins that set off checkpoint targets, effectively putting the brakes on immune defense systems that would otherwise attack cancer cells.
Currently, there are agents approved and in use that target 3 specific checkpoints, CTLA-4, PD-1, and PDL-1. The first clinical trial of a checkpoint inhibitor started in 2000, and it has been followed by a wave of similar studies that continue to this day.1 The United States Food and Drug Administration (FDA) approved the first 2 checkpoint inhibitor drugs, ipilimumab and pembrolizumab, in 2011. Both were initially used to treat advanced melanoma. These drugs revolutionized the treatment of this cancer and are now used to treat a wide range of other cancers.
Melanoma is the deadliest of the skin cancers, with 99,780 new diagnoses and 7,650 deaths expected in the United States in 2022.2 The melanoma mortality rate was stable between 1989 and 2013; but a dramatic decline, almost 6% per year, occurred between 2013 and 2018.1,3,4 This change was due to the shift from the older chemotherapy drugs to the PD-1 inhibitors and other checkpoint inhibitors. These drugs have made a huge difference in survival as they came into widespread use.
The 2 anti-PD-1 drugs, pembrolizumab and nivolumab, have had remarkable efficacy alone or, for nivolumab, in combination with ipilimumab.5 For metastatic melanoma, a 2019 report suggests that the previous long-term survival rate of 5% increased to 50% with the use of anti-PD-1 plus anti-CTLA-4 combined treatment.6
By August 2018, 6 checkpoint inhibitor drugs had been approved for use, and the estimated percentage of cancer patients eligible for these drugs increased from 1.5% in 2011 to 43.6% in 2018.7 Estimates as to how many patients respond to treatment vary widely. It’s common to read that these drugs “provide long-term clinical benefits to nearly 40% of patients with advanced melanoma.” Some papers put the response rate higher, at 40% to 60%.8 Yet a 2020 paper from the National Institutes of Health (NIH) is far more conservative: “Checkpoint inhibitor therapy for metastatic melanoma has dramatically improved outcomes. Currently, 20 to 40% of treated patients achieve lengthy remissions.” Some estimates put the response rate quite a bit lower. In a 2020 assessment, JAMA stated that only “up to 12.5% of patients respond…” Furthermore, they point out that these numbers “…were best-case estimates, and post-marketing studies for several of these drugs have failed to show improvement in overall survival or progression-free survival.”7 The phenomenon of variability in outcomes shows up often when reading about these drugs. There are factors that affect efficacy that we don’t understand yet.
Bolte et al, the authors of this current study on the Mediterranean diet, use an optimistic estimate that “Durable responses are observed for 40% to 60% of patients, depending on treatment.” When these drugs work, they can be incredibly effective, nearly miraculous. For example, recall how President Jimmy Carter was treated with pembrolizumab, a PD-1 inhibitor, for metastatic melanoma in his brain back in 2016 and left the hospital seemingly cured.
In recent years attention has focused on how the response rate to these drugs might be increased. Considerable attention has focused on the impact of the intestinal microbiome. It is well-accepted that antibiotics reduce the odds that patients will respond to immune checkpoint treatment.9 The working theory to explain this is that the gut microbiome plays an essential role in regulating the immune response.
Two landmark trials were published in Science in November 2020, Davar et al10 and Baruch et al.11 In both these small clinical trials, fecal transplants were used to transfer gut microbiomes from patients who had responded well to checkpoint treatment to patients who had previously failed to respond. One-third of the formerly nonresponsive patients then showed significant benefit when retreated with checkpoint inhibitors.
Patients consuming high-fiber diets were 5 times more likely to respond to anti-PD-1 therapy than those on a low-fiber diet.
Initial attempts to identify the essential components in the gut microbiome that regulate the resultant immune responses after checkpoint inhibitor treatments have been elusive. One study after another has compared the microbiomes of responders to treatment with nonresponders in the hope of identifying key differences. There were early hints that Bifidobacteria were more common in responders, a finding that prompted us to encourage patients undergoing treatment take oral Bifidobacter probiotics.12 Akkermansia muciniphila were also an early nominee for playing an important role.13 Then Ruminococcaceae were added to the list14 and then Bacteroides intestinalis.15
A problem became apparent over time. None of these suspected good bacteria was consistent from study to study—nor were attempts to identify specific bacterial genetics that differed significantly between responders and nonresponders. Consider what a challenge this would be: The human genome consists of about 23,000 genes, while the gut microbiome contains more than 3 million genes, any of which alone or in some combination might influence the immune reactions of the host.
The idea that fecal microbiome transplants might benefit a large minority of patients who had failed initial treatment was exciting but problematic. Ethical considerations and the ground rules for fecal transplants disallow using fecal material obtained from cancer patients for transplant, at least outside of clinical trials. This is a sensible precaution as we know that some cancers are associated with dysbiotic gut biomes.
In June 2023, John McCulloch et al published a comprehensive and relatively large study involving sophisticated analysis of data collected from responders vs nonresponders.16 Data from 94 patients in the Pittsburgh Cohort were then combined with data from 4 earlier cohorts (Gopalakrishnan,15 Matson,17 Frankel,18 and Peters19) and used to draw some broad conclusions. “Collectively, our findings hint that an unfavorable gut microbiome enriched in Gram negative bacteria may promote an LPS-dominated inflammatory signature in the gut, resulting in systemic inflammation manifested by elevated NLR [neutrophil to lymphocyte ratio] and poor response to antiPD-1,” wrote the investigators. They also concluded that “Two microbial signatures, enriched for Lachnospiraceae spp. and Streptococcaceae spp., were associated with favorable and unfavorable clinical response, respectively, and with distinct immune-related adverse effects.”
But were there single specific bacteria that needed to be present for a positive response? Not exactly. The reality sounds more complex than we imagined. McCulloch et al concluded that the human microbiome varies by geography, diet, and other factors and that multiple different biomes might suffice to favor immune responses when stimulated by checkpoint inhibitors. The bottom line is that it seems that having the ideal gut microbiome in supplement form is still a long way off.17
Nevertheless, we do have some good information about how to advise patients to create a gut microbiome that might improve their response to treatment. What is interesting about these suggestions is that they are familiar advice for any naturopathic physician to share with patients.
This new study by Bolte et al suggests that compliance with the Mediterranean diet guidelines may be the most critical information for us to convey to patients as they undergo treatment. Below is a summary graph of the study’s results:
This graph compares the probability of an overall response to treatment with the Mediterranean diet score calculated from dietary questionnaires. As described in previous articles, the Mediterranean diet is scored on a scale of 0 to 5 based on consumption levels of key components of the diet. This graph suggests that the chance of a response to treatment increases from less than 0.4 to almost 0.8 with dietary changes. That’s almost a doubling in the probability of a response and in the same range of effect as what has been demonstrated with fecal microbiome transplants.
Such a significant difference in benefit between differing dietary patterns is plausible. Some of the gut bacteria that have been suggested in past studies as being associated with the immune checkpoint blockade response are associated with dietary intake of fiber.
Diet can determine whether any of several types of gut bacteria predictive of a response to ICB will be present.20 Bifidobacteria, for example, degrade dietary fiber to short-chain fatty acids (SCFAs). High-fiber intake and fecal SCFA concentrations are clearly associated with ICB treatment response in both mice and humans.12,21 SCFA levels appear to be predictive of treatment response.
In a 2020 paper, Nomura et al reported on 52 patients with solid tumors treated with nivolumab or pembrolizumab and then divided into 2 groups, responders and nonresponders; their pretreatment fecal SCFAs were compared. Slightly less than 30% of patients responded to treatment. High concentrations of certain SCFAs were associated with longer progression-free survival. These included acetic acid (hazard ratio [HR], 0.29), propionic acid (HR, 0.08), butyric acid (HR, 0.31), valeric acid (HR, 0.53; 95% confidence interval [CI], 0.29–0.98), and plasma isovaleric acid (HR, 0.38).
So perhaps it isn’t what bacteria are present as much as their output of SCFAs that is responsible for the difference between responders and nonresponders?22 Thus, it could be a combination of diet and gut microbiome that determines the reaction to treatment.
SCFAs result from bacterial fermentation of fiber. This adds greater relevance to Spencer et al’s 2021 study that assessed dietary and supplement effects in 128 melanoma patients treated with ICB. The benefit of higher fiber diets was evident in these patients. Low-fiber intake in that study was defined as less than 20 grams per day, while adequate fiber was considered more than 20 grams per day. Patients consuming high-fiber diets were 5 times more likely to respond to anti-PD-1 therapy than those on a low-fiber diet. Every 5-gram increase in daily dietary fiber intake corresponded with a 30% lower risk of progression or death in these patients.
There was another striking finding in Spencer’s data: Patients taking probiotic supplements (as nearly 30% of the patients in her study were) had a significantly lower rate of response.
While these data were initially discovered in a human clinical trial, Spencer went back to confirm her findings in mice. In the mice, supplementation with either Acidophilus or even Bifidobacter led to much more rapid tumor growth rates than in control mice who were given distilled water instead of the probiotics.24
These findings suggest we should advise cancer patients who may be treated with either a checkpoint inhibitor or ICB to not only follow a Mediterranean-style diet but also to not take probiotics. There may be a few obscure exceptions to this prohibition.
A clinical trial treating renal cell cancer with ICB (nivolumab plus ipilimumab) was published in 2022; in it, some of the patients were also given a specific probiotic called CBM588. This is a strain of Clostridium butyricum, sometimes referred to as MIYAIRI 588, which was isolated by Dr Chikaji Miyairi in 1933 from human fecal material. It has been used as a probiotic in Asian countries (Japan, Korea, and China) for many years. It is difficult to find in the United States. Aside from several clinical trials that City of Hope Hospital is conducting, this product remains rare in the United States.
In that 2022 study, CBM588 use was associated with significantly longer progression-free survival (12.7 months vs 2.5 months, HR 0.15, 95% CI, 0.05–0.47, P=0.001). The response rate may have been higher for the patients who took this probiotic, but it did not reach statistical significance (58% vs 20%, P=0.06).23 While the odds are low that any of your patients will be taking this supplement, we should not advise against its use.
The other possible exception to our no-probiotic advisement might be homemade kefir. The rice-like grains formed during the fermentation process appear to enhance anti-PD-1 therapy in mice.24 (The bacteria responsible for this granular formation are not present in commercially made kefir.) Translating these findings from mice to humans is a long reach though. Mice and humans have only about 4% of their gut microbiomes in common,25 so predicting the effect of kefir in humans is difficult.
While a great deal of information now exists about how various diets alter the gut microbiome, we should be hesitant to proscribe specific diets based only on what is presumed to be the desirable microbiome.26 This is why the Bolte study is so important. It clearly associates compliance with the Mediterranean diet with the clear human end points of overall response and progression-free survival.
While on the topic of what advice to offer these patients, a few other considerations deserve mention. Proton pump inhibitor (PPI) therapy has been consistently associated with reduced response to these immunotherapy treatments. While it’s apparent that these drugs alter the gut biome, it is not clear exactly why they have such a negative impact. In a study of 140 melanoma patients, taking PPIs cut the response rate to ICBs by nearly half.27
At this point, the best advice we might provide to patients who have been treated with any of these checkpoint-targeting immunotherapies is for them to strive to eat a Mediterranean-style diet, avoid probiotic supplements, and wean off PPI drugs if they are currently using them.