The research team, led by Damir Zubac together with Timo Sonntag, Diana Kranjc and Freerk T. Baumann, recruited 19 patients undergoing acute immune checkpoint inhibitor treatment and 21 age-matched healthy controls from the Cologne area. Participants were screened to exclude hypertension, diabetes, chronic lung disease and obesity, and were required to be free of brain, bone or lung metastases. The patients, who had been diagnosed roughly 14 months earlier and had been on immunotherapy for about six months, most commonly had malignant melanoma or breast cancer, with the majority treated with the PD-1 antibodies pembrolizumab or nivolumab. Crucially, the two groups were well matched: there were no significant differences in age, height, body mass, body mass index, resting blood pressure, oxygen saturation or body composition measured by bioelectrical impedance.

Each participant completed two laboratory visits. In the first, they performed a ramp-based cardiopulmonary exercise test on a cycle ergometer to exhaustion, with breath-by-breath measurement of pulmonary gas exchange to determine peak oxygen uptake, peak power output, the gas exchange threshold and the oxygen uptake efficiency slope. In the second session, they completed the Multidimensional Fatigue Inventory questionnaire, a validated 20-item instrument covering general, physical, mental fatigue, reduced activity and reduced motivation, followed by a 30-minute steady-state cycling test at 90 percent of each individual’s gas exchange threshold, during which energy expenditure and gross cycling efficiency were calculated.

The results painted a clear picture of diminished physical capacity in the immunotherapy group. Patients showed significantly lower pulmonary ventilation, oxygen uptake, carbon dioxide production, stroke index and oxygen uptake efficiency slope during maximal testing, all with large effect sizes. Their exercise tests ended roughly four minutes sooner than those of controls, and the peak power output they achieved was dramatically lower, with an effect size of 2.35. Importantly, patients reported the same subjective effort at exhaustion, rating around 19 on the standard scale, meaning they were not simply holding back. The pattern of reduced oxygen pulse alongside only modest differences in maximal heart rate suggests that the limitation lies primarily in the heart’s ability to augment stroke volume and cardiac output during maximal effort, a finding the authors note could hint at early cardiotoxicity, though no direct cardiac diagnostics were performed.

Fatigue, however, was the most dramatic group difference of all. Using age- and sex-standardized cutoffs derived from a representative German community sample, 63 percent of the immunotherapy patients met criteria for clinically relevant chronic fatigue, compared with just 15 percent of the healthy controls. Across all five dimensions of the questionnaire, patients scored substantially worse: general fatigue was elevated by 38 percent, physical fatigue by nearly 40 percent, and mental fatigue, reduced activity and reduced motivation were all similarly increased with moderate to large effect sizes.

Yet when the researchers ran linear regression models asking whether peak oxygen uptake or peak power output, adjusted for age, could predict general fatigue scores, the answer was a consistent no. In healthy controls, the models explained between 8 and 13.5 percent of the variance in fatigue, and in immunotherapy patients only between 4.3 and 5.6 percent, none of it statistically significant. Age, too, failed to predict fatigue, despite the common assumption that aging drives both declining fitness and increasing tiredness. Multicollinearity diagnostics were clean, so the null result was not a statistical artifact of correlated predictors.

A particularly novel observation emerged from the submaximal endurance test. Although every single participant, patient and control alike, managed to complete the full 30 minutes of moderate-intensity cycling at 90 percent of their gas exchange threshold, the patients did so at significantly lower gross efficiency. The average workload was 122 watts for controls versus 88 watts for patients, and the patient group converted metabolic energy into mechanical work measurably less efficiently, with a mean group difference of just over two percentage points. Because similar data in checkpoint inhibitor populations do not otherwise exist, the authors can only hypothesize that impaired mitochondrial coupling in skeletal muscle, a phenomenon previously demonstrated with in vivo phosphorus magnetic resonance spectroscopy in older adults, may underlie the inefficiency. Supporting this idea, recent muscle biopsy work has shown that even a single chemotherapy dose induces mitochondrial dysfunction in breast cancer patients.

The disconnect between how tired patients feel and how their bodies perform on objective testing carries real clinical weight. Structured exercise is recommended as a first-line therapy for cancer-related fatigue, and prior work in immunotherapy-treated patients has shown that exercise interventions can reduce fatigue scores by roughly 8 to 10 points on validated scales. The present findings suggest that a clinically meaningful level of fatigue does not preclude moderate-to-vigorous aerobic exercise: all 19 fatigued patients completed a half hour of cycling under controlled conditions. This supports the feasibility of individualized aerobic exercise prescriptions, calibrated to objective test results rather than to how patients subjectively rate their energy.

The authors are careful to frame the study as exploratory. The cross-sectional design, the modest sample size of 40 analyzed participants, the heterogeneity of cancer diagnoses and prior treatments, and the absence of direct cardiac or mitochondrial measurements all limit how far the conclusions can be pushed. The findings characterize people receiving checkpoint inhibitor therapy rather than proving causal effects of the drugs themselves, and the regression models leave open the question of what actually drives chronic fatigue, with circulating inflammatory cytokines, muscle metabolic changes and neurotransmitter disruption all implicated in the wider literature.

Even so, the study marks a step beyond earlier feasibility and adherence work in this population. It provides the first data on work efficiency in patients under acute checkpoint inhibitor therapy, offers a physiological rationale for prescribing aerobic exercise at the gas exchange threshold, and points future investigators toward more comprehensive, non-invasive assessments of the oxygen transport pathway, such as flow-mediated dilation and vascular occlusion testing, to finally unravel why immunotherapy patients feel so tired even when their exercise capacity tells only part of the story.

Subject of Research: The relationship between self-reported chronic fatigue and objectively measured cardiopulmonary exercise parameters in adult cancer patients receiving immune checkpoint inhibitor therapy

Article Title: Chronic, self‐reported fatigue in adult immunotherapy patients and healthy controls: No association with cardiopulmonary exercise testing parameters

Article References: Zubac, D., Sonntag, T., Kranjc, D., & Baumann, F. T. (2026). Chronic, self‐reported fatigue in adult immunotherapy patients and healthy controls: No association with cardiopulmonary exercise testing parameters. Physiological Reports, 14(17), Article e71090. https://doi.org/10.14814/phy2.71090

Image Credits: AI Generated

DOI: 10.14814/phy2.71090

Keywords: immunotherapy, immune checkpoint inhibitors, chronic fatigue, cancer-related fatigue, cardiopulmonary exercise testing, VO2 peak, cycling efficiency, exercise oncology, cardiorespiratory fitness, MFI-20, mitochondrial dysfunction, supportive care