1 Introduction

Colon cancer is one of the most common cancers in men and women worldwide, and the incidence is increasing [1]. As life expectancy of colon cancer patients increases steadily following improved early detection and treatment [2], quality of life after cancer treatment becomes more relevant for patients and clinicians.

One of the most common and burdensome complaints during and after cancer treatment is cancer-related fatigue (CRF) [3-5]. Patients with CRF experience a distressing, persistent, subjective sense of physical, emotional and/or cognitive tiredness or exhaustion which is not proportional to recent activity, is hardly relieved by resting, and interferes with usual functioning [6]. As a result, cancer survivors suffering from CRF may be unable to regain their pre-cancer activity level and may face difficulties with daily life activities such as reading, getting out of bed, carrying out household chores, working or socialising [7]. CRF is most prominent during treatment but may persist for many years for a subset of colon cancer survivors [5].

While research efforts to understand CRF have increased in the past decades the pathophysiology of CRF remains unclear. A variety of potential biological mechanisms have been proposed, of which dysregulation of the immune system has been gaining much attention [8, 9]. Upregulation of pro-inflammatory cytokines as a result of chemotherapy treatment has been suggested to lead to CRF via central nervous system signalling [8, 10]. Longitudinal studies in colorectal cancer patients have shown that patients receiving chemotherapy score higher on CRF [11], and long-term CRF remains more common, compared to patients who did not receive chemotherapy [12]. Moreover, a recent study in breast cancer patients suggests that obese patients suffer more CRF compared to healthy-weight patients before, during and after chemotherapy [13]. Since obesity also can promote upregulation of pro-inflammatory cytokines and downregulation of anti-inflammatory cytokines [14], obese cancer patients may be at risk of more severe and prolonged CRF, especially when receiving chemotherapy.

To our knowledge, the interaction between chemotherapy and obesity has not yet been investigated in patients with colon cancer. Therefore, the aim of these secondary analyses on the PROCORE study was to evaluate the interaction between chemotherapy and Body Mass Index (BMI) on the level and trajectory of fatigue over time in colon cancer patients following their diagnosis. Exploring the interactions between chemotherapy and BMI may provide opportunities for focused (preventive/lifestyle-related) interventions to improve the quality of life of these cancer survivors.

2 Materials and Methods

3 Results

In our analyses 333 eligible patients with colon cancer were included. Table 1 shows baseline clinical and demographic characteristics. Of all patients, 94 were treated with chemotherapy (28.2%). Relatively more males were included (60.0%). The majority of patients were overweight (47.7%) or obese (19.5%). Patients treated with and without chemotherapy were comparable with regard to age, BMI and education level. Between the different BMI categories patients were comparable with regards to age, education level and chemotherapy treatment. All patients completed baseline questionnaire, 292 (88%), 259 (78%) and 247 (74%) completed T2, T3 and T4, respectively.

3.1 The Association Between Time and CRF

Intraclass correlation coefficients of the unconditional model showed that 56% of the variance was at the person level and the remaining 43% was at the time level. These results indicate that scores on CRF differed sufficiently between patients and over time to justify a two-level model. Next, we entered possible covariates (age, sex, education level) into the model. Only sex was found to be significantly different (p = 0.0035) indicating that women scored higher (10.1, 95% CI 9.4–10.8) on fatigue than men (9.0, 95% CI 8.4–9.7). Next, we computed an unconditional growth model by entering time as a fixed effect into the model. This model fitted the data better than the unconditional model and showed that time impacted the level of fatigue (p < 0.001). In this model, fatigue significantly increased from pre-treatment (9.2, 95% CI 8.6–9.6) to post-surgery (10.4, 95%CI 10.1–11.3) and decreased during follow-up at 1 (9.59, 95% CI 9.0–10.2) and 2 years (9.6, 95% CI 8.9–10.2; Figure 1, Table S1).

Moreover, the conditional growth model with time, sex, chemotherapy, BMI and the interaction terms showed that this model fitted the data as good as the unconditional growth model. A significant effect of time (p < 0.001), sex (p = 0.003), BMI (p = 0.009) and the interaction between time and chemotherapy (p = 0.047), and between time and BMI (p = 0.041) on fatigue levels was found. No significant three-way interaction between time, chemotherapy and BMI was found (p = 0.39).

3.2 The Interaction Between Time and Chemotherapy

Chemotherapy-treated patients reported more fatigue during treatment phase (11.5, 95%CI 10.5–12.4) compared to pre-treatment (9.2, 95%CI 8.3–10.1), 1 year (10.0, 95%CI 9.0–11.0) and 2 year (10.2, 95%CI 9.2–11.2) follow-up (Figure 2 and Table S2). No significant differences over time are seen in the group that did not receive chemotherapy.

3.3 The Interaction Between BMI and Chemotherapy

The significant interaction between time and BMI shows that patients with a higher BMI reported a different fatigue trajectory compared to patients with a lower BMI (Figure 3, Table S3). Post hoc pairwise comparisons revealed that while no differences in fatigue are present between the BMI groups at baseline, differences arise during and after treatment. Overweight patients reported more fatigue during treatment phase compared to healthy weight patients (mean difference 1.6, p = 0.04). During follow-up, scores of overweight patients decrease again, while obese patients reported more fatigue 1 year (mean difference 2.2, p = 0.04) and 2 years after diagnosis (mean difference 2.6, p = 0.01) compared to patients with a healthy BMI.

3.4 Chemotherapy, BMI and the Trajectory of Fatigue

The three-way interaction between time, chemotherapy and BMI was not significant. However, post hoc analyses showed that the level of CRF at a specific assessment point does depend on BMI and chemotherapy. See Table 2 for an overview of the estimated effect of the covariates and interactions on the mean fatigue score. Chemotherapy-treated patients with a BMI ≥ 30 reported the highest levels of fatigue specifically at 2 years after diagnosis (Figure 4 and Table S4).

TABLE 2. Estimated effect of the covariates and interactions on the mean fatigue score relative to the reference group.

	Estimate	95% CI	p
Intercepta	13.41	11.2, 15.6	0.000
T1	−3.59	−5.6, −1.6	0.000
T2	−0.80	−2.9, 1.3	0.456
T3	−1.16	−3.3, 1	0.296
Sex	−1.29	−2.1, −0.5	0.003
Chemotherapy	−3.20	−5.8, −0.6	0.015
BMI 18–25	−4.62	−7.3, −1.9	0.001
BMI 25–30	−3.06	−5.6, −0.5	0.019
T1 * Chemotherapy	3.55	1.1, 6	0.004
T2 * chemotherapy	1.21	−1.3, 3.7	0.344
T3* chemotherapy	1.17	−1.4, 3.8	0.378
T1 * BMI 18–25	3.94	1.4, 6.5	0.002
T1 * BMI 25–30	3.81	1.4, 6.2	0.002
T2 * BMI 18–25	2.91	0.3, 5.5	0.029
T2 * BMI 25–30	3.22	0.7, 5.7	0.012
T3 * BMI 18–25	1.42	−1.3, 4.1	0.301
T3 * BMI 25–30	1.33	−1.2, 3.9	0.307
T1 * Chemotherapy * BMI 18–25	−0.12	−3.1, 2.8	0.937
T1 * Chemotherapy * BMI 25–30	−0.93	−3.7, 1.8	0.505
T2 * Chemotherapy * BMI 18–25	1.03	−2.1, 4.1	0.512
T2 * Chemotherapy * BMI 25–30	0.56	−2.4, 3.5	0.706
T3 * Chemotherapy * BMI 18–25	2.24	−1, 5.5	0.174
T3 * Chemotherapy * BMI 25–30	1.63	−1.4, 4.7	0.290
T4 * Chemotherapy * BMI 18–25	3.15	−0.1, 6.4	0.057
T4 * Chemotherapy * BMI 25–30	2.80	−0.3, 5.9	0.073

• Note: Post hoc results from the final conditional growth model by including sex, time, chemotherapy, baseline BMI, and the interaction terms of time by chemotherapy by BMI. The values in the table represent the estimated effect of the covariates and interactions on the mean fatigue score relative to the reference group: patients with a BMI ≥ 30 who received chemotherapy at T4. T1: at diagnosis, T2: 4 weeks after surgery, T3: 1 year post diagnosis, T4: 2 years after diagnosis.
• Abbreviation: BMI, body mass index.
• ^a BMI > 30 who received chemotherapy with mean fatigue score of 12.76 at T4.

These patients reported a mean score of 12.8 (95%CI: 10.6–14.9) at T4, 3.6 point higher compared to baseline T1 (9.2, 95%CI: 7.3–11.0), p < 0.001. Patients with a BMI of 18–25 and 25–30 who received chemotherapy reported a significantly lower mean fatigue score of 8.1 (95%CI: 5.5–10.9, p < 0.001) and 9.7 (95%CI:7.2–12.3, p = 0.019), respectively, at T4. Moreover, chemotherapy-treated patients with a BMI ≥ 30 reported significantly higher levels of fatigue at T4 than their non-chemotherapy treated counterparts (9.6, 95%CI: 7.0–12.2, p = 0.015).

4 Discussion

CRF is a serious long-term effect of cancer treatment that dramatically decreases the quality of life of cancer survivors [6]. Previous research suggests that a high BMI and chemotherapy may interact to exacerbate CRF, but to what extent this applies to colon cancer patients remained unclear.

When analysing colon cancer patients as one group, fatigue increased during treatment phase, but recovered to pre-treatment levels during follow-up. Yet, when assessing the effects of chemotherapy and BMI, our study reveals several interesting observations. First, as expected, general fatigue was higher during treatment phase for chemotherapy-treated patients. As T2 was sent out 4 weeks after surgery, it is likely that not all patients scheduled for chemotherapy had started treatment when answering the questionnaire, potentially underrepresenting the effect of chemotherapy. Nonetheless, the interaction between time and chemotherapy showed that chemotherapy-treated patients were more fatigued during the treatment phase, while non-chemotherapy-treated patients showed a stable pattern. This is in line with existing literature in colorectal cancer patients [11, 12]. Second, overweight patients reported more fatigue during treatment phase compared to healthy weight patients. Obese patients remained more fatigued during follow-up compared to patients with a healthy BMI. This corroborates with earlier reported studies in colorectal cancer survivors that reported an association between higher BMI and fatigue [11, 24]. Thirdly, obese chemotherapy-treated patients did not recover to pre-treatment levels but rather continued to feel fatigued. Two years after treatment, this group of patients reported more fatigue compared to overweight (3.1 points higher) and healthy weight patients (4.6 points higher) as well as patients who did not receive chemotherapy at all (3.2 points higher). Given that a two-point difference on the MFI subscales is considered clinically meaningful, the observed long-term differences are substantial and of significant clinical relevance. These differences are highly pertinent for both patients and clinicians, as they may reflect considerable changes in functional status of cancer survivors [25]. The results are comparable with a recent study among 565 women with breast cancer treated with chemotherapy, reporting that obese patients (BMI ≥ 30) suffer more CRF compared to normal weight patients before, during and after chemotherapy [13]. Interestingly, BMI did not affect fatigue prior to treatment in our study. Differences may arise from differences in the time of measurement. In the breast cancer study, patients were assessed 7 days prior to chemotherapy and may have already been exposed to surgery and radiotherapy [13], while in our study patients had not received any treatment prior to the first measurement (T1). A high BMI itself does not appear to have a significant effect on MFI scores in healthy women [26]. The data in our study implies that the effect of BMI on fatigue is exacerbated when the patients are treated with chemotherapy, especially in the long term.

How the interplay of obesity and chemotherapy observed in this study results in an exacerbation of CRF in colon cancer patients remains to be clarified. Several lines of research have confirmed that obesity leads to chronic low-grade inflammation of the adipose tissue [27]. Weight gain and obesity cause a phenotypic switch of the adipose tissue, resulting in the release of proinflammatory cytokines [27]. Pro-inflammatory cytokines act on the central nervous system through several routes [28] and can elicit local inflammation resulting in altered brain functioning [29]. Conditions of chronic inflammation exacerbate sickness-like behaviours, characterised by poor sleep, reduced food intake, fatigue and pain complaints, social withdrawal, anhedonia, and memory and learning difficulties [28] that develop in response to acute peripheral inflammation, symptoms that are often seen with CRF. This phenomenon has been shown in several animal models with inflammatory conditions, and might be due to an effect called ‘priming’: exposure to a priming stimulus leads to an exaggerated production of pro-inflammatory cytokines when exposed to a triggering stimulus [28]. Preclinical experiments have shown that obese animals show a prolonged and more pronounced behavioural response when exposed to an inflammatory challenge [30, 31]. Chemotherapy might pose such an inflammatory challenge—alteration of cytokine profiles have been reported after standard doses of chemotherapy regimens [3, 32]. Disruption of the balance between pro– and anti-inflammatory cytokines following chemotherapy treatment may lead to a more intensified inflammatory response in obese cancer patients and result in chronic CRF. One may speculate that this may further be accelerated as obese patients may accumulate more drugs in fat tissues, resulting in a different release and altered inflammatory responses. Importantly, the effect of inflammation on brain functioning may be chronic, even if peripheral inflammation is not detectable anymore [33]. For this study, we analysed previously collected data from the PROCORE study. No data was currently available to identify biological pathways. Other mechanisms may play a role, such as alteration of gut microbiome, circadian rhythm disruption or HPA-axis dysregulation following chemotherapy [9]. Future (pre)clinical studies might examine the mechanisms of (neuro) inflammatory changes following chemotherapy treatment as well as the persistence of these changes by collecting biomarkers and proactive monitoring. Future research may also focus on lifestyle interventions to avoid fatigue exacerbation.

The value of this study includes the homogeneous sample of colon cancer patients as well as longitudinal design. The study has some limitations. No data was available on the chemotherapy type or dosing schedule. Higher body weight likely resulted in higher absolute dosing, as patients are dosed according to body surface area. Fatigue may therefore result from exposure to a higher dose of chemotherapy which we have not been able to account for. Moreover, weight and height were self-reported and may be subject to response bias. In addition, it is not known why patients were lost to follow-up, as such the effect of it on our analysis remains unclear. Furthermore, a very small number of patients received chemotherapy prior to the baseline questionnaire, which could have impacted the strength of our findings. Finally, the present results may not be generalisable to a broader group of (gastrointestinal) cancer patients. Due to the relatively low number of stage IV patients who participated in the study, our findings should be extrapolated to this group with care. Nevertheless, this study shows that further investigation of the interaction between chemotherapy, BMI and CRF is indispensable.

5 Conclusion

Our study indicates that colon cancer patients with a BMI ≥ 30 who receive chemotherapy treatment have a higher risk of remaining fatigued over longer periods, in contrast to patients that do not receive chemotherapy treatment or are not obese. Regular monitoring for CRF in this group as well as proactive communication with these patients on the possibility that their BMI may impact chronic fatigue may be needed. Moreover, lifestyle modifying coaching, including dietary and exercise interventions, may be beneficial prior, during or after treatment.

6 Précis

Results from this study indicate that colon cancer patients with a BMI≥ 30 develop chronic fatigue after treatment with chemotherapy. Two years after treatment, these patients reported significantly higher fatigue levels compared to patients with a lower BMI or patients who did not receive chemotherapy at all.

Author Contributions

Anneke Kastelein: conceptualisation, methodology, writing – original draft, visualisation. Floortje Mols: methodology, investigation, resources, writing – review and editing. Laura Kervezee: validation, writing – review and editing. Niels H. Chavannes: supervision, writing – review and editing. Hans Gelderblom: supervision, writing – review and editing. Jacques Neefjes: supervision, writing – review and editing. Chris Hinnen: conceptualisation, methodology, formal analysis, writing – original draft.

Ethics Statement

The PROCORE study was approved by the certified Medical Ethic Committee of Medical Research Ethics Committees United (registration number: NL51119.060.14).

Conflicts of Interest

The authors declare no conflicts of interest.