Response rate and patient characteristics

Between January 2013 and 15 May 2019, a total of 839 patients underwent surgery for cancer of the oesophagus in Sweden. Of those, 636 patients were alive at the one-year follow-up, 113 were unreachable and one patient was excluded owing to cognitive impairment, leaving a total of 522 patients eligible for inclusion in the OSCAR cohort and 438 eligible for the current study as 84 were included in the cohort before the EORTC QLQ-CAX24 questionnaire was introduced. Among those identified as eligible, 173 (39%) declined to participate for reasons namely, cancer recurrence (n = 25) or other reasons (because they did not want/had the energy to participate (n = 119), were too sick (n = 28), other reasons (n = 1)). In total, 253 (58% of eligible) completed the EORTC QLQ-CAX24 questionnaire, and 232 had complete data concerning WLGS before operation and clinical data from medical charts, hence were included in the analysis.

Socio-demographic and clinical characteristics of the patients included in the present study is provided in Table 1 and no statistical test were performed to compare these characteristics. Briefly, the majority of patients were men (92.7%), with a mean age of 67 years. The majority of patients had adenocarcinoma or high-grade dysplasia (83.6%), around 79.3% received neoadjuvant therapy before surgery. High grade postoperative complications were developed in 35% of patients and 82% received enteral/parenteral nutrition support.

Table 1 Patient and clinical characteristics of patients operated for oesophageal cancer (n = 232) and categorized by preoperative weight loss grading system.Preoperative and postoperative weight loss grades

The proportion of patients with weight loss grades 0–4 before surgery compared to after surgery are represented in Fig. 2. The preoperative weight loss grades categorized as per the WLGS showed that from the 232 patients included, WLGS 0 had the largest proportion (34.5%), followed by WLGS 3 (22.4%), WLGS 2 (19%), WLGS 1 (14.2%), and the smallest proportion were in the highest grade of weight loss, WLGS 4 (9.9%) [Fig. 2]. Whilst, according to the WLGS assessed at six months after surgery, WLGS 3 (44.5%) had the highest proportion, followed by WLGS 4 (32.3%), WLGS 2 (10.5%), WLGS 1 (8.3%), and the lowest proportion were in the lowest grade of WLGS 0 (4.4%) [Fig. 2].

Fig. 2: Distribution of patients as per pre- and post-operative weight loss grading system (WLGS) (0–4).

Descriptive statistics presented as proportion (%) of patients in the exposure groups of pre- and post-operative weight loss grading system.

The clinical characteristics among the weight loss grades indicated that more patients with WLGS 0 (31.3%) had a higher comorbidity score than those in the other categories of WLGS 1–4 (15.1% to 23.1%). More patients (43.5%) with WLGS 4 had a higher pathological tumour stage (III-IV) than those with the least WLGS 0 (31%). As many as 91.3% of patients in the WLGS 4 received enteral or parenteral nutrition support compared to 77.5% in the WLGS 1 (Table 1). More patients (80%) with WLGS 4 at six months after surgery had received neoadjuvant therapy compared to 60% in WLGS 0 (data not shown). More patients (47%) with WLGS 1 at six months after surgery had higher co-morbidities than those in other WLGS (0, 2–4) ranging from 17.6% to 25% (data not shown).

Preoperative weight loss grades and cancer cachexia scores

Among the five dimensions of the QLQ-CAX24 questionnaire, the highest mean scores (MS) from the five weight loss grades were reported in the eating and weight loss worry dimension (MS 30 to 38) and food aversion dimension (MS 28 to 34) [Table 2]. Conversely, the lowest scores were reported in the physical decline dimension (MS 9 to 20). Among the four single items, the highest MS were received for heartburn/indigestion (MS 41 to 48) followed by dry mouth (MS 35 to 39) [Table 2]. The multivariable analysis of the cancer cachexia scores among the five grades of weight loss indicated that significant differences with clinical relevance were observed in the physical decline dimension and the single item, information on weight loss (Figs. 3, 4). The MD in the physical decline dimension were clinically relevant and statistically significant between WLGS 1 and WLGS 4 (MD −10, 95% Confidence Interval [CI] −20 to −1), and WLGS 2 and WLGS 4 (MD −11, 95% CI: −20 to −2) [Fig. 3]. Likewise, in the single item information on weight loss, clinically relevant and statistically significant differences in MS were observed between WLGS 2 and WLGS 3 (MD −18, 95% CI: −33 to −3), WLGS 0 and WLGS 2 (MD 16, 95% CI: 2 to 30), and WLGS 0 and WLGS 4 (MD 26, 95% CI: 8 to 44) [Fig. 4]. In none of the other dimensions and single items were significant differences observed in the cachexia scores between the five grades of weight loss [Table 2, Fig. 3].

Table 2 Adjusted mean scores on the EORTC QLQ-CAX24 (outcome) reported by patients at one year after surgery for oesophageal cancer for the pre-operative weight loss grading system 0–4 (exposure).Fig. 3: Adjusted mean scores on the EORTC QLQ-CAX24 (outcome) for the two exposure groups.

Exposure groups are a pre-operative weight loss grading system (WLGS) and b post-operative weight loss grading system. Multivariable linear regression models were used to calculate the association between pre- and post-operative WLGS and the scores from EORTC QLQ-CAX24 and adjusted for predefined confounders. Circles indicate clinically relevant MS ( ≥10 points) and statistically significant differences as per the multivariable linear regression models (p < 0.050) between a pre-operative WLGS 1–4 and 2–4 in the physical decline scale, b post-operative WLGS 2–4 and 3–4 in the eating and weight loss worry scale and WLGS 0−4 in the physical decline scale. MS Mean scores, EORTC European Organisation for Research and Treatment of Cancer, QLQ-CAX24 Cancer-cachexia specific questionnaire, WLGS Weight loss grading system.

Fig. 4: Adjusted mean scores for the single item (weight loss information) in EORTC QLQ-CAX24 (outcome) for the two exposure groups.

Exposure groups are a pre-operative weight loss grading system (WLGS) and b post-operative weight loss grading system. Multivariable linear regression models were used to calculate the association between pre- and post-operative WLGS and the scores for the single item, ‘weight loss information’ from the EORTC QLQ-CAX24 and adjusted for predefined confounders. *Clinically relevant MS ( ≥10 points) and statistically significant differences as per the multivariable linear regression models (p < 0.050) between pre-operative WLGS 2–3, 0–2 and 0–4 in the single item information on weight loss. MS Mean scores, EORTC European Organisation for Research and Treatment of Cancer, QLQ-CAX24 Cancer-cachexia specific questionnaire, WLG Weight loss grading system.

Postoperative weight loss grades and cancer cachexia scores

Among the dimensions of the QLQ-CAX24 as per the five grades of postoperative WLGS at six months after surgery, the highest MS were reported in the eating and weight loss worry dimension (MS 21 to 38) and food aversion dimension (MS 21 to 33) [Table 3]. Whereas, the lowest scores were reported in the physical decline dimension (MS 2 to 15). Likewise, among single items, the highest MS were received for heartburn/indigestion (MS 33 to 46) and dry mouth (MS 32 to 46) [Table 3]. In almost all the dimensions of QLQ-CAX24 except for the loss of control and in all the four single items, clinically relevant differences were seen between at least two or more WLGS groups, however, these did not reach statistical significance (Table 3, Figs. 3, 4). The linear regression model between the exposure of WLGS at months after surgery and the outcome of cachexia scores showed MD with clinical relevance and statistical significance in the dimensions of eating and weight loss worry and physical decline (Figs. 3, 4). In the eating and weight loss worry dimension, clinically relevant and statistically significant differences were noticed in MS between WLGS 2 and WLGS 4 (MD −17, 95% CI: −32 to −3), and WLGS 3 and WLGS 4 (MD −11, 95% CI: −21 to −2) [Fig. 3]. Further, in the physical decline dimension, MS differed statistically significantly between WLGS 0 and WLGS 4 (MD −14, 95% CI: −25 to −2) [Fig. 3].

Table 3 Adjusted mean scores on the EORTC QLQ-CAX24 (outcome) reported by patients at one year after surgery for oesophageal cancer for the post-operative weight loss grading system (exposure).