The Influence of Comorbidity on Health-Related Quality of Life After Esophageal Cancer Surgery

This was a prospective, single-center cohort study of esophageal cancer patients who underwent curatively intended esophagectomy at St Thomas’ Hospital, London, UK. Patient recruitment commenced on 1 November 2011 and ended on 28 February 2015. For the 6-month analysis, patients were included until 31 August 2014 and the 6-month follow-up ended on 28 February 2015. Patients were identified by both the surgical team and the study coordinators during clinical consultations, at presentation in multidisciplinary case meetings, or following screening of medical records. Detailed information about patient characteristics, including weight and height, treatment, and predefined comorbidities was prospectively collected on the basis of a detailed and predefined study protocol. The American Society of Anesthesiologists classification for physical status score (ASA) was assigned by the anesthesiologist when structurally reviewing the patient’s physical status at the time of esophagectomy. HRQoL data were collected before surgery and 6 months after surgery using self-report questionnaires developed and validated by the European Organization for Research and Treatment of Cancer (EORTC).16

Exposures were the number of comorbidities, ASA classification, and specific comorbidities. The ASA classification is a readily available and widely accepted system used before surgical procedures to assess the physical fitness of a patient on the basis of their comorbidities. It was first introduced in 1941 and has been developed into a six-category physical status system:17,18 ASA I, a healthy patient; ASA II, a patient with mild systemic disease (e.g. well-controlled diabetes mellitus or hypertension, mild lung disease, and cancer); ASA III, a patient with severe systemic disease (e.g. poorly controlled diabetes mellitus, chronic obstructive pulmonary disease [COPD], morbid obesity, or history of myocardial infarction); ASA IV, a patient with severe systemic disease that is a constant threat to life (e.g. ongoing cardiac ischemia, severe valve dysfunction, or sepsis); ASA V, a moribund patient who is not expected to survive without the operation; and ASA VI, a declared brain-dead patient whose organs are being removed for donor purposes.19 It is a partly subjective assessment, made by the anesthesiologist prior to surgery, but it is a well-established system that has been shown to be a reliable way of classifying preoperative physical status.20,21 For the purposes of the present study, patients were categorized into two ASA categories: ASA I–II, representing otherwise healthy patients, and ASA III–IV, representing patients with substantial comorbidity. However, in practice, these categories represented ASA II and ASA III, respectively, because no ASA I patients were included since all had a cancer diagnosis, there were very few ASA IV patients, and patients in ASA groups V–VI were never considered for surgery.

The number of significant and predefined comorbidities were also analyzed and patients were categorized as having: (1) no comorbidity; (2) one comorbidity; (3) two comorbidities; or (4) three or more comorbidities. Comorbidities included in the variable ‘number of comorbidities’ were the same as the specific comorbidities described below.

Finally, the following specific comorbidities assessed before surgery were analyzed separately: (1) cardiac disease (including angina, previous myocardial infarction, and heart failure); (2) hypertension (requiring medication); (3) diabetes; (4) pulmonary disease (COPD and asthma); (5) obesity (body mass index ≥ 30 just before surgery); and (6) other comorbidities (used only for adjustment in the statistical model, including other [non-esophageal] cancers, cerebrovascular disease, kidney failure, and mental disorders, including depression, rheumatoid arthritis or other significant diseases, defined by experienced physicians while manually reviewing the medical charts).

HRQoL was assessed by the EORTC questionnaires QLQ-C30 and QLQ-OG25 before surgery and 6 months postoperatively. The QLQ-C30 is a general cancer-related HRQoL questionnaire that incorporates 30 questions distributed on nine multi-item scales measuring functions (global quality of life, physical, role, cognitive, emotional, and social functioning) and symptoms (fatigue, pain, nausea, and vomiting), and six single items measuring general symptoms (dyspnea, appetite loss, insomnia, constipation, diarrhea, and financial impact).16 The QLQ-OG25 is an esophagogastric-specific module that comprises 25 items distributed on six multi-item scales (dysphagia, eating restrictions, reflux, odynophagia, pain, and anxiety) and 10 single items (eating with others, dry mouth, trouble with taste, body image, trouble swallowing saliva, choking when swallowing, trouble with coughing, trouble talking, weight loss, and hair loss).22 Items were scored on a 4-point Likert-type scale: (1) not at all; (2) a little; (3) quite a bit; and (4) very much, except for the global quality-of-life scale, which was scored on a 7-point scale ranging from (1) very poor to (7) excellent. A priori, specific aspects of HRQoL were selected based on previous literature on how comorbidities affect HRQoL,13,23 and included global quality of life, physical function, emotional function, social function, fatigue, pain, and dyspnea from the QLQ-C30; and reflux, anxiety, and trouble with coughing from the QLQ-OG25. The remaining aspects were not analyzed in order to reduce multiple testing errors.

Responses were transformed linearly to a 0- to 100-point scale according to the EORTC scoring manual, and missing responses were handled accordingly.24 High scores on the global quality-of-life and function scales correspond to better HRQoL, while higher scores on symptom scales and single items reflect worse symptoms and poorer HRQoL. Mean score differences (MDs) were calculated between exposure groups by subtracting the exposure group score from the reference category score, and over time by subtracting the preoperative score from the score at 6 months follow-up. Linear regression models were used to assess exposure groups in relation to MDs with 95% confidence intervals (CIs) for the selected HRQoL scales and items. HRQoL scores before surgery were compared between patients’ number of comorbidities (0 = reference, 1, 2, or ≥ 3), ASA scores (I–II = reference, III, or IV), and specific comorbidities (absence of specific comorbidity = reference, cardiac disease, hypertension, diabetes, pulmonary disease, or obesity). Comparisons of the HRQoL scores 6 months after surgery were analyzed regarding the number of comorbidities and ASA score groups. Linear regression analysis further examined changes in HRQoL scores over time between the assessments before surgery and 6 months after surgery. Only the number of comorbidities and ASA score groups were assessed, since the specific comorbidities were too infrequent to allow for a robust analysis. Based on previous research, an MD between assessments or between exposure groups of ≥ 10 was considered clinically relevant.25 For all models, adjustments were made for potential confounding by age (categorized into < 65 years or ≥ 65 years), sex (male or female), tumor stage (0–II or III–IV), tumor histology (squamous cell carcinoma or adenocarcinoma), neoadjuvant treatment (yes or no), surgical approach (transhiatal esophagectomy, thoracoabdominal esophagectomy, or laparoscopic esophagectomy), postoperative complications (yes or no, regarding predefined complications occurring within 30 days of surgery), tobacco smoking status (non-smoker or current smoker at the time of surgery) and other comorbidities (all comorbidities were included except the one being analyzed). To increase the power in the regression models, missing values were used as separate categories since a sensitivity analysis comparing these models with models excluding all patients with missing data showed similar results.

At a presurgery hospital visit or admission to hospital, patients were invited to participate, at which point full study information was provided and written informed consent obtained. Ethical approval was granted by the National Research Ethics Service Committee London—London Bridge (REC reference 11/LO/0335) and the National Research Ethics Service, West Midlands (REC reference 13/WM/0131).

Among 171 patients who underwent esophagectomy for esophageal cancer at Thomas’ Hospital, London, UK, during the study period, and who were thus eligible for the study, 12 (7%) declined participation, 18 (11%) did not complete the questionnaires, and 5 (3%) had missing information on tobacco smoking (4 patients) or tumor stage (1 patient), leaving 136 (80%) patients for the presurgery analysis. At 6 months after surgery, 103 patients were eligible to participate, among whom 23 (22%) did not complete the questionnaires or had died, leaving 80 (78%) patients for final analysis of changes in HRQoL between the period before and 6 months after surgery. Characteristics of the responders and non-responders to the questionnaires were similar, although complications were less common among non-responders (25% vs. 45%). Compared with the patients who responded to the follow-up questionnaire, patients who did not respond had similar baseline values in all aspects, except for a stronger male predominance (96% vs. 74%) and poorer global quality of life (MD −  14, 95% CI −  25 to −  3). Characteristics of the patients who participated in the presurgery assessment are presented in Table 1. Most patients were male (79%), had adenocarcinoma (90%), and received neoadjuvant therapy (87%). Characteristics were generally equally common among patients with different numbers of comorbidities; however, patients aged ≥ 65 years had more comorbidities.

Comorbidity was associated with poorer HRQoL before surgery, on several aspects. Patients with three or more comorbidities had clinically and statistically significantly poorer global quality of life (MD −  14, 95% CI −  27 to −  2), worse physical function (MD −  12, 95% CI −  22 to −  2), and more fatigue (MD 18, 95% CI 6–31) compared with patients without comorbidity. However, comorbidities were not associated with poorer HRQoL on emotional function, reflux, or trouble with coughing (Table 2).

Compared with patients with an ASA score of I–II, patients with an ASA score of III–IV experienced worse global quality of life (MD −  11, 95% CI − 19 to − 2), worse physical function (MD − 16, 95% CI − 23 to − 10), worse social function (MD − 13, 95% CI − 23 to − 2), more problems with fatigue (MD 15, 95% CI 6–23), and more pain (MD 13, 95% CI 4–22).

For specific comorbidities, patients with cardiac disease had clinically and statistically significantly worse global quality of life (MD − 15, 95% CI − 26 to − 3) and more dyspnea (MD 17, 95% CI 5–29) compared with those without these comorbidities. A clinically relevant difference (MD ≥ 10) was also seen for social function, fatigue, pain, and anxiety for such patients but these MDs were not statistically significant (Table 3). Patients with pulmonary comorbidities experienced clinically, but not statistically, significantly more trouble with coughing (MD 11, 95% CI 0–22) and more dyspnea (MD 10, 95% CI − 2 to 22). Patients with diabetes, hypertension, and obesity did not report poorer HRQoL on any of the aspects, except worse social function among patients with diabetes (MD 14, 95% CI 0 − 28) [Table 3].

All patients deteriorated in several aspects of HRQoL during the period before and 6 months after surgery (Table 4). For global quality of life, patients without comorbidity had the greatest deterioration (MD − 22, 95% CI − 38 to − 8). For patients in the comorbidity groups, there were clinically, but not statistically, significant deteriorations in global quality of life for patients with two comorbidities (MD − 12) and three or more comorbidities (MD − 14). Patients with comorbidities had poorer global quality of life before surgery (i.e. worse baseline scores), therefore there was no difference in mean scores 6 months after surgery between patients with or without comorbidities. For pain and reflux, both the no comorbidity group and the comorbidity groups had clinically and statistically significant deterioration between the period before and 6 months after surgery, but mean scores at 6 months were similar, although starting from different levels (Table 4). For physical function, patients with no comorbidity, two comorbidities, and three or more comorbidities all deteriorated (MD − 18, 95% CI − 32 to − 5; MD − 22, 95% CI − 34 to − 10; MD − 18, 95% CI − 34 to − 3, respectively). Similar results were found for fatigue, and patients with three or more comorbidities also had a clinically relevantly worse mean score at 6 months compared with patients with no comorbidity (mean score 62 vs. 43). For trouble with coughing, patients with two and three or more comorbidities deteriorated (MD 25, 95% CI 9–40; MD 16, 95% CI clinically relevantly worse mean 4–35, respectively), and the mean scores 6 months after surgery were statistically significantly worse compared with patients with no comorbidity and one comorbidity. No changes in emotional function and anxiety were observed over time (Table 4).

Patients with ASA I–II had similar deterioration in most HRQoL aspects as those with ASA III–IV (Table 5). For example, for global quality of life, both groups had lower scores (MD − 15, 95% CI − 28 to − 2; and MD − 13, 95% CI − 29 to 2, respectively) and experienced similar mean scores at 6 months after surgery. For physical function, social function, and fatigue, both groups deteriorated, but patients with ASA III–IV had either clinically relevant or statistically significantly worse mean scores 6 months after surgery than ASA I–II (Table 5).