a prospective longitudinal study examining the quality of life of patients with esophageal carcinoma

A Prospective Longitudinal Study Examining theQuality of Life of Patients with Esophageal Carcinoma

Jane M. Blazeby, M.D.1

John R. Farndon, M.D.1

Jenny Donovan, Ph.D.2

Derek Alderson, M.D.1

1 University Department of Surgery, Bristol RoyalInfirmary, Bristol, United Kingdom.

2 Department of Social Medicine, University ofBristol, Bristol, United Kingdom.

An earlier version of this article was presented atthe Association of Upper Gastrointestinal Surgeonsin Bristol, United Kingdom, September 1998, andpublished in abstract form in the British Journal ofSurgery (1999;86:419).

Jane M. Blazeby was supported by a South andWest NHS Research Training Fellowship.

The authors thank Mr. K. Jeyasingham and Mr.C. P. Forrester-Wood for allowing them to studypatients in their care.

Address for reprints: Jane M. Blazeby, M.D., Uni-versity Department of Surgery, Bristol Royal Infir-mary, Bristol BS2 8HW, United Kingdom.

Received September 14, 1999; revision receivedDecember 10, 1999; accepted December 10,1999.

BACKGROUND. Quality of life (QL) measurement provides detailed information

about outcome from the patient’s perspective. This study assessed the impact on

short and long term QL of esophagectomy and palliative treatment in patients with

esophageal carcinoma.

METHODS. Consecutive patients undergoing potentially curative esophagectomy or

purely palliative treatment completed the European Organization for Research and

Treatment of Cancer (EORTC) QLQ-C30 and the dysphagia scale from the EORTC

QLQ-OES24 before treatment and at regular intervals until death or for 3 postop-

erative years. Median scores were calculated for patients surviving more than 2

years after surgery (n 5 17), for patients surviving less than 2 years after esopha-

gectomy (n 5 38), and for patients undergoing palliative treatment (n 5 37).

RESULTS. Baseline functional and symptom QL scores were similar in both groups

of patients undergoing esophagectomy, and these were better than scores from

patients selected for palliative treatment. Six weeks after esophagectomy, patients

reported worse functional, symptom, and global QL scores than before treatment.

In patients who survived at least 2 years, QL scores returned to preoperative levels

within 9 months, but patients who died within 2 years of surgery never regained

their former QL. In both groups, dysphagia improved after surgery and the im-

provement was maintained until death or for the duration of the study. Patients

undergoing palliative treatment reported gradual deterioration in most aspects of

QL until death.

CONCLUSIONS. Esophagectomy has a negative impact on QL; this effect is transient

for patients who survive for 2 or more years. This finding should be considered

when selecting patients for surgery. Cancer 2000;88:1781–7.

© 2000 American Cancer Society.

KEYWORDS: quality of life, esophageal carcinoma, European Organization for Re-search and Treatment of Cancer (EORTC) QLQ-C30, esophagectomy, intubation.

The optimum management of patients with carcinoma of theesophagus remains controversial. Patients often are elderly and

have significant comorbid disease. By the time they are symptom-atic, the condition may be locally advanced. Treatments such asradical surgery or chemoradiation are associated with high com-plication rates and consequent mortality. Two-year survival ratesafter surgery are between 20% and 50%, and only 10 –30% ofpatients are alive 5 years after surgery.1,2 Treatment selectionshould therefore be based on a critical assessment of the patient’sgeneral health, disease stage, and knowledge of the likely outcomeof intervention. Basic data on morbidity, mortality, and survival areavailable, but few studies have prospectively assessed quality of life(QL) for patients with esophageal carcinoma using valid multidi-mensional tools, and no long term QL data have been reported.3–10

1781

© 2000 American Cancer Society

Accurate data indicating how different treatmentsimpact on short and long term QL can inform pa-tients and doctors and assist in treatment decisions.The purpose of this longitudinal study was to exam-ine QL using a valid cancer specific questionnaire ina cohort of patients with esophageal carcinomatreated according to routine clinical practice.

PATIENTS AND METHODSBetween November 1993 and May 1995, baseline clin-ical data on 92 consecutive new patients presentingwith carcinoma of the esophagus to the Bristol RoyalInfirmary and Frenchay Hospital, Bristol, United King-dom, were prospectively entered into a data base.Standard assessments of general health and tumorstage included a full blood count, blood chemistry,electrocardiography, pulmonary function tests, chestradiography, endoscopy, and biopsy. Computed to-mography and/or magnetic resonance imaging wereperformed to exclude distant metastases. Patientswithout major comorbid disease or evidence of distanthaematogenous spread underwent esophagectomywith conventional lymphadenectomy. Patients unfitfor esophagectomy or with distant metastatic diseasewere treated by intubation with plastic tubes or ex-panding metal stents, or with palliative chemoradio-therapy.

Fifty-five patients underwent esophagectomy per-formed by three surgeons. Five of these patients re-ceived neoadjuvant chemotherapy as part of the Med-ical Research Council’s OEO2 trial.11 Adjuvantchemotherapy or radiotherapy was not administered.Four patients underwent laparotomy and intubationonly, as unsuspected liver metastases were present.Twenty-six patients were treated by intubation, andthree of these received a course of palliative radiother-apy. Seven patients underwent primary palliative ra-diotherapy or chemotherapy. Informed consent of allpatients and ethical committee permission were ob-tained for this study.

For the purpose of this study, patients wereretrospectively divided into three groups: Group 1,patients surviving longer than 2 years after esopha-gectomy; Group 2, patients who died within 2 yearsof esophagectomy; and Group 3, patients undergo-ing treatment with purely palliative intent. Patientsselected for esophagectomy (Groups 1 and 2) hadsimilar preoperative characteristics, althoughpathologic staging showed that patients in Group 1had less tumor burden than patients in Group 2(Table 1). Patients in Group 3, selected for palliativetreatment alone, were older, with lower perfor-mance scores and greater tumor burden than pa-tients in Group 1 or 2 (Table 1).

TABLE 1Patient Characteristics and Median Scores

Group 1 (n 5 17)Alive 2 yrs aftersurgery

Group 2 (n 5 38)Died within 2 yrsof surgery

Group 3 (n 5 37)Treatment withpalliative intent

Age (IQR) 63 (58–74) 71 (51–79) 74 (67–80)Gender

Male/Female 9/8 24/14 23/14Histologic diagnosis

Adenocarcinoma 11 21 22Squamous cell carcinoma 6 17 15

Tumor length (IQR) 5 (3–8) 5 (5–6) 5 (5–7)Distance from incisors (IQR) 36 (30–38) 35 (30–37) 36 (30–38)Dysphagia grade (IQR)24 3 (2–3) 3 (2–3) 4 (3–4)Karnofsky score (IQR) 90 (80–90) 90 (80–90) 70 (60–80)Comorbid disease

ASA grade (IQR)25 2 (2–3) 2 (2–3) 3 (3–3)Neoadjuvant treatment

Chemotherapy 2 3 —TNM pathologic stage26

I 0 0 —IIA 8 3 —IIB 1 3 —III 8 32 —IV 0 0 25

IQR: interquartile range.

1782 CANCER April 15, 2000 / Volume 88 / Number 8

Quality-of-Life AssessmentPatients who could speak or read English and whoconsented to participate in the QL study were asked tocomplete questionnaires 1–2 weeks before surgery orpalliative treatment. Further QL assessments wereperformed 4 – 6 weeks after esophagectomy, every 3months in the first postoperative year, every 6 monthsduring the second year, and annually thereafter. Aftertreatment with palliative intent, assessments wereperformed monthly. Data were collected until death orfor a minimum of 3 postoperative years. Baseline andfirst postoperative assessments of patients undergoingesophagectomy were performed in the patients’homes. Subsequent questionnaires were mailed un-less additional help was required. All patients receiv-ing purely palliative treatment were visited at homefor every assessment. Questionnaires were checked foromitted answers before completion of the interviews.Postal questionnaires were checked for missing data.If answers were absent, patients were contacted andasked the omitted questions. The European Organiza-tion for Research and Treatment of Cancer (EORTC)QLQ-C30 (version 1.0) and the dysphagia scale fromthe esophageal cancer module (the EORTC QLQ-OES24) were completed at each assessment.12,13 TheEORTC QLQ-C30 contains scales and items addressingfunctional aspects of QL and symptoms that com-monly occur in patients with cancer, as well as itemsassessing the financial impact of malignant diseaseand global QL. As the esophageal carcinoma specificmodule was under development at the time of thisstudy, only results from the three-item dysphagia

scale were analyzed.13,14 Patients with missing formswere excluded from the analysis of the absent assess-ment point. Results from patients in Groups 2 and 3are not shown after 6 and 12 months, as the numberswere too small to produce meaningful data. Medianvalues were calculated at each time point for eachgroup of patients. Comparisons between the 2 groupsof patients who underwent esophagectomy (Groups 1and 2) were made 6 months after surgery using theMann–Whitney U test.

RESULTSMorbidity, Mortality, and SurvivalTwenty-six patients had major complications within30 days of surgery, and 9 of these patients died (Table2). Five patients died within 30 days of palliative treat-ment and 7 complications occurred (Table 2). Seven-teen patients (Group 1) survived longer than 2 yearsafter surgery. Thirty-eight patients (Group 2) diedwithin 2 years of esophagectomy and 37 patients(Group 3) were treated with purely palliative intent.The majority of deaths within 2 years of surgery werecaused by disease recurrence (n 5 22). Five patientshad fatal cardiorespiratory problems, and in 2 casesthe cause of death was unknown. The median survivaltimes of patients in Groups 1, 2, and 3 were 46, 7.5,and 2.5 months, respectively.

Quality of LifeQuestionnaires were completed at each time pointexcept for five early postoperative assessments, whichwere not completed because patients felt unwell.

TABLE 2Thirty-Day Morbidity and Mortality after Esophagectomy and Intubation

Group 1 (n 5 17)Alive 2 yrs aftersurgery

Group 2 (n 5 38)Died within 2 yrsof surgery

Group 3 (n 5 37)Treatment withpalliative intent

MorbidityPulmonary complications 2 2 2Anastomotic leak — 2 —Wound hematoma/infection 2 — —Upper gastrointestinal bleed — 1 2Recurrent nerve palsy — 2 —Deep vein thrombosis 1 1 1Tracheoesophageal fistula 1 — —Early anastomotic stricture 1 2 —Early stent blockage — — 1Esophageal perforation — — 1

MortalityAnastomotic leak — 5 —Cardiac failure — 2 —Respiratory failure — 1 1Gastric necrosis — 1 —Disease progression — — 4

Quality of Life for Esophageal Carcinoma Patients/Blazeby et al. 1783

Three patients in Group 2 did not take part in the QLstudy; 1 was unable to speak or read English, 1 under-went emergency esophagectomy, and 1 was inadver-tently omitted.

Figures 1– 8 show median functional and symp-tom QL scores for the 3 groups of patients for the mainQL domains in the EORTC QLQ-C30 and the dyspha-gia scale from the EORTC QLQ-OES24. Baseline func-tional and symptom scores were similar for patients inGroups 1 and 2.

Six weeks after esophagectomy, patients inGroup 1 (2-year survivors) reported a decrease inmost functional aspects of QL and an increasednumber of symptoms. These subsequently im-proved such that at 9 months after surgery, QLscores had returned to preoperative levels. Sixweeks after surgery, patients in Group 2 (survivorsfor less than 2 years) also reported worse QL in alldomains, but these mostly never recovered. Theonly exception was the global QL scale (Fig. 4), inwhich patients in Groups 1 and 2 eventually had

similar scores, although patients in Group 2 hadmuch worse scores between Months 3 and 8.Throughout the study period, QL scores in the func-tional scales and the global QL scale were higher inGroup 1 than Group 2, and Group 1 patients re-ported fewer symptom scores than Group 2. Six

FIGURE 4. Median global quality of life (QL) scores are shown by treatment

group A high score is equivalent to a better QL.

FIGURE 5. Median fatigure scores are shown by treatment group. A high

score is equivalent to more symptoms.

FIGURE 1. Median physical function scores are shown by treatment group.

A high score is equivalent to a better quality of life (QL).

FIGURE 2. Median emotional function scores are shown by treatment group.

A high score is equivalent to a better quality of life (QL).

FIGURE 3. Median social function scores are shown by treatment group. A

high score is equivalent to a better quality of life (QL).


months after treatment, Group 2 had significantlyworse physical, emotional, and social function thanthose surviving 2 years (Group 1) and significantlymore problems with appetite loss (P , 0.01). Theexception was dysphagia scores, which were similarin Groups 1 and 2 (Fig. 8).

Patients in Group 3, undergoing palliative treat-ment, reported worse baseline functional QL scoresand an increased number of symptoms comparedwith all patients undergoing esophagectomy (Figs.1– 8). Following treatment with palliative intent, mostaspects of QL were maintained until death.

Results from the cognitive and role-functionscales are not shown, but they indicated similar trendsto Figures 1– 4. Similarly, the dyspnea and nausea-and-vomiting scales show similar trends to Figures5–7. Scores from items assessing sleep problems, con-stipation, diarrhea, and financial difficulties are notpresented here because very few patients reportedthese in any group.

DISCUSSIONThis longitudinal study shows that patients who survivefor at least 2 years after esophagectomy regain theirpreoperative QL within 6–9 months, whereas patientswho die within 2 years experience an irreversible dimin-ishment in QL. Patients selected for palliative treatmentinitially have a poorer QL than patients selected for sur-gery; but after treatment, their QL scores are similar tothose of patients who die within 2 years of surgery. Threeother studies have prospectively examined QL for pa-tients undergoing esophagectomy using valid instru-ments, and they have reported similar results to thoseobtained from patients in Group 1 in this study: mostaspects of QL get worse in the early postoperative period,but this improves with time.5–7 These studies had shortfollow-up (less than 12 months) and patients with recur-rence or complications were excluded from analysis. Thedata from patients in Group 2 in this study (who diedwithin 2 years of esophagectomy) are particularly impor-tant because they reveal that apart from dysphagia, allaspects of QL deteriorate until death. This means thatclinicians should try to identify those features likely to beassociated with poor survival in order to counsel patientsand plan more appropriate management strategies. He-matogenous metastases, incomplete excision with mac-roscopic residual tumor, and extensive lymph node in-volvement are all known to be associated with shortsurvival times.2 It was not possible to distinguish differ-ences in these features before surgery in patients inGroup 1 and Group 2 in this study. Endoluminal ultra-sonography combined with fine-needle aspiration ofsuspicious lymph nodes may improve selection of pa-tients for surgery, and preoperative laparoscopy can de-tect small peritoneal metastases.15–17

The analysis of longitudinal QL data in the contextof missing information due to attrition is problem-

FIGURE 6. Median pain scores are shown by treatment group A high score

is equivalent to more symptoms.

FIGURE 7. Median appetite loss scores are shown by treatment group A high


FIGURE 8. Median dysphagia scores are shown by treatment group A high



atic.18,19 In this study, patients who lived longer than 2years were deliberately separated from those who diedwithin 2 years of surgery in an attempt to minimizethe biases that occur when assessments only includehealthy survivors who are able to complete question-naires. This approach only partially reduced the diffi-culties of data analysis, as Groups 2 and 3 still hadrapidly diminishing patient numbers. Other methodsof analysis, such as the Q-TWiST, can be used tocompare survival times between treatment groupswhile adjusting for differences in QL.20 This approachwas not implemented in the current study because ofthe difficulties in dividing the time with symptomsand time free of symptoms or toxicity.

Patients’ responses to QL questionnaires may beinfluenced by the locations in which data are col-lected.21 In this study, all baseline and first posttreat-ment assessments were performed at patients’ homes.Thereafter, Group 3 patients continued with homeassessments, whereas Groups 1 and 2 received postalquestionnaires. After an initial decrease in QL, pa-tients in Group 1 reported better postoperative QLscores than patients in Groups 2 or 3. These differ-ences cannot be explained by the methods of datacollection alone because patients in Groups 1 and 2underwent similar QL assessments. Whenever possi-ble, questionnaires should be completed in a con-trolled setting throughout a study.18

Assessment of the patient’s perception of out-come is acknowledged as an accepted part of clinicalresearch, and QL questionnaires are currently amongthe most important tools for this endeavor. TheEORTC QLQ-C30 has been validated for use withmany patients who have cancer, including carcinomaof the esophagus, but does not contain questionsabout dysphagia or the psychosocial consequences ofeating difficulties.12,22 The esophageal carcinomamodule (EORTC QLQ-OES24) addresses QL issues thatare relevant to patients with malignant dysphagia, butthe full module was not available at the time of thisstudy.13,14 Dysphagia was the only aspect of QL thatimproved after esophagectomy. In 1980, Earlam statedthat “since the original symptom is dysphagia, it ispresumed that removal of this should produce a goodquality of life, but there is no objective evidence avail-able.”23 A few studies have now questioned this as-sumption by simultaneously measuring dysphagiagrade and QL.3–5,22 The findings are mixed, but overallthey indicate that the proportion of variance of QLthat can be accounted for by dysphagia is relativelylow (15–20% or less).22 Further studies using theEORTC QLQ-C30 and the esophageal carcinoma spe-cific module, EORTC QLQ-OES24, will provide data to

allow comparisons between site specific and genericaspects of QL.

Esophagectomy has a significant negative impacton self-assessed QL; this impact is transient for pa-tients who survive for 2 years or longer (Group 1).Patients who die within 2 years of surgery (Group 2),however, suffer an irreversible loss of QL apart fromthe relief of dysphagia. Patients selected for palliativetreatment (Group 3) report QL scores similar to thoseof patients who undergo surgery with curative intentbut die within 2 years of surgery (Group 2). Every effortshould be made to identify patients whose survivalwith surgery alone will be poor, so that suitable neo-adjuvant or palliative treatment can be offered. Untilthis can be reliably achieved, selection of patients foresophagectomy should include consideration of thepotential effects of surgery on QL.

REFERENCES1. Muller JM, Erasmi H, Stelzner M, Zieren U, Pichlmaier H.

Surgical therapy of oesophageal carcinoma. Br J Surg 1990;77:845–57.

2. Akiyama H, Tsurumaru M, Udagawa H, Kajiyama Y. Oe-sophageal cancer. Curr Probl Surg 1997;34:765– 834.

3. Barr H, Krasner N. Prospective quality-of-life analysis afterpalliative photoablation for the treatment of malignant dys-phagia. Cancer 1991;68:1660 – 4.

4. Loizou LA, Rampton D, Atkinson M, Robertson C, Bown SG.A prospective assessment of quality of life after endoscopicintubation and laser therapy for malignant dysphagia. Can-cer 1992;70:386 –91.

5. van Knippenberg FCE, Out JJ, Tilanus HW, Mud HJ, HopWCJ, Verhage F. Quality of life in patients with resectedoesophageal cancer. Soc Sci Med 1992;35:139 – 45.

6. O’Hanlon D, Harkin M, Daya K, Sergeant T, Hayes N, GriffinSM. Quality of life assessment in patients undergoing treat-ment for oesophageal carcinoma. Br J Surg 1995;82:1682–-5.

7. Zieren HU, Jacobi CA, Zieren J, Muller JM. Quality of lifefollowing resection of oesophageal carcinoma. Br J Surg1996;83:1772–5.

8. Bamias A, Hill ME, Cunningham D, Norman AR, Ahmed FY,Webb A, et al. Epirubicin, cisplatin, and protracted venousinfusion of 5-fluorouracil for oesophagogastric adenocarci-noma. Cancer 1996;77:1978 – 85.

9. Webb A, Cunningham D, Scarffe JH, Norman A, Joffe JK,Hughes M, et al. Randomised trial comparing epirubicin,cisplatin and fluorouracil versus fluorouracil, doxorubicinand methrotrexate in advanced oesophagogastric cancer.J Clin Oncol 1997;15:261–7.

10. Farndon MA, Wayman J, Clague MB, Griffin SM. Cost-effec-tiveness in the management of patients with oesophagealcancer. Br J Surg 1998;85:1394 – 8.

11. Girling DJ. British Medical Research Council protocol OE02:randomised controlled clinical trial of surgery with or with-out preoperative chemotherapy in the treatment of resect-able cancer of the oesophagus. Cambridge, England: CancerTrials Office, 1993.


12. Aaronson NK, Ahmedzai S, Bergman B, Bullinger M, Cull A,Duez NJ, et al. The European Organization for Research andTreatment of Cancer QLQ-C30: a quality of life instrumentfor use in international clinical trials in oncology. J NatlCancer Inst 1993;85:365–76.

13. Blazeby JM, Alderson D, Winstone K, Steyn R, Hammerlid E,Arraras J, et al. Development of a EORTC questionnairemodule to be used in quality of life assessment for patientswith oesophageal cancer. Eur J Cancer 1996;32:1912–7.

14. Blazeby JM, Conroy T, Alderson D, Farndon JR, Arraras J,Hammerlid E, et al. An international field study to test theEORTC QLQ-C30 and oesophageal cancer module (EORTCQLQ-OES24) among patients with oesophageal cancer.EORTC Protocol No. 15961/40973, 1997.

15. Vickers J, Alderson D. Oesophageal cancer staging usingendoscopic ultrasonography. Br J Surg 1998;85:994 – 8.

16. Vilmann P, Hancke S, Henriksen FW, Jacobsen GK. En-dosonographically guided fine needle aspiration biopsy ofmalignant lesions in the upper gastrointestinal tract. Endos-copy 1993;25:523–7.

17. Watt I, Stewart I, Anderson D, Bell G, Anderson JR. Laparos-copy, ultrasound and computed tomography in cancer ofthe oesophagus and gastric cardia: a prospective compari-son for detecting intra-abdominal metastases. Br J Surg1989;76:1036 –9.

18. Hopwood P, Stephens RJ, Machin D, for the MRC LungCancer Working Party. Approaches to the analysis of qualityof life data: experiences gained from a Medical Research

Council Lung Cancer Working Party palliative chemother-apy trial. Qual Life Res 1994;3:339 –52.

19. Curran D, Molenberghs G, Fayers PM, Machin D. Incom-plete quality of life data in randomized trials: missing forms.Stat Med 1998;17:697–709.

20. Gelber RD, Cole BF, Gelber S, Goldhirsch A. Comparingtreatments using quality-adjusted survival: the Q-TWiSTmethod. Am Stat 1995;49:161–9.

21. Weinberger M, Oddone EZ, Samsa G, Landsman P. Arehealth related quality of life measures affected by the modeof administration? J Clin Epidemiol 1996;49:135– 40.

22. Blazeby JM, Williams MH, Brookes ST, Alderson D, FarndonJR. Quality of life measurement in patients with oesophagealcancer. Gut 1995;37:505– 8.

23. Earlam R, Cunha-Melo JR. Oesophageal squamous cell car-cinoma. 1. A critical review of surgery. Br J Surg 1980;67:381–90.

24. Mellow MH, Pinkas H. Endoscopic laser therapy for malig-nancies affecting the esophagus and gastroesophageal junc-tion. Arch Intern Med 1985;145:1443– 6.

25. Vacanti CJ, van Houten RJ, Hill RC. A statistical analysis ofthe relationship of physical status to post operative mortal-ity in 68,388 cases. Anesth Analg 1970;49:564.

26. Hermanek P, Sobin LH. Digestive system tumours. In: Her-manek P, Sobin LH, editors. International Union AgainstCancer. TNM classification of malignant tumours. Berlin:Springer-Verlag, 1992.


a prospective longitudinal study examining the quality of life of patients with esophageal carcinoma

Documents