Yung KC, Piccirillo JF. The Incidence and Impact of Comorbidity Diagnosed After the Onset of Head and Neck Cancer. Arch Otolaryngol Head Neck Surg. 2008;134(10):1045-1049. doi:10.1001/archotol.134.10.1045
Copyright 2008 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2008
To investigate the incidence and prognostic impact of comorbidities diagnosed after the onset of head and neck cancer.
Retrospective review of medical records.
One hundred eighty-three patients diagnosed as having head and neck cancer at Washington University School of Medicine from January 1, 1997, through December 31, 1998.
Main Outcome Measures
We reviewed medical records for demographic, tumor, treatment, and comorbidity data. Comorbid ailments at diagnosis and last follow-up or death were coded using the Adult Comorbidity Evaluation–27.
Of the 183 patients, 53 (29.0%) were found to have a baseline comorbidity score of none; 58 (32.0%) of mild; 53 (29.0%) of moderate; and 19 (10.4%) of severe. At last follow-up or death, scores were none for 30 patients (16.4%), mild for 52 (28.4%), moderate for 43 (23.5%), and severe for 58 (31.7%). Comorbidity scores at baseline (P = .002) and last follow-up (P = .001) were associated with 5-year survival. The prognostic impact of comorbidity scores at baseline and last follow-up were analyzed using Cox proportional hazards analysis. Individual comorbid ailments after diagnosis included myocardial infarction, coronary artery disease/angina, solid tumor, psychiatric disease, chronic obstructive pulmonary disease, hypertension, and alcohol abuse.
These findings are consistent with previous research demonstrating that comorbidity at diagnosis is strongly correlated with prognosis. This study also showed that the burden of comorbidity changes after diagnosis. There appeared to be a prognostic gradient based on comorbidity severity at baseline and outcome despite adjusting for age, sex, and cancer stage. Additional studies with larger numbers of patients and longer follow-up should be performed to investigate the importance of comorbidities that arise after diagnosis and may be a result of treatment.
Patients with cancer (hereinafter referred to as cancer patients) often have other medical conditions in addition to their index cancer. The presence of other medical conditions is especially true for those with head and neck cancer, who are likely to have significant tobacco and alcohol use. These other medical conditions are referred to as comorbidities. Although a distinct feature from the cancer itself, comorbidity is an important attribute of the cancer patient.1- 3 Comorbidity directly influences the care of cancer patients, selection of initial treatment, and evaluation of treatment effectiveness. Recent studies have shown that comorbidity is an important feature of head and neck cancer.4- 6 The inclusion of comorbidity information will improve the value of cancer statistics and the care of cancer patients. The prognostic importance of comorbidity on survival has been shown for patients with cancers of the larynx,7 lung,8 breast,9,10 and rectum11,12 and those with leukemia.13,14 In these studies, comorbidity was found to be an important feature in estimating a patient's outcome. In some instances, comorbidity was more important to the care of the patient than the cancer itself. Comorbidity can interact with the index cancer to create a more lethal situation than that caused by the index cancer alone.
Previous studies examined the presence and prognostic impact of comorbidities at the time of diagnosis of head and neck cancer.6,15,16 However, it is expected that patients continue to develop comorbid illnesses after diagnosis and primary treatment. The outcomes for patients who survive the initial period after diagnosis and treatment may be more dependent on their comorbidities than on their malignant neoplasm.
We performed this study to examine the incidence and impact of comorbidities that occur after the diagnosis of head and neck cancer. The results of this type of research may provide more individualized prognostic information for patients who survive beyond the initial period of diagnosis and treatment.
We performed a retrospective review of the medical records of patients with head and neck cancer. Using the Oncology Data Services database of the Barnes-Jewish Hospital, St Louis, we identified 265 patients diagnosed as having squamous cell carcinoma (SCC) of the oral cavity, oropharynx, and larynx from January 1, 1997, through December 31, 1998, and treated at Washington University Medical Center (WUMC). Of these 265 patients, the medical records for 55 (20.8%) were unavailable. Of the remaining 210, 23 patients (11.0%) were excluded owing to lack of sufficient medical information to determine baseline comorbidity. These 23 patients included those who presented to WUMC for treatment of recurrence, secondary tracheoesophageal fistulas, or other reasons. Finally, among the remaining 187 patients, 4 (2.1%) were excluded secondary to lack of complete 5-year survival information. These patients were lost to follow-up mostly because of relocation to another state or country. The final study population consisted of the remaining 183 patients.
The research protocol was approved by the Human Studies Committee at WUMC. The patients' medical records were obtained from the Department of Otolaryngology medical records department and from the Health Information Management Department of Barnes-Jewish Hospital. The information included notes from clinic visits with the otolaryngologist, radiation oncologist, and medical oncologist (if applicable) and hospital admissions. In each patient, we determined demographic data, tumor site and stage, treatment given, tumor recurrence, date of last follow-up or death, and cancer status at the last follow-up or death. Comorbidities at the time of diagnosis of head and neck cancer and at the last follow-up were coded using the Adult Comorbidity Evaluation–27 (ACE-27), a validated comorbidity index.17
The ACE-27 is a 27-item comorbidity index for use with cancer patients; it grades specific diseases and conditions into 1 of 3 groups (grade 1, 2, or 3) according to the severity of organ decompensation and prognostic impact. After the patient's individual comorbid ailments are classified, an overall comorbidity score (none, mild, moderate, or severe) is assigned on the basis of the highest-ranked single ailment. In the case in which 2 or more grade 2 conditions occur in different organ systems or disease groupings, the overall comorbidity score is designated as severe.
Standard descriptive statistical techniques were used to describe the population, the incidence of comorbid ailments before and after the diagnosis of cancer, the relationship between the treatment received and development of new comorbidities, and survival outcomes. To determine whether observed differences were statistically significant, a χ2 test of significance was used in which significance was established at the level of P < .05. The independent prognostic impact of comorbidity scores at baseline and outcome, controlling for age, sex, race, and stage of tumor, was assessed using Cox proportional hazards analysis. The overall 5-year survival rate from the time of diagnosis was selected to be the primary outcome measure. Hazard ratios and 95% confidence intervals for key factors were determined. All statistical analyses were performed using SAS statistical software (version 9.1; SAS Institute Inc, Cary, North Carolina).
We reviewed the medical records of 183 patients who were diagnosed as having and were treated for SCC of the oral cavity, oropharynx, or larynx from January 1, 1997, through December 31, 1998. As described in Table 1, patients were predominantly white and male (84.2% and 71.6%, respectively). The mean (SD) age was 62.2 (12.2) years. Seventy patients (38.3%) had SCC of the larynx; 57 (31.1%) had SCC of the oral cavity; and 56 (30.6%) had SCC of the oropharynx. Thirty-four patients (18.6%) had stage I disease; 30 (16.4%) had stage II disease; 41 (22.4%) had stage III disease; and 78 (42.6%), stage IV disease (Table 1).
Most patients were treated with surgery only (37.2%) or a combination of surgery and radiotherapy (35.0%). The remaining patients were generally given radiotherapy only (10.9%) or a combination of chemotherapy and radiotherapy (12.0%). Only 3 patients (1.6%) were treated with chemotherapy only and 6 (3.3%) with surgery, radiotherapy, and chemotherapy (Table 1).
Twenty-seven comorbid ailments were evaluated with the ACE-27 instrument. Comorbidities that developed after the diagnosis of cancer of the oral cavity, oropharynx, or larynx included myocardial infarction, coronary artery disease/angina, solid tumor (excluding primary cancer, persistent disease, or recurrent disease), psychiatric disease, chronic obstructive pulmonary disease, hypertension, arrhythmias, stroke, diabetes mellitus, and alcohol abuse. The frequency and severity distribution of 10 selected key individual ailments at diagnosis and last follow-up are shown in Table 2. Although the ACE-27 instrument identifies many comorbidities in all the organ systems, these ailments were selected on the basis of the highest frequency at follow-up, with solid tumor and psychiatric disease having the highest incidence (21 and 16 patients, respectively).
The 20 patients who received radiotherapy only developed 21 new ailments for a mean of 1.05 (range, 0-5) new comorbidities per patient. The 68 patients who received surgery only developed 44 new ailments, for a mean of 0.68 (range, 0-5) new comorbidities per patient. The group of 3 patients who received chemotherapy only developed a single new comorbidity for a mean of 0.33 (range, 0-1) new comorbidities per patient. The group who received surgery and radiotherapy (n = 64) had a mean of 0.67 (range, 0-3) new comorbidities per person, whereas the one receiving chemotherapy and radiotherapy (n = 22) had a mean of 1.09 (range, 0-4) new comorbidities per person. The 6 patients who received surgery, radiotherapy, and chemotherapy developed 4 new ailments for a mean of 0.67 (range, 0-3) per person. The mean numbers of new comorbid ailments per patient within each of these groups were not significantly different (F = 1.06; P = .39). Because of the small population of some treatment groups, several combinations of group consolidation were attempted to narrow the field to 1 radiotherapy, 1 surgery, and 1 chemotherapy group. All possible combinations were used, such as radiotherapy as radiotherapy only; surgery as surgery only plus surgery and radiotherapy plus surgery, radiotherapy, and chemotherapy; chemotherapy as chemotherapy only plus chemotherapy and radiotherapy; radiotherapy as radiotherapy only plus radiotherapy and surgery plus radiotherapy, surgery, and chemotherapy; surgery as surgery only; chemotherapy as chemotherapy only, and so forth. However, in all instances, statistical testing did not show any significant difference between the mean numbers of new comorbid ailments.
The distribution of overall comorbidity severity at the time of diagnosis was none in 53 (29.0%), mild in 58 (31.7%), moderate in 53 (29.0%), and severe in 19 (10.4%). At the last follow-up or death, the distribution was none in 30 (16.4%), mild in 52 (28.4%), moderate in 43 (23.5%), and severe in 58 (31.7%). This change in the overall comorbidity scores was statistically significant (χ2 = 185; P < .001). At the 5-year follow-up, 89 patients (48.6%) were still alive. The mean (SD) duration of follow-up was 52 (34) months (range, 1-108 months).
We analyzed the association of survival with comorbidity severity at diagnosis and the last follow-up or death using Cox proportional hazards regression. Unadjusted comorbidity scores at baseline and outcome were associated with 5-year survival (P = .002 and P = .001, respectively). The model was then adjusted for age, race, sex, and tumor stage. The combined prognostic impact of mild, moderate, and severe comorbidity (relative to none) at baseline and follow-up is shown for the entire cohort (Table 3). More than a third of the cohort (36.1%) exhibited a change in overall comorbidity score. The adjusted hazard ratio (95% confidence interval [CI]) relative to none for no change from baseline to outcome was 2.6 (1.1-6.2) for mild, 2.8 (1.2-6.8) for moderate, and 6.7 (2.7-16.7) to severe. For the 6 patients with improvement from moderate to mild, the adjusted hazard ratio was 2.0 (95% CI, 0.4-9.4). For the group of patients whose scores worsened, the adjusted hazard ratio for changes from none to mild was 1.4 (95% CI, 0.4-5.3); for none to moderate, 1.7 (0.4-6.7); for none to severe, 3.4 (1.1-10.1); for mild to moderate, 2.7 (0.8-8.9); for mild to severe, 2.9 (1.1-7.9); and for moderate to severe, 3.3 (1.3-8.6).
Past studies examined comorbidities at the time of diagnosis of head and neck cancer. This study is unique in that it reports on comorbid ailments that develop after diagnosis and primary treatment. In this study, 39.3% of the patients had moderate to severe comorbidities at the time of diagnosis. A previous study by one of us (J.F.P.) showed that only patients with cancers of the lung and colorectum had a higher rate of comorbid illness than did patients with head and neck cancer, whereas patients with cancers of the prostate, breast, and gynecological sites had lower rates.18
The impact of comorbidity on survival has been shown to be significant even after controlling for other important factors such as age, sex, race, site of cancer, and tumor stage. The prognostic impact of comorbidity is believed to be a result of the physiological burden of chronic disease and its treatment.2,19 It has been considered whether the negative prognostic impact of severe comorbidity is actually caused by the use of less ideal or aggressive therapy. However, previous research demonstrated that for patients with head and neck cancer, the prognostic impact was found not to be a result of different treatment but rather an independent prognostic effect.4,20
There was an impressive increase in severity of comorbidity at the last follow-up or death. Patients with a comorbidity score of none decreased from 29.0% at diagnosis to 16.4% at outcome, whereas patients with a score of severe increased from 10.4% to 31.7%. Only 6 patients exhibited a decrease in comorbidity, from moderate to mild.
Cox proportional hazards regression modeling was used to test the independent contribution of the combination of scores at baseline and outcome on survival. There appears to be a prognostic gradient based on the combined severity of the scores at baseline and last follow-up in that the adjusted mortality risk generally increases with increasing levels of comorbidity. In the case of the 6 patients with moderate comorbidity scores that improved after diagnosis and treatment, the hazard ratio of 2.0 was lower than the 2.8 ratio for those whose scores remained moderate. This ratio is also lower than the 2.6 ratio of those 38 patients whose scores remained mild. It is unclear whether this is because of small sample sizes or whether the improvement in comorbidity score imparts an actual prognostic advantage.
We initially hypothesized that incidence of new comorbid ailments may be related to treatment. Surgical treatment places the patient at risk for immediate operative and postoperative morbidity and mortality, but once the patient has recovered from surgery, long-term systemic effects are rare. In contrast, radiotherapy and chemotherapy commonly have long-term effects, such as cardiovascular disorders or treatment-induced malignant neoplasms. In addition, patients are often offered radiotherapy and/or chemotherapy regimens if severe comorbid conditions preclude surgery. If the treatment populations are different and the long-term effects of the treatments are consistent, we may find a correlation between new ailments and treatment type. However, no relationship was found in this study between the type of treatment received and incidence of new ailments. It is possible that the mean duration of follow-up of 52 months is not a sufficient time to allow for treatment-induced illnesses to arise; the small sample size may also play a role.
Limitations of this retrospective study, in addition to the sample size and the limited duration of follow-up, include the fact that all patients were treated at only 1 institution, a large academic medical center. Treatment preferences of these otolaryngologists, radiation oncologists, and medical oncologists may be different from those at other institutions. Also, patients often travel long distances for the initial workup and treatment, returning to their local physicians for long-term care. These patients most likely account for a great percentage of those whose medical records were unavailable or who were lost to follow-up. Finally, the comorbidity scores were acquired from a review of the clinical records of the cancer specialists and the medical records from hospital admissions. More accurate comorbidity information may be gathered from the patient's primary care physician. Although this was not feasible for the present study, future research should explore this resource.
A significant number of patients continue to develop comorbid ailments after the initial diagnosis of head and neck cancer. In this study, patients developed psychiatric disease or a secondary malignant solid tumor after the onset of head and neck cancer with the highest incidence; whether this is a result of patient predisposition or an adverse treatment effect is unknown. However, physician awareness can assist in earlier diagnoses of these issues. Moreover, we believe that comorbidity data should be garnered both before and after diagnosis to provide the patient with more accurate prognostic information and to allow for a more accurate assessment of the impact and adverse effects of antineoplastic therapy.
Correspondence: Jay F. Piccirillo, MD, Clinical Outcomes Research Office, Department of Otolaryngology–Head and Neck Surgery, Washington University School of Medicine, 660 S Euclid Ave, St Louis, MO 63110 (PiccirilloJ@ent.wustl.edu).
Submitted for Publication: May 3, 2007; final revision received December 1, 2007; accepted January 5, 2008.
Author Contributions: Drs Yung and Piccirillo had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Yung and Piccirillo. Acquisition of data: Yung. Analysis and interpretation of data: Yung and Piccirillo. Drafting of the manuscript: Yung and Piccirillo. Critical revision of the manuscript for important intellectual content: Piccirillo. Statistical analysis: Yung and Piccirillo. Study supervision: Piccirillo.
Financial Disclosure: None reported.
Previous Presentation: This report was presented as a scientific poster at the American Academy of Otolaryngology–Head and Neck Surgery Annual Meeting; September 17-20, 2006; Toronto, Ontario, Canada.