|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 24, No 25 (September 1), 2006: pp. 4177-4183 © 2006 American Society of Clinical Oncology. DOI: 10.1200/JCO.2006.06.2901 Disadvantage of Men Living Alone Participating in Radiation Therapy Oncology Group Head and Neck Trials
From the Departments of Radiation Oncology and Population Sciences, Fox Chase Cancer Center; Statistical Headquarters, Radiation Therapy Oncology Group; Abramson Cancer Center, University of Pennsylvania, Philadelphia, PA; Department of Interdisciplinary Oncology, H. Lee Moffitt Cancer Center, Tampa, FL; Department of Radiation Oncology, M.D. Anderson Cancer Center, Houston TX; Department of Radiation Oncology, University of Alabama-Birmingham, Birmingham, AL; and Radiological Associates of Sacramento, Sacramento, CA Address reprint requests to Andre A. Konski, MD, MBA, MA, Department of Radiation Oncology, Fox Chase Cancer Center, 333 Cottman Ave, Philadelphia, PA 19111; e-mail: andre.konski{at}fccc.edu
Purpose: This study evaluated whether males without partners were disadvantaged for survival in Radiation Therapy Oncology Group (RTOG) head and neck cancer clinical trials. Methods: Patients treated on three RTOG trials were studied. The Cox proportional hazards model was used to determine if sex and the interaction between sex and marital/partner status were independent prognostic variables for overall survival controlling for Karnofsky performance status, tumor stage, nodal stage, primary site, and protocol treatment. Results: A total of 1,901 patients (1,509 men) were entered onto the three RTOG trials, with 1,822 (1,438 men) analyzable patients. Prognostic variables independent of disease-related variables for survival in multivariate analyses restricted to men were age, marital/partner status, and income. Conclusion: The apparent disadvantage of unpartnered men is striking, even after controlling for disease and other demographic variables. Possible explanations could easily be tested in observational studies, leading to evaluation of simple interventions to improve their outcome.
The effect on outcome of variables other than those classically associated with disease process has been investigated in patients with cancer and other diseases. Race and socioeconomic status have been shown to have a prognostic effect for patients with prostate and colorectal cancer.1-3 Race was not a consistently negative prognostic factor, however, in free and open-access health care systems. Johnstone et al4 reported that race appeared to confer a negative prognosis only in patients with advanced disease at presentation in an analysis of Department of Defense Center for Prostate Disease Research records. African American veterans with squamous cell carcinoma of the distal esophagus were found to have increased mortality compared with white veterans, but African American veterans with adenocarcinoma of the distal esophagus had similar survival to that of white veterans with adenocarcinoma of the distal esophagus.5 Marital and partner status also has been found to be a prognostic factor for all-cause mortality in cancer, as in other diseases.6-9 However, marital status and sex are frequently evaluated as independent main effects, rather than in interaction, despite evidence that the benefits of being married may be different for men versus women.9-12 The interpretation of such interaction effects, when they are found, takes precedence over main effects, and interaction effects may have different practical public health and clinical implications than the individually considered main effects.6,13 In a previous study, we showed that among patients with metastatic prostate and breast cancer, married men and women and single women receiving higher radiation doses (30 Gy) had significantly longer time to re-treatment than single men, and that in contrast to these other groups, there was no difference in re-treatment rates over time in single men receiving 30 v 8 Gy.14 We were interested in exploring whether the apparent disadvantage of men without partners extended to a population with curative disease. We therefore examined the main effects of sex and marital status on survival, with particular attention to the question of whether there was a disadvantage for men without partners.
Study and Patient Inclusion Criteria Male patients with locally advanced head and cancer treated on three Radiation Therapy Oncology Group (RTOG) clinical trials form the basis of this primary analysis, and female patients from these trials form the basis of secondary analyses. They were chosen because the sociodemographic data were collected on the same form (A5) and the treatment results have been reported previously.15-17 This form was completed by the patient, staff member, or others such as a family member. The eligibility criteria for entry differed. None of these trials used sex or marital status as a stratification variable before patients were randomly assigned to their treatment.
RTOG 9003
RTOG 9111
RTOG 9703 Table 1 shows the variables on the A5 form and their subdivisions. Income information was collected as interval data rather than as a specific value; thus, it was not possible to adjust the income to US dollars for the 139 Canadian participants and so the reported category was used. All but eight of the Canadian participants were entered between 1992 and 1997.
Prolonged RT was included as a study variable because it has been published previously that treatment delays adversely affect outcome in patients with head and neck cancer.18,19 The criteria for scoring delays were as follows. For RTOG 9003, the RT duration was classified as unacceptable for patients treated on RTOG 9003 arm 1, standard RT, if the elapsed days of treatment were  64 and 59 for arms 2, 3, and 4. For RTOG 9111, the RT duration was classified as unacceptable for patients treated on RTOG 9111 arm if the elapsed days of treatment were more than 63 for all three arms. For RTOG 9703, the RT duration was classified as unacceptable for patients treated on RTOG 9703 arms 1 and 3 if the elapsed days of treatment were more than 55, and more than 96 days for arm 2.
Statistical Methodology Treatment intensity was examined in two ways. The first was whether the patient received the standard, once-a-day RT schedule. Three treatment options in RTOG 9003 and one option in RTOG 9703 had an altered RT schedule, whereas all others had the standard RT schedule. The second was whether the patient received concurrent chemotherapy during RT. One treatment option in RTOG 9111 and all three treatment options in RTOG 9703 had concurrent chemotherapy, whereas the other arms did not. The OS rates were estimated using the Kaplan-Meier method and compared with the log-rank test.20,21 The multivariate Cox model was used to determine if any sociodemographic variables had prognostic impact on survival.19 Imputation of missing values for variables was performed with the Markov Chain Monte Carlo algorithm to minimize the potential bias from excluding patients from the analysis.22 Two SAS procedures (MI and MIANALYZE) were used to generate the imputed datasets and combine the results of analyses carried out on them (SAS/STAT software, SAS OnlineDoc 9.1.3; SAS Institute Inc, Cary, NC). Protocol was one of the variables used in the imputations. Imputation initially creates 10 distinct data sets. All variables listed in Tables 1 and 2 were evaluated as possible covariates for the Cox model with the exception of employment status and family/friends with cancer. They were all considered as dichotomous variables. Various cut points were evaluated. For the first imputed data set, a Cox model was fitted using a forward stepwise procedure. For the second data set, a second model was fitted. This process was then repeated until the 10th model was fitted. Then the covariates for the 10 fitted models were compared and seven of them were found to be significant, with similar hazard ratios (HRs). An eighth significant covariate, household precancer income, was also found to be significant but its cut point (eg, $15,000) differed among the models. The $15,000 interval was chosen because it appeared in the most models. Next, those eight covariates were used in the 10 imputed datasets to generate the HRs reported in Table 3. Imputation was then used to create another 10 distinct data sets, which were used to generate the HRs for the same eight covariates. They were virtually identical to those generated with the first 10 data sets.
Primary analyses are restricted to 1,438 men with follow-up information who met the original protocol eligibility: 59.4% came from RTOG 9003, 27.9% came from RTOG 9111, and 12.7% came from RTOG 9703. At the time of the analysis, 856 men had died. The median follow-up was 27.2 months (range, 0.13 to 117.2 months) for all patients and 49 months (range, 2.5 to 117.2 months) for surviving patients. Table 1 shows the pretreatment sociodemographic variables; Table 2 shows other pretreatment characteristics. Approximately half were currently married or had another live-in partner. Missing values were almost exclusively encountered with the sociodemographic variables. Of the 1,438 males analyzed, 1,047 patients (72.8%) had all sociodemographic data, 344 patients (26.1%) had some sociodemographic data, and 47 patients (3.3%) had no sociodemographic data. Sixty-four percent of patients with all data completed the A5 form, whereas 40% with some missing data completed it. Univariate survival comparisons stratified by protocol were made between patients with and without data for each variable (Table 4). In every instance but one, the deletion of the patients with missing data for a variable would significantly change the study population. The group of patients with missing data for a variable tended to have significantly higher percentage of patients with unfavorable Karnofsky performance score (KPS; 60 to 80) than patients with complete data with one exception of reported precancer income. Of 70 males with missing marital status, 54% had unfavorable KPS in contrast to 34% for males with marital status information. Of 345 patients without precancer income, 38% had unfavorable KPS in contrast to 34% for males with precancer income information. The distributions for three other unfavorable prognostic variables (T3-4, N3, age older than 60) between patients with and without information about a sociodemographic variable were found to be similar.
Univariate survival analyses of all but two variables listed in Tables 1 and 2 (family/friends with cancer and employment status) were performed. All of the variables were found to be significant, with high HRs (> 2.0) for KPS and N stage. In particular, the HR for marital status was 1.61 (95% CI, 1.4 to 1.84) for the combined trials. Its HRs for the three trials separately were as follows: 1.73 (95% CI, 1.47 to 2.04) for RTOG 9003; 1.44 (95% CI, 1.05 to 1.97) for RTOG 9111; and 1.14 (95% CI, 0.73 to 1.78) for RTOG 9703. Prolonged RT was associated with a significantly lower survival rate (HR, 1.45; P = .0039 starting from study day 120, which allows the patients sufficient time to complete RT). The final survival Cox model is listed in Table 3. Three sociodemographic variables (age, marital status, and income) had independent prognostic value. Their impact was not as great as those for the disease and patient general condition variables. Prolonged RT did not have independent prognostic value. There was a high degree of correlation between two variables, primary site and concurrent chemotherapy, in the final fitted survival model and the two trials (RTOG 9111 and 9703), respectively. When two protocol-related variables are substituted for primary site and concurrent chemotherapy, the estimated effects in the modified Cox model for marital status, T stage, KPS, N stage, and so on were almost identical. In Table 5, the survival results are listed by marital status. For survival in women, the associated HR of 1.07 (95% CI, 0.78 to 1.47) is close to one that suggests no difference between the two status groups. Figure 1 shows the survival for the four groups defined by sex and marital status without adjustment for other factors. It is striking that the males not currently married or having a live-in partner showed much poorer survival than the other three groups, which have similar outcomes.
We have reported previously on the effects of nontreatment-related variables on outcomes of RTOG 9003 patients, focusing on the impact of education level on treatment outcome.23 Multivariate analysis revealed education level was significant for predicting both OS and locoregional control when comparing attended college/technical school compared with all other education levels. Education level is a well-known proxy for income level, with individuals with higher education levels attaining higher income levels.24,25 In these analyses, we built on this work by adding patients from two additional RTOG trials. We found in this expanded sample that socioeconomic status was associated with better survival. Patients with household income levels before the diagnosis of cancer above $15,000 had improved survival compared with patients with income levels below $15,000. It may be hypothesized that patients with higher income may be in better baseline health to tolerate the rigors of head and neck cancer treatment or may have the financial resources to afford supportive care. However, in these analyses, we gave particular attention to men and their marital/partner status. Main effects of sex and marital status occurred for survival, but interpretation of these main effects must be tempered by the existence of an interaction effect. Examination of this effect revealed that there was a consistent and striking disadvantage of men who were not married or living with a partner. These unpartnered men had a disadvantage on entering the trials in terms of having later stage and lower KPS. However, additional deleterious effects on survival were observed even after controlling for these initial differences. Interpretation of the results of these analyses needs to take into account a number of considerations. First, although our results regarding the interaction between marital/partner status and sex are provocative, the effects of demographic variables were consistently less than those for disease and patient general condition variables. The virtue of considering demographic variables lies in the possibility that results could lead to suggestions for modifications of care that could improve survival, and not in the relative strength of influence of demographic factors. Second, the patients in these analyses were participants in clinical trials, such that the standardized treatment reduces the likelihood that differences in the treatment received was the primary source of the effect of not having a partner. However, the effects of men not being partnered were consistent from staging at entry to survival duration. It is possible that in the absence of the standardization of treatment and increased attention imposed by participation in a clinical trial, the unpartnered status of these men could interfere even more with the quality and intensity of care they received. Effects on survival of unpartnered status on mortality of men under conditions of routine care could thus be underestimated. Third, the combining of three RTOG clinical trials provided more than 1,800 patients with sufficient data for analyses. There were a sufficient number of deaths and representation of partnered versus unpartnered males to provide adequate statistical power. The combining of the trials also provided 384 women for analyses. However, they had 76% fewer deaths than the men. The estimated marital effect in the women for the three end points was less than that seen in men, and the effect was small (7%), which suggests that there may be no meaningful difference. The statistical power to detect a significant interaction effect between marital status and sex using the number of deaths observed and the estimated interaction effect was 0.30 and 0.46 for survival and local control, respectively. With such low statistical power, no definitive statement can be made about this interaction from this study. Fourth, we combined having a spouse with having an unmarried live-in partner, rather than simply considering legal marital status. The RTOG A5 form includes a question about married/other live-in status instead of a question only about whether the patient is married or not. Cohabitation status and not marital status may be the important variable to measure. Lund et al7 observed 1,265 men and women age 50, 60, and 70 years and analyzed the association of mortality with cohabitation status and marital status. They found cohabitation status was a stronger predictor of mortality than marital status. Finally, we examined the effects of marital/partner status, but we could not examine the effects of relationship quality for those who have a partner. It has been argued that some of the deleterious effects of unsupportive relationships on health may exceed the effects of not having a partner.26 A recent prospective study documented the negative effects of poor marital functioning on survival among patients with chronic heart failure.27 Consistent with the results of that study, there is other evidence that women may be more susceptible to the effects of poor quality in intimate relationships.10 There are a number of implications of these considerations. In the future, attention needs to be directed to both the availability of a spouse/live-in partner and the quality of that relationship. There may be some threshold of quality of an intimate relationship, below which the advantage of having that relationship is eliminated, particularly for women. Marital status has been reported to have a significant influence on treatment delay in an analysis of 70 patients with colorectal cancer in Germany.28 Married men with metastatic prostate cancer were found to have a slower rate of physical decline but a greater decline in emotional and social functioning than unmarried men.29 Unmarried men were found to have a greater decline in physical functioning. It was hypothesized by the authors that married men's slower rate of decline in physical functioning was due to the patients not wanting to be a burden on their families, whereas the greater decline in emotional and social functioning was due to their perceived dependence on family members and leaving loved ones.
Multiple regression analysis showed marital status to be a highly significant variable for general and breast cancer–specific emotional health in 302 women age Particularly in light of similar results of another recent study, our findings provide impetus for the development of interventions for male cancer patients without partners.14 Additional study of the marital status x sex interaction is warranted in other disease sites, but we cannot presume the means by which unpartnered status presents a disadvantage for men will be the same. Development and evaluation of appropriate compensatory interventions will depend on identifying what aspects of males not having a partner are relevant to the outcome of care to each particular cancer. Possible targets of interventions potentially could range from adherence, symptom management, and timely and effective response to emergent medical problems, to the more general provision of social support and structure to everyday life. One potential advantage of focusing on unpartnered status among men is that unlike disadvantages in survival associated with other demographic variables, it has the potential promise of having a remedy with educational or nursing interventions in the clinical setting.
The authors indicated no potential conflicts of interest.
Supported by the Pennsylvania Commonwealth Universal Research Enhancement (CURE) Program Grant No. ME-02-149. Presented at the 40th Annual Meeting of the American Society of Clinical Oncology, New Orleans, LA, June 5-8, 2004. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Wrigley H, Roderick P, George S, et al: Inequalities in survival from colorectal cancer: A comparison of the impact of deprivation, treatment, and host factors on observed and cause specific survival. J Epidemiol Community Health 57:301-309, 2003 2. Wudel LJ Jr, Chapman WC, Shyr Y, et al: Disparate outcomes in patients with colorectal cancer: Effect of race on long-term survival. Arch Surg 137:550-556, 2002 3. Godley PA, Schenck AP, Amamoo MA, et al: Racial differences in mortality among Medicare recipients after treatment for localized prostate cancer. J Natl Cancer Inst 95:1702-1710, 2003 4. Johnstone PA, Kane CJ, Sun L, et al: Effect of race on biochemical disease-free outcome in patients with prostate cancer treated with definitive radiation therapy in an equal-access health care system: Radiation oncology report of the Department of Defense Center for Prostate Disease Research. Radiology 225:420-426, 2002 5. Dominitz JA, Maynard C, Billingsley KG, et al: Race, treatment, and survival of veterans with cancer of the distal esophagus and gastric cardia. Med Care 40:I14-I26, 2002 (suppl 1)[Medline] 6. Yates BC: The relationships among social support and short- and long-term recovery outcomes in men with coronary heart disease. Res Nurs Health 18:193-203, 1995[Medline] 7. Lund R, Due P, Modvig J, et al: Cohabitation and marital status as predictors of mortality: An eight year follow-up study. Soc Sci Med 55:673-679, 2002[CrossRef][Medline] 8. Vercelli M, Lillini R, Capocaccia R, et al: Cancer survival in the elderly: Effects of socio-economic factors and health care system features (ELDCARE project). Eur J Cancer 42:234-242, 2006[CrossRef][Medline] 9. Lai H, Lai S, Krongrad A, et al: The effect of marital status on survival in late-stage cancer patients: An analysis based on surveillance, epidemiology, and end results (SEER) data, in the United States. Int J Behav Med 6:150-176, 1999[CrossRef][Medline] 10. Kiecolt-Glaser JK, Newton TL: Marriage and health: His and hers. Psychol Bull 127:472-503, 2001[CrossRef][Medline] 11. Litwak E, Messeri P: Organizational theory, social support and mortality rates: A theoretical convergence. Am Social Rev 54:49-66, 1993 12. House J, Umberson D, Landis K: Structures and processes of social support. Annu Rev Sociol 14:293-318, 1988[CrossRef] 13. Trief PM, Wade MJ, Britton KD, et al: A prospective analysis of marital relationship factors and quality of life in diabetes. Diabetes Care 25:1154-1158, 2002 14. Konski A, DeSilvio M, Hartsell W, et al: Continuing evidence for poorer treatment outcomes for single male patients: Re-treatment data from RTOG 97-14. Int J Radiat Oncol Biol Phys 63:S192, 2005 15. Forastiere AA, Goepfert H, Maor M, et al: Concurrent chemotherapy and radiotherapy for organ preservation in advanced laryngeal cancer. N Engl J Med 349:2091-2098, 2003 16. Fu KK, Pajak TF, Trotti A, et al: A Radiation Therapy Oncology Group (RTOG) phase III randomized study to compare hyperfractionation and two variants of accelerated fractionation to standard fractionation radiotherapy for head and neck squamous cell carcinomas: First report of RTOG 9003. Int J Radiat Oncol Biol Phys 48:7-16, 2000[CrossRef][Medline] 17. Garden AS, Harris J, Vokes EE, et al: Preliminary results of Radiation Therapy Oncology Group 97-03: A randomized phase II trial of concurrent radiation and chemotherapy for advanced squamous cell carcinomas of the head and neck. J Clin Oncol 22:2856-2864, 2004 18. Pajak TF, Laramore GE, Marcial VA, et al: Elapsed treatment days: A critical item for radiotherapy quality control review in head and neck trials—RTOG report. Int J Radiat Oncol Biol Phys 20:13-20, 1991[Medline] 19. Cox JD, Pajak TF, Marcial VA, et al: Interruptions adversely affect local control and survival with hyperfractionated radiation therapy of carcinomas of the upper respiratory and digestive tracts: New evidence for accelerated proliferation from Radiation Therapy Oncology Group Protocol 8313. Cancer 69:2744-2748, 1992 20. Silliman RA, Dukes KA, Sullivan LM, et al: Breast cancer care in older women: Sources of information, social support, and emotional health outcomes. Cancer 83:706-711, 1998 21. Mantel N: Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 50:163-170, 1966[Medline] 22. Allison P: Missing Data Analysis: Sage University Papers Series on Quantitative Applications in Social Science. Thousand Oaks, CA, Sage, 2001 23. Konski A, Berkey BA, Kian Ang K, et al: Effect of education level on outcome of patients treated on Radiation Therapy Oncology Group Protocol 90-03. Cancer 98:1497-1503, 2003 24. Akin JS, Garfinkel I: School expenditures and the returns to schooling. J Human Resources 12:460-481, 1977[CrossRef] 25. Mattila P: Determinants of male school enrollments: A time series analysis. Rev Econ Stat 64:244-251, 1982 26. Coyne JC, DeLongis A: Going beyond social support: The role of social relationships in adaptation. J Consult Clin Psychol 54:454-460, 1986[CrossRef][Medline] 27. Coyne JC, Rohrbaugh MJ, Shoham V, et al: Prognostic importance of marital quality for survival of congestive heart failure. Am J Cardiol 88:526-529, 2001[CrossRef][Medline] 28. Langenbach MR, Schmidt J, Neumann J, et al: Delay in treatment of colorectal cancer: Multifactorial problem. World J Surg 27:304-308, 2003[CrossRef][Medline] 29. Melmed GY, Kwan L, Reid K, et al: Quality of life at the end of life: Trends in patients with metastatic prostate cancer. Urology 59:103-109, 2002[CrossRef][Medline] 30. Ghori FY, Gutterman-Litofsky DR, Jamal A, et al: Socioeconomic factors and the presentation, management, and outcome of patients with differentiated thyroid carcinoma. Thyroid 12:1009-1016, 2002[CrossRef][Medline] 31. Cassileth BR, Walsh WP, Lusk EJ: Psychosocial correlates of cancer survival: A subsequent report 3 to 8 years after cancer diagnosis. J Clin Oncol 6:1753-1759, 1988 Submitted February 28, 2006; accepted July 12, 2006.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2006 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
