Cancer aggressiveness and mortality in men of exceptional age

Abstract

Listen

Translate article

11051 Background: Information on the characteristics of cancer in people ≥ 85 is limited, particularly in men. Methods: We evaluated the type, grade and extent of cancer among the 22,071 men in the Physicians’ Health Study by age at diagnosis (dx) (<65, 65–74, 75–84 and ≥85). All cases of cancer, deaths and cause of death were confirmed by medical record review. To investigate the relationship between age at dx and risk of cancer death, we matched newly diagnosed cancer patients to reference subjects by age and a modified Charlson comorbidity score. Participants were followed for all cause mortality. We estimated hazard ratios (HR) for death by age at dx using Cox proportional hazards models and adjusted for potential confounders. Results: Over a mean follow-up of 20.5 years, 5,623 incident cancers were confirmed. Prostate cancer remained the most common cancer across all age groups. Melanoma and lung cancer became less common with age, while unknown cancers and gastrointestinal cancers other than colorectal (other GI) became more common. There was no linear trend toward higher or lower grade across the four age groups for individual cancer types. For men ≥ 85 the frequency of metastatic cancer at dx increased for prostate (5.8% vs 14.6% p=0.01) and decreased for other GI tumors (63.8% vs 43.5% p=0.05). Cancer as a cause of death decreased among the entire cohort from 44.1% in men aged 55–64 to 20.5% in men ≥ 85, and among those with cancer it decreased from 93.6% to 52.8%. In the matched cohort analysis, the HR for death from all cancers combined declined markedly across categories of increasing age at cancer dx from 10.9 (95%CI:6.0–19.9) in men < 55 to 1.9 (95%CI:1.5–2.4) in men ≥ 85. There was a similar decline in the HR with increasing age for cancer death from lymphoma, melanoma, prostate and colorectal cancers, whereas the HR of lung, other GI and urinary tumors remained stable. Conclusions: In this prospective cohort of apparently healthy U.S. male physicians, characteristics of cancer in men ≥ 85 varied considerably with tumor type and may reflect changes in cancer detection or biology with age. Cancer specific mortality decreased markedly with increasing age of diagnosis for most cancers. This is likely explained by competing risks of death which outpace that of cancer, but may also suggest decreased cancer aggressiveness in advanced age. No significant financial relationships to disclose.