The signal of suicide rates seen from a distance in patients with pancreatic cancer


  • Donna B. Greenberg MD

    Corresponding author
    1. Psychiatric Oncology, Massachusetts General Hospital Cancer Center, Boston, Massachusetts
    2. Department if Psychiatry, Massachusetts General Hospital, Boston, Massachusetts
    3. Department of Psychiatry, Harvard Medical School, Boston, Massachusetts
    • Massachusetts General Hospital, 55 Fruit Streeet, WRN 605, Boston, MA 02114
    Search for more papers by this author
    • Fax: (617) 726-5946

  • See related original article on pages 642-7, this issue.


Although suicide is rare in patients with pancreatic cancer, the higher rate of suicide signals that the anguish of clinical depression is common. Older men, more alone, with a history of substance abuse and depression, soon after diagnosis, are more at risk for suicide; attention to psychiatric syndromes is an integral part of cancer care.

The epidemiological satellite view of patients who were diagnosed with pancreatic adenocarcinoma between 1995 and 2005 reported by Tiraga et al in this issue of Cancer1 sheds light on the emotional course of a grave cancer. Pancreatic cancer long has been linked to depression and distress. The authors document a disproportionate standardized mortality ratio (SMR) for suicide >10 times that in the general population among patients with pancreatic cancer. What do we make of this signal?

The paradox is that suicide among patients with pancreatic cancer actually is quite rare. Most oncologists, even those who are caring for patients with pancreatic cancer alone, could believe that no patients kill themselves when faced with this diagnosis. Among greater than 36,000 patients, there were 30 who killed themselves, such that suicide was the recorded cause of death. Of those 30 patients, 28 were men, and they were more apt to be men without a spouse—single, divorced, or widowed. All but 4 patients were aged >60 years.

In the general United States population, suicide occurs most often in older white men.2 Substance abuse and major depressive disorder are major risk factors. Suicide attempts are all the more successful with firearms. Risk factors for depression are personal lifetime history, a family history of depressive disorder, prior suicide attempts, lack of social support, and stressful life events.

The diagnosis of cancer, needless to say, is a stressful life event. Among US patients who are diagnosed with cancer overall, the risk of suicide doubles,3 and older white men have the greatest risk, as in the general population. High relative risks of suicide are different for different tumor types. The highest rates in the most common tumors recorded in the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute from 1973 to 2002 were for tumors of the lung and bronchus (SMR, 5.74), stomach (SMR, 4.68), oral cavity and pharynx (SMR, 3.66), and larynx (SMR, 2.83)3; however, no rates were as high as the rate for pancreatic cancer.

Patients face the existential plight most acutely in the first 100 days after diagnosis.4 The newness of the diagnosis and the perception of advanced disease can be overwhelming and can influence the risk of suicide. This pattern also is observed among men with prostate cancer. Suicide was the cause of death for 148 of 342,500 patients who were diagnosed with prostate cancer between 1979 and 2004 who were followed in the SEER Program. Patients who were diagnosed with prostate cancer at a mean age of 70 years, similar to the mean age at diagnosis among patients with pancreatic cancer, had a greater risk of suicide during the first year, and especially during the first 3 months after diagnosis5: the SMR adjusted for age was 1.4 compared with all US men, which was much less marked (but still elevated) than that in patients with pancreatic cancer. The highest relative risk of suicide was during the first weeks after diagnosis. Similarly, the first week after diagnosis, followed by the first month, and the first year, had the highest relative risk of suicide in prostate cancer patients in Sweden.6 The risk, however, is mediated by the perception of advanced disease. In a more recent population-based cohort study in Sweden, the risk of suicide was twice as high among men with locally advanced or metastatic prostate cancer compared with men of the same age; but the relative suicide risk was not increased in men who had early, nonpalpable lesions identified by prostate-specific antigen testing.7

In Night Falls Fast, Jamison points out that suicide is a particularly awful way to die: The mental suffering that leads up to suicide is long, intense, and unpalliated; the death often violent and shocking; and the patient's suffering private and inexpressible, leaving family and friends to deal with furious loss and guilt.8 In that book, Jamison summarizes that the causes of suicide lie, for the most part, in an individual's predisposing temperament and genetic vulnerabilities, in severe psychiatric illness, and in acute psychological stress.8 Beyond the newness and gravity of a cancer diagnosis, this also is true in cancer hospitals.9

Like sepsis is to bacterial infection, suicide is the worst outcome of a serious, treatable condition—major depressive disorder. When depressed, a man feels small, worthless, and hopeless. Hope is a function of self-esteem; with cancer, he may feel that his potential for being someone who matters has been exhausted. This narrow, rigid thinking may be unrelated to the perception of others who love him. Depressive disorder darkens perception even for an individual who has no medical illness. Styron reported that the inward turned, mental anguish punctuated by bursts of anxiety that he experienced in clinical depression was almost indescribable.10

Pancreatic cancer has a serious prognosis. The overall 5-year survival rate has been <5%, and <20% of patients present with localized, potentially curable tumors.11 Of those who committed suicide, only 2 patients had localized disease.1 Those authors noted that clinical depression is more common in patients with pancreatic cancer than in other cancers. Depression and pancreatic cancer are associated in the general population; in men, clinical depression seems to precede pancreatic cancer more than it does other gastrointestinal malignancies.12 Among patients with newly diagnosed pancreatic cancer in 1 modern series of 130 patients, 38% of patients had elevated depression scores on the Beck Depression Inventory (BDI). Pain also was common and was severe in 29% of patients and moderate in 37% of patients.13

Individuals with depression and pain are more apt to have suicidal thoughts. The desire for hastened death, for death sooner than natural disease progression, and passive or active suicidal wishes in the setting of advanced disease are associated with depressive disorder and its feature of hopelessness. Chochinov reported that 8.5% of patients who were admitted to the hospital for terminal illness in Canada wished to die sooner. Those patients were more apt to be clinically depressed and were more apt to be hopeless.14, 15 In the ambulatory setting of a regional cancer center in Scotland, a recent survey of consecutive outpatients indicated that 8% of 3000 patients had thoughts of being better off dead or of hurting themselves in some way in the previous 2 weeks.16 Those patients who had suicidal thoughts were more apt to have clinically significant emotional distress, substantial pain, and, to a lesser extent, older age.

In studying the SEER data on pancreatic cancer, the relative risk of suicide was greater in patients who underwent surgery. In 8 patients, suicide followed an operative procedure. Five patients who committed suicide had been offered an operation but declined; they were among the 15% of patients who declined an operation. In 16 patients who committed suicide, no operation was recommended, presumably because of more advanced disease. The authors asked clinicians to closely observe the clinical presentation of those who undergo surgery. From this epidemiological distance, we do not know the detailed clinical course of these patients. They took the option for cure; they may have been more action oriented, more at risk of delirium, more worried about recurrence, more disappointed if disease was not cured, more intoxicated; or they may have had an extensive lifetime history of affective disorder. Care of depression and care of pain are part of humane medical care for serious illness. We stand by, treat what can be reversed, and use our skill for comfort.

Because patients feel hopeless when they are clinically depressed and may not speak up and because it seems natural that someone who has just been diagnosed with pancreatic cancer would be upset, the presence of the serious, relapsing syndrome of major depressive disorder may not be recognized. There is no blood test that provides a marker for the syndrome.17

The medical staff do tend to review and treat what is abnormal. Using the BDI (a touch-screen, 13-item, self-report measure) as a laboratory test and informing the healthcare team of a score that suggests moderate or severe depression, researchers have alerted the cancer care team to the need for treatment. In an Australian multicenter study, McLachlan et al made a difference in the number of depressed patients who remained moderately or severely depressed by reporting BDI scores to the team. More patients were referred for psychosocial evaluation for high scores; at 6 months, those patients in the group who had such a depression report sent to the team were less depressed than patients in the control group.18 Collaborative care for clinical depression in patients with cancer at a regional cancer center in Scotland (typically, a 7-session intervention from specially trained cancer nurses with supervision from psychiatry) reduced symptoms of depression and anxiety over usual care with benefit that was sustained at 6 months and at 12 months.19 Successful treatment of depression, usually a combination of antidepressant medication and psychosocial intervention, can reduce the wish for hastened death in the seriously ill.20

Attention to psychiatric syndromes in patients with pancreatic cancer is an integral part of care. Men who are isolated and who have a history of substance abuse, depression, and suicide attempts are particularly at risk for suicide; however, depression itself is morbid and common and is mixed with torturous anxiety. These patients deserve the time and expertise for a clinical assessment of psychiatric disorder. Treatment with antidepressant medication and psychosocial intervention should be considered with a close eye on their effectiveness. The early months after diagnosis are the most important to target. While we work to improve the outcome in pancreatic cancer, let us pay attention to psychiatric anguish as an integral part of medical care. From the epidemiologic study by Tiraga et al, there is a signal.1 The early months after a diagnosis of pancreatic cancer should provoke regular monitoring for the presence or absence of depressive disorder and sustained collaborative care.


The author made no financial disclosures.