An autopsy study from the time when phenothiazine medications were new, roughly 50 years ago, provided a view of death by natural causes (Hussar 1966). Autopsy data from 1,275 chronically hospitalized patients with schizophrenia in a Veterans Affairs (VA) hospital who died at age 40 or older found that heart disease and cancer were the most common causes of death (similar to findings in the general 40+ population of that time). Pneumonia was somewhat overrepresented as a cause of death. Undetermined causes and aspiration of food were among the top eight causes of death in the patients with schizophrenia but not in the general population. The authors speculated that phenothiazines could be responsible, although they noted that other investigators at the time found no association between the new medications and death rates.
Harris and Barraclough (1998) reviewed mortality rates in cohorts of patients with various mental disorders between 1966 and 1995. Cohorts of patients with schizophrenia had an SMR (x 100) of 137 for natural causes, with significant increases in expected deaths from a variety of medical conditions (infectious, endocrine, circulatory, respiratory, digestive, and genitourinary disorders). Several subsequent studies have found a somewhat higher SMR for death by natural causes, with values of 232 (x 100, 95% CI 176-300) in one report (Brown et al. 2000), and 2.0 in males and 1.9 in females (95% CI 1.8-2.2 for males, and 1.8-2.0 for females) in another (Osby et al. 2000a). In a separate analysis of their data, the latter authors found that the SMR related to natural causes of death increased when rates were determined by 5-year periods (males increasing from 1.7 in 1976-1980 to 4.4 in 1991-1995, P=0.02; females showing an increase from 1.7 to 2.1 over the same interval, P=0.04) (Osby et al. 2000b).
Reviewing individual studies for increased risk of specific disorders is somewhat problematic. Many studies report categories of diseases from the International Classification of Diseases (ICD) rather than specific disorders. Also, findings vary between studies, perhaps because the SMR will be greatly influenced by a small number of observed cases when the expected rate is low. Given that reviews of studies, some including institutionalized patients, will present increased risks of certain diseases (infectious diseases, for instance), examining individual studies may be helpful in understanding risks for more recent periods. The failure to demonstrate increased risk of cancer when all causes of cancer are aggregated has led to much speculation. The persistence and increase in both deaths from and risk for disorders such as cardiovascular disease and diabetes is also important.
When all types of cancer are included, mortality for patients with schizophrenia ranges from very similar to nonsignificantly elevated over expected. Specific results include SMR of 100 (x 100, meaning that the number observed was the same as the number expected) (Harris and Barraclough 1998), 1.1 (Newman and Bland 1991), 1.1 for males and 1.3 for females (Osby et al. 2000a), 1.46 (Brown et al. 2000), and 1.4 (Allebeck 1989). A decreased rate of lung cancer in males has been noted to offset increases in other types of cancer, such as breast cancer in females (Mortensen and Juel 1990) or digestive cancers (Newman and Bland 1991). In the latter study, the SMR for lung cancer was 0.7 (not significantly decreased) and for digestive cancer was 1.7 (P<0.05). Limited access to cigarettes for institutionalized patients has been raised as a possible explanation for reduced rates of lung cancer (Harris and Barraclough 1998).
To investigate the reported decrease in mortality from lung cancer, Masterson and O'Shea (1984) looked at rates of smoking in 100 chronic inpatients with schizophrenia in about 1984 and at cause of death in 122 patients who died in the same hospital between 1974 and 1983. The proportion of deaths from respiratory cancers was not significantly different from that for the general population (3.3% vs. 4.3%). The authors expected a higher rate of respiratory cancers in the patients with schizophrenia because the rate of smoking was higher in patients compared with the general public and because the rate of death from pneumonia (presumably linked to smoking) accounted for more deaths in patients with schizophrenia compared with the general population (18.8% vs. 6.5%, P<0.001). The authors speculated that the failure to see an increase in lung cancer deaths suggested antitumor activity by phenothiazine medication. However, an alternative explanation may be that patients die from other causes such as cardiovascular disease before they reach the expected age of death from lung cancer. Despite this possible antitumor activity, breast cancer caused proportionately more deaths in the schizophrenic patients (8.3% vs. 3.4% in the general population, P<0.04). The authors also speculated that being nulliparous and having elevated prolactin, both associated with schizophrenic illness and treatment, while practicing inadequate breast care (insufficient self-examinations and incomplete reporting of lumps) could have caused the increase in breast cancer.
Two incidence studies on cancer in patients with schizophrenia further demonstrate varying results between geographic locations for different types of cancer. The results varied between cohorts in the World Health Organization's three-cohort study (Gulbinat et al. 1992). The Danish cohort showed decreased rates of cancer (standardized incidence ratio [SIR] 0.67, 95% CI 0.60-0.75), the Nagasaki cohort revealed increased risk of cancer (male SIR 1.43, not significant, with 95% CI 0.85-2.26; female SIR 1.67, 95% CI 1.17-2.44), and the Honolulu cohort overall demonstrated no unusual cancer incidence (females of Japanese origin did have elevated risk, with an SIR of 1.73, 95% CI 1.03-2.74). Different patterns of risk were seen between cohorts when the data were analyzed by site of malignancy. The Danish group had low rates of lung, prostate, cervical, and ovarian cancers, whereas breast cancer was increased in the Nagasaki and Honolulu Japanese females and cervical cancer in the Nagasaki females. The authors concluded that patients with schizophrenia do not demonstrate any consistent increase or decrease in risk for cancer. In contrast, the authors of a records linkage study between the Finnish cancer and psychiatric illness registries found a modest but significantly elevated SIR for cancer (1.17, 95% CI 1.09-1.25) in patients with schizophrenia (Lichtermann et al. 2001), with the incidence of lung cancer twice as high in patients with schizophrenia (SIR 2.17, 95% CI 1.78-2.60). They also noted that the incidence of cancer of the gallbladder was elevated (SIR 2.07, 95% CI 1.03-3.70) and wondered about the role of obesity and poor diet in patients with schizophrenia, given that these are risk factors for gallstones, and hence possibly for gallbladder cancer. The authors did not find a significantly increased incidence of breast cancer (SIR 1.15, 95% CI 0.98-1.34) but did find a higher rate of uterine cancer (SIR 1.75, 95% CI 1.19-2.48), suggesting obesity, low levels of physical activity, and lower parity as possible causes. The inconsistency of these findings with the other cancer incidence and mortality studies in schizophrenia dictates caution in accepting these as correct. For instance, still another study suggests increased risk of breast cancer. Breast cancer was found at a higher rate on mammograms done for all female psychiatric patients at the Buffalo Psychiatric Center compared with women who had been referred for mammograms to the same radiology service because of possible breast disease (Halbreich et al. 1996). Thus, one cannot draw a firm conclusion about whether any type of cancer is increased or decreased in patients with schizophrenia.
On the other hand, many other diseases do clearly contribute to increased mortality in schizophrenia, including respiratory, endocrine, and cardiovascular diseases. In their review, Harris and Barraclough (1998) found an SMR for respiratory disease in patients with schizophrenia of 214 in males (95% CI 191-238) and 249 in females (95% CI 223-278). Subsequent studies confirm this rate, with an SMR of 317 (95% CI 116-690) in one (Brown et al. 2000); 9.5 in males (95% CI 3.8-23.8) and 8.3 in females (95% CI 2.58-26.84) in a second (Joukamaa et al. 2001); and 3.2 in males (95% CI 2.4-4.2) and 2.7 in females (95% CI 2.1-3.4) in a third (Osby et al. 2000a).
Although an elevated SMR for endocrine disorders was not reported in the review article (Harris and Barraclough 1998), several separate studies have found an increase in SMR. In one, the SMR was 2.7 for males (95% CI 1.6-4.2) and 2.0 for females (95% CI 1.1-3.4) (Osby et al. 2000a). In the other, the SMR for all endocrine disorders was 1,166 (x 100, 95% CI 3792,721), with diabetes having an SMR of 996 (x 100, 95% CI 205-2,911) (Brown et al. 2000). A future increase in the SMR for diabetes might be anticipated, given the probable association between use of atypical anti-psychotic medications and diabetes (Muench and Carey 2001).
Cardiovascular disease causes a large number of deaths in both the general population and patients with schizophrenia. In the review by Harris and Barraclough, circulatory diseases caused 2,313 deaths out of the 5,591 deaths from natural causes, leading to an SMR in males of 110 (x 100, 95% CI 104-116) and in females of 102 (x 100, not significant, 95% CI 96-108) (Harris and Barraclough 1998). Although the increase in risk is relatively small, the number of excess deaths is substantial (230 more than expected compared with a total of excess deaths from natural causes of 1,250). Subsequent studies found higher cardiovascular mortality rates. In one, circulatory disease had an SMR of 249 (x 100, 95% CI 164-363), and more specifically, cardiovascular disease had an SMR of 187 (95% CI 102-298) (Brown et al. 2000). Only males demonstrated an increase in relative risk for cardiovascular deaths in another study (RR 2.92, 95% CI 1.65-5.2). In a third, the SMR for cardiovascular disease was 2.3 for males (95% CI 2.0-2.6) and 2.1 in females (95% CI 1.9-2.4) (Osby et al. 2000a).
The increase in SMR from cardiovascular disease seen in these studies compared with older reports might reflect an important trend of increasing risk for patients with schizophrenia relative to the comparison group. The data in the last study reported above were reanalyzed by 5-year periods from 1976 to 1995 (Osby et al. 2000b). The SMR, reflecting mortality rates above those in comparable people from the general population, rose sharply, with the SMR for males increasing from 1.7 in 1976-1980 to 8.3 in 1991-1995 (P<0.001) and for females from 1.7 to 5.0 over the same period (P<0.002). If in fact risks of mortality from schizophrenia are increasing, then increased attention to the risks of cardiovascular disease in schizophrenia such as weight gain, smoking, hyperlipidemia, hypertension, and diabetes will be needed (Fontaine et al. 2001; Muench and Carey 2001). (For further information, please see other chapters in this volume: Chapter 3, "Obesity in Patients With Schizophrenia"; Chapter 4, "Cardiovascular Illness and Hyperlipidemia in Patients With Schizophrenia"; Chapter 5, "Nicotine and Tobacco Use in Schizophrenia"; and Chapter 6, "Glucose Intolerance and Diabetes in Schizophrenia.")
These data demonstrate that a number of medical conditions, especially cardiovascular disease and diabetes, contribute to the increased mortality related to natural causes seen in patients with schizophrenia. These diseases, and others, might also be expected to cause increased morbidity.
Was this article helpful?
A Hard Hitting, Powerhouse E-book That Is Guaranteed To Change The Way You Look At Your Health And Wellness... Forever. Everything You Know About Health And Wellness Is Going To Change, Discover How You Can Enjoy Great Health Without Going Through Extreme Workouts Or Horrendous Diets.