Rates of adult schizophrenia following prenatal exposure to the Chinese famine of 1959-1961

CONTEXT: Schizophrenia is a common major mental disorder. Intrauterine nutritional deficiency may increase the risk of schizophrenia. The main evidence comes from studies of the 1944-1945 Dutch Hunger Winter when a sharp and time-limited decline in food intake occurred. The most exposed cohort conceived during the famine showed a 2-fold increased risk of schizophrenia. OBJECTIVE: To determine whether those who endured a massive 1959-1961 famine in China experienced similar results. DESIGN, SETTING, AND PARTICIPANTS: The risk of schizophrenia was examined in the Wuhu region of Anhui, one of the most affected provinces. Rates were compared among those born before, during, and after the famine years. Wuhu and its surrounding 6 counties are served by a single psychiatric hospital. All psychiatric case records for the years 1971 through 2001 were examined, and clinical and sociodemographic information on patients with schizophrenia was extracted by researchers who were blinded to the nature of exposure. Data on number of births and deaths in the famine years were available, and cumulative mortality was estimated from later demographic surveys. MAIN OUTCOME MEASURES: Evidence of famine was verified, and unadjusted and mortality-adjusted relative risks of schizophrenia were calculated. RESULTS: The birth rates (per 1000) in Anhui decreased approximately 80% during the famine years from 28.28 in 1958 and 20.97 in 1959 to 8.61 in 1960 and 11.06 in 1961. Among births that occurred during the famine years, the adjusted risk of developing schizophrenia in later life increased significantly, from 0.84% in 1959 to 2.15% in 1960 and 1.81% in 1961. The mortality-adjusted relative risk was 2.30 (95% confidence interval, 1.99-2.65) for those born in 1960 and 1.93 (95% confidence interval, 1.68-2.23) for those born in 1961. CONCLUSION: Our findings replicate the Dutch data for a separate racial group and show that prenatal exposure to famine increases risk of schizophrenia in later life.     

Carcinoma of the Lung: The Changing Pattern of Mortality in Ontario

Study of the mortality rates for carcinoma of the lung in men in Ontario between 1931 and 1959 reveals a rise from 3.7 in the early 1930's to 26.7 per 100,000 in the late 1950's. Analysis of age-specific mortality rates in five-year cohorts (groups of men born within five-year periods) shows that (1) mortality rates in each cohort rise rapidly after the age of 40, ascending in the later years of life almost as a straight line; (2) each succeeding cohort experiences an appreciably higher mortality rate than the preceding one. Mortality rates in individual cohorts in Ontario are compared with those in England; the shape of the cohort curves, and the rate of the increase in mortality from cohort to cohort, are almost identical. However, the picture in Ontario appears to be lagging some 10 years behind. The mortality rates for men in Ontario born around the year 1890 are almost identical with those shown by men born in England around 1880. The rates for carcinoma of the lung will almost inevitably continue to rise in Ontario for at least the next 10 to 20 years.     

Deaths due to sexually transmitted diseases. The forgotten component of reproductive mortality

Reproductive mortality has three principal components: deaths related to pregnancy, contraception, and sexually transmitted diseases (STDs). The last component is usually overlooked. In 1955, deaths due to STDs constituted a minimum of 32% of all reproductive mortality in the United States; in 1965 and 1975, the percentages were 32% and 20%, respectively. Pelvic inflammatory disease and syphilis account for most deaths due to STDs. In 1979, the mortality rate due to pelvic inflammatory disease was 0.29 deaths per 100,000 women aged 15 to 44 years; the corresponding figure for syphilis was 0.17. If cervical cancer is viewed as an STD, then deaths due to this cause alone (approximately 6,800 per year) would far outnumber deaths due to all other reproductive causes combined. Surveillance of reproductive mortality in the United States should be expanded to encompass deaths due to STDs.     

Postneonatal deaths in North Carolina, 1959-1963

The data presented indicate that the disturbing upward trend in infant mortality in North Carolina has been arrested and possibly reversed during the 1959 through 1963 period. Information obtained from death certificates indicates that infections accounted for slightly more than half (52.4%) of the postneonatal deaths occurring in the study periods. The most common type of infection was influenza and pneumonia, followed by gastroenteritis and colitis, infective and parasitic disease, meningitis, and acute respiratory infections, in that order of frequency. Infections were responsible for a greater percentage of the postneonatal deaths among nonwhite (58.5%) than amon white infants (40.7%). the postneonatal death rate from infections was 13.4 for nonwhite infants and 2.2 for white infants. The next most common cause of postneonatal mortality -- congenital malformations -- was relatively more important in the white race, being responsible for approximately 25% of white deaths and only 6% of nonwhite deaths. I11 defined and unknown causes ranked 3rd in importance, with postneonatal death rates of 3.0 for nonwhite and .4 for white infants. Accidents, wich ranked 4th, were responsible for approximately 10% of the postneonatal deaths in each race. In both races, the risk of postneonatal death was greater in infants born to younger mothers, partiuclarly those under age 20. For the infants of mothers under age 15, the postneonatal death rate was 3 times as high as for those of 20-24 year old mothers. Beginning with age 20, the risk of postneonatal mortality decreases gradually as maternal age increases up to 35 years, when it begins to rise again in the white race. In nonwhite races, the decline continoues to age 40. Infants born to young mothers of nonwhite races suffer relatively higher postneonatal mortality than do their white counterparts. The postneonatal mortality rate is lowest for 1st born infants of both races. Among nonwhites, it is highest for the 2nd born; in the white race, it rises with each successive birth, with the exception of the 5th. Postneonatal mortality among very small white infants (those weighing less thatn 1500 gm at birth) was some 7 times that of infants weighing more than 2500 gm; it was even higher in nonwhite races being nearly 2 1/2 times that of the white group and appoproximately 4 times higher than the rate for nonwhite infants weighing more than 2500 gm at birth. The risk of postneonatal death for nonwhite infants born illegitimately was 1 1/2 times as great for those born in wedlock. Among white infants, the risk was almost twice as great for those born out of wedlock.     

Postvaccine era measles epidemiology.

Measles incidence in the United States was calculated from estimates of the United States Immunization Survey for the years 1965 to 1975. These data were examined according to birth cohorts and cross-sectionally by age to determine the effect of live measles virus vaccine on measles epidemiology. The results showed a high rate of infection for children aged 1 year, followed by a decrease in the infection rates for preschool children, with rates increasing agin at ages 5 and 6 years. It was hypothesized that an effective vaccine-induced herb immunity exists in preschool children until they are of school age. When incidence rates were based on susceptible children, the highest rates occurred in the school-aged populations.     

Down's syndrome in South Australia.

In a survey of Down's syndrome in South Australia, 921 persons, both living and deceased, were identified; 717 individuals with the disorder were living in South Australia. Cytogenetic confirmation of the diagnosis had been made in 774 cases. From 1955 to 1977, the over-all incidence of Down's syndrome at birth was found to be 1.175/1000 live births. The incidence of Down's syndrome was significantly lower over the last five years of this period than for the first 18 years; thus it appears that the incidence of Down's syndrome in South Australia is falling. Analysis of maternal age changes with time has not revealed any changes to the maternal age-specific rates for Down's syndrome, although the rate for mothers aged 25 years or younger appears to be falling. The proportion of Down's syndrome babies born to women aged 35 years or more has decreased from 65.7% for those born before 1950 to 30.4% for those born from 1975 to 1977; similarly, the median maternal age has fallen from 37.12 years to 28.25 years. Regression analyses of maternal age rates for Down's syndrome by single years have produced figures suitable for genetic counselling. A plea is made that Down's syndrome should become a notifiable condition.     

1960年(饥荒中期)出生人群成年后糖尿病患病率显著性增高

目的:本研究探讨生命早期食物短缺及营养不良对成年后糖尿病患病危险的影响,为国家制订糖尿病干预政策提供科学依据.方法:本研究运用重庆某医院体检中心资料,选取1959～1964年出生的4640例成人作为研究对象进行调查.其中将1959～1960年出生人群作为研究组,1962～1964年出生人群作为对照组,进行回顾性研究,比较两组人群平均血糖值和糖尿病的患病率以及分布状况.结果:本次调查的结果显示1960年出生人群的平均血糖(Blood slouse,BG)水平显著高于非饥荒年出生人群(P<0.05).饥荒年出生人群糖尿病患病率显著高于非饥荒年出生人群(P<0.05). 1960年出生人群糖尿病患病率最高为4.55%.1959、1960、1961年出生人群相对于非饥荒年出生人群的标化后相对危险度为3.33、3.48和3.45.结论:生命早期经历了食物短缺和营养不良等灾害可能会增加成年后患糖尿病的危险性.     

Decline in US childhood cancer mortality. 1950 through 1980

Cancer mortality among children in the United States, 1950 through 1979, as evaluated by death certificate diagnoses, revealed dramatic declines primarily in the second half of the 30-year interval. The numbers of deaths of persons younger than 15 years, 1965 through 1979, as compared with the number expected at 1950 rates, fell 50% for leukemia, 32% for non-Hodgkin's lymphoma, 80% for Hodgkin's disease, 50% for bone sarcoma, 68% for kidney cancer, and 31% for all other cancer. There were 17,411 fewer deaths from childhood cancer from 1965 through 1979 than expected at the 1950 rate. Leukemia mortality declined by 8,073 deaths and kidney tumor mortality by 2,393. In data subsequently received for 1980, the decline in rates persisted for leukemia and non-Hodgkin's lymphoma, but the rates for the other four cancer categories seem to have reached a plateau. The reduction in mortality is attributed to improved therapy.     

Family planning and public health in Georgia: an enlarged commitment

County clinics reached an estimated 4% of eligible women in 1964 in Georgia. Continuing high crude birth rate showed a large gap between those eligible and interested and those, particularly nonwhites, being served. Although family planning was offered in 133 of 157 counties, the statewide postpartum return rate was only 24% in 1963-4. The maternal mortality rate in Georgia in 1964 was 5.7/10000. Of the 57 maternal deaths, 14 were associated with abortion and 15 with parities of 5 or more. Infant mortality rates were also higher than national figures, 29.2/1000. In November, 1974 State Health Department officials and prominent doctors reorganized the program to offer newer, more modern methods, e.g. oral contraceptives and IUDs, summarized the terms of state subsidization, and enabled clinics to consider systematically all the medical requirements for use of modern methods. Because of cost efficiency and ease of use, the IUD was the mainstay of the program. County acceptance increased steadily since June 1965. By 1966, 96 counties had plans for expanding services. By the end of 1965, 2434 devices were inserted, 832 women accepted oral contraceptives, and 4649 women were using traditional methods.     

CLINICOP ATHOLOGICAL ANALYSIS OF CAUSES OF PERINATAL DEATH

The perinatal mortality rates and causes of deaths in our hospital within the three 5-year periods (1955-1959, 1976-1980, 1981-1985) were reported as well as the total number of births (16,846), deaths (457), and autopsies (393, autopsy rate 85.9%). The perinatal mortality for the three 5-year periods was 44.5%, 23.8%, and 17.2% respectively; it declined more significantly in 1981-1985 than in 1976-1980. Anoxia was the first cause of death for the three 5-year periods. Other causes in sequence in 1955-1959 were traumatic intracranial hemorrhage and pulmonary diseases, in 1976-1980 malformation and pulmonary diseases, and in 1981-1985 anoxia, pulmonary diseases and hyaline membrane disease. Results suggest that accurate analysis of causes of deaths depends on meticulous systematic fetal and neonatal autopsy, including macerated fetuses, extensive discussion by pathologists, obstetricians and neonatalogists, and indispensable placental examination.     

The contribution of mild and moderate preterm birth to infant mortality. Fetal and Infant Health Study Group of the Canadian Perinatal Surveillance System

CONTEXT: The World Health Organization defines preterm birth as birth at less than 37 completed gestational weeks, but most studies have focused on very preterm infants (birth at <32 weeks) because of their high risk of mortality and serious morbidity. However, infants born at 32 through 36 weeks are more common and their public health impact has not been well studied. OBJECTIVE: To assess the quantitative contribution of mild (birth at 34-36 gestational weeks) and moderate (birth at 32-33 gestational weeks) preterm birth to infant mortality. DESIGN, SETTING, AND PARTICIPANTS: Population-based cohort study using linked singleton live birth-infant death cohort files for US birth cohorts for 1985 and 1995 and Canadian birth cohorts (excluding Ontario) for 1985-1987 and 1992-1994. MAIN OUTCOME MEASURES: Relative risks (RRs) and etiologic fractions (EFs) for overall and cause-specific early neonatal (age 0-6 days), late neonatal (age 7-27 days), postneonatal (age 28-364 days), and total infant death among mild and moderate preterm births vs term births (at >/=37 gestational weeks). RESULTS: Relative risks for infant death from all causes among singletons born at 32 through 33 gestational weeks were 6.6 (95% confidence interval [CI], 6.1-7.0) in the United States in 1995 and 15.2 (95% CI, 13.2-17.5) in Canada in 1992-1994; among singletons born at 34 through 36 gestational weeks, the RRs were 2.9 (95% CI, 2.8-3.0) and 4.5 (95% CI, 4.0-5.0), respectively. Corresponding EFs were 3.2% and 4.8%, respectively, at 32 through 33 gestational weeks and 6.3% and 8.0%, respectively, at 34 through 36 gestational weeks; the sum of the EFs for births at 32 through 33 and 34 through 36 gestational weeks exceeded those for births at 28 through 31 gestational weeks. Substantial RRs were observed overall for the neonatal (eg, for early neonatal deaths, 14.6 and 33.0 for US and Canadian infants, respectively, born at 32-33 gestational weeks; EFs, 3.6% and and 6. 2% for US and Canadian infants, respectively) and postneonatal (RRs, 2.1-3.8 and 3.0-7.0 for US and Canadian infants, respectively, born at 32-36 gestational weeks; EFs, 2.7%-5.8% and 3.0%-7.0% for the same groups, respectively) periods and for death due to asphyxia, infection, sudden infant death syndrome, and external causes. Except for a reduction in the RR and EF for neonatal mortality due to infection, the patterns have changed little since 1985 in either country. CONCLUSIONS: Mild- and moderate-preterm birth infants are at high RR for death during infancy and are responsible for an important fraction of infant deaths. JAMA. 2000;284:843-849     

Maternal and Perinatal Mortality Associated with Cesarean Section in Alberta (1955-1959)

A province-wide study of perinatal mortality was initiated in Alberta (population 1,283,000) in 1955. The period 1955-1959 covered 182,028 total births and 4219 perinatal deaths of which 260 were from 3813 Cesarean sections.

The perinatal mortality rate in Cesarean-section births in rural hospitals (101.4 per thousand Cesarean births) was compared with that for urban hospitals (55.7 per thousand).

Examination of the indications for primary Cesarean section in which a perinatal death occurred showed that hemorrhage accounted for 54 out of 85 of these deaths in rural hospitals, and 49 out of 110 similar urban deaths. Of 33 perinatal deaths associated with elective repeat sections, 17 were of premature babies.

Eleven of the 85 maternal deaths during 1955-1959 were associated with Cesarean section, a maternal mortality rate of 28.8 per 10,000 Cesarean section births. Preventable factors were present in 8 of the 11 cases. Hemorrhage was the primary cause of death.

     

An Analysis of the Increase in Lung Cancer in Canada

Lung cancer mortality in Canada over the period 1936-1964 is reviewed and a forecast is presented of future trends in the death rates, based on cohort analyses. Since 1936 the annual increases in mortality have been greater among individuals over 65 years of age, but in this group no single five-year age-group has contributed the major part to the general increase. Cohort analyses show (a) that the rate of increase of lung cancer has been much slower in generations born after 1906, (b) that the actual death rate will rise more slowly in the future, and (c) that the death rate may become stable within 15 years.     

Time trends in autism and in MMR immunization coverage in California

CONTEXT: Considerable concern has been generated in the lay and medical communities by a theory that increased measles-mumps-rubella (MMR) immunization among young children may be the cause of an apparent marked increase in autism occurrence. OBJECTIVE: To determine if a correlation exists in secular trends of MMR immunization coverage among young children and autism occurrence. DESIGN, SETTING, AND PARTICIPANTS: Retrospective analyses of MMR immunization coverage rates among children born in 1980-1994 who were enrolled in California kindergartens (survey samples of 600-1900 children each year) and whose school immunization records were reviewed to retrospectively determine the age at which they first received MMR immunization; and of autism caseloads among children born in these years who were diagnosed with autism and were enrolled in the California Department of Developmental Services regional service center system. MAIN OUTCOME MEASURES: Measles-mumps-rubella immunization coverage rates as of ages 17 months and 24 months and numbers of Department of Developmental Services system enrollees diagnosed with autism, grouped by year of birth. RESULTS: Essentially no correlation was observed between the secular trend of early childhood MMR immunization rates in California and the secular trend in numbers of children with autism enrolled in California's regional service center system. For the 1980-1994 birth cohorts, a marked, sustained increase in autism case numbers was noted, from 44 cases per 100 000 live births in the 1980 cohort to 208 cases per 100 000 live births in the 1994 cohort (a 373% relative increase), but changes in early childhood MMR immunization coverage over the same time period were much smaller and of shorter duration. Immunization coverage by the age of 24 months increased from 72% to 82%, a relative increase of only 14%, over the same time period. CONCLUSIONS: These data do not suggest an association between MMR immunization among young children and an increase in autism occurrence.     

Pulmonary embolism in surgical patients: 1959-79

The Edinburgh surgical statistics (audit) have been analysed for the years 1959, 1964, 1969, 1974, and 1979 to determine the trends in pulmonary embolism in surgical patients who died. There was a total of 61,038 operations, 1528 postoperative deaths, 804 necropsies, and 158 reported pulmonary emboli. The incidence of embolism diagnosed clinically and at necropsy fell throughout the period. This fall held good after corrections for necropsy rates, prognosis, and proportions of major operations. Although the overall necropsy rate fell from 58% to 40%, in patients expected to have a good prognosis the rate rose from 68% to 75%. Necropsy-proved embolism in "good prognosis" patients fell from 0.5% to 0.15% per 100 major operations. The main reduction has taken place since most surgeons in the area adopted methods of prophylaxis against venous thrombosis, but a direct relationship is not proved by this study.     

Relationship of baseline serum cholesterol levels in 3 large cohorts of younger men to long-term coronary, cardiovascular, and all-cause mortality and to longevity

CONTEXT: Based on observational and interventional data for middle-aged cohorts (aged 40-64 years), serum cholesterol level is known to be an established major risk factor for coronary heart disease (CHD). However, findings for younger people are limited, and the value of detecting and treating hypercholesterolemia in younger adults is debated. OBJECTIVE: To evaluate the long-term impact of unfavorable serum cholesterol levels on risk of death from CHD, cardiovascular disease (CVD), and all causes. DESIGN, SETTING, AND PARTICIPANTS: Three prospective studies, from which were selected 3 cohorts of younger men with baseline serum cholesterol level measurements and no history of diabetes mellitus or myocardial infarction. A total of 11,017 men aged 18 through 39 years screened in 1967-1973 for the Chicago Heart Association Detection Project in Industry (CHA); 1266 men aged 25 through 39 years examined in 1959-1963 in the Peoples Gas Company Study (PG); and 69,205 men aged 35 through 39 years screened in 1973-1975 for the Multiple Risk Factor Intervention Trial (MRFIT). MAIN OUTCOME MEASURES: Cause-specific mortality during 25 (CHA), 34 (PG), and 16 (MRFIT) years of follow-up; mortality risks; and estimated life expectancy in relation to baseline serum cholesterol levels. RESULTS: Death due to CHD accounted for 26%, 34%, and 28% of all deaths in the CHA, PG, and MRFIT cohorts, respectively; and CVD death for 34%, 42%, and 39% of deaths in the same cohorts, respectively. Men in all 3 cohorts with unfavorable serum cholesterol levels (200-239 mg/dL [5.17-6.18 mmol/L] and >/=240 mg/dL [>/=6.21 mmol/L]) had strong gradients of relative mortality risk. For men with serum cholesterol levels of 240 mg/dL or greater (>/=6.21 mmol/L) vs favorable levels (<200 mg/dL [<5.17 mmol/L]), CHD mortality risk was 2.15 to 3.63 times greater; CVD disease mortality risk was 2.10 to 2.87 times greater; and all-cause mortality was 1.31 to 1.49 times greater. Hypercholesterolemic men had age-adjusted absolute risk of CHD death of 59 per 1000 men in 25 years (CHA cohort), 90 per 1000 men in 34 years (PG cohort), and 15 per 1000 men in 16 years (MRFIT cohort). Absolute excess risk was 43.6 per 1000 men (CHA), 81.4 per 1000 men (PG), and 12.1 per 1000 men (MRFIT). Men with favorable baseline serum cholesterol levels had an estimated greater life expectancy of 3.8 to 8.7 years. CONCLUSIONS: These results demonstrate a continuous, graded relationship of serum cholesterol level to long-term risk of CHD, CVD, and all-cause mortality, substantial absolute risk and absolute excess risk of CHD and CVD death for younger men with elevated serum cholesterol levels, and longer estimated life expectancy for younger men with favorable serum cholesterol levels. JAMA. 2000;284:311-318     

Recent trends in mortality due to chronic obstructive pulmonary disease (COPD) in Mexico, 1980-2002

BACKGROUND: Chronic obstructive pulmonary disease (COPD) is characterized by airflow limitation that is not fully reversible. Changes in classification have a major impact on reported mortality rates. METHODS: Between 1980 and 2002, 230,463 COPD cases were studied (age group 35-94 years); 134,579 men; 95,884 women. RESULTS: The crude mortality rate varied from 37 (1980) to 61 per 100,000 men (2002), (increase: 65%). The crude mortality rate for females increased from 27 (1980) to 43 deaths per 100,000 women (2002), (increase: 56%). The trends of the absolute values by birth year and age groups are higher as age increases. As age of death comes down and the birth cohort increases, the absolute values decrease. The Mexican states located in the northern and central areas present a higher risk for dying. There is an increment coefficient of 93 cases per increment year in males (age group 35-74), and 61 cases per increment year in females (age group 35-74 years). For the age group 35-94 years, the annual increase for males is 288 cases. These results were statistically significant, and the regression model was validated by residual analysis. CONCLUSIONS: The oldest cohorts of the studied population showed the highest COPD mortality absolute values. The geographic risk of dying from COPD is concentrated in two regions: a) the three Mexican states of higher economic income at the northern frontier to the U.S. and b) those Mexican states surrounding the main producer of tobacco (Nayarit).     

Improvement in outcome for very low birthweight children: apparent or real?

OBJECTIVE: To determine whether improvement in the survival rate of infants with a birthweight of less than 1501 g was accompanied by an increase in the rate of neurological impairment or disability among the survivors. DESIGN, SETTING AND PATIENTS: Two cohorts of consecutive very low birthweight infants (birthweight less than 1501 g) in one tertiary perinatal centre were followed prospectively to eight years of age; for both cohorts, comparison groups of children of birthweight more than 1501 g were randomly selected from hospital births. INTERVENTIONS: The first cohort was born before the introduction of assisted ventilation (1966-1970), the second after assisted ventilation was well established (1980-1982). MAIN OUTCOME MEASURES: Comparisons between cohorts, at eight years of age, of the survival rates and the rates of severe sensorineural impairments and disabilities. RESULTS: The survival rate for very low birthweight infants to eight years of age almost doubled between these cohorts, from 37.1% to 67.8% (odds ratio [OR], 3.4; 95% confidence interval [CI], 2.5-4.7; chi 2 = 57.6; P much less than 0.0001). The biggest gain was the increase in non-disabled survivors at eight years of age, from 52.6% in the first cohort to 80.8% in the second cohort (OR, 3.5; 95% CI, 2.2-5.7; chi 2 = 26.7; P less than 0.0001). Furthermore, the rate of severe disabilities in survivors fell substantially, from 13.6% to 4.1% (OR, 0.31; 95% CI, 0.14-0.69; chi 2 = 8.3; P less than 0.01). Of specific impairments, the rate of severe sensorineural deafness fell substantially (3.2% to 0%: OR, 0.14, 95% CI, 0.02-0.81; chi 2 = 4.8; P less than 0.05), as did the rate of severe intellectual impairment (13.0% to 2.7%: OR, 0.25; 95% CI, 0.11-0.57; chi 2 = 10.7; P less than 0.002). Only the rate of cerebral palsy increased, but not significantly (2.6% to 6.8%; OR, 2.6; 95% CI, 0.89-7.6; chi 2 = 3.0). CONCLUSIONS: It has been possible to improve the survival rate of very low birthweight infants over time without increasing the number of severely disabled survivors. Whether the long-term outcome for these infants is continuing to improve with more recent advances in perinatal care remains to be determined.     

Maternal mortality in British Columbia in 1971-86.

We reviewed the 56 maternal deaths in British Columbia in 1971-78 and 1979-86 identified through the provincial Ministry of Health and compared the findings with data for the two preceding 8-year periods. The maternal death rate, defined as the number of deaths directly or indirectly related to pregnancy or delivery per 100,000 live births, decreased from 42 in 1955-62 to 5 in 1979-86. In the same interval the number of direct obstetric deaths decreased from 100 to 10 and the number of indirect deaths from 29 to 8. The number of deaths due to abortion decreased from 32 to 1. There was no change in the number of deaths among North American Indians. There was also no change in the number of deaths due to hypertension, most of which were avoidable; these findings have stimulated intensive teaching efforts to increase recognition and improve management of the problem. Review of maternal deaths can help identify deficiencies in the quality of care and can direct measures aimed at further reducing the maternal death rate.