American Journal of Gastroenterology ISSN 0002-9270


C 2007 by Am. Coll. of Gastroenterology doi: 10.1111/j.1572-0241.2007.01157.x
Published by Blackwell Publishing

CLINICAL REVIEWS

Time Trends of Ulcer Mortality in Non-European Countries

Amnon Sonnenberg, M.D., M.Sc.
Portland VA Medical Center and the Oregon Health and Science University, Portland, Oregon

AIMS: The aim of the present study was to follow the time trends of mortality from gastric and duodenal
ulcer in countries outside Europe and compare them with previous reports of ulcer mortality from
Europe and North America.
METHODS: Mortality data from eight different countries were analyzed, including Argentina, Australia, Chile,
Hong Kong, Japan, Mexico, Singapore, and Taiwan. The age-standardized death rates of individual
countries were followed from 1971 to 2004. Japan and Australia had mortality data for more than
50 yr that provided the opportunity to conduct a birth-cohort analysis.
RESULTS: The data from all countries were characterized by a decline in gastric and duodenal ulcer mortality.
Except for Hong Kong, mortality from gastric ulcer was higher than mortality from duodenal ulcer. In
Japan and Australia, mortality from gastric and duodenal ulcer displayed time trends that were
consistent with an underlying birth-cohort phenomenon. The risk of dying from gastric and duodenal
ulcer increased in consecutive generations born between the mid- and the end of the nineteenth
century and decreased in all subsequent generations. The peak mortality from gastric ulcer occurred
among generations born in 1875, whereas peak mortality from duodenal ulcer occurred among
generations born 10–20 yr later.
CONCLUSIONS: The ubiquitous decline in ulcer mortality in countries from different parts of the world is likely to be
associated with a worldwide decline in the occurrence of H. pylori infection. The events
accompanying the receding infection in developed countries must have similarly affected
populations exposed to increasing standards of hygiene.
(Am J Gastroenterol 2007;102:1101–1107)

INTRODUCTION
The decline in Helicobacter pylori (H. pylori) infection in
the general population has led to a decline in the occurrence
of gastric and duodenal ulcer in Europe and North America
(1, 2). The receding trends of gastric and duodenal ul-
cer among the populations of various European countries,
Canada, and the United States have been well documented
by innumerous published articles (2–7). The time trends of
peptic ulcer disease (PUD) outside Europe and North Amer-
ica are less well known. Although PUD is associated with a
low case-fatality rate, mortality has been proven in the past
to be a reliable morbidity parameter to predict time trends
of PUD (8). Mortality data are available for many countries,
which otherwise do not have any other statistics available
to follow disease trends over prolonged time periods. The
unequivocal outcome of disease-related death is generally
less subjective to bias of ascertainment or standards of health
statistics, which affect other morbidity parameters and render
their comparisons across different times and countries more
difficult. The present study was aimed at analyzing the time
trends of mortality from PUD in non-European countries and
test whether such trends would reveal similar patterns as pre-
viously described for European countries.
METHODS
Data Sources
Initially, mortality data from eight different non-European
countries were analyzed. Table 1 lists the individual coun-
tries and the time periods for which separate mortality data on
gastric and duodenal ulcer were available. In a second step,
mortality data of gastric and duodenal ulcer from England
and the United States were added to the analysis for com-
parison. To be included in the study, the number of deaths
from gastric ulcer and duodenal ulcer had to be broken down
by 5- or 10-yr age groups, individual years of death, and
the vital statistics had to cover a time period of 20 yr or
more. Deaths were recorded by cause according to the 6th
through 10th revision of the International Classification of
Diseases (ICD). The data were made available by the na-
tional statistical offices of the individual countries included
in the present study. The more recent data sets were supplied
as Excel spreadsheets, whereas the older data were supplied
as photocopies from the annual compilations of causes of
death. The resident populations of the eight countries, broken
down by 5-yr age groups, were also provided by the national
statistical offices.

1101
1102 Sonnenberg
Table 1. Countries Included in the Analysis
Country Population in 2000 Period of Available Data
Argentina 36,784,000 1980–2004
Australia 19,153,000 1950–2004
Chile 15,398,000 1982–2003
Hong Kong 6,665,000 1981–2004
Japan 126,923,000 1933–2004
Mexico 97,483,000 1985–2004
Singapore 3,263,000 1980–2005
Taiwan 22,277,000 1971–2004
England 52,947,000 1921–2004
United States 281,675,000 1946–2002
Data Analyses
In the analysis of time trends, mortality from gastric or duo-
denal ulcer was expressed as deaths per million living popu-
lation. The data from each 5-yr period were lumped together.
To adjust for the changing age distribution among different
countries and times, the death rates from consecutive 5-yr
periods were age-standardized by the method of direct stan-
dardization (9). The age distribution of the WHO standard
world population from 2000 served as external standard (10).
For applying the method of direct standardization, the age-
specific death rates had to be calculated separately for each
country and 5-yr time period. For instance, the number of
deaths from gastric ulcer among Mexicans aged 45–54 dur-
ing the time period 1991–1995 was divided by the number
of all residents in Mexico of the same age group and living
during the same time period, 1991–1995. The number of ul-
cer deaths per time period was treated as a Poisson variable.
The difference between two death rates was considered sta-
tistically significant if their 95% confidence intervals did not
overlap (11).
Only Japan and Australia had data on mortality from gas-
tric and duodenal ulcer for a consecutive time period of more
than 50 yr. For the birth-cohort analysis of mortality data from
these two countries, the age-specific death rates were calcu-
lated for consecutive 10-yr periods and 10-yr age groups. For
instance, the total number of deaths from 1971 until 1980
among subjects aged 45–54 was divided by the correspond-
ing number of the total resident population of the same age
group and living during the same time period when the deaths
occurred. The death rates were expressed per million living
population. The age-specific death rates were plotted against
the period of death as period-age contours and against the
period of birth as cohort-age contours. In the plots, each age
group is labeled by its central year, for instance, 20 indicat-
ing the age group 15–24 and 30 indicating the age group
25–34. The periods of death are labeled by the mid-year of
death, for instance 1975 instead of 1971–1980. Similarly, the
periods of birth are labeled by their mid-year of birth. The
age-standardized cohort mortality ratio (SCMR) was calcu-
lated as a summary statistic of the overall mortality associated
with each consecutive birth cohort (12). This method of in-
direct standardization adjusts for the changing contribution
of different age groups to the mortality of different cohorts.
The SCMR expressed as a percentage was plotted against the
period of birth.
RESULTS
Figure 1 shows the time trends of six countries with sepa-
rate data on gastric and duodenal ulcer mortality available
for the past 20–35 yr. Three countries were from Asia and
three countries from South America. The data from all six
countries were characterized by a decline in gastric and duo-
denal ulcer mortality. The difference between the first and
the last rate was statistically significant in all countries and
for both ulcer types. Except for Hong Kong, mortality from
gastric ulcer was higher than mortality from duodenal ulcer.
The gap between gastric and duodenal ulcer mortality was
largest in Mexico. It generally appeared larger in the three
countries from South America than Asia.
Japan had separate data available for gastric and duodenal
ulcer mortality since 1933 and Australia since 1950. These
were the only two countries with statistical data covering a
time period sufficiently long to subject them to an analysis by
birth cohorts. The left panel of Figure 2 contains the period-
age contours of gastric ulcer in Japan. Mortality from gastric
ulcer was much higher among the older than younger age
groups. To present the time trends of all age groups in a
single graph, a logarithmic y-axis was chosen. The period-
age contours appeared to be arranged in a fan-like pattern.
The fan-like pattern originated from the progressively steeper
decline in the time trends of mortality among consecutive age
groups. In the two oldest age groups, death rates rose during
the first three decades before starting a smooth decline. The
period-contour of the next youngest age group, dying at age
65–74 yr, showed a shorter initial rise followed by a more
pronounced decline than in the two oldest age groups. As the
age of death decreased, the decline in the period-age contours
became gradually steeper.
In the middle panel, the same age-specific death rates as
those depicted in the left panel of Figure 2 were plotted against
the period of birth. For instance, individuals who died aged
90 between 1935 and 2005 were born between 1845 and 1915.
Hence, the oldest period-age contour, running between 1935
and 2005 in the left panel of Figure 1, changed into a cohort-
age contour running between 1855 and 1915 in Figure 2.
Similarly, individuals who died aged 20 between 1945 and
2005 were born between 1925 and 1985. A vertical line per-
pendicular to the abscissa in the middle panel would intersect
the cohort-age contours of individuals born during the same
time period, but who died from gastric ulcer at a different age
and during different time periods. Plotted against the period
of birth rather than period of death, the death rates became re-
aligned as cohort-age contours, which covered a much longer
time of 140 yr (1845–1985) as compared to the 70 yr (1935–
2005) covered by the period-age contours. The conversion
from period-age to cohort-age contours pulled the individual
 Time Trends of PUD Mortality 1103

140 250

120 Taiwan GU Singapore

200

Deaths per million

Deaths per million

100

80 150

60 GU DU
100
40
50
20
DU
0 0
1970 1980 1990 2000 1980 1985 1990 1995 2000 2005

160 140

140 Hong Kong 120 Argentina

120
Deaths per million

Deaths per million

DU 100
100 GU
GU 80
80
60
60
DU
40
40

20 20

0 0
1980 1985 1990 1995 2000 2005 1980 1985 1990 1995 2000 2005

250 400

Chile Mexico
200
300 GU
Deaths per million

Deaths per million

150 GU
200
100

100
50 DU
DU

0 0
1980 1990 2000 1985 1995 2005

Figure 1. Age-adjusted death rates from gastric and duodenal ulcer in different countries.

age-specific curves apart but left their shapes unchanged. The
individual cohort-age contours aligned in a pattern that re-
sembled a hyperbola with an initial rise and a subsequent
decline associated with consecutive periods of birth. This
overall pattern of the cohort-age contours suggested that, ir-
respective of age at death, the highest mortalities from gastric
ulcer in Japan occurred among generations born around 1875.
The logarithmic scale of the y-axis allowed one to com-
pare the relative changes in consecutive period-age contours
within a single graph. Because the logarithmic plot tended to
compress the large changes in death rates among the older
age groups and expand small changes among the younger
age groups, it emphasized the recent decline at the expense
of partly concealing the initial rise. The underlying birth-
cohort phenomenon became more evident if the individual
cohort-age contours rates were summarized as an SCMR, as
displayed in the right panel of Figure 2. Every point of the
SCMR curve in Figure 2 represented an approximation of
the average death rate among individuals belonging to differ-
ent age groups but being born during the same time period.
As evidenced by the SCMR curve, the average risk of dying
from gastric ulcer increased among Japanese cohorts born
between 1855 and 1875 and declined in all subsequent gen-
erations born between 1885 and 1985.
Similar analyses as outlined above for gastric ulcer were
also performed for the Japanese data on mortality from duo-
denal ulcer (Fig. 3). Similarly to gastric ulcer, the period-
age contours of duodenal ulcer appeared to be arranged in
a fan-like pattern. The switch from period-age to cohort-age
contours was again accompanied by the emergence of a hy-
perbola with its peak located around 1885. The birth-cohort
pattern became even more obvious by plotting of the SCMR
curve as in the right panel of Figure 3. Compared with gastric
ulcer, mortality from duodenal ulcer peaked among Japanese
generations born 10 yr later. Since the turn of the century,
mortality from duodenal ulcer has continued to drop in a
seemingly exponential fashion.
In Australia, separate mortality data for gastric and duode-
nal ulcer were available since 1950. Figure 4 contains the
birth-cohort analysis for gastric ulcer, and Figure 5 con-
tains the birth-cohort analysis for duodenal ulcer. Similarly
to Japan, mortality from gastric and duodenal ulcer both dis-
played overall time trends that were consistent with an un-
derlying birth-cohort phenomenon. The risk of dying from
1104 Sonnenberg

10000 400

350

Standardized cohort mortality ratio

1000
Deaths per million living 300
90
250
100 80
70 200
60
10 150
50

40 100
1
30 50
20
0.1 0
1930 1950 1970 1990 1840
2010 1880 1920 1960 20001840 1880 1920 1960 2000
Period of death Period of birth Period of birth

Figure 2. Birth-cohort analysis of gastric ulcer mortality in Japan. Left panel: age-specific mortality rate at different years of death (logarithmic
scale). Middle panel: age-specific mortality rate at different years of birth (logarithmic scale). Right panel: cohort mortality ratio standardized
by age (linear scale).

gastric and duodenal ulcer increased in consecutive Aus-
tralian generations born between the mid- and the end of the
nineteenth century and decreased in all subsequent genera-
tions. The peak mortality from gastric ulcer occurred among
generations born in 1875, while peak mortality from duodenal
ulcer occurred among generations born 10–20 yr later. Fig-
ure 6 compares the temporal relationship between the SCMR
curves of gastric and duodenal ulcer. The curves from Japan
and Australia are the same as shown above in Figures 2–5.
Mortality data of gastric and duodenal ulcer from England
and the United States have been added for comparison. The
data from England and the United States were subjected to
the same birth-cohort analysis as the data from Japan and
Australia. In all four countries alike, the rise in mortality
from gastric ulcer preceded the rise in mortality from duode-
nal ulcer by about one to two decades.
DISCUSSION
Similarly to previous studies of ulcer epidemiology from Eu-
rope and North America (2–7), the present analysis of vital
statistics from South America and Asia revealed a marked
decline in the mortality from gastric and duodenal ulcer. The
decline persisted until the most recent time period and was
seen in all countries without exception. Mortality from gas-
tric and duodenal ulcer is usually associated with ulcers that
perforated into the peritoneal cavity or bled without effective
means to stop the bleeding. Smaller studies from individ-
ual medical centers have occasionally claimed that although
the overall ulcer prevalence has declined the occurrence of
ulcers complicated by bleeding or perforation has remained
unchanged or even increased (13–16). Such contention is not
supported by the present data, and these previous observa-
tions are likely to have stemmed from an underlying selection
bias. A larger proportion of patients with complicated than
uncomplicated ulcers are hospitalized and treated for PUD.
All time series have reported a proportional increase in the
fraction of older patients with ulcer complications, which is
consistent with the underlying birth-cohort phenomenon, as
depicted in Figures 2–5.
The maintenance of vital statistics, as well as other types
of health and social statistics, is usually one of the hallmarks
of a developed country with a well-functioning health-care
system. The selection of the countries in the present study
was, therefore, biased toward a group of wealthier and more

1000 250
Standardized cohort mortality ratio

200
100
Deaths per million living

90

80
150
70
10
60
100
50

1
40 50

30
20
0.1 0
1930 1950 1970 1990 1840
2010 1880 1920 1960 20001840 1880 1920 1960 2000
Period of death Period of birth Period of birth

Figure 3. Birth-cohort analysis of duodenal ulcer mortality in Japan. Left panel: age-specific mortality rate at different years of death
(logarithmic scale). Middle panel: age-specific mortality rate at different years of birth (logarithmic scale). Right panel: cohort mortality
ratio standardized by age (linear scale).
 Time Trends of PUD Mortality 1105

1000 250

Standardized cohort mortality ratio

100 90 200
Deaths per million living
80

10 70 150
60

1 50 100
40

0.1 30 50
20

0.01 0
1940 1960 1980 2000 2020
1850 1890 1930 1970 1850 1890 1930 1970
Period of death Period of birth Period of birth

Figure 4. Birth-cohort analysis of gastric ulcer mortality in Australia. The layouts of the panels are similar to those of the birth-cohort
analysis of gastric ulcer from Japan.
developed countries that were characterized by health sys-
tems of similar standards to most European and North Amer-
ican countries. The decline in mortality from gastric and duo-
denal ulcer may not apply to the health statistics of other less
developed countries outside Europe and North America. The
mortality statistics of the World Health Organization show
PUD mortality to be still high in many countries from Africa,
Asia, and South America (17). Unfortunately, the vital statis-
tics of these countries did not distinguish between gastric
and duodenal ulcer, but used only an abbreviated list of ICD
codes with a single code for all ulcer types. Moreover, data
were only available for single years rather than prolonged
observation periods.
The countries studied here are diverse in terms of their
age distributions, ethnic compositions, health-care services,
usage of antibiotics or nonsteroidal anti-inflammatory drugs,
and nicotine consumption. The methods of direct and indirect
standardization used in the present analysis adjust for vary-
ing age distributions of populations during different time pe-
riods and among different nations, but are unable to account
for the other potential variations. In spite of such potentially
confounding influences, with the exception of gastric ulcer
mortality in Mexico, the death rates from all countries were
of similar magnitude. The Mexican data may reflect patterns
of a country where large fractions of the population still lead
a rural life and live in poverty. Different from other coun-
tries, Hong Kong was characterized by higher death rates for
duodenal than gastric ulcer. The reasons for this behavior are
presently unknown. They are unlikely to reflect differences
in census data, record keeping, or diagnostic accuracy. All
countries of the present analysis have had a long tradition in
conducting decennial censuses. Gastric and duodenal ulcers
both represent well-known and easy-to-diagnose diseases that
have been described by medical textbooks for more than
100 yr. Moreover, all countries have used the ICD codes to
record deaths, and similar types of computerized systems
were utilized to accumulate statistical data.
The birth-cohort patterns that underlie the mortality statis-
tics from Japan and Australia are identical to those observed
in Europe, Canada, and the United States (2, 5, 8). Similarly
to the European and North American data, the rise and fall in
gastric ulcer preceded a similar behavior of duodenal ulcer by
a time gap of 10–20 yr. With data from more than 15 different
countries subjected to a birth-cohort analysis of peptic ulcer,
there has not been a single exception to this rule. Although
the birth-cohort analysis utilizes data that were generated only
after 1933 or 1950, it opens a window to look at the trends
of ulcer disease extending over a time period of 140 yr. The
birth-cohort analysis of Japan and Australia, thus, allows one
to follow the decline for a much longer time period than in the
other six countries. It shows unequivocally that the decline
in ulcer mortality started much earlier and continued much
longer than suggested by the data of Figure 1 alone. It also
reveals a unique and marked rise in the risk of dying from

1000 180

160
Standardized cohort mortality ratio

100 90
140
Deaths per million living

80
120
10 70
100
60
80
1 50
40 60

40
0.1 30
20
20

0.01 0
1940 1960 1980 2000 1850
2020 1890 1930 1970 1850 1890 1930 1970
Period of death Period of birth Period of birth

Figure 5. Birth-cohort analysis of duodenal ulcer mortality in Australia. The layouts of the panels are similar to those of the birth-cohort
analysis of duodenal ulcer from Japan.
1106 Sonnenberg

400 250

Japan Australia
Standardized cohort mortality ratio

Standardized cohort mortality ratio

350
200
300
GU
250 150

200

100
150

100
DU 50
50

0 0
1830 1870 1910 1950 1990 1850 1890 1930 1970
Period of birth Period of birth

250 250

England USA
Standardized cohort mortality ratio

Standardized cohort mortality ratio

200 200

150 150

100 100

50 50

0 0
1830 1870 1910 1950 1990 1850 1890 1930 1970
Period of birth Period of birth

Figure 6. Time trends of standardized cohort mortality ratio (SCMR) of gastric ulcer (GU, solid lines) and duodenal ulcer (DU, broken lines)
from four different countries. The curves for Japan and Australia are the same as shown above in Figures 2–5. Data from England and the
United States are shown for comparison.
gastric and duodenal ulcer among all generations born before
1875 or 1885.
The similarity among the birth cohort patterns of various
European countries, Canada, the United States, as well as
Japan and Australia, suggests that the underlying mechanisms
must have been ubiquitous and applied to many different
countries alike. The underlying mechanisms could not have
involved national characteristics or historical events that were
peculiar to one or a few individual countries, such as timing of
wars, famine, or other socioeconomic upheavals. They must
have been truly universal and be intimately associated with
the natural history of PUD and its underlying infection with
H. pylori.
Various clinical and epidemiologic studies have suggested
that peptic ulcer occurs preferentially in subjects who contract
the disease during childhood or adolescence (18). While early
acquisition of H. pylori shortly after birth is primarily associ-
ated with pangastritis and gastric cancer, first acquisition of
this infection during childhood and adolescence is more likely
to result in gastric and duodenal ulcer, respectively. There is
also epidemiologic evidence to suggest that the incidence of
H. pylori infection started to fall in the general population
after 1800 secondary to increasing standards of hygiene in
developed countries (1). When the infection was still ram-
pant, the majority of subjects became infected as toddlers.
As the infection receded in the general population, the frac-
tion of subjects who acquired H. pylori infection during late
childhood or adolescence increased among consecutive gen-
erations. The superimposition of a declining infection trend
and a rising age trend has resulted in a bell-shaped peak of ul-
cer occurrence among consecutive generations born between
1850 and 1950 (19). The short delay between the birth-cohort
patterns of gastric and duodenal ulcers is a reflection of the
more pronounced age delay in duodenal than gastric ulcer.
The interaction between increasing hygiene and delayed in-
fection could have been operative in different populations of
developing societies in European, as well as non-European
countries alike. The present analysis shows that time trends of
mortality from gastric and duodenal ulcer in non-European
countries mirror those observed in Europe and North America
when the same epidemiologic techniques are applied. Lon-
gitudinal cross-sectional studies using aggregate data have

limitations for exploring causal associations, however, and
other study designs would be needed to establish causal as-
sociations.
In conclusion, the present analysis shows that mortality
from gastric and duodenal ulcer has decreased in the devel-
oped countries of Asia and South America similarly to in
Europe and in North America. For the very few countries
with long time series of more than 50 yr, it can be demon-
strated that the time trends of ulcer mortality were shaped
by underlying birth-cohort patterns that are identical to the
ones found in European countries, Canada, and the United
States. The ubiquitous decline in ulcer mortality of peptic
ulcer in countries from different parts of the world is likely to
be associated with a worldwide decline in the occurrence of
H. pylori infection. The events accompanying the receding
infection in developed countries must have similarly affected
populations exposed to increasing standards of hygiene.
Reprint requests and correspondence: Amnon Sonnenberg, M.D.,
M.Sc., Gastroenterology, Portland VA Medical Center P3-GI, 3710
SW U.S. Veterans Hospital Road, Portland, OR 97239.
Received September 21, 2006; accepted January 9, 2007.
REFERENCES
1. Taylor DN, Parsonnet J. Epidemiology and natural his-
tory of Helicobacter pylori infection. In: Blaser MJ, Smith
PD, Greenberg HB, et al, eds. Infections of the gas-
trointestinal tract. New York: Raven Press, 1995:551–
63.
2. Sonnenberg A. Temporal trends and geographical vari-
ations of peptic ulcer disease. Aliment Pharmacol Ther
1995;9(Suppl 2):3–12.
3. Primatesta P, Goldacre MJ, Seagroatt V. Changing patterns
in the epidemiology and hospital care of peptic ulcer. Int J
Epidemiol 1994;23:1206–17.
4. Munnangi S, Sonnenberg A. Time trends of physician visits
and treatment patterns of peptic ulcer disease in the United
States. Arch Intern Med 1997;157:1489–94.
5. Cucino C, Sonnenberg A. The long-term time trends of ul-
cerative colitis and peptic ulcer are interrelated. Am J Gas-
troenterol 2002;97:2657–62.
6. El-Serag HB, Sonnenberg A. Opposing time trends of peptic
ulcer and reflux disease. Gut 1998;43:327–33.
7. Kang JY, Elders A, Majeed A, et al. Recent trends in hospital
admissions and mortality rates for peptic ulcer in Scotland
1982–2002. Aliment Pharmacol Ther 2006;24:65–79.
8. Susser M, Stein Z. Civilization and peptic ulcer. Lancet
1962;1:115–9.
Time Trends of PUD Mortality 1107
9. Kahn HA, Sempos CT. Statistical methods in epidemiology
(Monographs in epidemiology and biostatistics, Vol. 12).
New York: Oxford University Press, 1989:85–95.
10. Ahmad OB, Boschi-Pinto C, Lopez AD, et al. Age stan-
dardization of rates: A new WHO standard. GPE Discussion
Paper Series: No. 31, EIP/GPE/EBD, World Health Orga-
nization, 2000. Available at: http://www.who.int/ncd
surveillance/infobase/web/InfoBaseCommon/Help/
PopUpHelp.aspx?id=293&lang=EN. Accessed September
19, 2006.
11. Ingelfinger JA, Mosteller F, Thibodeau LA, et al. Biostatis-
tics in clinical medicine. New York: MacMillan Publishing
Company, 1983;150–6.
12. Gardner MJ, Osmond C. Interpretation of time trends in
disease rates in the presence of generation effects. Stat Med
1984;3:113–30.
13. van Leerdam ME, Vreeburg EM, Rauws EA, et al. Acute
upper GI bleeding: Did anything change? Time trend anal-
ysis of incidence and outcome of acute upper GI bleed-
ing between 1993/1994 and 2000. Am J Gastroenterol
2003;98:1494–9.
14. Bardhan KD, Williamson M, Royston C, et al. Admis-
sion rates for peptic ulcer in the Trent Region, UK, 1972–
2000. Changing pattern, a changing disease? Dig Liver Dis
2004;36:577–88.
15. Ohmann C, Imhof M, Ruppert C, et al. Time-trends in the
epidemiology of peptic ulcer bleeding. Scand J Gastroen-
terol 2005;40:914–20.
16. Post PN, Kuipers EJ, Meijer GA. Declining incidence
of peptic ulcer but not of its complications: A nation-
wide study in The Netherlands. Aliment Pharmacol Ther
2006;23:1587–93.
17. World Health Organization. Death and DALY estimates for
2002 by cause for WHO Member States. Available at: http://
www.who.int/entity/healthinfo/statistics/bodgbddeathdaly
estimates.xls. Accessed September 19, 2006.
18. Graham DY. Helicobacter pylori: Its epidemiology and its
role in duodenal ulcer disease. J Gastroenterol Hepatol
1991;6:105–13.
19. Sonnenberg A. Causes underlying the birth-cohort phe-
nomenon of peptic ulcer: Analysis of mortality data 1911–
2000, England and Wales. Int J Epidemiol 2006;35:1090–7.
CONFLICT OF INTEREST
Guarantor of the article: Amnon Sonnenberg, M.D., M.Sc.
Specific author contributions: Amnon Sonnenberg, M.D.,
M.Sc.: Study concept and design, acquisition of data, analy-
sis, writing of manuscript, and statistical expertise.
Financial support: None
Potential competing interests: None