888
Original article
Ubiquitous occurrence of birth-cohort patterns in inflammatory bowel disease Amnon Sonnenberg Background and aims The aim of the present study was to demonstrate the ubiquitous occurrence of the birthcohort phenomenon of inflammatory bowel disease among US whites and nonwhites, as well as males and females. Methods Mortality from Crohn’s disease and ulcerative colitis in the USA between 1950 and 2010 were analyzed to discern underlying birth-cohort patterns affecting both their time trends. Age-standardized cohort mortality ratio was used as a summary statistic to represent the overall mortality associated with consecutive birth-cohorts.
Conclusion The birth-cohort patterns indicate that exposure to two separate risk factors must have occurred in both diseases during an early period of life. In the USA, these exposures have changed over historical times similarly in both sexes and different ethnic groups. Eur J Gastroenterol Hepatol 26:888–893 © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. European Journal of Gastroenterology & Hepatology 2014, 26:888–893 Keywords: birth-cohort pattern, Crohn’s disease, death rates, ethnicity, time trends, ulcerative colitis
Results The cohort-age contours of Crohn’s disease aligned to form one hyperbola with an initial rise between 1865 and 1935 and a subsequent decline. This pattern was confirmed by the time trends of the corresponding standardized cohort mortality ratio values. In ulcerative colitis, the individual cohort-age contours also aligned into one hyperbola that appeared shifted towards earlier generations by about 30 years when compared with Crohn’s disease. Similar trends were observed in men and women or whites and nonwhites analyzed separately.
Portland VA Medical Center and the Oregon Health and Science University, Portland, Oregon, USA
Introduction
recorded by underlying cause of death according to the International Classification of Diseases. In its 6th (1950–1957) and 7th revision (1958–1967), Crohn’s disease and UC were coded as 572.0 and 572.2, respectively. In the 8th revision (1968–1978) they were coded as 563.0 and 563.1, respectively. In the 9th revision (1979–1999) they were coded as 555 and 556, respectively. Since the 10th revision (starting in 2000), Crohn’s disease and UC have been coded as K50 and K51, respectively. For the purpose of the present analysis, none of the additional four-digit or four-letter subcodes of the more recent 9th and 10th International Classification of Disease revisions were considered.
The time trends of mortality from inflammatory bowel disease (IBD) are influenced by an underlying birthcohort pattern. The risk of IBD rose among consecutive generations born during the 19th century, peaked among those born around the turn of the century, and then declined in all subsequent generations born during the 20th century [1]. This pattern is found similarly in the vital statistics of several different European countries, as well as the USA [2]. Because of its large population size, the US is particularly well suited to study such temporal patterns also within its different demographic strata. Using US vital statistics, the present analysis was aimed as an extension of the previous study to test whether the IBD birth-cohort phenomenon also applies to different race and sex groups.
Methods Data sources
Resident US population and number of deaths from Crohn’ disease and ulcerative colitis (UC), stratified by age, sex, and ethnicity, were made available by the US National Center of Health Statistics. Mortality data were available from 1950 until 2010; stratification by sex was available only since 1962. Deaths were 0954-691X © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
Correspondence to Amnon Sonnenberg, MD, MSc, Gastroenterology, Portland VA Medical Center P3-GI, 3710 SW U.S. Veterans Hospital Road, Portland, Oregon 97239, USA Tel: + 1 503 220 8262 x56679; fax: + 1 503 220 3426; e-mail: [email protected] Received 17 February 2014 Accepted 7 April 2014
Data analyses
Average annual age-specific death rates were calculated for consecutive 10-year periods and 10-year age groups. For instance, the total number of deaths from 1971 until 1980 among individuals aged 45–54 was divided by the corresponding number of the total resident population of the same age group and living during the same time period when the deaths occurred. The age-specific death rates were plotted against the period of birth as cohort-age contours. In the plots, each age group was labeled by its central year, for instance, 40 indicating the DOI: 10.1097/MEG.0000000000000118
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Birth-cohort patterns in IBD Sonnenberg 889
age group 35–44 and 50 indicating the age group 45–54. The periods of birth were labeled by the mid year of birth, for instance 1925 instead of 1916–1935. The analyses were performed separately for Crohn’s disease, UC, and both diseases analyzed jointly as IBD. An age-standardized cohort mortality ratio (SCMR) was used as a summary statistic of the overall mortality associated with each consecutive birth-cohort as previously described [3]. Briefly, the SCMR was calculated similarly to the standardized mortality ratio as the ratio of observed (O) over expected (E) number of deaths in each birthcohort: SCMR = O/E. The SCMR was expressed as a percentage and plotted against the period of birth. The numbers of observed and expectedP deaths P summed over all birth-cohorts are equal, that is O ¼ E, with an P O overall average ratio of P E ¼ 1 or 100%. The individual
SCMR value indicates to what extent the SCMR associated with an individual birth-cohort lies above or falls below the overall average ratio of 100%.
Results Between 1950 and 2010 there were 47 000 deaths from IBD, with 27 000 secondary to UC and 20 000 secondary to Crohn’s disease (Table 1). Of all deaths, 93 and 7% occurred in whites and nonwhites, respectively, with similar proportions in UC and Crohn’s disease. About 70% of IBD-related deaths in nonwhites occurred in African Americans. During the period between 1950 and 2010 the average annual death rate of UC fell steadily from 0.46 to 0.11 per 100 000, whereas the death rate of Crohn’s disease increased from 0.08 to 0.20 per 100 000. Similar time trends were found in whites and nonwhites, as well as males and females (Fig. 1). These seemingly steady behaviors masked underlying divergent trends amongst individual age groups. In both diseases alike, the older age groups showed an initial rise followed by a smooth decline, whereas the younger age groups were characterized by a steady steep decline. Such divergent time trends of consecutive period-age contours are generally suggestive of an underlying birth-cohort pattern. The upper right graph of Fig. 2 shows the death rates of Crohn’s disease among whites plotted against the period of birth as cohort-age contours. The cohort-age contours of Crohn’s disease aligned themselves to form
Table 1
All races White Nonwhite Black
US deaths from inflammatory bowel disease 1950–2010 IBD
UC
CD
46 469 43 124 3345 2324
26 526 24 613 1913 1145
19 943 18 511 1432 1179
CD, Crohn’s disease; IBD, inflammatory bowel disease; UC, ulcerative colitis.
one hyperbola with an initial rise between 1865 and 1935 and a subsequent decline during more recent times. This pattern was also confirmed by the time trends of the corresponding SCMR values shown in the left upper graph of Fig. 2. In UC among US whites, the individual cohort-age contours were also aligned as one hyperbola that appeared shifted left towards earlier generations when compared with Crohn’s disease. A logarithmic scale was chosen to be able to plot data from different age groups in a single graph. Because of the logarithmic scale of the y-axis, however, the decline in the younger age groups became somewhat exaggerated at the expense of compressing the initial rise in the older age groups. The linear plot of the corresponding SCMR values clearly showed an initial rise and subsequent decline in the risk for UC associated with consecutive birth-cohorts. Similar patterns as in UC were also observed in the combined cohort-age contours of Crohn’s disease and UC, as well as their combined SCMR values (in the bottom two graphs of Fig. 2). The left upper graph of Fig. 3 contains the three SCMR curves among US whites from Fig. 2. The right upper graph contains for comparison the three SCMR curves among nonwhites. In the two lower graphs of Fig. 3, the SMCR curves are stratified by sex group. In both race and sex groups the time trends of IBD were characterized by similar birth-cohort patterns. The risk of dying from IBD increased initially among generations born between 1865 and 1895, and decreased in all subsequent generations. Compared with Crohn’s disease, the rise and fall of UC were shifted by 30 years towards older generations. The recent fall of Crohn’s disease was more pronounced in white than nonwhite birth-cohorts. Otherwise, no racespecific or sex-specific variations were noted.
Discussion The present analysis of US mortality from Crohn’s disease and UC shows that their time trends are both shaped by underlying birth-cohort patterns. These patterns apply similarly to whites and nonwhites, as well males and females. A birth-cohort pattern implies that the risk for developing IBD is acquired during childhood or early adolescence and that such risk is retained for the remainder of one’s life [4]. As the risk exposure changes over historical times, the changing susceptibility for the disease among consecutive birth-cohorts comes to reflect the time dependent variations in risk acquisition. Crohn’s disease is generally more common in females, whereas UC tends to be slightly more common in males [5,6]. Both types of IBD are more common in whites than nonwhites [6,7]. Irrespective of such demographic variations, however, the temporal variations in risk acquisition seem to have similarly occurred in both ethnic and sex groups. Previous analyses of the time trends of IBD have suggested that at least two separate environmental risk
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
890
European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 8
Fig. 1
0.6
0.6 White
0.5
Nonwhite
0.5 0.4
0.4 IBD
0.3
0.3 CD
0.2
UC
Death rate per 100 000
0.1
0.2 0.1 0
0 1955
1965
1975
1985
1995
2005
1955
1965
1975
1985
1995
2005
0.6
0.6 Male
0.5 0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0 1955
1965
1975
1985
1995
Female
0.5
0 2005 Period of death
1955
1965
1975
1985
1995
2005
Time trends of mortality from Crohn’s disease (CD), ulcerative colitis (UC), and inflammatory bowel disease (IBD) in US whites, nonwhites, males, and females.
factors must contribute to the occurrence of Crohn’s disease and UC [1,8,9]. One risk factor is primarily responsible for the development of IBD in general and, taken on its own, is sufficient to cause UC. A second risk factor in the presence of the first factor results in the phenotypic expression of IBD as Crohn’s disease. The birth-cohort analysis partly supports this hypothesis. The temporal behavior of the first risk factor is reflected by the birth-cohort patterns of UC alone and both types of IBD combined. The temporal behavior of the second risk factor is reflected by the separate birth-cohort pattern of Crohn’s disease. The birth-cohort patterns indicate that the exposure to both risk factors must occur during an early period of life. These exposures have changed over historical times similarly in both sex groups and different ethnic groups residing within the USA. If the history of gastric cancer and peptic ulcer with their similar underlying birth-cohort patterns is of any guidance, the underlying factors are probably infectious in origin [10]. Rising and falling trends of some yet unknown childhood infection presently appear the most likely explanation for the peculiar time trends of UC and Crohn’s disease. The birth-cohort phenomenon implies that in any large group of patients with inflammatory bowel, their
frequency distribution rises and falls in relationship with the patients’ birthdates. The vital statistics provide an easy access to study a large population with a diagnosis of IBD. For the purpose of the birth-cohort analysis, however, the occurrence of death itself is largely irrelevant. In other words, the birth-cohort pattern is a characteristic of the disease itself rather than its mortality. The birthdate characterizes a patient’s risk to harbor IBD similarly to their sex, ethnicity, or place of residence. If another health attribute or morbidity parameter besides mortality were available to accumulate a large IBD population over an extended time period, such data would be also characterized by an underlying birth-cohort pattern. The means of data accumulation, for instance, through statistics of prevalence, incidence, physician visits, hospitalization, or disability, would not influence the risk of birthdate on the occurrence IBD. The prerequisites for a database to reveal a birth-cohort phenomenon pertain solely to a large size, as well as a long and continuous time period of data collection. Only a small fraction of IBD patients die from their disease. Some patients die from colorectal cancer, primary sclerosing cholangitis, thrombembolic or infectious complications without mention of IBD on their death
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Birth-cohort patterns in IBD Sonnenberg 891
Fig. 2
4
200% 90 80
1
150% 70 60 100%
50 40
0.1
30
50% 20
CD
CD 0%
0.01 1
5 85
1
5 88
1
5 91
1
5 94
75 19
1
5 85
85 18
1
5 91
1
5 94
75 19
4
Death rate per 100 000
80 70
1
150%
60 50
40 100%
0.1 50% 30
UC
UC
20
Standardized cohort mortality ratio
200% 90
0%
0.01 18
55
18
85
19
15
19
45
75 19
18
55
85 18
19
15
19
45
75 19
4
200% 90 80 70
1
60
150% 50
40 100%
0.1 30
50% 20
IBD
IBD 0%
0.01 18
55
18
85
19
15
19
75 19
45
18
55
85 18
19
15
19
45
75 19
Period of birth Age-specific death rates of Crohn’s disease (CD), ulcerative colitis (UC), and inflammatory bowel disease (IBD) in US whites, data plotted logarithmically as cohort-age contours (left panels) and as standardized cohort mortality ratio (right panels).
certificate. Because the actual means of data accumulation is secondary to the birth-cohort analysis, it is largely irrelevant as to whether the vital statistics capture all or only a small fraction of diseased patients as long as the database is sufficiently large. For these reasons it also
does not matter to what extent mortality data reflect on underlying trends in the incidence of IBD. As treatment of IBD continues to improve, it may become increasingly a nonfatal disease with normal life expectancy.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
892 European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 8
Fig. 3
200%
200% White
150%
Nonwhite
IBD
150%
100%
100% UC
50%
Standardized cohort mortality ratio
50% CD
0% 1855
1885
1915
1945
1975
200%
0% 1855
1885
1915
1945
200% Male
Female
150%
150%
100%
100%
50%
50%
0% 1855
1975
1885
1915
1945
1975
0% 1855
1885
1915
1945
1975
Period of birth Standardized cohort mortality ratios of Crohn’s disease (CD), ulcerative colitis (UC), and inflammatory bowel disease (IBD) in US whites, nonwhites, males, and females.
With the advance of endoscopic and radiologic imaging techniques, the diagnosis of inflammatory disease has become more accurate. Although there is little evidence to support such contention, one may also speculate that IBD coding practices have changed over time. However, such period effects exerted by changes in medical practice are independent of cohort effects and do not influence the association between the period of birth and disease mortality or other morbidity parameters. Over time, the birth-cohort effects result initially in a shift of disease occurrence towards older patients and subsequently in its overall decline. At the
present time the ongoing age-shift is more pronounced in Crohn’s disease and the overall decline is more pronounced in UC. The interaction between the period effects (of therapy, diagnosis or coding) on one hand and the varying cohort effects (of UC and Crohn’s disease) on the other hand is responsible for the temporal behavior depicted by Fig. 1. Conclusion
The present analysis suggests that the occurrence of IBD is influenced by an underlying birth-cohort pattern with an initial rise and subsequent decline during the past
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Birth-cohort patterns in IBD Sonnenberg 893
150 years. This pattern is observed similarly among different sex and ethnic groups within the USA. The fact that the birthdate influences a patient’s risk to harbor IBD also suggests that some highly influential risk factors for developing the disease must have been acquired during childhood. This insight may provide an important clue in the search for the yet unknown etiology of IBD.
Acknowledgements Conflicts of interest
There are no conflicts of interest.
References 1 Sonnenberg A, Koch TR. Period and generation effects on mortality from idiopathic inflammatory bowel disease. Dig Dis Sci 1989; 34:1720–1729. 2 Sonnenberg A. Birth-cohort patterns in Crohn’s disease and ulcerative colitis. Eur J Gastroenterol Hepatol 2014; 26:19–25.
3
Gardner MJ, Osmond C. Interpretation of time trends in disease rates in the presence of generation effects. Stat Med 1984; 3:113–130. 4 MacMahon B, Pugh TE. Epidemiology – principles and methods. Boston, MA: Little, Brown, and Company; 1970. pp. 184–198. 5 Brant SR, Nguyen GC. Is there a gender difference in the prevalence of Crohn’s disease or ulcerative colitis? Inflamm Bowel Dis 2008; 14 (Suppl 2): S2–S3. 6 Loftus EV Jr. Clinical epidemiology of inflammatory bowel disease: incidence, prevalence, and environmental influences. Gastroenterology 2004; 126:1504–1517. 7 Mahid SS, Mulhall AM, Gholson RD, Eichenberger MR, Galandiuk S. Inflammatory bowel disease and African Americans: a systematic review. Inflamm Bowel Dis 2008; 14:960–967. 8 Delcò F, Sonnenberg A. Commonalities in the time trends of Crohn’s disease and ulcerative colitis. Am J Gastroenterol 1999; 94:2171–2176. 9 Sonnenberg A. Time trends of mortality from Crohn’s disease and ulcerative colitis. Int J Epidemiol 2007; 36:890–899. 10 Sonnenberg A. Differences in the birth-cohort patterns of gastric cancer and peptic ulcer. Gut 2010; 59:736–743.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Original article
Ubiquitous occurrence of birth-cohort patterns in inflammatory bowel disease Amnon Sonnenberg Background and aims The aim of the present study was to demonstrate the ubiquitous occurrence of the birthcohort phenomenon of inflammatory bowel disease among US whites and nonwhites, as well as males and females. Methods Mortality from Crohn’s disease and ulcerative colitis in the USA between 1950 and 2010 were analyzed to discern underlying birth-cohort patterns affecting both their time trends. Age-standardized cohort mortality ratio was used as a summary statistic to represent the overall mortality associated with consecutive birth-cohorts.
Conclusion The birth-cohort patterns indicate that exposure to two separate risk factors must have occurred in both diseases during an early period of life. In the USA, these exposures have changed over historical times similarly in both sexes and different ethnic groups. Eur J Gastroenterol Hepatol 26:888–893 © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins. European Journal of Gastroenterology & Hepatology 2014, 26:888–893 Keywords: birth-cohort pattern, Crohn’s disease, death rates, ethnicity, time trends, ulcerative colitis
Results The cohort-age contours of Crohn’s disease aligned to form one hyperbola with an initial rise between 1865 and 1935 and a subsequent decline. This pattern was confirmed by the time trends of the corresponding standardized cohort mortality ratio values. In ulcerative colitis, the individual cohort-age contours also aligned into one hyperbola that appeared shifted towards earlier generations by about 30 years when compared with Crohn’s disease. Similar trends were observed in men and women or whites and nonwhites analyzed separately.
Portland VA Medical Center and the Oregon Health and Science University, Portland, Oregon, USA
Introduction
recorded by underlying cause of death according to the International Classification of Diseases. In its 6th (1950–1957) and 7th revision (1958–1967), Crohn’s disease and UC were coded as 572.0 and 572.2, respectively. In the 8th revision (1968–1978) they were coded as 563.0 and 563.1, respectively. In the 9th revision (1979–1999) they were coded as 555 and 556, respectively. Since the 10th revision (starting in 2000), Crohn’s disease and UC have been coded as K50 and K51, respectively. For the purpose of the present analysis, none of the additional four-digit or four-letter subcodes of the more recent 9th and 10th International Classification of Disease revisions were considered.
The time trends of mortality from inflammatory bowel disease (IBD) are influenced by an underlying birthcohort pattern. The risk of IBD rose among consecutive generations born during the 19th century, peaked among those born around the turn of the century, and then declined in all subsequent generations born during the 20th century [1]. This pattern is found similarly in the vital statistics of several different European countries, as well as the USA [2]. Because of its large population size, the US is particularly well suited to study such temporal patterns also within its different demographic strata. Using US vital statistics, the present analysis was aimed as an extension of the previous study to test whether the IBD birth-cohort phenomenon also applies to different race and sex groups.
Methods Data sources
Resident US population and number of deaths from Crohn’ disease and ulcerative colitis (UC), stratified by age, sex, and ethnicity, were made available by the US National Center of Health Statistics. Mortality data were available from 1950 until 2010; stratification by sex was available only since 1962. Deaths were 0954-691X © 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
Correspondence to Amnon Sonnenberg, MD, MSc, Gastroenterology, Portland VA Medical Center P3-GI, 3710 SW U.S. Veterans Hospital Road, Portland, Oregon 97239, USA Tel: + 1 503 220 8262 x56679; fax: + 1 503 220 3426; e-mail: [email protected] Received 17 February 2014 Accepted 7 April 2014
Data analyses
Average annual age-specific death rates were calculated for consecutive 10-year periods and 10-year age groups. For instance, the total number of deaths from 1971 until 1980 among individuals aged 45–54 was divided by the corresponding number of the total resident population of the same age group and living during the same time period when the deaths occurred. The age-specific death rates were plotted against the period of birth as cohort-age contours. In the plots, each age group was labeled by its central year, for instance, 40 indicating the DOI: 10.1097/MEG.0000000000000118
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Birth-cohort patterns in IBD Sonnenberg 889
age group 35–44 and 50 indicating the age group 45–54. The periods of birth were labeled by the mid year of birth, for instance 1925 instead of 1916–1935. The analyses were performed separately for Crohn’s disease, UC, and both diseases analyzed jointly as IBD. An age-standardized cohort mortality ratio (SCMR) was used as a summary statistic of the overall mortality associated with each consecutive birth-cohort as previously described [3]. Briefly, the SCMR was calculated similarly to the standardized mortality ratio as the ratio of observed (O) over expected (E) number of deaths in each birthcohort: SCMR = O/E. The SCMR was expressed as a percentage and plotted against the period of birth. The numbers of observed and expectedP deaths P summed over all birth-cohorts are equal, that is O ¼ E, with an P O overall average ratio of P E ¼ 1 or 100%. The individual
SCMR value indicates to what extent the SCMR associated with an individual birth-cohort lies above or falls below the overall average ratio of 100%.
Results Between 1950 and 2010 there were 47 000 deaths from IBD, with 27 000 secondary to UC and 20 000 secondary to Crohn’s disease (Table 1). Of all deaths, 93 and 7% occurred in whites and nonwhites, respectively, with similar proportions in UC and Crohn’s disease. About 70% of IBD-related deaths in nonwhites occurred in African Americans. During the period between 1950 and 2010 the average annual death rate of UC fell steadily from 0.46 to 0.11 per 100 000, whereas the death rate of Crohn’s disease increased from 0.08 to 0.20 per 100 000. Similar time trends were found in whites and nonwhites, as well as males and females (Fig. 1). These seemingly steady behaviors masked underlying divergent trends amongst individual age groups. In both diseases alike, the older age groups showed an initial rise followed by a smooth decline, whereas the younger age groups were characterized by a steady steep decline. Such divergent time trends of consecutive period-age contours are generally suggestive of an underlying birth-cohort pattern. The upper right graph of Fig. 2 shows the death rates of Crohn’s disease among whites plotted against the period of birth as cohort-age contours. The cohort-age contours of Crohn’s disease aligned themselves to form
Table 1
All races White Nonwhite Black
US deaths from inflammatory bowel disease 1950–2010 IBD
UC
CD
46 469 43 124 3345 2324
26 526 24 613 1913 1145
19 943 18 511 1432 1179
CD, Crohn’s disease; IBD, inflammatory bowel disease; UC, ulcerative colitis.
one hyperbola with an initial rise between 1865 and 1935 and a subsequent decline during more recent times. This pattern was also confirmed by the time trends of the corresponding SCMR values shown in the left upper graph of Fig. 2. In UC among US whites, the individual cohort-age contours were also aligned as one hyperbola that appeared shifted left towards earlier generations when compared with Crohn’s disease. A logarithmic scale was chosen to be able to plot data from different age groups in a single graph. Because of the logarithmic scale of the y-axis, however, the decline in the younger age groups became somewhat exaggerated at the expense of compressing the initial rise in the older age groups. The linear plot of the corresponding SCMR values clearly showed an initial rise and subsequent decline in the risk for UC associated with consecutive birth-cohorts. Similar patterns as in UC were also observed in the combined cohort-age contours of Crohn’s disease and UC, as well as their combined SCMR values (in the bottom two graphs of Fig. 2). The left upper graph of Fig. 3 contains the three SCMR curves among US whites from Fig. 2. The right upper graph contains for comparison the three SCMR curves among nonwhites. In the two lower graphs of Fig. 3, the SMCR curves are stratified by sex group. In both race and sex groups the time trends of IBD were characterized by similar birth-cohort patterns. The risk of dying from IBD increased initially among generations born between 1865 and 1895, and decreased in all subsequent generations. Compared with Crohn’s disease, the rise and fall of UC were shifted by 30 years towards older generations. The recent fall of Crohn’s disease was more pronounced in white than nonwhite birth-cohorts. Otherwise, no racespecific or sex-specific variations were noted.
Discussion The present analysis of US mortality from Crohn’s disease and UC shows that their time trends are both shaped by underlying birth-cohort patterns. These patterns apply similarly to whites and nonwhites, as well males and females. A birth-cohort pattern implies that the risk for developing IBD is acquired during childhood or early adolescence and that such risk is retained for the remainder of one’s life [4]. As the risk exposure changes over historical times, the changing susceptibility for the disease among consecutive birth-cohorts comes to reflect the time dependent variations in risk acquisition. Crohn’s disease is generally more common in females, whereas UC tends to be slightly more common in males [5,6]. Both types of IBD are more common in whites than nonwhites [6,7]. Irrespective of such demographic variations, however, the temporal variations in risk acquisition seem to have similarly occurred in both ethnic and sex groups. Previous analyses of the time trends of IBD have suggested that at least two separate environmental risk
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
890
European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 8
Fig. 1
0.6
0.6 White
0.5
Nonwhite
0.5 0.4
0.4 IBD
0.3
0.3 CD
0.2
UC
Death rate per 100 000
0.1
0.2 0.1 0
0 1955
1965
1975
1985
1995
2005
1955
1965
1975
1985
1995
2005
0.6
0.6 Male
0.5 0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0 1955
1965
1975
1985
1995
Female
0.5
0 2005 Period of death
1955
1965
1975
1985
1995
2005
Time trends of mortality from Crohn’s disease (CD), ulcerative colitis (UC), and inflammatory bowel disease (IBD) in US whites, nonwhites, males, and females.
factors must contribute to the occurrence of Crohn’s disease and UC [1,8,9]. One risk factor is primarily responsible for the development of IBD in general and, taken on its own, is sufficient to cause UC. A second risk factor in the presence of the first factor results in the phenotypic expression of IBD as Crohn’s disease. The birth-cohort analysis partly supports this hypothesis. The temporal behavior of the first risk factor is reflected by the birth-cohort patterns of UC alone and both types of IBD combined. The temporal behavior of the second risk factor is reflected by the separate birth-cohort pattern of Crohn’s disease. The birth-cohort patterns indicate that the exposure to both risk factors must occur during an early period of life. These exposures have changed over historical times similarly in both sex groups and different ethnic groups residing within the USA. If the history of gastric cancer and peptic ulcer with their similar underlying birth-cohort patterns is of any guidance, the underlying factors are probably infectious in origin [10]. Rising and falling trends of some yet unknown childhood infection presently appear the most likely explanation for the peculiar time trends of UC and Crohn’s disease. The birth-cohort phenomenon implies that in any large group of patients with inflammatory bowel, their
frequency distribution rises and falls in relationship with the patients’ birthdates. The vital statistics provide an easy access to study a large population with a diagnosis of IBD. For the purpose of the birth-cohort analysis, however, the occurrence of death itself is largely irrelevant. In other words, the birth-cohort pattern is a characteristic of the disease itself rather than its mortality. The birthdate characterizes a patient’s risk to harbor IBD similarly to their sex, ethnicity, or place of residence. If another health attribute or morbidity parameter besides mortality were available to accumulate a large IBD population over an extended time period, such data would be also characterized by an underlying birth-cohort pattern. The means of data accumulation, for instance, through statistics of prevalence, incidence, physician visits, hospitalization, or disability, would not influence the risk of birthdate on the occurrence IBD. The prerequisites for a database to reveal a birth-cohort phenomenon pertain solely to a large size, as well as a long and continuous time period of data collection. Only a small fraction of IBD patients die from their disease. Some patients die from colorectal cancer, primary sclerosing cholangitis, thrombembolic or infectious complications without mention of IBD on their death
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Birth-cohort patterns in IBD Sonnenberg 891
Fig. 2
4
200% 90 80
1
150% 70 60 100%
50 40
0.1
30
50% 20
CD
CD 0%
0.01 1
5 85
1
5 88
1
5 91
1
5 94
75 19
1
5 85
85 18
1
5 91
1
5 94
75 19
4
Death rate per 100 000
80 70
1
150%
60 50
40 100%
0.1 50% 30
UC
UC
20
Standardized cohort mortality ratio
200% 90
0%
0.01 18
55
18
85
19
15
19
45
75 19
18
55
85 18
19
15
19
45
75 19
4
200% 90 80 70
1
60
150% 50
40 100%
0.1 30
50% 20
IBD
IBD 0%
0.01 18
55
18
85
19
15
19
75 19
45
18
55
85 18
19
15
19
45
75 19
Period of birth Age-specific death rates of Crohn’s disease (CD), ulcerative colitis (UC), and inflammatory bowel disease (IBD) in US whites, data plotted logarithmically as cohort-age contours (left panels) and as standardized cohort mortality ratio (right panels).
certificate. Because the actual means of data accumulation is secondary to the birth-cohort analysis, it is largely irrelevant as to whether the vital statistics capture all or only a small fraction of diseased patients as long as the database is sufficiently large. For these reasons it also
does not matter to what extent mortality data reflect on underlying trends in the incidence of IBD. As treatment of IBD continues to improve, it may become increasingly a nonfatal disease with normal life expectancy.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
892 European Journal of Gastroenterology & Hepatology 2014, Vol 26 No 8
Fig. 3
200%
200% White
150%
Nonwhite
IBD
150%
100%
100% UC
50%
Standardized cohort mortality ratio
50% CD
0% 1855
1885
1915
1945
1975
200%
0% 1855
1885
1915
1945
200% Male
Female
150%
150%
100%
100%
50%
50%
0% 1855
1975
1885
1915
1945
1975
0% 1855
1885
1915
1945
1975
Period of birth Standardized cohort mortality ratios of Crohn’s disease (CD), ulcerative colitis (UC), and inflammatory bowel disease (IBD) in US whites, nonwhites, males, and females.
With the advance of endoscopic and radiologic imaging techniques, the diagnosis of inflammatory disease has become more accurate. Although there is little evidence to support such contention, one may also speculate that IBD coding practices have changed over time. However, such period effects exerted by changes in medical practice are independent of cohort effects and do not influence the association between the period of birth and disease mortality or other morbidity parameters. Over time, the birth-cohort effects result initially in a shift of disease occurrence towards older patients and subsequently in its overall decline. At the
present time the ongoing age-shift is more pronounced in Crohn’s disease and the overall decline is more pronounced in UC. The interaction between the period effects (of therapy, diagnosis or coding) on one hand and the varying cohort effects (of UC and Crohn’s disease) on the other hand is responsible for the temporal behavior depicted by Fig. 1. Conclusion
The present analysis suggests that the occurrence of IBD is influenced by an underlying birth-cohort pattern with an initial rise and subsequent decline during the past
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Birth-cohort patterns in IBD Sonnenberg 893
150 years. This pattern is observed similarly among different sex and ethnic groups within the USA. The fact that the birthdate influences a patient’s risk to harbor IBD also suggests that some highly influential risk factors for developing the disease must have been acquired during childhood. This insight may provide an important clue in the search for the yet unknown etiology of IBD.
Acknowledgements Conflicts of interest
There are no conflicts of interest.
References 1 Sonnenberg A, Koch TR. Period and generation effects on mortality from idiopathic inflammatory bowel disease. Dig Dis Sci 1989; 34:1720–1729. 2 Sonnenberg A. Birth-cohort patterns in Crohn’s disease and ulcerative colitis. Eur J Gastroenterol Hepatol 2014; 26:19–25.
3
Gardner MJ, Osmond C. Interpretation of time trends in disease rates in the presence of generation effects. Stat Med 1984; 3:113–130. 4 MacMahon B, Pugh TE. Epidemiology – principles and methods. Boston, MA: Little, Brown, and Company; 1970. pp. 184–198. 5 Brant SR, Nguyen GC. Is there a gender difference in the prevalence of Crohn’s disease or ulcerative colitis? Inflamm Bowel Dis 2008; 14 (Suppl 2): S2–S3. 6 Loftus EV Jr. Clinical epidemiology of inflammatory bowel disease: incidence, prevalence, and environmental influences. Gastroenterology 2004; 126:1504–1517. 7 Mahid SS, Mulhall AM, Gholson RD, Eichenberger MR, Galandiuk S. Inflammatory bowel disease and African Americans: a systematic review. Inflamm Bowel Dis 2008; 14:960–967. 8 Delcò F, Sonnenberg A. Commonalities in the time trends of Crohn’s disease and ulcerative colitis. Am J Gastroenterol 1999; 94:2171–2176. 9 Sonnenberg A. Time trends of mortality from Crohn’s disease and ulcerative colitis. Int J Epidemiol 2007; 36:890–899. 10 Sonnenberg A. Differences in the birth-cohort patterns of gastric cancer and peptic ulcer. Gut 2010; 59:736–743.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
