the lancet

Articles
Differences in cancer rates among adults born between 1920
and 1990 in the USA: an analysis of population-based cancer
registry data
Hyuna Sung, Chenxi Jiang, Priti Bandi, Adair Minihan, Miranda Fidler-Benaoudia, Farhad Islami, Rebecca L Siegel, Ahmedin Jemal
Summary
Background Trends in cancer incidence in recent birth cohorts largely reflect changes in exposures during early life
and foreshadow the future disease burden. Herein, we examined cancer incidence and mortality trends, by birth
cohort, for 34 types of cancer in the USA.
Methods In this analysis, we obtained incidence data for 34 types of cancer and mortality data for 25 types of cancer
for individuals aged 25–84 years for the period Jan 1, 2000, to Dec 31, 2019 from the North American Association of
Central Cancer Registries and the US National Center for Health Statistics, respectively. We calculated birth cohortspecific incidence rate ratios (IRRs) and mortality rate ratios (MRRs), adjusted for age and period effects, by nominal
birth cohort, separated by 5 year intervals, from 1920 to 1990.
Findings We extracted data for 23 654 000 patients diagnosed with 34 types of cancer and 7 348 137 deaths from
25 cancers for the period Jan 1, 2000, to Dec 31, 2019. We found that IRRs increased with each successive birth cohort
born since approximately 1920 for eight of 34 cancers (pcohort<0·050). Notably, the incidence rate was approximately
two-to-three times higher in the 1990 birth cohort than in the 1955 birth cohort for small intestine (IRR 3·56 [95% CI
2·96–4·27]), kidney and renal pelvis (2·92 [2·50–3·42]), and pancreatic (2·61 [2·22–3·07]) cancers in both male and
female individuals; and for liver and intrahepatic bile duct cancer in female individuals (2·05 [1·23–3·44]).
Additionally, the IRRs increased in younger cohorts, after a decline in older birth cohorts, for nine of the remaining
cancers (pcohort<0·050): oestrogen-receptor-positive breast cancer, uterine corpus cancer, colorectal cancer, non-cardia
gastric cancer, gallbladder and other biliary cancer, ovarian cancer, testicular cancer, anal cancer in male individuals,
and Kaposi sarcoma in male individuals. Across cancer types, the incidence rate in the 1990 birth cohort ranged from
12% (IRR1990 vs 1975 1·12 [95% CI 1·03–1·21] for ovarian cancer) to 169% (IRR1990 vs 1930 2·69 [2·34–3·08] for uterine corpus
cancer) higher than the rate in the birth cohort with the lowest incidence rate. The MRRs increased in successively
younger birth cohorts alongside IRRs for liver and intrahepatic bile duct cancer in female individuals, uterine corpus,
gallbladder and other biliary, testicular, and colorectal cancers, while MRRs declined or stabilised in younger birth
cohorts for most cancers types.
Lancet Public Health 2024;
9: e583–93
Surveillance and Health Equity
Science, American Cancer
Society, Atlanta, GA, USA
(H Sung PhD, C Jiang MPH,
P Bandi PhD, A Minihan MPH,
F Islami PhD, R L Siegel MPH,
A Jemal PhD); Cancer
Epidemiology and Prevention
Research, Cancer Care Alberta,
Alberta Health Services,
Calgary, AB, Canada
(M Fidler-Benaoudia PhD);
Department of Oncology and
Community Health Sciences,
Cumming School of Medicine,
University of Calgary, Calgary,
AB, Canada
(M Fidler-Benaoudia)
Correspondence to:
Dr Hyuna Sung, Surveillance and
Health Equity Science, American
Cancer Society,
Atlanta, GA 30303, USA
[email protected]
Interpretation 17 of 34 cancers had an increasing incidence in younger birth cohorts, including nine that previously
had declining incidence in older birth cohorts. These findings add to growing evidence of increased cancer risk in
younger generations, highlighting the need to identify and tackle underlying risk factors.
Funding American Cancer Society.
Copyright © 2024 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY-NC
4.0 license.
Introduction
We previously reported that incidence rates increased in
successively younger birth cohorts for eight cancer
types—six of which are obesity-related—alongside
steeper or exclusive increases in young adults (age
25–49 years) over time in the USA.1 Some studies have
since reported increases in cancer incidence among
young adults (usually defined as diagnosis age <50 years)
in the USA and elsewhere.2–5 A recent analysis of data
from 13 registries in the USA suggested that individuals
born between 1965 and 1980 (also known as Generation X)
might have higher incidence rates of some cancers
(eg, thyroid, colorectal, kidney, uterine corpus, and
www.thelancet.com/public-health Vol 9 August 2024
leukaemia) as well as an increased risk of all leading
cancers combined.6 However, a comprehensive analysis
that examines both cancer incidence and mortality trends
by birth cohort or year of birth in contemporary
generations has yet to be done. Trends in cancer
incidence in young generations or young adults (age
<50 years) largely reflect increased exposure to
carcinogenic factors during early life or young adulthood
compared with previous generations,7 and foreshadow
future disease burden as these young cohorts carry their
increased risk into older age, when cancers most
frequently occur.7 To quantify cancer risk among younger
generations of adults compared with their older
e583
Articles
Research in context
Evidence before this study
We searched PubMed for articles published in English from
Jan 1, 2000, to June 30, 2024, using the terms “cancer
incidence”, “cancer mortality”, “young adults”, “early-onset”,
“age-period-cohort”, “birth cohort”, “cohort effect”, and “period
effect”. Studies were considered if they examined cancer
incidence or mortality trends using an age-period-cohort
modelling approach and presented parameters quantifying the
differences in cancer risk by year of birth on the basis of
population-based cancer registry or mortality data. We also
reviewed the titles and abstracts of the articles in the references
of identified studies. Cancer incidence trends by birth cohort
have been extensively studied in the USA and elsewhere.
Previous studies that examined birth cohort trends in cancer
incidence have mostly focused on cancers of the lung, female
breast, testis, and stomach, and more recently on colorectal
cancer. We identified three previous studies that examined birth
cohort trends for a comprehensive list of cancers in the USA
(21–30 cancer types) and Canada (13 cancer types) with the
latest birth years of 1983 and 1988. Evidence suggests that
incidence rates have increased in successively younger birth
cohorts for multiple obesity-related cancers (colorectum,
uterine corpus, gallbladder and other biliary, kidney and renal
pelvis, and pancreas in both the USA and Canada) alongside
steeper or exclusive increases in young adults (age 25–49 years)
over time. We did not identify any studies that reported cancer
mortality trends by birth cohort for a comprehensive list of
cancers.
Added value of this study
Based on nationally representative population-based data, this
study provides cancer incidence trends by year of birth between
1920 and 1990 for 34 cancer types in the USA, with 25 of these
counterparts, we aimed to examine trends in incidence
rates of 34 cancers and mortality rates of 25 cancers for
2000 to 2019 by birth cohort spanning from 1920 to 1990
in the USA using nationally representative high-quality
data.
Methods
Study design and data sources
For more on the NAACCR
database see https://www.
naaccr.org/cina-public-use-dataset/
See Online for appendix
For more on the SEER Program
see https://seer.cancer.gov/
e584
In this analysis of population-based registry data, we
obtained incidence data for the period Jan 1, 2000 to
Dec 31, 2019, for the 34 most common cancer types
diagnosed at ages 25–84 years (ie, diagnosed in at least
200 000 individuals during the study period) from the
North American Association of Central Cancer Registries
(known as NAACCR); these data covered 94% of the
US population, with six states and the District of Columbia
excluded due to incompleteness of data (appendix p 7). For
female breast cancer, we obtained incidence data from
17 Surveillance, Epidemiology, and End Results (SEER)
Program registries to enable subtype-specific analysis.
accompanied by mortality trends. Incidence rates continued to
increase in successively younger generations for 17 of the
34 cancer types studied with a significant birth cohort effect,
expanding the list of cancers from previous findings
(colorectum, uterine corpus, gallbladder and other biliary,
kidney and renal pelvis, pancreas, myeloma, non-cardia gastric,
leukaemia, and testis) by eight additional cancers (cardia
gastric, small intestine, oestrogen receptor-positive breast,
ovary, female liver, female non-HPV-associated oral and
pharynx, male anus, and male Kaposi sarcoma). Mortality
trends increased in conjunction with the incidence trends for
liver (female), uterine corpus, gallbladder, testicular, and
colorectal cancers.
Implications of all the available evidence
The rising cancer incidence for many cancer types in
successively younger generations suggests increases in the
prevalence of carcinogenic exposures during early life or young
adulthood, which have yet to be elucidated. There is a need for
investigation of various mechanisms contributing to the
heightened incidence of cancer, and hence cancer risk, among
more recent birth cohorts and for a comprehensive life course
framework approach in epidemiological research to address
substantial knowledge gaps regarding birth-cohort-specific
exposures and cancer risk trajectories. Additionally, the
evidence underscores the importance of developing
intervention strategies that align with the social and cultural
context, values, and preferences of the young generations.
Without effective population-level interventions, as younger
generations age, an overall increase in cancer burden could
occur in the future, halting or reversing decades of progress
against the disease.
Cases were classified per the International Classification
of Disease Oncology, third edition, and further into ana­
tomical, histological, and molecular subtypes (ie, oestrogen
receptor-positive and oestrogen receptor-negative for
breast cancer; cardia and non-cardia for gastric cancer;
adenocarcinoma and squamous cell carcinoma for
oesophageal cancer; and HPV-related and non-HPVrelated for oral cavity and pharynx cancers;8,9 appendix p 8).
Nationwide mortality data were obtained from the US
National Center for Health Statistics. Mortality analyses
were restricted to 25 cancer types, classified by ICD-10,
with exclusion of Kaposi sarcoma because of a low
number of deaths, and breast, oesophagus, oral cavity and
pharynx, and gastric cancers because of the absence of
subtype classification (appendix pp 3–4). The population
estimates (denominators used to calculate rates) were
provided by the US Census Bureau for each study year.
Sex and age information were extracted from all databases.
Data on race and ethnicity were available but were not
analysed.
www.thelancet.com/public-health Vol 9 August 2024
Articles
This study used de-identified publicly available data,
which is considered non-human participants research
under the Common Rule (US federal regulation 45
CFR §46); thus, institutional review board approval and
informed consent were not needed.
Statistical analysis
To evaluate birth cohort trends in cancer rates, adjusted
for age and period effects, we fitted age-period-cohort
models to the incidence or mortality rates of each cancer
type using weighted least squares, assuming Poissondistributed counts and including overdispersion para­
meters for potential extra-Poisson variation.10 A change in
the birth cohort trend usually indicates changes in
exposure prevalence, leading to varying risks of developing
cancers for individuals of the same generation. By
contrast, a period effect reflects systematic changes in
cancer ascertainment or the influence of newly introduced
or improved medical interventions, affecting all age
groups simultaneously during the same period.
To obtain stable parameter estimates across all cancer
types, 5-year intervals were used to categorise age groups
(age 25–29 years up to age 80–84 years) and period of
cancer diagnosis or death (2000–04, 2005–09, 2010–14, and
2015–19). Nominal birth cohorts (eg, 1990 birth cohort)
were constructed by subtracting the mid-year age (eg, 27 for
25–29 years) from the mid-year period (eg, 2017 for
2015–19), resulting in 15 partially overlapping birth cohorts,
referenced by mid-year of birth and spaced by 5 years
(from 1920 up to 1990). Such that, for instance, the
1990 birth cohort largely represents the experiences of
those born in approximately 1990, with 80% of cases or
deaths in this cohort involving individuals born between
1987 and 1993 (appendix pp 3–4).10 First, we generated
cohort rate ratio curves, which are composite curves that
include backward projections for older cohorts and forward
projections for younger cohorts.10 A ratio of the net
incidence or mortality rate in each cohort relative to the
reference cohort, also known as incidence rate ratio (IRR)
and mortality rate ratio (MRR), respectively, was computed
for each birth cohort.10 We first analysed cancer types for
which IRRs increased with each successive birth cohort,
and for these analyses we chose the 1955 birth cohort as
the reference because it is midway among the birth cohorts
examined. We then analysed cancers for which there was a
reversal in IRR trend between older and younger cohorts
(such that decreasing trends in IRR for older cohorts
began to increase for younger cohorts), and for these
analyses, the birth cohort with the lowest incidence rate of
all cohorts was selected as an alternative reference, with
the most recent reversal birth year being selected as the
reference if there was more than one change in trend.
Next, we calculated the average annual percentage changes
(AAPCs) from 2000 to 2019 in model-based estimates of
incidence or mortality rates that were specific to each age
group.10 We used a Wald test to assess the significance of
the non-linear components of the cohort rate ratio curves
www.thelancet.com/public-health Vol 9 August 2024
(birth cohort deviation) and the heterogeneity in AAPCs
across age groups. Finally, we visualised trends in observed
age-specific incidence or mortality rates by birth cohort
(canonical plots) to inspect so-called non-parallelism,
which is indicative of the birth cohort effect. All age-periodcohort analyses were conducted overall and by sex.
Two-sided p values of less than 0·05 were considered to
be statistically significant. Not adjusting for multiple
comparisons was preferred because the analyses were
exploratory rather than testing hypotheses, and the data
under evaluation (cancer incidence and mortality) were
actual observations in nature rather than random
numbers.11 Correcting for multiple comparisons could
increase the type 2 errors, potentially causing important
findings (eg, subtle trends or those based on a relatively
small number of cases or deaths) to be overlooked. We
tabulated the number of cancer cases and deaths along
with population estimates using SEER*Stat (version 8.4.2)
and an established web tool for age-period-cohort
analysis10 to obtain estimable functions of the age-periodcohort model.
Role of the funding source
The funder had no role in study design, data collection,
data analysis, data interpretation, writing of the
manuscript, or the decision to submit for publication.
Results
We extracted data for 23 654 000 patients diagnosed with
34 types of cancer and 7 348 137 deaths from 25 cancers
for the period Jan 1, 2000, to Dec 31, 2019 (frequencies by
cancer type are in the appendix [pp 5–6]).
Birth cohort trends in cancer incidence and mortality
rates for 11 of 34 cancers for which IRRs increased with
each successive birth cohort born since approximately 1920
and across the 1950 to 1990 birth cohorts are shown in
figure 1 (all estimates and 95% CIs are in the
appendix [pp 9–17]). Data for male and female individuals
combined were presented if IRRs increased in both sexes;
otherwise, sex-specific IRRs with increasing trends were
provided (all sex-specific data are in the appendix [p 28]).
Compared with individuals in the 1955 birth cohort, the
incidence rate was approximately two-to-three times
higher in the 1990 birth cohort for cancers of the small
intestine (IRR 3·56 [95% CI 2·96–4·27]), thyroid (3·29
[2·91–3·73]), kidney and renal pelvis (2·92 [2·50–3·42]),
and pancreas (2·61 [2·22–3·07]). Additionally, female
individuals in the 1990 birth cohort had higher incidence
of liver and intrahepatic bile duct (IRR 2·05 [1·23–3·44])
and those in the 1985 birth cohort had a higher incidence
of non-HPV-associated oral or pharyngeal (1·26
[1·14–2·61]) cancers than did those in the 1955 birth cohort.
The birth cohort effect on the incidence trends, tested for
deviations in IRRs from linear trends, was significant for
eight of 11 cancers (pcohort<0·050), and was non-significant
for thyroid, soft tissue including heart, and female
oesophageal adenocarcinoma (appendix pp 25–26).
e585
Articles
Small intestine
3·0
3·56
2·27
Cohort rate ratio
3·29
2·06
0·90
0·82
0·75
0·95
0·55 0·57
0·62
0·76
0·69
2·36
2·17
1·70
1·60
1·44 1·48 1·44 1·39
1·40
1·35
1·12
1·01 0·99 1·02 1·00
1·07
0·99
1·00
0·90 0·96
0·88 0·81
0·84
2·92
2·75
2·55
1·86
2·0
1·0
0·9
0·8
0·7
0·6
0·5
0·4
Kidney and renal pelvis
Thyroid*
Incidence
Mortality
4·0
1·19
1·05 1·07 1·02 1·02 1·00
1·05 1·061·02 1·08 1·05 1·02
1·08
0·86
0·77
1·00 0·97
0·94 0·99
1·80
1·34
1·21
1·44
1·10
1·09
1·00
1·00
0·94
0·99
0·860·90
0·81
0·79
0·73 0·76 0·78
0·84
0·73
0·62
0·53
0·39
0·35
0·3
Pancreas
2·53
0·89
0·70 0·70
0·72 0·71
0·44
Liver and intrahepatic bile duct (female)†
Myeloma†
4·0
3·0
2·611
2·08
8
Cohort rate ratio
2·0
1·75 1·73
1·59
9
1·22
1·28
8
6
1·11 1·16
1·0
0·9
0·8
0·7
0·6
0·5
0·4
0·85
1·05
0·94
0·93 0·95 0·97 0·99 1·00
1·000·96
0·941·00
0·89
0·88
0·89
0·83
0·86
0·76 0·79
1·00
0·89 0·92
0·69 0·72
1·07 1·04 1·02
2·05
1·56 1·59 1·59 1·54
1·36
0·49
4
0·44
0·40
4
0·30
0·3
0·53
0·78
0·64
0·74
0·58
0·66
0·71
0·75
0·81
1·13
0·85 0·90
1·07
1·00
1·00
0·95
0·88
1·35
1·21
7
0·78
7
0·71
0·54
0·35
Oesophageal adenocarcinoma (female)*†‡
1·27
1·222
1
1·33
0 1·09
1·04
0 1·05
0
1·00
0·72
1·41
1·52
0·63
0·45
0·40
1·62
6
1·78
0·57
0·46
Gastric cardia‡
Leukaemia
4·0
3·0
Cohort rate ratio
2·0
1
1·12
1 1·16
1·0
0·9
0·8
0·7
0·6
0·5
0·4
1·70 1·70 1·68 1·61
1·67
1·28
1·46
1·
1·4
1·17
8
8 0·89
0·850·85
8 8 0·88
9 9
0·92
9 0·960·97
1·00
0·95
9
0·94
940
9
1·07
10
07
1·01 1·02 1·00 1·00 1·011 0·98
10
00
0·990·99
09
98 1·00
1·19
1 1·17
0·90 0·920·94
1·48
1·31
1
1·15
1
1·0
1
0
0·950·980·980·98 1·00
1·06
0·91
911
1·10
11
0·82
0
0·
1·17
11
0·73
0·7
0
7
12
1·2
1·27
0·67
6
14
40
11·37
3 1·40
1·45
1·4
4
0·62
0·6
6
0·53
0·5
5 0·50
0·55
Non-HPV-related oral cavity and pharynx (female)†‡
19
2
19 0
2
19 5
3
19 0
19 35
4
19 0
4
19 5
5
19 0
19 55
6
19 0
6
19 5
7
19 0
19 75
8
19 0
8
19 5
90
0·3
Soft tissue including heart*
Birth cohort
4·0
3·0
Cohort rate ratio
2·0
1·0
0·9
0·8
0·7
0·6
0·5
0·4
0·89 0·91 0·93 0·920·920·920·92
1·00
1·07 1·06 1·04 1·13
1·16
1·26
2
0·890·94
0·79
7
0·95 0·99 0·991·00 0·991·00
8
0·85
8 0·89
0·92 0·95 0·97
1·04 1·02 1·04
1·00 1·000·98 1·05 1·00
1
1·09 1·17
1·08 1·12
1·17 1·18
1·04 1·01
19
2
19 0
2
19 5
3
19 0
19 35
4
19 0
4
19 5
5
19 0
19 55
6
19 0
6
19 5
7
19 0
19 75
8
19 0
8
19 5
90
19
2
19 0
2
19 5
3
19 0
19 35
4
19 0
4
19 5
5
19 0
19 55
6
19 0
6
19 5
7
19 0
19 75
8
19 0
8
19 5
90
0·3
Birth cohort
Birth cohort
Figure 1: Birth cohort incidence and mortality rate ratio trends from 1920 to 1990 for 11 cancers with a consistent increase in incidence across birth cohorts
in the USA, 2000–19
The shaded areas indicate 95% CIs. Data are for both male and female individuals combined, unless otherwise stated. The rate in the 1955 birth cohort was used as
the reference to calculate the relative risk. The panels are sorted by the magnitudes of the IRR for the latest birth cohort. IRR=incidence rate ratio. *The test for nonlinear components of the IRR curves (birth cohort deviation) is not significant (pcohort>0·050). †Due to the relatively small number of cases or deaths, the youngest
birth cohorts were truncated. ‡Mortality data were not available.
e586
www.thelancet.com/public-health Vol 9 August 2024
Articles
Among these 11 cancers, mortality data were available for
eight cancers, and, by contrast with the incidence trends,
birth cohort trends in mortality rates decreased
Average annual percentage change (%)
Average annual percentage change (%)
6
5
4
3
2
1
0
–1
–2
–3
–4
–5
–6
Average annual percentage change (%)
Small intestine
6
5
4
3
2
1
0
–1
–2
–3
–4
–5
–6
6
5
4
3
2
1
0
–1
–2
–3
–4
–5
–6
4·22
3·27 3·48 3·40 3·22
2·70
2·01 1·84
1·05
1·79 1·95 1·90 1·63
0·39
0·31
–0·24
–1·26
–0·71
3·18 3·23
3·58 3·75 3·56 3·45 3·36
3·24 3·19 3·41 3·10
1·41 1·60
0·75
–0·09 0·02 0·11
–0·25 –0·23 –0·01
0·39 0·54
0·09
–0·40 –0·39 –0·22
0·27
0·85
2·78
3·80 3·93
3·26
3·39
3
2·47
1·47
1·24
1·82
0·20
1·32
0·65
0·97 0·95 1·00 0·88
–0·57
–1·79
Liver and intrahepatic bile duct (female)†
–2·42
–2·15 –1·92 –1·93
–1·62
–1·12
–0·65
–0·25
Myeloma†
5·53
4·86
4·34
2·47
1·31
0·20
0·72 0·77 0·90 1·00 0·90 0·85 0·71 0·80 0·88
–0·18
–1·05
–1·17
1·95
2·61
–0·77
–0·17
0·22 0·33 0·38 0·37 0·44
1·34
0·63
4·15
7
3·70
3·35 3·49
2·27
11·37
0·50
0·91
2·95
33·43
0
2·50
11·84
·
2·73
2·52
9
8 1·98
7 1·83
1·78
1·80 2·02
1·54 1·33
2·06
1·60
1·18 1·11 1·10 1·16 1·25 1·37
–0·04
–0·77
–2·80
2·01
2·34
0·13
Oesophageal adenocarcinoma (female)*†‡
1·85
0·98
0·44
–2·14 –2·14 –2·26
Gastric cardia‡
0·49 0·56 0·59 0·54 0·41
4 1·25
1·21 1·46
1·04
0·36
–2·43 –2·43 –2·27
–2·07
–1·52
Leukaemia
1·13 1·28
0·12 0·28 0·35 0·35
0·85
0·86
0·47
0·13 0·17
0·33 0·37 0·40
–0·03 –0·09 –0·07 –0·11
–2·10
2
–2·09
30
Non-HPV-related oral cavity and pharynx (female)†‡
Average annual percentage change (%)
Kidney and renal pelvis
Thyroid*
Incidence
Mortality
Pancreas
6
5
4
3
2
1
0
–1
–2
–3
–4
–5
–6
(eg, myeloma and leukaemia), plateaued (eg, small
intestine and thyroid), or fluctuated (eg, pancreas and
kidney and renal pelvis) in the 1955 to 1990 birth cohorts
1·21
2
0·69
6
0·97 1·09
0·30
3 0·23
0·50
0·05 0·21
0·02
70
50
40
60
Age at diagnosis (years)
0·31
80
30
3 0·07
0·50 0·30
0 0·16 –0·12 0·04 0·01 0·33 0·41
70
50
40
60
Age at diagnosis (years)
–2·08
0 –2·20
2
40
50
–2·24
–2·46 –2·60
6
60
–1·60
70
–0·37
–0·91
80
Age at diagnosis (years)
0·75
7
0·58
5 0·72
4 0·38
4 0·45
0·49
3 0·41
4 0·36 0·53 0·57
–0·25 0·37
–0·11
30
Soft tissue including heart*
–1·76
7 –2·00
0
1·01
0·68
80
Figure 2: Age-specific average annual percentage changes in incidence or mortality rates from 2000 to 2019 for 11 cancers with a consistent increase in
incidence in the USA across birth cohorts from 1920 to 1990
The shaded areas indicate 95% CIs. Data are for both male and female individuals combined, unless otherwise stated. The vertical line at 50 years is indicative of the cutoff
for young adults, defined as individuals younger than 50 years. Average annual percentage changes were derived from the slope (derivative) of the cohort rate ratio curve
and are estimated trends in fitted rates over time specific to each age group. *The test for heterogeneity in average annual percentage changes (age deviation) is not
significant (page>0·05). †Due to the relatively small number of cases or deaths, the youngest age groups were truncated. ‡Mortality data were not available.
www.thelancet.com/public-health Vol 9 August 2024
e587
Articles
and intrahepatic bile duct and non-HPV-associated oral
and pharyngeal cancers, with rapid increases in
incidence among individuals aged 30–39 years and
55–64 years. By contrast, increases in incidence rates did
not vary significantly across age groups for thyroid
cancer, female oesophageal adenocarcinoma, and soft
tissue including heart cancer (page>0·050), consistent
with the age invariant period effect (pperiod<0·050; all test
results are in the appendix [pp 25–26]). Mortality rates
either decreased or stabilised in young adults (aged
25–49 years), except for female liver and intrahepatic bile
duct cancer, which increased by 1·95% (95% CI
0·59–3·34) annually among those aged 35–39 years
(page<0·050; all test results are in the appendix [p 27];
for all cancers, except for female liver and intrahepatic bile
duct cancer, for which, compared with those in the
1955 birth cohort, the 1980 birth cohort had an MRR of 1·36
(95% CI 1·09–1·69; all MRR estimates and 95% CIs are in
the appendix [pp 18–24]).
Age-specific AAPCs in incidence or mortality rates
from 2000 to 2019 for the 11 identified cancers are shown
in figure 2 (sex-specific results are shown in the
appendix [p 29]). Among young adults (aged 25–49 years),
the most rapid annual increases in incidence rates were
seen for pancreatic (4·34%, age 25–29 years), small
intestine (4·22%, age 25–29 years), and kidney and renal
pelvis (3·93%, age 35–39 years) cancers. In female
individuals, bimodal patterns were observed for liver
Uterine corpus, not otherwise specified
4·0
2·82
Cohort rate ratio
3·0
2·0
1·14
1·0
0·9
0·8
0·7
0·6
0·5
0·4
1·04 1·00 1·07
0·97 0·97 1·00
1·04
1·19
1·12
1·30
1·24
1·48 1·55
1·64
1·82
1·34 1·401·40
1·45
2·23
2·06
1·60
Oestrogen receptor-positive breast*
3·12
Gallbladder and other biliary†
3·49
2·69
2·42
2·09
1·82
1·02 1·07
1·04
4 1·04
1·14 1·10
1·09
1·12
1·20
1·26
1·38
1·73
1·50
1·86
1·20 1·14
1·06 1·00
1·01 1·02 1·03
1·08
1·14 1·23
1·02 1·02 1·00 1·00
0·95
9
0·89 0·90
8
0·85 0·84
1·00
3
1·32 1·36
4
1·45
1·66
1·11
0·93
1·06
0·81 0·80
Incidence
Mortality
0·3
Testis
Colon and rectum
Gastric non-cardia*
4·0
3·38
2·79
3·0
Cohort rate ratio
2·64
2·0
1·28
1·0
0·9
0·8
0·7
0·6
0·5
0·4
1·15
1·14
0·97
1·31
1·09 1·06
1·07
0·92
1·04
1·00
1·47 1·49 1·45
1·14 1·21
1·24 1·23
1·001·00
1·49 1·56 1·55
1·38 1·41
1
1·06 1·09 1·11
1·62
1·43
1·31
1·28 1·31
1·21
1·12
1·02 1·00 1·06
1·18
1·26
6
0
1·40
1·59
59
8 1·78
1·7
778 1·82
1·78
2·29
2·31
1·86
1·96
1·67
1·48
4
1·51
1·42
1·37
3
1·26
1·23
1·17
1·10
1·39
1·06
1·01
1·00
1·22
2
1·23
3
1·17
1·16
1·15
1·09
1·06 1·00
1·01 1·04
0·3
Anus, anal canal, and anorectum (male)†
Ovary
4·0
Cohort rate ratio
3·0
2·0
1·0
0·9
0·8
0·7
0·6
0·5
0·4
3·20
2
2·87
8
9
2·95
6
2·68
2·53
2·25
4
2·49
2·04
Kaposi sarcoma (male)*
3·38
2·25
1·83
2·47
1·98
1·61
1·75
1·40
1·55
1·27
1·38
1·22
1
1·22
1
1·12
1·061·00
0
1·02 1·03
1·17
1·09 1·02 1·00
9 0·88
0·91
8 0·92
1·17
0·98
0·72 0·76
0·79
0·85
0·99
0·82
0·92
2·57
2·38
2·48
2·14
1·60
1·41
3·03
1·25
1·01
1·08
1·00 0·99
1·00
2·51
2·39
2·35
1·61
1·68
1·22
1·21
1·06
1·02
1·00
0·69
0·63
0·56
0·43
0·45
0·33
19
2
19 0
2
19 5
3
19 0
19 35
4
19 0
4
19 5
5
19 0
19 55
6
19 0
6
19 5
7
19 0
19 75
8
19 0
8
19 5
90
19
2
19 0
2
19 5
3
19 0
19 35
4
19 0
4
19 5
5
19 0
19 55
6
19 0
6
19 5
7
19 0
19 75
8
19 0
8
19 5
90
19
2
19 0
2
19 5
3
19 0
19 35
4
19 0
4
19 5
5
19 0
19 55
6
19 0
6
19 5
7
19 0
19 75
8
19 0
8
19 5
90
0·3
Birth cohort
Birth cohort
Birth cohort
Figure 3: Birth cohort incidence and mortality rate ratios trends from 1920 to 1990 for nine cancers with reversing trends in incidence compared with older
birth cohorts in the USA, 2000–19
The shaded areas indicate 95% CIs. Data are for both male and female individuals combined, unless cancer is sex-specific or otherwise stated. The birth cohort with
the lowest incidence rate was used as the reference to calculate the relative risk. The panels are sorted by the magnitudes of the IRR for the latest birth cohort.
IRR=incidence rate ratio. *Mortality data were not available. †Due to the relatively small number of cases or deaths, the youngest birth cohorts were truncated.
e588
www.thelancet.com/public-health Vol 9 August 2024
Articles
colorectal cancers; and in the 1975 to 1980 birth cohorts for
ovarian cancer, anus, anal canal and anorectum cancer
(hereon referred to as anal cancer) in male individuals, and
Kaposi sarcoma in male individuals (all pcohort<0·050;
appendix pp 25–26). Compared with the birth cohort with
the lowest incidence rate, incidence rates in the 1990 birth
cohort were 12% (IRR1990 vs 1975 1·12 [95% CI 1·03–1·21] for
ovarian cancer) to 169% (IRR1990 vs 1930 2·69 [2·34–3·08] for
uterine corpus cancer) higher (appendix pp 9–17). Among
these nine cancers, mortality data were available for six
cancers. The birth cohort trends in mortality rates generally
mirrored those of the incidence rates (sex-specific results
are in the appendix [pp 18–24]) with the steepest increase
in MRR estimated for uterine corpus cancer
age-specific mortality estimates are shown in the
appendix [pp 18–24]).
Birth cohort trends for the nine of 34 cancers for which
IRRs increased in younger birth cohorts after previous
short-term or long-term declines in older birth cohorts are
shown in figure 3. Data for male and female individuals
combined were presented if IRRs increased after a
previous decrease in both sexes, otherwise, sex-specific
IRRs were provided (sex-specific results are in the
appendix [p 28]). The reversal in IRRs occurred in the
1930 to 1935 birth cohorts for uterine corpus, oestrogen
receptor-positive breast, and gallbladder and other biliary
cancers; in the 1940 birth cohort for testicular cancer; in
the 1950 to 1955 birth cohorts for non-cardia gastric and
Average annual percentage change (%)
Average annual percentage change (%)
6
5
4
3
2
1
0
–1
–2
–3
–4
–5
–6
Average annual percentage change (%)
Uterine corpus,not otherwise specified
6
5
4
3
2
1
0
–1
–2
–3
–4
–5
–6
6
5
4
3
2
1
0
–1
–2
–3
–4
–5
–6
Oestrogen receptor-positive breast*
Gallbladder and other biliary†
Incidence
Mortality
2·93 2·99 2·83
2·13 1·93
2·66 2·81 2·48
1·79
1·87
1·36 1·50
1·49
0·87
2·14
1·77
0·47 0·82
1·80
2·27 1·93
1·33
0·61
1·46
0·52
2·18 2·12
1·60 1·48
4 1·43
4
0·60
6
0
0·86
8 0·02
0·89
8
0·23
–0·03
0
1·21
–0·54
1·42 1·42 1·29
0·48
–0·42 –0·45
1·81
8
1·33
0·64
0·22
0·68 0·88
2·78
–0·49
0 1·82
2·02
8 1·60
0·98
0·89
1·59 1·51 1·46
0·77
–1·02
–0·84
1·11 0·91
–0·64
0·44 0·20
–0·45 –0·63
–1·31
–0·07 –0·14
–0·78–1·15
–1·17 –1·21
–1·61 –1·61
0·21
–0·99
–0·11
–2·12
–1·28
–1·94
–2·72
–3·00
–3·05
–3·41 –3·32
–3·73
Anus, anal canal, and anorectum (male)†
Ovary
0·12
1·11
0·59
0·02
–0·38
–1·22
0·79
Colon and rectum
–0·27 –0·18
–0·38
–1·06 –0·82
3
2·36
0·76
1·19
0·57
Gastric non-cardia*
2·39
1·39
1·19 1·34
1·04
–0·22
–0·65
Testis
1·54
–4·10 –4·01 –3·88
Kaposi sarcoma (male)*
5·04
3·64 3·88
3·53
3·28
2·00
0·70
7
0·15
1
–0·57
–1·27
–0·10
–1·47
–0·43
–1·10
1
–1·85
–2·19
1
–1·66
–2·08
–2·47 –2·37
–2·50 –2·37 –2·36 –2·44 –2·45
–2·00
–1·74 –1·67
3·03
4·09
3·63
3·25 3·34
2·32
1·48 1·31
0·37
1·05
1·49
0·93
0·87
0·17
–0·16
–0·12
–1·15
–2·19 –2·11 –2·27
–3·15
–2·54
40
50
60
70
80
Age at diagnosis (years)
30
40
50
60
Age at diagnosis (years)
70
80
–2·16
–2·13
–2·82
30
–3·09
–5·76
–5·88
30
–0·77
–2·32
40
50
60
70
80
Age at diagnosis (years)
Figure 4: Age-specific average annual percentage changes in incidence or mortality rates from 2000 to 2019 for nine cancers with reversing trends in
incidence compared with older birth cohorts in the USA
The shaded areas indicate 95% CIs. Data are for both male and female individuals combined, unless cancer is sex-specific or otherwise stated. The vertical line at
50 years is indicative of the cutoff for young adults, defined as individuals younger than 50 years. Average annual percentage changes were derived from the slope
(derivative) of the cohort rate ratio curve and represent estimated trends in fitted rates over time specific to each age group. *Mortality data were not available. †Due
to the relatively small number of cases or deaths, the youngest age groups were truncated.
www.thelancet.com/public-health Vol 9 August 2024
e589
Articles
(MRR1990 vs 1930 3·49 [95% CI 2·49–4·88; appendix pp 18–24]).
The birth cohort effect on the mortality trends, tested for
deviations in MRRs from linear trends, was significant for
all six cancers (pcohort<0·050; appendix 27).
Age-specific AAPCs in incidence or mortality rates
from 2000 to 2019 for these nine cancers are shown in
figure 4 (sex-specific AAPC incidence and mortality
results are in the appendix [pp 9–24]). For uterine corpus,
oestrogen receptor-positive breast, and gallbladder and
other biliary cancers, incidence rates increased across
almost all age groups, with younger age groups having
faster increases than older age groups. For non-cardia
gastric and colorectal cancers, increases in incidence
were confined to younger age groups (age 30–59 years for
non-cardia gastric and 30–54 years for colorectal cancer).
Male individuals aged 25–29 years had a 3·53% (95% CI
1·25 to 5·87) annual increase in anal cancer and a nonsignificant 0·87% (–0·09 to 1·84) annual increase for
Kaposi sarcoma, contrasting with adjacent older age
groups (30–49 years) who had rapid declines in incidence.
A similar pattern was found for testicular cancer with
male individuals aged 25–34 years having a 0·68%
(95% CI 0·42 to 0·94) to 0·88% (0·66 to 1·09) annual
increase in incidence. For ovarian cancer, the decreasing
incidence trends lessened with progressively younger
age groups; and the incidence rate increased in those
aged 25–29 years by 0·70% per year (95% CI 0·19–1·21).
Age-specific mortality rates increased in some young
adult age groups (25–49 years) for cancers of the
gallbladder and other biliary and colorectum. Uterine
corpus cancer increased in age-specific mortality across
all age groups, with greater increases in younger groups,
and testicular cancer and anal cancer in male individuals
had a bimodal increase in age-specific mortality, one
among younger adults (age 25–34 years for testicular and
age 30–34 years for anal cancer in male individuals) and
another among older cohorts (age 55–64 years for
testicular and age 45–84 years for anal cancer in male
individuals).
Birth cohort trends for the remaining 14 cancers
decreased or were relatively stable over successive birth
cohorts from approximately 1955 to 1990 (appendix p 30).
Unlike other cancers, the IRRs for vulva and HPVrelated oral and pharyngeal cancers decreased after the
1960 birth cohort and for melanoma decreased after the
1980 birth cohort, after having shown a steady increase
in incidence across older birth cohorts. Other notable
observations are the sharp downturn in IRRs for cervical
cancer between the 1985 and 1990 birth cohorts and the
decelerations of long-term declines in IRRs for smokingrelated cancers (eg, lung, larynx, and oesophageal
squamous cell carcinoma) for birth cohorts born around
1975. Sex-specific patterns were similar to each other
except for the aforementioned smoking-associated
cancers and melanoma, with a lag of 10 to 30 years in the
female birth cohorts exhibiting these trends compared
with male birth cohorts (appendix p 32).
e590
The corresponding age-specific AAPCs across
2000 to 2019 for the 14 cancers with decreasing or stable
birth cohort trends are in the appendix (p 31). In young
adults (age 25–49 years), incidence rates for all these
cancers generally decreased or stabilised, with mortality
rates often declining more sharply than incidence rates
(eg, lung cancer and lymphomas). A sharp decline in
the incidence of cervical cancer was seen in females
aged 25–29 years (–1·45% per year), which was
steeper than declines among those aged 30–39 years
(–0·49% to –0·63% per year).
Consistent with the trends based on the fitted rates
(ie, the modelled rates), trends in observed age-specific
incidence increased in some age groups among young
adults (age 25–49 years) for the 20 cancers with increasing
IRRs (appendix p 34; corresponding mortality data are in
the appendix [p 35]), showing that the birth cohort
findings were not an artifact of modelling. Age-specific
incidence rates for the 14 cancers with decreasing or
stable IRRs are shown in the appendix (p 36; corre­
sponding mortality trends are in the appendix [p 37]).
Discussion
Based on a comprehensive model-based cohort analysis of
nationwide, high-quality, cancer incidence and mortality
data, we found that incidence rates for 17 of the 34 cancer
types studied increased in progressively younger birth
cohorts in the USA, after adjusting for age and period
effect (pcohort<0·050). These findings expand on our
previous research,1 adding eight cancer types (cardia
gastric, small intestine, oestrogen receptor-positive breast,
ovary, liver and intrahepatic bile duct and non-HPVassociated oral and pharynx in female individuals, and
anus and Kaposi sarcoma in male individuals) to the list of
cancers previously identified (colorectum, uterine corpus,
gallbladder and other biliary, kidney and renal pelvis,
pancreas, myeloma, non-cardia gastric, testis, and
leukaemia). For liver and intrahepatic bile duct in female
individuals, uterine corpus, gallbladder and other biliary,
testicular, and colorectal cancers, mortality trends mirrored
incidence trends, suggesting the increase in incidence in
younger generations is substantial enough to outweigh
improvements in cancer survival. The increasing cancer
incidence rates over the successive younger generations
up to those born around 1990 suggest there have been
increases in the prevalence of carcinogenic exposures
during early life or young adulthood in this generation,
which have yet to be elucidated.
Ten of 17 cancers with increasing incidence in younger
birth cohorts are obesity-related cancers (colorectum,
kidney and renal pelvis, gallbladder and other biliary,
uterine corpus, pancreas, cardia gastric, oestrogen
receptor-positive breast, ovary, myeloma, and liver and
intrahepatic bile duct),12 suggesting a potential role of
obesity in emerging cancer trends in recent generations.
Since the late 1970s, the obesity epidemic in the USA has
affected individuals across all age groups; however, the
www.thelancet.com/public-health Vol 9 August 2024
Articles
most rapid rise has been seen among individuals aged
2–19 years,13 indicative of birth cohort effects. This
hypothesis is supported by analyses of nationally
representative survey data from the USA indicating
increasing odds of developing obesity, type 2 diabetes, and
metabolic syndrome among generations born after those
born in 1946–64 (known as the Baby Boomer generation),14–17
and that has also been linked to the increase in premature
cardiometabolic mortality since 201118 and the decrease in
life expectancy since 2014.19 Accumulating evidence
underscores the substantial influence of bodyweight in
early life, as highlighted by a recent study associating
excess bodyweight during early adulthood (ages
18–40 years) with up to 18 cancers.20
The effect of exposures to other suspected risk factors
(eg, unhealthy diet, sedentary lifestyle, altered sleep
patterns, and environmental chemicals) during early life
and young adulthood on cancer risk,3,21 as well as birth
cohort trends in the prevalence of these factors, remain
poorly understood. Some dietary patterns (eg, high in
saturated fats, refined grain, ultra-processed foods, and
sugar, as well as meat vs plant diets) and sedentary
behaviours during adolescence and young adulthood have
been associated with increased risk of some cancers, most
consistently for colorectal and breast cancers.3,21 The
constellation of gastrointestinal tract cancers (including
those not necessarily associated with obesity—eg, small
intestine and non-cardia gastric cancers) that are
increasing in incidence in recent birth cohorts highlight
the potential involvement of an altered microbiome, given
substantial changes in dietary patterns and antibiotics use
in particular over the past few decades.22,23 Evidence from
multiple sources, ranging from murine models to
prospective cohort studies, has linked specific microbes or
associated dietary patterns to oral and gastrointestinal
tract cancer carcinogenesis.24–26 Microbial dysbiosis via
various exposures, leading to heightened oxidative stress
and inflammation, is thought to promote epigenetic
alterations and somatic mutations via interactions with
host genetics and the immune system.24,25
Historical changes in reproductive patterns probably
had at least a small role in shaping birth cohort trends
for cancers in female individuals. In particular, the nadir
of the IRR curves for oestrogen receptor-positive
breast and uterine cancers among women born in
approximately 1930 corresponds to the peak fertility rate
during 1946–64 (ie, the birth of the Baby Boomer
generation). The subsequent increase (in the
1930–50 birth cohorts) and deceleration, and then
increase in IRRs align with trends in fertility rates,
which reduced (1957–76), stabilised, and then decreased.
Reasons for the reversal in the birth cohort trend
observed for ovarian cancer since 1975 are unclear;
however, it might in part reflect the decline in oral
contraceptive use from the early 1970s to the mid-1990s,
considering the inverse association of oral contraceptive
use and ovarian cancer risk.27
www.thelancet.com/public-health Vol 9 August 2024
For liver and non-HPV-associated oral and pharyngeal
cancers, incidence rates increased in women in
successive birth cohorts born since approximately 1970,
whereas rates decreased or stabilised in men. Increasing
birth cohort trends in women are consistent with lifetime
trajectories in alcohol-related behaviours among those
born in the late 1970s and early 1980s, as well as
acceleration in binge drinking from 1990 to 2010, when
individuals in this birth cohort would have been aged
30–49 years.28 Conversely, the stabilisation of liver and
intrahepatic bile duct cancer incidence and the decline in
the incidence of oral and pharyngeal cancer among
recent male birth cohorts might be attributable, at least
in part, to decreased alcohol consumption and smoking
among male individuals aged 18–25 years.28 The effect of
the increase in the prevalence of hepatitis C virus
infection since approximately 2010, which dispro­
portionately affects young adults,29 on trends in liver
cancer remains to be determined, given the latency from
viral exposure to the onset of hepatocellular carcinoma.
The rising incidence rate of anal cancer and Kaposi
sarcoma in recent male birth cohorts is a reversal from
previous trends. The partially similar patterns of both
malignancies might indicate the influence of shared risk
factors, such as HIV infection.30 The escalating incidence
rate in the 1945–65 birth cohorts in men is probably
indicative of the AIDS epidemic during the 1980s, and its
peak in 1993, when these birth cohorts were aged
20–49 years.31 The subsequent sharp decrease in the birth
cohort incidence of these two cancers aligns with the swift
decline in AIDS incidence, reflecting major advances in
antiviral therapies during the mid-1990s.32 Such treatment
advances probably manifested as a period effect, affecting
incidence rates across all age groups. The increase in
incidence of anal cancer and Kaposi sarcoma in male
individuals among birth cohorts born since 1975–80 is
somewhat unexpected and highlights the need for further
investigation, especially considering the increase in the
number of men with HIV aged 25–34 years, contrasting
with a continued decrease in older adults33 and increased
screening (for anal cancer) in high-risk individuals.34
The accelerated downturn in the birth cohort trend of
cervical cancer incidence shows the effectiveness of HPV
vaccination, particularly in women born in approximately
1990, who were approximately age 16 years when HPV
vaccination was first approved in the USA for women
and girls aged 9–26 years in 2006.35 The strong downward
birth cohort trends for mainly smoking-associated
cancers (eg, lung, larynx, and oesophageal squamous cell
carcinoma) probably reflect the rapid decline in the
prevalence of smoking among younger birth cohorts;
however, signs of slowed progress in the youngest birth
cohorts (approximately from 1975 birth cohort onward)
warrant continuous monitoring with additional years of
data.
Declines in birth cohort trends for mortality rates,
observed for many cancers that showed increases in the
e591
Articles
birth cohort trends for incidence, should be interpreted
with caution. The seemingly contradictory pattern of
increasing incidence and decreasing mortality could be
because of early detection or treatment advancements, or
a combination of both, which improves cancer survival.
For example, for pancreatic cancer, the improvement in
5-year survival among young adults (aged <50 years;
from 16·5% in 2000 to 37·2% in 2016 in the USA36)
might have outpaced the effect of increasing incidence.
For cancers with decreasing birth cohort trends for both
incidence and mortality, mortality often declined more
sharply than incidence, reflecting both successful
primary prevention efforts and treatment breakthroughs
during the past few decades (eg, lung cancer).37,38
Additionally, the increased use of diagnostic testing and
overdiagnosis probably contributed to the rise in
incidence rates of cancers that are highly sensitive to
diagnostic scrutiny (eg, thyroid and kidney cancers),
without increasing mortality trends.
The findings of increased cancer incidence in recent
birth cohorts for 17 cancer types have important public
health implications. The heightened cancer risk in recent
birth cohorts has already translated into an increasing
cancer morbidity and mortality burden in young adults,
posing unique challenges to the health-care system and
community. Improving awareness among health-care
providers and the general public about the signs and
symptoms of cancer among young adults is crucial for
early detection and treatment. Because most existing
cancer care systems in the USA are not tailored to the
unique needs of young adults with cancer, targeted
programmes and services are needed. Additionally, our
findings should prompt a more vigorous application of
the life course framework in epidemiological studies to
fill significant gaps in understanding birth-cohortspecific exposures and the trajectories of cancer risk,
beginning as early as adolescence, childhood, infancy,
and even prenatal stages,39 prioritising modifiable
exposures. Funding should also support research on
the development of interventions that align with the
influential social and cultural context, values, and
preferences of younger generations. Without effective
population-level inter­ventions, the maturation of younger
generations could lead to an overall increase in cancer
burden in the future, halting or reversing decades of
progress against cancer.
The strengths of this study include the use of highquality representative population-based data; an analysis
of both incidence and mortality trends; the use of ageperiod-cohort modelling; and a comprehensive and
systematic examination of various cancers.
The major limitation of this study stems from the
so-called identifiability issue that is inherent to all
age-period-cohort modelling studies, where changes in
disease incidence cannot be separately attributed to each
element because of their co-linear nature.40 Nevertheless,
statistical testing of cohort deviations, adjusted for age
e592
and period effects, uniquely captured changes in cancer
incidence attributable to the birth cohort effect and
signifies generational differences in cancer risk.
Furthermore, all interpretations were based on modelderived quantities; the cohort rate ratio curve is a
composite curve, integrating observed data with projected
rates under the assumption that the age-period-cohort
model provides an adequate description of the data.6,10
Partly because of the absence of data in cancer registries,
we were unable to assess the contribution of secular
changes in plausible risk factors or diagnostic practices to
changes in cancer incidences in younger generations.
Assessing incidence trends by cancer stage and method
of detection (eg, patient symptoms or signs, screening, or
incidental findings) could help differentiate the effect on
incidence trends of diagnostic changes versus
generational population risk changes. Additionally, some
cancers are likely to have been misclassified. In particular,
the classification of HPV-associated versus non-HPVassociated oral and pharyngeal cancers was solely
determined by histology without information on risk
factors such as HPV infection and smoking status.8,9
Furthermore, the accuracy of cancer mortality statistics
based on death certificates varies by cancer type, although
the concordance with population-based cancer registry
data is generally high for common cancers.41 Additionally,
the fact that our study focused specifically on US adult
populations might limit the generalisability of our
findings. Finally, aggregated data obscures health
disparities among diverse socioeconomic and demo­
graphic groups, including by race and ethnicity,
warranting future studies to identify high-risk popu­
lations and intervention priorities.
In conclusion, controlling for age and period effects
on cancer incidence rates, each successive generation
born during the second half of the 20th century has had
increased incidences of many common cancer types of
heterogeneous aetiologies compared with preceding
generations in the USA. Our findings highlight the
importance of early lifetime exposures and highlight
crucial opportunities to prevent a substantial fraction of
cancer occurrence through modification of enviro­
nmental and lifestyle risk factors. Extensive efforts are
needed to identify underlying risk factors responsible for
these trends to inform prevention strategies.
Contributors
HS contributed to study conceptualisation, data curation, investigation,
methods, project administration, supervision, validation, and writing of
the original manuscript draft and review and editing. CJ contributed to
investigation, method, data curation, formal analysis, validation, design
of graphs, and reviewing and editing the manuscript. AJ contributed to
study conceptualisation, investigation, supervision, and reviewing and
editing the manuscript. AM, FI, MF-B, PB and RLS contributed to the
investigation and reviewing and editing the manuscript. HS and CJ
accessed and verified the underlying data. All authors had full access to
all the data in the study, and final responsibility for the decision to
submit for publication.
Declaration of interests
We declare no competing interests.
www.thelancet.com/public-health Vol 9 August 2024
Articles
Data sharing
The NAACCR Incidence Data, the SEER Database, and the US Mortality
Database (listed in the appendix [p 38]) are non-confidential and publicly
accessible data, and can be accessed via a signed Data Use Agreement
and distributed through SEER*Stat (version 8.4.2).
20
Acknowledgments
This study was funded by the Intramural Research Department of the
American Cancer Society. We thank all cancer registry staff for their
diligence in collecting cancer information, without which this research
could not have been conducted. We thank Philip S Rosenberg from
National Cancer Institute for the helpful discussion on the statistical
methods used in the study.
22
References
1
Sung H, Siegel RL, Rosenberg PS, Jemal A. Emerging cancer trends
among young adults in the USA: analysis of a population-based
cancer registry. Lancet Public Health 2019; 4: e137–47.
2
Koh B, Tan DJH, Ng CH, et al. Patterns in cancer incidence among
people younger than 50 years in the US, 2010 to 2019.
JAMA Netw Open 2023; 6: e2328171.
3
Ugai T, Sasamoto N, Lee HY, et al. Is early-onset cancer an
emerging global epidemic? Current evidence and future
implications. Nat Rev Clin Oncol 2022; 19: 656–73.
4
di Martino E, Smith L, Bradley SH, et al. Incidence trends for
twelve cancers in younger adults—a rapid review. Br J Cancer 2022;
126: 1374–86.
5
Heer EV, Harper AS, Sung H, Jemal A, Fidler-Benaoudia MM.
Emerging cancer incidence trends in Canada: the growing burden
of young adult cancers. Cancer 2020; 126: 4553–62.
6
Rosenberg PS, Miranda-Filho A. Cancer incidence trends in
successive social generations in the US. JAMA Netw Open 2024;
7: e2415731.
7
Doll R. Progress against cancer: an epidemiologic assessment.
The 1991 John C. Cassel Memorial Lecture. Am J Epidemiol 1991;
134: 675–88.
8
Kreimer AR, Chaturvedi AK, Alemany L, et al. Summary from an
international cancer seminar focused on human papillomavirus
(HPV)-positive oropharynx cancer, convened by scientists at IARC
and NCI. Oral Oncol 2020; 108: 104736.
9
Tota JE, Best AF, Zumsteg ZS, Gillison ML, Rosenberg PS,
Chaturvedi AK. Evolution of the oropharynx cancer epidemic in the
United States: moderation of increasing incidence in younger
individuals and shift in the burden to older individuals. J Clin Oncol
2019; 37: 1538–46.
10 Rosenberg PS, Check DP, Anderson WF. A web tool for age-periodcohort analysis of cancer incidence and mortality rates.
Cancer Epidemiol Biomarkers Prev 2014; 23: 2296–302.
11 Rothman KJ. No adjustments are needed for multiple comparisons.
Epidemiology 1990; 1: 43–46.
12 Lauby-Secretan B, Scoccianti C, Loomis D, Grosse Y, Bianchini F,
Straif K. Body fatness and cancer—viewpoint of the IARC Working
Group. N Engl J Med 2016; 375: 794–98.
13 Fryar CD, Carroll MD, Afful J. Prevalence of overweight, obesity,
and severe obesity among children and adolescents aged 2–19 years:
United States, 1963–1965 through 2017–2018. NCHS Health E-Stats,
2020. https://www.cdc.gov/nchs/data/hestat/obesity-child-17-18/
obesity-child.htm (accessed May 24, 2024).
14 Reither EN, Hauser RM, Yang Y. Do birth cohorts matter?
Age-period-cohort analyses of the obesity epidemic in the United
States. Soc Sci Med 2009; 69: 1439–48.
15 Robinson WR, Keyes KM, Utz RL, Martin CL, Yang Y. Birth cohort
effects among US-born adults born in the 1980s: foreshadowing future
trends in US obesity prevalence. Int J Obes (Lond) 2013; 37: 448–54.
16 Fishman EI, Stokes A, Preston SH. The dynamics of diabetes
among birth cohorts in the U.S. Diabetes Care 2014; 37: 1052–59.
17 Zheng H, Echave P. Are recent cohorts getting worse? Trends in US
adult physiological status, mental health, and health behaviors
across a century of birth cohorts. Am J Epidemiol 2021; 190: 2242–55.
18 Shah NS, Lloyd-Jones DM, Kandula NR, et al. Adverse trends in
premature cardiometabolic mortality in the United States, 1999 to
2018. J Am Heart Assoc 2020; 9: e018213.
19 Olshansky SJ, Passaro DJ, Hershow RC, et al. A potential decline in
life expectancy in the United States in the 21st century. N Engl J Med
2005; 352: 1138–45.
www.thelancet.com/public-health Vol 9 August 2024
21
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Recalde M, Pistillo A, Davila-Batista V, et al. Longitudinal body
mass index and cancer risk: a cohort study of 2.6 million Catalan
adults. Nat Commun 2023; 14: 3816.
Murphy CC, Zaki TA. Changing epidemiology of colorectal
cancer—birth cohort effects and emerging risk factors.
Nat Rev Gastroenterol Hepatol 2024; 21: 25–34.
Browne AJ, Chipeta MG, Haines-Woodhouse G, et al. Global
antibiotic consumption and usage in humans, 2000–18: a spatial
modelling study. Lancet Planet Health 2021; 5: e893–904.
Whelan K, Bancil AS, Lindsay JO, Chassaing B. Ultra-processed
foods and food additives in gut health and disease.
Nat Rev Gastroenterol Hepatol 2024; 21: 406–27.
Zepeda-Rivera M, Minot SS, Bouzek H, et al. A distinct
Fusobacterium nucleatum clade dominates the colorectal cancer
niche. Nature 2024; 628: 424–32.
Nguyen LH, Cao Y, Hur J, et al. The sulfur microbial diet is
associated with increased risk of early-onset colorectal cancer
precursors. Gastroenterology 2021; 161: 1423–32.e4.
Mima K, Kosumi K, Baba Y, Hamada T, Baba H, Ogino S.
The microbiome, genetics, and gastrointestinal neoplasms: the
evolving field of molecular pathological epidemiology to analyze the
tumor-immune-microbiome interaction. Hum Genet 2021;
140: 725–46.
Webb PM, Green AC, Jordan SJ. Trends in hormone use and
ovarian cancer incidence in US White and Australian women:
implications for the future. Cancer Causes Control 2017; 28: 365–70.
Keyes KM, Jager J, Mal-Sarkar T, Patrick ME, Rutherford C,
Hasin D. Is there a recent epidemic of women’s drinking? A critical
review of national studies. Alcohol Clin Exp Res 2019; 43: 1344–59.
Ryerson AB, Schillie S, Barker LK, Kupronis BA, Wester C. Vital
signs: newly reported acute and chronic hepatitis C cases - United
States, 2009–2018. MMWR Morb Mortal Wkly Rep 2020; 69: 399–404.
Zhang ER, Pfeiffer RM, Austin A, et al. Impact of HIV on anal
squamous cell carcinoma rates in the United States, 2001–2015.
J Natl Cancer Inst 2022; 114: 1246–52.
Centers for Disease Control and Prevention (CDC). HIV and
AIDS—United States, 1981-2000. MMWR Morb Mortal Wkly Rep
2001; 50: 430–34.
National Institute of Allergy and Infectious Diseases. Antiretroviral
drug discovery and development. https://www.niaid.nih.gov/diseasesconditions/antiretroviral-drug-development#:~:text=In%20March%20
1987%2C%20AZT%20became,reverse%20transcriptase%20
inhibitors%2C%20or%20NRTIs (accessed May 24, 2024).
US Centers for Disease Control and prevention. HIV surveillance
report. HIV diagnoses; 2008–2023; 25–34; all races/ethnicities; both
sexes; all transmission categories; United States. https://gis.cdc.
gov/grasp/nchhstpatlas/charts.html (accessed May 24, 2024).
Haas CB, Engels EA, Palefsky JM, et al. Severe anal intraepithelial
neoplasia trends and subsequent invasive anal cancer in the United
States. J Natl Cancer Inst 2024; 116: 97–104.
Markowitz LE, Dunne EF, Saraiya M, Lawson HW, Chesson H,
Unger ER. Quadrivalent human papillomavirus vaccine:
recommendations of the Advisory Committee on Immunization
Practices (ACIP). MMWR Recomm Rep 2007; 56: 1–24.
SEER*Explorer: An interactive website for SEER cancer statistics.
Surveillance, Epidemiology, and End Results Program, National
Cancer Institute, 2024. https://seer.cancer.gov/statistics-network/
explorer/ (accessed June 2, 2024).
Rittmeyer A, Barlesi F, Waterkamp D, et al. Atezolizumab versus
docetaxel in patients with previously treated non-small-cell lung
cancer (OAK): a phase 3, open-label, multicentre randomised
controlled trial. Lancet 2017; 389: 255–65.
Howlader N, Forjaz G, Mooradian MJ, et al. The effect of advances
in lung-cancer treatment on population mortality. N Engl J Med
2020; 383: 640–49.
Wagner C, Carmeli C, Jackisch J, et al. Life course epidemiology
and public health. Lancet Public Health 2024; 9: e261–69.
Holford TR. Understanding the effects of age, period, and cohort on
incidence and mortality rates. Annu Rev Public Health 1991; 12: 425–57.
German RR, Fink AK, Heron M, et al. The accuracy of cancer
mortality statistics based on death certificates in the United States.
Cancer Epidemiol 2011; 35: 126–31.
e593