Abstract
The burden of lung cancer in Hong Kong continues to rise. We analyzed trends in lung cancer incidence and associations with age, period, and cohort from 1985 to 2019, made projections up to 2030 and examined the drivers of lung cancer incidence. We used age-period-cohort modeling to estimate age, period, and cohort effects on lung cancer incidence rates in Hong Kong between 1985 and 2019. We projected lung cancer incidence in Hong Kong from 2020 to 2030 using Bayesian age-period-cohort analysis with an integrated nested Laplace approximation. We decomposed changes in the number of lung cancer cases into population growth, population aging, and epidemiologic changes. From 1985 to 2019, the number of lung cancer incident cases in Hong Kong continued to rise, yet the age-standardized incidence rates have declined for both sexes while have fluctuated for females over the past two decades. The overall annual percentage change from 1985 to 2019 was -2.29 (95% CI, -2.53 to -2.05) for males and -0.86 (95% CI, -1.06 to -0.65) for females. Age-specific annual percentages for both sexes showed a decreasing trend in all age groups and were more pronounced for females older than 65 years and males younger than 65 years. Period effects for both sexes showed a similar monotonic downward pattern, with the downward trend slowing for females after 2000. The cohort effect declined monotonically for males and started to plateau for females after the 1945 birth cohort. It was projected that the incident cases of lung cancer in Hong Kong would continue to increase, with 4,435 male cases and 3,561 female cases in 2030. Demographic decomposition suggested that population growth and population aging play an important role in the change of lung cancer cases. Much progress has been made in reducing the incidence of lung cancer in Hong Kong, but this has been offset by demographic changes that will continue to increase the incident cases of lung cancer in Hong Kong, especially among females. There is an urgent need for continued public health policies and clinical programs for risk factor control and necessary screening.
Keywords: Lung cancer, Hong Kong, age-period-cohort analysis, projection, decomposition
Introduction
Lung cancer has been one of the most common causes of cancer deaths in Hong Kong [1,2], where 98% are Chinese ethnicity. Considering the strong causal relationship between smoking and multiple adverse health outcomes [3-6], the Hong Kong government has successfully implemented a series of tobacco control interventions since the 1980s [7,8]. The age-standardized incidence rate of lung cancer in Hong Kong has declined notably, particularly in males, but the incidence in females has not changed much over the past two decades [7]. Despite this, the burden of lung cancer in Hong Kong remains high. According to the latest survey data released by the Hong Kong Cancer Registry (HKCaR) (https://www3.ha.org.hk/cancereg/allages.asp), 3,424 men and 2,151 women were registered with lung cancer in Hong Kong in 2019. The survey data shows that lung cancer remains the most common cancer among men and the third most common cancer among women in Hong Kong and that the number of lung cancer cases in Hong Kong have continued to increase in recent years, especially among women.
Consistent with studies in other populations [2,3,5,7,9], tobacco smoking is the well-documented significant risk factor of lung cancer in Hong Kong men, and smoking and exposure to cooking emissions from Chinese-style frying are the main risk factors for lung cancer in Hong Kong women. Smoking prevalence and lifestyle habits vary across birth cohorts [8,10], leading to the varying risk of lung cancer incidence in different birth cohorts. The effect of temporal variation and birth cohort on lung cancer incidence can be separated and analyzed by age-period-cohort models, which are more suitable for elucidating possible determinants of temporal trends [11-13]. The age effect refers to the association of physiological changes in aging with cancer incidence independent of birth cohort and calendar period. Birth cohort effects represent intergenerational changes in the prevalence of causal factors, such as specific lifestyles or behaviors, which may have a delayed effect on lung cancer incidence. Period effects reflect changes in factors that affect all individuals simultaneously over a given period, such as changes in diagnostic methods and the introduction of screening programs. Understanding each factor of the age, period, and cohort effects may provide a more accurate understanding and projection of disease incidence, which may help guide public health policy, resource allocation, and the design of screening programs.
This study aimed to evaluate the relationship between age, period, and cohort with the incidence of lung cancer in Hong Kong. We used age-period-cohort (APC) modeling to assess the associations between age, period, and birth cohort and the incidence of lung cancer in Hong Kong. In addition, we projected the future incidence of lung cancer in Hong Kong from 2020 to 2030 by conducting a predictive model that included age, period, and cohort effects. Further, we analyzed the key drivers of the net change in lung cancer incident cases in Hong Kong, including population growth, population aging, and epidemiological shifts.
Material and methods
Data source
We obtained rates and incident cases of lung cancer in Hong Kong by sex, age group, and calendar year from 1985 to 2019 from the Hong Kong Cancer Registry (HKCaR). HKCaR was established in 1963 as a population-based cancer registry and is a member of the International Association of Cancer Registries. The primary mission of HKCaR is to report the cancer incidence and mortality rates in Hong Kong by collecting, consolidating, and validating preliminary demographic data, information on the topography and histology of all cancers diagnosed in Hong Kong. HKCaR has established quality control measures to check the completeness, accuracy, and validity of the data. The percentage of morphologically verified cases has increased significantly from 55% in 1983 to nearly 90% in 2019, and the percentage of cases confirmed by death certificates has decreased from 13.3% in 1983 to 0.2% in 2019, having achieved the highest standards for developed countries as depicted by the World Health Organization (WHO) International Agency for Research on Cancer (IARC). In addition, HKCaR has applied a series of validation rules to the data registered to ensure the accuracy and validity of the data before publication. All International Classification of Diseases (9th and 10th revision) codes pertaining to lung cancer (162 and C33-34) are included. As very few lung cancer cases were diagnosed at ages below 30, we only consider data of 12 five-year age groups ranging from 30-34 years to 85 years or older. Population data of Hong Kong from 1985 to 2030 were obtained from the United Nations (UN) World Population Prospects 2019 revision (https://population.un.org/wpp). Age-standardized incidence estimates were weighted to the age distribution of the WHO’s world standard population based on world average population between 2000-2025 (http://www.who.int/healthinfo/paper31.pdf) by the direct method.
Age-period-cohort analysis
For age-period-cohort (APC) analyses, we arranged the incident lung cancer cases and population data of Hong Kong into 12 age intervals (from 30-34 years to 85 plus) and 7-period intervals (from 1985-1989 to 2015-2019). The cohort was defined using the difference between the midpoint of the age interval and period interval. Net drift and local drift were essential parameters in the APC models. Net drift indicates the overall annual percentage change in the expected age-adjusted rate. Local drift indicates estimated annual percentage change over time specific to age group. The longitudinal age curve indicates the expected age-specific rate in a reference cohort adjusted for period effects. The period (or cohort) rate ratio (RR), represents the ratio of age-specific rates in period (or cohort) relative to the reference period (or cohort). For relative rate measurements, the reference cohort was the 1945 birth cohort, and the reference period interval was from 2000 to 2004. The Wald chi-square test was adopted to test the significance of the estimable parameters and functions. The estimable parameters were obtained by the APC Web Tool (https://analysistools.cancer.gov/apc). All statistical tests were two-sided.
Projection analysis
Age-specific incident cases of lung cancer in Hong Kong from 2020 to 2030 were projected using the Bayesian age-period-cohort (BAPC) analysis with integrated nested Laplace approximations [14]. The model showed better coverage and accuracy than other prediction methods [15]. We prepared age-specific incident cases of lung cancer (from 1985 to 2019) and Hong Kong population data (from 1985 to 2030), followed by an 11-year (from 2020 to 2030) retrospective projection using the BAPC function in the R package BAPC, version 0.0.34.
Decomposition
Further, using 1985 as the reference year, we decomposed the drivers of the increase in the number of lung cancer cases in Hong Kong from 1986 to 2030 into three components: population growth, population aging, and changes in age-specific incidence rate. Changes in age-specific incidence rates represent epidemiological changes, including all differences in incidence that cannot be explained by population growth and population aging, such as changes in the prevalence of smoking and other risk exposures [16]. The net changes in these three components were equal to the difference in the total incident cases. We performed the decomposition using a validated algorithm [17-19] that is robust to the order of decomposition and the choice of the reference year. All statistical analyses were performed with R software (version 3.6.3; R Foundation for Statistical Computing).
Results
Trends in incidence rates of lung cancer
A total of 147,668 patients (97,435 male patients [66.0% of total patients] and 50,233 female patients [34.0%]) were included in our analysis. From 1985 to 2019, the incident cases of lung cancer have continued to rise in Hong Kong, from 2,133 and 1,052 to 3,421 and 2,148 for males and females, respectively (Figure 1A). There was a slight increase in the crude incidence rate of lung cancer for both sexes. However, in the same time period, marked declines were observed in the age-standardized incidence rate for both sexes, with a greater decline for males. The age-standardized incidence rate decreased from 93.1 per 100,000 persons in 1985 to 48.8 per 100,000 persons in 2019 for males and decreased from 39.3 per 100,000 persons in 1985 to 27.1 per 100,000 persons in 2019 for females (Figure 1B). However, the age-standardized incidence rate for females has fluctuated over the last two decades, with relatively little change.
Figure 1.
Changes in incidence rate and incident cases of lung cancer for males and females in Hong Kong, 1985-2019. A. Incident cases for males and females. B. Age-standardized incidence rate and crude incidence rate of lung cancer for males and females.
Age-period-cohort analysis
Net drift indicates the overall annual percentage change (Figure 2), which is the APC analog of the estimated average percentage change of the age-standardized rate. There were marked sex differences in net drift. The net drift was -2.29% (95% confidence interval, CI: -2.53 to -2.05) for males and -0.86% (95% CI: -1.06 to -0.65) for females, reflecting decreases in incidence for both sexes from 1985 to 2019. Local drift reflects additional age-specific changes in lung cancer incidence trends. Values were less than zero in almost all age groups, indicating a decline in the incidence of lung cancer for both sexes in all age groups. The decline was more pronounced among women in the older age groups (>65 years) and men in the younger age groups (<65 years) (Figure 2).
Figure 2.
Local drifts with net drift values for males and females for lung cancer incidence in Hong Kong from 1985 to 2019. Shaded areas indicate 95% confidence interval.
The age-adjusted incidence rate of lung cancer increases with the patient age until around 80 years of age, after which it decreased slightly (Figure 3). Incidence rates were highest in those around 80 years of age and were much higher in men than in women, with the gap increasing with age. We performed curve estimates on the longitudinal age curves and found that they were all fitted with a second-order polynomial regression (Regression coefficients were significant at the P<0.0001 level, and adjusted R-squared were 0.999 for both sexes).
Figure 3.
Fitted longitudinal age curves of lung cancer incidence (per 100,000 person-years) and the corresponding 95% confidence interval for males and females in Hong Kong.
Period effects for both sexes followed similar monotonic declined patterns throughout the study period, with more quickly decreasing for males than for females and the trend slowing down after 2000 for females (Figure 4). Cohort effects followed different patterns across the sexes, with a consistent monotonic decline for males. On the other hand, females showed a monotonic decrease before the 1940-1944 birth cohort and appeared to be plateauing in the subsequent birth cohorts (Figure 5). The Wald tests showed that all net drifts, local drifts, cohort effects, and period effects were statistically significant for both sexes (P<0.0001 for all) (Table 1).
Figure 4.
Relative risk of each period compared with the reference adjusted for age and nonlinear cohort effects and the corresponding 95% confidence interval for males and females.
Figure 5.
Relative risk of each cohort compared with the reference adjusted for age and nonlinear period effects and the corresponding 95% confidence interval.
Table 1.
Wald Chi-square tests for estimable parameters in the APC model
| Null Hypothesis | Males | Females | ||
|---|---|---|---|---|
|
|
|
|||
| Chi-squre | P-value | Chi-squre | P-value | |
| NetDrift =0 | 349.79 | <0.001 | 65.49 | <0.001 |
| All Age Deviations =0 | 1460.28 | <0.001 | 1000.59 | <0.001 |
| All Period RR =1 | 351.66 | <0.001 | 74.53 | <0.001 |
| All Cohort RR =1 | 1034.75 | <0.001 | 474.93 | <0.001 |
| All Local Drifts = Net Drift | 56.48 | <0.001 | 159.49 | <0.001 |
Projection
We next performed a projection analysis to estimate the future incidence of lung cancer in Hong Kong. We found the number of lung cancer cases in Hong Kong would continue to increase, with 4,435 male cases and 3,561 female cases in 2030 (Table 2). The age-standardized incidence rate will continue to decline for males and may increase slightly for females (Figures 6, 7). The largest increases in incidence were expected to be in the female population over 60 years of age and in the male population over 70 years of age (Tables 2, 3).
Table 2.
Estimated incident cases of age-specific lung cancer in Hong Kong males from 1985 to 2030
| Year | Incident cases of age-specific and total lung cancer | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||||||
| 30-34 | 35-39 | 40-44 | 45-49 | 50-54 | 55-59 | 60-64 | 65-69 | 70-74 | 75-79 | 80-84 | 85+ | Total | |
| 1985 | 15 | 36 | 42 | 108 | 219 | 295 | 400 | 371 | 344 | 182 | 90 | 31 | 2133 |
| 1990 | 16 | 20 | 51 | 83 | 185 | 297 | 400 | 426 | 402 | 293 | 145 | 50 | 2368 |
| 1995 | 25 | 31 | 57 | 93 | 107 | 206 | 409 | 455 | 456 | 334 | 229 | 70 | 2472 |
| 2000 | 9 | 19 | 64 | 100 | 149 | 193 | 314 | 479 | 519 | 412 | 260 | 168 | 2686 |
| 2005 | 10 | 22 | 50 | 101 | 177 | 207 | 278 | 417 | 497 | 497 | 304 | 168 | 2728 |
| 2010 | 6 | 18 | 37 | 84 | 210 | 261 | 336 | 365 | 483 | 491 | 381 | 242 | 2914 |
| 2015 | 2 | 10 | 40 | 59 | 155 | 308 | 382 | 461 | 430 | 451 | 382 | 244 | 2924 |
| 2019 | 5 | 19 | 29 | 67 | 157 | 313 | 507 | 584 | 572 | 453 | 376 | 339 | 3421 |
| 2020 | 4 | 12 | 28 | 64 | 135 | 292 | 495 | 566 | 590 | 432 | 434 | 396 | 3448 |
| 2021 | 4 | 12 | 27 | 63 | 132 | 282 | 502 | 592 | 621 | 459 | 421 | 419 | 3534 |
| 2022 | 4 | 11 | 27 | 62 | 130 | 271 | 502 | 624 | 647 | 504 | 400 | 441 | 3623 |
| 2023 | 4 | 11 | 27 | 61 | 130 | 259 | 498 | 657 | 672 | 560 | 381 | 459 | 3719 |
| 2024 | 4 | 11 | 26 | 59 | 130 | 249 | 489 | 686 | 701 | 614 | 377 | 470 | 3816 |
| 2025 | 4 | 10 | 26 | 58 | 129 | 242 | 478 | 708 | 737 | 659 | 395 | 474 | 3920 |
| 2026 | 4 | 10 | 26 | 57 | 127 | 237 | 464 | 719 | 774 | 695 | 425 | 485 | 4023 |
| 2027 | 4 | 10 | 25 | 57 | 125 | 235 | 447 | 722 | 818 | 726 | 470 | 490 | 4129 |
| 2028 | 4 | 10 | 24 | 56 | 123 | 235 | 429 | 717 | 863 | 757 | 523 | 492 | 4233 |
| 2029 | 4 | 10 | 24 | 56 | 121 | 234 | 413 | 707 | 904 | 793 | 573 | 496 | 4335 |
| 2030 | 3 | 10 | 24 | 55 | 119 | 233 | 402 | 693 | 935 | 838 | 616 | 507 | 4435 |
Figure 6.
Trends and projected incidence rates of lung cancer for males in Hong Kong. Dots represent fitted points. Each lighter shade of blue represents an additional 10% CI.
Figure 7.
Trends and projected incidence rates of lung cancer for females in Hong Kong. Dots represent fitted points. Each lighter shade of blue represents an additional 10% CI.
Table 3.
Estimated incident cases of age-specific lung cancer in Hong Kong females from 1985 to 2030
| Year | Incident cases of age-specific and total lung cancer | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||||||
| 30-34 | 35-39 | 40-44 | 45-49 | 50-54 | 55-59 | 60-64 | 65-69 | 70-74 | 75-79 | 80-84 | 85+ | Total | |
| 1985 | 7 | 11 | 15 | 33 | 62 | 99 | 163 | 181 | 184 | 141 | 89 | 67 | 1052 |
| 1990 | 3 | 13 | 19 | 18 | 44 | 79 | 118 | 167 | 204 | 186 | 112 | 105 | 1068 |
| 1995 | 9 | 31 | 25 | 37 | 40 | 65 | 109 | 156 | 180 | 199 | 162 | 116 | 1129 |
| 2000 | 6 | 21 | 41 | 47 | 53 | 59 | 113 | 147 | 209 | 218 | 182 | 163 | 1259 |
| 2005 | 6 | 13 | 28 | 55 | 106 | 97 | 96 | 150 | 227 | 251 | 187 | 185 | 1401 |
| 2010 | 10 | 11 | 35 | 78 | 136 | 146 | 154 | 140 | 181 | 242 | 218 | 213 | 1564 |
| 2015 | 7 | 16 | 30 | 85 | 145 | 193 | 231 | 232 | 165 | 230 | 220 | 262 | 1816 |
| 2019 | 3 | 21 | 33 | 84 | 140 | 250 | 324 | 299 | 290 | 202 | 211 | 291 | 2148 |
| 2020 | 7 | 19 | 41 | 76 | 131 | 230 | 309 | 333 | 314 | 199 | 234 | 324 | 2217 |
| 2021 | 7 | 19 | 41 | 77 | 132 | 232 | 324 | 358 | 347 | 221 | 225 | 337 | 2320 |
| 2022 | 7 | 19 | 42 | 77 | 133 | 231 | 339 | 383 | 377 | 257 | 211 | 350 | 2426 |
| 2023 | 7 | 19 | 44 | 79 | 136 | 228 | 353 | 409 | 407 | 303 | 199 | 362 | 2546 |
| 2024 | 7 | 19 | 44 | 80 | 138 | 227 | 364 | 434 | 439 | 350 | 197 | 371 | 2670 |
| 2025 | 7 | 19 | 45 | 81 | 140 | 227 | 370 | 457 | 474 | 394 | 210 | 377 | 2801 |
| 2026 | 7 | 18 | 45 | 83 | 143 | 229 | 373 | 479 | 511 | 435 | 235 | 381 | 2939 |
| 2027 | 7 | 18 | 45 | 86 | 145 | 233 | 372 | 502 | 549 | 474 | 274 | 380 | 3085 |
| 2028 | 6 | 19 | 44 | 88 | 147 | 238 | 369 | 523 | 586 | 512 | 323 | 380 | 3235 |
| 2029 | 6 | 18 | 43 | 90 | 150 | 242 | 368 | 539 | 622 | 554 | 374 | 387 | 3393 |
| 2030 | 6 | 18 | 43 | 91 | 153 | 247 | 369 | 550 | 657 | 601 | 422 | 404 | 3561 |
Decomposition analysis
Finally, we conducted a demographic decomposition to analyze the potential driver of the incident cases of lung cancer from 1986 to 2030 in Hong Kong. Our results indicated that demographic factors (i.e., population growth and population aging) play an essential role in changing of the incident cases of lung cancer in Hong Kong. Although Hong Kong has made significant progress in preventing the incidence of lung cancer, population aging and population growth have led to a marked increase in the incident cases of lung cancer (Figures 8, 9; Tables 4, 5).
Figure 8.
Contribution of changes in population aging, population growth, and age-specific incidence rate to changes in incident cases from 1986 to 2030 for Hong Kong males, using 1985 as the reference year. Data in the right of the blue dashed line were the decomposition based on the projected data.
Figure 9.
Contribution of changes in population aging, population growth, and age-specific incidence rate to changes in incident cases from 1986 to 2030 for Hong Kong females, using 1985 as the reference year. Data in the right of the blue dashed line were the decomposition based on the projected data.
Table 4.
Contribution of changes in population aging, population growth, and age-specific incidence rate to the net change of incident cases of lung cancer for males in Hong Kong from 1986 to 2030, using 1985 as the reference year
| Year | Due to population aging | Percentage change due to population aging | Due to population growth | Percentage change due to population growth | Due to age-specific incidence rate | Percentage change due to age-specific incidence rate | Net change | Percentage change from 1985 |
|---|---|---|---|---|---|---|---|---|
| 1986 | 11 | 0.5 | 77 | 3.6 | -147 | -6.9 | -59 | -2.8 |
| 1990 | 33 | 1.5 | 433 | 20.3 | -230 | -10.8 | 235 | 11.0 |
| 1995 | 86 | 4.0 | 814 | 38.2 | -562 | -26.3 | 339 | 15.9 |
| 2000 | 280 | 13.1 | 1108 | 51.9 | -835 | -39.1 | 553 | 25.9 |
| 2005 | 498 | 23.4 | 1260 | 59.1 | -1163 | -54.5 | 595 | 27.9 |
| 2010 | 857 | 40.2 | 1431 | 67.1 | -1507 | -70.7 | 781 | 36.6 |
| 2015 | 1117 | 52.3 | 1594 | 74.7 | -1919 | -90.0 | 791 | 37.1 |
| 2019 | 1447 | 67.8 | 1838 | 86.2 | -1997 | -93.6 | 1288 | 60.4 |
| 2020 | 1545 | 72.4 | 1873 | 87.8 | -2103 | -98.6 | 1315 | 61.7 |
| 2021 | 1621 | 76.0 | 1922 | 90.1 | -2142 | -100.4 | 1401 | 65.7 |
| 2022 | 1696 | 79.5 | 1973 | 92.5 | -2179 | -102.2 | 1490 | 69.9 |
| 2023 | 1772 | 83.1 | 2027 | 95.0 | -2213 | -103.8 | 1586 | 74.4 |
| 2024 | 1851 | 86.8 | 2082 | 97.6 | -2250 | -105.5 | 1683 | 78.9 |
| 2025 | 1935 | 90.7 | 2138 | 100.2 | -2286 | -107.2 | 1787 | 83.8 |
| 2026 | 2018 | 94.6 | 2190 | 102.7 | -2317 | -108.6 | 1890 | 88.6 |
| 2027 | 2101 | 98.5 | 2241 | 105.1 | -2346 | -110.0 | 1996 | 93.6 |
| 2028 | 2185 | 102.4 | 2290 | 107.4 | -2375 | -111.3 | 2100 | 98.5 |
| 2029 | 2268 | 106.3 | 2335 | 109.5 | -2401 | -112.6 | 2202 | 103.2 |
| 2030 | 2352 | 110.3 | 2376 | 111.4 | -2425 | -113.7 | 2302 | 107.9 |
Table 5.
Contribution of changes in population aging, population growth, and age-specific incidence rate to the net change of incident cases of lung cancer for females in Hong Kong from 1986 to 2030, using 1985 as the reference year
| Year | Due to population aging | Percentage change due to population aging | Due to population growth | Percentage change due to population growth | Due to age-specific incidence rate | Percentage change due to age-specific incidence rate | Net change | Percentage change from 1985 |
|---|---|---|---|---|---|---|---|---|
| 1986 | -1 | 0.0 | 41 | 3.9 | -39 | -3.7 | 2 | 0.2 |
| 1990 | -14 | -1.3 | 219 | 20.8 | -189 | -18.0 | 16 | 1.5 |
| 1995 | -58 | -5.5 | 443 | 42.1 | -308 | -29.3 | 77 | 7.3 |
| 2000 | -61 | -5.8 | 652 | 62.0 | -384 | -36.5 | 207 | 19.7 |
| 2005 | -11 | -1.1 | 835 | 79.3 | -474 | -45.1 | 349 | 33.2 |
| 2010 | 82 | 7.8 | 1012 | 96.2 | -582 | -55.3 | 512 | 48.7 |
| 2015 | 184 | 17.5 | 1235 | 117.4 | -654 | -62.2 | 764 | 72.6 |
| 2019 | 313 | 29.7 | 1451 | 137.9 | -668 | -63.5 | 1096 | 104.2 |
| 2020 | 361 | 34.3 | 1499 | 142.5 | -694 | -66.0 | 1165 | 110.7 |
| 2021 | 404 | 38.4 | 1556 | 147.9 | -692 | -65.8 | 1268 | 120.5 |
| 2022 | 450 | 42.8 | 1614 | 153.5 | -691 | -65.7 | 1374 | 130.6 |
| 2023 | 500 | 47.5 | 1678 | 159.5 | -684 | -65.0 | 1494 | 142.0 |
| 2024 | 553 | 52.6 | 1741 | 165.5 | -676 | -64.2 | 1618 | 153.8 |
| 2025 | 610 | 58.0 | 1804 | 171.5 | -665 | -63.2 | 1749 | 166.3 |
| 2026 | 671 | 63.8 | 1869 | 177.6 | -653 | -62.0 | 1887 | 179.4 |
| 2027 | 735 | 69.8 | 1933 | 183.7 | -634 | -60.3 | 2033 | 193.3 |
| 2028 | 803 | 76.3 | 1994 | 189.6 | -614 | -58.4 | 2183 | 207.5 |
| 2029 | 876 | 83.3 | 2056 | 195.5 | -592 | -56.2 | 2341 | 222.5 |
| 2030 | 955 | 90.8 | 2120 | 201.6 | -566 | -53.8 | 2509 | 238.5 |
Compared to 1985, the incident cases of lung cancer in Hong Kong increased by 1,288 and 1,096 cases in males and females respectively in 2019, representing an increase of 60.4% and 104.2%, respectively. The change in epidemiology resulted in a decrease of 93.6% and 63.5% in the incident cases for males and females. The drivers of incident cases of lung cancer were population aging (67.8% and 29.7% increase for males and females respectively compared to 1985) and population growth (86.2% and 137.9% increase for males and females respectively compared to 1985) (Tables 4, 5).
The projection and decomposition analysis suggested that the trend will continue. In 2030, Hong Kong will have 107.9% and 238.5% more incident lung cancer cases for males and females respectively than in 1985, with the contribution of population aging at 110.3% and 90.8%, population growth at 111.4% and 201.6%, and epidemiological changes at -113.7% and -53.8% respectively (Tables 4, 5).
Discussion
This study used APC analysis to address why incident cases of lung cancer continue to increase in Hong Kong. We showed that, while age-standardized incidence rates declined for both sexes, the decline was relatively small for males >65 years and for females <65 years. In addition, the period effects showed monotonic decreases in both sexes. In contrast, the cohort effect showed a monotonic decline for males and appeared to be plateauing for females after the 1945 birth cohort. Moreover, demographic factors were the main drivers of incident cases of lung cancer in Hong Kong. Although Hong Kong has made significant progress in reducing the incidence of lung cancer, these gains have been offset by demographic changes. Therefore, there is a need for continued epidemiological and clinical research on the disease and its treatment.
One of the main values of this study is the APC analysis, which quantitatively describes the relative age, period, and cohort effects. We showed an encouraging downward trend in period effects and cohort effects for both sexes. These trends were attributed to the efforts to reduce tobacco use in Hong Kong over the past decades. Smoking was the most critical risk factor associated with lung cancer [3], but the attributable risk was much higher in men than in women (45.8% vs. 6.2% in Hong Kong [9]). Hong Kong has been working on reducing the health burden of tobacco use with notable success [20]. Beginning with a tobacco-focused health ordinance in 1982, Hong Kong started a progressive approach to tobacco control that included a multipronged strategy to reduce the supply and demand for tobacco use. Further, in 2007, Hong Kong implemented legislation to protect against secondhand smoke exposure in indoor workplaces and in public places [20]. Smoking prevalence among those aged 15 or older in Hong Kong fell progressively from 23.3% in 1982 to 11.1% (18.1% in men and 3.2% in women) in 2019 [8,21]. The period effect of lung cancer is consistent with the declining trend in smoking prevalence. However, heated tobacco products have become increasingly popular in recent years, especially among adolescents, and can be a pathway to smoking and a health risk for youth [22,23]. Adolescents are less aware of the risks associated with heated tobacco products [22,24], so stricter regulation of heated tobacco products is needed to limit adolescent exposure to heated tobacco products and enhance health education.
However, we found that the period effect for lung cancer in Hong Kong women has leveled off in the last two decades, probably due to the smaller proportion of lung cancer attributable to cigarette smoking and small change in smoking prevalence among Hong Kong women. For example, in another study, only 16% of lung cancers in Hong Kong women can be attributed to smoking, while 22% to environmental radon exposure and 26% to exposure to cooking fume [1]. In addition, the prevalence of smoking among women in Hong Kong has changed a little over the past few decades and has even increased notably among women as the population has grown [25,26]. Our study highlights the need for Hong Kong women to lower their exposure to cooking emissions from Chinese-style frying and maintain a low smoking prevalence, leading to good health gains.
The cohort effect reflects early lifestyle factors or carcinogenic exposure. The development of lung cancer usually requires long-term exposure and a long incubation period, such as smoking habits and secondhand smoke exposure [4,27,28]. Declines in smoking prevalence and exposure to secondhand smoke and sex differences in lung cancer attributable to smoking could explain the sex differences in lung cancer cohort effects in Hong Kong. We found an encouraging declining cohort effect among Hong Kong males. In contrast, the cohort effect was weaker in Hong Kong females born after the 1945 birth cohort, which may persist in the future. This weak cohort effect among women may be related to the multiple attributions of lung cancer in Hong Kong women. Other studies have indicated that smoking and secondhand smoke exposure are not the most important causes of lung cancer in women [1].
Although we could estimate period effect and cohort effect as period RRs and cohort RRs, respectively, it is not appropriate to interpret them entirely separately [12,29,30]. The reason is that the period effect leads to cohort effect when they affect all age groups simultaneously; different cohorts are born in different periods and, therefore, inevitably have a confounding impact on period effects.
Another main value of this study is the prediction of trends in lung cancer incidence in Hong Kong. We showed that although there was a clear downward trend in age-standardized lung cancer incidence rates in Hong Kong, suggesting that Hong Kong has made great progress in preventing the occurrence of lung cancer. However, incident lung cancer cases in Hong Kong have continued to increase gradually over the past three decades. This inconsistency reflects the critical role of demographic change in the burden of lung cancer. Although tobacco reduction measures and other cancer-related lifestyle changes have reduced the incidence of lung cancer [7,20,31], demographic shifts have offset the effect of such epidemiological changes. Population growth and population aging have emerged as the main drivers of lung cancer cases in Hong Kong. Our projections suggest that this trend will continue as Hong Kong’s population continues to age and grow.
Our results are generally consistent with previous epidemiological studies of lung cancer incidence in Hong Kong [2,32]. In particular, we found a significant attenuation of the period and cohort effects in Hong Kong women. Although other studies have shown an association between lung cancer incidence and age [2,3,6,32], after controlling for cohort and period effects, we found a second-order polynomial relationship between lung cancer incidence and age, and this age effect may be an increase in cancer risk with increasing age. It has been suggested that there is a roughly 20 to 30-year gap between smoking and the development of lung cancer [33], highlighting the health benefits of quitting smoking early. We estimate a similar relationship between exposure to cooking emissions and lung cancer development in women. Therefore there are significant health benefits for women from the early reduction of exposure to cooking emissions.
This retrospective analysis has some limitations. First, the quality of the early data of HKCaR was not good enough; for example, it was not until 1993 that the percentage of morphologically verified cases exceeded 60%. Second, HKCaR is limited because it does not contain specific data on whether patients smoke, comorbidities, and medications. It would be helpful to understand how these comorbidities coincide with the incidence of lung cancer to adjust our analysis. Third, data on the future structure and size of the population in Hong Kong obtained from United Nations World Population Prospects may be subject to biases, adding uncertainty to the projection. Finally, several modeling approaches were used, including projections and population decomposition, which inevitably introduced some degrees of uncertainty.
Conclusions
Age-standardized lung cancer incidence rates in Hong Kong showed a consistent decline in males but little change in females. Differences in smoking prevalence, exposure to secondhand smoke, and cooking emissions from Chinese-style frying may be associated with the period and cohort effects of lung cancer incidence in Hong Kong. With population growth and aging, lung cancer cases in Hong Kong are expected to continue to increase, especially among females. Further research on this trend and epidemiological assessment should be conducted.
Acknowledgements
This work was supported by Huilan Public Welfare Fund Project (No. HL-HS2020-45) and Beijing Xisike Clinical Oncology Research Foundation (No. Y-QL202102-0175).
Disclosure of conflict of interest
None.
References
- 1.Chiu YL, Wang XR, Qiu H, Yu IT. Risk factors for lung cancer: a case-control study in Hong Kong women. Cancer Causes Control. 2010;21:777–785. doi: 10.1007/s10552-010-9506-9. [DOI] [PubMed] [Google Scholar]
- 2.Wang XR, Chiu YL, Qiu H, Au JS, Yu IT. The roles of smoking and cooking emissions in lung cancer risk among Chinese women in Hong Kong. Ann Oncol. 2009;20:746–751. doi: 10.1093/annonc/mdn699. [DOI] [PubMed] [Google Scholar]
- 3.Alberg AJ, Samet JM. Epidemiology of lung cancer. Chest. 2003;123(Suppl):21S–49S. doi: 10.1378/chest.123.1_suppl.21s. [DOI] [PubMed] [Google Scholar]
- 4.GBD 2019 Respiratory Tract Cancers Collaborators. Global, regional, and national burden of respiratory tract cancers and associated risk factors from 1990 to 2019 a systematic analysis for the global burden of disease study 2019. Lancet Respir Med. 2021;9:1030–1049. doi: 10.1016/S2213-2600(21)00164-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Zhang Y, Luo G, Etxeberria J, Hao Y. Global patterns and trends in lung cancer incidence: a population-based study. J Thorac Oncol. 2021;16:933–944. doi: 10.1016/j.jtho.2021.01.1626. [DOI] [PubMed] [Google Scholar]
- 6.de Groot P, Munden RF. Lung cancer epidemiology, risk factors, and prevention. Radiol Clin North Am. 2012;50:863–876. doi: 10.1016/j.rcl.2012.06.006. [DOI] [PubMed] [Google Scholar]
- 7.Mok TSK, Zee B, Chan ATC, Kwan WH, Johnson PJ. A population survey on tobacco consumption and lung cancer risk in Hong Kong. Int J Cancer. 2002:194–194. [Google Scholar]
- 8.Koplan JP, An WK, Lam RM. Hong Kong: a model of successful tobacco control in China. Lancet. 2010;375:1330–1331. doi: 10.1016/S0140-6736(10)60398-4. [DOI] [PubMed] [Google Scholar]
- 9.Chan-Yeung M, Koo LC, Ho JC, Tsang KW, Chau WS, Chiu SW, Ip MS, Lam WK. Risk factors associated with lung cancer in Hong Kong. Lung Cancer. 2003;40:131–140. doi: 10.1016/s0169-5002(03)00036-9. [DOI] [PubMed] [Google Scholar]
- 10.Fernandez E, Schiaffino A, Borras JM, Shafey O, Villalbi JR, La Vecchia C. Prevalence of cigarette smoking by birth cohort among males and females in Spain, 1910-1990. Eur J Cancer Prev. 2003;12:57–62. doi: 10.1097/00008469-200302000-00009. [DOI] [PubMed] [Google Scholar]
- 11.Kuzmickiene I, Everatt R. Trends and age-period-cohort analysis of upper aerodigestive tract and stomach cancer mortality in Lithuania, 1987-2016. Public Health. 2021;196:62–68. doi: 10.1016/j.puhe.2021.05.001. [DOI] [PubMed] [Google Scholar]
- 12.Jacobs D, Huang H, Olino K, Weiss S, Kluger H, Judson BL, Zhang Y. Assessment of age, period, and birth cohort effects and trends in Merkel cell carcinoma incidence in the United States. JAMA Dermatol. 2021;157:59–65. doi: 10.1001/jamadermatol.2020.4102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Rosenberg PS, Check DP, Anderson WF. A web tool for age-period-cohort analysis of cancer incidence and mortality rates. Cancer Epidemiol Biomarkers Prev. 2014;23:2296–2302. doi: 10.1158/1055-9965.EPI-14-0300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Riebler A, Held L. Projecting the future burden of cancer: Bayesian age-period-cohort analysis with integrated nested Laplace approximations. Biom J. 2017;59:531–549. doi: 10.1002/bimj.201500263. [DOI] [PubMed] [Google Scholar]
- 15.Knoll M, Furkel J, Debus J, Abdollahi A, Karch A, Stock C. An R package for an integrated evaluation of statistical approaches to cancer incidence projection. BMC Med Res Methodol. 2020;20:257. doi: 10.1186/s12874-020-01133-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Roth GA, Forouzanfar MH, Moran AE, Barber R, Nguyen G, Feigin VL, Naghavi M, Mensah GA, Murray CJ. Demographic and epidemiologic drivers of global cardiovascular mortality. N Engl J Med. 2015;372:1333–1341. doi: 10.1056/NEJMoa1406656. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Cheng X, Yang Y, Schwebel DC, Liu Z, Li L, Cheng P, Ning P, Hu G. Population ageing and mortality during 1990-2017: a global decomposition analysis. PLoS Med. 2020;17:e1003138. doi: 10.1371/journal.pmed.1003138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Cheng X, Tan L, Gao Y, Yang Y, Schwebel DC, Hu G. A new method to attribute differences in total deaths between groups to population size, age structure and age-specific mortality rate. PLoS One. 2019;14:e0216613. doi: 10.1371/journal.pone.0216613. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Du J, Wu X, Li L, Cao W, Sun S. Bayesian age-period-cohort prediction of mortality of type 2 diabetic kidney disease in China: a modeling study. Front Endocrinol. 2021;12:767263. doi: 10.3389/fendo.2021.767263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Thach TQ, McGhee SM, So JC, Chau J, Chan EK, Wong CM, Hedley AJ. The smoke-free legislation in Hong Kong: its impact on mortality. Tob Control. 2016;25:685–691. doi: 10.1136/tobaccocontrol-2015-052496. [DOI] [PubMed] [Google Scholar]
- 21.Ho SY, Chen J, Leung LT, Mok HY, Wang L, Wang MP, Lam TH. Adolescent smoking in Hong Kong: prevalence, psychosocial correlates, and prevention. J Adolesc Health. 2019;64:S19–S27. doi: 10.1016/j.jadohealth.2019.01.003. [DOI] [PubMed] [Google Scholar]
- 22.Jiang N, Cleland CM, Wang MP, Kwong A, Lai V, Lam TH. Perceptions and use of e-cigarettes among young adults in Hong Kong. BMC Public Health. 2019;19:1123. doi: 10.1186/s12889-019-7464-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Chan SSC, Cheung YTD, Wong YMB, Kwong A, Lai V, Lam TH. A brief smoking cessation advice by youth counselors for the smokers in the Hong Kong quit to win contest 2010: a cluster randomized controlled trial. Prev Sci. 2018;19:209–219. doi: 10.1007/s11121-017-0823-z. [DOI] [PubMed] [Google Scholar]
- 24.Wu YS, Wang MP, Ho SY, Li HCW, Cheung YTD, Tabuchi T, Kwong AC, Lai V, Lam TH. Heated tobacco products use in Chinese adults in Hong Kong: a population-based cross-sectional study. Tob Control. 2020;29:277–281. doi: 10.1136/tobaccocontrol-2018-054719. [DOI] [PubMed] [Google Scholar]
- 25.Li HC, Chan SS, Lam TH. Smoking among Hong Kong Chinese women: behavior, attitudes and experience. BMC Public Health. 2015;15:183. doi: 10.1186/s12889-015-1529-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lau EM, Lee P, Lynn H, Sham A, Woo J. The epidemiology of cigarette smoking in Hong Kong Chinese women. Prev Med. 2003;37:383–388. doi: 10.1016/s0091-7435(03)00148-8. [DOI] [PubMed] [Google Scholar]
- 27.Luo Q, Yu XQ, Wade S, Caruana M, Pesola F, Canfell K, O’Connell DL. Lung cancer mortality in Australia: projected outcomes to 2040. Lung Cancer. 2018;125:68–76. doi: 10.1016/j.lungcan.2018.09.001. [DOI] [PubMed] [Google Scholar]
- 28.Pope CA 3rd, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K, Thurston GD. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA. 2002;287:1132–1141. doi: 10.1001/jama.287.9.1132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wang Z, Hu S, Sang S, Luo L, Yu C. Age-period-cohort analysis of stroke mortality in China: data from the global burden of disease study 2013. Stroke. 2017;48:271–275. doi: 10.1161/STROKEAHA.116.015031. [DOI] [PubMed] [Google Scholar]
- 30.Zou Z, Cini K, Dong B, Ma Y, Ma J, Burgner DP, Patton GC. Time trends in cardiovascular disease mortality across the BRICS: an age-period-cohort analysis of key nations with emerging economies using the global burden of disease study 2017. Circulation. 2020;141:790–799. doi: 10.1161/CIRCULATIONAHA.119.042864. [DOI] [PubMed] [Google Scholar]
- 31.Li HC, Chan SS, Wan ZS, Wang MP, Lam TH. An evaluation study of a gender-specific smoking cessation program to help Hong Kong Chinese women quit smoking. BMC Public Health. 2015;15:986. doi: 10.1186/s12889-015-2326-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Shibuya K, Inoue M, Lopez AD. Statistical modeling and projections of lung cancer mortality in 4 industrialized countries. Int J Cancer. 2005;117:476–485. doi: 10.1002/ijc.21078. [DOI] [PubMed] [Google Scholar]
- 33.Torre LA, Siegel RL, Jemal A. Lung cancer statistics. Adv Exp Med Biol. 2016;893:1–19. doi: 10.1007/978-3-319-24223-1_1. [DOI] [PubMed] [Google Scholar]