Mortality Related to Chronic Obstructive Pulmonary Disease and Co-morbidities in the United States, A Multiple Causes of Death Analysis

ABSTRACT

Chronic obstructive pulmonary disease (COPD) mortality based on the underlying cause of death (UCOD) underestimates disease burden. We aimed to determine the current COPD mortality rate, trends and the distribution of co-morbidities using United States (US) multiple-cause of death (MCOD) records.

All 38,905,575 death certificates of decedents aged ≥45 years in the United States were analyzed for 1999–2015. COPD was defined by ICD–10 codes J40–J44 and J47 based either on the UCOD or up to 20 contributing causes coded. Annual age–standardized COPD death rates were computed by age, gender and race/ethnicity for those with any mention of COPD.

In 2015, COPD was mentioned in 11.59% (292,572 deaths) in MCOD, compared to 11.13% (243,617 deaths) in 1999, a 4% increase. However, it was reported as the UCOD for only 5.56% and 4.97% in 2015 and 1999 respectively, an 11% increase. The most common UCOD in subjects with any mention of COPD was respiratory disorders in 49% of males and 55% of females. The relative change in death rates differed between MCOD and UCOD. For example, among non-Hispanic white females aged 65–74 years the UCOD rate per 100,000 (95% CI) decreased from 163 (160–166) to 147 (145–150), average annual percent decrease (AAPD) –0.26, while the MCOD rate decreased from 308 (304–311) to 263 (260–267), AAPD –0.87.

Statistics based on UCOD understated the burden of COPD in the United States. MCOD rates were twice as high as UCOD rates. The relative change in death percent or rates differed between MCOD and UCOD. MCOD analysis should be repeated periodically to help evaluate the burden of COPD-related mortality.

KEYWORDS:

• COPD
• deaths
• tobacco

Introduction

Globally in 2015, 3.2 million people (95% uncertainty interval [UI] 3.1 million to 3.3 million) died from chronic obstructive pulmonary disease (COPD) worldwide, an increase of 11.6% (95% UI 5.3 to 19.8) compared with 1990. Deaths from COPD were eight times more common than deaths from asthma (1). In the United States today, COPD is the third leading cause of death and affects 16 million Americans diagnosed with the disease and millions more who likely do not know they have it. It costs the healthcare system more than $32 billion a year and can lead to long-term disability, and significantly affect the quality of life (2).

The Global Burden of Disease Study (GBD) for 2015 estimated that 482,000 deaths from COPD occurred in high-income countries (highest quintile of social development index) in 2015 and age standardized death rates declined 26% in high income countries between 1990 and 2015 (3). Among high-income regions, GBD estimates for 2016 indicated the age-standardized COPD death rates for females were highest in high-income North America (US, Canada and Greenland), intermediate in Australasia, Southern Latin America and Western Europe, and lowest in high-income Asia/Pacific. Rates for males were highest in Southern Latin America, intermediate in the high-income North America, Australasia and Western Europe, and lowest in high-income Asia/Pacific (3).

COPD mortality rates worldwide are obtained from standard mortality statistics founded on internationally adopted algorithms that identify the underlying cause of death (UCOD) from all the conditions reported on a death certificate (1,3). For each death certificate a single underlying cause of death is coded and, in the United States, up to 20 additional multiple causes of death are coded using the International Classification of Diseases (ICD10). The underlying cause of death is defined by the World Health Organization (WHO) as “the disease or injury which initiated the train of events leading directly to death, or the circumstances of the accident or violence which produced the fatal injury” (4). Only a few studies have shown that assessments of COPD mortality based only on the UCOD derived from death certificates underestimate disease burden (5,6). The only analysis in the United States of multiple-cause of death (MCOD) records from 1979–1993 showed that of all the decedents who had a diagnosis of COPD or asthma listed on their death certificates, only 43.3% had COPD or asthma as an UCOD (5). No analysis of more recent US MCOD data has been published. Therefore, in this study, the burden of COPD mortality in the United States 1999–2015 was analyzed, as well as the trends and patterns of reporting of COPD and its co-morbidities in death certificates, using MCOD records.

Materials and methods

The Multiple Cause of Death data were obtained from CDC WONDER including national mortality from death certificates and population data spanning the years 1999–2015 (7). The Multiple Cause of Death data are formed by the Division of Vital Statistics, National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention (CDC) and the United States Department of Health and Human Services (US DHHS). Use of these data is exempt from patient consent and IRB approval.

The mortality data are based on data from all death certificates filed in the 50 states and the District of Columbia. Mortality data from the death certificates are coded by the states and provided to NCHS through the Vital Statistics Cooperative Program or coded by NCHS from copies of the original death certificates provided to NCHS by the State registration offices (6).

All 38,905,575 death certificates of decedents aged ≥45 years in the United States were analyzed through 1999–2015. The UCOD was selected by the Automated Classification of Medical Entities software (7, 8). Causes of death are classified in accordance with the International Classification of Disease.

COPD was defined based either on the UCOD or on any mention of COPD in death certificates (MCOD) among subjects aged ≥45 years. In order to be comparable with previous analyses for example Marcon et al. (6), we used the following ICD-10 codes: bronchitis, not specified as acute or chronic (J40); simple and mucopurulent chronic bronchitis (J41); unspecified chronic bronchitis (J42); emphysema (J43); other chronic obstructive pulmonary disease (J44); bronchiectasis (J47). In the United States, ICD-10 was in use for mortality coding from 1999–2015, although ICD-9 CM continued to be used for hospital discharge diagnosis until recently. The population estimates are bridged-race estimates based on Bureau of the Census estimates of total United States, State and county resident populations. The year 2011–2015 population estimates are bridged-race postcensal estimates of the July 1 resident population.

The plan of analysis was as follows: Annual age-standardized COPD death rates were computed for 45–85+ year-old subjects. The 2000 standard populations were obtained from the National Center for Health Statistics. Beginning with the 1999 data year, NCHS adopted the year 2000 projected population of the United States as the standard population for use in age adjusting death rates. Proportional mortality (percentage of all registered deaths) and age- (5-year classes between 40 and 89 years, and ≥90 years) and gender-specific mortality rates were computed for COPD selected as the UCOD, and for COPD mentioned anywhere in a certificate (MCOD). In other words, proportional mortality both in UCOD and MCOD-based analyses was tabulated by gender and race. Also tabulated was the distribution of the most common UCOD in subjects with any mention of COPD in their death certificates in the recent years (2013–2015). The UCOD/MCOD ratio was computed by age and gender. Also tabulated was the UCOD and MCOD rate per 100,000 across ages, races and gender. The CDC statistical software was used for analysis referenced (7).

Results

In 1999–2015,^{Figure 1, Figure 2, Figure 3, Figure 4}among 2,254,556 deaths with UCOD ICD10 codes J40-J47 (Chronic lower respiratory diseases), 6,544 deaths were coded J40 (Bronchitis not specified as acute or chronic), 112 deaths were J41 (Simple or mucopurulent chronic bronchitis), 5,465 were J42 (Unspecified chronic bronchitis) and 16,293 were J47 (Bronchiectasis). In comparison, 1,717,340 were coded J44.9 (Chronic obstructive pulmonary disease, unspecified), the largest category.

Between 1999 and 2015, the age-standardized death rates for mortality from COPD as underlying cause of death in people age 45 years and older decreased from 125.6 per 100,000 to 116.5 per 100,000 per year (–8% change). Further analysis showed marked variation in trends between males and females. Among males aged 45 years and older, the overall aged standardized death rates for COPD mortality decreased from 165.3 per 100,000 to 141 per 100,000 (–27% change), while in females the age standardized death rate in the same age group increased from 102.7 per 100,000 to 104.2 per 100,000 (5% change).

In 2015, COPD was selected as the UCOD for 5.56% of the all cause deaths in people aged 45 and above, and this contrasts with the MOCD analysis that showed COPD was mentioned in 11.59% (292,572 deaths) in the same year. Similarly, in 1999 COPD accounted for 4.97% as the UCOD, while it was mentioned as MCOD in11.59% (292,572 deaths). Between 1999 and 2015 in all the decedents who had a diagnosis of COPD listed on their death certificates, only 48.89% and 50.34% respectively had a diagnosis of COPD as an UCOD (a 6% increase).

Figure 1 shows that the UCOD/MCOD ratio was highest in the 75–84 and 85+ age groups. For example, for white men, UCOD/MCOD ratio was 0.51 for both age groups compared to <0.40 for the age group 45 -54. Among whites the UCOD/MCOD ratio was higher in women, however in blacks the ratio was similar in men and women. The ratio was higher in whites than blacks in both genders.

Figure 1. Ratio of underlying cause of death to multiple cause of death (UCOD/MCOD) among non-Hispanic black and white males and females aged 45–85+ years.

Figures 2 and 3 show the trends in UCOD and MCOD rate per 100,000 by race and gender. The UCOD and MCOD rates/100,000 showed a decline in men but an increase in women. The relative change in death rates differed between MCOD and UCOD. For example, amongst non-Hispanic white females aged 65–74 years the UCOD rate per 100,000 (95% CI) decreased from 163 (160–166) to 147 (145–150), average annual percent 155 decrease (AAPD) -0.26, while the MCOD rate decreased from 308 (304–311) to 263 (260–267), AAPD -0.87. Table 1 shows UCOD and MCOD rates per 100,000 increased with age in both races and genders and were higher in men than women and whites than blacks. At age 65–74, UCOD/MCOD was highest in white women (0.56) and lowest in black men (0.45).

Figure 2. Age-standardized COPD mortality rates per 100,000/year based on underlying cause of death and multiple cause of death analyses in the United States between 1999 and 2015, among non-Hispanic black and white males aged 45–85+ years.

Figure 3. Age-standardized COPD mortality rates per 100,000/year based on underlying cause of death and multiple cause of death analyses in the United States between 1999 and 2015, among non-Hispanic black and white females aged 45–85+ years.

Figure 4. Proportional mortality (percentage of all registered deaths) from COPD both in UCOD and MCOD-based analyses among non-Hispanic black and white males and females aged 45–85+ years.

Proportional mortality (percentage of all registered deaths) from COPD both in UCOD and MCOD-based analyses was higher in women than men among the white race but the black women had a slightly lower proportional mortality than black men. (Table 1 and Figure 4). For example, in white men aged 65–74, proportional mortality using UCOD was 8% and proportional mortality using MCOD was 16% while in white women of the same age, UCOD was 10% and the MCOD was 18%, respectively. The proportional mortality percent for MCOD was higher than the percent for UCOD in men and women of both races.

Table 1. Number of deaths, proportional mortality (% of all deaths) and mortality rates (/100,000/year) for COPD selected as the underlying cause of death (UCOD) or mentioned in any position of death certificates (MCOD) of non–Hispanics: United States, 2013–2015.

	Number of deaths			Proportional mortality		Rate /100,000		Ratio
	UCOD	MCOD	All causes	UCOD	MCOD	UCOD	MCOD	UCOD/MCOD
White men
45–54	4,542	12,579	220,280	2%	6%	8.6	29.2	0.36
55–64	21,356	50,740	463,125	5%	11%	32.1	118.6	0.42
65–74	47,855	100,059	633,965	8%	16%	76.0	343.5	0.48
75–84	62,161	121,976	787,979	8%	15%	171.3	855.5	0.51
85+	45,814	89,215	810,783	6%	11%	311.9	1698.2	0.51
All^*	181,728	374,569	2,916,132	6%	13%	142.6	291.8	0.49
White women
45–54	5,289	11,616	141,508	4%	8%	8.1	26.6	0.46
55–64	19,587	38,408	292,203	7%	13%	41.4	85.7	0.51
65–74	47,754	84,910	473,857	10%	18%	132.5	264.0	0.56
75–84	68,452	119,379	752,240	9%	16%	297.0	651.9	0.57
85+	67,097	119,867	1,350,979	5%	9%	590.9	1197.1	0.56
All^*	208,179	374,180	3,010,787	7%	12%	121.7	218.9	0.56
Black men
45–54	622	1879	53,552	1%	4%	12.1	24.6	0.33
55–64	2578	6815	103,355	2%	7%	43.7	109.4	0.38
65–74	4218	9709	99,262	4%	10%	148.4	329.9	0.45
75–84	3922	8570	84,194	5%	10%	373.8	649.0	0.46
85+	2329	4776	53,612	4%	9%	670.1	1211.8	0.49
All^*	13669	31749	393,975	3%	8%	103.1	232.2	0.43
Black women
45–54	739	1,797	40,677	2%	4%	10.5	20.8	0.41
55–64	2,383	5,399	74,090	3%	7%	49.9	72.6	0.44
65–74	3,216	6,863	81,632	4%	8%	164.3	162.3	0.47
75–84	3,686	7,379	95,974	4%	8%	436.0	342.8	0.50
85+	2,981	5,912	114,438	3%	5%	872.1	618.5	0.50
All^*	13,005	27,350	406,811	3%	7%	61.7	128.6	0.48

Abbreviations: UCOD – Underlying cause of death; MCOD – Multiple cause of death

^* = All age groups 45–85+

Table 2 and Figures 5a and 5b show the distribution of the most common UCOD in subjects with any mention of COPD in their death certificates in the recent years (2013–2015). Deaths from respiratory diseases, including COPD, were 49% and 55% in males and females, respectively followed by circulatory disorders which were 24% in males and 21% in females.

Table 2. Underlying cause of death (UCOD) according to the International Classification of Diseases, 10th Edition (ICD-10) in certificates mentioning COPD: subjects aged ≥45 years, Unites States, 2013–2015.

UCD – ICD chapter	UCD – ICD Chapter Code	Deaths	Crude rate	Percent
Women
Neoplasms	C00-D48	26,6894	25.8	12.49
Endocrine, nutritional and metabolic diseases	E00-E88	42,653	4.1	2.00
Mental and behavioral disorders	F01-F99	29,359	2.8	1.37
Diseases of the nervous system	G00-G98	34,961	3.4	1.64
Diseases of the circulatory system	I00-I99	456,957	44.1	21.38
Diseases of the respiratory system	J00-J98	1,182,472	114.2	55.34
Diseases of the digestive system	K00-K92	37,070	3.6	1.73
Diseases of the genitourinary system	N00-N98	18,344	1.8	0.86
External causes of morbidity and mortality	V01-Y89	26,228	2.5	1.23
Total		2,136,929	206.3
Men
Neoplasms	C00-D48	363,687	40.1	16.12
Endocrine, nutritional and metabolic diseases	E00-E88	47,745	5.3	2.12
Mental and behavioral disorders	F01-F99	25,723	2.8	1.14
Diseases of the nervous system	G00-G98	31,919	3.5	1.41
Diseases of the circulatory system	I00-I99	546,696	60.3	24.23
Diseases of the respiratory system	J00-J98	1,108,938	122.4	49.15
Diseases of the digestive system	K00-K92	41,402	4.6	1.83
Diseases of the genitourinary system	N00-N98	19,551	2.2	0.87
External causes of morbidity and mortality	V01-Y89	31,683	3.5	1.40
Total		2,256,403	249.1

Figure 5. Distribution of the most common UCOD in subjects with any mention of COPD in their death certificates in the recent years (2013–2015) among non-Hipsanic, white and black (a) females and (b) males.

Discussion

Our study showed that the burden of COPD-related mortality in the United States from 1999 to 2015 estimated by means of a MCOD-based analysis was twice greater compared to estimates using standard UCOD statistics. This underestimation may be because COPD diagnosis is understated on death certificates and when mentioned, it is not selected as the UCOD.

Our finding updates a previous analysis of US data by Mannino et al. which showed that mortality related to obstructive lung disease was underestimated when UCOD is used instead of MCOD (5). Our results are also in agreement with other international studies. However, among countries mention of COPD greatly varies both in UCOD and MCOD analyses (9,10,11,12). For instance, a study in France by Furham et al. showed COPD was the underlying cause of 1.4% of deaths (deaths from COPD) and was mentioned among MCOD on the death certificate in 3.0% (12). Similarly, in England obstructive lung diseases were the underlying cause of death for under 5% of all deaths; however, it was mentioned on the death certificate for 8% of deaths (13). In addition, a recent study from North Eastern Italy found that the COPD mortality rates were three times greater in the MCOD analyses when compared to UCOD (6).

Interestingly, we found MCOD rate of deaths from COPD decreased among men, but the rate continued to increase among both black and white women. Increasing COPD trend among women in the United States was reported in previous studies (14,15,16,17). This increasing mortality trend in women from COPD is probably following the growing COPD burden seen among women in the United States (5,18). The annual number of COPD deaths in the United States is now higher in women than in men (18). This trend is consistent with some reports (5,18) and differs from others (6,19). One might ask why COPD is increasing in women in the United States. Possible explanations include: gender difference in susceptibility to the lung-damaging effects of cigarette smoking and more severe disease in those with COPD (14,15,16). Intersections between gender, smoking and lung function are complicated, especially among subjects with COPD. Focusing on subgroups of the study sample may improve our ability to detect gender differences in COPD (14). Additionally, the extent to which both susceptibility and vulnerability contribute and interact to explain sex and gender differences for COPD development and its severity and associated mortality remains largely unexplored.

In our study, the most common UCOD in subjects with any mention of COPD in their death certificates were diseases of the respiratory system, closely followed by diseases of the cardiovascular system in both genders. This is to be expected: not only do these diseases share a common risk factor (tobacco exposure), but COPD is a risk factor for respiratory failure and cardiovascular mortality. Our results also agree with previous studies which reported similar finding (20,21). These finding highlights the need for detailed epidemiology of chronic respiratory diseases (CRDs) risk factors and co-morbidities to plan preventive interventions to reduce the burden of chronic respiratory diseases on population health (11).

Finally, our analysis showed the greatest UCOD and MCOD rates were seen among older non-Hispanic white males. This supports the hypothesis that racial differences in disease patterns may exist. Blacks may have lower incidence of lower-lung emphysema because of genetic differences, such as alpha 1 anti-trypsin deficiency, which occurs at higher frequency among Whites of European descent (22,23). Black lives might have been lost to smoking-related cardiovascular/cancer mortality at earlier ages before they could die with COPD. There is a possibility of a higher prevalence of under diagnoses of COPD in blacks due to lesser access to care in blacks than whites. These factors could help explain the lower mortality rates in blacks than whites at older ages, therefore producing the observed racial gap (24). Nevertheless, further investigation is needed to determine the factors involved in inducing and modifying the different patterns of COPD mortality expressed by these two races.

Strengths and limitations

The benefits of using the MCOD data on COPD include coverage of the entire United States, assessment of mortality trends and the ability to assess comorbidity. Unfortunately, the use of death certificate data for epidemiologic studies also has certain drawbacks. Death certificate data are not independently authenticated, and some degree of misclassification unavoidably occurs. Some of the ICD-10 codes used might have also included patients who do not suffer from COPD.

In addition, death certificates do not include data on the severity of COPD, and exposure to risk factors such as occupational exposures and family history. Tobacco use data are inconsistently collected on death certificates. CDC datasets do not include US national data on COPD incidence for correlation with mortality trends.

Conclusions

Our findings support the premise that UCOD analysis underestimates COPD burden. MCOD analysis should be adopted to fully evaluate the burden of COPD-related mortality in the United States. Finally, the lack of decline in COPD-related mortality rates among women confirms the findings from other studies and signifies an important health disparity. Targeted public health policies and intervention are needed to reduce tobacco use and decrease occupational and environmental exposures. These measures if put in place may help reduce the number of people who die from COPD.

Conflicts of interest

Authors have no conflicts of interest.

Acknowlegdments

Authors have met the following criteria:

1.	Made substantial contributions to conception and design, or acquisition of data, or analysis and interpretation of data (JO, RG)
2.	Has drafted the submitted article or revised it critically for important intellectual content (JO, AM, RG, PW)
3.	Has provided final approval of the version to be published. (JO, AM, RG)
4.	Has agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. (RG)
5.	The authors acknowledge the assistance of Ms K.E. Dodd with proof reading.