Pulmonary hypertension mortality trends in United States 1999–2019

Abstract

Purpose:

Pulmonary hypertension (PH) is a heterogenous, often progressive disorder leading to right heart failure and death. Previous analyses show stable PH mortality rates from 1980 to 2001 but increasing from 2001 to 2010 especially among women and non-Hispanic (NH) Black. This study seeks to identify recent trends in PH mortality in the United States from 1999 to 2019.

Methods:

Mortality rates among individuals more than or equal to 15 years of age were obtained from the Centers for Disease Control and Prevention’s (CDC) Wide-Ranging Online Data for Epidemiology Research (WONDER) database. ICD-10 codes were used to identify individuals with PH.

Results:

Between 1999 and 2019, PH was included as a cause on 429,105 recorded deaths. The average age-adjusted PH mortality rate was 7.9 per 100,000 individuals and increased by 1.9% per year. Higher age-adjusted mortality rates were experienced by females and NH Black persons. The crude mortality rate was 105.4 per 100,000 among those decedents 85 or older. From 1999 to 2019, mortality in PH and left heart disease co-occurrence increased at nearly double the annual rate of the overall PH group.

Conclusions:

Despite therapeutic advances for selected PH subgroups, the overall age-adjusted PH mortality rate increased significantly from 1999 to 2019 and previously reported racial disparities have persisted. These findings emphasize the need for additional study to improve outcomes in PH.

Introduction

Pulmonary hypertension (PH) is a disorder of elevated pressure in the pulmonary circulation that is often progressive, eventually leading to death from right heart failure [1]. PH has numerous underlying causes that impact disease risk and treatment. Accordingly, PH is divided into five clinical groups: 1) pulmonary arterial hypertension (PAH), 2) PH due to left heart disease, 3) PH due to lung diseases, 4) PH due to pulmonary artery obstruction (most often thromboembolic), and 5) PH due to unclear and/or multifactorial mechanisms [1,2].

Research studies and resulting advances for patient management, have primarily focused on Group 1 PH, a rare condition with prevalence of 5–25 per million adults [3,4]. More broadly though, PH is common, occurring in up to 5% of persons over 65, and is dominated by PH related to underlying left heart and lung diseases [3,5–7]. Observational data indicate that PH, regardless of subtype, increases mortality, and some reports suggest higher mortality from Group 2 and Group 3 PH compared to Group 1 PH [6–10]. Thus, PH of all subtypes, but particularly those associated with common heart and lung diseases, have major implications for individual patients, healthcare systems, and public health.

Previous analyses of mortality data compiled from United States death certificates showed that age-adjusted mortality rate associated with PH was stable from 1980 to 2001 but has since gradually increased through 2010 from 5.2 to 6.5 per 100,000, an observation largely concentrated in women [11–13]. Additionally, these studies noted higher age-adjusted mortality in Non-Hispanic (NH) Blacks compared to NH Whites (9.1 versus 6.5 per 100,000 in 2010) [12]. We analyzed mortality rates associated with PH in the United States from 1999 to 2019 to update and extend the previous observations and to better understand current age, gender, and race disparities, as well as trends among common PH subtypes.

Methods

Data extraction

Mortality rates (deaths per 100,000) in the United States for individuals more than 15 years old from 1999 to 2019 were obtained from the Centers for Disease Control and Prevention’s (CDC) Wide-Ranging Online Data for Epidemiology Research (WONDER) database [14]. Briefly, this resource compiles mortality data from all death certificates issued in the United States and includes a single underlying cause of death, up to 20 separate conditions contributing to death (multiple causes of death), and demographic information such as age, race, gender, year of death, and information about the decedent’s location of residence (region, state and county). Since 1999, causes of death have been reported using International Classification of Diseases, 10th Revision (ICD-10). WONDER provides tables of summarized data including total counts of deaths, reference population counts, and crude and age-adjusted mortality rates which can be summarized across one or more variables. Crude mortality rates are calculated based on population data of the entire United States derived from the Census Bureau. The age-adjusted mortality rates are standardized to the 2000 census population using direct standardization method.

We used the following ICD-10 codes to identify individuals with PH as one of the multiple causes of death: I27.0, I27.2, I27.8, and I27.9. Details of the ICD-10 codes are in the Supplement (Table A1). Group 1 PH was defined as deaths that included the I27.0 diagnosis code, indicating “primary pulmonary hypertension”. We extracted estimated age-adjusted rates, standard errors, and 95% confidence intervals (CI) by 10-year age groups, race and ethnicity, gender, and co-morbidities. Separately, we extracted data where death certificates included both PH and the following comorbidities that may be associated with PH: left heart disease, lung disease, liver disease, obesity, diabetes mellitus and connective tissue disease. These categories were not mutually exclusive. Details of ICD-10 codes used for these conditions are reported in the Supplement (Table A1). Race and ethnicity data are provided separately in WONDER. For our analyses, we combined race and ethnicity into the following groups: non-Hispanic (NH) white, NH Blacks, NH Asian and Pacific islander (NH-API), NH American Indian or Alaskan Native (NH-AIAN), and Hispanic which could have been of any coded race.

Statistical analysis

Trends in mortality rates were analyzed using joinpoint modeling (Joinpoint Regression Program version 4.9.0.0; National Cancer Institute) [15,16]. This technique employs regression analysis to identify trend segments with the best fit and smallest number of inflection points (called joinpoints) at which the direction or magnitude of the trend changes. For this analysis, annual age-adjusted or crude mortality rates (with standard errors) from CDC WONDER were analyzed by joinpoint regression. The average annual percentage of change (AAPC) was estimated for the entire 1999–2019 period from this regression and 95% confidence intervals (CI) were calculated based on standard error of the regression estimates. Additionally, the annual percent change (APC) was calculated for each trend segment. Using related underlying methods, null hypothesis testing of whether APCs or AAPCs were non-zero was performed in Joinpoint. Furthermore, for comparing strata (e.g., male vs. female), tests of parallelism and coincidence were performed in Joinpoint. Results of these tests were considered significant if the test’s P-value was < .05. This work was performed with publicly available, summary data without identifiable or person-level information and therefore is not considered human subjects research and Institutional Review Board review was not required.

Results

There were 429,105 deaths recorded between 1999 and 2019 where PH was included among the listed multiple causes of death. The overall age-adjusted mortality rate associated with PH in the United States for this period was 7.9 per 100,000 individuals (95% CI 7.9–7.9; Table 1). During this period overall mortality associated with PH increased on average by 1.9% per year (average annual percent change, AAPC; Table 1, Fig. 1).

Table 1.

Total deaths, crude rate, age adjusted rate, and average annual percentage change of pulmonary hypertension deaths from 1999 to 2019 for ages more than 15 years by gender, race, and comorbidities.

	1999–2019
	Total deaths	Crude rate / 100,000(95% CI)	Age-adjusted rate / 100,000 (95% CI)	AAPC (95% CI)*
Total	429,105	8.3 (8.3, 8.4)	7.9 (7.9, 7.9)	1.9 (1.4, 2.4)*
Female	261,345	9.9 (9.9, 9.9)	8.3 (8.2, 8.3)	2.3 (1.9, 2.7)*
Male	167,760	6.7 (6.7, 6.7)	7.4 (7.4, 7.5)	1.1 (0.7, 1.5)*
NH American Indian/Alaska Native	2148	5.3 (5.1, 5.5)	7.1 (6.8, 7.4)	2.8 (2.0, 3.6)*
NH Asian/Pacific Islander	7446	2.8 (2.7, 2.8)	3.7 (3.6, 3.8)	1.9 (1.4, 2.5)*
NH Blacks	56,176	9.0 (8.9, 9.1)	11.0 (10.9, 11.1)	2.1 (1.9, 2.2)*
NH White	343,308	9.9 (9.8, 9.9)	7.9 (7.9, 7.9)	2.0 (1.3, 2.7)*
Hispanic or Latino	19,165	2.6 (2.6, 2.7)	4.6 (4.5, 4.7)	2.1 (1.8, 2.4)*
Left heart disease†	207,757	4.0 (4.0, 4.1)	3.8 (3.8, 3.8)	3.9 (3.4, 4.5)*
Obesity†	19,520	0.4 (0.4, 0.4)	0.3 (0.3, 0.3)	4.2 (3.3, 5.1)*
Diabetes mellitus†	55,035	1.1 (1.1, 1.1)	1.0 (1.0, 1.0)	3.7 (3.1, 4.4)*
Liver disease†	8986	0.2 (0.2, 0.2)	0.2 (0.2, 0.2)	4.3 (3.3, 5.2)*
Lung disease†	170,540	3.3 (3.3, 3.3)	3.1 (3.1, 3.1)	−0.1 (−0.9, 0.8)
Connective tissue disease†	11,173	0.2 (0.2, 0.2)	0.2 (0.2, 0.2)	2.0 (−3.3, 7.6)*

^*P value <.05.

AAPC indicates average annual percentage change; CI = confidence interval; NH = non-Hispanic.

^†The categories are not mutually exclusive.NH, Non-Hispanic.

Fig. 1.

Age adjusted mortality rate (per 100,000) from 1999 to 2019 for Pulmonary Hypertension (PH), Pulmonary arterial hypertension (PAH), PH with left heart disease, and PH with chronic lung disease as contributing cause of death among individuals of age 15 years and older. The data for PAH from 1999 to 2002 was not included due to coding anomaly. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Age groups

The mortality rate with PH per 100,000 population is greater as age increases (Table 2). The crude mortality rate (per 100,000 population) increased from less than 1 among those aged 15–24 to 105.4 among those decedents aged more than or equal to 85 (Table 2). There was a significant decrease in mortality rate in the 15–24 years group (AAPC −2.8%) and a significant increase in mortality for age groups 55–64 years (AAPC + 0.3%), 75–84 years (AAPC + 2.0%) and 85+ years (AAPC + 5.0%) (Table 2, Fig. 2). Over the study period, there has been a trend for mortality rates to decrease or remain stable in younger age groups (those 15–54) and increase in older age groups (Table 2, Fig. 2).

Table 2.

Total deaths, crude mortality rate, average annual percentage change of pulmonary hypertension associated deaths from 1999 to 2019 for ages 15 and above by age groups.

Age groups	1999–2019
	Total deaths	Crude rate (95% Cl)	AAPC (95% Cl)
15–24 y	1871	0.2 (0.2, 0.2)	−2.8 (−3.6, −1.9)*
25–34 y	4167	0.5 (0.5, 0.5)	−1.0 (−1.9, 0.0)
35–44 y	10,110	1.1 (1.1, 1.2)	−0.5 (−1.6, 0.6)
45–54 y	24,786	2.8 (2.8, 2.8)	0.2 (0.0, 0.4)
55–64 y	51,187	7.1 (7.0, 7.1)	0.3 (0.0, 0.6)*
65 –74 y	90,110	18.9 (18.7, 19.0)	0.0 (−0.3, 0.4)
75–84 y	127,973	45.4 (45.1, 45.6)	2.0 (1.3, 2.8)*
85+y	118,901	105.4 (104.8, 106.0)	5.0 (4.4, 5.6)*

^*P value < .05.

AAPC average annual percentage change.

Fig. 2.

Age specific mortality rate (per 100,000) from 1999 to 2019 for Pulmonary Hypertension (PH), as contributing cause of death among individuals of age 15 years and older by 10-year age groups. Panel A: Ages 55–85+ years. Panel B: Ages 15–54 years.(For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Gender

Between 1999 and 2019, the average age-adjusted mortality rates with PH for females and males were 8.3 and 7.4 cases per 100,000 population, respectively (P < .05 for coincidence) (Table 1). In addition to having higher mortality rates, females experienced a greater increase in age-adjusted mortality rate during this period compared to males with AAPC + 2.3% versus + 1.1%, respectively (P < .05 for parallelism, Table 1, Fig. 3).

Fig. 3.

Age adjusted mortality rate (per 100,000) from 1999–2019 for Pulmonary Hypertension (PH), as contributing cause of death among individuals of age 15 years and older by gender.(For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Race

The age-adjusted mortality rate associated with PH per 100,000 population for 1999–2019 varied significantly by race. The rate for persons who were NH White was 7.9, NH Black 11.0, NH Asian/Pacific Islander (API) was 3.7, NH American Indian/Alaska Native (AIAN) was 7.1 (Table 1). The rate among Hispanic persons was 4.6 (Table 1). NH Black females had the highest age-adjusted mortality rate (12.2 per 100,000 population) of any gender and race grouping (Table 3).

Table 3.

Age-adjusted mortality rates associated with pulmonary hypertension for females and males by race and ethnicity from 1999 to 2019.

Gender	Race	Total Deaths	Age-adjusted Rate / 100,000 (95% CI)
Female	NH American Indian or Alaska Native	1312	7.8 (7.4, 8.3)
Female	NH Asian or Pacific Islander	4368	3.8 (3.7, 3.9)
Female	NH Black	36,980	12.2 (12.0, 12.3)
Female	NH White	206,255	8.1 (8.1, 8.1)
Female	Hispanic	11,955	5.0 (5.0, 5.1)
Female	All races and ethnicities	260,870 *	8.3 (8.2, 8.3) **
Male	NH American Indian or Alaska Native	836	6.3 (5.8, 6.8)
Male	NH Asian or Pacific Islander	3078	3.7 (3.5, 3.8)
Male	NH Black	19,196	9.4 (9.2, 9.5)
Male	NH White	137,053	7.6 (7.6, 7.6)
Male	Hispanic	7210	3.9 (3.8, 4.0)
Male	All races and ethnicities	167,373 *	7.4 (7.4, 7.4) **
Male and females	All races and ethnicities	409,078 *	8.1 (8.1, 8.1) **

^*Cases where Hispanic or Non-Hispanic ethnicity was not reported were excluded.NH = Non-Hispanic.

^**Bold values denote subtotal rows. E.g. for Female with total deaths of 260,870, this is the subtotal of deaths for all races and ethnicities which are broken down in the rows above.

There has been a significant increase in mortality rate between 1999 and 2019 for all race-ethnicity groups. The AAPC from 1999 to 2019 for all races and ethnicity groups ranged between +2.0 and +2.8% (Table 1, Fig. 4). There was a steady increase in mortality throughout the study period for persons who were NH Black, NH-API, NH-AIAN and Hispanic (Table E2, Fig. 4) with no inflection points detected.

Fig. 4.

Mortality rate (per 100,000) from 1999 to 2019 for Pulmonary Hypertension (PH), as contributing cause of death among individuals of age 15 years and older by race and/or ethnicity.(For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Comorbidities associated with PH

From 1999 to 2019, the rates of age-adjusted mortality with death certificates indicating both PH and left- heart disease, or PH and chronic lung diseases were similar, at 3.8 and 3.1 per 100,000 persons, respectively (Table 1). The mortality rate from the combination of PH and left heart disease increased significantly over this period (AAPC + 4.0%) while the mortality rate for the combination of PH and lung disease remained unchanged (AAPC −0.1%, P > .5 for a non-zero slope) (Table 1, Fig. 1). There were also significant increases in mortality rates in individuals with the combinations of PH and obesity, diabetes mellitus, liver disease, and connective tissue disease from 1999 to 2019 (Table 1).

Group 1 PH associated mortality rate increased significantly from 1999 to 2002 (APC +4.6%), declined from 2002 to 2005 (APC −70.6%), and increased from 2005 to 2019 (APC + 4.0%) (Table A4, Fig. 1).

Discussion

This updated analysis of U.S. death data indicates that the age-adjusted mortality rate associated with PH increased significantly between 1999 and 2019 – increasing about 2% per year. This trend has continued through a period of expanding medical therapies for Group 1 PH. Additionally, this report provides information regarding age, gender, race and ethnicity, and disease subsets with high numbers of deaths with PH that is relevant to population health efforts.

The rising trend of PH mortality is most pronounced among the older population. For persons aged 75 or older, PH mortality is markedly higher and rising at a much faster pace when compared to younger age groups. While this could be a phenomenon of the aging U.S. population, or increased recognition of PH, it likely reflects that PH is a late complication of the accumulation of chronic cardiopulmonary conditions and is common among those 65 and older [6].

Our study notes significant differences in PH mortality between race and ethnicity groups, yet prior studies of this topic have been mixed. REVEAL, the largest Group 1 PH registry in the United States in which enrollment is primarily via specialty referral centers, no differences in mortality are seen based on race or ethnicity [17]. Yet data obtained in healthcare settings indicate Black race is an independent risk factor for PH [18,19]. Our analysis shows that NH Black persons have higher age-adjusted mortality from PH compared to all other racial and ethnic groups. While this has been demonstrated using CDC WONDER data previously [12], our analysis suggests that the gap is widening in recent years. This conflicts with death data more broadly suggesting that racial differences in all-cause mortality have decreased between 2000 and 2017 [20]. Increased prevalence of heart failure, hypertension and diabetes among Black persons, all of which promote PH, may explain the observed higher mortality rate in our study [21–24]. Alternatively, a separate study of racial disparities in PH mortality found that socioeconomic status fully explained the observed differences [18], suggesting that social determinants of health may underly our observations.

Age-adjusted mortality with PH in females is higher and increasing at a more rapid rate when compared to males over the study period reported here. PH, primarily of the Group 1 PH subtype, has known gender disparities [25]. Group 1 PH has higher prevalence among females with modern registries being 60%–80% female. Since Group 1 PH is rare compared to other subtypes of PH, our study suggests that other forms of PH may also be more common—or more commonly recognized—among women at the time of death. Indeed, other studies have shown that diastolic heart dysfunction, a major cause of Group 2 PH due to left heart disease, disproportionately affects women [26]. While significant research has been dedicated to sex differences of Group 1 PH [25], the higher number among women of overall deaths with PH in this study suggest that additional work is needed to understand gender differences in other forms of PH.

Most PH is related to underlying chronic cardiopulmonary conditions [6]. Our analysis shows that the age-adjusted mortality rate associated with PH and left heart disease, the most common cause of PH, has increased significantly from 1999 to 2019. PH due to lung diseases, the second most common cause of PH, has remained stable between 1999 and 2019, potentially explained by the general decline in deaths from chronic lung diseases over the last twenty years [27,28]. Yet, PH due to both chronic lung disease and left heart disease deserve ongoing focus as the total number of deaths with PH in these conditions remain very high and reflect a major public health problem.

Our results must be interpreted within the context of limitations of the data source analytic methods. First, our data include the number of deaths within the U.S. population but cannot evaluate risk of death among persons diagnosed with PH. Therefore, higher mortality in one group or increased mortality rate over time may be attributed to 1) increased disease incidence, 2) increases in recognition and diagnosis which may appear as increased disease incidence, 3) increase in mortality for the condition, or 4) some combination of these factors. This becomes especially relevant with respect to data regarding age, gender, and racial disparities. While having their own limitations due to referral bias, existing multicenter registries of Group 1 and 4 PH, relatively rare in the population, may be better suited to understand disparities and trends in these conditions. Second, mortality data in CDC WONDER are compiled from death certificates which have recognized limitations including possible misclassification bias [29–31]. Because PH often results from underlying cardiopulmonary conditions that may be more clinically obvious, we suspect that PH may be present more often than reported in this data source. Third, we recognize that our evaluation of death due to a combination of PH and other diseases such as left heart and lung diseases is prone to misclassification of both conditions. Finally, the code I27.2 (“Other secondary PH”) was added in 2003 leading to a marked increase in cases attributed to that code and a decrease in cases attributed to I27.0 (“Primary PAH”) over the ensuing two years. Recognizing these limitations and acknowledging the lack of prospective registries focused on non-Group 1 PH, this analysis provides important epidemiological insights in mortality of distinct PH subgroups. We believe that data from broader sources like CDC WONDER can provide valuable insight into conditions that less often included in disease-focused studies.

In summary, publicly reported mortality data from the United States show increasing frequency of deaths associated with PH over the last 20 years (1999–2019). Deaths with PH are experienced more in women, NH Black persons, and older populations. Furthermore, these differences are worsening over time, including over the 2010–2019 period which has not been previously reported. PH paired with left- heart disease seems to be a significant driver of this increasing population health problem. While dedicated prospective epidemiology studies of PH, especially non-Group 1 PH, have been infrequent, we believe that the current findings indicate a call to action. To address this deepening population health concern, there is a need to prioritize research spanning the spectrum of preclinical and translational science focused on common causes of PH.

Abbreviations list

PH

pulmonary hypertension

PAH

pulmonary arterial hypertension

LHD

left heart disease

CTEPH

chronic thromboembolic pulmonary hypertension

NH

Non-Hispanic

NH-API

Non-Hispanic Asian and Pacific islander

NH-AIAN

Non-Hispanic NH American Indian or Alaskan native

CDC

centers for disease control and prevention

WONDER

wide-ranging online data for epidemiology research

APC

annual percent change

AAPC

average annual percentage of change