Incidence and Healthcare Burden of Pertussis among Older Adults with and without Pre-Existing Chronic Obstructive Pulmonary Disease or Asthma in South Korea

Abstract A retrospective cohort study was conducted to examine trends in the incidence and burden of pertussis among adults ≥50 years in South Korea, with/without pre-existing chronic obstructive pulmonary disease (COPD) or asthma. The nationwide Health Insurance Review and Assessment Service (HIRA) database was used to identify patients ≥50 years diagnosed with pertussis (2009–2018). Mean annual incidence of pertussis per 100 000 persons and overall mean incidence rate ratios (IRR) were calculated for patients with pre-existing COPD or asthma versus those with neither. Incremental healthcare costs (all-cause and pertussis-related) and healthcare utilisation (number of outpatient visits, emergency room visits, and number and length of hospitalisations) up to 12 months after, compared to 3 months before pertussis diagnosis, were also measured for each group (matched on sex, age, and Charlson Comorbidity Index). Of 1011 pertussis cases, 175 had asthma, 96 had COPD (not mutually exclusive), and 796 had neither. Overall mean pertussis incidence was 2.5, 3.4, and 0.5 for adults with pre-existing COPD, asthma, and those with neither. IRR (95% confidence interval) of pertussis for adults with pre-existing COPD and asthma was 4.9 (4.0–6.1) and 6.7 (5.7–7.9). Both COPD-pertussis and asthma-pertussis groups had higher mean incremental all-cause costs and length of hospitalisations than the general-pertussis group 3 months following pertussis diagnosis. In conclusion, individuals ≥50 years in South Korea with pre-existing COPD or asthma were at an increased risk of being diagnosed with pertussis and had higher healthcare resource utilisation than those without these conditions.


Introduction
Pertussis, or "whooping cough", is a highly contagious respiratory infection caused by Bordetella pertussis [1].Despite the availability and wide-spread use of pertussis vaccines, an estimated 50 million cases of pertussis and 300 000 deaths occur annually worldwide [2].
A recent resurgence of pertussis in countries with high childhood pertussis vaccination coverage has been reported, with waning immunity in the years post-vaccination cited as one of the main driving forces of this resurgence [3][4][5].Epidemiological data from Europe and the United States (US) indicate an increasing proportion of cases in older individuals [6].Resurgence of pertussis has also been reported in South Korea among children and infants as well as among older adults, particularly those ≥60 [7,8].However, as clinical presentation of pertussis is usually atypical among adult pertussis patients, the burden of pertussis in adults may be underestimated [2].
Healthcare burden of pertussis in adults is substantial; a US study estimated average costs of $1800 and $14 500 in outpatient and inpatient settings respectively across episodes of pertussis per patient [9].Adults with underlying comorbidities or pre-existing respiratory conditions, including chronic obstructive disease (COPD) or asthma, may be at increased risk of pertussis diagnosis and of complications and subsequent increased healthcare burden following pertussis diagnosis [10][11][12][13].
There is a lack of local data in South Korea on whether pre-existing respiratory conditions such as COPD and asthma may increase risk of contracting and being diagnosed with pertussis.This is especially significant as the prevalence of COPD in South Korea, estimated as 9.0-38.8%among adult aged ≥50 between 2007 and 2015, is considerably high (estimated worldwide prevalence of 11.7%) potentially due to high rates of smoking and air pollution, amongst other factors [14,15].Prevalence of diagnosed asthma in adults (estimated as 3.7% in 2014) is slightly lower than global estimates (4.5%) [16,17].However, the prevalence of asthma in South Korea has been reported to be increasing, particularly amongst adults ≥60 years [18].
To better understand the burden of pertussis in adults in South Korea, we conducted a retrospective cohort study to examine trends in the incidence and healthcare resource utilisation of diagnosed pertussis cases among adults ≥50 years, with or without pre-existing COPD or asthma.

Data source
Individuals diagnosed with pertussis were identified using the South Korea national health insurance claims database operated by the Health Insurance Review and Assessment Service (HIRA), with a data period from January 1, 2008-December 31, 2019.This database contains comprehensive information pertaining to healthcare services for almost 50 million beneficiaries (98% of the total South Korean population).All diagnoses in this database are coded using the Korea Standard Classification of Diseases, 7 th revision (KCD-7), which is based on the International Classification of Diseases, 10 th revision (ICD-10) [19].
Due to the HIRA's maximum data size extraction policy per request, the HIRA-National Patient Sample (HIRA-NPS) database was used to estimate the denominators for the incidence calculations.The HIRA-NPS is an annually sampled population-based cohort that includes a stratified random sample of 3% of the entire population [20].From the HIRA-NPS, a number of individuals ≥50 years each year were calculated and divided by 3% to estimate total population at risk.

Inclusion criteria and study cohorts
Individuals were included in this analysis if aged ≥50 years and diagnosed with pertussis (ICD-10 diagnosis code A37: "Whooping cough").Individuals with a history of pertussis within 12 months preceding the pertussis diagnosis were excluded to only include incident cases.As individuals were required to have a 12-month period prior and following pertussis diagnosis to sufficiently capture changes in health care resource utilisation, pertussis cases diagnosed in 2008 and 2019 were also excluded.
Included pertussis cases were grouped into general-pertussis, COPD-pertussis, and asthma-pertussis groups based on the presence of pre-existing COPD or asthma in the 12-months preceding date of pertussis diagnosis.Patients were classified into the COPD-pertussis group if they had ≥1 diagnoses of COPD or emphysema (ICD-10: J43-J44 as primary or secondary diagnosis) and ≥1 prescription of a COPD medication.Patients were classified into the asthma-pertussis group if they had 1 inpatient or emergency room (ER) visit or ≥2 outpatient visits with a diagnosis of asthma (ICD-10: J45-J46 as primary or secondary diagnosis) and ≥1 prescription for an asthma medication (included medications listed in Table 1, supplementary material).All other pertussis cases were classified into the general-pertussis group.

Healthcare burden
Healthcare resource utilisation (including number of outpatient visits, number and length of hospitalisations, and number of ER visits) and direct medical costs were measured for each group, as captured in the HIRA database.Medical costs were divided into all-cause medical costs and pertussis-related medical costs.The latter included costs of inpatient or ER visits with a primary diagnosis of pertussis, outpatient visits with a primary or secondary diagnosis of pertussis, and antibiotics (Table 1, supplementary material) used for pertussis treatment.

Statistical analysis
Baseline clinical characteristics were assessed during the 12-months prior to pertussis diagnosis and summarised in frequency for categorical and mean for continuous variables; the median was reported for data with skewed distribution.To compare the characteristics between the groups, ANOVA (analysis of variance) test was used for continuous variables and chi-square test or Fisher's exact test (expected frequency <5) for binary variables.A p-value of <0.05 denoted statistical significance.
Cumulative annual incidence of pertussis per 100 000 persons was calculated according to calendar year.Annual incidence was then averaged to represent 10-year mean pertussis incidence, stratified by age groups and sex.Mean pertussis incidence per 100 000 persons over 10 years from the asthma-pertussis and COPD-pertussis groups were compared with the general (i.e.non-COPD/asthma)-pertussis group to compute incidence rate ratios (IRRs) with corresponding 95% confidence intervals (CI) based on Poisson distribution.
Healthcare resource utilisation and direct medical costs were measured 3 months pre-index date (baseline period); 0-3 months post-index date, >3-6 months post-index date, >6-9 months post-index date, and >9-12 months post-index date.Index date was defined as the earliest date of pertussis diagnosis (ICD-10: A37) minus 15 days to allow for workup of diagnosis [12] (Figure 1).To evaluate the excess burden of pertussis, incremental all-cause costs were calculated as the difference in healthcare costs for post-index and baseline time periods.These incremental all-cause costs were then compared for COPD-pertussis and asthma-pertussis groups versus the general pertussis group.Pertussis-related costs were calculated as the sum of pertussis-related costs for each time interval post-index and also compared across groups.
To account for differential characteristics across groups, analysis of healthcare resource utilisation and medical costs was conducted using a matched cohort study design.Here, pertussis patients in the COPD-pertussis and asthma-pertussis groups were matched with up to 3 pertussis patients without *up to 3 pertussis patients without pre-existing COPD or asthma were matched on sex, age (+/-10 years range), and CCi score excluding COPD or asthma (+/-1 range).‡ COPD and asthma were excluded from CCi score calculation.† Comorbidities associated with high direct and indirect medical costs among individuals ≥50 years in south Korea.§ index date was defined as the date of pertussis diagnosis minus 15 days; costs were adjusted to 2018 KrW and converted from south Korean Won (KrW) to us dollars (usD) using the 2018 exchange rate (1 usD = 1115.7 KrW).asthma-pertussis: pertussis cases with pre-existing asthma; CCi: Charlson comorbidity index; COPD: chronic obstructive pulmonary disease; COPD-pertussis: pertussis cases with pre-existing COPD; KrW: south Korean won; sD: standard deviation; usD: united states dollars.
Figure 1.study design for the evaluation of excess burden of pertussis.*to evaluate the excess burden of pertussis, incremental all-cause costs were calculated as the difference in healthcare costs between post-index and baseline time periods.these incremental all-cause costs were then compared between COPD-pertussis and asthma pertussis groups versus the general pertussis group, matching for differences in sex, age, and CCi score (excluding COPD or asthma).Pertussis-related costs were calculated as the sum of pertussis-related costs for each time post-index and also compared across groups.
COPD or asthma from the general-pertussis group based on sex, age, and Charlson Comorbidity Index (CCI) score excluding COPD or asthma (+/-1 score range).All health resource utilisation and direct medical costs were reported as mean with standard deviation (SD).Costs were adjusted to 2018 KRW (South Korean won) using the healthcare component of the consumer price index from the KOrean Statistical Information Service (KOSIS) [21] and converted from South Korean Won (KRW) to US dollars (USD) using the 2018 exchange rate (1 USD = 1115.7 KRW) [22].

Additional analyses
As the COPD-pertussis and asthma-pertussis groups were not mutually exclusive, pertussis cases could have both pre-existing COPD and asthma.As these patients may be different from a pathophysiological perspective to patients with either COPD or asthma only, we conducted an additional analysis of the incidence of pertussis, incremental healthcare resource utilisation and direct medical costs for this group using the methods described above.
In addition, a sensitivity analysis was conducted for the calculation of incremental healthcare resource utilisation and costs following pertussis diagnosis, where the workup for diagnosis period was extended to 30 and 60 days prior to the pertussis diagnosis.This was to account for other research which has reported the median delay for pertussis diagnosis to be longer than the 15 days used originally [23].

Study population
Among individuals ≥50 years in South Korea, 1011 were diagnosed with pertussis between 2009-2018.During the same period, 359 cases of pertussis among patients ≥50 years were reported to the Korea Disease Control and Prevention Agency (KDCA) [24].
Of the 1011 pertussis cases, 796 did not have a history of COPD or asthma (general-pertussis group), while 96 had pre-existing COPD (COPD-pertussis group) and 175 had pre-existing asthma (asthma-pertussis group; Figure 2).These groups included 56 patients with pre-existing COPD and asthma.
In the matched analysis, 91 cases in the COPD-pertussis group were matched with 269 cases in the general-pertussis group, and 169 cases in the asthma-pertussis group were matched with 503 cases in the general-pertussis group.After matching, baseline characteristics were comparable between each group (Table 1).

Pertussis incidence
Overall, the 10-year mean incidence of pertussis per 100 000 persons ≥50 years was 0.61.Mean annual incidence was highest among individuals ≥75 years (0.84) and mean incidence of pertussis in women was higher than in men (0.70 versus 0.51).Annual incidence of pertussis also increased over time from 0.36 in 2009 to 1.48 in 2018 (Table 2, supplementary material).
Mean annual incidence of pertussis per 100,000 persons was highest among individuals with pre-existing asthma (3.35), followed by individuals with COPD (2.47) versus those with neither condition (0.50; Tables S3-S5, supplementary material).Similar trends of increasing annual pertussis incidence over the study period were seen for individuals with COPD (0.84-1.13) and asthma (1.58-11.20)peaking in 2017 for the COPD-pertussis group and 2018 for the asthma-pertussis group.
Overall mean IRR (95% CI) of pertussis relative to individuals without COPD or asthma was 6.70 (5.69-7.89)for individuals with pre-existing asthma and 4.94 (4.00-6.11)for individuals with pre-existing COPD (Table 2).In both individuals with COPD and asthma, IRRs of pertussis were higher in men compared with women.

Healthcare resource utilisation
At 3-month post-index date, the mean length of all-cause hospital stays increased by 8.1 days for the COPD-pertussis group versus 5.4 days for the matched group compared to the pre-index date.However, the mean number of outpatient visits only showed a marginal increase of 0.8 for the COPD-pertussis group compared with 1.5 for the matched group.Similar patterns were observed for the asthma-pertussis group (Tables S6 and S7, supplementary material).

Medical costs
Individuals with pre-existing COPD and asthma still had higher incremental costs versus the matched patients without COPD or asthma (Figure 3).This was mainly driven by increases in the all-cause incremental cost for ambulatory *56 patients had both pre-existing COPD and asthma.hira: health insurance review and assessment service; COPD: chronic obstructive pulmonary disease.COPD-pertussis and asthma-pertussis groups are not mutually exclusive.care and hospitalisation.For example, at the first 3-month post-index date (when most cost was incurred), the mean incremental cost of ambulatory care and hospitalisation for the COPD-pertussis and asthma-pertussis groups was 2 421 USD and 1594 USD, compared with 1686 USD and 1153 USD for matched general-pertussis groups, respectively (Table 7, supplementary material).
No obvious difference in pertussis-related healthcare costs between pertussis patients with and without pre-existing COPD or asthma was observed after matching (Figure 4, Table 8, supplementary material).

Additional analyses
Incidence and IRR of pertussis compared to the general population for the patients with both pre-existing COPD and pertussis are presented in Table 9 (supplementary material).Overall 10-year mean incidence was 3.82 per 100,000 persons, and IRR (95% CI) of pertussis relative to those without COPD or asthma was 7.64 (5.83-10.02).After matching, patterns in healthcare resource utilisation and medical costs among the COPD and asthma-pertussis group resembled those of the COPD-pertussis and asthma-pertussis groups.This included longer hospitalisation stays, and increased all-cause costs of both ambulatory care and hospitalisation within the first 3 months following pertussis diagnosis (Table 10, supplementary material).Furthermore, no obvious difference in pertussis-related healthcare costs among patients with COPD and asthma versus those without were observed (Table 11, supplementary material).
Similarly, the sensitivity analysis conducted where the workup period for pertussis diagnosis was extended to 30 and n: number of pertussis cases.incidence of pertussis was calculated per 100 000 persons.95% Cis for irrs were based on Poisson distribution.Ci: confidence interval; COPD: chronic obstructive pulmonary disease; irr: incidence rate ratio.
Figure 3. mean incremental all-cause direct medical costs for the asthma-and COPD-pertussis versus the general-pertussis group.incremental cost and healthcare resource utilisation was calculated as differences between each post-index period and baseline period (0-3 months pre-index date) for each group.Costs were adjusted to 2018 KrW using the healthcare component of the Korean Consumer Price index and converted from south Korean Won (KrW) to us dollars (usD) using the 2018 exchange rate (1 usD = 1115.7 KrW). the index date was defined as the date of pertussis diagnosis minus 15 days to allow for the inclusion of healthcare resource utilisation and medical costs related to workup pertussis diagnosis.asthma-pertussis: pertussis cases with pre-existing asthma; COPD: chronic obstructive pulmonary disease; COPD-pertussis: pertussis cases with pre-existing COPD; General-pertussis: pertussis cases without pre-existing COPD or asthma; KrW: south Korean won; usD: united states dollars.
60 days prior to pertussis diagnosis also had incremental healthcare resource utilisation and cost results consistent to that of the main analysis which had a workup period of 15 days prior to pertussis diagnosis.With increased all-cause costs within the first 3 months for the COPD-pertussis and asthma-pertussis groups following index date and no obvious difference in pertussis-related healthcare costs between COPD-pertussis or asthma-pertussis groups and their matched general-pertussis groups (Tables S12-S17, supplementary material).

Discussion
This retrospective database analysis was conducted to estimate the incidence and healthcare burden of diagnosed pertussis cases among individuals ≥50 years, with or without pre-existing COPD or asthma.The number of pertussis cases identified in this study more than doubled that reported to the KDCA from the South Korean population over the same time period [24].This difference may in part be due to the stricter pertussis case definition used by the KDCA, which only includes cases confirmed by diagnostic testing, whereas cases captured in the HIRA database are based on medical diagnosis codes.The most commonly used diagnostic testing method for pertussis used in South Korea is PCR.However, this has found to yield lower rates of confirmation when used alone, potentially as the sensitivity of PCR diagnosis is high during early onset of disease but declines considerably after the first two weeks [25].Serological testing has proven useful for the diagnosis of pertussis infection after 3-4 weeks; however, these are not available for routine clinical practice in South Korea.Furthermore, adult patients presenting with severe cough are also less likely to go attend primary care clinics, and due to the atypical presentation of symptoms in this population, it is possible that pertussis is not suspected and hence diagnostic tests are less likely to be performed [26].Nevertheless, given the substantial difference between these two figures, underreporting of pertussis to the KDCA cannot be ruled out.
The overall 10-year mean annual incidence of pertussis among adults aged ≥50 years in South Korea was estimated to be 0.61 per 100 000 persons, relatively lower than that reported for adults in other countries [27,28].However, as only diagnosed cases were included, the true incidence of pertussis may have been underestimated [25].Consistent with the shift in pertussis cases to older adults reported in several countries, this study found that annual incidence of pertussis among adults aged ≥50 increased from 2009 to 2018.Similarly, the KDCA has reported that adults ≥50 years accounted for an increasing proportion of diagnosed pertussis cases (3.7% of in 2010 [1/27] and 38.2% [47/123] in 2020) [24].
Annual incidence of pertussis was higher in individuals with pre-existing asthma and/or COPD, compared with those without underlying respiratory conditions, consistent with findings in the US and Europe [12,13].Healthcare burden generally increased for all groups in the 3 months following pertussis diagnosis, especially for individuals with pre-existing COPD and asthma, consistent with findings in the US [12].This increase was mainly driven by increased medical costs for hospitalisation and ambulatory care, .mean pertussis-related medical costs among asthma and COPD-pertussis groups relative to the general pertussis group.incremental pertussis-related healthcare costs were calculated within the 3 months post index date.Costs were adjusted to 2018 KrW using the healthcare component of the Korean Consumer Price index and converted from south Korean Won (KrW) to us dollars (usD) using the 2018 exchange rate (1 usD = 1115.7 KrW). the index date was defined as the date of pertussis diagnosis minus 15 days to allow for the inclusion of healthcare resource utilisation and medical costs related to workup pertussis diagnosis.asthma-pertussis: pertussis cases with pre-existing asthma; COPD-pertussis: pertussis cases with pre-existing COPD; COPD: chronic obstructive pulmonary disease; General-pertussis: pertussis cases without pre-existing COPD or asthma; KrW: south Korean won; usD: united states dollars.
suggesting the need for more intensive medical care following pertussis diagnosis.
Pertussis-related medical costs were similar between these individuals and their matched cases, indicating that the excess costs were likely to be related to increased medical costs for concomitant conditions.This trend was consistent even in an additional analysis which looked at patients with both pre-existing COPD and asthma.However, as pertussis may have aggravated concomitant conditions, it is difficult to make a clear distinction between pertussis-related and pertussis-unrelated costs.
Following the first 3 months after diagnosis, incremental costs and healthcare resource utilisation decreased across groups.This may be due to a relative increase in healthcare resource utilisation prior in the baseline period.While a 15-day workup period was incorporated to account for costs leading up to pertussis diagnosis, the pre-index period is not necessarily a pertussis-free state, with studies of pertussis burden demonstrating a clear increase in healthcare resource utilisation in the month before diagnosis [13].However, a sensitivity analysis which extended the workup period to 30 and 60 days found similar patterns of healthcare resource utilisation.
Notably, mean number of outpatient visits was relatively high across groups and time periods.This is in line with the high number of general outpatient visits in South Korea reported by the Organisation for Economic Cooperation and Development (OECD), which found that the number of consultations per-year in Korea was almost three times higher than average [29].The high number of outpatient visits may, in part, be due to the conveniently accessible and reimbursed healthcare available in South Korea.
This study represents the first of its kind in South Korea to investigate the incidence and healthcare burden of pertussis in adults, and the potential effects of pre-existing respiratory disease.A robust methodology was used as well as a comprehensive dataset, courtesy of the HIRA database which is internationally recognised and representative of the entire South Korean population [19].Furthermore, consideration of bias was accounted for by employing a matched cohort design.Additional analyses which looked specifically at patients with both COPD and asthma, as well as which extended the workup period, both found similar patterns to the main analysis further supporting these results.
However, when interpreting the increased risk of pertussis diagnosis among patients with pre-existing respiratory conditions, a potential for ascertainment bias should also be accounted for (i.e. as these patients are more likely to frequently follow-up with healthcare providers, they are more likely to be diagnosed with pertussis).
Pertussis may also be initially misdiagnosed as asthma or acute exacerbations of COPD, influencing the subsequent annual incidence of pertussis among these groups.To combat this, stricter definitions of COPD and asthma were used and this also aligned with previous studies using the HIRA database [30,31].
It should also be noted that both COPD and asthma have been reported to be under-diagnosed in South Korea [18,32].Given this, pertussis patients who may have in fact had pre-existing COPD or asthma may have been misclassified as having neither.This could have led to underestimation of the incidence of pertussis in the COPD and asthma groups and overestimation of the incidence among the general group.
As for healthcare resource utilisation, although a matched cohort design was utilised, other factors not matched for or included in the database (e.g.smoking status) might have influenced comparability between groups.A substantial number of pertussis patients also had both COPD and asthma which may have influenced medical resource utilisation.Lastly, it should be noted that the HIRA database does not include information on non-reimbursed procedures, medications, and over-the-counter medicines.However, it is expected that patients experiencing respiratory problems are more likely to visit the hospital.

Conclusions
Incidence of pertussis among individuals ≥50 years in South Korea has been increasing in the last decade.Overall, individuals with pre-existing COPD or asthma were at an increased risk of being diagnosed with pertussis and had higher healthcare burden following pertussis diagnosis (Figure 5, plain language summary).However, given the likely underdiagnosis of pertussis in South Korea, additional research is needed to better characterise the burden of pertussis in older adults and in order to inform public health policy makers on formulating effective public health strategies for pertussis prevention.

Figure 2 .
Figure 2. study flowchart identifying study cohort.*56patients had both pre-existing COPD and asthma.hira: health insurance review and assessment service; COPD: chronic obstructive pulmonary disease.COPD-pertussis and asthma-pertussis groups are not mutually exclusive.

Figure 4
Figure 4. mean pertussis-related medical costs among asthma and COPD-pertussis groups relative to the general pertussis group.incremental pertussis-related healthcare costs were calculated within the 3 months post index date.Costs were adjusted to 2018 KrW using the healthcare component of the Korean Consumer Price index and converted from south Korean Won (KrW) to us dollars (usD) using the 2018 exchange rate (1 usD = 1115.7 KrW). the index date was defined as the date of pertussis diagnosis minus 15 days to allow for the inclusion of healthcare resource utilisation and medical costs related to workup pertussis diagnosis.asthma-pertussis: pertussis cases with pre-existing asthma; COPD-pertussis: pertussis cases with pre-existing COPD; COPD: chronic obstructive pulmonary disease; General-pertussis: pertussis cases without pre-existing COPD or asthma; KrW: south Korean won; usD: united states dollars.

Table 1 .
baseline characteristics of the pertussis cases and pre-pertussis diagnosis medical costs.

Table 2 .
Pertussis incidence among individuals with pre-existing COPD and asthma compared with those without (2009-2018).