Health and economic burden associated with 15-valent pneumococcal conjugate vaccine serotypes in children in the United States

Abstract

Aims

V114 is an investigational 15-valent pneumococcal conjugate vaccine (PCV) containing the 13 Streptococcus pneumoniae serotypes in 13-valent PCV (PCV13) plus two additional serotypes. This study quantified the health and economic burden of invasive pneumococcal disease (IPD) and acute otitis media (AOM) caused by V114 types among children in the United States.

Materials and methods

A Markov model estimated the number of V114-type IPD and AOM cases and costs in a hypothetical, unvaccinated US birth cohort over 20 years. Three time periods were analyzed using time-specific epidemiological data to determine the number of IPD and AOM cases associated with all 15 serotypes in V114. The time periods were: (1) pre-PCV7 (1999); (2) pre-PCV13 (2009); (3) post-PCV13 (2017). Costs were estimated from a societal perspective (2018 US dollars) and discounted at 3%.

Results

The model estimated 18,983 IPD cases and 5.4 million AOM cases associated with V114 serotypes pre-PCV7, 4,697 IPD cases and 3.0 million AOM cases pre-PCV13, and 948 IPD cases and 0.2 million AOM cases post-PCV13. Total discounted costs associated with V114 serotypes were $1.7 billion pre-PCV7, $730 million pre-PCV13, and $75 million US dollars post-PCV13.

Limitations

Post-meningitis sequelae, cases of non-bacteremic pneumonia, and direct non-medical costs were not included.

Conclusions

IPD and AOM cases and costs were estimated in a hypothetical US birth cohort followed for 20 years at three time periods. In all three periods, the serotypes targeted by V114 contributed to significant morbidity and costs. New pediatric pneumococcal vaccines must continue to retain serotypes in licensed vaccines to maintain disease reduction while extending coverage to non-vaccine serotypes.

Keywords:

• Invasive pneumococcal disease
• acute otitis media
• pneumococcal conjugate vaccine
• Markov model
• health and economic burden

JEL CLASSIFICATION CODES:

• I10
• I19

Introduction

Streptococcus pneumoniae is a Gram-positive bacterium that causes both invasive and non-invasive disease manifestations¹. Invasive pneumococcal disease (IPD) manifestations include bacteremic pneumonia, bacteremia without focus, and meningitis¹. In the United States (US), the annual incidence of IPD in 2017 was seven cases per 100,000 children aged <5 years². IPD is associated with substantial disease burden, including high case fatality rates (CFRs), and risk of developing meningitis and post-meningitis sequelae (PMS), such as hearing loss or developmental delay^1,3.

Non-invasive manifestations of pneumococcal disease include non-bacteremic pneumococcal pneumonia (NBPP) and acute otitis media (AOM)¹. Although difficult to diagnose⁴, the incidence and burden of NBPP are understood to be considerably higher than those of IPD^5,6. AOM also has a higher incidence rate (358.7 cases per 1,000 in US children aged ≤2 years in 2016)⁷ and is one of the most common causes of medical visits, antibiotic prescriptions, and hearing impairment in children^8,9. AOM is also associated with substantial medical costs⁸.

Pneumococcal diseases also lead to considerable indirect costs from productivity losses due to premature disability and death in children ($3.1 billion in 2004 [2007 US dollars]), and work loss in parents caring for children with pneumococcal disease ($914 million in 2004 [2007 US dollars])⁵.

The introduction of pneumococcal conjugate vaccines (PCVs) has been effective in reducing the burden of pneumococcal diseases. PCV7, which targets seven S. pneumoniae serotypes, was introduced into infant vaccination schedules in 2000¹⁰. Its use has led to a substantial reduction in IPD and AOM cases in the US^11,12. However, the subsequent increase in disease caused by non-PCV7 serotypes necessitated the development of new vaccines containing additional serotypes to maintain disease reduction¹³. In 2010, PCV13, which targets 13 S. pneumoniae serotypes, was licensed and recommended for use in infants in the US¹⁴. Following PCV13 introduction, the disease caused by non-PCV13 serotypes has increased, while select PCV13 vaccine serotypes, such as serotype 3, continue to persist^15–18. New vaccines must therefore continue to retain original PCV serotypes to maintain disease reduction while extending coverage to emerging disease-causing serotypes.

A 15-valent PCV, V114, is currently in clinical development^19,20. V114 covers all serotypes included in PCV13 (serotypes 1, 3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F, and 23F), as well as two additional serotypes, 22F and 33F. Serotypes 22F and 33F are two of the most common non-PCV13 serotypes found in IPD in children¹⁵, causing 8% and 9% of residual IPD cases in children aged <5 years in the US, respectively, during 2017²¹. Serotype 33 F has high invasive disease potential compared with other non-PCV13 serotypes and is associated with multidrug resistance^22,23.

The objective of this study was to quantify the health and economic burden of IPD and AOM attributable to the 15 serotypes targeted by V114 in a hypothetical, unvaccinated birth cohort over 20 years in the US. To highlight the importance of maintaining disease reduction while addressing the current disease burden, IPD and AOM cases and costs attributable to V114 serotypes were calculated at three time periods using time-specific epidemiological data to determine the number of IPD and AOM cases associated with all 15 serotypes in V114. The time periods were: (1) pre-PCV7 (1999), (2) pre-PCV13 (2009), and (3) post-PCV13 (2017 or the most current year).

Methods

Analytical approach

This analysis was conducted from a societal perspective using a probabilistic Markov state transition model, which simulated clinical events and costs over a 20-year time horizon.

Model structure

As illustrated in Figure 1, the model has four health states: no pneumococcal disease, IPD (meningitis or bacteremia, including bacteremic pneumonia), AOM, and death. An unvaccinated birth cohort enters the model and is at risk of developing IPD or AOM. The probability of an infant developing pneumococcal disease varies by age over the time horizon of the model. The model assumes that pneumococcal disease events are mutually exclusive; however, infants can experience both IPD and AOM during each model cycle. Following IPD or AOM episodes, infants can recover and return to the no pneumococcal disease health state. Only infants with IPD face a risk of death (depending on the specific manifestation of IPD and age). The model tracks the birth cohort up to 20 years of age or until death, whichever occurs earlier, with a cycle length of 1 year.

Figure 1. Markov pathway. Abbreviations. AOM, acute otitis media; IPD, invasive pneumococcal disease.

The societal perspective was used in this analysis. Therefore, costs include both direct medical costs and indirect costs from productivity losses in children and parents. The model did not include transmission dynamics of pneumococcal infections from person to person, nor were PMS or NBPP considered.

Model inputs

Target population

The target population of the study was a 2018 cohort of unvaccinated newborns in the US. The cohort size of 3.8 million newborns was based on the overall number of births in the US in 2018 according to the US Centers for Disease Control and Prevention (CDC) National Vital Statistics report²⁴.

Epidemiologic inputs

Epidemiological data obtained from published literature and publicly available reports were used as inputs (Table 1). Age-specific IPD incidence rates and serotype distributions were taken from CDC Active Bacterial Core (ABC) surveillance data^2,11,25,26. Meningitis and bacteremia (including bacterial pneumonia) comprised 15% and 85% of IPD cases, respectively, in the pre-PCV7 era, 9% and 91% in the pre-PCV13 era, and 13% and 87% in the post-PCV13 era^27,28. CFRs were 10% for meningitis²⁹ and 3% for bacteremia²⁸; these were applied across all periods to reflect current access to medical care and treatment options.

Table 1. Epidemiological inputs^†.

	Age group^‡ (years)	Pre-PCV7 period			Pre-PCV13 period			Post-PCV13 period
		PCV7 serotypes	PCV13 not PCV7 serotypes	V114 not PCV13 serotypes	PCV7 serotypes	PCV13 not PCV7 serotypes	V114 not PCV13 serotypes	PCV7 serotypes	PCV13 not PCV7 serotypes	V114 not PCV13 serotypes
IPD incidence rate (100,000 person-years)^39	<1	138.9	6.9	2.4	0.4	25.7	4.5	0.5	2.0	1.4
	1	175.4	8.7	3.1	0.3	22.2	3.9	0.3	1.4	1.0
	2–4	25.8	2.6	0.5	0.3	9.6	1.2	0.5	1.1	0.7
	5–20	2.5	0.9	0.1	0.1	2.3	0.2	0.2	0.3	0.1
Pneumococcal AOM incidence rate (100,000 person-years)^12^,^16^,^31^,^38^†	<2	24,381.7	5,257.3	304.8	562.1	12,647.7		81.3	975.2	325.1
	2–4	7,822.2	1,686.7	97.8	198.3	4,462.7		32.1	385.6	128.5
	5–20	2,538.5	547.4	31.7	74.3	1,670.9		11.6	138.6	46.2

† The IPD serotype-specific incidence rate was calculated by multiplying incidence with serotype distribution data in the reference in pre-PCV7, pre-PCV13, and post-PCV periods. The pneumococcal AOM serotype-specific incidence was calculated by multiplying the proportion of Streptococcus pneumoniae among AOM incidence and serotype distribution data in the reference in pre-PCV7, pre-PCV13, and post-PCV periods. ^‡Age groups are different for IPD and AOM due to different methodologies used in the literature.

PCV7 serotypes are 4, 6B, 9V, 14, 18C, 19F, and 23F; PCV13 not PCV7 serotypes are 1, 3, 5, 6A, 7F, and 19A; V114 not PCV13 serotypes are 22F and 33F.

Abbreviations. AOM, acute otitis media; IPD, invasive pneumococcal disease; PCV, pneumococcal conjugate vaccine.

Age-specific AOM incidence rates were obtained from a retrospective study of AOM-associated ambulatory visits¹², and serotype distribution was taken from published surveillance and observational studies^16,30. The proportion of AOM cases associated with S. pneumoniae was 44% in the pre-PCV7 era, 47% in the pre-PCV13 era, and 21% in the post-PCV13 era³¹.

Cost inputs

Total costs per episode of IPD or AOM comprised direct medical costs and indirect costs (Table 2). Costs were expressed in 2018 US dollars and discounted at 3%. Direct costs were retrieved from a published report on the health and economic impact of PCV13³². Indirect costs included both productivity losses in children associated with premature death from IPD, and work loss among caregivers of children with IPD or AOM, both of which were calculated using the human capital approach³³. Lost lifetime income due to premature death was calculated using age-specific life expectancy values obtained from life tables and age-specific average income per year. Work loss in parents caring for children with pneumococcal disease was calculated using the number of work days missed multiplied by the daily wage rate and labor force participation rate in the US in 2016³⁴. The numbers of work days missed were 7 days per episode for bacteremia³⁵, 11 days for meningitis²⁹, and 1 day for AOM. The annual income of individuals under 20 years of age was assumed to be zero.

Table 2. Cost inputs (2018 US dollars).

	Medical cost per episode^†	Indirect cost per episode‡
Meningitis	$22,703^32	$1,170^29
Bacteremia	$4,332^32	$745^35
AOM	$135^32	$106^47

^†Costs from Rubin et al. 2010 were inflated to 2018 dollars using the consumer price index for medical care (https://fred.stlouisfed.org/series/CPIMEDSL). ‡The indirect cost was calculated by multiplying the episode length from the source reference indicated, with the daily lost wage (calculated by multiplying $907 [2017 August Weekly earnings from https://www.bls.gov/news.release/archives/empsit_09012017.pdf] by 80.5% [men and women labor force participation] divided by 7).

Abbreviation. AOM, acute otitis media.

Model verification and validation

The validity of the model was checked during collaboration with experts and by comparing its structure with that of a previously published model³². Several tests were built into the model for verification and to ensure internal validity³⁶. For example, the sum of the distribution of persons in each compartment at each point of time and age was verified analytically to be equal to the size of the population at that age.

Model outputs

Modeled outcomes included the number of IPD cases and deaths, number of AOM cases, and total, direct, and indirect costs attributable to IPD and AOM. Cases, deaths, and costs were estimated overall and by vaccine-type serotype.

Sensitivity analysis

One-way sensitivity analysis, conducted by varying the incidence, CFR, and cost data by ±20%, and varying the discount rate to 0% and 5%, was performed to assess the impact of uncertainties of the key model parameters and assumptions on model results in the pre-PCV7 period.

Results

Clinical events

Invasive pneumococcal disease cases by serotype

The model estimated that 18,983 IPD cases in the pre-PCV7 period, 4,697 IPD cases in the pre-PCV13 period, and 948 IPD cases in the post-PCV13 period were attributable to V114 serotypes. The majority of IPD cases (16,321 cases; 86%) were attributable to PCV7 serotypes in the pre-PCV7 period (Table 3).

Table 3. IPD cases attributable to V114 serotypes in the pre-PCV7, pre-PCV13, and post-PCV13 periods.

Cases, n (%)	Pre-PCV7 period	Pre-PCV13 period	Post-PCV13 period
PCV7 serotypes	16,321 (86)	147 (3)	224 (24)
ST18C	1,863 (10)	41 (1)	11 (1)
ST4	1,421 (7)	9 (0)
ST6B	1,960 (10)
ST9V	1,343 (7)	19 (0)
ST14	6,326 (33)	27 (1)
ST19F	2,156 (11)	42 (1)	214 (23)
ST23F	1,252 (7)	9 (0)
PCV13, but not PCV7 serotypes	2,338 (12)	4,009 (85)	427 (45)
ST3	147 (1)	226 (5)	202 (21)
ST19A	514 (3)	2,081 (44)	214 (23)
ST6A	1,056 (6)	27 (1)
ST1	358 (2)	243 (5)
ST5	17 (0)	32 (1)
ST7F	246 (1)	1,400 (30)	11 (1)
V114, but not PCV13 serotypes	324 (2)	545 (12)	298 (31)
ST22F	185 (1)	229 (5)	181 (19)
ST33F	139 (1)	315 (7)	116 (12)
Total V114 serotypes	18,983 (100)	4,697 (100)	948 (100)

Abbreviations. IPD, invasive pneumococcal disease; PCV, pneumococcal conjugate vaccine; ST, serotype.

IPD cases associated with the six additional serotypes in PCV13 increased from 2,338 cases in the pre-PCV7 period to 4,009 cases in the pre-PCV13 period. The increase was primarily due to increases in serotype 19A from 514 (3%) to 2,081 (44%) cases, 7F from 246 (1%) to 1,400 (30%) cases, and serotype 3 from 147 (1%) to 226 (5%) cases. Serotype 6A decreased significantly between the pre-PCV7 and pre-PCV13 scenarios, from 1,056 (6%) to 27 cases (1%). Overall, there was a net increase of 1,671 IPD cases due to the six additional serotypes in PCV13 in the pre-PCV13 period compared with the pre-PCV7 period.

The number of cases associated with serotypes 22F and 33F combined was 324 (2%) in the pre-PCV7 period, 545 (12%) in the pre-PCV13 period, and 298 (31%) in the post-PCV13 period.

Mortality due to invasive pneumococcal disease

The number of estimated deaths associated with V114 serotypes was 769 in the pre-PCV7 period, 172 in the pre-PCV13 period, and 37 in the post-PCV13 period (Table 4). Most of these deaths (657; 85%) were attributable to PCV7 serotypes in the pre-PCV7 period. The numbers of deaths attributable to the six additional serotypes in PCV13 were 147 (85%) in the pre-PCV13 period and 17 (45%) in the post-PCV13 period.

Table 4. IPD deaths attributable to V114 serotypes in the pre-PCV7, pre-PCV13, and post-PCV13 periods.

Deaths, n (%)	Pre-PCV7 period			Pre-PCV13 period			Post-PCV13 period
	Bacteremia	Meningitis	Total	Bacteremia	Meningitis	Total	Bacteremia	Meningitis	Total
PCV7 serotypes	418 (87)	240 (83)	657 (85)	4 (3)	1 (3)	5 (3)	6 (24)	3 (24)	9 (24)
PCV13, but not PCV7 serotypes	60 (12)	34 (12)	94 (12)	109 (85)	38 (85)	147 (85)	5 (45)	3 (45)	17 (45)
ST5	0 (0)	0 (0)	1 (0)	1 (1)	0 (1)	1 (1)	0 (0)	0 (3)	0 (0)
ST3	4 (1)	2 (1)	6 (1)	6 (5)	2 (5)	8 (5)	6 (21)	3 (21)	8 (21)
ST7F	6 (1)	4 (1)	10 (1)	38 (30)	13 (30)	51 (30)	0 (0)	0 (1)	0 (0)
ST1	9 (2)	5 (2)	14 (2)	7 (5)	2 (5)	9 (5)	0 (0)	0 (0)	0 (0)
ST19A	13 (3)	8 (3)	21 (3)	57 (44)	20 (44)	76 (44)	8 (23)	3 (23)	6 (23)
ST6A	27 (6)	16 (5)	43 (6)	1 (1)	0 (1)	1 (1)	0 (0)	0 (0)	0 (0)
V114, but not PCV13 serotypes (ST22F and ST33F)	5 (1)	13 (5)	18 (2)	15 (12)	5 (12)	20 (12)	8 (31)	4 (31)	12 (31)
Total V114 serotypes	482 (100)	287 (100)	769 (100)	128 (100)	44 (100)	172 (100)	25 (100)	12 (100)	37 (100)

PCV7 serotypes are 4, 6B, 9V, 14, 18C, 19F, and 23F.

Abbreviations. IPD, invasive pneumococcal disease; PCV, pneumococcal conjugate vaccine; ST, serotype.

Acute otitis media cases by serotype

V114 serotypes were associated with 5.4 million AOM cases in the pre-PCV7 period, 3.0 million AOM cases in the pre-PCV13 period, and 217,461 AOM cases in the post-PCV13 period (Table 5). The majority of AOM cases (4.6 million; 85%) were attributable to PCV7 serotypes in the pre-PCV7 period.

Table 5. AOM cases attributable to V114 serotypes in the pre-PCV7, pre-PCV13, and post-PCV13 periods.

Cases, n (%)	Pre-PCV7 period	Pre-PCV13 period	Post-PCV13 period
PCV7 serotypes	4,572,053 (85)	109,613 (4)	16,728 (8)
ST18C
ST4
ST6B	844,392 (16)
ST9V	350,115 (7)
ST14	1,132,697 (21)
ST19F	1,441,644 (27)	109,613 (4)	16,728 (8)
ST23F	803,205 (15)
PCV13, but not PCV7 serotypes	782,613 (15)	2,849,930 (96)	133,822 (62)
ST3	164,761 (3)	274,034 (9)	66,911 (31)
ST19A	226,546 (4)	2,192,250 (74)	66,911 (31)
ST6A	391,306 (7)	383,647 (13)
ST1
ST5
ST7F
V114, but not PCV13 serotypes			66,911 (31)
ST22F
ST33F			66,911 (31)
Total V114 serotypes	5,354,666 (100)	2,959,544 (100)	217,461 (100)

Abbreviations. AOM, acute otitis media; PCV, pneumococcal conjugate vaccine; ST, serotype.

The number of AOM cases associated with the six additional serotypes in PCV13 increased from 0.8 million (15%) in the pre-PCV7 period to 2.8 million (96%) in the pre-PCV13 period. The increase was primarily due to an increase in serotype 19A from 226,546 (4%) to 2,192,250 (74%) cases, and serotype 3 from 164,761 (3%) to 274,034 (9%) cases. Serotype 6A decreased between the pre- and post-PCV periods, from 391,306 (7%) cases to 383,647 cases (13%). Overall, there was a net increase of 2.1 million AOM cases due to these six serotypes in the pre-PCV13 period compared with the pre-PCV7 period.

The proportion of cases attributable to serotypes 22F and 33F together was 0 cases (0%) in the pre-PCV7 period, 0 cases in the pre-PCV13 period, and 66,911 cases (31%) in the post-PCV13 period.

Economic impact by serotype

Total discounted costs (direct medical costs and indirect costs; US dollars) of IPD and AOM for all V114 serotypes were estimated to be $1.7 billion in the pre-PCV7 period, $730 million in the pre-PCV13 period, and $75 million in the post-PCV13 period (Table 6). IPD and AOM associated with PCV7 serotypes accounted for the majority of costs in the pre-PCV7 period ($1.4 billion; 86%).

Table 6. Discounted direct and indirect costs due to IPD and AOM in the pre-PCV7, pre-PCV13, and post-PCV13 periods (2018 US dollars).

$ in millions (%)	Pre-PCV7 period			Pre-PCV13 period			Post-PCV13 period
	Direct costs	Indirect costs	Total costs	Direct costs	Indirect costs	Total costs	Direct costs	Indirect costs	Total costs
PCV7 serotypes	647.2 (86)	766.1 (86)	1,413.3 (86)	13.4 (4)	12.9 (4)	26.3 (4)	3.2 (10)	6.4 (15)	9.6 (13)
PCV13, but not PCV7 serotypes	106.8 (14)	121.7 (141)	228.5 (14)	351.3 (96)	338.9 (94)	690.2 (95)	17.8 (58)	23.9 (54)	41.8 (56)
ST5	0.1 (0)	0.4 (0)	0.5 (0)	0.2 (0)	0.6 (0)	0.8 (0)	0.0 (0)	0.0 (0)	0.0 (0)
ST3	20.3 (3)	18.5 (2)	38.8 (2)	32.9 (9)	29.6 (8)	62.5 (9)	8.9 (29)	11.8 (27)	20.6 (28)
ST7F	1.5 (0)	5.1 (1)	6.5 (0)	7.0 (2)	26.1 (7)	33.1 (5)	0.1 (0)	0.2 (0)	0.3 (0)
ST1	2.0 (0)	7.3 (1)	9.3 (1)	1.1 (0)	4.5 (1)	5.7 (1)	0.0 (0)	0.0 (0)	0.0 (0)
ST19A	29.9 (4)	31.9 (4)	61.8 (4)	265.5 (72)	242.0 (67)	507.5 (70)	8.9 (29)	12.0 (27)	20.9 (28)
ST6A	53.1 (7)	58.6 (7)	111.8 (7)	44.5 (12)	36.0 (10)	80.6 (11)	0.0 (0)	0.0 (0)	0.0 (0)
V114, but not PCV13 serotypes (22 F and 33 F)	2.1 (0)	6.7 (1)	8.8 (1)	3.1 (1)	10.2 (3)	13.3 (2)	9.5 (31)	13.7 (31)	23.2 (31)
Total V114 serotypes	756.1 (100)	894.5 (100)	1,650.6 (100)	367.7 (100)	362.0 (100)	729.8 (100)	30.5 (100)	44.0 (100)	74.5 (100)

PCV7 serotypes are 4, 6B, 9V, 14, 18C, 19F, and 23F.

Abbreviations. AOM, acute otitis media; IPD, invasive pneumococcal disease; PCV, pneumococcal conjugate vaccine; ST, serotype; US, United States.

The six additional serotypes in PCV13, but not in PCV7, increased from $229 million (14%) in the pre-PCV7 period to $690 million (95%) in the pre-PCV13 period. This increase was primarily due to increases in serotype 19A from $62 million (4%) to $508 million (70%), serotype 3 from $39 million (2%) to $63 million (9%), and serotype 7F from $7 million (0.4%) to $33 million (5%). Overall, there was a net increase of $462 million due to these six serotypes in the pre-PCV13 period compared with the pre-PCV7 period.

Total costs associated with serotype 22F and 33F together were $9 million (1%) in the pre-PCV7 period, $13 million (2%) in the pre-PCV13 period, and $23 million (31%) in the post-PCV13 period.

Sensitivity analysis

The one-way sensitivity analysis indicated that the discounted total cost was sensitive to uncertainties around all key parameters. It was most affected by the discount rate, whereby total IPD and AOM costs increased by 64–189% and decreased by 18–40% for 0% and 5% discount rates, respectively (Table 7). When the incidence of IPD and AOM varied by 20%, total costs increased and decreased by 20% accordingly. When the CFR for meningitis and bacteremia varied by 20%, the total cost varied by 4–15%. When the direct medical cost of an IPD and AOM episode varied by 20%, total costs varied by 5–16%.

Table 7. One-way sensitivity analyses results (2018 US dollars).

$ in millions (% change from base cost)	Discounted total costs
	PCV7 serotypes	PCV13, but not PCV7 serotypes	V114, but not PCV13 serotypes
Base case	1,413	229	9
Incidence −20%	1,131 (−20%)	183 (−20%)	7 (−20%)
Incidence +20%	1,696 (20%)	274 (20%)	11 (20%)
CFR −20%	1,349 (−5%)	219 (−4%)	7 (−15%)
CFR +20%	1,478 (5%)	238 (4%)	10 (15%)
Cost −20%	1,195 (−15%)	192 (−16%)	8 (−5%)
Cost +20%	1,631 (15%)	265 (16%)	9 (5%)
Discount factor 0%	2,393 (69%)	374 (64%)	25 (189%)
Discount factor 5%	1,166 (−18%)	191 (−17%)	5 (−40%)

PCV7 serotypes are 4, 6B, 9V, 14, 18C, 19F, and 23F; PCV13, but not PCV7 serotypes are 1, 3, 5, 6A, 7F, and 19A; V114, but not PCV13 serotypes are 22F and 33F.

Abbreviations. CFR, case fatality rate; PCV, pneumococcal conjugate vaccine; US, United States.

Discussion

In this modeling analysis, the 15 pneumococcal serotypes included in V114 contributed to a substantial number of IPD and AOM cases and deaths during the pre-PCV7, pre-PCV13, and post-PCV13 periods in a hypothetical unvaccinated US birth cohort over 20 years. V114 serotypes were associated with high total discounted costs during all three time periods. Although the overall number of cases and costs reduced over time, the proportion attributable to the six additional serotypes in PCV13 increased in the pre-PCV13 period compared with the pre-PCV7 period, primarily driven by serotypes 19A, 7F, and 3. Similarly, the proportion of cases and costs attributed to the two unique serotypes in V114 increased in the pre-PCV13 and post-PCV13 periods compared with the pre-PCV7 period.

Our findings are consistent with the literature. Prior to the introduction of PCVs, the majority of IPD and AOM cases were caused by serotypes in PCV7^27,37,38. Based on US ABC surveillance reports, PCV7 serotypes were responsible for >80% of all IPD cases in children aged <5 years prior to the introduction of PCV7²⁷. Two-thirds of AOM cases have been attributed to PCV7 serotypes in the US prior to PCV7 introduction³⁰. Following the introduction of PCV7, there was a significant increase in the incidence of IPD and AOM associated with non-PCV7 serotypes due to serotype replacement, in particular serotypes 3, 7F, and 19A^16,39–42. Several studies have also highlighted the persistence of select vaccine serotypes after PCV13 introduction, particularly serotype 3^17,18.

The next generation of pediatric pneumococcal vaccines continues to maintain serotypes that were highly prevalent and invasive prior to the introduction of earlier-generation PCVs, while extending coverage to non-vaccine serotypes currently causing disease. This modeling analysis demonstrates the importance of this strategy by estimating cases and deaths attributable to licensed and non-vaccine-type serotypes (22F and 33F) in a hypothetical cohort of US children followed from birth to 20 years of age.

Substantial clinical and economic burden continues to be prevented by serotypes contained in previously licensed vaccines, particularly those included in PCV7. In the pre-PCV7 period, more than 16,000 PCV7-type IPD cases and nearly 4.6 million PCV7-type AOM cases were predicted to occur in a single birth cohort, with a total cost of $1.4 billion. More than 2,000 IPD cases, and nearly 0.8 million AOM cases, were attributable to the six additional serotypes in PCV13, with a total cost of $229 million. The disease burden associated with these six additional serotypes further increased in the pre-PCV13 period. Excluding these serotypes from next-generation pediatric pneumococcal vaccines could have severe morbidity and mortality consequences, should these serotypes return. In addition, the near elimination of these serotypes from nasopharyngeal carriage through vaccination has prevented transmission and disease in older unvaccinated age groups^43,44.

Evidence of serotype persistence suggests a need for vaccines with greater effectiveness against select serotypes^45,46. In this simulation, serotype 3 still caused 202 IPD cases and 66,911 AOM cases in the post-PCV13 period. Finally, despite the substantial reduction of disease following the introduction of PCVs, the residual disease caused by non-vaccine serotypes remains. In the US, non-PCV13 serotypes accounted for 75.7% of IPD for children <5 years of age in 2017²¹. In this simulation, in a single birth cohort, serotypes 22F and 33F were associated with 298 IPD cases and 66,911 AOM cases over 20 years in the post-PCV13 period.

The value of next-generation pediatric pneumococcal vaccines should therefore not be limited to disease prevented by the non-vaccine serotypes only. This analysis demonstrates that the full value of next-generation vaccines should account for both those serotypes in the previously licensed vaccines, as well as the additional non-vaccine type serotypes.

A limitation of this analysis was that PMS and the prevention of NBPP were not considered. The availability of data for the pre-PCV7 period was limited, particularly for AOM. Furthermore, the analysis did not include direct non-medical costs borne by families or caregivers, such as transportation and lodging. Indirect costs associated with productivity loss were estimated using conservative values for earnings; for caregivers, only absenteeism, not presenteeism, was included in the productivity loss calculation. For children, the average income for those under 20 years of age was assumed to be zero. Consequently, overall estimates of health and economic burden of IPD and AOM associated with V114 serotypes were conservative.

Conclusions

This modeling analysis showed that PCV7 serotypes contributed to a high amount of morbidity and costs during both pre-PCV7 and pre-PCV13 eras in a hypothetical US birth cohort followed for 20 years. Three of the additional serotypes in PCV13 (3, 7F, and 19A) were associated with substantial IPD-related morbidity after the introduction of PCV7; serotype 19A was also associated with substantial AOM-related morbidity and overall costs pre-PCV13. Unique V114 serotypes 22F and 33F were associated with additional morbidity and costs during all three periods. Investigational PCVs for pediatric populations must continue to retain and provide protection against serotypes in currently licensed vaccines to maintain disease reduction while expanding coverage to non-vaccine serotypes.

Transparency

Declaration of funding

This analysis was funded by Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Kenilworth, NJ, USA.

Declaration of financial/other interests

All authors are employees of Merck Sharp & Dohme Corp., a subsidiary of Merck & Co., Inc., Kenilworth, NJ, USA, who may own stock and/or hold stock options in Merck & Co., Inc., Kenilworth, NJ, USA.

JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Author contributions

TP, TH, TW, and KOE designed the study; TP and TH analyzed the study data; and TP, TH, and TW interpreted the study data. All authors critically reviewed the manuscript and approved the final version for submission.

Previous presentations

An earlier iteration of this model was presented as a poster at the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) annual conference, 18–22 May 2019, New Orleans, LA, USA.

Acknowledgements

The authors would like to thank Elamin Elbasha for technical advice on the pediatric Markov model. Medical writing support, including assisting authors with the development of the outline and initial draft and incorporation of comments was provided by Rachel Wright, PhD, and Gauri Saal, MA Economics, and editorial support, was provided by Annabel Ola and Ian Norton, all of Scion, London, supported by Merck & Co. Inc., Kenilworth, NJ, USA, according to Good Publication Practice guidelines (https://www.acpjournals.org/doi/10.7326/M15-0288). The Sponsor was involved in the collection, analysis, and interpretation of data, as well as data checking of information provided in the manuscript. However, ultimate responsibility for opinions, conclusions, and data interpretation lies with the authors.