Number needed to immunize to prevent RSV with extended half-life monoclonal antibody

Abstract

Background:

Respiratory syncytial virus (RSV) is one of the most important respiratory pathogens in young children. Infants <6 months of age and infants and young children with extreme pre-term birth, and cardiac and pulmonary co-morbidities experience the highest incidence of severe RSV disease. There are no licensed vaccines; immunoprophylaxis is recommended for the highest risk children. Extended half-life RSV monoclonal antibodies (EHL-mAbs) are under development intended for immunization of all infants and high-risk children <2 years of age. We modeled the anticipated public health benefits of RSV EHL-mAb immunization using the number needed to immunize (NNI).

Methods:

We combined RSV hospitalization, outpatient and outpatient lower respiratory tract infection (LRI) incidence estimates and a range of immunization efficacies to estimate the annual NNI. We calculated the absolute incidence rate reduction (ARR) by multiplying the incidence rates by immunization efficacy. NNI was calculated as the reciprocal of the ARR.

Results:

For an RSV EHL-mAb with 70% efficacy, 6–18 infants would need to be immunized to prevent one RSV-associated outpatient visit, and 13–33 infants would need to be immunized to prevent one RSV-associated LRI outpatient visit. To prevent one RSV-associated hospitalization, 37–85 infants 0–5 months of age, and 107–280 infants 6–11 months of age would need to be immunized.

Conclusions:

Public health benefits, such as disease cases averted due to immunization, are essential elements in consideration of candidate vaccines for a national immunization program. An RSV EHL-mAb of moderate efficacy could have high impact. These data provide an additional perspective for public health decision making.

1. Introduction

Respiratory syncytial virus (RSV) is one of the most important respiratory pathogens in young children worldwide [1]. Infants less than six months of age and infants and young children with extreme pre-term birth, chronic lung disease, and congenital heart disease experience the highest incidence of severe RSV disease. In the US, approximately 20% of infants have an outpatient visit [2] and 2% of all infants are hospitalized due to an RSV infection [3-5]. More than half of all RSV hospitalizations in young children occur in infants <6 months of age [5]. Rates of RSV-associated hospitalization in children >1 year are also high, though they decline rapidly with age [6,7]. Severe disease in young infants may be mediated by a combination of lung/airway and immunological immaturity [8]. Despite the considerable burden, only symptomatic treatment is available and there are no licensed vaccines. Immunoprophylaxis is recommended for only the highest risk infants and children. Palivizumab, a short-acting monoclonal antibody administered monthly during the RSV season, was approved in 1998 but due to moderate efficacy [9] and high price, palivizumab is recommended for use in infants with extreme prematurity and infants and children with serious cardiac or pulmonary disease ≤24 months of age [10]. However, there are a growing number of RSV immunization candidates under development intended for RSV prevention in all infants with two main strategies in late stage: active immunization of pregnant women with transplacental transfer of maternal antibody to the infant [11-13], and passive immunization with high potency, long-acting EHL-mAbs [14-16] that require one dose for seasonal prophylaxis. RSV EHL-mAbs allow for flexibility of administration so that infants can be protected during their first RSV season when they are at highest risk of infection and severe disease. Infants born during the RSV season can receive the EHL-mAb at or shortly after birth. Infants born before the RSV season but entering their first RSV season can receive the EHL-mAb at the start of the season; administration may be synchronized with the 2, 4- or 6-month health maintenance visits or in ad hoc visits or seasonal clinics.

Public health programs consider candidate vaccines based on clinical efficacy and safety evaluated in the context of randomized clinical trials. Other important measures of vaccine value are the vaccine-attributable incidence rate reduction and cost-effectiveness. Several cost-effectiveness studies have modeled the anticipated benefits of RSV immunization with EHL-mAbs and maternal vaccines with a range of potential vaccine efficacies [17-19]; these strategies were cost-effective under certain ranges of assumptions. No published studies to date have modeled the anticipated benefits of RSV immunization using the number needed to vaccinate (NNV). The NNV estimates impact by expressing the benefit in terms of number of people needed to vaccinate to prevent one disease-related event each year. NNV is the reciprocal of the annual vaccine attributable incidence rate reduction [20]. We describe a corollary measure for the NNV, the number needed to immunize (NNI), which is more relevant to the passive EHL-mAb immunization strategy. The NNI is the number of infants that need to be immunized to prevent one RSV-associated outpatient visit, outpatient lower respiratory infection (LRI) or hospitalization. We calculated the NNI for outpatient LRI because RSV-associated LRI is a clinically significant outcome and this outcome is a primary endpoint in RSV EHL-mAb clinical trials (NCT03979313, NCT02624947) [21,22]. We combined RSV incidence estimates and a range of potential immunization efficacies to estimate the annual NNI for infants.

2. Materials and methods

We assumed that the RSV EHL-mAb was administered at the start of the RSV season and effectiveness was maintained at a steady state for the duration of the season. Our approach for calculation of NNV was adapted from a pediatric influenza NNV study conducted by Lewis et al. [23] We selected incidence studies for which we could ascertain the rates per 1000 children of RSV outpatient visits and hospitalizations in children 0–5 and 6–11 months of age [2,4-5,7,24]. Data for incidence rates of outpatient lower respiratory infection (LRI) were extrapolated from two of the outpatient studies [25-26]. In order to explore a range of values for NNI, we selected studies that represented a range of incidence rates [2,4-5,7,24-26]. We varied immunization efficacy (IE) from a low of 50% to a high of 90%. We calculated the absolute incidence rate reduction (ARR) due to immunization by multiplying the incidence rate by IE. We calculated 95% Wilson confidence intervals (CIs) for the ARR [27]. NNI was calculated as the reciprocal of the ARR (1/ARR) [20]. NNI 95% CIs were calculated as the reciprocal of the ARR 95% CI ((Lower Limit (NNI) = 1/Upper Limit (ARR) and Upper Limit (NNI) = 1/Lower Limit (ARR)).

2.1. Incidence source data

2.1.1. Outpatient visits

We selected three outpatient incidence studies with a range of incidence rates (Table 1). Two of these were multi-site studies conducted in the Nashville, Tennessee, Rochester New York, and Cincinnati, Ohio, using the same surveillance methods but different years [2,7]. The third study was conducted in Nashville, Tennessee, with different methods in different years [24] (Table 1).

Table 1.

RSV-Associated Pediatric Practice Visit and Hospitalization Rates and Rate Reduction by Month of Age and Immunization Efficacy.

Incidence Source Data	IncidenceRate/ 1000	ARR (95% CI)			NNI (95% CI)
		50% Immunization Efficacy	70% Immunization Efficacy	90% Immunization Efficacy	50% Immunization Efficacy	70% Immunization Efficacy	90% Immunization Efficacy
Outpatient Studies
Fisher [24]
Age 0–5 months	79.8	39.9 (27.1–56.4)	55.9 (40.4–74.8)	71.8 (54.2–92.8)	25 (18–37)	18 (13–25)	14 (11–18)
Age 6–11 months	89.8	44.9 (29.9–66.1)	62.9 (44.4–86.4)	80.8 (59.2–106.2)	22 (15–33)	16 (12–23)	12 (9–17)
Hall [7]
Age 0–5 months	132.0	66.0 (49.43–87.3)	92.4 (73.1–117.2)	118.8 (95.9–145.0)	15 (11–20)	11 (9–14)	8 (7–10)
Age 6–11 months	177.0	88.5 (67.19–118.5)	123.9 (98.4–157.6)	159.3 (130.3–195.9)	11 (8–15)	8 (6–10)	6 (5–8)
Lively [2]
Age 0–5 months	215.7	108.0 (80.3–140.0)	151.0 (118.9–187.8)	194.1 (158.5–234.6)	9 (7–12)	7 (5–8)	5 (4–6)
Age 6–11 months	246.1	123.1 (93.7–157.2)	172.3 (137.4–210.2)	221.5 (182.2–262.1)	8 (6–11)	6 (5–7)	5 (4–6)
Outpatient LRI Studies
Fisher (Lee) [25]
Age 0–5 months	45.0	22.5 (13.23–35.6)	31.5 (20.1–46.0)	40.5 (27.1–56.4)	44 (28–76)	32 (22–50)	25 (18–37)
Age 6–11 months	42.7	21.4 (12.0–37.8)	29.9 (17.7–47.5)	38.5 (23.7–56.9)	47 (26–84)	33 (21–56)	26 (18–42)
Hall [26]
Age 0–5 months	81.8	40.9 (29.2–59.9)	57.3 (43.0–78.8)	73.7 (57.2–97.3)	24 (17–34)	17 (13–23)	14 (10–17)
Age 6–11 months	109.7	54.9 (37.2–78.2)	76.8 (56.4–104.4)	98.8 (78.1–132.4)	18 (13–27)	13 (10–18)	10 (8–13)
Hospitalization Studies
Hall [7]
Age 0–5 months	16.9	8.5 (7.3–9.8)	11.8 (10.4–13.5)	15.2 (13.6–17.0)	118 (102–138)	85 (74–96)	66 (59–74)
Age 6–12 months	5.1	2.6 (1.9–3.3)	3.6 (2.8–4.5)	4.6 (3.7–5.6)	392 (299–526)	280 (224–360)	218 (177–270)
Paramore [4]
Age 0–5 months	25.7	12.9 (12.5–13.2)	18.0 (17.6–18.4)	23.1 (22.7–23.6)	78 (76–80)	56 (54–57)	43 (42–44)
Age 6–12 months	9.0	4.5 (4.3–4.7)	6.3 (6.0–6.7)	8.1 (7.7–8.4)	222 (212–233)	159 (150–167)	123 (119–130)
Stockman [5]
Age 0–5 months	38.6	19.3 (14.8–25.3)	27.0 (21.4–33.7)	34.7 (28.4–42.3)	52 (39–68)	37 (30–47)	29 (24–35)
Age 6–12 months	13.4	6.7 (4.5–10.6)	9.4 (6.3–13.7)	12.1 (8.5–16.8)	149 (95–235)	107 (73–158)	83 (60–118)

ARR = Absolute Rate Reduction, CI = Confidence Interval, NNI = Number Needed to Immunize.

Fisher et al. enrolled a birth cohort of 1113 healthy full-term infants (for a total of 3615 child-years of observation) in the vaccine treatment and evaluation unit (VTEU) over a 20-year period from 1973 to 1993. The median age was 13.4 months (range of 2 weeks to 5 years). Parents were instructed to bring their child to the clinic if they developed runny nose, cough, fever, or symptoms suggesting ear infection. Patients were seen if ill. Nasal wash cultures for virus isolation were obtained for any of the following indications: URI accompanied by fever of ≥38.4 °C, signs or symptoms suggesting LRI, and acute otitis media. Specimens were tested by viral culture during RSV season defined as November 1 to April 30. During the 20-year period the VTEU conducted two RSV vaccine trials and trial participants were excluded [24]. Incidence rates of RSV-associated LRI from this study were published in a later paper [25] (Table 1).

Hall et al. conducted active, prospective outpatient surveillance in the New Vaccine Surveillance Network (NVSN) from November until April over 2 years from 2002 to 2004 in children 0–59 months who visited a pediatric practice with acute respiratory infection in three US sites (Nashville, Rochester and Cincinnati (2003–2004 only)). Nasal and throat swabs were collected from each enrolled patient. RSV was detected by a reverse-transcriptase polymerase chain reaction (RT-PCR) assay performed at each study site. ARI/fever outpatient visits were estimated between November and April in 2002 to 2004 using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9), codes from the National Ambulatory Medical Care Survey (NAMCS) and the National Hospital Ambulatory Medical Care Survey NHAMCS) for primary care practices. Rates of ARI/fever visits by month of age were multiplied by the proportion of RSV positive cases in NAMCS and NHAMCS according to month of age to estimate outpatient RSV rates [7]. The proportion of outpatient infants with LRI were published in a later paper by Hall et al. [26] Incidence rates of outpatient LRI were estimated by multiplying the proportion of infants with LRI by the outpatient incidence rates in each age group (Table 1).

Lively et al. performed a recent analysis of NVSN outpatient data collected between November and April in the 2004–2009 RSV seasons in children aged 0–23 months [2]. Case ascertainment methods for this study were the same as those described in the Hall et al. study [7] (Table 1).

2.1.2. Hospitalization

We selected three studies for comparison of hospitalization rates (Table 1). Two studies were nationally representative medical encounter database studies and one study was an active prospective surveillance study. The studies used different methods; study periods overlapped for some years.

Hall et al. conducted active, prospective inpatient surveillance in the NVSN over 4 years from October 2000 until September 2004 in children 0–59 months in counties surrounding Rochester and Nashville and from 2003 until 2004 in Cincinnati. Eligible children were those who received a diagnosis of acute respiratory infection. Children were enrolled within 48 h after admission from Sunday through Thursday. During the study period, surveillance hospitals cared for more than 95% of each county’s children. Nasal and throat swabs were collected from each enrolled patient. RSV was detected by a RT-PCR assay. The weighted number of hospitalizations for RSV infection per 1000 children were expressed as the weighted number of RSV hospitalizations divided by the number of children in the county population determined by the 2000 US Census, multiplied by 1000. Rates were calculated by weighting the observed number of enrolled RSV hospitalizations to account for sampling 4 days per week and eligible patients who were not enrolled, by age < 1 year (<12 months), 1 year (12–23 months), and 2 to <5 years (24–59 months), with sites as sampling strata [7] (Table 1).

Paramore et al. retrospectively analyzed cross-sectional data from year 2000 from the Healthcare Cost & Utilisation Project’s Nationwide Inpatient Sample, a nationally representative federally funded medical encounter database, for children <5 years. Identification of RSV infection-associated hospitalization was based on at least one of the following ICD-9 diagnosis codes 079.6 (RSV); 466.11 (acute bronchiolitis due to RSV infection); 466.19 (acute bronchiolitis due organisms other than RSV; and 480.1 (pneumonia due to RSV infection) [4]. Sampling weights were assigned to each record for the calculation of rates [4] (Table 1).

Stockman et al. analyzed data from the National Hospital Discharge Survey from 1997 to 2006 for children 0–59 months. RSV-associated hospitalizations were classified using the ICD-9-CM codes: 466.11 (RSV bronchiolitis), 480.1 (RSV pneumonia), 079.6 (RSV, other), and bronchiolitis (466.1), acute bronchiolitis due to an infectious organism (466.19), pneumonia (480–486); for the second group of ICD-9 codes, Stockman estimated proportions of bronchiolitis (30%) and pneumonia (20%) hospitalizations likely to be associated with RSV infection from NVSN data [7]. Rates of hospitalization were based on bridged race census estimates from 1997 to 2006 (Table 1) [5].

3. Results

3.1. Outpatient visits

Among 0–5-month-old infants, 9–25 (95% CI 7–37) need to be immunized at 50% IE, 7–18 (95% CI 5–25) at 70% IE, and 5–14 (95% CI 4–18) at 90% IE to prevent 1 RSV-associated outpatient visit (Table 1, Fig. 1). Among 6–11-month-old infants, 8–22 (95% CI 6–33) would need to be immunized at 50% IE, 6–16 (95% CI 5–23) at 70% IE, and 5–12 (95% CI 4–17) at 90% IE to prevent 1 RSV-associated outpatient visit (Table 1, Fig. 1). These results indicate for an EHL-mAb with 70% efficacy 6–18 infants would need to be immunized to prevent 1 RSV-associated outpatient visit in the first year of life.

Fig. 1.

Number Needed to Immunize to Avert One RSV-Associated Pediatric Outpatient Visit in Infants by Age Group and Immunization Efficacy.

3.2. Outpatient LRI

Among 0–5-month-old infant, 24–44 (95% CI 17–76) would need to be immunized at 50% IE, 17–32 (95% CI 13–50) at 70% IE, and 14–25 (95% CI 10–37) at 90% IE to prevent 1 RSV-associated LRI outpatient visit (Table 1, Fig. 2). Among 6–11-month-old infants, 18–47 (95% CI 13–84) would need to be immunized at 50% IE, 13–33 (95% CI 10–56) at 70% IE, and 10–26 (95% CI 8–42) at 90% IE to prevent 1 RSV-associated LRI outpatient visit (Table 1, Fig. 2). These results indicate for an EHL-mAb with 70% efficacy 13–33 infants would need to be immunized to prevent 1 RSV-associated LRI outpatient visit in the first year of life.

Fig. 2.

Number Needed to Immunize to Avert One RSV-Associated Outpatient Lower Respiratory Infection in Infants by Age Group and Immunization.

3.3. Hospitalization

Rates of RSV-associated hospitalization were much higher in 0–5-month-old infant compared to those 6–11 months of age, thus NNI to prevent hospitalization was lower among infants 0–5 months of age. Among 0–5-month-old infants, 52–118 (95% CI 39–138) would need to be immunized at 50% IE, 37–85 (95% CI 30–96) at 70% IE, and 29–66 (95% CI 24–74) at 90% IE to prevent 1 RSV-associated hospitalization (Table 1, Fig. 3). Among 6–11-month-old infants, 149–392 (95% CI 95–526) would need to be immunized at 50% IE, 107–280 (95% CI 73–360) at 70% IE, and 83–218 (95% CI 60–270) for 90% to prevent 1 RSV-associated hospitalization (Table 1, Fig. 3). These results indicate for an EHL-mAb with 70% efficacy 37–280 infants would need to be immunized to prevent 1 RSV-associated hospitalization in the first year of life. Approximately three times as many infants 6–11 months of age would need to be immunized to prevent 1 RSV-attributable hospitalization compared those 0–5 months of age.

Fig. 3.

Number Needed to Immunize to Avert One RSV-Associated Hospitalization in Infants by Age Group and Immunization Efficacy.

4. Discussion

This analysis demonstrates that infant immunization with an RSV EHL-mAb of even moderate efficacy could have a high impact. The estimated NNI for the RSV EHL-mAb was lower than the NNVs for several vaccines against other pathogens included in the routine childhood immunization schedule in the United States. For an influenza vaccine with 50% efficacy, 1000–3000 children 6–23 months of age would need to be vaccinated to prevent 1 influenza-attributable hospitalization, and 12–42 children vaccinated to prevent 1 influenza-attributable outpatient visit [23]. For a EHL-mAb with 50% efficacy, an estimated 149–392 infants 6–12 months of age would need to be immunized to prevent 1 RSV-attributable hospitalization and 8–22 infants to prevent 1 RSV-attributable outpatient visit. For live pentavalent rotavirus vaccine at 85% efficacy, an estimated 200 infants would need to be immunized to prevent 1 rotavirus-associated hospitalization [28]. For an RSV EHL-mAb with 70% efficacy, an estimated 37–280 infants would need to be immunized to prevent 1 RSV-attributable hospitalization. The NNV for pneumococcal disease is substantially higher than the NNI for the RSV EHL-mAb. An estimated 671 infants would need to be immunized with pneumococcal conjugate vaccine to prevent one case of invasive pneumococcal disease [29] and an estimated 448 children <36 months of age would need to be immunized to prevent 1 case of consolidated community acquired pneumonia [28].

We took a conservative approach to the estimation of NNI. We used a wide range of immunization efficacies given that the efficacy of the RSV EHL-mAb is not yet determined. We used 70% as the mid-range reference point given similar recent Phase 2 efficacy results for a EHL-mAb in development [30]. We varied incidence estimates as seasonal epidemics of RSV vary from year to year and the incidence of medically attended visits may vary by geography due to health care utilization and access. In addition, we estimated the NNI for outpatient LRI. LRI is clinically meaningful and the NNI to prevent outpatient LRI may resonate globally with policymakers whose incidence of outpatient visits and hospitalization may differ by geography but for whom outpatient LRI may be a more constant measure.

Our analysis has some limitations. One of the primary criticisms of NNV is that it does not account for potentially significant indirect or dynamic effects of vaccines on the broader population [31]. The RSV EHL-mAb is targeted at infants and high-risk children <2 years of age (<2% of population) and is projected to confer immunity in this population for six months or less. Household genomic studies indicate that infants are not a primary source of transmission in the household but instead likely to get infected from older children [32,33]. Evidence from a modeling study that evaluated the maternal immunizations strategy, indicated that an immunization strategy in a narrow age group with a limited duration of protection is unlikely to provide indirect protection [34]. Thus, the dynamic effect of immunization strategies on the broader population is expected to be limited. If the mAb conferred indirect protection, that would make our estimates even more conservative.

We selected incidence studies that had included population rates per 1000 infants/children, that represented a two to threefold range of incidence and that used different methods of incidence estimation to reduce the bias associated with a given method. However, all methods of RSV incidence estimation are imperfect, and underestimation of incidence has been described in studies using population-based active surveillance [35] and in studies using ICD coding for identification of cases [36]. Underestimation of incidence would decrease the NNI, thus the true RSV EHL-mAb NNI may be even lower than estimated. Overall, we believe our estimates are conservative.

RSV infection may be severe in infants and young children born with severe prematurity, chronic lung disease other respiratory conditions [3], and hemodynamically significant congenital heart disease, leading to high incidence of hospitalization and outpatient visits [37]. Incidence of RSV hospitalization in this high-risk population is 2–10 times that of the healthy pre-term and full-term population [38]. Children with conditions that place them at high risk of severe disease are included in overall incidence estimates. Given the high incidence of medically attended visits in high-risk children, we would expect the NNI to be very low within these risk groups. It would be worthwhile to explore the specific benefits of the EHL-mAb in children with high-risk conditions, including children ≥1 year of age for which the rates of severe RSV-associated outcomes are high [3].

Most children are infected with RSV in the first year of life [39] and the greatest burden of severe RSV infection occurs in infants, perhaps due to a combination of lung and immunological immaturity [8]. Because RSV infection does not induce durable immunity [40], repeated infections occur during childhood and adulthood [39]. Reinfections are generally milder than primary infections evidenced by milder symptoms and lower levels of viral shedding [41-42] and by lower rates of RSV-associated hospitalization after one year of age. EHL-mAbs have a limited duration of protection and infants immunized in the first season may be susceptible to primary infection in the season after immunization. An increase in clinically significant RSV disease in the year following immunization would decrease the overall impact of the EHL-mAb. However, follow-up studies of infants and children prophylaxed with palivizumab or motavizumab monoclonal antibody have demonstrated that children were not at risk for increased rates of severe RSV disease in the subsequent seasons [43-45]. Follow-up for 510 days after immunization is planned in an ongoing RSV EHL-mAb clinical trial [21] and may provide information about the risk of clinically significant RSV disease in the year after immunization.

5. Conclusion

Population health benefits such as disease cases averted due to immunization, are of prime importance when considering the introduction of a new immunization into a national immunization program. The NNI is useful as it expresses the vaccine-mediated reduction in the underlying disease burden in an unvaccinated population [46]. The NNI can also provide a conceptual measure for the cost of preventing one disease event as the product of the number of doses needed and the immunization cost per dose [20,47]. The NNI is easy to understand and communicate and may help express the value of immunization to a variety of different stakeholders for whom complex output derived from economic or mathematical models may be more difficult to digest. The RSV EHL-mAb NNI compares favorably with other childhood vaccines included in the routine vaccination schedule. Given the competition for and scarcity of public health resources, these data provide an additional insight for resource allocation and public health decision making.

Abbreviations:

ARR

Absolute Rate Reduction

CI

Confidence Interval

EHL-mAbs

Extended half-life RSV monoclonal antibodies

IE

Immunization efficacy

ICD-9

International Classification of Diseases, Ninth Revision, Clinical Modification

LL

Lower Limit

LRI

Lower respiratory infection

NAMCS

National Ambulatory Medical Care Survey

NHAMCS

National Hospital Ambulatory Medical Care Survey

NVSN

New Vaccine Surveillance Network

NNI

Number needed to immunize

NNV

Number needed to vaccinate

RT-PCR

Real-time polymerase chain reaction

RSV

Respiratory syncytial virus

UL

Upper Limit.