ACIP-Results of Randomized Trial of a New H1N1 LAIV … · Improved LAIV strain selection identified a new H1N1 ... The proportion of subjects with any antibody response ... Grade

Results of Randomized Trial of a New H1N1 LAIV Strain in US Children

Raburn Mallory, MD Senior Director Clinical DevelopmentMedImmune/AstraZeneca

February 21, 2018

Improved LAIV strain selection identified a new H1N1 strain that is more immunogenic in children• LAIV is not recommended in the US as H1N1 vaccine strains used in 2013-2014 and 2015-2016 had

reduced effectiveness

• A broad-based scientific investigation determined that H1N1 LAIV strains used in those seasons replicate less well compared to older more effective LAIV strains

• Assays measuring replicative fitness were incorporated into strain selection for 2017-2018 and a new H1N1 strain (A/Slovenia) selected

• A clinical trial was conducted in US children to determine if new A/Slovenia strain was more immunogenic compared to previous A/Bolivia strain used in 2015-2016

• In the study, the new A/Slovenia H1N1 strain induced antibody responses that were significantly higher than those seen with the A/Bolivia strain Immune responses were similar to those seen with a highly effective pre-pandemic LAIV H1N1 strain

• Clinical study results validate improved strain selection process

• New strain selection process applied to all future LAIV strains and data reviewed by FDA/EMA annually

2 LAIV = Live attenuated influenza vaccine

Pediatric shedding and immunogenicity study design

Days 1-27 Days 28-56

Dose 1200 subjects 2 to < 4 years randomized Baseline immunogenicity assessed

Dose 2Immunogenicity Assessed

Day 56Immunogenicity Assessed

Shedding: Days 2, 3, 4, 5 and 7

Shedding: Days 30, 32 and 34

LAIV3 = Trivalent live attenuated influenza vaccine; LAIV4 = Quadrivalent live attenuated influenza vaccine

Arm 1: LAIV3 2015-2016 (A/Bolivia) Arm 2: LAIV4 2015-2016 (A/Bolivia) Arm 3: LAIV4 2017-2018 (A/Slovenia)

Vaccine type/subtype Arm 1: LAIV3 2015-2016

Arm 2: LAIV4 2015-2016

Arm 3:LAIV4 2017-2018

A/H1N1 A/Bolivia A/Bolivia A/Slovenia

A/H3N2 A/Switzerland A/Switzerland A/New Caledonia/14

B/Yamagata lineage B/Phuket B/Phuket B/Phuket

B/Victoria lineage -- B/Brisbane B/Brisbane

4

Vaccine strains by study arm

The two quadrivalent arms (Arms 2 and 3) represent the commercial vaccine that was distributed in those seasons

Primary endpoint:• The proportion of subjects with strain-specific HAI antibody seroconversion (≥ 4-fold increase)

through Days 28 and 56 Response rate for A/H1N1 strains in LAIV4 formulations was pre-specified for significance testing*

Secondary Endpoints: • Immunogenicity Neutralizing antibody seroconversion rates (≥ 4-fold increase from baseline) Nasal Immunoglobulin A (IgA) (≥ 2-fold increase from baseline; ≥ 4-fold shown) The proportion of subjects with any antibody response (≥ 4-fold increase in HAI antibodies,

neutralizing antibodies or IgA antibodies)

• Shedding of vaccine virus

• Safety

Study endpoints

5 HAI = Hemagglutination inhibition * Testing pre-specified in Statistical Analysis Plan for study

Baseline subject demographics were balanced between the treatment groups

6

LAIV3 2015-2016 (A/Bolivia)

N = 67


N = 66

LAIV4 2017-2018 (A/Slovenia)

N = 67

All subjects

N = 200

Median age, months 36.00 34.85 35.30 35.30

Gender

Male 40 (60%) 34 (52%) 32 (48%) 106 (53%)

Female 27 (40%) 32 (48%) 35 (52%) 94 (47%)

Prior vaccination and serostatus

Not previously vaccinated 29 (43%) 29 (44%) 29 (43%) 87 (44%)

Seronegative* 30 (45%) 43 (66%) 39 (58%) 112 (56%)

* Baseline HAI antibody titer ≤ 10

Subject disposition

7




All subjects

Number of subjects randomized 67 66 67 200

Number of subjects completing study 65 (97.0%) 63 (95.5%) 67 (100%) 195 (97.5.0%)

Reasons for not completing study

Withdrawal by legal guardian 0 1 (1.5%) 0 1 (0.5%)

Lost to follow-up 2 (3.0%) 2 (3.0%) 0 4 (2%)

Shedding of A/H1N1 strains: All subjects

8

4549

139

5

15

3841

12

5

12

5

4348

37

18

5

16

8

0

20

40

60

80

100

Day 2 Day 3 Day 4 Day 5 Day 7 Day 30 Day 32 Day 34

Study Day

2015-2016 LAIV3 (A/Bolivia)2015-2016 LAIV4 (A/Bolivia)2017-2018 LAIV4 (A/Slovenia)

Dose 1 Dose 2

Subjects shedding A/H1N1 vaccine virus*Percentage

* Subjects shedding quantifiable virus

Shedding of A/H1N1 strains: Seronegative subjects

9

43

53

1310

713

37

47

147

16

7

41

64

49

23

8

1510

0

20

40

60

80

100

Day 2 Day 3 Day 4 Day 5 Day 7 Day 30 Day 32 Day 34

Study Day


Dose 1 Dose 2



Study Primary Endpoint: A/H1N1 HAI antibody seroconversion rates after Dose 1 and 2

10 Dose 1

p = 0.006*

1

0

10

20

30

40

50

60

70

80

90

00

10

19.2

5.4 8.6

23.4

38.9

0

10

20

30

40

50

60

70

80

90

100

All subjects Seronegative subjects


p = < 0.001*

Dose 2

23.7

48

12.518.4

45.2

75

All subjects Seronegative subjects

Subjects seroconvertingPercentage

* A/H1N1 HAI antibody responses in LAIV4 formulations for all subjects assessed for statistical significance per Statistical Analysis Plan

Antibody seroconversion rates after Dose 1

11

103.3

30.235.9

5.4 5.4

39.7 42.2

23.415.9

23.9

37.3

0

10

20

30

40

50

60

70

80

90

100

HAI antibodies* Neutralizing antibodies IgA antibodies Any antibody response

Antibody Response


p = 0.006*p = 0.081**

p = 0.076** p = 0.589**


* A/H1N1 HAI responses in LAIV4 formulations assessed for statistical significance per Statistical Analysis Plan ** Post hoc analyses adjusted for multiple comparisons using Benjamini-Hochberg procedure

Antibody seroconversion rates after Dose 2

12

23.7

15.3

28.1

39.1

12.5 10.7

3541.9

45.2

31.7

52.3

68.2

0

10

20

30

40

50

60

70

80

90

100

HAI antibodies Neutralizing antibodies IgA antibodies Any antibody response


Antibody Response

p = < 0.001*

p = 0.014**

p = 0.076**

p = 0.012**


* Only A/H1N1 HAI responses in LAIV4 formulations assessed for statistical significance per Statistical Analysis Plan** Post hoc analyses adjusted for multiple comparisons using Benjamini-Hochberg procedure

A/Slovenia HAI seroconversion rates similar to those seen for previous H1N1 strain (A/New Caledonia/99) with high levels of efficacy

13

45

18

75

0

10

20

30

40

50

60

70

80

90

100

HAI seroconversion rates all subjectsHAI seroconversion rates seronegative subjects

A/Bolivia (LAIV4) A/Slovenia (LAIV4)

Current 2017 study (LAIV4)

57 5559

84

A/New Caledonia/99Study 11

A/New Caledonia/99Study 22

92% Efficacy 81% Efficacy

Studies prior to 2009 pandemic (LAIV3)1 Vesikari Pediatrics 2006 and 2 Tam PIDJ 200713

A/H1N1 strain summary

14

•

•

•

The A/Slovenia strain was shed by a higher proportion of children from Day 4 through Day 7, after Dose 1 of the vaccine

The study met its primary endpoint: HAI seroconversion rates were significantly higher for the A/Slovenia strain than for the A/Bolivia strain

Similar results seen for neutralizing antibodies and for nasal IgA

A/Slovenia HAI seroconversion rates were similar to those seen in children the same age vaccinated with a highly efficacious pre-pandemic H1N1strain

A/H3N2 HAI seroconversion rates similar to those seen for three previous H3N2 strains with high levels of efficacy

15

66

40

10091

0

10

20

30

40

50

60

70

80

90

100

A/New Caledonia/14A/Switzerland


HAI seroconversion all subjects

HAI seroconversion seronegative subjects

52

69

59

9792 94

A/SydneyStudy 11

A/PanamaStudy 22

A/WuhanStudy 33

90% Efficacy 73% Efficacy 96% Efficacy

Previous studies (LAIV3)1 Tam PIDJ 2007; 2 Neto PIDJ 2009 and 3 Belshe NEJM 1998

Key changes made in the way LAIV strains are selected annually

16

•

Human nasal epithelial cell culture now used to assess the replicative fitness of new LAIV strainsPrevious strains were assessed for fitness in MDCK (dog kidney) cells and eggsFor post-pandemic H1N1 strains results were not predictive of human nasal cell replication

•

•

TCID50 assay added to quantify new LAIV strainsFFA measures infectivity of vaccine viruses through expression of antigen on the cell surface TCID50 measures spread of vaccine virus between cells following multiple rounds of replicationFor post-pandemic H1N1 strains TCID50 values were up to 3 logs lower than FFA resultsAs a result, post-pandemic H1N1 strains with reduced effectiveness may not have sustained intranasal replication to the level needed for optimal effectiveness

All future strains now only selected if they have high levels of replication in human nasal epithelial cells and when the FFA and TCID50 assays give similar results*

* TCID50 and FFA results to be submitted and reviewed by FDA and EMA as part of annual strain change process

Summary

• A broad-based scientific investigation determined that H1N1 LAIV strains used in 2013-2014 and 2015-2016 had reduced replicative fitness compared to older more effective vaccine strains

• New assays measuring how well strains replicate were incorporated into strain selection for 2017-2018 and a new H1N1 strain (A/Slovenia) was selected

• In a randomized trial in US children, the new A/Slovenia strain induced antibody responses that were significantly higher than those seen with the 2015-16 H1N1 strain Immune responses similar to those seen with a highly effective pre-pandemic LAIV H1N1 strain

• Clinical study results validate improved strain selection process

• New strain selection process applied to all future LAIV strains and data reviewed by FDA/EMA on an annual basis

• LAIV is an important vaccine option for providers, patients, and parents in the US and other countries where it continues to be recommended

17 LAIV = Live attenuated influenza vaccine

Thank you!

Safety data

19

Solicited symptoms following vaccination

20

0

20

40

60

80

100

Any symptom Runny/stuffynose

Cough Decreasedactivity

Decreasedappetite

Fever > 101 F Headache Sore throat Muscle aches

Solicited symptoms


0

20

40

60

80

100 Dose 1

Dose 2

Subjects with solicited symptomsPercentage

Adverse events occurred at similar rates across the arms

21


N = 67


N = 66


N = 67

All subjects

N = 200

At least one event 29 (43.3%) 31 (47.0%) 34 (50.7%) 94 (47.0%)

At least one related event

5 ( 7.5%) 2 ( 3.0%) 3 ( 4.5%) 10 ( 5.0%)

At least one ≥ Grade 3 event

0 0 1 ( 1.5%) 1 ( 0.5%)

Death 0 0 0 0

At least one serious event

0 0 0 0

Grade 3 event was strep throat: this occurred in a male subject 16 days after the second dose of vaccine and was considered not related to vaccination

H1N1 data

22

23

23

0 .0

1 .0

2 .0

3 .0

4 .0

5 .0

6 .0

7 .0

8 .0

S tu d y d a y

Lo

g1

0 v

ira

l c

op

ies

/mL

D a y 3D a y 2 D a y 4 D a y 5 D a y 7

Q /L A IV 2 0 1 7 -1 8Q /L A IV 2 0 1 5 -1 6T r i/L A IV 2 0 1 5 -1 62015-2016 LAIV3 (A/Bolivia)

2015-2016 LAIV4 (A/Bolivia)2017-2018 LAIV4 (A/Slovenia)

Limit of quantification

Viral titers of shed A/H1N1 strains after Dose 1*

* Subjects with viral titers above limit of detection

A/H1N1 HAI seroconversion rates, by vaccination status

24

Dose 12015-2016 LAIV3 (A/Bolivia)2015-2016 LAIV4 (A/Bolivia)2017-2018 LAIV4 (A/Slovenia)

p = 0.006*

p = < 0.001*Dose 2

23.7 21.9 25.912.5

6.3

20.8

45.237.1

55.6

0

20

40

60

80

100

All subjects Vaccinated subjects Vaccine-naïve subjects

10.0 9.1 11.15.4 6.7 3.8

23.4 22.2 25.0

0.0

20.0

40.0

60.0

80.0

100.0

All subjects Vaccinated subjects Vaccine-naïve subjects


* A/H1N1 HAI antibody responses in LAIV4 formulations for all subjects assessed for statistical significance per Statistical Analysis Plan

A/Slovenia post-dose 2 GMTs similar to those seen for previous H1N1 strain (A/New Caledonia) with high levels of efficacy

7.3

29.2

14.5

49.9

0

10

20

30

40

50

60

Post-dose 2 HAI GMTs seronegative subjectsPost-dose 2 HAI GMTs all subjects

A/Bolivia (LAIV4) A/Slovenia (LAIV4)


32

21.2

46.6 45.7

A/New CaledoniaStudy 11

A/New CaledoniaStudy 22

92% Efficacy 81% Efficacy

Studies prior to 2009 pandemic (LAIV3)1 Vesikari Pediatrics 2006 and 2 Tam PIDJ 2007GMT = Geometric mean titer

H3N2 data

26

Shedding of H3N2 strains

27

8579 76

66

52

34

5 2

71 7076

64

44

32

93

64

5248 45

22

31

18

6

0

20

40

60

80

100

Day 2 Day3 Day 4 Day 5 Day 7 Day 30 Day 32 Day 34

Study DayDose 1 Dose 2 * Subjects shedding quantifiable virus

2015-2016 LAIV3 (A/Switzerland)2015-2016 LAIV4 (A/Switzerland)2017-2018 LAIV4 (A/New Caledonia)


A/H3N2 antibody seroconversion rates after Dose 1 and Dose 2

28

54.2 57.6

76.6 79.7

66.1 69.982 85.5

40.330

50.8

67.7

0

20

40

60

80

100

2015-2016 LAIV3 (A/Switzerland)2015-2016 LAIV4 (A/Switzerland)2017-2018 LAIV4 (A/New Caledonia)

Antibody Response

51.760

81.387.5

64.3 66.1

85.5 87.5

31.3 30.238.8

53.7

0

20

40

60

80

100

HAI antibodies Neutralizing antibodies IgA antibodies Any antibody respons

Dose 1

Dose 2

e


A/H3N2 post-dose 2 GMTs similar to those seen for three previous H3N2 strains with high levels of efficacy

29

293

127

385

170

0

50

100

150

200

250

300

350

400

A/New CaledoniaA/Switzerland


Post Dose 2 HAI GMTs all subjects

Post Dose 2 HAI GMTs seronegative subjects

34

202

5644

211

73

90% Efficacy 73% Efficacy 96% Efficacy

Previous studies (LAIV3)A/SydneyStudy 11

A/PanamaStudy 22

A/WuhanStudy 33

1 Tam PIDJ 2007; 2 Neto PIDJ 2009 and 3 Belshe NEJM 1998GMT = Geometric mean titer

-20

0

20

40

60

80

100

Effe

ctiv

enes

s/ E

ffica

cy (%

)

6 mo-8 yrs 2-18 years 2 years 2-17 years 2-17 yearsLAIV IIV IIV

UK TNCC

LAIV LAIV2 LAIV LAIV LAIV LAIVIIV3 IIV IIV IIV IIV9-17 yrs 2-17 years

US CDC TNCC

Germany TNCC

Canada TNCC5

Con-solidated

LAIV IIV

Japan TNCC

Finland Cohort4

Japan RCT

<6 years

A/New Caledonia H3N2 strain was effective in 2016-2017, comparable to IIV1

1 Estimate for all strains regardless of match to vaccine, except where noted; LAIV estimate not available for US and IIV estimate not available for UK. 2 Estimate for matched strains 3 Presented at Japan Ministry of Health 25 Aug 2017; test-negative study conducted in children < 6 years of age given two doses of vaccine. 4 Efficacy estimates for A strains; >90% of As30trains were H3N2 strains. 5 Unadjusted estimate.

B strain data

31

Shedding of B/Yamagata strain

32

21

3129

20 20

4 4 3

27

35

31

23

14

6 5

26

34

40

36

28

8

0

10

20

30

40

50

60

Day 2 Day3 Day 4 Day 5 Day 7 Day 30 Day 32 Da

Study DayDose 1 Dose 2

2015-2016 LAIV3 (B/Phuket)2015-2016 LAIV4 (B/Phuket)2017-2018 LAIV4 (B/Phuket)

y 34

Subjects shedding B/Yamagata vaccine virus*Number


B/Yamagata seroconversion rates after Dose 1 and Dose 2

33

Dose 1

Dose 2

50 48.3

65.678.1

42.937.5

87.3 89.1

57.848.4

61.2

77.6

0

20

40

60

80

100

5061.7

70.3

85.9

53.6 57.1

83.693.5

54.861.2

77.389.4

0

20

40

60

80

100

2015-2016 LAIV3 (B/Phuket)2015-2016 LAIV4 (B/Phuket)2017-2018 LAIV4 (B/Phuket)


Antibody ResponseHAI antibodidi es Neutralizing antibodies IgA antibodies Any antibody responseN t li i tib di I A tib di A tib d

Shedding of B/Victoria strain

34

35

47

42

29 30

21

6

2

36

45

5149

38

21

12

4

0

10

20

30

40

50

60

Day 2 Day3 Day 4 Day 5 Day 7 Day 30 Day 32 Day 34

Study DayDose 1 Dose 2

2017-2018 LAIV4 (B/Brisbane)2015-2016 LAIV4 (B/Brisbane)

Subjects shedding B/Victoria vaccine virus*Number


B/Victoria seroconversion rates after Dose 1 and Dose 2

35


Antibody ResponseHAI antibodies Neutralizing antibodies IgA antibodies Any antibody response

Dose 1

Dose 2

14.3 19.6

76.2 76.6

35.9

17.5

55.2

74.6

0

20

40

60

80

100

2516.1

83.690.3

40.3

21.7

78.889.4

0

20

40

60

80

100

2017-2018 LAIV4 (B/Brisbane)2015-2016 LAIV4 (B/Brisbane)

HAI tib di N t li i tib di I A tib di A tib d

Root cause investigation

36

LAIV effectiveness investigation: Root Causes• Likely root cause = Reduced replicative fitness of H1N1pdm09 LAIV strains

• No support for the following as likely root causes:

Pre-existing immunity from prior vaccination

Quadrivalent-specific vaccine virus interference

Vaccine virus temperature stability and heat exposure during shipping

Manufacturing

Stability at 2-8 ⁰C

37

Unlikely Root Cause:Pre-existing immunity among vaccinated children

38

LAIV3 demonstrated low to no vaccine effectiveness against post-pandemic strains in 2010-2011 season 1,2

Given observations of reduced effectiveness with LAIV3, quadrivalent-specific interference considered an unlikely root cause of the reduced VE

Interference could be a contributing factor to the low effectiveness in context of reduced replicative fitness of post-pandemic LAIV strains

However, A/Slovenia strain in the quadrivalent formulation was more immunogenic than the A/Bolivia strain in the trivalent or quadrivalent formulations

Immunogenicity of A/Slovenia was similar to a highly effective pre-pandemic strain in the trivalent formulation

1. Chung JR, Flannery B, Thompson MG, et. al. Pediatrics 2016, 2. Helmeke et al. PLoS One. 2015., 3. Victor J, Lewis K, Diallo A, et. al. Lancet Glob Health. 2016.

Unlikely Root Cause:Vaccine virus interference from quadrivalent formulation

39

• LAIV3 demonstrated low to no vaccine effectiveness against post-pandemic strains in 2010-2011 season 1,2

• Given observations of reduced effectiveness with LAIV3, quadrivalent-specific interference considered an unlikely root cause of the reduced VE

• Interference could be a contributing factor to the low effectiveness in context of reduced replicative fitness of post-pandemic LAIV strains

• However, A/Slovenia strain in the quadrivalent formulation was more immunogenic than the A/Bolivia strain in the trivalent or quadrivalent formulations

• Immunogenicity of A/Slovenia was similar to a highly effective pre-pandemic strain in the trivalent formulation

1. Chung JR, Flannery B, Thompson MG, et. al. Pediatrics 2016, 2. Helmeke et al. PLoS One. 2015., 3. Victor J, Lewis K, Diallo A, et. al. Lancet Glob Health. 2016.

Documents

ACIP-Results of Randomized Trial of a New H1N1 LAIV … · Improved LAIV strain selection identified a new H1N1 ... The proportion of subjects with any antibody response ... Grade