Smallpox gone

Measles within

the next 10-15

years

The eradication of

measles infection

will increase child

mortality in

Africa

1Fra CDC public health images

Peter Aaby

Christine S Benn

Bandim

The eradication scenario• The ultimate dream in Public health: eradication

• Measles targetted within next 10-15 years

• Polio targetted within next 5-10 years

• Rubella may be targetted with measles

• To quote Gates: Smallpox gone; Polio 99% down; Measles deaths 98% down => Vaccines best buy in Global Health

2

Global Health

• Removing/reducing vaccinations and reduced outbreak control makes eradication efforts cost-effective

• This is clearly beneficial: Saving lives and money ................ .in the current paradigm where the only effect of vaccinations is to prevent against specific diseases

• What will happen if vaccines have other effects?

Measles eradication

• The immediate consequences:

1. WHO recommendations: The age of measles vaccination (MV) will be increased from 9 to 12 months – as in Latin America in 1996 when measles was eliminated.

2. It will become increasingly difficult to maintain

3

2. It will become increasingly difficult to maintain funding for MV activities

3. The supplementary immunisation activities (SIA) campaigns will be removed to save money

4. DTP/Penta/PCV is likely to be the last vaccinations to profile the immune system

Before-after measles vaccination (MV): Annual mortality rates in African community studies

6%

8%

10%

12%

14%Before

After

4

0%

2%

4%

6%

Bissau 6-35 mo Bandafassi 9-60 mo,

Senegal

Zaire 7-21 mo

Bissau: MV at 6 mo introduced 1979 – 3-fold reductionMeasles infection may have caused 10-20% of deaths!=> A beneficial effect unrelated to measles prevention

Reduction in mortality associated with MV and the

% deaths due to measles infection in unvaccinated children

405060708090

100

RCT

5

01020304050

Bss

au II

IB

issa

u IVSen

egal

Ban

glad

esh

Buru

ndiB

issa

u VB

issa

u VI

Bis

sau V

IIB

issa

u VII

IH

aiti

Ben

inSudan

+ Measles % Acute measles deaths

Not explained by prevention of acute (or delayed) measles infection!

Could it be selection bias?

Randomised/blind studies:

Measles vaccinated vs unvaccinated children

Design Control

group

Mortality (deaths/children) Mortality ratio

(MV+/MV-)Measles

Vaccinated

Not measles

vaccinated

Nigeria: Random, not

blind – 18 month

DTP 0%(0/26) 12%(3/27) 0.00 (0.0-2.5)

6

blind – 18 month

follow-up

0.00 (0.0-2.5)

Sudan: Random, not

blind – 5-9 months

Meningo-

coccal A+C

0.3%(1/340) 3.5%(6/170) 0.09 (0.0-0.7)

Bissau: Blind, not

random

2-years follow-up

Ineffective

measles

vaccine

5%(6/124) 13%(7/53) 0.32 (0.1-0.9)

Bissau: Random, not

blind – 3 months

during war

Inactivated

polio (IPV)

2%(4/211) 5%(11/222) 0.30 (0.1-0.9)

Surv

ival pro

babili

ty

EZ M

Control M

Control F

0.7

0.8

0.9

1.0

High-titre measles vaccine (HTMV), Bissau, 1986-90

7

Age (months)

Surv

ival pro

babili

ty

4 6 12 18 24 30 36 42 48 54

EZ F

0

0.6

0.7

Similar results in Senegal, Sudan and Haiti – HTMV withdrawn 1992

No explanation –but repeatable. How can an effective vaccine do this?

Introduction of DTP: Rural areas of Guinea-Bissau 1984-87

Surv

ival pro

babili

ty

0.96

0.98

1.00

DTP –

(N=868)

Children aged 2-8 moFollowed 6 mo

Unvaccinated: travelling; sick; days without vaccines

8

Follow-up (months)

Surv

ival pro

babili

ty

0 1 2 3 4 5 6

0.90

0.92

0.94 DTP +

DTP+ / DTP- 1.98 (1.03-3.8)

(N=967)

Diphtheria-Tetanus-Pertussis (DTP) effect strongest for girls

0

1

2

3

4

5

6

Cum

ula

ted M

ort

alit

y (

%)

0 1 2 3 4 5 6 7Age (months)

Early DTP

No early DTP

Females

0

1

2

3

4

5

6

Cu

mu

late

d M

ort

alit

y (

%)

0 1 2 3 4 5 6 7Age (months)

Early DTP

No early DTP

Males

Accumulated mortality curves for DTP-vaccinated vs no DTP at 2 months

9

355 356 353 304 11No early DTP (N=359)650 655 620 538 19Early DTP (N=668)

Number at risk

Age (months)

250 248 244 212 5No early DTP (N=256)524 531 498 418 14Early DTP (N=534)

Number at risk

Age (months)

DTP/noDTP MRR crude MRR adjusted

Girls 2.5 (0.9-6.5) 5.7 (2.1-16)

Boys 0.5 (0.2-1.2) 1.3 (0.5-3.1)

All 2.6 (1.4-5.1)

HTMV and DTP?

30

40

50

60

Nu

mb

er

of

de

ath

s

15

20

25

30

35

40

Nu

mb

er o

f d

eath

s

Girls

Boys

No DTP after HTMV DTP/IPV after HTMV

10

0

10

20

Senegal

HT

Senegal

routine

Bissau Gambia Sudan Congo

Nu

mb

er

of

de

ath

s

0

5

10

Senegal

HT

Senegal

Routine

Bissau Gambia Sudan

F/M ratio: 0.96 (0.7-1.3) F/M ratio: 1.93(1.3-2.8)

HTMV withdrawn for the wrong reason

Not RCT – but this ”proves” a causal biological process

Annual mortality in MV trials depending

on DTP status at enrolment

Study Girls Boys

No

DTP3

DTP3

< MV

MR No

DTP3

DTP3

< MV

MR

2-dose 7.5% 3.8% 1.97(1.0-3.7) 6.4% 6.0% 1.06(0.6-1.9)

11

2-dose

Bissau

7.5% 3.8% 1.97(1.0-3.7) 6.4% 6.0% 1.06(0.6-1.9)

Sudan 6.0% 2.8% 2.16(0.3-17.3) 1.4% 1.9% 0.71(0.1-7.9)

Congo 10.0% 2.8% 3.06(0.6-16.1) 10.6% 5.1% 2.06(0.5-9.22)

Total 2.10(1.2-3.7) 1.13(0.7-1.9)

DTP after measles vaccine associated with 2 fold higher mortality

40

60

80

10

0M

ort

alit

y r

ate

(p

er

10

00

ye

ars

)

Guinea-Bissau 1992-94

40

60

80

Mort

alit

y r

ate

(per

1000 y

ears

)

Gambia 1998-2002

What can be done to reduce the negative effect of DTP?

12

02

04

0M

ort

alit

y r

ate

(p

er

10

00

ye

ars

)

0 3 6 9 12 15 18 21 24 27 30 33 36Age (months)

Male Female

Bandwidth: 4 months0

20

Mort

alit

y r

ate

(per

1000 y

ears

)

0 3 6 9 12 15 18 21 24 27 30 33 36Age (months)

Male Female

Bandwidth 4 months

Increased female mortality in the age groups of DTP => Change the immunological profile with a live vaccine =>

RCT: Early Measles Vaccine at 4½ m

Testing non-specific effects of MV

Birth

BCGDTP/OPV

1 2 3 MV

36 mo9 mo6 10 14 weeks

Follow-upOPV

4½ mo

MV vs No vac

13

• Recruitment 2003-2007 – Follow-up to 2009

• 6,600 randomised to A) Edmonston-Zagreb (EZ) at 4½+9

mo, or B+C) no vaccine at 4½ mo and EZ MV or Schwarz

MV at 9 mo

• DTP3 four weeks before enrolment – to prevent the

problem of DTP after MV

• Study designed to test a 25% difference in mortality

RCT of two doses of Measles Vaccine

34

56

7accum

ula

ted m

ort

alit

y (

%)

Two-dose MV

One-dose MV

Children born before June 2004

34

56

7accum

ula

ted m

ort

alit

y (

%)

Two-dose MV

One-dose MV

Children born after June 2004

14

Two-dose standard MV at 4½ and 9 mo was fully protective and

reduced mortality with 50% - with 45% if measles was excluded

MRR=0.48 (0.26-0.87)

P-value=0.02

01

2accum

ula

ted m

ort

alit

y (

%)

750 660 594One-dose405 377 334Two-dose

Number at risk

0 1 2 3Age in Years

MRR=0.53 (0.28-1.00)

P-value=0.050

12accum

ula

ted m

ort

alit

y (

%)

1184 1062 940One-dose599 541 485Two-dose

Number at risk

0 1 2 3Age in Years

RCTs of two doses of measles vaccine vs standard policy of one dose at 9 month

Study Follow-up period

Mortality rate ratio

Sudan, 1989-92

Vaccine 2007

5-36 mo 0.60 (0.3-1.4)

Bissau, 1993-95 6-18 mo 0.66 (0.2-2.3)

15

IJE 2003

6-18 mo 0.66 (0.2-2.3)

Bissau, 2003-09

BMJ2010

4½-36 mo 0.50 (0.3-0.8)

Combined 0.53 (0.36-0.77)

Observational study campaigns

BMJ 1993

9-60 mo

4-8 vs 9-11 mo

0.41 (0.2-0.9)

Measles eradication I: age of MV from 9 to 12 months

and DTP becomes the last vaccination

0

1

2

3

4

5

6

Cu

mu

late

d M

ort

alit

y (

%)

Early DTP

No early DTP

Females

20

40

60

80

Mort

alit

y r

ate

(per

1000 y

ears

)

Gambia 1998-2002

16

0

355 356 353 304 11No early DTP (N=359)650 655 620 538 19Early DTP (N=668)

Number at risk

0 1 2 3 4 5 6 7Age (months)

0

0 3 6 9 12 15 18 21 24 27 30 33 36Age (months)

Male Female

Bandwidth 4 months

Instead of moving the age foreward as done in Bissau it will be

moved backwards => increased mortality, particularly for girls

Measles eradication II: Campaigns will be removed

200

250

300

350

400

Death

s p

er

1000 liv

e b

irth

sMV

MV MV

BCG

17

050

100

150

200

Death

s p

er

1000 liv

e b

irth

s

1980 1985 1990 1995 2000 20052007Year

Under-5 mortality in Bandim 1978-2007

Polio eradication

• The immediate consequences:

1. Change from OPV to IPV for fear of OPV related polio

cases and revision of OPV strains

2. Removal of OPV

Evidence:

18

1. OPV may have beneficial effects in natural experiments:

The hospital case fatality was 3-fold lower when DTP was

absent and only OPV given

Mortality lower among those who received OPV in campaign

2. OPV beneficial effects in RCT (ínterim results)

3. IPV associated with 52% (2-128%) higher female than

male mortality in RCT

Smallpox and vaccinia eradication

• The first example of eradication of a disease

(1977) and removal of a vaccine (1980)

• Due to the fear of what could happen when

measles was eradicated we started in 1998 to

examine what happened after the removal vaccinia

19

examine what happened after the removal vaccinia

1. In Bissau: we made scar surveys and

followed for mortality

2. In Copenhagen: We used school

health cards which had vaccination information

and could link to Danish health registers

Su

rviv

al p

rob

ab

ility

0.4

0.6

0.8

1.0

Live vaccines are good – what happens when removed? Vaccinia after smallpox eradication in Guinea-Bissau

Vaccinia+

20

Age (years)

Su

rviv

al p

rob

ab

ility

25 30 40 50 60 70 80 90 100

0.0

0.2

1

2

Mortality rate ratio for Scar/no scar: Study I (1998-2002) 0.60 (0.4-0.9); F 0.51(0.3-0.8); M 0.72(0.4-1.2)Study II (2003-5) 0.22 (0.1-0.6): F 0.19(0.1-0.6); M 0.40(0.0-3.7)

Protection against HIV for scar/no scar: Female: 46% (0-71%)

No vaccinia

Smallpox and BCG phased out between 1965-1976 in DenmarkWe used Copenhagen school health cards with information on vaccinations to link with Danish health registers

BCGBCG reduced lymphomas with 51%(7-74%)

Live vaccines are good – what happens when removed? Vaccinia after smallpox eradication in Denmark

21

BCG reduced lymphomas with 51%(7-74%) Vaccine 2009

Smallpox vaccineAsthma reduced with 45% (0-70%) – J Allergy Clin Imm 2003Hospitalisation for inf diseases reduced with 16% (2-28%) IJE 2011

BCG and smallpox vaccineReduced the risk of hospitalization for HIV-infection by 65% (12-86%)

1

1.5

2

2.5

3

3.5

Mo

rta

lity

Ra

tio

* ** *

**

* *

The eradication of measles will lead to DTP

being the last vaccination in childhood

22

0

0.5

BMJ 2000 Benin Bissau-war DTP

introduction

Hospital

fatality

Malawi

BCG DTP Measles V

Mo

rta

lity

Ra

tio

* Significant difference between DTP and BCG or between DTP and MV

MV at 4+9mo vs No vac(DTP3)+MV at 9mo

1

1.5

2

2.5

3

3.5

4

MV 4+9 mo

DTP3+MV 9 mo

0.33(0.1-0.9)

0.56(0.3-0.9)

0.50(0.3-0.8)

Mortality

rate

23

0

0.5

1

4-8 mo 9-36 mo 4-36 mo

Having DTP as last vaccinationThe only RCT suggests 3-fold higher mortality in infancy for DTP =>Globally this would mean 100,000s of children every year

The eradication scenario:What can be done?

• Global health’s eradication strategy is contradicted by all data

• INDEPTH sites are in best situation to resolve these contradictions: are there or are there not non-specific beneficial effects of vaccines

• Polio - OPV

– Test IPV vs OPV in RCT

– Document effect of OPV campaigns –

• Maybe have a phased implementation of a trial

– Gradual phasing out of OPV

24

– Gradual phasing out of OPV

• Measles vaccine

– Test 12 mo vs 9 mo in RCT

– Document effect of MV campaigns (SIA) of children aged 9-60 months

• Maybe phased implementation

• Comparison of effects above and below 9 months

– Gradual change in age of vaccination or phasing out

• Smallpox vaccine

– More surveys

Su

rviv

al p

rob

ab

ility

4 6 12 18 24 30 36 42 48 54

EZ M

EZ F

Control M

Control F

0

0.6

0.7

0.8

0.9

1.0

High-titre measles vaccine: 2-fold higher female mortality

Surv

ival pro

babili

ty

6 12 18 24 30 36 42 48 54 60

EZ M

EZ F

Con M

Con F

0

0.80

0.85

0.90

0.95

1.00

SENEGAL 1987-92, EZ-HTBISSAU 1986-90, EZ-HT

25

Age (months)

Lessons from high-titre measles vaccine (HTMV) trials:

• EZ HTMV was fully protective against measles => negative non-specific effect• Sex-differential effect

• Public health effects: 35% excess mortality from 4 to 60 months => at least ½ mill

annual deaths in Africa

WHO introduced HTMV 1989 and withdrew it in 1992 =>

Interpretation: Too much of a good thing => Major donors: Money for new vaccines!

=> We looked for important NSEs of other vaccines

Age (months)

MV at 4+9mo vs No vac(DTP3)+MV at 9mo by Vitamin A-at-birth status

0.8

1

1.2

1.4

1.6

1.8

2

MV 4+9 mo

DTP3+MV 9 mo

0.98(0.6-1.5) 0.34(0.2-0.7)Mortality

rate

26

0

0.2

0.4

0.6

0.8

Vitamin A Placebo

DTP3+MV 9 mo

P=0.012

Vitamin A may have a fundamental impact on the NSEs=> Only those who did not receive VAS-at-birth

Non-specific effects (NSE) of standard MV at 4½ and 9 months of age:

General reduction in childhood mortality

27

childhood mortality

• Vaccines stimulate the immune system affecting susceptibility

• The NSE are often more important than specific effects

• Vaccination programmes should take the NSE into consideration: age at vaccination,

number of vaccinations, sequence of vaccinations

• Reconsider assumptions –

•Focus: specific diseases or immune deviations

•Effects may differ for boys and girls

•Interventions interact

• INDEPTH in a unique position to pursue these problems

• => EU is (hopefully) going to fund a multicentre trial of early MV

Impact of preventing measles infection:

Reduction in mortality with or without measles

cases in the survival analysis

405060708090

100

%

28

010203040

BIS

SAU

III

BIS

SAU

IVSE

NE

GA

LB

AN

GL

AB

UR

UN

DI

BIS

SAU

VB

ISSA

U V

IB

ISSA

U V

IIB

ISSA

U V

III

SUD

AN

+ Measles - Measles

%

Hypothesis: Non-specific beneficial immunestimulation