- Email: [email protected]
Overview of Recent Clinical Trials of Acellular Pertussis Vaccines
Biologicals (1999) 27, 79–86 Article No. biol.1999.0184, available online at http://www.idealibrary.com on
Overview of Recent Clinical Trials of Acellular Pertussis Vaccines Elizabeth Miller Public Health Laboratory Service Communicable Disease Surveillance Centre 61, Colindale Avenue, London, NW9 5EQ
Abstract. The evidence from pre-licensure studies does not suggest that there are clinically important differences in reactogenicity between acellular vaccines. The merits of different acellular products will therefore have to be compared on efficacy criteria. Ideally, acellular vaccines with the minimum antigen content necessary to ensure optimum protection should be used in order to avoid administration of superfluous antigens to children and to simplify licensing and batch release procedures. On the basis of the evidence so far available it seems unlikely that monocomponent pertussis toxin (PT) vaccines provide optimal protection and that multicomponent vaccines are needed to achieve a level of disease control that approaches that of a good whole-cell vaccine. It is unclear whether all two component vaccines containing PT and filamentous haemagglutinin (FHA) have similar efficacy but on the available evidence the safest option for policy makers would seem to be to use a vaccine with at least three components, PT+FHA+pertactin. There is now good evidence that the five component vaccine which contains agglutinogens 2 and 3 in addition to PT/FHA and pertactin provides the best protection and is the only acellular vaccine whose efficacy matches that of a good whole cell vaccine. However, the public health advantage of the five component vaccine over other acellular vaccines may not become apparent until they have been in routine use for some decades and their ability to protect against transmission as well as clinical pertussis has emerged. The decision to replace an effective whole-cell vaccine by an acellular vaccine for primary immunisation needs careful consideration. Apart from the probable sacrifice of efficacy for reduced reactogenicity (at least for vaccines which do not contain agglutinogens 2 and 3) there is the question of value for money and the ease with which acellular DTP vaccines can be combined with conjugate polysaccharide vaccines such as Haemophilus influenzae type b. Whatever the decision of policy makers, the need for continued follow up of trial cohorts and active surveillance of the efficacy and safety of those acellular vaccines that are introduced into routine use must © 1999 The International Association for Biologicals be accorded a high priority.
Introduction The last few years have seen the culmination of a major international e#ort to evaluate the safety and e#icacy of acellular pertussis vaccines. The reactogenicity and immunogenicity results of the various Phase II trials in which acellular vaccines with widely di#erent antigen contents have been compared with whole cell vaccines have been unequivocal.1,2 All the trials have shown an improved safety profile with acellular vaccines for the common systemic symptoms and local reactions, whether given for primary immunisation or boosting. Antibody responses to the acellular antigens 1045–1056/99/020079+08$30.00/0
are either equivalent, or in many cases superior, to those generated by whole cell vaccines. The position with respect to the relative e#icacy and frequency of rare adverse events of di#erent acellular vaccines and their performance compared with whole cell vaccines is, however, less clear. This is due to di#erences between Phase III trials in design, case definition, schedule, study population and the whole cell vaccine used as a control. This paper reviews the results of the Phase III trials completed to date in the light of the epidemiological and biological questions that will need to be answered in order to formulate national policy on the use of acellular vaccines. 1999 The International Association for Biologicals
80
E. Miller
What policy makers need to know The principal aim of a Phase III vaccine trial is to provide evidence for regulatory authorities that the vaccine is e#icacious and that its safety profile is acceptable when compared to similar licensed vaccines. For this purpose, the results of each e#icacy trial stand alone and comparisons between trials are not indicated. For policy makers, however, the questions that need to be answered are di#erent. It is essential for policy makers to form a view on which acellular vaccine o#ers the best protection and whether there are di#erences between acellular vaccines in the incidence of common and rare adverse events. The optimum schedule must also be decided in the light of the performance of the vaccine, the age specific incidence of pertussis and the existing primary immunization schedule employed in that country for other paediatric vaccines. For countries with an existing successful whole cell vaccination programme it is essential to establish the e#icacy of the various acellular vaccines relative to the whole cell vaccine currently used and to quantify the gain in terms of reduced reactogenicity of switching from that whole cell to an acellular vaccine. Answering such questions inevitably necessitates making comparisons between trials and taking decisions on the basis of data which are incomplete and epidemiologically imperfect.
Overview of trial design The ideal scientific method for comparing the performance of vaccines is the double blind randomised trial in which two or more vaccines are evaluated, preferably with a placebo arm to allow absolute as well as relative vaccine e#icacy (VE) to be estimated. A double blind randomized trial with only one vaccine and a placebo arm gives unbiased estimates of vaccine e#icacy but is less useful for making decisions about the relative merits of di#erent products. Non-randomized vaccine studies can also provide estimates of absolute and relative e#icacy, or relative risk (RR) of disease, but need careful design and analysis to ensure freedom from bias due to allocation of vaccine, ascertainment and categorization of cases and the e#ect of confounding factors. The design of the randomized Phase III trials of acellular vaccines conducted to date is shown in Table 1, and the design of the nonrandomized studies in Table 2. All pertussis vac-
cines were given combined with diphtheria and tetanus vaccines and, in some trials, Haemophilus influenzae b (Hib) vaccine was given at the same time but at a di#erent site. One of the main factors influencing the e#icacy estimates of pertussis vaccine obtained from randomized trials or observational studies is the case definition employed in the analysis. Even within the same trial, the e#icacy estimates can vary from over 80% to around zero depending on the case definition used.3 In order to reduce the variation between trials arising from the use of di#erent case definitions, a meeting was held by WHO in 1991 at which those involved in sponsoring and conducting the various acellular vaccine trials agreed to use a common case definition in the primary analysis.4 The so-called ‘‘WHO case definition’’ was chosen to ensure high specificity and to reflect clinically significant pertussis infections and is as follows: d21 days days of paroxysmal cough plus B. pertussis isolated and/or serological evidence of pertussis and/or household contact with culture positive case within 28 days of onset. Despite the attempts to ensure uniformity through the ‘‘WHO case definition’’ confusion exists as to whether the 21 days of paroxysmal cough requires paroxysms on 21 consecutive days or paroxysms spanning a period of 21 days. It was also agreed at the WHO meeting that secondary analyses should be carried out with case definitions based on di#erent clinical and laboratory criteria to help elucidate the mechanism of action of the vaccine, for example whether protection against infection was as solid as protection against severe disease.
Role of different antigens in protection Monocomponent PT vaccines
There have been two trials in which a monocomponent PT vaccine has been evaluated (Table 1). The first Swedish trial, together with the subsequent unblinded follow up of the trial cohort, provided convincing evidence that the PT vaccine made by Biken (JNIH-7) protected less well against infection and clinically significant disease than the two component PT/FHA vaccine (JNIH-6) made by the same company.5,6 The di#erence between vaccines was maintained even after the exclusion of cases confirmed by FHA serology alone, a method which is likely to be more sensitive in children who
Clinical trials of acellular pertussis vaccines
81
Table 1. Double blind randomized trials of acellular pertussis vaccines Trial: Outcome measure (ref) Two or more acellular vaccines Sweden 86/87 VE, RR (5)
Pertussis vaccines
Manufacturer
Schedule
PT PT/FHA
Biken Biken
2 doses at 9 and 11 months
Sweden 92/95 VE, RR (16)
PT/FHA PT/FHA/69 kDa/Aggs 2,3 US whole cell
SKB Connaught Connaught
3 doses at 2, 4 and 6 months
Sweden 94/96 RR (17)
*PT/FHA PT/FHA/69 kDa/Aggs 2,3 recomb PT/FHA/69 kDa UK whole cell
SKB Connaught Biocine Evans
3 doses at 3, 5 and 12 months or 2, 4 and 6 months
Italy 92/95 VE, RR (15)
PT/FHA/69 kDa recomb PT/FHA/69 kDa US whole cell
SKB Biocine Connaught
3 doses at 2, 4 and 6 months
Single acellular vaccine Sweden Go¨ teborg 91/94 VE (7)
PT
Amvax
3 doses at 3, 5 and 12 months
Senegal 90/94 RR **(VE) (14)
PT/FHA French whole cell vaccine
Merieux Merieux
3 doses at 2, 4 and 6 months
Germany Erlangen 91/94 RR **(VE) (18)
PT/FHA 69 kDa/Agg 2 US whole cell
Lederle/Takeda Lederle
3 doses at 2, 4 and 6 months plus booster at 15–18 months
*Group unblinded and given 2 doses of the 3 or 5 component acellular DTP before completion of follow-up. **(VE) from open, non-randomised unvaccinated group.
Table 2. Non-randomized acellular pertussis vaccine studies Study: Outcome measure (ref)
Design
Pertussis vaccines
Manufacturer
Schedule
Mainz 92/94 VE, *RR (13)
Prospective, ‘‘blinded’’ PT/FHA/69 kDa SKB Behringwerke 3 doses at household contact study German whole 3, 4, 5 months cell vaccine plus booster in 2nd year of life
Munich 93/95 VE, *RR (12)
Case-control study
PT/FHA German whole cell vaccine
Biken-Connaught Behringwerke
3 doses at 3, 4, 5 months plus booster at 15–25 months
*RR estimates not included in primary analysis.
have not been immunized with an FHA containing vaccine3 (Table 3). The e#icacy estimates from the Go¨ teborg trial with the PT vaccine made by Amvax7 are shown in Table 4. Direct comparisons between the e#icacy estimates for the Biken PT vaccine and the Amvax PT vaccine are di#icult because of di#erences
between studies in schedule and the serological criteria used for confirmation of cases; however for culture positive cases with d21 days paroxysmal cough the e#icacy estimate for the two dose course of the Biken vaccine given at 9 and 11 months of age was 83% (63–92)3 compared with 76% (66–84) for the three dose course of the Amvax vaccine. The
82
E. Miller
Table 3. Cases confirmed by PT serology: first Swedish Trial with Biken vaccines JNIH-7 PT (n=1395)
JNIH-6 PT/FHA (n=1380)
RR PT:PT/FHA
50 34 17 6 20 5 12 9
25 20 11 10 12 4 6 8
1·95 1·65 1·48 0·59 1·60 1·21 1·92 1·06
All cases With paroxysmal cough With d14 days paroxysmal cough With d21 days paroxysmal cough With >5 spasms/day maximum With d15 spasms/day With vomiting With whooping
Table 4. E#icacy estimates from Go¨ teborg trial: 1724 PT AMVAX recipients Type of case Between dose 2 and 3* WHO definition d 7 days cough plus lab. confirmation After dose 3 WHO definition WHO cases excluding those confirmed by FHA serology alone d7 days cough plus lab. confirmation
No. cases
VE
14 28
55 (12–78) 41 (14, 59)
72 52
71 (63–78) 78 (71–84)
121
54 (43–63)
*From 30 days after dose 2 until 29 days after dose 3.
Table 5. Number of cases and range of VE estimates for PT/FHA vaccines Trial/study: vaccine
*Swedish Trial: SKB *Senegal Trial: Merieux **German case-control study: Biken
d21 days paroxysmal cough
d21 days cough
% paroxysmal
No.
VE
No.
VE
159 12
59 (51–66) 85 (66–93)
194 49
54 (46–62) 53 (23–71)
82% 25%
4
96 (78–99)
29
82 (68–90)
14%
*Laboratory confirmation as per WHO definition. **No serological confirmation, culture positive cases and/or epidemiologically linked cases only.
e#icacy of the Amvax vaccine after the first two doses given at 3 and 5 months of age was low and the third dose, which was given su#iciently late to act as a booster, appeared to be important. From the available data, it seems reasonable to conclude that there are no major di#erences between the Amvax and Biken products and that a monocomponent PT vaccine is probably inadequate. The view that PT alone is su#icient as the active component of an acellular vaccine and by itself could achieve pertussis elimination8 has so far received little support.9–11
PT/FHA vaccines
Three PT/FHA vaccines have been evaluated in blinded randomized controlled trials (Table 1), one of which (the Biken) has also been evaluated in a non-randomized observational study (Table 2). The e#icacy estimates for the three dose primary course12,13,14 are summarized in Table 5. It is evident that the precise clinical criteria used to define the quality of the cough can have profound consequences for the e#icacy estimates obtained. This e#ect is unlikely to be due biological di#erences between the vaccines and is probably a
Clinical trials of acellular pertussis vaccines
83
Table 6. Results of phase III acellular pertussis trials: Sweden and Italy: WHO case definition Vaccine E#icacy (95% CI) Swedish Trial
Italian Trial
Acellular vaccines (Manufacturer) PT/FHA (SKB) PT/FHA/69kDa/Agg 2+3 (Connaught) PT/FHA/69kDa (SKB) recomb PT/FHA/69 kDa (Biocine)
58 (50–65) 85 (80–89) — —
— — 84 (76–89) 84 (76–90)
Whole-cell vaccine U.S. Connaught Laboratories
48 (37–57)
36 (14–52)
recording artefact. Caution should therefore be exercised when making comparison between trials in e#icacy estimates even when the WHO case definition has been employed. It is also evident that quoting e#icacy estimates based on the WHO case definition without giving an indication of the total proportion of clinically significant, confirmed cases meeting that case definition in the trial can give a misleading picture of the probable public health value of a vaccine. Based on the e#icacy estimates for d21 days cough, neither the Merieux nor the SKB PT/FHA vaccines appears adequate. The higher e#icacy estimates for the Biken than the SKB and Merieux vaccines may reflect the absence of serologically confirmed cases. The exclusion of such cases generally raises e#icacy estimates as shown by the secondary analyses of the first Swedish trial. However, the higher e#icacy estimates for culture confirmed than for serologically confirmed cases were only evident those with <21 days cough.3 Another explanation of the higher e#icacy estimates for the Biken vaccine could be bias arising from the non-blinded non-randomized study design. However, the e#icacy estimate for culture positive/epidemiologically linked cases with d30 days cough for the Biken vaccine in the first Swedish trial (the closest comparison that can be made between studies) was 78% (57–88), similar to that for the Biken vaccine in the later case-control study. The possibility that the Biken PT/FHA vaccine provides better protection than the SKB and Merieux products cannot be excluded. Interestingly, the e#icacy estimate for culture positive cases with d21 days paroxysmal cough for the SKB PT/FHA vaccine was only 68% (60–75) (P Olin, personal communication), consistent with a lower e#icacy for this two component vaccine than for the Biken product.
Vaccines with three or more components
The results of the Swedish and Italian trials carried out in 1992/5 provided unequivocal evidence that the SKB vaccine containing only PT and FHA was substantially less protective than vaccines containing 69 kDa (pertactin) in addition to PT and FHA15,16 (Table 6). Taken together, these results strongly support a role for pertactin in protecting against clinical pertussis, although it has been suggested that lot variation between the two (PT/FHA) and three (PT/FHA/pertactin) component SKB vaccines used respectively in Sweden and Italy may have been responsible for the di#erences in clinical protection observed. However this seems unlikely as a second lot of the SKB PT/FHA vaccine used in the third Swedish trial also performed poorly compared to the vaccines in the remaining arms.17 This trial also provided evidence that the addition of purified agglutinogens 2 and 3 to a three component vaccine improved protection (Table 7). Relative efficacy of acellular and whole-cell vaccines
The US whole cell vaccine (Connaught) used in the Swedish and Italian trials performed very poorly (Table 6) but this has not been the experience with other whole-cell vaccines. The results of the five other trials which have provided relative e#icacy or relative risk estimates for acellular versus wholecell vaccines are summarized in Table 7 and 8. Four di#erent whole cell vaccines were used, three European (Evans Medical [ex Wellcome], Merieux and Behringwerke) and one American (Lederle).18 These results show that the Connaught whole-cell vaccine is atypical and that, with the exception of the five component vaccine, acellular vaccines are less e#icacious than a good whole cell vaccine.
84
E. Miller
Table 7. Relative risk (95% CI) of pertussis in acellular pertussis to U.K. whole-cell group in third Swedish trial according to case definition Vaccine
Culture positive pertussis
PT/FHA/69 kDa PT/FHA/69 kDa/Agg 2+3
Parentally reported pertussis
Any duration
d21 days cough
Certain diagnosis
Any reported case
2·55 (1·50–4·33) 1·40 (0·78–2·52)
1·38 (0·71–2·69) 0·85 (0·41–1·79)
1·48 (1·20–1·82) 1·12 (0·90–1·40)
1·22 (1·09–1·38) 0·98 (0·87–1·11)
Note: this trial did not include serologically confirmed cases.
Table 8. Relative risk and VEs (95% CI) of laboratory confirmed pertussis with d 21 days cough in acellular: whole-cell groups Erlangen (PT/FHA/69 kDa/Agg 2) d21 days cough plus either paroxysms, whooping or vomiting Senegal (PT/FHA) d21 days cough d21 days cough plus paroxysms
RR 2·1 (3·3)*
VE DTaP 81·5 (74·9)**
VE DTPwP 90·8 (86·1)**
1·59 (1·27–1·98) 3·18 (2·10–4·81)
58 (29–75) 74 (48–87)
77 (60–87) 91 (86–96)
Mainz (PT/FHA/69 kDa) WHO case definition
(RR) (4·8)
Munich (PT/FHA) d21 days cough d21 days cough plus paroxysms
NA NA
88·7 (76·6–94·6) 97·6 (83·1–99·7) 82 (68–90) 96 (87–99)
96 (78–99) 97 (79–100)
* Upper 95% CI. ** Lower 95% CI. (RR) – Not calculated by authors.
Moreover, data from the Senegal trial during the period of blinded follow up suggests a decline in protection with the Merieux acellular vaccine (PT/ FHA) after 18 months but not with whole cell vaccine.14 Booster immunization with the two component Merieux vaccine in the second year of the life is now recommended.14 In contrast, no decline in e#icacy has been reported with other acellular vaccines, based on post trial, unblinded, follow up of the trial cohorts.6,19,20 Unfortunately, none of these post trial studies provided comparative e#icacy data with a good whole cell vaccine. Better information on long term persistence of protection and the need for boosting is needed for acellular vaccines. Relative frequency of rare adverse events after acellular and whole-cell vaccines
One of the main driving forces behind the development of acellular vaccines was the perceived need for an improved safety profile, particularly with respect to the rare but potentially serious adverse
events. Although the risk of brain damage with whole-cell vaccine is now considered remote, other events such as high fever, convulsions, prolonged crying and hypotonic–hyporesponsive episodes (HHEs) have continued to cause concern. In the Swedish and Italian trials, prolonged crying and fever d40C were significantly more common after the US whole cell vaccine than either the acellular DTP or DT vaccines15,16 but no di#erences between vaccine groups were found in convulsions within 48 h of vaccination (Table 9). HHEs were more frequent in whole cell than acellular vaccine recipients in the second Swedish and Italian trials (P=0·008) (Table 9). In contrast, in the very large third Swedish trial the incidence of HHEs in acellular and whole-cell arms was similar (P=0·06) but convulsions were more frequent after whole cell vaccine (P<0·001). In other trials no HHEs have been reported in either the whole cell or acellular groups. At present, the pathogenesis of HHEs remains unclear and failure to detect any cases in some e#icacy trials suggest that the case definition
Clinical trials of acellular pertussis vaccines
85
Table 9. Number (frequency per 10 000 doses) of serious adverse events within 48 h of whole-cell DTP, acellular DTP, and DT vaccines: all doses combined, second Swedish and Italian trials together and third Swedish trial
Trial and Vaccine Group Sweden 2 plus Italy DT Connaught whole-cell DTP All acellular DTPs Sweden 3 U.K. whole-cell DTP All acellular DTPs
Fever d40C within 24 hours
Convulsion within 48 hours
Hypotonic hyporesponsive episode
12207 19662 15347
9 (7·3) 60 (30·5) 24 (15·6)
2 (1·6) 4 (2·0) 3 (2·0)
2 (1·6) 14 (7·1) 1 (0·7)
61400 184100
37 (6·0) 37 (2·0)
13 (2·1) 12 (0·7)
34 (5·5) 67 (3·6)
Number of doses given
Note. HHES within 48 hours in second Swedish and Italian trials and within 3 days in third Swedish trial.
is imprecise and sensitive to the method of follow up. In conclusion, the relative safety of acellular and whole cell vaccines with respect to the more severe or alarming adverse events is di#icult to quantify and is likely to depend on the whole cell vaccine and schedule used.21 Active post-licensure surveillance of acellular vaccines using methods such as record linkage22 will be essential to confirm freedom from rare adverse e#ects under conditions of routine use.
References 1. Decker MD, Edwards K, Steinho# MC et al. Comparison of 13 acellular pertussis vaccines; adverse reactions. Pediatrics 1995; 96: 557–566. 2. Edwards KM, Meade BD, Decker MD et al. Comparison of 13 acellular pertussis vaccines: overview and serologic response. Pediatrics 1995; 96: 548–557. 3. Storsaeter J, Hallander H, Farrington CP et al. Secondary analyses of the e#icacy of two acellular pertussis vaccines evaluated in a Swedish phase III trial. Vaccine 1990; 8: 457–461. 4. WHO meeting on case definition of pertussis. Geneva, January 10–11, 1991. MIM/EPI/PERT/91·1. Geneva: World Health Organisation. 1991. 5. Ad Hoc Group for the Study of Pertussis Vaccines. Placebo-controlled trial of two acellular pertussis vaccines in Sweden—protective e#icacy and adverse events. Lancet 1988; 1: 955–60. [Erratum, Lancet 1988;1:1238]. 6. Storsaeter J, Olin P. Relative e#icacy of two acellular pertussis vaccines during three years of passive surveillance. Vaccine 1992; 10:(3); 142–144. 7. Trollfors B, Taranger J, Lagergard T et al. A placebocontrolled trial of a pertussis-toxoid vaccine. N Engl J Med 1995; 333: 1045–1050.
8. Schneerson R, Robbins JB, Taranger J et al. A toxoid vaccine for pertussis as well as diphtheria? Lessons to be relearned. Lancet 1996; 348: 1289–1292. 9. von Konig CHW, Schmitt HJ. Toxoid vaccine for pertussis (corres.). Lancet 1997; 349: 136. 10. He Q, Viljanen MK, Arvilommi H, Mertsola J. Toxoid vaccine for pertussis (corres). Lancet 1997; 349: 137. 11. Preston NW, Matthews RC. Toxoid vaccine for pertussis (corres). Lancet 1997; 349: 137. 12. Liese JG, Meschievitz CK, Harzer E et al. E#icacy of a two-component acellular pertussis vaccine in infants. Pediatr Infect Dis 1997; 16: 1038–1044. 13. Schmitt H-J, von Konig CHW, Neiss A et al. E#icacy of acellular pertussis vaccine in early childhood after household exposure. JAMA 1996; 275: 37–41. 14. Simondon F, Preziosi M-P, Yam A et al. A randomized double-blind trial comparing a two-component acellular to a whole-cell pertussis vaccine in Senegal. Vaccine 1997; 15: 1606–1612. 15. Greco D, Salmaso S, Mastrantonio P et al. A controlled trial of two acellular vaccines and one wholecell vaccine against pertussis. N Engl J Med 1996; 334: 341–8. 16. Gustafsson L, Hallander HO, Olin P et al. A controlled trial of a two-component acellular, a five-component acellular, and a whole-cell pertussis vaccine. N Engl J Med 1996; 334: 349–355. 17. Olin P, Rasmussen F, Gustafsson L et al. Randomised controlled trial of two-component, three-component, and five-component acellular pertussis vaccines compared with whole-cell pertussis vaccine. Lancet 1997; 350: 1569–1577. 18. Heininger U, Cherry J, Stehr K et al. Comparative e#icacy of the Lederle/Takeda acellular pertussis component DTP (DTaP) vaccine and the Lederle whole-cell component DTP vaccine in German children after household exposure. Pediatrics 1998; 102: 546–553. 19. Taranger J, Trollfers B, Lagergard TG et al. Unchanged e#icacy of a pertussis toxoid vaccine
86
E. Miller
throughout the two years after the third vaccination of infants. Pediatr Infect Dis J 1997; 16: 180–184. 20. Salamso S, Mastrantonio P, Wassilak SGF et al. Persistence of protection by immunization in infancy with two three-component acellular pertussis vaccines. Vaccine 1998; 13: 1270–75. 21. Miller E, Ashworth LAE, Redhead K, Thornton C, Waight PA, Coleman T. E#ect of schedule on
reactogenicity and antibody persistence of acellular and whole-cell vaccines: value of laboratory tests as predictors of clinical performance. Vaccine 1997; 15: 51–60. 22. Miller E, Waight P, Farrington P. Safety assessment post-licensure. Dev Biol Stand 1998; 95: 235–243.
Overview of Recent Clinical Trials of Acellular Pertussis Vaccines Elizabeth Miller Public Health Laboratory Service Communicable Disease Surveillance Centre 61, Colindale Avenue, London, NW9 5EQ
Abstract. The evidence from pre-licensure studies does not suggest that there are clinically important differences in reactogenicity between acellular vaccines. The merits of different acellular products will therefore have to be compared on efficacy criteria. Ideally, acellular vaccines with the minimum antigen content necessary to ensure optimum protection should be used in order to avoid administration of superfluous antigens to children and to simplify licensing and batch release procedures. On the basis of the evidence so far available it seems unlikely that monocomponent pertussis toxin (PT) vaccines provide optimal protection and that multicomponent vaccines are needed to achieve a level of disease control that approaches that of a good whole-cell vaccine. It is unclear whether all two component vaccines containing PT and filamentous haemagglutinin (FHA) have similar efficacy but on the available evidence the safest option for policy makers would seem to be to use a vaccine with at least three components, PT+FHA+pertactin. There is now good evidence that the five component vaccine which contains agglutinogens 2 and 3 in addition to PT/FHA and pertactin provides the best protection and is the only acellular vaccine whose efficacy matches that of a good whole cell vaccine. However, the public health advantage of the five component vaccine over other acellular vaccines may not become apparent until they have been in routine use for some decades and their ability to protect against transmission as well as clinical pertussis has emerged. The decision to replace an effective whole-cell vaccine by an acellular vaccine for primary immunisation needs careful consideration. Apart from the probable sacrifice of efficacy for reduced reactogenicity (at least for vaccines which do not contain agglutinogens 2 and 3) there is the question of value for money and the ease with which acellular DTP vaccines can be combined with conjugate polysaccharide vaccines such as Haemophilus influenzae type b. Whatever the decision of policy makers, the need for continued follow up of trial cohorts and active surveillance of the efficacy and safety of those acellular vaccines that are introduced into routine use must © 1999 The International Association for Biologicals be accorded a high priority.
Introduction The last few years have seen the culmination of a major international e#ort to evaluate the safety and e#icacy of acellular pertussis vaccines. The reactogenicity and immunogenicity results of the various Phase II trials in which acellular vaccines with widely di#erent antigen contents have been compared with whole cell vaccines have been unequivocal.1,2 All the trials have shown an improved safety profile with acellular vaccines for the common systemic symptoms and local reactions, whether given for primary immunisation or boosting. Antibody responses to the acellular antigens 1045–1056/99/020079+08$30.00/0
are either equivalent, or in many cases superior, to those generated by whole cell vaccines. The position with respect to the relative e#icacy and frequency of rare adverse events of di#erent acellular vaccines and their performance compared with whole cell vaccines is, however, less clear. This is due to di#erences between Phase III trials in design, case definition, schedule, study population and the whole cell vaccine used as a control. This paper reviews the results of the Phase III trials completed to date in the light of the epidemiological and biological questions that will need to be answered in order to formulate national policy on the use of acellular vaccines. 1999 The International Association for Biologicals
80
E. Miller
What policy makers need to know The principal aim of a Phase III vaccine trial is to provide evidence for regulatory authorities that the vaccine is e#icacious and that its safety profile is acceptable when compared to similar licensed vaccines. For this purpose, the results of each e#icacy trial stand alone and comparisons between trials are not indicated. For policy makers, however, the questions that need to be answered are di#erent. It is essential for policy makers to form a view on which acellular vaccine o#ers the best protection and whether there are di#erences between acellular vaccines in the incidence of common and rare adverse events. The optimum schedule must also be decided in the light of the performance of the vaccine, the age specific incidence of pertussis and the existing primary immunization schedule employed in that country for other paediatric vaccines. For countries with an existing successful whole cell vaccination programme it is essential to establish the e#icacy of the various acellular vaccines relative to the whole cell vaccine currently used and to quantify the gain in terms of reduced reactogenicity of switching from that whole cell to an acellular vaccine. Answering such questions inevitably necessitates making comparisons between trials and taking decisions on the basis of data which are incomplete and epidemiologically imperfect.
Overview of trial design The ideal scientific method for comparing the performance of vaccines is the double blind randomised trial in which two or more vaccines are evaluated, preferably with a placebo arm to allow absolute as well as relative vaccine e#icacy (VE) to be estimated. A double blind randomized trial with only one vaccine and a placebo arm gives unbiased estimates of vaccine e#icacy but is less useful for making decisions about the relative merits of di#erent products. Non-randomized vaccine studies can also provide estimates of absolute and relative e#icacy, or relative risk (RR) of disease, but need careful design and analysis to ensure freedom from bias due to allocation of vaccine, ascertainment and categorization of cases and the e#ect of confounding factors. The design of the randomized Phase III trials of acellular vaccines conducted to date is shown in Table 1, and the design of the nonrandomized studies in Table 2. All pertussis vac-
cines were given combined with diphtheria and tetanus vaccines and, in some trials, Haemophilus influenzae b (Hib) vaccine was given at the same time but at a di#erent site. One of the main factors influencing the e#icacy estimates of pertussis vaccine obtained from randomized trials or observational studies is the case definition employed in the analysis. Even within the same trial, the e#icacy estimates can vary from over 80% to around zero depending on the case definition used.3 In order to reduce the variation between trials arising from the use of di#erent case definitions, a meeting was held by WHO in 1991 at which those involved in sponsoring and conducting the various acellular vaccine trials agreed to use a common case definition in the primary analysis.4 The so-called ‘‘WHO case definition’’ was chosen to ensure high specificity and to reflect clinically significant pertussis infections and is as follows: d21 days days of paroxysmal cough plus B. pertussis isolated and/or serological evidence of pertussis and/or household contact with culture positive case within 28 days of onset. Despite the attempts to ensure uniformity through the ‘‘WHO case definition’’ confusion exists as to whether the 21 days of paroxysmal cough requires paroxysms on 21 consecutive days or paroxysms spanning a period of 21 days. It was also agreed at the WHO meeting that secondary analyses should be carried out with case definitions based on di#erent clinical and laboratory criteria to help elucidate the mechanism of action of the vaccine, for example whether protection against infection was as solid as protection against severe disease.
Role of different antigens in protection Monocomponent PT vaccines
There have been two trials in which a monocomponent PT vaccine has been evaluated (Table 1). The first Swedish trial, together with the subsequent unblinded follow up of the trial cohort, provided convincing evidence that the PT vaccine made by Biken (JNIH-7) protected less well against infection and clinically significant disease than the two component PT/FHA vaccine (JNIH-6) made by the same company.5,6 The di#erence between vaccines was maintained even after the exclusion of cases confirmed by FHA serology alone, a method which is likely to be more sensitive in children who
Clinical trials of acellular pertussis vaccines
81
Table 1. Double blind randomized trials of acellular pertussis vaccines Trial: Outcome measure (ref) Two or more acellular vaccines Sweden 86/87 VE, RR (5)
Pertussis vaccines
Manufacturer
Schedule
PT PT/FHA
Biken Biken
2 doses at 9 and 11 months
Sweden 92/95 VE, RR (16)
PT/FHA PT/FHA/69 kDa/Aggs 2,3 US whole cell
SKB Connaught Connaught
3 doses at 2, 4 and 6 months
Sweden 94/96 RR (17)
*PT/FHA PT/FHA/69 kDa/Aggs 2,3 recomb PT/FHA/69 kDa UK whole cell
SKB Connaught Biocine Evans
3 doses at 3, 5 and 12 months or 2, 4 and 6 months
Italy 92/95 VE, RR (15)
PT/FHA/69 kDa recomb PT/FHA/69 kDa US whole cell
SKB Biocine Connaught
3 doses at 2, 4 and 6 months
Single acellular vaccine Sweden Go¨ teborg 91/94 VE (7)
PT
Amvax
3 doses at 3, 5 and 12 months
Senegal 90/94 RR **(VE) (14)
PT/FHA French whole cell vaccine
Merieux Merieux
3 doses at 2, 4 and 6 months
Germany Erlangen 91/94 RR **(VE) (18)
PT/FHA 69 kDa/Agg 2 US whole cell
Lederle/Takeda Lederle
3 doses at 2, 4 and 6 months plus booster at 15–18 months
*Group unblinded and given 2 doses of the 3 or 5 component acellular DTP before completion of follow-up. **(VE) from open, non-randomised unvaccinated group.
Table 2. Non-randomized acellular pertussis vaccine studies Study: Outcome measure (ref)
Design
Pertussis vaccines
Manufacturer
Schedule
Mainz 92/94 VE, *RR (13)
Prospective, ‘‘blinded’’ PT/FHA/69 kDa SKB Behringwerke 3 doses at household contact study German whole 3, 4, 5 months cell vaccine plus booster in 2nd year of life
Munich 93/95 VE, *RR (12)
Case-control study
PT/FHA German whole cell vaccine
Biken-Connaught Behringwerke
3 doses at 3, 4, 5 months plus booster at 15–25 months
*RR estimates not included in primary analysis.
have not been immunized with an FHA containing vaccine3 (Table 3). The e#icacy estimates from the Go¨ teborg trial with the PT vaccine made by Amvax7 are shown in Table 4. Direct comparisons between the e#icacy estimates for the Biken PT vaccine and the Amvax PT vaccine are di#icult because of di#erences
between studies in schedule and the serological criteria used for confirmation of cases; however for culture positive cases with d21 days paroxysmal cough the e#icacy estimate for the two dose course of the Biken vaccine given at 9 and 11 months of age was 83% (63–92)3 compared with 76% (66–84) for the three dose course of the Amvax vaccine. The
82
E. Miller
Table 3. Cases confirmed by PT serology: first Swedish Trial with Biken vaccines JNIH-7 PT (n=1395)
JNIH-6 PT/FHA (n=1380)
RR PT:PT/FHA
50 34 17 6 20 5 12 9
25 20 11 10 12 4 6 8
1·95 1·65 1·48 0·59 1·60 1·21 1·92 1·06
All cases With paroxysmal cough With d14 days paroxysmal cough With d21 days paroxysmal cough With >5 spasms/day maximum With d15 spasms/day With vomiting With whooping
Table 4. E#icacy estimates from Go¨ teborg trial: 1724 PT AMVAX recipients Type of case Between dose 2 and 3* WHO definition d 7 days cough plus lab. confirmation After dose 3 WHO definition WHO cases excluding those confirmed by FHA serology alone d7 days cough plus lab. confirmation
No. cases
VE
14 28
55 (12–78) 41 (14, 59)
72 52
71 (63–78) 78 (71–84)
121
54 (43–63)
*From 30 days after dose 2 until 29 days after dose 3.
Table 5. Number of cases and range of VE estimates for PT/FHA vaccines Trial/study: vaccine
*Swedish Trial: SKB *Senegal Trial: Merieux **German case-control study: Biken
d21 days paroxysmal cough
d21 days cough
% paroxysmal
No.
VE
No.
VE
159 12
59 (51–66) 85 (66–93)
194 49
54 (46–62) 53 (23–71)
82% 25%
4
96 (78–99)
29
82 (68–90)
14%
*Laboratory confirmation as per WHO definition. **No serological confirmation, culture positive cases and/or epidemiologically linked cases only.
e#icacy of the Amvax vaccine after the first two doses given at 3 and 5 months of age was low and the third dose, which was given su#iciently late to act as a booster, appeared to be important. From the available data, it seems reasonable to conclude that there are no major di#erences between the Amvax and Biken products and that a monocomponent PT vaccine is probably inadequate. The view that PT alone is su#icient as the active component of an acellular vaccine and by itself could achieve pertussis elimination8 has so far received little support.9–11
PT/FHA vaccines
Three PT/FHA vaccines have been evaluated in blinded randomized controlled trials (Table 1), one of which (the Biken) has also been evaluated in a non-randomized observational study (Table 2). The e#icacy estimates for the three dose primary course12,13,14 are summarized in Table 5. It is evident that the precise clinical criteria used to define the quality of the cough can have profound consequences for the e#icacy estimates obtained. This e#ect is unlikely to be due biological di#erences between the vaccines and is probably a
Clinical trials of acellular pertussis vaccines
83
Table 6. Results of phase III acellular pertussis trials: Sweden and Italy: WHO case definition Vaccine E#icacy (95% CI) Swedish Trial
Italian Trial
Acellular vaccines (Manufacturer) PT/FHA (SKB) PT/FHA/69kDa/Agg 2+3 (Connaught) PT/FHA/69kDa (SKB) recomb PT/FHA/69 kDa (Biocine)
58 (50–65) 85 (80–89) — —
— — 84 (76–89) 84 (76–90)
Whole-cell vaccine U.S. Connaught Laboratories
48 (37–57)
36 (14–52)
recording artefact. Caution should therefore be exercised when making comparison between trials in e#icacy estimates even when the WHO case definition has been employed. It is also evident that quoting e#icacy estimates based on the WHO case definition without giving an indication of the total proportion of clinically significant, confirmed cases meeting that case definition in the trial can give a misleading picture of the probable public health value of a vaccine. Based on the e#icacy estimates for d21 days cough, neither the Merieux nor the SKB PT/FHA vaccines appears adequate. The higher e#icacy estimates for the Biken than the SKB and Merieux vaccines may reflect the absence of serologically confirmed cases. The exclusion of such cases generally raises e#icacy estimates as shown by the secondary analyses of the first Swedish trial. However, the higher e#icacy estimates for culture confirmed than for serologically confirmed cases were only evident those with <21 days cough.3 Another explanation of the higher e#icacy estimates for the Biken vaccine could be bias arising from the non-blinded non-randomized study design. However, the e#icacy estimate for culture positive/epidemiologically linked cases with d30 days cough for the Biken vaccine in the first Swedish trial (the closest comparison that can be made between studies) was 78% (57–88), similar to that for the Biken vaccine in the later case-control study. The possibility that the Biken PT/FHA vaccine provides better protection than the SKB and Merieux products cannot be excluded. Interestingly, the e#icacy estimate for culture positive cases with d21 days paroxysmal cough for the SKB PT/FHA vaccine was only 68% (60–75) (P Olin, personal communication), consistent with a lower e#icacy for this two component vaccine than for the Biken product.
Vaccines with three or more components
The results of the Swedish and Italian trials carried out in 1992/5 provided unequivocal evidence that the SKB vaccine containing only PT and FHA was substantially less protective than vaccines containing 69 kDa (pertactin) in addition to PT and FHA15,16 (Table 6). Taken together, these results strongly support a role for pertactin in protecting against clinical pertussis, although it has been suggested that lot variation between the two (PT/FHA) and three (PT/FHA/pertactin) component SKB vaccines used respectively in Sweden and Italy may have been responsible for the di#erences in clinical protection observed. However this seems unlikely as a second lot of the SKB PT/FHA vaccine used in the third Swedish trial also performed poorly compared to the vaccines in the remaining arms.17 This trial also provided evidence that the addition of purified agglutinogens 2 and 3 to a three component vaccine improved protection (Table 7). Relative efficacy of acellular and whole-cell vaccines
The US whole cell vaccine (Connaught) used in the Swedish and Italian trials performed very poorly (Table 6) but this has not been the experience with other whole-cell vaccines. The results of the five other trials which have provided relative e#icacy or relative risk estimates for acellular versus wholecell vaccines are summarized in Table 7 and 8. Four di#erent whole cell vaccines were used, three European (Evans Medical [ex Wellcome], Merieux and Behringwerke) and one American (Lederle).18 These results show that the Connaught whole-cell vaccine is atypical and that, with the exception of the five component vaccine, acellular vaccines are less e#icacious than a good whole cell vaccine.
84
E. Miller
Table 7. Relative risk (95% CI) of pertussis in acellular pertussis to U.K. whole-cell group in third Swedish trial according to case definition Vaccine
Culture positive pertussis
PT/FHA/69 kDa PT/FHA/69 kDa/Agg 2+3
Parentally reported pertussis
Any duration
d21 days cough
Certain diagnosis
Any reported case
2·55 (1·50–4·33) 1·40 (0·78–2·52)
1·38 (0·71–2·69) 0·85 (0·41–1·79)
1·48 (1·20–1·82) 1·12 (0·90–1·40)
1·22 (1·09–1·38) 0·98 (0·87–1·11)
Note: this trial did not include serologically confirmed cases.
Table 8. Relative risk and VEs (95% CI) of laboratory confirmed pertussis with d 21 days cough in acellular: whole-cell groups Erlangen (PT/FHA/69 kDa/Agg 2) d21 days cough plus either paroxysms, whooping or vomiting Senegal (PT/FHA) d21 days cough d21 days cough plus paroxysms
RR 2·1 (3·3)*
VE DTaP 81·5 (74·9)**
VE DTPwP 90·8 (86·1)**
1·59 (1·27–1·98) 3·18 (2·10–4·81)
58 (29–75) 74 (48–87)
77 (60–87) 91 (86–96)
Mainz (PT/FHA/69 kDa) WHO case definition
(RR) (4·8)
Munich (PT/FHA) d21 days cough d21 days cough plus paroxysms
NA NA
88·7 (76·6–94·6) 97·6 (83·1–99·7) 82 (68–90) 96 (87–99)
96 (78–99) 97 (79–100)
* Upper 95% CI. ** Lower 95% CI. (RR) – Not calculated by authors.
Moreover, data from the Senegal trial during the period of blinded follow up suggests a decline in protection with the Merieux acellular vaccine (PT/ FHA) after 18 months but not with whole cell vaccine.14 Booster immunization with the two component Merieux vaccine in the second year of the life is now recommended.14 In contrast, no decline in e#icacy has been reported with other acellular vaccines, based on post trial, unblinded, follow up of the trial cohorts.6,19,20 Unfortunately, none of these post trial studies provided comparative e#icacy data with a good whole cell vaccine. Better information on long term persistence of protection and the need for boosting is needed for acellular vaccines. Relative frequency of rare adverse events after acellular and whole-cell vaccines
One of the main driving forces behind the development of acellular vaccines was the perceived need for an improved safety profile, particularly with respect to the rare but potentially serious adverse
events. Although the risk of brain damage with whole-cell vaccine is now considered remote, other events such as high fever, convulsions, prolonged crying and hypotonic–hyporesponsive episodes (HHEs) have continued to cause concern. In the Swedish and Italian trials, prolonged crying and fever d40C were significantly more common after the US whole cell vaccine than either the acellular DTP or DT vaccines15,16 but no di#erences between vaccine groups were found in convulsions within 48 h of vaccination (Table 9). HHEs were more frequent in whole cell than acellular vaccine recipients in the second Swedish and Italian trials (P=0·008) (Table 9). In contrast, in the very large third Swedish trial the incidence of HHEs in acellular and whole-cell arms was similar (P=0·06) but convulsions were more frequent after whole cell vaccine (P<0·001). In other trials no HHEs have been reported in either the whole cell or acellular groups. At present, the pathogenesis of HHEs remains unclear and failure to detect any cases in some e#icacy trials suggest that the case definition
Clinical trials of acellular pertussis vaccines
85
Table 9. Number (frequency per 10 000 doses) of serious adverse events within 48 h of whole-cell DTP, acellular DTP, and DT vaccines: all doses combined, second Swedish and Italian trials together and third Swedish trial
Trial and Vaccine Group Sweden 2 plus Italy DT Connaught whole-cell DTP All acellular DTPs Sweden 3 U.K. whole-cell DTP All acellular DTPs
Fever d40C within 24 hours
Convulsion within 48 hours
Hypotonic hyporesponsive episode
12207 19662 15347
9 (7·3) 60 (30·5) 24 (15·6)
2 (1·6) 4 (2·0) 3 (2·0)
2 (1·6) 14 (7·1) 1 (0·7)
61400 184100
37 (6·0) 37 (2·0)
13 (2·1) 12 (0·7)
34 (5·5) 67 (3·6)
Number of doses given
Note. HHES within 48 hours in second Swedish and Italian trials and within 3 days in third Swedish trial.
is imprecise and sensitive to the method of follow up. In conclusion, the relative safety of acellular and whole cell vaccines with respect to the more severe or alarming adverse events is di#icult to quantify and is likely to depend on the whole cell vaccine and schedule used.21 Active post-licensure surveillance of acellular vaccines using methods such as record linkage22 will be essential to confirm freedom from rare adverse e#ects under conditions of routine use.
References 1. Decker MD, Edwards K, Steinho# MC et al. Comparison of 13 acellular pertussis vaccines; adverse reactions. Pediatrics 1995; 96: 557–566. 2. Edwards KM, Meade BD, Decker MD et al. Comparison of 13 acellular pertussis vaccines: overview and serologic response. Pediatrics 1995; 96: 548–557. 3. Storsaeter J, Hallander H, Farrington CP et al. Secondary analyses of the e#icacy of two acellular pertussis vaccines evaluated in a Swedish phase III trial. Vaccine 1990; 8: 457–461. 4. WHO meeting on case definition of pertussis. Geneva, January 10–11, 1991. MIM/EPI/PERT/91·1. Geneva: World Health Organisation. 1991. 5. Ad Hoc Group for the Study of Pertussis Vaccines. Placebo-controlled trial of two acellular pertussis vaccines in Sweden—protective e#icacy and adverse events. Lancet 1988; 1: 955–60. [Erratum, Lancet 1988;1:1238]. 6. Storsaeter J, Olin P. Relative e#icacy of two acellular pertussis vaccines during three years of passive surveillance. Vaccine 1992; 10:(3); 142–144. 7. Trollfors B, Taranger J, Lagergard T et al. A placebocontrolled trial of a pertussis-toxoid vaccine. N Engl J Med 1995; 333: 1045–1050.
8. Schneerson R, Robbins JB, Taranger J et al. A toxoid vaccine for pertussis as well as diphtheria? Lessons to be relearned. Lancet 1996; 348: 1289–1292. 9. von Konig CHW, Schmitt HJ. Toxoid vaccine for pertussis (corres.). Lancet 1997; 349: 136. 10. He Q, Viljanen MK, Arvilommi H, Mertsola J. Toxoid vaccine for pertussis (corres). Lancet 1997; 349: 137. 11. Preston NW, Matthews RC. Toxoid vaccine for pertussis (corres). Lancet 1997; 349: 137. 12. Liese JG, Meschievitz CK, Harzer E et al. E#icacy of a two-component acellular pertussis vaccine in infants. Pediatr Infect Dis 1997; 16: 1038–1044. 13. Schmitt H-J, von Konig CHW, Neiss A et al. E#icacy of acellular pertussis vaccine in early childhood after household exposure. JAMA 1996; 275: 37–41. 14. Simondon F, Preziosi M-P, Yam A et al. A randomized double-blind trial comparing a two-component acellular to a whole-cell pertussis vaccine in Senegal. Vaccine 1997; 15: 1606–1612. 15. Greco D, Salmaso S, Mastrantonio P et al. A controlled trial of two acellular vaccines and one wholecell vaccine against pertussis. N Engl J Med 1996; 334: 341–8. 16. Gustafsson L, Hallander HO, Olin P et al. A controlled trial of a two-component acellular, a five-component acellular, and a whole-cell pertussis vaccine. N Engl J Med 1996; 334: 349–355. 17. Olin P, Rasmussen F, Gustafsson L et al. Randomised controlled trial of two-component, three-component, and five-component acellular pertussis vaccines compared with whole-cell pertussis vaccine. Lancet 1997; 350: 1569–1577. 18. Heininger U, Cherry J, Stehr K et al. Comparative e#icacy of the Lederle/Takeda acellular pertussis component DTP (DTaP) vaccine and the Lederle whole-cell component DTP vaccine in German children after household exposure. Pediatrics 1998; 102: 546–553. 19. Taranger J, Trollfers B, Lagergard TG et al. Unchanged e#icacy of a pertussis toxoid vaccine
86
E. Miller
throughout the two years after the third vaccination of infants. Pediatr Infect Dis J 1997; 16: 180–184. 20. Salamso S, Mastrantonio P, Wassilak SGF et al. Persistence of protection by immunization in infancy with two three-component acellular pertussis vaccines. Vaccine 1998; 13: 1270–75. 21. Miller E, Ashworth LAE, Redhead K, Thornton C, Waight PA, Coleman T. E#ect of schedule on
reactogenicity and antibody persistence of acellular and whole-cell vaccines: value of laboratory tests as predictors of clinical performance. Vaccine 1997; 15: 51–60. 22. Miller E, Waight P, Farrington P. Safety assessment post-licensure. Dev Biol Stand 1998; 95: 235–243.