POST-PARTUM RUBELLA IMMUNISATION: A CONTROLLED TRIAL OF TWO VACCINES

990

POST-PARTUM RUBELLA IMMUNISATION: A CONTROLLED TRIAL OF TWO VACCINES NICHOLAS A. BLACK JOHN B. KURTZ ARTHUR LACEY

AUDREY PARSONS NORMAN MCWHINNEY RICHARD T. MAYON-WHITE

Oxford Regional Health Authority; Virology Department and Public Health Laboratory, John Radcliffe Hospital, Oxford; and John Radcliffe Maternity Hospital, Oxford

Summary

The effectiveness of two rubella vaccines 27/3 and Cendehill) in women

(RA

vaccinated post partum was compared. Significantly more women who received RA 27/3 gave satisfactory seroconversion reponses than women who received Cendehill (97·6% vs 82·2%). This difference was reflected in the geometric mean titres (43·6 vs 17·0). More women who received RA 27/3 had minor side-effects, but the difference was not significant. The serological response to both vaccines was not affected by the concurrent administration of anti-D immunoglobulin. In view of these findings, the replacement of Cendehill vaccine with RA 27/3 vaccine for women vaccinated post partum should be considered. Introduction THERE has been

concern

for several years about the

antibody response to the two rubella vaccines available in the United Kingdom, particularly when they are administered to women post partum. The geometric mean titre (GMT) after RA 27/3 (’Almevax’) was higher than that after Cendehill vaccine in

vaccinated post

partumland

in the seroconversion rates for the two vaccines were similar. Two of these studies2,3 were based on serological findings 6-8 weeks after vaccination, when it has been suggested that antibody production in response to Cendehill is still rising.5However, the other studiesl,4 were carried out around the time of the peak response to Cendehill, 10-12 weeks. women

schoolgirls.3,4 Despite these differences,

By 1979 many workers therefore believed that the vaccine incorporating the RA 27/3 strain "provided a much better immunological stimulus than other vaccines".6 In the USA Cendehill vaccine was discarded (a decision taken by the manufacturers rather than through medical discussion) in favour ofRA 27/3.7In the UK both vaccines are still in use. A small retrospective study of the response of women vaccinated post partum in the UK 6 weeks after immunisation suggested that RA 27/3 gave a higher seroconversion rate (96.6% vs 86%) and higher GMT (J. O’H. Tobin, unpublished). The interpretation of these results is difficult for several reasons-the data were obtained from nine centres, some of which only ever used one of the vaccines; several laboratories carried out the serological measurement; and in one centre the Cendehill vaccine used was subsequently withdrawn owing to lack of potency. However, the results suggested a need for a prospective, randomised, clinical trial, which we now report. This study also considers two other features of post-partum rubella vaccination. First, the reported higher incidence with RA 27/3 vaccine of unwanted side-effects, including arthralgia,8 sore throat, and rash, 1,4 and secondly, the possibility that anti-D immunoglobulin may inhibit the serological response to rubella vaccine.

Subjects and Methods

Study Design All rubella-seronegative women who gave birth in the John Radcliffe Maternity Hospital, Oxford, between January, 1981, and October, 1982, were invited to enter the study. Antenatal serological status had been established by means of single radial haemolysis and haemagglutination-inhibition techniques. In addition to post-partum vaccination before discharge (routine hospital policy) the women were offered a test 6 weeks after vaccination for confirmation of seroconversion. Any woman refusing this additional venepuncture at her postnatal examination was excluded from the study. The

received either Cendehill

RA 27/3 vaccine, which thus allowing for seasonal variation in the incidence of natural infection. The vaccine in use was not revealed to the subjects before they agreed to participate in the study. The following information was noted for each woman: name, address, age, antenatal antibody status (trace or no trace of antibodies), type of vaccine, date of vaccination, and whether or not anti-D was also given. were

women

provided alternately

on a

or

monthly basis,

We attempted to have serum samples collected at the postnatal visit, approximately 6 weeks after vaccination, by either the hospital or the general practitioner. However, this was not possible in some cases for administrative reasons, resulting in delays of up to 12 weeks. All serological testing was carried out by the Oxford virology laboratory. At the time of vaccination, the subjects were given a simple questionnaire on adverse reactions to the vaccine which they were asked to complete and return at their postnatal visit.

Laboratory Analysis The antenatal

serum samples were tested by single radial with antigen haemolysis9 prepared from the Judith Strain (Virus Reference Laboratory, Colindale), sheep red blood cells, complement, and the undiluted serum sample. All samples giving a zone of haemolysis less than 15 IU standard serum were checked by the haemagglutination inhibition technique10 with the same antigen, day-old-chick red blood cells, and serum samples absorbed with kaolin in U-bottomed plates. Patients whose samples showed titres of less than the lowest dilution (1/10) and a <15 IU zone of haemolysis were deemed to need vaccination. This group included those with no detectable antibody (nil) or with a low level (trace). The postnatal serum samples were tested by the haemagglutination-inhibition test, and in any demonstrating a trace of antibody this was confirmed by single radial haemolysis. Samples from patients with low postvaccination titres were retested in parallel with their antenatal samples. In addition, the serum samples of all patients with low antenatal titres were tested in parallel with their postvaccination samples.

Subjects Of the 298 women 153 received RA 27/3 and 145 received Cendehill. 57 (19%) women could not be included in the analysis owing to lack of postvaccination serum samples, insufficient data, or lack of written evidence of antenatal antibody status. These women did not appear to differ systematically from those included in the analysis, and approximately half came from each group, leaving 123 who received RA 27/3 and 118 Cendehill. The mean ages of the two groups were similar (28’3±SEM 5’0in the RA 27/3 group and 29 - 2±5.4in the Cendehill group), though there was slight excess of older women in the Cendehill group (table I). The timing of the postvaccination serum samples was similar in the two groups (65% within 7 weeks, 25% between 7 and 13 weeks, and 10%

over

13

weeks).

Statistical Analysis The relative effectiveness of the two vaccines was assessed by of the seroconversion rates and the resulting GMTs.

analysis

991 TABLE I-AGE DISTRIBUTION OF WOMEN RECEIVING RA

27/3 OR

TABLE II—SEROCONVERSION AFTER POST-PARTUM VACCINATION

CENDEHILL RUBELLA VACCINE POST PARTUM

WITH RA

27/3 OR CENDEHILL

Differences between groups: for all subjects (nil+trace) X’=15-1, Idf, p<0. 01; for subjects with prevaccination trace X2 =9’ 73, ldf, p<0.01.

Differences in outcome were compared by a simple chi-squared test for conversion rates and a t test for GMTs. The logarithms of titres less than 1/10 were scored as 0. The data were also analysed according to whether or not a trace of antibody was present. The response of a subgroup who also received anti-D is also presented.

TABLE 111-TITRES BEFORE AND AFTER VACCINATION IN WOMEN RECEIVING RA

27/3 OR CENDEHILL

VACCINE POST PARTUM

Results 97-6% of women who received RA 27/3 showed satisfactory responses (table II), compared with only 82’ 2% 62 women had of those who received Cendehill (p<0.01). less titres than 1/10 but their serum samples prevaccination showed a trace of haemagglutination inhibition and single radial haemolysis less than 15 IU. It was in these women that the largest difference between the vaccines was found. A good response to RA 27/3 was noted in 24 of 26 women (92’3%), compared with 20 of 36 (55-6%) with Cendehill (p<0.01). All these differences persisted when the analysis was standardised for age. The GMT after RA 27/3 was higher than that after Cendehill (table III), whether or not the women had a preexisting trace of antibody. The greatest GMT (56’ 2) was found in the 97 recipients ofRA 27/3 who had no pre-existing antibody. The 36 women who had a trace of antibody and received Cendehill vaccine had the lowest GMT (5’5), owing to the effect of the 16 women whose antibody levels stayed at a

On

log titres t= 5 . 65;

TABLE

239

df; p<0 . 10 .

IV-FREQUENCY OF ADVERSE REACTIONS REPORTED

trace.

105 (44%) vaccinees completed the questionnaire about six

particular adverse reactions. Similar proportions of both groups reported symptoms (54% of the RA 27/3 group and 47% of the Cendehill group). However, the RA 27/3 women have more symptoms than the Cendehill women (73 compared with 50). This difference was largely confined to three symptoms - soreness at the injection site, reddish rash, and sore throat (table IV). The frequency of symptoms was higher in women who showed a good serological response to the vaccine. tended

to

A subgroup of 50 women received anti-D immunoglobulin time as rubella vaccine. The seroconversion rates shown by these women were similar to the rates for all women receiving the same vaccine. 96’3% of recipients ofRA 27/3 and anti-D immunoglobulin seroconverted, compared with 97-9% of women receiving RA 27/3 only. The rates of seroconversion for Cendehill recipients were 73,9 for those also receiving anti-D and 83 -. 2% for those receiving Cendehill only (X2 0 -73; p>0 . 05). A second .postvaccination antibody test was carried out in 28 women who attended an antenatal clinic with a subsequent pregnancy a mean of 16 months (range 9-27 months) later. Of the 15 women who had received RA 27/3, the titre had dropped in 6 (40%) and had remained the same in 7 (47%), whereas in those who were vaccinated with Cendehill, the titre had dropped in only 3 (23%), had remained the same in 6 (46%), and had risen in 4 (31%). at the same

=

Discussion In this study the serological response to RA 27/3 vaccine better in terms of both the numbers responding and the height of response than that to Cendehill vaccine. However, before the appropriate application of this finding can be made we must consider the time after vaccination at which effectiveness is assessed, the definition of a good serological response, and the long-term persistence of immunity to was

rubella. Most of the postvaccination serum samples were collected within 7 weeks of immunisation, when it has been suggested that the response to Cendehill vaccine has not yet reached its peak.5 This timing could account for some of the difference in GMTs between the vaccines, but it cannot explain the lower seroconversion rate with Cendehill. 7 weeks after vaccination evidence of seroconversion could reasonably be expected, even if the actual titre was a prepeak level. The difference between the two vaccines is due partly to the small response to Cendehill in women with a pre-existing trace of antibody. It is possible that the trace reflects sufficient

992

immunity, although a titre of

15 IU/ml has been accepted as a minimum level for antenatal screening tests. Indeed, the fact that 6 women without detectable antibody before vaccination failed to show a good response (1 to RA 27/3, 5 to Cendehill) need not necessarily indicate vaccine failure. They may have had a low level of antibody, providing adequate protection to the fetus, which could have been detected by more sensitive techniques, such as enzyme-linked immunosorbent assay or radioimmunoassay. The women who had no more than a trace of antibody after vaccination would have to be considered as seronegative in a subsequent pregnancy. We believe, therefore, that the greater response to RA 27/3 does have practical benefits in a programme of vaccination for adult women. The difference in GMTs between the two vaccines has been noted previously.S,11,12 Long-term follow-up studies have suggested that these differences disappear within 1-4 yçars,12-14 although only one study included women vaccinated post partum. 14 However, a study of young women vaccinated 6-8 years previously (not post partum) suggested that persistence of antibodies was greater after RA 27/3 than after Cendehill.ls Although our sample of women whose antibody status was assessed at the time of a subsequent pregnancy is small, the findings confirm previous reports that the response to Cendehill may take longer to peak and that reinfection with wild virus may be more frequent than with RA 27/3. Our findings also confirm that initial high titres tend to fall back during the following months. However, despite these flunctuations in antibody titres over time, the difference in the seroconversion rates persists. The continued use of Cendehill vaccine ’in women post partum should therefore be reviewed and its replacement with RA 27/3 vaccine considered. We thank the medical and nursing staff of the John Radcliffe Maternity Hospital for their help and cooperation in this study, the women who took part, and Mrs E. Coles for typing the manuscript.

Correspondence should be addressed to N. A. B., Oxfordshire Health Authority, Manor House, Off Headley Way, Headmgton, Oxford.

EARLY INTRODUCTION OF COPPER-CONTAMINATED ANIMAL MILK FEEDS AS A POSSIBLE CAUSE OF INDIAN CHILDHOOD CIRRHOSIS M. S. TANNER S. A. BHAVE

Department of Child Health, University of Leicester, Clinical Sciences Building, Leicester Royal Infirmary; and Department of Paediatrics, King Edward Memorial Hospital, Rasta Peth, Pune 411 011, India Brass and copper household utensils are a possible source of the gross hepatic copper accumulation characteristic of Indian childhood cirrhosis (ICC). In 107 families with a child with ICC, the use of copper or brass for water storage (97%) and of brass for milk storage (90%) or milk boiling (67%) resembled that of neighbouring village control families. However, the feeding history of 132 children with ICC differed from that of 70 children with other hepatic disorders and 311 children in 2 rural control groups. No child with ICC was exclusively breast-fed whereas 10%, 32%, and 25% of the control children were. Duration of breast-feeding was shorter in children with ICC and animal milk was introduced earlier. 57% were started on animal milk before 3 months of age. The age at introduction of animal milk correlated with the age at presentation with ICC. 66 urban Pune children had a feeding history similar to those with ICC, but in these families brass vessels were not used for milk. Experimentally, milk took up copper from utensils more avidly than water did. The copper concentration in milk samples obtained from ICC households, and those obtained experimentally, would supply a copper intake 6 to 20 times greater than that of the breast-fed infant, and similar to that producing copper toxicosis in the lamb. The hypothesis that early introduction of copper contaminated animal milk is of aetiological importance explains many of the epidemiological features of ICC.

Summary

Introduction

REFERENCES

JO’H. Rubella vaccination of post-partum women and of adolescents in the North West of England. Can J Public Health (Monograph suppl) 1971; 62: 64-67. 2. Sharpe DS, MacDonald H. Use of medroxyprogesterone acetate as a contraceptive in conjunction with early post partum rubella vaccination. Br Med J 1973; iv: 443-46. 3. Fox JP, Rainey HS, Hall CG, Ray CG, Patterson MJ. Rubella vaccine in post-pubertal women: experience in Western Washington State. JAMA 1976; 236: 837-43. 4. Menser MA, Forrest JM, Bransby RD, Collins E. Rubella vaccination in Australia. 2 Experience with the RA/27 rubella vaccine and results of a double blind trial in 1. Tobin

schoolgirls. Med J Aust 1978; ii: 85-88. 5. Just M, Berger-Hernandez R, Burgin-Wolff A. Serum antibodies 9 years after Cendehill rubella immunisation. Lancet 1977; ii: 1349-50. 6. Editorial. Rubella: who needs a blood test? Lancet 1979; i: 1329-31. 7. Ingalls TH. Rubella: which vaccine? Lancet 1979; ii: 791. 8. Grillner L, Hedstrom CE, Berstrom H, Forssman L, Rigner A, Lycke E. Vaccination against rubella of newly delivered women. Scand J Infect Dis 1973; 5: 237. 9. Kurtz JB, Mortimer PP, Mortimer PR, Morgan-Capner P, Shafi MS, White GBB. Rubella antibody measured by radial haemolysis. Characteristics and performance of a simple screening method for use in diagnostic laboratories. J Hyg(Camb) 1980; 84: 213-22. 10. Public Health Laboratory Service. 11.

12.

13. Black

14.

15.

Haemagglutination test for the detection of rubella

antibody. J Hyg (Camb) 1978; 81: 373-82. Herrmann KL, Halstead SB, Brandling-Bennett AD, Witte JJ, Wiebenga NH, Eddins DL. Rubella immunization. Persistence of antibody four years after a large-scale field trial. JAMA 1976; 235: 2201-04. Zealley H. Rubella screening and immunization of schoolgirls. A long term evaluation. Br J Prevent Soc Med 1974; 28: 54-59. FL, Lamm SH, Emmons JE, Pinheiro FP. Reactions to rubella vaccine and persistence of antibody in virgin-soil population after vaccination and wild-virus induced immunization. J Infect Dis 1976; 133; 393-98. MacDonald H, Tobin JO’H, Craddock-Watson JE, Lomax J. Antibody litres in women six to eight years after the administration of RA/3 and Cendehill rubella vaccines. J Hyg (Camb) 1978; 80: 337-47. O’Shea S, Best JM, Banatvala JE, Marshall WC, Dugeon JA. Rubella vaccination: J persistence of antibodies for up to 16 years. Br Med 1982; 285: 253-55.

A. H. KANTARJIAN A. N. PANDIT

INDIAN childhood cirrhosis (ICC) is a fatal disorder of children aged 1 to 3, which occurs predominantly in rural middle-income Hindu families throughout India. 1-3 Very high hepatic copper concentrations have been demonstrated in ICC histologically,4-8 by atomic absorption spectrophotometry9-12 and by electron probe analysis.13 An increase of dietary copper caused by the use of copper and brass utensils has been suggested 14-16 but not confirmed. This prompted a study of feeding patterns in children with ICC, children with other liver disorders, and healthy children from three communities.

Methods

During a prospective study of chronic liver disease in children in (formerly Poona),17,18 132 children with ICC (101 boys, 31 girls) were compared with 70 children with other liver disorders (49 boys, 21 girls). A feeding history was obtained by the admitting paediatrician. A home visit and second interview were performed by a social worker, who also examined the house, water supply, and utensils used for cooking and storage. Three groups of control Pune

homes and healthy children were similarly assessed: 50 houses around the home of 1 case of ICC seen in the Narayan Peth area of Pune; 35 houses in the vicinity of Miraj, 150 km south of Pune where the incidence of ICC is reportedly low and in which no cases were seen; and 5-10 homes in the vicinity of each of 18 cases ofICC