AMERICAN JOURNAL OF EPIDEMIOLOGY
Vol. 104, No. 2
Copyright © 1976 by The Johns Hopkins University School of Hygiene and Public Health
Printed in U.S.A.
VACCINE-ASSOCIATED POLIOMYELITIS IN THE UNITED STATES, 1961-19721 LAWRENCE B. SCHONBERGER,2 JOHN E. McGOWAN, JR., AND MICHAEL B. GREGG Schonberger, L. B., J. E. McGowan, Jr.. and M. B. Gregg (CDC, Atlanta, GA 30333). Vaccine-associated poliomyelitis in the United States, 1961-1972. Am J Epidemiol 104:202-211, 1976. Although poliovirus vaccines have led to a dramatic reduction in the incidence of poliomyelitis in the United States, there is evidence that vaccine-related cases have occurred in both vaccine recipients (recipient cases) and their contacts (contact cases) and that the incidence and epidemiologic characteristics of these cases in the 12-year period 1961-1972 have changed. Before 1965, there were 63 recipient cases compared with only 16 between 1965 and 1972. These latter cases were in younger persons and were more frequently associated with trivalent oral poliovirus vaccine (TOPV) than the former cases. Both groups were predominantly male and had intervals between immunization and onset of illness that clustered between 7 and 21 days. Only three contact cases were recognized before 1965, compared with 39 cases in 1965-1972. The increase was in both children and adults. Unlike recipient cases, contact cases were almost as frequent in females as in males. Evidence that the contact cases were vaccine-related included the apparent clustering between 20 and 29 days of the intervals from vaccine administration to onset of illness and the significantly increased frequency of longer intervals for contact cases in non-household members compared with contact cases in household members. The decrease in recipient cases and the increase in contact cases in 1965-1972 compared with the previous 3-year period are explained by the general curtailing of routine vaccination of adults after 1964, a switch from monovalent oral polio vaccine to TOPV, improved recognition of contact cases and a shift in emphasis from mass vaccination campaigns and communitywide programs to routine vaccination of infants. Immunization; poliomyelitis; poliovirus vaccine, oral; vaccination; vaccines
tenuated oral poliovirus vaccine (OPV) Widespread use of inactivated poliovirus ^suited in virtually comvaccine (IPV) after 1955 and of live, at- P l e t e c o n t r o 1 of poliomyelitis in the United States. The number of reported paralytic Received for publication November 17, 1975, and cases dropped from 13,850 in 1955 to 829 in infinalform February 3, 1976. 1961 and declined further to fewer than 35 Abbreviations: AC0PV, Advisory Committee on j 1969-1972 v(1). F J INTRODUCTION
after 1961 h a s
Oral Polio Vaccine; IPV, inactivated poliovirus vaccine; MOPV, monovalent oral polio vaccine; OPV, oral poliovirus vaccine; TOPV, trivalent oral polio
' Although poliovirus vaccines h a v e led to a d r a m a t i c reduction in t h e incidence of
'From the Viral Diseases Division, Bureau of Epidemiology, Center for Disease Control, US Public Health Service, US Department of Health, Education and Welfare, Atlanta, GA 30333. 2 Reprint Requests to Dr. Schonberger, Enteric and Neurotropic Viral Diseases Branch, Bureau of Epidemiology, Center for Disease Control, Atlanta, GA 30333.
The authors gratefully acknowledge the valuable consultation and laboratory support in the National Poliomyelitis Surveillance System of Dr. Milford H. Hatch, Chief of the Enteric Virology Division, and Dr. James H. N'akano, Chief of the Viral Exanthem Division, Bureau of Laboratories, Center for Disease Control.
202
203
VACCINE-ASSOCIATED POLIOMYELITIS
polio, their use has not been completely without risk (2). Since 1961, some of the polio cases reported to the Center for Disease Control (CDC) occurred in temporal association with administration of OPV. Cases have been recognized in persons who recently received the vaccine and in persons with a history of close contact with a recent recipient (3). Review of these cases for the period 1961-1972 reveals that epidemiologic characteristics have changed greatly. This paper summarizes these changes and examines available data on the incidence of paralytic disease and its association with the type and amount of OPV distributed. METHODS
Sources of data. Data were derived from CDC's National Poliomyelitis Surveillance System, which has previously been described (4). In addition, information is included on six cases of vaccine-associated paralytic polio not officially reported to CDC. Viral isolation procedures and serologic tests were performed by individual state laboratories or by the Enteric Virology Division, Bureau of Laboratories, CDC. Generic characterization studies on poliovirus isolates, including use of temperature markers, modified Wecker tests, and McBride kinetic neutralization tests, were also done at CDC (5-7). Cases in recent OPV recipients (recipient cases). In July 1964, reported cases in which onset of illness occurred within 30 days after ingestion of OPV were reviewed by a special Advisory Committee on Oral Polio Vaccine (ACOPV). A total of 57 cases between 1961 and July 1964 were classified as compatible with vaccine induction on the basis of the following criteria (8, 9): 1) Significant residual lower motor neuron paralysis. 2) No evidence of upper motor neuron disease, definite sensory loss, or progression or recurrence of paralytic illness one month or more after onset. 3) Onset of illness between 4 and 30 days
following feeding of the specific vaccine in question, with onset of paralysis not sooner than 6 days after the feeding. 4) Occurrence in an area where polio was not epidemic. 5) Laboratory data not inconsistent with respect to multiplication of the vaccine virus fed. Since July 1964, CDC has continued to use these criteria to classify recipient cases. Cases in contacts of OPV recipients (contact cases). Polio cases in close contacts of OPV recipients have also been reviewed as part of the polio surveillance program. A case of this type is considered compatible with vaccine induction when criteria 1, 2, 4, and 5 above for recipient cases are met, plus: 1) Onset of illness between 4 and 60 days and onset of paralysis 6 to 60 days after ingestion of the specific vaccine by the contact. 2) Contact between the case and the recipient within 30 days before the onset of illness. RESULTS
Recipient cases Review of possible vaccine-associated cases by the ACOPV placed 57 compatible cases in this category. After that review in July 1964, information was obtained on an additional six cases with onset of illness before 1965 and 16 more cases in 1965-1972, bringing the total number of cases to 79 for 1961-1972. Cases have been reported from 32 states; however, between 1965 and 1972, only three states reported more than a single case, and no state reported more than three cases. A total of 44 (77 per cent) of the recipient cases reviewed by the ACOPV in 1964 occurred in persons over 14 years of age. Because of this, vaccination recommendations were revised in 1964 to exclude routine vaccination of persons "over school age (about 18 years)." Since 1965, only 2 of the 16 recipient cases were in persons over
204
SCHONBERGER, McGOWAN AND GREGG
14 years of age (table 1). Although the age distribution has changed, the strong preponderance of males (81 per cent) which was noted for cases in 1961-1964 (9) has continued. Cases in the period 1961-1964 were most frequently associated with monovalent oral polio vaccine (MOPV) types 1 and 3; only 8 per cent of the cases in this period were associated with administration of trivalent oral polio vaccine (TOPV). By contrast, more than half the recipient cases in the period 1965-1972 were associated with the administration of TOPV. The median interval between vaccine administration and onset of illness was 14 days; 62 of the patients (78 per cent) became ill 7 to 21 days after vaccine administration (figure 1). Almost 30 per cent of the recipient cases reviewed by the ACOPV in 1964 had received three or more injections of IPV, and about two-thirds had received no IPV. Of the 22 recipient cases reported since that time, none had received three or more injections of IPV, and 19 had received no IPV. Three of these latter cases, however, became ill within 30
days after receiving their second dose of TOPV. No seasonal variation was noted for the recipient cases. For 17 of the 22 recipient cases reported since July 1964, culture specimens (predominantly stool samples) were collected within four weeks after onset of illness. At least one poliovirus type was isolated in 15 of these 17 cases; three cases had poliovirus type 1, three had poliovirus type 2, seven had poliovirus type 3, one had both poliovirus types 2 and 3 and one case had all three poliovirus types. No non-polio enteroviruses were isolated from any of these patients. Generic characterization studies were performed on isolates from 14 cases. All were "vaccine-like" strains, that is, were probably vaccine-derived. Paired serum specimens were tested for 10 of the recipient cases. A diagnostic fourfold rise in poliovirus titer of poliovirus neutralizing and/or complement-fixing antibody was demonstrated in three cases. In three cases, convalescent titers of less than 1:8 were reported despite the isolation of vaccine-
TABLE l
Cases of paralytic poliomyelitis in OPV recipients by age, sex,,and type of OPV received, United States, 1961-1972 Vaccine received U fTA / 1PAQPC 1
rt e
B
vyearsj
1961-1964 0-4 5-9 10-14 15-19 20-29 30-39 40+ Total Sex 1965-1972 0-4 15-19 40+ Total
MOPV-1
MOPV-2
2
2t 1 4 1 2 4 16 11M:5F
1
1 2 2M:0F
TOPV
MOPV-3
6 1 1 5 6 11 10 40 32M:8F
1
6
1
6
1
1 3 5 4M:1F
7 1 1 9
Total
9 4 2 9 7 14 18 63*
O C A
6M:3F 3M:1F 0M-.2F 8M:1F 6M:1F 13M:1F 13M:5F 49M:14F
14 1 1 16*
* Three cases in 1963 and one case each in 1971 and 1972 were not officially reported. t Type 3 poliovirus was implicated also in one case.
11M:3F 1M:OF 1M:OF 13M:3F
205
VACCINE-ASSOCIATED POLIOMYELITIS 14
VACCINE RECEIVED ^ MOPV-I ( 3 MOPV-2 Q MOPV-3
1961 — 1964
12
0 TOPV
108
^
6 4 2 0
3'
6
9
12
15
18
21
24
27
30
PERIOD BETWEEN VACCINATION AND ONSET! 3-DAY PERIODS )
FIGURE 1. Recipient cases of poliomyelitis, by period between vaccination and onset of illness, United States, 1961-1964 and 1965-1972. * Excludes one agammaglobulinemic patient with a probable 40-day interval (MOPV-1) or a possible 5-day interval (MOPV-3).
like poliovirus from appropriate stool specimens collected 16, 23 or 44 days after ingestion of vaccine. In two of these cases, complement-fixation tests only were reported. The third case was found to be agammaglobulinemic, the only immunologically deficient recipient case reported in the period. Contact cases Before 1965, three cases of paralytic polio in contacts of recent OPV recipients were classified as compatible with induction by the vaccine. Two were in household contacts (persons who regularly share the same home for sleeping) and one was in a next-door neighbor. In the period 1965-1972, 39 additional cases were identified, 28 in household contacts, and 11 in non-household contacts. The 42 patients resided in 19 states and the District of Columbia. Only three states reported more than three cases, and no state reported more than five cases. Unlike the recipient cases, contact cases
were almost as frequent in females as in males (table 2). Cases occurred either in children four months to eight years of age (43 per cent) or in adults 18 to 43 years of age (57 per cent) with no cases in persons aged 9 to 17. In approximately 80 per cent of these cases, the vaccine administered to the contact was TOPV. Median interval between vaccine administration and onset of illness for contact cases was 26 days (figure 2); half of the contact cases became ill 20 to 29 days after the implicated vaccine administration. Seventy-nine per cent of the 30 patients who were household contacts of the vaccinee became ill within 30 days after vaccine administration, while only 5 of 12 non-household contacts became ill within 30 days after vaccination (p < .05 by chi-square determination). Twenty-six of the contact cases had received no previous vaccination against polio, and four patients had received one dose of MOPV of a different type than that implicated in their illness. Five patients
206
SCHONBERGER, McGOWAN AND GREGG TABLE 2
Cases of paralytic polio in contacts of recent OPV recipients by age, sex, and type of OPV given to recipients. United States, 1961-1972 Age (years) 0-4 5-9 10-14 15-19 20-29 30-39 40+ Total
OPV received MOPV-1
M.OPV 3
MOPV-2
TOPV
Total
Unknown
Sex
9 3
15 3
8M:7F 1M:2F
1
1 8 11
3
33
1 10 12 1 42
1M:OF 4M:6F 8M:4F 0M:lF 22M:20F
2
It
* MOPV-3 also implicated in one case. t One case was not officially reported. 1961-1964
1965-1972
E3 M O P V 0 MOPV-2
•
MOPV-3
0 TOPV
• UNKNOWN 6-
••>•»••
42 0 5
—i—
10
15
20
W-f>;:::::i{^:.:-»-. 25
30
35
40
45
50
55
60
PERIOD BETWEEN VACCINATION OF RECIPIENT AND ONSET (5-DAYPERIODS)
FIGURE 2. Contact cases of poliomyelitis, by period between vaccination and onset of illness, United States, 1961-1964 and 1965-1972. * A close contact received MOPV-3 also 18 days prior to onset; Type 1 poliovirus was isolated from the patient.
had received at least three doses of IPV, and one patient had received all three types of MOPV. Another patient had previously ingested four doses of TOPV; this patient was one of the total of three contact cases with a known immunologic deficiency. The remaining contact cases had received some vaccinations against polio but less than three doses of IPV or OPV. As with the classic seasonal increase of "wild" poliovirus disease in May through October, almost 70 per cent of the contact cases for the period 1965-1972 occurred in these six months (figure 3).
Specimens for viral culture (predominantly stool samples) were collected within four weeks after onset of illness in 40 of the 42 contact cases; a single poliovirus type was isolated from 35; 34 of these were characterized antigenically, and 30 were "vaccine-like." Coxsackievirus A4 was isolated from the stool of one patient in addition to "vaccine-like" poliovirus type 3. No other non-polio enteroviruses were isolated. Poliovirus type 3 was isolated directly from postmortem brain and spinal cord specimens in one case with thymic abnormality (10). From the stool of another
VACCINE-ASSOCIATED POLIOMYELITIS
5Li 3
3 2 I J
F
M
A
M
J
J
A
S
0
N
0
FIGURE 3. Contact cases of paralytic polio by
month of onset, United States, 1965-1972.
207
MOPV or through TOPV administration) but was incriminated in only three of 13 contact cases in which the recipient previously had not ingested all three types of attenuated poliovirus (p < .025). Twentytwo contact cases could not be included in this analysis because no single poliovirus type was incriminated or because the vaccination history of the recipient was unreported. Age and sex of the single vaccine recipients associated with 29 and 28 contact cases, respectively, were reported. Twenty of the recipients were under one year of age, 25 were under three years of age, and no recipient was older than eight years. Eighteen were boys, and 10 were girls. Of the 24 contact cases in adults, 23 were associated with vaccine recipients living in the same household. By contrast, only seven of the 18 cases in children were associated with vaccine administration to a household contact. Furthermore, household cases were associated with recipients under one year of age significantly more often than non-household cases (p < .05).
patient, both "vaccine-like" poliovirus types 1 and 3 were isolated. Paired serum specimens were obtained from 35 of the contact cases; 21 showed fourfold changes in titer of neutralizing and/or complement-fixing antibody; no case had a convalescent neutralizing antibody titer below 1:40. A poliovirus type-specific etiology was incriminated for 40 of the 42 contact cases by viral isolation (35 cases) and/or by an elevated convalescent serum titer of neutralizing or complement-fixing antibody to only a single poliovirus type (five cases). Ratio of cases to doses of vaccine Though such incrimination may not condistributed stitute proof of an etiologic role, it is The experience of recipients and their noteworthy that the incriminated poliocontacts with respect to developing vacvirus type was related to the interval becine-associated paralytic disease can be tween vaccine administration and onset of illness: of the eight cases in which the onset TABLE 3 of illness occurred 20 days or less after the Incriminated poliovirus types* in contact cases by interval between onset of illness and vaccine contact received vaccine, type 2 poliovirus administration to recipient for contact. was incriminated in six and type 3 in none. United States, 1961-1972 Of the six cases for which the interval Incriminated poliovirus type between vaccine administration and onset Interval (in days) was 41 days or more, type 3 poliovirus was 1 2 Unknown 3 incriminated in four and type 2 in none 1 0 1 8 4-20 6 (table 3). 11 4 12 1 28 21-40 The incriminated poliovirus type was 0 2 4 0 6 41-60 also related to the vaccination history of 17 Total 2 42 7 16 the vaccine recipient. Type 3 poliovirus * Incriminated by isolation of the poliovirus from was incriminated in six of seven contact stool predominantly (35 cases) and/or by an elevated cases in which the recipient previously convalescent serum titer of neutralizing or compleingested all three types of poliovirus (either ment-fixing antibody to only a single poliovirus type through sequential administration of (5 cases).
208
SCHONBERGER, McGOWAN AND GREGG
expressed in terms of ratios of cases per million doses of vaccine distributed (table 4). These statistics are useful for comparing, trends. Because vaccination recommendations were changed in 1964 to exclude routine vaccination of adults, separate ratios have been calculated for the period 1961-1964 and 1965-1972. During the period 1961-1964, MOPV distribution greatly exceeded that for TOPV. By contrast, in the period 1965-1972, almost 89 per cent of the OPV distributed was TOPV. For recipient cases, lowest ratios were noted for MOPV type 2, and ratios for this vaccine and for MOPV type 1 remained relatively constant in the two periods shown. By contrast, the ratio noted for MOPV type 3 rose from 0.38 to 0.80 cases per million doses, and the ratio for TOPV declined from 0.18 to 0.05. For all OPV combined, as shown in the bottom two lines of table 4, the ratio of recipientassociated cases has decreased from 0.18 to 0.08. Ratios calculated in similar fashion for contact cases are also compared for the period before and after the change in
vaccination recommendations (table 4). Overall, the ratio for occurrence of contact cases is lower than that for recipient cases for MOPV types 1 and 3, equally low for MOPV type 2, and higher for TOPV. For all OPV combined, the ratio for vaccineassociated polio in contacts has risen sharply for the period 1965-1972 in comparison with the previous three-year period, with a ratio in the latter period of approximately two cases to 10 million doses of vaccine distributed. DISCUSSION
The temporal association of a case of paralytic polio with the administration of OPV has been observed in a large number of countries, and relevant data were reviewed in 1969 by a special committee sponsored by the World Health Organization. The committee concluded in its memorandum that "the data presented here, taken as a whole, provide evidence that vaccine-related cases of poliomyelitis have occurred" (11). That the cases reported in the present article represent more than just an incidental association between onset of
TABLE 4
Ratios of vaccine-associated paralytic polio to doses of OPV distributed, United States, 1961-1972* Vaccine MOPV-1 MOPV-2 MOPV-3 All MOPV TOPV All OPV
Period
Estimated doses distributed in millions!
Recipient cases
Recipient cases: million doses
Contact cases
Contact cases: million doses
1961-64 1965-72 1961-64 1965-72 1961-64 1965-72 1961-64 1965-72 1961-64 1965-72 1961-64 1965-72
109' 8.76 104 6.96 105 7.46 318 23.2 28.2 182 346 205
16t
0.147 0.114 0.019 0 0.381 0.804 0.182 0.302 0.177 0.049 0.182 0.078
2 0 2 2 0 3 4 0 33 3 39$
It
0.009 0.228 0 0.287 0.019 0 0.009 0.172 0 0.181 0.009 0.190
1 2 0 40 6 58 7 5 9 63 16
* Three recipient cases in 1963, one recipient case each in 1971 and 1972 and one contact case in 1965 were not officially reported. t Type 3 poliovirus implicated also in one case. % Includes two cases for whom type of vaccine administered to recipient is unknown. § Sources of distribution data: State health departments and PHS regional office before June 1962 and the Biologic Surveillance Unit, CDC, afterwards.
VACCINE-ASSOCIATED POLIOMYELITIS
illness and administration of OPV is suggested by the apparent cluster of recipient cases with onset occurring 7 to 21 days after vaccine administration (figure 1), the grouping of contact cases with intervals between vaccine administration and onset of illness from 20 to 29 days (figure 2), and the more frequent longer intervals between vaccine administration and onset of illness for contact cases in non-household members. There are a number of defects in a surveillance program of this type. First, the time interval used to define compatible cases is derived largely from poliovirus excretion patterns and histories of "wild" poliovirus disease following exposure in people without immunologic deficiencies. Not all cases of vaccine-induced disease may fit this arbitrary definition. A hypogammaglobulinemic infant from New York in 1972, for example, contracted possible vaccine-induced polio 75 days after his second dose of TOPV. In addition, a number of cases of polio have been reported in which "vaccine-like" poliovirus has been recovered from stool culture, but for which there is no history of recent vaccine ingestion or contact with recent vaccinees. These cases also are excluded from the above analysis. Third, the apparent increased occurrence of contact cases in May through October, the same period in which enteroviral activity is greatest in the United States, suggests that some of the contact cases may actually have been due to infection with wild poliovirus or to non-polio enteroviruses. It must be remembered, however, that factors responsible for seasonal variation in enteroviral disease are not well understood. Transmission of both wild and vaccine strains of poliovirus may be enhanced during this period. These and many other problems in recognizing and reporting vaccine-associated polio, combined with the absence of satisfactory estimates of the number of persons who actually received polio vaccine or of the number of susceptible people in con-
209
tact with vaccine recipients, prevent an accurate determination of the risk that recipients and their contacts will develop vaccine-associated polio. Nevertheless, the figures presented here indicate a number of trends: 1) Reported recipient cases declined and contact cases increased during the period 1965-1972 compared with the period 1961-1964. 2) Within both of these periods, the comparable ratio of contact cases to each given type of vaccine was similar; that is, the number of contact cases depended much more on the period in which a vaccine was used than on its type. 3) The increase in contact cases was marked for both children and adults. The decrease in ratio of recipient cases to doses of vaccine distributed may be due for the most part to the general curtailing in late 1964 of routine vaccination of adults against poliomyelitis, as 48 of the 63 recipient cases before 1965 were adults, compared with 2 of the 16 recipient cases reported in the period 1965-1972. A second factor may be the change to TOPV as the major immunizing agent used. In the period 1965-1972, a marked difference was seen between the ratio of recipient cases to distribution of TOPV and with both the comparable ratios for distribution of all MOPV types and particularly for distribution of MOPV type 3. The greater than 16-fold difference between these recipient case ratios associated with distribution of TOPV and MOPV type 3 supports a hypothesis that the effect of poliovirus type 3 vaccine is modified by the simultaneous administration of poliovirus types 1 and 2 in the trivalent vaccine. This hypothesis is supported further by the fact that poliovirus type 3 was not incriminated by culture and/or antibody response in most of the TOPV recipient cases. It must be remembered, however, that TOPV, unlike MOPV type 3, has often been the sole agent used for the primary vaccination series as well as for booster doses. Consequently, at least part of the difference in recipient case ratios for TOPV and MOPV
210
type 3 may be explained by a greater previous exposure and immunity to poliovirus type 3 vaccine expected among recipients of TOPV. A suggested third factor in the decreased ratio of recipient cases to doses of vaccine distributed is the introduction after 1964 of a possibly less virulent type 3 poliovirus. That this is not a major factor, however, is suggested first by the continued increased ratio (.804) of recipient cases to million doses of MOPV type 3 distributed after 1964; four of the six cases in this ratio occurred after 1966. Second, type 3 poliovirus was incriminated in only one of five TOPV recipient cases that occurred before 1965; the ratio of this type 3 case to the millions of doses of TOPV distributed was only .035. In 1964, the ACOPV recommended that the order of administration of MOPV be changed from types 1, 3, and 2 to types 2,1, and 3. The reason offered for this change was that "giving type 2 vaccine first should theoretically further diminish the slight risks associated with the other vaccine types" (8). It is thus noteworthy that during the period 1965-1972 none of the seven recipient cases following administration of MOPV had been vaccinated in accordance with these revised recommendations; they had never received MOPV type 2. It is possible that the increase in reported contact cases resulted both from the shift in emphasis from mass vaccination campaigns to routine vaccination of infants and from a presumed improvement in recognition of contact cases, an improvement which would be facilitated by the reduction in total polio cases following the 1961-1964 polio vaccination campaigns. The shift from mass campaigns would increase the number of susceptibles in contact with each case. During mass campaigns, for instance, many contacts of a given recipient would themselves have received vaccine. This would either protect them from becoming infected with another
poliovirus, or if not, and paralysis occurred, could lead to their being classified as recipient rather than contact cases. The shift to vaccination of infants is apparent from the progressive increase since 1964 in the percentage of infants under one year of age who have received at least one dose of OPV (12-15). As children under two years of age appear to spread vaccine poliovirus to their contacts more readily than older children (16), an increase in the percentage of infants receiving OPV may increase the likelihood of spread to contacts. Indeed, almost 76 per cent of the associated recipients in contact cases were under two years of age, and contact cases have occurred in precisely the segments of the population most intimately in contact with young children: adults and older children among household contacts, and children of approximately the same age among non-household contacts. REFERENCES
1. Center for Disease Control: Neurotropic Diseases Surveillance—Poliomyelitis, Annual Poliomyelitis Summary 1972, Atlanta, Georgia, 1974 2. Nathanson N, Langmuir AD: The Cutter incident. Am J Hyg 78:16-81, 1963 3. Hopkins CC, Dismukes WE, Glick TH, et al: Surveillance of paralytic poliomyelitis in the United States. JAMA 210:694-700, 1969 4. Morris L, Witte JJ, Gardiner P, et al: Surveillance of poliomyelitis in the United States, 1962-1965. Pub Health Rep 82:417-428, 1967 5. Nakano JH, Gelfand HM: The use of a modified Wecker technique for the serodifferentiation of Type 1 polioviruses related and unrelated to Sabin's vaccine strain: I. Standardization and evaluation of the test. Am J Hyg 75:363-376, 1962 6. Nakano JH, Gelfand HM, Cole JT: The use of a modified Wecker technique for serodifferentiation of Type 1 polioviruses related and unrelated to Sabin's vaccine strain: II. Antigenic segregation of isolates from specimens collected in field studies. Am J Hyg 78:214-226, 1963 7. Nakano JH, Gelfand HM, Cole JT: Antigenic segregation of type 3 poliovirus isolates related and unrelated to Sabin's vaccine strain with use of modified Wecker and McBride techniques. Am J Epidemiol 83:130-145, 1966 8. Oral Poliomyelitis Vaccines: Report of special advisory committee on oral poliomyelitis vaccines to the Surgeon General of the Public Health Service. JAMA 190:49-51, 1964 9. Henderson DA, Witte JJ, Morris L, et al: Para-
211
10. 11.
12.
13.
lytic diseases associated with oral polio vaccines. JAMA 190:41-48, 1964 Riker JB, Brandt CD, Chandia R, et al: Vaccineassociated poliomyelitis in a child with thymic abnormality. Pediatrics 48:923-929, 1971 Evidence on the safety and efficacy of live poliomyelitis vaccines currently in use, with special reference to type 3 poliovirus. Bull WHO 40:925-942, 1969 United States Public Health Service: US Immunization Survey, 1962, Published in Supplement to Polio Surveillance Unit Report No. 276, Communicable Disease Center, Atlanta, Georgia, March 29, 1963 United States Public Health Service: US Immu-
nization Survey, 1964, Published in Supplement to Polio Surveillance Unit Report No. 287, Communicable Disease Center, Atlanta, Georgia, June 1, 1965 14. United States Public Health Service: US Immunization Survey, 1965, Published in Supplement to Polio Surveillance Unit Report No. 288, Communicable Disease Center, Atlanta, GA, June 1, 1966 15. Center for Disease Control: US Immunization Survey, 1971, Atlanta, GA, 1971, p 15 16. Gard, S: Poliomyelitis, Fifth International Poliomyelitis Conference, Copenhagen, Denmark, July 26-28, 1960. Philadelphia, PA, JB Lippincott Co, 1961, p 413
Vol. 104, No. 2
Copyright © 1976 by The Johns Hopkins University School of Hygiene and Public Health
Printed in U.S.A.
VACCINE-ASSOCIATED POLIOMYELITIS IN THE UNITED STATES, 1961-19721 LAWRENCE B. SCHONBERGER,2 JOHN E. McGOWAN, JR., AND MICHAEL B. GREGG Schonberger, L. B., J. E. McGowan, Jr.. and M. B. Gregg (CDC, Atlanta, GA 30333). Vaccine-associated poliomyelitis in the United States, 1961-1972. Am J Epidemiol 104:202-211, 1976. Although poliovirus vaccines have led to a dramatic reduction in the incidence of poliomyelitis in the United States, there is evidence that vaccine-related cases have occurred in both vaccine recipients (recipient cases) and their contacts (contact cases) and that the incidence and epidemiologic characteristics of these cases in the 12-year period 1961-1972 have changed. Before 1965, there were 63 recipient cases compared with only 16 between 1965 and 1972. These latter cases were in younger persons and were more frequently associated with trivalent oral poliovirus vaccine (TOPV) than the former cases. Both groups were predominantly male and had intervals between immunization and onset of illness that clustered between 7 and 21 days. Only three contact cases were recognized before 1965, compared with 39 cases in 1965-1972. The increase was in both children and adults. Unlike recipient cases, contact cases were almost as frequent in females as in males. Evidence that the contact cases were vaccine-related included the apparent clustering between 20 and 29 days of the intervals from vaccine administration to onset of illness and the significantly increased frequency of longer intervals for contact cases in non-household members compared with contact cases in household members. The decrease in recipient cases and the increase in contact cases in 1965-1972 compared with the previous 3-year period are explained by the general curtailing of routine vaccination of adults after 1964, a switch from monovalent oral polio vaccine to TOPV, improved recognition of contact cases and a shift in emphasis from mass vaccination campaigns and communitywide programs to routine vaccination of infants. Immunization; poliomyelitis; poliovirus vaccine, oral; vaccination; vaccines
tenuated oral poliovirus vaccine (OPV) Widespread use of inactivated poliovirus ^suited in virtually comvaccine (IPV) after 1955 and of live, at- P l e t e c o n t r o 1 of poliomyelitis in the United States. The number of reported paralytic Received for publication November 17, 1975, and cases dropped from 13,850 in 1955 to 829 in infinalform February 3, 1976. 1961 and declined further to fewer than 35 Abbreviations: AC0PV, Advisory Committee on j 1969-1972 v(1). F J INTRODUCTION
after 1961 h a s
Oral Polio Vaccine; IPV, inactivated poliovirus vaccine; MOPV, monovalent oral polio vaccine; OPV, oral poliovirus vaccine; TOPV, trivalent oral polio
' Although poliovirus vaccines h a v e led to a d r a m a t i c reduction in t h e incidence of
'From the Viral Diseases Division, Bureau of Epidemiology, Center for Disease Control, US Public Health Service, US Department of Health, Education and Welfare, Atlanta, GA 30333. 2 Reprint Requests to Dr. Schonberger, Enteric and Neurotropic Viral Diseases Branch, Bureau of Epidemiology, Center for Disease Control, Atlanta, GA 30333.
The authors gratefully acknowledge the valuable consultation and laboratory support in the National Poliomyelitis Surveillance System of Dr. Milford H. Hatch, Chief of the Enteric Virology Division, and Dr. James H. N'akano, Chief of the Viral Exanthem Division, Bureau of Laboratories, Center for Disease Control.
202
203
VACCINE-ASSOCIATED POLIOMYELITIS
polio, their use has not been completely without risk (2). Since 1961, some of the polio cases reported to the Center for Disease Control (CDC) occurred in temporal association with administration of OPV. Cases have been recognized in persons who recently received the vaccine and in persons with a history of close contact with a recent recipient (3). Review of these cases for the period 1961-1972 reveals that epidemiologic characteristics have changed greatly. This paper summarizes these changes and examines available data on the incidence of paralytic disease and its association with the type and amount of OPV distributed. METHODS
Sources of data. Data were derived from CDC's National Poliomyelitis Surveillance System, which has previously been described (4). In addition, information is included on six cases of vaccine-associated paralytic polio not officially reported to CDC. Viral isolation procedures and serologic tests were performed by individual state laboratories or by the Enteric Virology Division, Bureau of Laboratories, CDC. Generic characterization studies on poliovirus isolates, including use of temperature markers, modified Wecker tests, and McBride kinetic neutralization tests, were also done at CDC (5-7). Cases in recent OPV recipients (recipient cases). In July 1964, reported cases in which onset of illness occurred within 30 days after ingestion of OPV were reviewed by a special Advisory Committee on Oral Polio Vaccine (ACOPV). A total of 57 cases between 1961 and July 1964 were classified as compatible with vaccine induction on the basis of the following criteria (8, 9): 1) Significant residual lower motor neuron paralysis. 2) No evidence of upper motor neuron disease, definite sensory loss, or progression or recurrence of paralytic illness one month or more after onset. 3) Onset of illness between 4 and 30 days
following feeding of the specific vaccine in question, with onset of paralysis not sooner than 6 days after the feeding. 4) Occurrence in an area where polio was not epidemic. 5) Laboratory data not inconsistent with respect to multiplication of the vaccine virus fed. Since July 1964, CDC has continued to use these criteria to classify recipient cases. Cases in contacts of OPV recipients (contact cases). Polio cases in close contacts of OPV recipients have also been reviewed as part of the polio surveillance program. A case of this type is considered compatible with vaccine induction when criteria 1, 2, 4, and 5 above for recipient cases are met, plus: 1) Onset of illness between 4 and 60 days and onset of paralysis 6 to 60 days after ingestion of the specific vaccine by the contact. 2) Contact between the case and the recipient within 30 days before the onset of illness. RESULTS
Recipient cases Review of possible vaccine-associated cases by the ACOPV placed 57 compatible cases in this category. After that review in July 1964, information was obtained on an additional six cases with onset of illness before 1965 and 16 more cases in 1965-1972, bringing the total number of cases to 79 for 1961-1972. Cases have been reported from 32 states; however, between 1965 and 1972, only three states reported more than a single case, and no state reported more than three cases. A total of 44 (77 per cent) of the recipient cases reviewed by the ACOPV in 1964 occurred in persons over 14 years of age. Because of this, vaccination recommendations were revised in 1964 to exclude routine vaccination of persons "over school age (about 18 years)." Since 1965, only 2 of the 16 recipient cases were in persons over
204
SCHONBERGER, McGOWAN AND GREGG
14 years of age (table 1). Although the age distribution has changed, the strong preponderance of males (81 per cent) which was noted for cases in 1961-1964 (9) has continued. Cases in the period 1961-1964 were most frequently associated with monovalent oral polio vaccine (MOPV) types 1 and 3; only 8 per cent of the cases in this period were associated with administration of trivalent oral polio vaccine (TOPV). By contrast, more than half the recipient cases in the period 1965-1972 were associated with the administration of TOPV. The median interval between vaccine administration and onset of illness was 14 days; 62 of the patients (78 per cent) became ill 7 to 21 days after vaccine administration (figure 1). Almost 30 per cent of the recipient cases reviewed by the ACOPV in 1964 had received three or more injections of IPV, and about two-thirds had received no IPV. Of the 22 recipient cases reported since that time, none had received three or more injections of IPV, and 19 had received no IPV. Three of these latter cases, however, became ill within 30
days after receiving their second dose of TOPV. No seasonal variation was noted for the recipient cases. For 17 of the 22 recipient cases reported since July 1964, culture specimens (predominantly stool samples) were collected within four weeks after onset of illness. At least one poliovirus type was isolated in 15 of these 17 cases; three cases had poliovirus type 1, three had poliovirus type 2, seven had poliovirus type 3, one had both poliovirus types 2 and 3 and one case had all three poliovirus types. No non-polio enteroviruses were isolated from any of these patients. Generic characterization studies were performed on isolates from 14 cases. All were "vaccine-like" strains, that is, were probably vaccine-derived. Paired serum specimens were tested for 10 of the recipient cases. A diagnostic fourfold rise in poliovirus titer of poliovirus neutralizing and/or complement-fixing antibody was demonstrated in three cases. In three cases, convalescent titers of less than 1:8 were reported despite the isolation of vaccine-
TABLE l
Cases of paralytic poliomyelitis in OPV recipients by age, sex,,and type of OPV received, United States, 1961-1972 Vaccine received U fTA / 1PAQPC 1
rt e
B
vyearsj
1961-1964 0-4 5-9 10-14 15-19 20-29 30-39 40+ Total Sex 1965-1972 0-4 15-19 40+ Total
MOPV-1
MOPV-2
2
2t 1 4 1 2 4 16 11M:5F
1
1 2 2M:0F
TOPV
MOPV-3
6 1 1 5 6 11 10 40 32M:8F
1
6
1
6
1
1 3 5 4M:1F
7 1 1 9
Total
9 4 2 9 7 14 18 63*
O C A
6M:3F 3M:1F 0M-.2F 8M:1F 6M:1F 13M:1F 13M:5F 49M:14F
14 1 1 16*
* Three cases in 1963 and one case each in 1971 and 1972 were not officially reported. t Type 3 poliovirus was implicated also in one case.
11M:3F 1M:OF 1M:OF 13M:3F
205
VACCINE-ASSOCIATED POLIOMYELITIS 14
VACCINE RECEIVED ^ MOPV-I ( 3 MOPV-2 Q MOPV-3
1961 — 1964
12
0 TOPV
108
^
6 4 2 0
3'
6
9
12
15
18
21
24
27
30
PERIOD BETWEEN VACCINATION AND ONSET! 3-DAY PERIODS )
FIGURE 1. Recipient cases of poliomyelitis, by period between vaccination and onset of illness, United States, 1961-1964 and 1965-1972. * Excludes one agammaglobulinemic patient with a probable 40-day interval (MOPV-1) or a possible 5-day interval (MOPV-3).
like poliovirus from appropriate stool specimens collected 16, 23 or 44 days after ingestion of vaccine. In two of these cases, complement-fixation tests only were reported. The third case was found to be agammaglobulinemic, the only immunologically deficient recipient case reported in the period. Contact cases Before 1965, three cases of paralytic polio in contacts of recent OPV recipients were classified as compatible with induction by the vaccine. Two were in household contacts (persons who regularly share the same home for sleeping) and one was in a next-door neighbor. In the period 1965-1972, 39 additional cases were identified, 28 in household contacts, and 11 in non-household contacts. The 42 patients resided in 19 states and the District of Columbia. Only three states reported more than three cases, and no state reported more than five cases. Unlike the recipient cases, contact cases
were almost as frequent in females as in males (table 2). Cases occurred either in children four months to eight years of age (43 per cent) or in adults 18 to 43 years of age (57 per cent) with no cases in persons aged 9 to 17. In approximately 80 per cent of these cases, the vaccine administered to the contact was TOPV. Median interval between vaccine administration and onset of illness for contact cases was 26 days (figure 2); half of the contact cases became ill 20 to 29 days after the implicated vaccine administration. Seventy-nine per cent of the 30 patients who were household contacts of the vaccinee became ill within 30 days after vaccine administration, while only 5 of 12 non-household contacts became ill within 30 days after vaccination (p < .05 by chi-square determination). Twenty-six of the contact cases had received no previous vaccination against polio, and four patients had received one dose of MOPV of a different type than that implicated in their illness. Five patients
206
SCHONBERGER, McGOWAN AND GREGG TABLE 2
Cases of paralytic polio in contacts of recent OPV recipients by age, sex, and type of OPV given to recipients. United States, 1961-1972 Age (years) 0-4 5-9 10-14 15-19 20-29 30-39 40+ Total
OPV received MOPV-1
M.OPV 3
MOPV-2
TOPV
Total
Unknown
Sex
9 3
15 3
8M:7F 1M:2F
1
1 8 11
3
33
1 10 12 1 42
1M:OF 4M:6F 8M:4F 0M:lF 22M:20F
2
It
* MOPV-3 also implicated in one case. t One case was not officially reported. 1961-1964
1965-1972
E3 M O P V 0 MOPV-2
•
MOPV-3
0 TOPV
• UNKNOWN 6-
••>•»••
42 0 5
—i—
10
15
20
W-f>;:::::i{^:.:-»-. 25
30
35
40
45
50
55
60
PERIOD BETWEEN VACCINATION OF RECIPIENT AND ONSET (5-DAYPERIODS)
FIGURE 2. Contact cases of poliomyelitis, by period between vaccination and onset of illness, United States, 1961-1964 and 1965-1972. * A close contact received MOPV-3 also 18 days prior to onset; Type 1 poliovirus was isolated from the patient.
had received at least three doses of IPV, and one patient had received all three types of MOPV. Another patient had previously ingested four doses of TOPV; this patient was one of the total of three contact cases with a known immunologic deficiency. The remaining contact cases had received some vaccinations against polio but less than three doses of IPV or OPV. As with the classic seasonal increase of "wild" poliovirus disease in May through October, almost 70 per cent of the contact cases for the period 1965-1972 occurred in these six months (figure 3).
Specimens for viral culture (predominantly stool samples) were collected within four weeks after onset of illness in 40 of the 42 contact cases; a single poliovirus type was isolated from 35; 34 of these were characterized antigenically, and 30 were "vaccine-like." Coxsackievirus A4 was isolated from the stool of one patient in addition to "vaccine-like" poliovirus type 3. No other non-polio enteroviruses were isolated. Poliovirus type 3 was isolated directly from postmortem brain and spinal cord specimens in one case with thymic abnormality (10). From the stool of another
VACCINE-ASSOCIATED POLIOMYELITIS
5Li 3
3 2 I J
F
M
A
M
J
J
A
S
0
N
0
FIGURE 3. Contact cases of paralytic polio by
month of onset, United States, 1965-1972.
207
MOPV or through TOPV administration) but was incriminated in only three of 13 contact cases in which the recipient previously had not ingested all three types of attenuated poliovirus (p < .025). Twentytwo contact cases could not be included in this analysis because no single poliovirus type was incriminated or because the vaccination history of the recipient was unreported. Age and sex of the single vaccine recipients associated with 29 and 28 contact cases, respectively, were reported. Twenty of the recipients were under one year of age, 25 were under three years of age, and no recipient was older than eight years. Eighteen were boys, and 10 were girls. Of the 24 contact cases in adults, 23 were associated with vaccine recipients living in the same household. By contrast, only seven of the 18 cases in children were associated with vaccine administration to a household contact. Furthermore, household cases were associated with recipients under one year of age significantly more often than non-household cases (p < .05).
patient, both "vaccine-like" poliovirus types 1 and 3 were isolated. Paired serum specimens were obtained from 35 of the contact cases; 21 showed fourfold changes in titer of neutralizing and/or complement-fixing antibody; no case had a convalescent neutralizing antibody titer below 1:40. A poliovirus type-specific etiology was incriminated for 40 of the 42 contact cases by viral isolation (35 cases) and/or by an elevated convalescent serum titer of neutralizing or complement-fixing antibody to only a single poliovirus type (five cases). Ratio of cases to doses of vaccine Though such incrimination may not condistributed stitute proof of an etiologic role, it is The experience of recipients and their noteworthy that the incriminated poliocontacts with respect to developing vacvirus type was related to the interval becine-associated paralytic disease can be tween vaccine administration and onset of illness: of the eight cases in which the onset TABLE 3 of illness occurred 20 days or less after the Incriminated poliovirus types* in contact cases by interval between onset of illness and vaccine contact received vaccine, type 2 poliovirus administration to recipient for contact. was incriminated in six and type 3 in none. United States, 1961-1972 Of the six cases for which the interval Incriminated poliovirus type between vaccine administration and onset Interval (in days) was 41 days or more, type 3 poliovirus was 1 2 Unknown 3 incriminated in four and type 2 in none 1 0 1 8 4-20 6 (table 3). 11 4 12 1 28 21-40 The incriminated poliovirus type was 0 2 4 0 6 41-60 also related to the vaccination history of 17 Total 2 42 7 16 the vaccine recipient. Type 3 poliovirus * Incriminated by isolation of the poliovirus from was incriminated in six of seven contact stool predominantly (35 cases) and/or by an elevated cases in which the recipient previously convalescent serum titer of neutralizing or compleingested all three types of poliovirus (either ment-fixing antibody to only a single poliovirus type through sequential administration of (5 cases).
208
SCHONBERGER, McGOWAN AND GREGG
expressed in terms of ratios of cases per million doses of vaccine distributed (table 4). These statistics are useful for comparing, trends. Because vaccination recommendations were changed in 1964 to exclude routine vaccination of adults, separate ratios have been calculated for the period 1961-1964 and 1965-1972. During the period 1961-1964, MOPV distribution greatly exceeded that for TOPV. By contrast, in the period 1965-1972, almost 89 per cent of the OPV distributed was TOPV. For recipient cases, lowest ratios were noted for MOPV type 2, and ratios for this vaccine and for MOPV type 1 remained relatively constant in the two periods shown. By contrast, the ratio noted for MOPV type 3 rose from 0.38 to 0.80 cases per million doses, and the ratio for TOPV declined from 0.18 to 0.05. For all OPV combined, as shown in the bottom two lines of table 4, the ratio of recipientassociated cases has decreased from 0.18 to 0.08. Ratios calculated in similar fashion for contact cases are also compared for the period before and after the change in
vaccination recommendations (table 4). Overall, the ratio for occurrence of contact cases is lower than that for recipient cases for MOPV types 1 and 3, equally low for MOPV type 2, and higher for TOPV. For all OPV combined, the ratio for vaccineassociated polio in contacts has risen sharply for the period 1965-1972 in comparison with the previous three-year period, with a ratio in the latter period of approximately two cases to 10 million doses of vaccine distributed. DISCUSSION
The temporal association of a case of paralytic polio with the administration of OPV has been observed in a large number of countries, and relevant data were reviewed in 1969 by a special committee sponsored by the World Health Organization. The committee concluded in its memorandum that "the data presented here, taken as a whole, provide evidence that vaccine-related cases of poliomyelitis have occurred" (11). That the cases reported in the present article represent more than just an incidental association between onset of
TABLE 4
Ratios of vaccine-associated paralytic polio to doses of OPV distributed, United States, 1961-1972* Vaccine MOPV-1 MOPV-2 MOPV-3 All MOPV TOPV All OPV
Period
Estimated doses distributed in millions!
Recipient cases
Recipient cases: million doses
Contact cases
Contact cases: million doses
1961-64 1965-72 1961-64 1965-72 1961-64 1965-72 1961-64 1965-72 1961-64 1965-72 1961-64 1965-72
109' 8.76 104 6.96 105 7.46 318 23.2 28.2 182 346 205
16t
0.147 0.114 0.019 0 0.381 0.804 0.182 0.302 0.177 0.049 0.182 0.078
2 0 2 2 0 3 4 0 33 3 39$
It
0.009 0.228 0 0.287 0.019 0 0.009 0.172 0 0.181 0.009 0.190
1 2 0 40 6 58 7 5 9 63 16
* Three recipient cases in 1963, one recipient case each in 1971 and 1972 and one contact case in 1965 were not officially reported. t Type 3 poliovirus implicated also in one case. % Includes two cases for whom type of vaccine administered to recipient is unknown. § Sources of distribution data: State health departments and PHS regional office before June 1962 and the Biologic Surveillance Unit, CDC, afterwards.
VACCINE-ASSOCIATED POLIOMYELITIS
illness and administration of OPV is suggested by the apparent cluster of recipient cases with onset occurring 7 to 21 days after vaccine administration (figure 1), the grouping of contact cases with intervals between vaccine administration and onset of illness from 20 to 29 days (figure 2), and the more frequent longer intervals between vaccine administration and onset of illness for contact cases in non-household members. There are a number of defects in a surveillance program of this type. First, the time interval used to define compatible cases is derived largely from poliovirus excretion patterns and histories of "wild" poliovirus disease following exposure in people without immunologic deficiencies. Not all cases of vaccine-induced disease may fit this arbitrary definition. A hypogammaglobulinemic infant from New York in 1972, for example, contracted possible vaccine-induced polio 75 days after his second dose of TOPV. In addition, a number of cases of polio have been reported in which "vaccine-like" poliovirus has been recovered from stool culture, but for which there is no history of recent vaccine ingestion or contact with recent vaccinees. These cases also are excluded from the above analysis. Third, the apparent increased occurrence of contact cases in May through October, the same period in which enteroviral activity is greatest in the United States, suggests that some of the contact cases may actually have been due to infection with wild poliovirus or to non-polio enteroviruses. It must be remembered, however, that factors responsible for seasonal variation in enteroviral disease are not well understood. Transmission of both wild and vaccine strains of poliovirus may be enhanced during this period. These and many other problems in recognizing and reporting vaccine-associated polio, combined with the absence of satisfactory estimates of the number of persons who actually received polio vaccine or of the number of susceptible people in con-
209
tact with vaccine recipients, prevent an accurate determination of the risk that recipients and their contacts will develop vaccine-associated polio. Nevertheless, the figures presented here indicate a number of trends: 1) Reported recipient cases declined and contact cases increased during the period 1965-1972 compared with the period 1961-1964. 2) Within both of these periods, the comparable ratio of contact cases to each given type of vaccine was similar; that is, the number of contact cases depended much more on the period in which a vaccine was used than on its type. 3) The increase in contact cases was marked for both children and adults. The decrease in ratio of recipient cases to doses of vaccine distributed may be due for the most part to the general curtailing in late 1964 of routine vaccination of adults against poliomyelitis, as 48 of the 63 recipient cases before 1965 were adults, compared with 2 of the 16 recipient cases reported in the period 1965-1972. A second factor may be the change to TOPV as the major immunizing agent used. In the period 1965-1972, a marked difference was seen between the ratio of recipient cases to distribution of TOPV and with both the comparable ratios for distribution of all MOPV types and particularly for distribution of MOPV type 3. The greater than 16-fold difference between these recipient case ratios associated with distribution of TOPV and MOPV type 3 supports a hypothesis that the effect of poliovirus type 3 vaccine is modified by the simultaneous administration of poliovirus types 1 and 2 in the trivalent vaccine. This hypothesis is supported further by the fact that poliovirus type 3 was not incriminated by culture and/or antibody response in most of the TOPV recipient cases. It must be remembered, however, that TOPV, unlike MOPV type 3, has often been the sole agent used for the primary vaccination series as well as for booster doses. Consequently, at least part of the difference in recipient case ratios for TOPV and MOPV
210
type 3 may be explained by a greater previous exposure and immunity to poliovirus type 3 vaccine expected among recipients of TOPV. A suggested third factor in the decreased ratio of recipient cases to doses of vaccine distributed is the introduction after 1964 of a possibly less virulent type 3 poliovirus. That this is not a major factor, however, is suggested first by the continued increased ratio (.804) of recipient cases to million doses of MOPV type 3 distributed after 1964; four of the six cases in this ratio occurred after 1966. Second, type 3 poliovirus was incriminated in only one of five TOPV recipient cases that occurred before 1965; the ratio of this type 3 case to the millions of doses of TOPV distributed was only .035. In 1964, the ACOPV recommended that the order of administration of MOPV be changed from types 1, 3, and 2 to types 2,1, and 3. The reason offered for this change was that "giving type 2 vaccine first should theoretically further diminish the slight risks associated with the other vaccine types" (8). It is thus noteworthy that during the period 1965-1972 none of the seven recipient cases following administration of MOPV had been vaccinated in accordance with these revised recommendations; they had never received MOPV type 2. It is possible that the increase in reported contact cases resulted both from the shift in emphasis from mass vaccination campaigns to routine vaccination of infants and from a presumed improvement in recognition of contact cases, an improvement which would be facilitated by the reduction in total polio cases following the 1961-1964 polio vaccination campaigns. The shift from mass campaigns would increase the number of susceptibles in contact with each case. During mass campaigns, for instance, many contacts of a given recipient would themselves have received vaccine. This would either protect them from becoming infected with another
poliovirus, or if not, and paralysis occurred, could lead to their being classified as recipient rather than contact cases. The shift to vaccination of infants is apparent from the progressive increase since 1964 in the percentage of infants under one year of age who have received at least one dose of OPV (12-15). As children under two years of age appear to spread vaccine poliovirus to their contacts more readily than older children (16), an increase in the percentage of infants receiving OPV may increase the likelihood of spread to contacts. Indeed, almost 76 per cent of the associated recipients in contact cases were under two years of age, and contact cases have occurred in precisely the segments of the population most intimately in contact with young children: adults and older children among household contacts, and children of approximately the same age among non-household contacts. REFERENCES
1. Center for Disease Control: Neurotropic Diseases Surveillance—Poliomyelitis, Annual Poliomyelitis Summary 1972, Atlanta, Georgia, 1974 2. Nathanson N, Langmuir AD: The Cutter incident. Am J Hyg 78:16-81, 1963 3. Hopkins CC, Dismukes WE, Glick TH, et al: Surveillance of paralytic poliomyelitis in the United States. JAMA 210:694-700, 1969 4. Morris L, Witte JJ, Gardiner P, et al: Surveillance of poliomyelitis in the United States, 1962-1965. Pub Health Rep 82:417-428, 1967 5. Nakano JH, Gelfand HM: The use of a modified Wecker technique for the serodifferentiation of Type 1 polioviruses related and unrelated to Sabin's vaccine strain: I. Standardization and evaluation of the test. Am J Hyg 75:363-376, 1962 6. Nakano JH, Gelfand HM, Cole JT: The use of a modified Wecker technique for serodifferentiation of Type 1 polioviruses related and unrelated to Sabin's vaccine strain: II. Antigenic segregation of isolates from specimens collected in field studies. Am J Hyg 78:214-226, 1963 7. Nakano JH, Gelfand HM, Cole JT: Antigenic segregation of type 3 poliovirus isolates related and unrelated to Sabin's vaccine strain with use of modified Wecker and McBride techniques. Am J Epidemiol 83:130-145, 1966 8. Oral Poliomyelitis Vaccines: Report of special advisory committee on oral poliomyelitis vaccines to the Surgeon General of the Public Health Service. JAMA 190:49-51, 1964 9. Henderson DA, Witte JJ, Morris L, et al: Para-
211
10. 11.
12.
13.
lytic diseases associated with oral polio vaccines. JAMA 190:41-48, 1964 Riker JB, Brandt CD, Chandia R, et al: Vaccineassociated poliomyelitis in a child with thymic abnormality. Pediatrics 48:923-929, 1971 Evidence on the safety and efficacy of live poliomyelitis vaccines currently in use, with special reference to type 3 poliovirus. Bull WHO 40:925-942, 1969 United States Public Health Service: US Immunization Survey, 1962, Published in Supplement to Polio Surveillance Unit Report No. 276, Communicable Disease Center, Atlanta, Georgia, March 29, 1963 United States Public Health Service: US Immu-
nization Survey, 1964, Published in Supplement to Polio Surveillance Unit Report No. 287, Communicable Disease Center, Atlanta, Georgia, June 1, 1965 14. United States Public Health Service: US Immunization Survey, 1965, Published in Supplement to Polio Surveillance Unit Report No. 288, Communicable Disease Center, Atlanta, GA, June 1, 1966 15. Center for Disease Control: US Immunization Survey, 1971, Atlanta, GA, 1971, p 15 16. Gard, S: Poliomyelitis, Fifth International Poliomyelitis Conference, Copenhagen, Denmark, July 26-28, 1960. Philadelphia, PA, JB Lippincott Co, 1961, p 413
