Feb., 1952

The Relation Between Recent InjectionsAnd Paralytic Poliomyelitis in Children*

MORRIS GREENBERG, M.D., M.S.P.H., F.A.P.H.A., HAROLDABRAMSON, M.D., HELEN M. COOPER, AND

HELEN E. SOLOMONDirector, Epidemiologist, and Assistants, Bureau of Preventable Diseases,

Department of Health, New York, N. Y.

SEVERAL factors have been sus-pected of predisposing individuals

to a higher incidence or a more severeform of poliomyelitis. Among thesemay be mentioned trauma," 2 tonsillec-tomy,3' 4 fatigue,5 6, 16 and preg-nancy.7 S Reports of an apparent rela-tion between prior immunization andpoliomyelitis were made by DeTeys-sien 9 in 1921, and later by Jermu-lowicz,'0 Molinari," Banerjea,'2 Mazel,et al.,'3 Brain,'4 Verjaal,15 and Rus-sell.'6 These reports were based onobservations made in one or severalcases. During 1950 a number of Britishand Australian investigators reported onstudies in groups of cases 17-23 which in-dicated that in children recently inocu-lated with pertussis vaccine, diphtheriatoxoid, or tetanus toxoid the injectedlimb was more frequently the site ofparalysis than was the case in childrennot recently inoculated. Hill andKnowelden,24 in a careful statisticalstudy, corroborated these findings. Theyalso observed that a history of recentimmunization was obtained more fre-quently in cases of poliomyelitis than inmatched controls and considered it pos-sible that recent inoculations may haveshifted some of the cases of poliomyelitis

*Presented before a Joint Session of the Epidemi-ology, Health Officers, Laboratory, and Maternal andChild Health Sections of the American Public HealthAssociation at the Seventy-ninth Annual Meeting inSan Francisco, Calif., October 31, 1951.

from the abortive to the clinicallyrecognized group. In this country,Anderson and Skaar25 have confirmedthe British observation of a correlationbetween the site of injection and the siteof paralysis.On the other hand, Breen and Ben-

jamin26 found that, out of a total of356 cases under age 5 in the London out-breaks of 1947 and 1949, only 58 hada history of recent inoculation and inonly 33 was a real association estab-lished between paralysis and injection.They doubted that the associationcolored the epidemiological picture of theLondon outbreaks. McLeod 27 studiedan outbreak in Belfast in 1950. He ob-tained data on immunization from 73paralytic children, of whom 33 had beeninoculated at various times before theonset of their paralysis. He was unableto demonstrate a relationship betweenthe site of injection and the site ofparalysis; but he admitted that his ex-perience was limited.

METHODThe present study includes all cases

of poliomyelitis reported to the NewYork City Department of Health in1949 and 1950 among children 5 yearsof age and under from whom an accu-rate history of previous injections couldbe obtained. Upon the notification ofa case, a complete epidemiological his-

[142]

POLIOMYELITIS

tory was obtained by a physician withina month of onset. This included theclinical picture, results of spinal fluidexaminations, and outcome of the case.In addition, a history of injections re-ceived by every member of the family inthe age groups studied was obtained.This included the-dates of inoculationsand types of materials used. A monthafter onset, cases were revisited bypublic health nurses, who checked on

paralysis. Necessary corrections wereentered on the original cards. During1950, the investigation was extended toinclude children from 6 to 10 years ofage, and the physician also visited aneighboring family in which no case hadoccurred and obtained a similar inocula-tion history of all children 10 years ofage or under. These histories were usedin the selection of a control group.The data obtained from controls in

1950 were entered on cards and arrangedby age and sex. The case histories forthat year were similarly arranged. Foreach case a card was pulled at randomfrom the corresponding control groupand the date of onset of the case enteredon it. The time span between thisassigned date and the date of inocula-tion provided intervals among the con-trols comparable to the interval betweeninoculation and onset among the cases.

There were more controls than cases ineach group except for age 5, female, andage 8, male. Two controls were lackingin the former and five in the latter.They were obtained from control groupsof corresponding age and sex in 1951.Thus, for 1949 and 1950 data were

available on the time interval betweeninoculation and onset of poliomyelitis inthe cases, and for 1950 on inoculationsand the presumed onsets in the matchedcontrols. The agents used in the in-jections were also known for both cases

and controls.The data from cases in 1949 and 1950

were arranged by time interval betweeninoculation and onset. The data for the

1950 controls were similarly arranged.All cases and controls showing an in-terval of not more than 3 months be-tween inoculation and onset were furtherchecked for accuracy by communicatingwith the physician who had given theinjection and obtaining from his recordsthe dates of and materials used for theinoculations. Confirmations were ob-tained in about 75 per cent of thosequestioned. The differences between thestatement of the mother and the doctor'srecords were surprisingly few. Wheresuch a difference existed, the doctor'srecord was accepted. Where no recordwas kept by the doctor, the mother'sstatement was accepted if her reliabilitywas judged to be good. Otherwise thecase was not used in the study.

All data were entered on coding sheetsand punch cards made. These wereprocessed by tabulating machines.

Since no data on immunization wereobtained on children over 5 years ofage in 1949, the analyses for the twoyears are limited to children 5 years ofage or under. It should be pointed outthat the results apply to the 0-5 yearage group, but not necessarily to eachyear of age in the group. Furtheranalysis is needed to clarify this. Al-though the figures for the 6-10 yearage group are not used in the two yearanalysis, their inclusion would make nosignificant statistical change. The dis-cussion of immunization histories incases and controls is limited to the year1950 and therefore includes all childrenaged 10 years or under.

RESULTSThe total number of cases studied was

1,300, of which 681 occurred in 1949 and619 in 1950. Table 1 indicates theirdistribution by age, sex, paralysis, andhistory of previous inoculation. Of thetotal number, 879 had received their lastinjection with DPT, 259 with penicillin,and 117 with other or unknown agents.The term DPT signifies an injection with

Vol. 42 143

AMERICAN JOURNAL OF PUBLIC HEALTH

TABLE 1

Cases of Poliomyelitis by Age, Sex, Paralysis, and History of Previous Inoculation,New York City, 1949-1950

Sex I nterval Between Inoculation and Poliomyelitis, in Months

Fe- Not In- At Un-Male male Total 0-1 1-2 2-3 3-6 6-12 12+ oculated Onset known51 39 90 41 11 9 4 3 - 19 1 248 37 85 39 11 8 3 3 - 18 1 23 2 5 2 - 1 1 - - 1 - -

573413160

376 949 79 39 32 90 112 516 26 23 32293 706 56 27 25 78 86 375 20 17 2283 243 23 12 7 12 26 141 6 6 10

155 106 261 12 4 4 9 26 185 - 12 984 59 143 6 4 3 6 15 98 - 7 471 47 118 6 - 1 3 11 87 - 5 5

779 521 1,300 132 54 45 103 141 701 45 36 43

diphtheria toxoid, pertussis vaccine, ortetanus toxoid, singly or in any combi-nation. About 41 per cent had receiveda triple mixture of diphtheria toxoid,pertussis vaccine, and tetanus toxoid;27 per cent diphtheria toxoid only; 15per cent pertussis vaccine alone; and10 per cent a mixture of pertussis vac-cine and diphtheria toxoid. Four percent had been inoculated with tetanustoxoid only, and an equal percentagewith a mixture of diphtheria and tetanustoxoids. Thus, about 80 per cent hadbeen inoculated with diphtheria toxoid,65 per cent with pertussis vaccine, and50 per cent with tetanus toxoid. Thedata were insufficient to permit deter-mination of the influence of each of theseantigenic agents alone.There were 132 children injected not

more than a month before onset, 343between a month and a year priorthereto, and 701 more than a year before.In the latter group, information wasobtained about the type of injection, butno attempt was made to learn the exactdate of inoculation and the limb in-jected, since it was felt that the in-formation would not be sufficiently re-liable. The 36 children who had beeninjected on the day of onset had re-

ceived penicillin in all but one case,

probably as a therapeutic measure be-

fore a definite diagnosis was made. In43 instances no information was ob-tained on the interval between injectionand onset of poliomyelitis. In the totalgroup there were 779 boys and 521 girls.

It will be noted that in all age groups,by reference to columns 5 to 9 of Table1, there is a concentration of cases inthe month following the last injection, aswas pointed out by Anderson andSkaar.25 Thus, in age group T-1 year,41 cases occurred within a month afterthe last inoculation; 11 between 1 and2 months; 9 between 2 and 3 months,4 between 3 and 6 months, an averageof 1 per month; and 3 between 6 and 12months, an average of 1 case in 2months. Such a concentration suggestsa relationship between recent injectionand poliomyelitis.

Injected and Paralyzed Limbs-In-formation about the exact site of the lastinjection was available in 160 childrenin the age group 0-5 years (Table 2).Of these, 80 had been injected withDPT, 31 with penicillin and 49 withother agents.Among those inoculated with DPT, 2 7

had received their last injection withina month of onset of poliomyelitis and53 between a month and a year before.Thirty-three per cent of the former and13 per cent of the latter were paralyzed

AgeIn years

0-1ParalyzedNot Paralyzed

1-SParalyzedNot Paralyzed

6-10ParalyzedNot Paralyzed

Total

144 Feb., 1952

POLIOMYELITIS

TABLE 2

Relation Between Injected and Paralyzed Limb After Inoculation with DPT, Penicillin, andOther Agents, in Children 5 Years of Age or Under, New York City, 1949-1950

Agent Used

DPT

Penicillin

Other agents

All agents

Interval Between LastInjection and Onset of

Poliomyelitis, in Months

1- 22- 31-12

0-11- 22- 31-12

0- 11- 22- 31-12

0-11- 22- 31-12

in the injected limb, a statistically sig-nificant difference.Among the children who had received

penicillin within a year of onset of polio-myelitis, a similar difference was foundbetween those injected not more than amonth before and those injected betweena month and a year before. However,the numbers were small and the dif-ference was not statistically significant.Among the 49 children injected withother agents, only 1 had received aninjection within a month of onset ofpoliomyelitis. When all agents werecombined, there were 34 children in-jected not more than a month beforeonset and 12, or 35 per cent of these,were paralyzed in the injected limb. Onthe other hand, 126 children had beeninjected between 1 month and 1 yearbefore onset, and 20, or 16 per cent,were paralyzed in the same limb. Thedifference is statistically significant.

Injection into Unspecified Arms orLegs-In some cases there was informa-tion that the injections with DPT hadbeen given in the arm, but whether rightor left was not known. The number ofthese children who developed paralysisin the limbs was ascertained. There

Number In-jected in

Known Site

2713S

53

6112

'5

66

48

343016

126

Paralyzed in Injected Limb

Number Per cent

9207

3417

0006

1261

20

33150

13

50365028

13

35206

16

were 18 among those inoculated notmore than a month before onset of polio-myelitis and 43 among those inoculatedfrom 1 to;, 12 months before onset.Paralysis of the arms occurred among56 per cent of the former and 21 percent of the latter, a statistically signifi-cant difference (Table 3).There were too few children injected

with DPT in the legs to permit adequatecomparison between the recently and themore remotely injected cases.

All injections of penicillin had beengiven in the legs. There were 134 chil-dren so injected within a year of onsetof poliomyelitis. No significant differencein the percentage of paralyzed legs wasfound between those recently injectedand those injected some time before(Table 3).Weekly Interval Between Injection

and Onset-When the cases occurringwithin a month after injection werearranged by weeks, no marked pre-ponderance of cases was noted in anyone week. However, in the DPT-injectedchildren some bunching of cases wasnoted in the 2nd week after injection.Of the total of 55 cases occurring in the4 weeks after the last injection, 18, or

Vol. 42 145

AMERICAN JOURNAL OF PUBLIC HEALTH

TABLE 3

Retion Between Injected and Paralyzed Limbs in Cases Injected with DPT in the Arms andPenicin in the Legs, Which Arm or Leg Not Specified, in Children 5 Years of Age or

Under, New York City, 1949-1950Ixterdl Between Lastinjection axd Onset of

Poliomyelitis, in Months

0-11- 22- 31-12

0-11- 22- 31-12

Number Injected

2486

57

5298

82

Number ParalyzedIn Limbs

1856

43

34ss

49

TABI 4

Cases of Poliomyelitis Occurring in the 4 Weeks FoUowing Injection with AU Agents, by DaysAfter Injections, Children 0-5 Years of Age, New York City, 1949-1950

DPT

Days After Total Numberinjection Cases Paralyzed

1- 7 11 98-14 18 18

1S-21 14 1222-28 12 10

Total 55 49

33 per cent, occurred in the 2nd week.In the 1st, 3rd, and 4th weeks after thelast injection with DPT, there were 11,14, and 12 cases, respectively. Further-more, all 18 cases that occurred in the2nd week were paralytic, while in theother three weeks the paralytic caseswere respectively 9, 12, and 10 (Table4).

Evaluation of the group last injectedwith penicillin is somewhat complicated.A large number of cases, 32 out of atotal of 58, occurred in the 1st weekafter the last injection of penicillin, mostof them in the first 3 days. One wonderswhether many of these injections weregiven therapeutically, after the childbecame with poliomyelitis and beforea diagnosis was made. If this is thecase, then the apparent adverse effect ofrecent injections of penicillin noted in

Other Agents

Total NumberCases Paralysed

Total

Total NumberCases Paralyed

32 25 - - 43 3411 7 1 1 30 268 5 1 1 23 18

7 4 - - 19 14

58 41 2 2 115 92

some of the tables may not be of anyimportance. In none of the tables wasthe difference between the effect of re-cent and remote injections of penicillinlarge enough to be of statisticalsignificance. However, even the slightdifferences may be wiped out if some ofthe injections were given after onset butbefore symptoms were sufficiently char-acteristic for a diagnosis to be made.The cases injected in a known site in

the month preceding onset of polio-myelitis were also arranged by week ofonset after injection. The numbers ofinjected limbs paralyzed in each weekwere too small to allow the emergenceof any particular pattern (Table 5).

Ratio of Arm to Leg Paralysis-Since four-fifths of all DPT injectionshad been given in the arms and all peni-cillin injections in the legs, it appeared

Agext Used

DPT

Picillin

NuaberPara&yzedin-

Arms10 (56%)1 (20%)2 (33%>9 (21%)

Legs29 (85%)4 (80%)4 (80%o)

36 (90%o)

Penicillin

Tota NumberCases Paralysed

Feb., 1952146

POLIOMYELITIS

TABLE 5S

Relation Between Injected and Paralyzed Limb in Cases of Pokomyelitis, by Week AfterInjection, Chidren 05 Years of Age, New York City, 1949-1950

Number Ixiected in Known Site

OtherPenicillin Agents

Paralyzed in Injected Limb

OtherTotal DPT Penicillin Agents Total

1- 7 6 38-14 7 2

1S-21 5 -

22-28 7 1

Total 25 6

9 21 9 1- 6 2- 8 3

32 8 3

- 31 - 2_ 21 4

- 11

TABLE 6

Ratio of Leg to Arm Paralysis in Children 5 Years of Age or Under, Inoculated with DPT orPenicillin, or Not Inoculated, New York City, 1949-1950

Agent Used

DPT

Penicillin

None

Interval Between Last In-jection and Onset of

Poliomyelitis, in Months

0-1

1- 22- 31-12

0- I1- 22- 31-12

Not Inoculated

profitable to compare the ratios of legto arm paralysis in the two groups,whether the site of injection was knownor not. It will be noted in Table 6 thatamong the children not inoculated, threeto four times as many were paralyzed inthe legs as in the arms. This held truealso for children inoculated with DPTmore than a year before onset and be-tween 1 month and 1 year before onset.However, among those children injectedwith DPT not more than a month beforeonset, almost as many arms as legs wereparalyzed, indicating a definite influenceof the injecting agent.

In the penicillin-injected group, whereall injections were given in the legs, agreater percentage of leg paralysisoccurred among the children inoculatedwithin a month than among those in-oculated more than a month beforeonset of poliomyelitis. The difference

Number of ChildrenParalyzed in Limb

, A

Number of LimbsParalyzed

Arms Legs L/A Arms Legs LIA

22 33 1.5 27 44 ..64 13 3.3 4 20 5.06 13 2.2 8 19 2.4

24 83 3.5 29 111 3.8

9 33 3.7 10 49 4.97 9 1.3 8 11 1.41 5 5.0 2 8 4.0

23 53 2.3 28 69 2.5

10 31 3.1 12 42 3.5

was not statistically significant and, asalready indicated, there is doubt thatall the injections preceded onset ofpoliomyelitis.

In both the DPT and the penicillingroups the ratios of arm to leg paralysiswere approximately the same, whetherone compared paralyzed children orparalyzed limbs. In the group injectedwith other agents, the figures were toosmall in some categories to make com-parisons worth while.

Severity of Cases-Studies on theseverity of poliomyelitis following ton-sillectomy indicate that the bulbar typeis many times more frequent in cases

that have had a recent tonsillectomythan in those that have had none or thathad one in the more distant past.8 4 Itwas of interest to determine whetherthis held true also for children recently

injected. In Table 7 the cases that had

Days AfterInjection DPT

147Vol. 42

AMERICAN JOURNAL OF PUBLIC HEALTH

received an injection of DPT or peni-cillin were arranged according to clinicaltypes. Those called spinal paralytic hadparalysis resulting from involvement ofthe spinal cord only; those called bulbarhad involvement of a cranial nerve ormedullary center, alone or combined withparalysis of the trunk or a limb, andaccounted for almost all the deaths. Forcomparison, the cases that had neverbeen inoculated were similarly arranged.It will be seen that the percentage dis-tribution of cases inoculated with DPTwithin a month of onset of poliomyelitisaccording to nonparalytic, spinal para-lytic, and bulbar types hardly differsfrom the percentage distribution of casesinoculated 1 month to 1 year beforeonset, or of those not inoculated.

Similar comparison made in children

previously injected with penicillin alsoshows no significant difference inseverity of the poliomyelitis, whetherthe injections had been given recentlyor some time in the past. An analysismade separately for children under 1year of age, for those between 1 and 5years, and for those in the 6 to 10 yearage group also showed no significantdifferences in the types of clinicalpoliomyelitis, whatever the agent usedand whatever the time interval betweeninjection and onset of poliomyelitis.

Anderson and Skaar25 also found noincrease in bulbar cases in the recentlyinjected group. However, they dividedtheir spinal paralytic cases into severeand mild paralysis. They were able todemonstrate a more severe paralysis inthe children immunized in the previous

TABLE 7

Cases of Poliomyeltis According to Clinical Type and History of Previous Inoculation withDPT or Penicillin, Children 5 Years of Age or Under, New York City, 1949-1950

Interval Between Last Inoculation and Poliomyelitis, in Months__- N I

Agent Used Type of Case

NonparalyticDPT Spinal "DPT ~Bulbar

Total

NonparalyticPenicillin Bulbar "

Total

0-1

No. %

7 1244 758 13

Not In-1-2 2-3 1-12 oculated

No. % No. % No. % No. %

5 23 2 9 20 15 7 1615 68 13 62 90 68 33 732 9 6 29 22 17 5 11

59 100 22 100 21 100 132 100

17 29 7 35 4 40 30 2837 64 13 65 5 50 62 574 7 0 - 1 10 16 15

58 100 20 100 10 100 108 100

45 100

TABLE 8

Comparison of the Histories of Injections with DPT or Penicilin of Cases of Poliomyelitis,Children 10 Years of Age or Under, and in Their Matched Controls, New York City, 1950

Interval Between LastInjection and Onset of

Poliomyelitis, in Months

0- 11- 22- 31-12

0- 11- 22- 31-12

Agent Used

DPT

Peniclin

Cases26 (33%)8 (10%)7 ( 9%)

52

27 (24%)-7 ( 6%)9 ( 9%)

84

Controls

6 (10%)10 (16%)5 ( 8%)

57

13 (13%)12 (12%)8 ( 8%)

91

Feb.,, 1952148

POLIOMYELITIS

month than in those inoculated 2 to 6months before. We did not havesufficient information on severity ofparalysis to make such an analysis.The other agents injected not more

than a month before onset of polio-myelitis were too few in number to makediscussion of them of value.

Controls-During the 1950 investiga-tion, the physician who obtained theimmunization history of the family inwhich a case occurred also obtained asimilar immunization history from aneighboring family where no case hadoccurred. By matching controls tocases, as indicated earlier in this paper,one coUma compare 1jections that cases -

controls had receivedpast. The distributitypes of immunizingtially the same in cas

Diphtheria toxoidPertussis vaccineTetanus toxoidDiphtheria toxoid

and pertussis vaccineDiphtheria and

tetanus toxoidsDiphtheria and

tetanus toxoids andpertussis vaccine

Total

The results are showwill be noted that fipenicillin a significanage of cases than of

injected within a month of onset ofpoliomyelitis than in the preceding 11months. In the group injected withother agents, the numbers were too smallfor valid comparison. When the totalgroup of 619 cases and a similar numberof controls were compared, the resultswere similar (Table 9). Only in thegroups injected not more than a monthpreceding the onset of poliomyelitis wasthere a significant difference between thepercentages of cases and controls thathad received injections, two and a halftimes greater in the former than in thelatter.

me number ot in- DISCUSSIONnd their matched This investigation corroborates therecently and in the published findings of other investigatorson of the different that there is a relationship between re-agents was essen- cent inoculation with diphtheria toxoid,es and in controls: tetanus toxoid or pertussis vaccine

(DPT) and the development of para-Cases Controls lytic poliomyelitis. This is indicated by

No. % No. % the fact that a significantly larger per-45 11 43 10 centage of children was paralyzed in the12 3 12 36 2 7 2 injected limb when the last injection

178 44 212 49 had been received not more than amonth preceding onset of poliomyelitis5 1 4 1 than when received from a month to ayear before. Also, where the precisesite of injection was not known, but

405 100 425 100 where it was known that the injectionof DPT had been given in the arms, a

mn in Table 8. It similar difference occurred in the per-or both DPT and centage of arm paralysis, depending ontly larger percent- whether the inoculation prior to onsetcontrols had been of poliomyelitis was recent or remote.

TABLE 9Comparison of the Histories of Injections with All Agents in Cases of Poliomyelitis, Childrrn

10 Years of Age or Under, and in Their Matched Controls, New York City, 1950Interval Between Last Injection and Onset of Paliomyelitis, in Months

At OnsrtNot or

Total 0-1 1-2 2-3 1-12 12+ Injected Unknown

No. % No. % No. % No. % No. % No. % No. % No. %Cases 619 100 56 9.0 19 3.0 21 3.4 163 26.3 334 54.0 14 2.3 52 8.4Controls 619 100 21 3.4 25 4.0 15 2.4 173 27.9 358 57.8 18 2.9 49 7.9

149Vol. 42

AMERICAN JOURNAL OF PUBLIC HEALTH

Furthermore, in all inoculations withDPT, 85 per cent of which had beengiven in the arms, the usual ratio of legto arm paralysis changed from about 3to 1 to approximately 1 to 1 when theinoculation had been given not morethan a month before onset. If the in-terval between injection and onset wasgreater, or if there had been no inocu-lation, the usual ratio prevailed.The separation of cases that occurred

'in the month following the last injectionof DPT, by weeks of incidence, showedsome bunching of cases in the 2nd week.This suggested that the effect of the in-jection was felt mainly during the in-cubation period.The results of injections with peni-

cillin were not clear-cut. An analysis ofthe cases that followed within a monthafter the last injection of penicillin indi-cated a concentration of cases in thefirst few days after the injection. Thereis a strong probability that these injec-tions were given therapeutically afteronset but before a diagnosis was made.If these cases are omitted, even theslight differences that were occasionallynoted between the effects of recent andmore distant injections of penicillin losetheir significance. It is pertinent tomention that in the next to last columnof Table 1 there are noted 36 cases ofpoliomyelitis that had been injected onthe day of onset. They were not usedin the comparison of recent and remoteinjections. All but one of the injectionswere of penicillin. Here, too, the sup-position is strong that the injections weregiven therapeutically, after onset.An analysis of the histories of previous

inoculations, obtained from cases ofpoliomyelitis in 1950 and from theirmatched controls, indicated that a sig-nificantly larger percentage of casesthan of controls had received injectionsin the month preceding onset, or pre-sumed onset, of poliomyelitis. This wastrue both for DPT and for penicillin.It suggests, but does not prove, that

there may be an increased incidence ofrecognizable poliomyelitis in recently in-jected children. The possibility isenhanced by the fact that, as noted be-fore in Table 1, a concentration of casesis seen in the month following the lastinjection. Whether this applies to peni-cillin- as well as to DPT-injected chil-dren depends, as previously stated, onwhether injections of penicillin, givenpresumably before onset of poliomyelitis,had not really been given before diag-nosis but after onset. Further studyshould elucidate this problem.

It is true that in all comparisonsmade, the actual numerical differences,even when statistically significant, weresmall. However, for DPI, the consis-tency of the trend and the fact that anumber of other investigators foundsimilar differences tend to strengthenthe hypothesis that a definite relation-ship exists between recent injections andpoliomyelitis. Hill and Knowelden24and Anderson and Skaar 25 have stressedthe fact that the relationship exists onlyfor recent injections. As in the case oftonsillectomy, there is no evidence inour investigations, either, that injectionsreceived in the more distant past haveany influence whatever on the localiza-tion or occurrence of poliomyelitis. Itshould be noted in Tables 2, 3, 6, 8,and 9 that the differences found by acomparison of the intervals 0-1 and1-12 months after last inoculation donot obtain when similar compari-sons are made between the interval 1-2or 2-3 months after the last injectionand the rest of the year. Obviously, theinfluence of the last injection does notextend beyond a month following thelast injection.When the injected cases were sepa-

rated according to clinical types, therewas no evidence that recent injectionshave any influence on the bulbar typeof poliomyelitis, and therefore on deaths.The percentage distribution of cases re-cently injected with DPT was little

Feb., 1952ISO

POLIOMYELITIS

different from the distribution of non-injected cases. No attempt was madeto divide the spinal paralytic cases intomild and severe ones. There is someindication in the work of others25 thatrecent injections tend to favor a moresevere form of spinal paralysis.The practical question that presents

itself to health officers and physiciansgenerally as a result of these and otherpublished data is, what should theirattitude be toward routine and otherinoculations during the season of in-creased incidence of poliomyelitis? Thereare no published data to indicate therelative incidence of poliomyelitis in re-cently injected as compared with non-injected children. Korns, Albrecht, andLocke have estimated that the incidenceof poliomyelitis in persons inoculatednot more than 2 months before onsetof poliomyelitis is about twice that of theuninoculated.28 In New York City themedian of annually reported cases ofpoliomyelitis is about 600, and less than3 per cent of these occur in childrenunder 1 year of age. In 1950, a littleover 1,000 cases were reported. Byactual count, 27 of these were in infantsunder 1 year of age and 18 of them hadreceived an injection of some kind dur-ing the year. Ten infants had beeninoculated with DPT not more than 2months before onset.

Presumably, if no inoculations hadbeen given to infants under 1 year ofage in 1950, half of the 10 cases wouldhave come down with poliomyelitis. Theextra hazard from inoculations was

therefore limited to 5 children in a year

when 1,000 cases were reported. In an

average year, when about 600 cases are

reported, the extra hazard might involve3 children, or 1 child in about 50,000of that age in the population. The riskis probably even lower, since for 6 to 8months of the year children are hardlyexposed to poliomyelitis. In smallercommunities the extra hazard applies to-even fewer infants. It becomes obvious

that although the risk to infants under ayear of age from recent injections withDPT is scientifically a true one, in prac-tice it is more academic than real.Infants under 6 months of age constituteabout one-half of one per cent of allcases of poliomyelitis. The extra hazardto them is so small as to be negligible.

It appears reasonable to continueroutine immunizations in infants under6 months of age irrespective of the inci-dence of poliomyelitis in the community.In a nonepidemic year, routine immuni-zations may even be permitted in infantsunder 1 year of age. In older children,routine inoculations should be deferreduntil after the poliomyelitis season. Inthe presence of a severe epidemic it maybe wise to defer routine inoculationsamong infants over 6 months of age aswell as among older children. Thereshould be no suspension of therapeuticinoculations at any time, first, becausethere is insufficient evidence that theireffect is like that of injections with anti-gens; and second, because the risk ofwithholding them is greater than theextra risk of the injection. This holdstrue, also, for immediate prophylacticinjections, such as booster doses of tet-anus toxoid in an injured child, anti-rabic injections, typhoid vaccine forcontacts to a case, diphtheria toxoid fornurses beginning service on a communi-cable disease ward, etc. In all these sit-uations there is a greater risk in with-holding the injection than in giving it.

CONCLUSIONS1. The inoculation histories of 1,300 children

that had poliomyelitis in New York City in1949 and 1950 were analyzed.

2. A relationship was shown to exist be-tween site of injection and site of paralysis inchildren injected not more than a month be-fore onset of poliomyelitis with diphtheriatoxoid, pertussis vaccine, or tetanus toxoid, orany combination of the three.

3. This relationship was not clearly shownto follow recent injections of penicillin orother agents.

4. No increase in bulbar cases or deathsresulted from previous injections.

151Vol. 42

152 AMERICAN JOURNAL OF PUBLIC HEALTH Feb., 1952

5. A comparison of 619 children with polio-myelitis and their matched controls indicatedthat a larger percentage of cases than of con-trols had had recent injections. This suggestedthat the incidence of recognizable poliomyelitismay be greater in children recently injectedthan in those not injected or injected in thepast.

6. The extra hazard of poliomyelitis as aresult of recent inoculations is small in chil-dren under 1 year of age and is negligible ininfants under 6 months of age.

7. It appears reasonable to continue im-munizations in infants under 6 months of ageat all times. During nonepidemic years, im-munizations may be given to infants under 1year of age throughout the year. Routineimmunizations may well be suspended in olderchildren during the poliomyelitis season. Inepidemic years the suspension might be ex-tended to infants between 6 months and 1 yearof age.

8. Therapeutic and immediate prophylacticinjections should not be discontinued at anytime.

ACKNOWLEDGMENT-We are greatly indebtedto Ralph Muckenfuss, M.D., Assistant Com-missioner, New York City Department ofHealth, for many helpful suggestions. LouisPincus, Senior Statistician, performed thetabulations and helped with technical advice.

REFERENCES1. Parker, H. L. Acute Anterior Poliomyelitis: Re-

lationship of Injury to Localization of Paralysis.Irish J. M. Sc. 136:303-305 (July 7), 1937.

2. Horstmann, D. M., and Paul, J. M. The Incuba-tion Period in Human Poliomyelitis and Its Im-plications. J.A.M.A. 135:11-14 (Sept. 6), 1947.

3. Fischer, A. E., Stillerman, M., and Marks, H. H.Relation of Tonsillectomy and of Adenoidectomyto the Incidence of Poliomyelitis, with SpecialReference to Bulbar Form. Am. I. Dis. Child. 61:305-321 (Feb.), 1941.

4. Siegel, M., Greenberg, M., and Magee, M. C.Tonsillectomy and Poliomyelitis. I. Studies onIncidence in 1949. 1. Pediat. 537-547 (May),1951; II. Frequency of Bulbar Paralysis, 1944-1949. Idem., 548-558.

5. Horstmann, D. M. Acute Poliomyelitis. Relationof Physical Activity at the Time of Onset to theCourse of the Disease. J.A.M.A. 142:236-241(Jan. 28), 1950.

6. Brahdy, M. B., and Katz, S. H. Effect of Trans-portation on Severity of Acute Poliomyelitis.J.A.M.A. 146:772-774 (June 30), 19S.

7. Aycock, W. L. The Frequency of Poliomyelitis in

Pregnancy. New England J. Med. 45:405-408(Sept. 11), 1941.

8. Fox, M. J., and Belfus, F. H. Poliomyelitis inPregnancy. Am. J. Obst. & Gyncc. 59:1134-1139(May), 1950.

9. DeTeyssien, M. Acute Poliomyelitis Following anInjection of Anti-Cholera Vaccine. J. de mid. deBordeaux 92:77-78 (Feb.), 1921.

10. Jermulowicz, M. Two Cases of Poliomyelitis Fol-lowing Anti-Smallpox Vaccination. Rev. neurol.37:92-93 (Jan.), 1930.

11. Molinari, E. Poliomyelitis: Common ProdromalSymptoms and Therapy of Paralysis FollowingAnti-Smallpox Vaccination: Colibacillary Etiology.Gazz. d. osp. 55:226-232 (Feb. 25), 1934.

12. Banerjea, J. C. Acute Anterior Poliomyelitis Fol-lowing Vaccination in an Infant. Indian J. Pediat.2:177 (Apr.), 1935.

13. Mazel, P., Meeserman, F., Camelin, A., andGuibert, A. A Case of Poliomyelitis Following anInjection of Multiple Vaccine (TAB and Diph-theria Anatoxin). Soc. de mid. mil. franc., Bull.mens. 33:314-317 (June), 1939.

14. Brain, R. Discussion on Recent Experience ofAcute Encephalomyelitis and Allied Conditions.Proc. Roy. Soc. Med. 36:319-320 (Apr.), 1943.

15. Verjaal, A. Poliomyelitis After Vaccination andInjury. Nederl. tijdschr. v. geneesk. 91:3137-3141(Nov.), 1947.

16. Russell, W. R. Paralytic Poliomyelitis: EarlySymptoms and Effect of Physical Activity onCourse of Disease. Brit. M. J. 1:465-471 (Mar.19), 1949.

17. Martin, J. K. Local Paralysis in Children AfterInjections. Arch. Dis. Childhood 25:1-14 (Mar.),1950.

18. Geffen, D. H. The Incidence of Paralysis Occurringin London Children Within Four Weeks After Im-munization. M. Officer 83:137-140 (Apr. 8), 1950.

19. McCloskey, B. P. The Relation of ProphylacticInoculations to the Onset of Poliomyelitis. Lancet1:659-663 (Apr. 8), 1950.

20. Burnet, F. M. et al. Discussion on PoliomyelitisFollowing Inoculations. Proc. Roy. Soc. Med. 43:775-782 (Nov.), 1950.

21. McCallum, F. 0. Clinical Poliomyelitis in Associa-tion with Peripheral Inoculation of ProphylacticLBrit. M. J. 2:6-7 (July 1), 1950.

22. Banks, H. S., and Beale, A. J. Poliomyelitis andInoculation against Whooping Cough and Diph-theria. Brit. M. J. 2:251-252 (July 29), 1950.

23. Bousefield, G. Paralysis Following ImmunizingInjections: Measures Calculated To Diminish theRisk. Lancet 1:1028-1032 (May 12), 1951.

24. Hill, A. B., and Knowelden, J. Inoculation andPoliomyelitis: A Statistical Investigation in Eng-land and Wales in 1949. Brit. M. J. 2:1-6 (July1), 1950.

25. Anderson, G. W., and Skaar, A. E. PoliomyelitisOccurring After Antigen Injections. Pediatrics 7:741-759 (June), 1951.

26. Breen, G. E., and Benjamin, B. Poliomyelitis inLondon in 1949. Brit. M. 1. 2:1473-1475 (Dec.30), 1950.

27. McLeod, W. J. Poliomyelitis and Diphtheria Im-munization in Belfast. Brit. M. J. 1:736-738 (Apr.7), 1951.

28. Korns, R. F. Immunization During PoliomyelitisSeason. Health News 28:11 (July), 1951.