Original Article HumanandExperimentalToxicology 30(9)1420–1428 Infant mortality rates regressed ªTheAuthor(s)2011 Reprintsandpermission: against number of vaccine doses sagepub.co.uk/journalsPermissions.nav DOI:10.1177/0960327111407644 routinely given: Is there a het.sagepub.com biochemical or synergistic toxicity? Neil Z Miller and Gary S Goldman Abstract Theinfant mortality rate(IMR)is oneofthemostimportant indicatorsof thesocio-economicwell-being and publichealthconditionsofacountry.TheUSchildhoodimmunizationschedulespecifies26vaccinedosesfor infantsagedlessthan1year—themostintheworld—yet33nationshavelowerIMRs.Usinglinearregression, the immunization schedules of these 34 nations were examined and a correlation coefficient of r ¼ 0.70 (p<0.0001)wasfoundbetweenIMRsandthenumberofvaccinedosesroutinelygiventoinfants.Nationswere alsogroupedintofivedifferentvaccinedoseranges:12–14,15–17,18–20,21–23,and24–26.ThemeanIMRsof all nations within each group were then calculated. Linear regression analysis of unweighted mean IMRs showed a high statistically significant correlation between increasing number of vaccine doses and increasing infantmortalityrates,withr¼0.992(p¼0.0009).UsingtheTukey-Kramertest,statisticallysignificantdiffer- encesinmeanIMRswerefoundbetweennationsgiving12–14vaccinedosesandthosegiving21–23,and24–26 doses.Acloserinspectionofcorrelationsbetweenvaccinedoses,biochemicalorsynergistictoxicity,andIMRs is essential. Keywords infantmortalityrates,suddeninfantdeath,SIDS,immunizationschedules,childhoodvaccines,drugtoxicology, synergistic effects, linear regression model Introduction adecreaseinimmunefunction.Animpairedimmune function often leads to an increased susceptibility to The infant mortality rate (IMR) is one of the most infection.2 It is well established that infections, no important measures of child health and overall matter how mild, have adverse effects on nutritional development in countries. Clean water, increased status. Conversely, almost any nutritional deficiency nutritional measures, better sanitation, and easy will diminish resistance to disease.3 accesstohealthcarecontributethemosttoimproving DespitetheUnitedStatesspendingmorepercapita infantmortalityratesinunclean,undernourished,and on health care than any other country,4 33 nations impoverished regions of the world.1–3 In developing havebetterIMRs.SomecountrieshaveIMRsthatare nations,IMRsarehighbecausethesebasicnecessities less than half the US rate: Singapore, Sweden, and forinfantsurvivalarelackingorunevenlydistributed. Japan are below 2.80. According to the Centers for Infectious and communicable diseases are more DiseaseControlandPrevention(CDC),‘‘Therelative common in developing countries as well, though position of the United States in comparison to coun- sound sanitary practices and proper nutrition would tries with the lowest infant mortality rates appears do much to prevent them.1 to be worsening.’’5 The World Health Organization (WHO) attributes 7 out of 10 childhood deaths in developing countries to five main causes: pneumonia, diarrhea, measles, Correspondingauthor: malaria,andmalnutrition—thelattergreatlyaffecting NeilZMiller,POBox9638,SantaFe,NM87504,USA all the others.1 Malnutrition has been associated with Email:[email protected] Miller N Z and Goldman G S 1421 Table 1.2009Infantmortality rates,top34 nations8 nations routinely give to their infants and their infantmortalityrates,alinearregressionanalysiswas Rank Country IMR performed. 1 Singapore 2.31 2 Sweden 2.75 3 Japan 2.79 Methods and design 4 Iceland 3.23 Infant mortality 5 France 3.33 6 Finland 3.47 The infant mortality rate is expressed as the 7 Norway 3.58 number of infant deaths per 1000 live births. 8 Malta 3.75 According to the US Central Intelligence Agency 9 Andorra 3.76 (CIA), which keeps accurate, up-to-date infant 10 CzechRepublic 3.79 11 Germany 3.99 mortality statistics throughout the world, in 2009 12 Switzerland 4.18 there were 33 nations with better infant mortality 13 Spain 4.21 rates than the United States (Table 1).8 The US 14 Israel 4.22 infant mortality rate of 6.22 infant deaths per 15 Liechtenstein 4.25 1000 live births ranked 34th. 16 Slovenia 4.25 17 SouthKorea 4.26 Immunization schedules and vaccine doses 18 Denmark 4.34 19 Austria 4.42 A literature review was conducted to determine the 20 Belgium 4.44 immunization schedules for the United States and all 21 Luxembourg 4.56 33nationswithbetterIMRsthantheUnitedStates.9,10 22 Netherlands 4.73 The total number of vaccine doses specified for 23 Australia 4.75 24 Portugal 4.78 infantsagedless than 1year was then determinedfor 25 United Kingdom 4.85 each country (Table 2). A vaccine dose is an exact 26 NewZealand 4.92 amount of medicine or drug to be administered. The 27 Monaco 5.00 number of doses a child receives should not be con- 28 Canada 5.04 fused with the number of ‘vaccines’ or ‘injections’ 29 Ireland 5.05 given. For example, DTaP is given as a single injec- 30 Greece 5.16 tion but contains three separate vaccines (for 31 Italy 5.51 diphtheria, tetanus, and pertussis) totaling three vac- 32 SanMarino 5.53 33 Cuba 5.82 cine doses. 34 United States 6.22 Nations organized into data pairs CIA.Countrycomparison:infantmortalityrate(2009).TheWorld Factbook.www.cia.gov(Datalastupdated13April2010).8 The34nationswereorganizedintodatapairsconsist- ing of total number of vaccine doses specified for their infants and IMRs. Consistent with biostatistical There are many factors that affect the IMR of any conventions, four nations—Andorra, Liechenstein, given country. For example, premature births in the Monaco, and San Marino—were excluded from the United States have increased by more than 20% dataset because they each had fewer than five infant between1990and2006.Pretermbabieshaveahigher deaths, producing extremely wide confidence inter- risk of complications that could lead to death within vals and IMR instability. The remaining 30 (88%) the first year of life.6 However, this does not fully of the data pairs were then available for analysis. explainwhytheUnitedStateshasseenlittleimprove- ment in its IMR since 2000.7 Nations organized into groups Nations differ in their immunization requirements for infants aged less than 1 year. In 2009, five of the Nations were placed into the following five groups 34 nations with the best IMRs required 12 vaccine based on the number of vaccine doses they routinely doses, the least amount, while the United States give their infants: 12–14, 15–17, 18–20, 21–23, and required 26 vaccine doses, the most of any nation. 24–26 vaccine doses. The unweighted IMR means Toexplorethecorrelationbetweenvaccinedosesthat of all nations as a function of the number of vaccine 1422 Human and Experimental Toxicology 30(9) Table2.SummaryofInternationalImmunizationSchedules:vaccinesrecommended/requiredpriortooneyearofagein 34nations Totalb Group Nation Vaccines prior toone yearof age doses (range ofdoses) Sweden DTaP (2),Polio (2),Hib(2), Pneumo (2) 12 1(12–14) Japan DTaP (3),Polio (2),BCG 12 Iceland DTaP (2),Polio (2),Hib(2), MenC(2) 12 Norway DTaP (2),Polio (2),Hib(2), Pneumo (2) 12 Denmark DTaP (2),Polio (2),Hib(2), Pneumo (2) 12 Finland DTaP (2),Polio (2),Hib(2), Rota(3) 13 Malta DTaP (3),Polio (3),Hib(3) 15 2(15–17) Slovenia DTaP (3),Polio (3),Hib(3) 15 South Korea DTaP (3),Polio (3),HepB (3) 15 Singapore DTaP (3),Polio (3),HepB (3), BCG,Flu 17 New Zealand DTaP (3),Polio (3),Hib(2), HepB (3) 17 Germany DTaP (3),Polio (3),Hib(3), Pneumo (3) 18 3(18–20) Switzerland DTaP (3),Polio (3),Hib(3), Pneumo (3) 18 Israel DTaP (3),Polio (3),Hib(3), HepB (3) 18 Liechtensteina DTaP (3),Polio (3),Hib(3), Pneumo (3) 18 Italy DTaP (3),Polio (3),Hib(3), HepB (3) 18 SanMarinoa DTaP (3),Polio (3),Hib(3), HepB (3) 18 France DTaP (3),Polio (3),Hib(3), Pneumo (2),HepB (2) 19 CzechRepublic DTaP (3),Polio (3),Hib(3), HepB (3),BCG 19 Belgium DTaP (3),Polio (3),Hib(3), HepB (3),Pneumo (2) 19 United Kingdom DTaP (3),Polio (3),Hib(3), Pneumo (2),MenC (2) 19 Spain DTaP (3),Polio (3),Hib(3), HepB (3),MenC (2) 20 Portugal DTaP (3),Polio (3),Hib(3), HepB (3),MenC (2),BCG 21 4(21–23) Luxembourg DTaP (3),Polio (3),Hib(3), HepB (2),Pneumo (3), Rota(3) 22 Cuba DTaP (3),Polio (3),Hib(3), HepB (4),MenBC (2),BCG 22 Andorraa DTaP (3),Polio (3),Hib(3), HepB (3),Pneumo (3), MenC(2) 23 Austria DTaP (3),Polio (3),Hib(3), HepB (3),Pneumo (3), Rota(2) 23 Ireland DTaP (3),Polio (3),Hib(3), HepB (3),Pneumo (2), MenC(2), BCG 23 Greece DTaP (3),Polio (3),Hib(3), HepB (3),Pneumo (3), MenC(2) 23 Monacoa DTaP (3),Polio (3),Hib(3), HepB (3),Pneumo (3), HepA,BCG 23 Netherlands DTaP (4),Polio (4),Hib(4), Pneumo (4) 24 5(24–26) Canada DTaP (3),Polio (3),Hib(3), HepB (3),Pneumo (3), MenC(2), Flu 24 Australia DTaP (3),Polio (3),Hib(3), HepB (4),Pneumo (3), Rota(2) 24 United States DTaP (3),Polio (3),Hib(3), HepB (3),Pneumo (3), Rota(3), Flu(2) 26 aThesefournationswereexcludedfromtheanalysisbecausetheyhadfewerthanfiveinfantdeaths. bDTaPisadministeredasasingleshotbutcontainsthreeseparatevaccines(fordiphtheria,tetanus,andpertussis).Thus,DTaPgiven threetimesininfancyisequivalenttoninevaccinedoses. Immunizationschedulesarefor2008–2009.9,10 doses were analyzed using linear regression. The resultsfromtheTukey-Kramertest,aposttestforthe Pearson correlation coefficient (r) and coefficient of overall linear trend was performed. determination (r2) were calculated using GraphPad Prism, version 5.03 (GraphPad Software, San Diego, Results CA, USA, www.graphpad.com). Additionally, the Nations organized into data pairs Fstatisticandcorrespondingpvalueswerecomputed to test if the best fit slope was statistically signifi- A scatter plot of each of the 30 nation’s IMR versus cantly non-zero. The Tukey-Kramer test was used to vaccinedosesyieldedalinearrelationshipwithacor- determine whether or not the mean IMR differences relation coefficient of 0.70 (95% CI, 0.46–0.85) and betweenthegroupswerestatisticallysignificant.Fol- p < 0.0001 providing evidence of a positive correla- lowing the one-way ANOVA (analysis of variance) tion:IMRandvaccinedosestendtoincreasetogether. Miller N Z and Goldman G S 1423 7 best fit line: y = 0.148x + 1.566, r = 0.7 (p < 0.0001) 7 best fit line: y = 0.157x +1.34, r = 0.992 (p = 0.0009) slope F−statistic: F = 27.2, p < 0.0001 slope F−statistic: F = 173.9, p = 0.0009 6 e 6 ate y rat Infant mortality r(deaths/1000) 45 best fit line an infant mortalit(deaths/1,000) 45 18−20 21−23 24−26 3 95% CI Me 3 best fit line 12−14 15−17 2 2 9 12 15 18 21 24 27 13 16 19 22 25 Number of vaccine doses Vaccine Doses Figure 1. 2009 Infant mortality rates and number of Figure 2. 2009 Mean infant mortality rates and mean vaccine dosesfor30 nations. numberof vaccine doses(five categories). Discussion The F statistic applied to the slope [0.148 (95% CI, Basic necessities for infant survival 0.090–0.206)]issignificantlynon-zero,withF¼27.2 (p<0.0001;Figure1). It is instructive to note that many developing nations requiretheirinfantstoreceivemultiplevaccinedoses and have national vaccine coverage rates (a percent- ageofthetargetpopulationthathasbeenvaccinated) Nations organized into groups of 90% or better, yet their IMRs are poor. For exam- The unweighted mean IMR of each category was ple, Gambiarequires its infants toreceive 22vaccine computed by simply summing the IMRs of each doses during infancy and has a 91%–97% national nationcomprisingagroupanddividingbythenumber vaccine coverage rate, yet its IMR is 68.8. Mongolia of nations in that group. The IMRs were as follows: requires 22 vaccine doses during infancy, has a 3.36 (95% CI, 2.74–3.98) for nations specifying 95%–98% coverage rate, and an IMR of 39.9.8,9 12–14 doses (mean 13 doses); 3.89 (95% CI, These examples appear to confirm that IMRs will 2.68–5.12) for 15–17 doses (mean 16 doses); remain high in nations that cannot provide clean 4.28 (95% CI, 3.80–4.76) for 18–20 doses (mean water, proper nutrition, improved sanitation, and bet- 19 doses); 4.97 (95% CI, 4.44–5.49) for 21–23 doses ter access to health care. As developing nations (mean 22doses);5.19(95% CI,4.06–6.31)for24-26 improve in all of these areas a critical threshold will doses (mean 25 doses; Figure 2). Linear regression eventuallybereachedwherefurtherreductionsofthe analysis yielded an equation of the best fit line, infant mortality rate will be difficult to achieve y ¼ 0.157x þ 1.34 with r ¼ 0.992 (p ¼ 0.0009) and because most of the susceptible infants that could r2¼0.983.Thus,98.3%ofthevariationinmeanIMRs have been saved from these causes would have been isexplainedbythelinearmodel.Again,theFstatistic saved. Further reductions of the IMR must then be yieldedasignificantlynon-zeroslope,withF¼ 173.9 achievedinareasoutsideofthesedomains.Asdevel- (p¼0.0009). opingnationsascendtohighersocio-economicliving The one-way ANOVA using the Tukey-Kramer standards, a closer inspection of all factors contribut- test yielded F ¼ 650 with p ¼ 0.001, indicating the ing to infant deaths must be made. five mean IMRs corresponding to the five defined dose categories are significantly different (r2 ¼ Crossing the socio-economic threshold 0.510).Tukey’smultiplecomparisontestfoundstatis- ticalsignificanceinthedifferencesbetweenthemean Itappearsthatatacertainstageinnations’movement IMRs of those nations giving 12–14 vaccine doses upthesocio-economicscale—afterthebasicnecessi- and (a) those giving 21–23 doses (1.61, 95% CI, ties for infant survival (proper nutrition, sanitation, 0.457–2.75) and (b) those giving 24–26 doses (1.83, clean water, and access to health care) have been 95% CI, 0.542–3.11). met—a counter-intuitive relationship occurs between 1424 Human and Experimental Toxicology 30(9) the number of vaccines given to infants and infant anaverageannualrateof8.6%.However,othercauses mortality rates: nations with higher (worse) infant of sudden unexpected infant death (SUID) increased. mortality rates give their infants, on average, more Forexample,thepostneonatalmortalityratefrom‘suf- vaccinedoses.Thispositivecorrelation,derivedfrom focationinbed’(ICD-9codeE913.0)increasedduring the data and demonstrated in Figures 1 and 2, elicits this same period at an average annual rate of 11.2%. an important inquiry: are some infant deaths associ- The postneonatal mortality rate from ‘suffocation- ated with over-vaccination? other’ (ICD-9 code E913.1-E913.9), ‘unknown and unspecified causes’ (ICD-9 code 799.9), and due to A closer inspection of infant deaths ‘intentunknown’intheExternalCausesofInjurysec- tion(ICD-9codeE980-E989),allincreasedduringthis Many nations adhere to an agreed upon International periodaswell.18(InAustralia,Mitchelletal.observed Classification of Diseases (ICD) for grouping infant thatwhentheSIDSratedecreased,deathsattributedto deathsinto130 categories.11–13 Among the34nations asphyxia increased.19 Overpeck et al. and others, analyzed, those that require the most vaccines tend to reported similar observations.)20,21 have the worst IMRs. Thus, we must ask important Acloserinspectionofthemorerecentperiodfrom questions:isitpossiblethatsomenationsarerequiring 1999 to 2001 reveals that the US postneonatal SIDS too many vaccines for their infants and the additional ratecontinuedtodecline,buttherewasnosignificant vaccines are a toxic burden on their health? Are some change in the total postneonatal mortality rate. Dur- deaths that are listed within the 130 infant mortality ing this period, the number of deaths attributed to death categories really deaths that are associated with ‘suffocationinbed’and‘unknowncauses,’increased over-vaccination? Are some vaccine-related deaths significantly. According to Malloy and MacDorman, hidden within the death tables? ‘‘If death-certifier preference has shifted such that previously classified SIDS deaths are now classified Sudden infant death syndrome (SIDS) as ‘suffocation,’ the inclusion of these suffocation Prior to contemporary vaccination programs, ‘Crib deathsandunknownorunspecifieddeathswithSIDS death’ was so infrequent that it was not mentioned deaths then accounts for about 90 percent of the in infant mortality statistics. In the United States, decline in the SIDS rate observed between 1999 and national immunization campaigns were initiated in 2001 and results in a non-significant decline in the1960swhenseveralnewvaccineswereintroduced SIDS’’18 (Figure 3). and actively recommended. For the first time in his- tory,mostUSinfantswererequiredtoreceiveseveral Is there evidence linking SIDS to vaccines? doses of DPT, polio, measles, mumps, and rubella vaccines.14 Shortly thereafter, in 1969, medical certi- Although some studies were unable to find correla- fiers presented a new medical term—sudden infant tions between SIDS and vaccines,22–24 there is some deathsyndrome.15,16In1973,theNationalCenterfor evidencethatasubsetofinfantsmaybemoresuscep- Health Statistics added a new cause-of-death cate- tible to SIDS shortly after being vaccinated. For gory—for SIDS—to the ICD. SIDS is defined as the example, Torch found that two-thirds of babies who sudden and unexpected death of an infant which haddiedfromSIDShadbeenvaccinatedagainstDPT remains unexplained after a thorough investigation. (diphtheria–pertussis–tetanus toxoid) prior to death. Although there are no specific symptoms associated Of these, 6.5% died within 12 hours of vaccination; with SIDS, an autopsy often reveals congestion and 13% within 24 hours; 26% within 3 days; and 37%, edema of the lungs and inflammatory changes in the 61%,and70%within1,2,and3weeks,respectively. respiratory system.17 By 1980, SIDS had become the Torch also found that unvaccinated babies who died leading cause of postneonatal mortality (deaths of of SIDS did so most often in the fall or winter while infants from 28 days to one year old) in the United vaccinated babies died most often at 2 and 4 States.18 months—the same ages when initial doses of DPT In1992,toaddresstheunacceptableSIDSrate,the were given to infants. He concluded that DPT ‘‘may American Academy of Pediatrics initiated a ‘Back to be a generally unrecognized major cause of sudden Sleep’ campaign, convincing parents to place their infant and early childhood death, and that the risks infants supine, rather than prone, during sleep. From of immunization may outweigh its potential benefits. 1992 to 2001, the postneonatal SIDS rate dropped by Aneedforre-evaluationandpossiblemodificationof Miller N Z and Goldman G S 1425 may also be cases of SIDS reclassified within the Reclassification of SIDS Deaths to Suffocation in Bed and Unknown Causes ICD. Some of these infant deaths may be vaccine 90 related as well. This trend toward reclassifying ICD data is a great concern of the CDC ‘‘because inaccu- 00 80 77.4 77.1 75.4 rate or inconsistent cause-of-death determination and 0 00, 70 reporting hamper the ability to monitor national 1 r 61.6 h pe 60 57.1 trends,ascertain riskfactors,anddesignandevaluate at 50.9 programs to prevent these deaths.’’29 If some infant e D 50 deaths are vaccine related and concealed within the 40 various ICD categories for SUIDs, is it possible that 1999 2000 2001 other vaccine-related infant deaths have also been reclassified? SIDS Of the 34 nations that have crossed the socio- SIDS + Suffocation + Unknown Causes economic threshold and are able to provide the basic necessitiesforinfantsurvival—cleanwater,nutrition, Figure3.Reclassificationofsuddeninfantdeathsyndrome sanitation, and health care—several require their (SIDS) deaths to suffocation in bed and unknown causes. infantstoreceivearelativelyhighnumberofvaccine ThepostneonatalSIDSrateappearstohavedeclinedfrom 61.6 deaths (per 100,000 live births) in 1999 to 50.9 in doses and have relatively high infant mortality rates. 2001. However, during this period there was a significant Thesenationsshouldtakeacloserlookattheirinfant increase in postneonatal deaths attributed to suffocation death tables to determine if some fatalities are possi- in bed and due to unknown causes. When these sudden bly related to vaccines though reclassified as other unexpectedinfantdeaths(SUIDs)arecombinedwithSIDS causes. Of course, all SUID categories should be re- deaths,thetotalSIDSrateremainsrelativelystable,result- inspected. Other ICD categories may be related to ingina non-significant decline. vaccinesaswell.Forexample,anewlive-virusorally administered vaccine against rotavirus-induced current vaccination procedures is indicated by this 1 study.’’25Walkeretal.found‘‘theSIDSmortalityrate diarrhea—Rotarix —was licensed by the European Medicine Agency in 2006 and approved by the US intheperiodzerotothreedaysfollowingDPTtobe7.3 Food and Drug Administration (FDA) in 2008. times that in the period beginning 30 days after immunization.’’26FineandChenreportedthatbabies However, in a clinical study that evaluated the safety of the Rotarix vaccine, vaccinated babies died at a died at a rate nearly eight times greater than normal within3daysaftergettingaDPTvaccination.27 higher rate than non-vaccinated babies—mainly due to a statistically significant increase in pneumonia- Ottavianietal.documentedthecaseofa3-month-old related fatalities.30(One biologicallyplausible expla- infantwhodiedsuddenlyandunexpectedlyshortlyafter nation is that natural rotavirusinfection might havea beinggivensixvaccinesinasingleshot:‘‘Examination protective effect against respiratory infection.)31 of the brainstem on serial sections revealed bilateral Although these fatalities appear to be vaccine related hypoplasiaofthearcuatenucleus.Thecardiacconduc- and raise a nation’s infant mortality rate, medical tion system presented persistent fetal dispersion and certifiers are likely to misclassify these deaths as resorptive degeneration. This case offers a unique pneumonia. insight into the possible role of hexavalent vaccine in SeveraladditionalICDcategoriesarepossiblecan- triggeringalethaloutcomeinavulnerablebaby.’’With- didates for incorrect infant death classifications: out a full necropsy study in the case of sudden, unex- unspecified viral diseases, diseases of the blood, pected infant death, at least some cases linked to vaccinationarelikelytogoundetected.28 septicemia, diseases of the nervous system, anoxic brain damage, other diseases of the nervous system, diseases of the respiratory system, influenza, and Reclassified infant deaths unspecified diseases of the respiratory system. All It appears as though some infant deaths attributed to of these selected causes may be repositories of SIDSmaybevaccinerelated,perhapsassociatedwith vaccine-related infant deaths reclassified as common biochemical or synergistic toxicity due to over- fatalities. All nations—rich and poor, industrialized vaccination.Someinfants’deathscategorizedas‘suf- and developing—have an obligation to determine focation’ordueto‘unknownandunspecifiedcauses’ whether their immunization schedules are achieving 1426 Human and Experimental Toxicology 30(9) their desired goals. Progress on reducing infant of the nations in this study had coverage rates in the mortality rates should include monitoring vaccine 90%–99% range for the most commonly recom- schedulesandmedicalcertificationpracticestoascer- mended vaccines—DTaP, polio, hepatitis B, and Hib tain whether vaccine-related infant deaths are being (when these vaccines were included in the schedule). reclassified as ordinary mortality in the ICD. Therefore,thisfactorisunlikelytohaveimpactedthe analyses.9 Howmanyinfantscanbesavedwithanimproved IMR? Minority races Slight improvements in IMRs can make a substantial IthasbeenarguedthattheUSIMRispoorincompar- difference.In2009,therewereapproximately4.5mil- isontomanyothernationsbecauseAfrican–American lionlivebirthsand28,000infantdeathsintheUnited infants are at greater risk of dying relative to White States, resulting in an infant mortality rate of 6.22/ infants, perhaps due to genetic factors or disparities 1000.Ifhealthauthoritiescanfindawaytoreducethe inlivingstandards.However,in2006theUSIMRfor rate by 1/1000 (16%), the United States would rise infants of all races was 6.69 and the IMR for White in international rank from 34th to 31st and about infantswas5.56.13In2009,thisimprovedratewould 4500 infants would be saved. havemovedtheUnitedStatesupbyjustonerankinter- nationally,from34thplaceto33rdplace.8Inaddition, theIMRsforHispanicsofMexicandescentandAsian– Limitations of study and potential AmericansintheUnitedStatesaresignificantlylower confounding factors thantheIMRforWhites.6Thus,diverseIMRsamong differentracesintheUnitesStatesexertonlyamodest This analysis did not adjust for vaccine composition, influence over the United States’ international infant national vaccine coverage rates, variations in the mortalityrank. infant mortality rates among minority races, preterm births, differences in how some nations report live births,orthepotentialforecologicalbias.Afewcom- Preterm births mentsabouteachofthesefactorsareincludedbelow. Preterm birth rates in the United States have steadily increased since the early 1980s. (This rise has been Vaccine composition tied to a greater reliance on caesarian deliveries, This analysis calculated the total number of vaccine inducedlabor,andmorebirthstooldermothers.)Pre- doses received by children but did not differentiate term babies are more likely than full-term babies to between the substances, or quantities of those sub- die within the first year of life. About 12.4% of US stances, in each dose. Common vaccine substances births are preterm. In Europe, the prevalence rate of includeantigens(attenuatedviruses,bacteria,toxoids), premature birth ranges from 5.5% in Ireland to preservatives (thimerosal, benzethonium chloride, 11.4% in Austria. Preventing preterm births is essen- 2-phenoxyethanol, phenol), adjuvants (aluminum tial to lower infant mortality rates. However, it is salts), additives (ammonium sulfate, glycerin, sodium important to note that some nations such as Ireland borate, polysorbate 80, hydrochloric acid, sodium and Greece, which have very low preterm birth rates hydroxide, potassium chloride), stabilizers (fetal (5.5% and 6%, respectively) compared to the United bovine serum, monosodium glutamate, human serum States,requiretheirinfantstoreceivearelativelyhigh albumin,porcinegelatin),antibiotics(neomycin,strep- number of vaccine doses (23) and have correspond- tomycin, polymyxin B), and inactivating chemicals ingly high IMRs. Therefore, reducing preterm birth (formalin, glutaraldehyde, polyoxyethylene). For the rates is only part of the solution to reduce IMRs.6,32 purposes of this study, all vaccine doses were equally weighted. Differences in reporting live births Infant mortality rates in most countries are reported Vaccine coverage rates using WHO standards, which do not include any ref- No adjustment was made for national vaccine cover- erence to the duration of pregnancy or weight of the age rates—a percentage of the target population that infant,butdodefinea‘livebirth’asababybornwith received the recommended vaccines. However, most any signs of life for any length of time.12 However, Miller N Z and Goldman G S 1427 four nations in the dataset—France, the Czech giventoinfants.Whennationsweregroupedintofive Republic,theNetherlands,andIreland—donotreport different vaccine dose ranges (12–14, 15–17, 18–20, live births entirely consistent with WHO standards. 21–23, and 24–26), 98.3% of the total variance in These countries add an additional requirement that IMR was explained by the unweighted linear livebabiesmustalsobeatleast22weeksofgestation regression model. These findings demonstrate a orweighatleast500grams.Ifbabiesdonotmeetthis counter-intuitive relationship: nations that require requirement and die shortly after birth, they are morevaccinedosestendtohavehigherinfantmortal- reportedasstillbirths.Thisinconsistencyinreporting ity rates. live births artificially lowers the IMRs of these Efforts to reduce the relatively high US IMR have nations.32,33AccordingtotheCDC,‘‘Therearesome been elusive. Finding ways to lower preterm birth differences among countries in the reporting of very rates should be a high priority. However, preventing smallinfantswhomaydiesoonafterbirth.However, premature births is just a partial solution to reduce itappearsunlikelythatdifferencesinreportingarethe infant deaths. A closer inspection of correlations primary explanation for the United States’ relatively between vaccine doses, biochemical or synergistic low international ranking.’’32 Nevertheless, when the toxicity,andIMRs,isessential.Allnations—richand IMRsofFrance,theCzechRepublic,theNetherlands, poor, advanced and developing—have an obligation and Ireland were adjusted for known underreporting to determine whether their immunization schedules of live births and the 30 data pairs retested for are achieving their desired goals. significance, the correlation coefficient improved from 0.70 to 0.74 (95% CI, 0.52–0.87). Acknowledgments The authors wish to thank Gerard Jungman, PhD, Paul G. Ecological bias King, PhD, and Peter Calhoun for their assistance in reviewingthe manuscript andsharing their expertise. Ecological bias occurs when relationships among individuals are inferred from similar relationships Funding observed among groups (or nations). Although most This research received no specific grantfrom any funding of the nations in this study had 90%–99% of their agencyinthepublic,commercial,ornot-for-profitsectors. infants fully vaccinated, without additional data we do not know whether it is the vaccinated or unvacci- References natedinfantswhoaredyingininfancyathigherrates. 1. Wegman ME. Infant mortality in the 20th century, However, respiratory disturbances have been docu- dramatic but uneven progress. J Nutr 2001; 131: mentedin closeproximity toinfant vaccinations, and 401S–408S. lethal changes in the brainstem of a recently vacci- 2. BeckMA.Theroleofnutritioninviraldisease.JNutri nated baby have been observed. Since some infants Biochem1996; 7:683–690. may be more susceptible to SIDS shortly after being 3. ScrimshawNSandSanGiovanniJP.Synergismofnutri- vaccinated, and babies vaccinated against diarrhea tion, infection, and immunity: an overview. Am J Clin diedfrompneumoniaatastatisticallyhigherratethan Nutr1997; 66:464S–477S. non-vaccinatedbabies,thereisplausiblebiologicand 4. AndersonGF,HussayPS,FrognerBK,andWatersHR. causalevidencethattheobservedcorrelationbetween HealthspendingintheUnitedStatesandtherestofthe IMRs and the number of vaccine doses routinely industrializedworld.HealthAffairs2005;24:903–914. giventoinfantsshouldnotbedismissedasecological 5. MacDorman MF and Mathews TJ. Recent trends in bias. infant mortality inthe United States. NCHS DataBrief (CDC),no9.Hyattsville,MD,USA:NationalCenterfor Conclusion Health Statistics,2008. The US childhood immunization schedule requires 6. KentMM.Prematurebirthshelptoexplainhigherinfant 26vaccinedosesforinfantsagedlessthan1year,the mortality rate. Population Reference Bureau. most in the world, yet 33 nations have better IMRs. www.prb.org/articles/2009/prematurebirth.aspx Using linear regression, the immunization schedules (accessed December2009). of these 34 nations were examined and a correlation 7. Xu Jiaquan, Kochaneck KD, and Tejada-Vera B. coefficient of 0.70 (p < 0.0001) was found between Deaths: preliminary data for 2007. Natl Vital Stat Rep IMRs and the number of vaccine doses routinely 2009;58:6. 1428 Human and Experimental Toxicology 30(9) 8. CIA.Countrycomparison:infantmortalityrate(2009). infant death syndrome: no increased risk after TheWorldFactbook.www.cia.gov(accessed13April immunisation. Vaccine 2007; 25:336–340. 2010). 23. Stratton K, Almario DA, Wizemann TM, and 9. WHO/UNICEF.ImmunizationSummary:AStatistical McCormick MC. Immunization safety review: vacci- Reference Containing Data Through 2008 (The 2010 nations and sudden unexpected death in infancy. Edition).www.childinfo.org. Washington DC, USA: National Academies Press, 10. Up-to-date European vaccination schedules may be 2003. foundhere: www.euvac.net (accessed 13April2010). 24. Silvers LE, Ellenberg SS, Wise RP, Varricchio FE, 11. WHO. International Classification of Diseases, 9th Mootrey GT, and Salive ME. The epidemiology of Revision.Geneva,Switzerland:WorldHealthOrgani- fatalitiesreportedtothevaccineadverseeventreport- zation,1979. ing system 1990-1997. Pharmacoepidemiol Drug Saf 12. WHO. International Statistical Classification of Dis- 2001; 10:279–285. eases and Related Health Problems, 10th Revision. 25. TorchWC.Diphtheria-pertussis-tetanus(DPT)immu- Geneva,Switzerland:WorldHealthOrganization,1992. nization: a potential cause of the sudden infant death 13. CDC. Table 31. Number of infant deaths and infant syndrome (SIDS). American Academy of Neurology, mortalityratesfor130selectedcauses,byrace:United 34thAnnualMeeting,Apr25-May1,1982.Neurology States,2006.Natl Vital StatRep2009; 57:110–112. 32(4): pt.2. 14. IannelliV.Immunizationtimeline.KeepKidsHealthy. 26. Walker AM, Jick H, Perera DR, Thompson RS, and keepkidshealthy.com (accessed 21 April2010). Knauss TA. Diphtheria-tetanus-pertussis immuniza- 15. Bergman AB. The ‘‘Discovery’’ of Sudden Infant tion and sudden infant death syndrome. Am J Public Death Syndrome. New York, NY,USA: Praeger Pub- Health 1987;77:945–951. lishers, 1986. 27. Fine PE and Chen RT. Confounding in studies of 16. MacDormanMFandRosenbergHM.Trendsininfant adverse reactions to vaccines. Am J Epidemiol 1992; mortality by cause of death and other characteristics, 136: 121–135. 1960-88 (vital and health statistics), Volume 20. 28. Ottaviani G, Lavezze AM, and Matturri L. Sudden Hyattsville,MD,USA:NationalCenterforHealthSta- infantdeathsyndrome(SIDS)shortlyafterhexavalent tistics,U.S. Government Printing,1993. vaccination: another pathology in suspected SIDS? 17. NationalCenterforHealthStatistics.VitalStatisticsof Virchows Archiv 2006;448: 100–104. the United States 1988, Volume II, Mortality, Part A. 29. CDC. About the sudden unexpected infant death Washington,DC,USA: PublicHealth Service, 1991. investigation (SUIDI) reporting form. Department 18. MalloyMH andMacDorman M.Changes inthe clas- of Health and Human Services (accessed 20 May sification of sudden unexpected infant deaths: United 2010). 1 States,1992-2001.Pediatrics 2005; 115:1247–1253. 30. GlaxoSmithKline.Rotarix (RotavirusVaccine,Live, 19. Mitchell E, Krous HF, Donald T, and Byard RW. Oral) Oral Suspension. Product insert fromthe manu- Changing trends in the diagnosis of sudden infant facturer (April 2008):6. death.AmJForensicMedPathol2000;21:311–314. 31. FDA. Center for biologics evaluation and research, 20. OverpeckMD,BrennerRA,CosgroveC,TrumbleAC, vaccinesandrelatedbiologicalproductsadvisorycom- Kochanek K, and MacDorman M. National under mittee meeting (20February 2008):127–128. ascertainment of sudden unexpected infant deaths 32. MacDorman MF and Mathews TJ. Behind interna- associated with deaths of unknown cause. Pediatrics tional rankings of infant mortality: how the United 2002;109: 274–283. States compares with Europe. NCHS data brief, no 21. ByardRWandBealSM.Haschangingdiagnosticpre- 23.Hyattsville,MD,USA:NationalCenterforHealth ferencebeenresponsiblefortherecentfallinincidence Statistics, 2009. of sudden infant death syndrome in South Australia? 33. Euro-Peristat Project, with SCPE, Eurocat, Euroneo- JPediatr ChildHealth1995; 31:197–199. stat. European Perinatal Health Report: Data for 22. Vennemann MM, Butterfass-Bahloul T, Jorch G, 2004 (The 2008 Edition): Table 3.1:40. www. BrinkmannB,FindeisenM,SauerlandC,etal.Sudden europeristat.com.