- Email: [email protected]
Intravenous immune globulin use in newborns
Intravenous
immune
globulin
use in newborns
Harry R. Hill, MD, Ann 0. Shigeoka, MD, Luis Antonia Robert D. Christensen, MD Srtit L-&Y, Cirv 1‘t& Yn&m~
infirnts.
over~~heltning mtmber
of defects
of ctpsonic studies
in the txwnute’s
host de~~xse
sJ.s/c’tn have
ubnorn~aiities
haw
found
detccruhle adults.
hacteriui
that doses side effects.
Limited
decrease
itmunrc~.s.
ittttrtunotherrtr\-
studies
morhidi~\~
prophyiactic~
have
indicuted itntnune
infec.ti0n.s
as high This
freutment Optimul
giob~riitt in ti~rtntrti
inf~x~riott
inf~tnts
iti prettzttw(~
imtnlrttc
,qtc~btrlitr
What first stimulated our interest in GBS disease was a 22-year-old primipara who had a perfectly normal pregnancy and delivery, resulting in the birth of a healthy male infant. Shortly thereafter. the infant began to have some respiratory distress. By 10 hours of age. the infant had marked respiratory distress. which on x-ray films was interpreted as hyaline membrane disease. Despite maxima1 respirator settings and all the facilities of the intensive care unit, the infant died at 36 hours of age. The fulminant nature of the infection suggested to us the need for more than conventional antimicrobial therapy. This initiated our studies on immunotherapy of neonatal infection. During our early investigations (1974 to 1976). we specifically looked at opsonic antibody in patients who had GBS disease. We found that the vast majority ot persons who became infected lacked opsonins to the specific strain with which they had been infected.’ In contrast, infants who did not become infected often had antibody to their colonizing strain. Thus antibody deficiency appeared to be at least one risk factor for developing disease. From the Division of Clmical Immunology and Allergy and Human Development of Aging. Department of Pediatrics. and the Department of Pathology, University of Utah, Salt Lake City, Utah Supported in part hv LJ.S. Public Health Service grants AIl?iSti and A119094.. Reprint requests: Hay R. Hill, MD, Division of Clinical Immunology and Allergy. Department of Pediatrics. University 01 Utah, SO North Medical Dr.. Salt Lake City. UT X4132 l/O/14663
I.1 :A11 Lw;!
IVIG:
globulin
humun
it@nts
inJitt1t.s anti therap?
tttct\
Ftrrthern~ore. of‘ itlfktiott may
tn
invol\~c~
(‘I I? I.W$f~ ,L’OL 19K9:84:hI;~_‘J.
Abbreviutions
CBS:
immttne
the devc~ioptnetlt
to WC
itzfctnt.s witkw
in tertn
inJ~tttts.
uftemplrd
giohuiin.
rhose
Itutmm
and ,sotne term
animal rhrou yh rhr
have
itntnune
fo ptwnuture
ttirh
tnrrx .set’\‘cJ lo ptwwrt
of pruterm
deftcretq
E.vperitnet~tai
can he ovcrc,ome
itttrm~c~nor4.v
intravcnotrs
A
including
.se~~~rrti in~wtigtrtors with
ir1~~l.s c~otnparuhle .sL(g,qes’ thaf
from
munagetnett~
rl?th intruvenou.t
Kr~c rttfl! tteotiofe.\
md
understood.
of‘~~~~!\‘morphot~uc~iertr
c~otttpotwnts
us 750 mg / kg (‘(ttt br udmini.stc~red
in the literature of‘preterm
and Jirnc~tiott
compl~wettt
MD, and
to wwre
been desr~rihrd,
rhut mutt>’ of’ rhe.\e de/kct.s
re.su1f.s itI I&
und mortoiir\
str.sceptihir~
t11i.s are not cwnpiefel!,
in the number
ie\~eis oj rhe mujor
of intruvenou.s
or ptwent
(ire unique/\. fitr
in our Iuborntor~
treat
otws,
The reuson.s
trtrd depressed
ct~lntit~isrrcrtiotl
premtrrirre
ittfec~tiotts
antibodies,
lerrkocytes,
.swrw
and e.specirril~
hacteriui
Gonzales,
I
used
Group B streptococcus Intravenous immune globulin
At that time, working with Dr. Robert Hall in Kansas City, we began to assessthe effects of transfusion of whole blood into infants who appeared to be septic.’ Fig. 1 shows that if an infant receives whole blood from an adult donor who has antibody, the infant’s level of opsonic activity after transfusion is raised above that in the pretransfusion specimen. Unfortunately. the donor blood might not contain antibody or may have lower levels than that of the patient, and this may actually reduce the level of opsonic activity as shown in the right-hand panels of Fig. 1. Another drawback in trying to use whole blood or plasma in our experience is that the level of antibody in most adults is relatively low so that a very high volume of blood must be given to bring about a change in antibody titer. In our study the transfusion volume needed to effect a significant increase in opsonic activity was at least 40% of the infant’s total blood volume (Fig. 2). Despite that, a retrospective analysis of a very small number of infants showed a mortality rate of 0% in nine infants who demonstrated an antibody rise after transfusion. In contrast, of six infants who had no antibody rise after transfusion, the mortality rate was 50% (p < 0.04’). We next turned to an animal model of neonatal GBS 617
618
Hill et
J. ALLERGY CLIN. IMMUNOL. OCTOBER 1989
al. Infant pe-tmmhsii
f-h
-
Umpsanized
FIG. 1. Effect of fresh, whole neutrophil chemiluminescence.
blood transfusion (From Shigeoka
disease in which bacteria and IVIG were administered to neonatal rats. We believe that this is an excellent model for studying neonatal infection because the infant rat’s complement and antibody levels are similar to those found in the human neonate, and infant rats also have abnormalities of white cell movement and white cell kinetics that are very similar to those in the human neonate. In 1981 we showed that a modified serum globulin preparation resulted in protection against GBS infection in this neonatal rat model.3 The level of protection attained with 100 mg/kg of the modified preparation (40% to 50%) exceeded that attained with an intramuscular -y-globulin preparation made from the same plasma. Subsequently, we showed significant protective activity in IVIG against both GBS and Escherichia co/i Kl, two of the major organisms infecting neonates. Moreover, we found that a pH 4.25 material (Gamimune-N, Cutter Biological, West Haven, Conn.) had significantly greater protective activity against GBS than did a reduced and alkylated preparation .4 In other studies we have looked specifically at the in vitro opsonic activity of IVIG against a variety of bacterial pathogens.5 These studies were performed by adding adult human neutrophils, a source of complement, and IVIG to test tubes, allowing the organisms to be phagocytized, and then plating them out and doing colony count analysis to determine the titer of antibody that results in significant bacterial killing. In addition, we have evaluated the ability of these antibody preparations, combined with the organisms, to consume complement.5 In our studies with GBS,
UnopMnized
on opsonic activity to GBS as measured AO, et al. Lancet 1978;1:636-8.)
by
we found that antibody, complement, and functional white cells are absolutely essential for bacterial killing; we believe that is the case with most organisms. As shown in Figs. 3 and 4, we compared complement consumption with opsonic activity of IVIG against two strains of GBS. The first was a more sensitive strain, type IIIs, for which the IVIG had fairly good opsonic activity, with titers of 1: 160 to 1: 320, which resulted in excellent consumption of complement (Fig. 3). The second was a more resistant group B strain, IIIa, which had a higher amount of sialic acid in its type-specific capsule material.‘j More antibody is required to protect against this organism. As shown, the opsonic activity of the IVIG drops off at lower titers, and the bacteria opsonized in IVIG did not consume as much complement as did the IIIs strain (Fig. 4). Fig. 5 shows the activity of IVIG against Klebsieflu pneumoniae. Both opsonic activity and complement consumption are excellent. In fact these IVIG preparations have excellent activity against most of the gram-negative organisms that we have studied.’ Serratia marcescens is an interesting organism in that either antibody or complement alone results in essentially 100% neutrophil-mediated bacterial killing. Moreover, the organism consumes complement adequately through either the classic or alternative complement pathway.’ A plot of complement consumption versus polymorphonuclear leukocyte-mediated killing results in essentially a straight line graph with an R value of 0.85. Thus with most bacterial pathogens, there is a clear need for both IVIG and a functional level of
VOLUME NUMBER
WIG in rewborns
84 4, PART 2
complement to promote adequate opsonization and killing. When protection versus complement consumption is compared in the animal models, the R value is an even more striking 1.O, which indicates a very significant correlation between consumption of complement and the ability to protect these animals.’ In our opinion, another important function of IVIG is its ability to affect the kinetics of leukocytes in neonates. An adult rat given GBS very rapidly controls bacterial invasion, has few symptoms of infection, and never dies. However, in the neonatal animal, there is a marked dropoff in the neutrophil count after GBS inoculation, which parallels in many cases what happens in the infected human neonate.7 Immature cells are released into the circulation, and the marrow is unable to keep up with the consumption of neutrophils that occurs when infection is established. The storage pool of cells that will develop into mature neutrophils rapidly decrease\ in the neonatal animal, resulting in severe marrow depletion. profound neutropenia, and death. fn contrast, animals given IVIG or monoclonal GBS antibody mobilize cells from the neutrophil storage pool faster, get them to the site of infection sooner. and survive the infection.8. ‘) In an attempt to explain this phenomenon, we recently purified ;I fragment from the third component ofcomplemtent, C3d,g. “’ The effect of giving neonatal rats GBS in association with this fragment is shown in Fig. 6. The fragment alone generates little accumulation of leukocytes in the peritoneal cavities of neonatal rats. Inoculation of GBS alone resulted in a significant accumulation of neutrophils in the peritoneal cavity but only after approximately 4 hours. However, when both C3d,g and the bacteria were introduced simultaneously, there was a marked and very rapid influx of neutrophils. On the basis of these findings. we believe a major activity of IVIG is to combine with the infecting organism. trigger the complement system, and produce these leukocytemobilizing fragments. Such factors cause more prompt release of polymorphonuclear leukocytes from the marrow into the peripheral circulation, from where they are called into the local site of invasion by chemotactic factors. In studies just completed, we have given infusions of IVIG or an albumin solution randomly into 20 premature human infants.” These infants weighed an average of approximately 1000 gm. We administered 750 mgikg of IVIG (pH 4.25) (Gamimune-N) to 10 infants to cletermine whether this has any major effect on a variety of parameters. Heart rate. respiratory rate, blood gases, temperature. and urine output were compared and showed essentially no difference between those who received
619
‘z100% 1 80
i
I I
.
”
WITHO~,A~TIBODY FIG. 2. Association between volume of blood transfused into human neonates and increases in opsonic antibody to GBS. (From Hill RH. Diagnosis and treatment of sepsis in the neonate. In: Root RK, Sande MA, eds. Septic shock: contemporary issues in infectious diseases. New York: Churchill Livingstone, 1985:219-32.)
750 mglkg of IVIG and those who received a similar volume of an albumin control. Numerous chemistries and blood counts were compared, and the only change detected was a major drop in the red cell count 15 minutes after infusion of this volume of IVIG. This was rapidly corrected and can be explained on a dilutional basis. Serum IgG levels were determined in newborns who received infusions of this dose of IVIG or the albumin control. ” Those who received the albumin control showed no increase in IgG levels, whereas there was approximately a threefold increase in recipients of IVIG. This brought the level of IgG up approximately to adult levels, which one expects to see in term neonates. The level of IgG dropped off rapidly initially as equilibration within the body occurred, but there was still significant elevation 8 days after IVIG infusion.” GBS type-specific opsonic antibody levels are shown in Fig. 7. Since the vast majority of neonates are unable to leave the intensive care unit without the administration of antibiotics, we had to devise a tech-
620
Hill et al.
J. ALLERGY
CLIN. IMMUNOL. OCTOBER 1989
80 60
0 10
20
40
80
160 320 6401280
CC
IGIV Titer FIG. 3. Opsonic activity and complement consumption of IVIG (pH 4.25) with an antibodysensitive (S) type Ill GBS strain. (From Yang KD, et al. J Infect Dis 1989;159:701-7, University of Chicago, publisher.)
10
20
40
80
160 320 6401280
CC
IGIV Tlter FIG. 4. Opsonic activity and complement consumption of IVIG (pli 4.25) with resistant (R) type Ill GBS strain. (From Yang KD, et al. J Infect Dis 1989;159:701-7, Chicago, publisher.)
nique to measure opsonins using killed organisms rather than live ones. For this reason, we used an a&dine orange-crystal violet visual assay of phagocytic uptake.” We observed a marked increase in opsonic activity for GBS after IVIG infusion that peaked at 15 minutes and persisted for 1 week or more (Fig. 7). This indicates that IVIG supplies functional antibody to the infants. Very limited data are presently available on the efficacy of treating infected human neonates with IVIG. In 1986 Sidiropoulos et al. I3reported the results of a study of 35 neonates with sepsis who had positive
an antibodyUniversity of
blood culture findings. When antibiotics alone were used, 4 of 15 patients died, for a 27% mortality rate. When IVIG and antibiotics were given, 2 of every 20 infants died, for a 10% mortality rate. However, the differences were not statistically significant. When only premature infants weighing <2500 gm were considered, the mortality rate was 44% in those who received antibiotics alone versus 8% in those given IVIG plus antibiotics (p < 0.04). At present this is the only study in the literature that addresses therapy of established human neonatal infection with IVIG.
VOLUME NUMBER
IVIG jr. newborns
84 4. PAR r 2
621
80 60
10
20
40
80
160 320 6401280
CC
IGIV Titer FIG. 5. Opsonic of K. pneumoniae. publisher.)
activity and complement consumption of WIG (pH 4.25) with a strain (From Yang KD, et al. J Infect Dis 1989;159:701-7, University of Chicago,
; HOURS
;
; POSTINOCULATION
FIG. 6. Effect of C3d,g fragments on mobilization of leukocytes to peritoneal cavity of GBSinfected and noninfected neonatal rats. (From Shigeoka AO, et al. Am J Pathol 1988;133: 623-9.)
Several studies on prophylaxis of neonatal bacterial infections with IVIG have appeared in the literature. A prophylaxis study in premature infants from Saudi Arabia was interesting in that about one-third of the organisms were Salmonella; other etiologic agents included Klebsiella and E. coli.14 Since these are somewhat different organisms from those usually seen in our United States newborn intensive care units, one should be very cautious about transposing
the results of this study to the United States. Nonetheless, in the group (100) given either one or two doses of IVIG (120 mg/kg), the infection rate was 4%. In those who received no IVIG, the rate was 16% (p < 0.005). In another study, Sidiropoulos et a1.15infused large amounts of IVIG into mothers and tried to determine whether the cord blood showed evidence of placental transfer of antibody. Infants <32 weeks’ gestation
622
Hill et al.
J. ALLERGY
CLIN. IMMUNOL. OCTOBER 1989
GBS+IVIG+BBC+PMNs
Ll
“:I!
,
0
GBS+5% Patient’s Serum +PMNs
,
,
15’ 48h 72h Time Postinfusion
,GBS+PMNs
1 week
FIG. 7. Opsonic antibody to type III GBS in the sera of premature of 750 mg/kg of WIG (pH 4.25).
generally had no significant elevation of IgG despite infusion of 24 gm of IVIG to the mother each day for 5 days. In infants >32 weeks’ gestation, supposedly after the placenta had matured, there was significant elevation of IgG in the infant when the mother had received infusions of high-dose IVIG. Whether this type of therapy is warranted remains to be seen, since 60% to 70% of neonatal infections occur in premature infants. Rather than infusing IVIG into the mother, we will probably have to treat the infant directly in most cases. Chirico et al. l6 reported that 500 mgl kg/ wk of IVIG had no significant effect on the prophylaxis of infection in a group of high-risk infants of diverse gestational age. However, in infants who weighed <1500 gm, a significant reduction in total infections (51% vs 77%) was noted. Moreover, a 5% versus a 20% sepsis rate and a 2% versus a 15% mortality rate were noted in infants who received IVIG versus those who did not. These differences were statistically significant (p < 0.05). In a recent report from Rome, Stabile et a1.17used Venogamma Polivalente, presumably an Italian preparation, prophylactically in neonates of cl500 gm. Six doses of 500 mg/kg were administered during the first 28 days of life. Forty infants were studied in the treated and untreated groups. The rates of proved sepsis (12.5% vs 7.5%) and total mortality (17.5% vs 17.5%) did not differ significantly between the groups. The only reported U.S. study on prophylaxis of neonates with IVIG at present is that of Clapp et al.‘* These investigators attempted to maintain an IVIG level in preterm neonates of >700 mg/dl through repeated infusions of IVIG. In infants who received either placebo (3/37) or no IVIG therapy (9/85), the rate of infection ranged from 8% to 10%; no cases of
infants before and after infusion
sepsis occurred among 41 infants given IVIG. Clapp et al.” indicated that approximately 500 persons would have to be studied before the results would reach significance. At least three large, multicenter, IVIG prophylaxis studies are currently underway in human neonates. Preliminary analysis of these studies is expected shortly. It is hoped that these studies may allow us to judge the efficacy of IVIG in the prophylaxis of bacterial infections in high-risk newborns. The role of IVIG in the therapy of established neonatal infection must still be determined. Another interesting potential use for IVIG in neonates was recently reported from Vienna. Eibl et a1.19 administered an oral IVIG preparation containing 73% IgA, 25% IgG, and 2% IgM (Ingalubin) in a dose of 300 mg/day in three divided doses randomly to neonates. The treated group (n = 88) showed a significant reduction in the incidence of necrotizing enterocolitis (6 vs 0; p = 0.01) and a slight reduction in pneumonia (9 vs 5) that was not statistically significant. In conclusion, systemic bacterial infection in neonates is often associated with antibody deficiency and an inadequate acute cellular inflammatory response. The administration of IVIG to neonatal animals corrects many of these abnormalities and markedly improves survival. IVIG in doses of 500 to 700 mg / kg may be administered to premature and term human neonates without significant side effects in most instances. Preliminary studies suggest that IVIG administration may be warranted in the treatment and prophylaxis of bacterial infection in premature and perhaps mature infants. However, further studies are needed to confirm the efficacy of such therapy or prophylaxis and establish the true indications and optimal dose required to improve survival and decrease
VOLUME NUMBER
IVIG in newborns
84 4. PART 2
morbidity in these patients. If we are going to continue to decrease the morbidity and mortality from such serious and overwhelming infections in immunocompromised hosts, such as the human neonate, new antibiotics are not the answer, but rather, we need to find ways to reconstitute the host’s own defense mechanism. The administration of IVIG may be one way to accomplish this goal.
14.
ii.
16. We thank Nancy Augustine for excellent technical support and Jeannette Rejali for secretarial help. Don Morse prepared the illustrations. REFERENCES 1. Hemming VB. Hall RT, Rhodes PG, Shigeoka AO. Hill HR. Assessment of group B streptococcal opsonins in human and rabbit serum by neutrophil chemiluminescence. J Clin Invest 1976;58: 1379-87. 2. Shigeoka AO. Hall RT, Hill HR. Blood transfusion in group B streptococcal sepsis. Lancet 1978;1:636-8. 3. Santos JI, Shigeoka AO, Hill HR. Protective efficacy of a modified immune serum globulin in experimental group B streptococcal infection. J Pediatr 1981;99:873-9. 4. Hill HR, Batbras JM. Protective and opsonic activities of a native, pH 4.25 intravenous immunoglobulin G preparation against common bacterial pathogens. Rev Infect Dis 1986; 8:S396-S400. 5. Yang KD, Bathras JM, Shigeoka AO, James J, Pincus S, Hill HR. Mechanisms of bacterial opsonization by immune globulin intravenous: correlation of complement consumption with opsonic activity and protective efficacy-. J Infect Dis 1989;159:701-7. 6 Shigeoka AO. Rote NS, Santos Jl, Hill HR. Assessment of the virulence factors of group B streptococci: I. Correlation with sialic acid content. J Infect Dis 1983;147:857-63. 1 Christensen RD, MacFarlane JL, Taylor NL, Hill HR, Rothstein G. Blood and marrow neutrophils during experimental group B streptococcal infection: quantification of stem cell. proliferal.ive, storage and circulating pools. Pediatr Res 1982:16:549-53. 8 Christensen RD, Rothstein G, Hill HR, Pincus SH. The effect of hybridoma antibody administration upon neutrophil kinetics during experimental type III group B streptococci sepsis. Pcdiatr Rea 1983;17:795-9. 9. Harper TE, Christensen RD, Rothstein G, Hill HR. Effect of intravenous immunoglobulin G on neutrophil mobilization induced by complement fragments during experimental group B streptococcal infection in rats. Rev Infect Dis 1983;8:S401S308. 10. Shigeoka AO, Gobel RJ, Janata J, Hill HR. Neutrophil mobilization induced by complement fragments during experimental group B streptococcal infection. Am J Pathol 1988;133:623-9. Il. Christensen RD, Hardman T, Thornton J, Hill HR. A rdndomized. double-blind, placebo-controlled investigation of the safety of intravenous immune globulin administration to preterm neonates. Perinatology (in press). 12. Yang KD, Augustine NH, Gonzales LA, Bohnsack JF. Hill HR. Effects of fibronectin on the interaction of polymorphonuclear !eukocytes with unopsonized and antibody-opsonired bacteria. J Infect Dis 1988;185:823-30. 13. Sidiropoulos D, Boehme U, von Murait G, Morel1 A, Bardndun
I?.
IX.
19.
623
S. lmmunoglobuhn supplementation in prevention or treatment of neonatal sepsis. Pediatr Infect Dis 1986;S(supp! ?):Sl93-4. Haque KN, Zaidi MH, Haque SK, Bahakim H. El-Hazmi M. El-Swailam M. Intravenous immunoglobulin for prevention of sepsisin preterm and Iow brrthweight infants. Pcdiatr Infect Dis 1986;5:622-5. Sidiropoulous D, Herrmann U Jr. Morcil I\. van Muralt G. Barandun S. Transplacental passage of intravenous immunoglobulin in the last trimester of prcgnancv J Pcdiatr 1986: 109:505-8. Chirico G. Rondini G. Plebani A, Chiara A. Maasa M. tigario AG Intravenous gamma globulin therapy for prophylaxis of infection in high-risk neonates, J Pediatr 19X7:10:437-42. Stabile A. Miceli Sopo S, Romanelli V, Pastore M. Pescaresi MA. Intravenous immunoglobulin for prophylaxis of neonatal sepsis in premature infants. Arch Dis Child 198X:63:341-3. Clapp DW, Baley JE, Kliegman RM. et al. Randomized trial of IVIG to prevent nosocomial infections m preterm infantspreliminary data [Abstract]. Pediatr Res I988:73:47 1. Eibl MM. Wolf HM. Fumkranz H, Roscnkranz A. Prevention of necrotizing cnterocolitia m low-birth-werght intents by IgAIpG feeding. K Engl I Med 1988:310: l-7
DISCUSSION Dr. Stiehm. I believe that the most significant fact about IVIG in ncwboms is that there is a large double-blind, controlled study underway at Cleveland and several other centers that are using albumin and IVIci infusions. A third group of patients, whose parents did not consent to treatment, arc receiving nothing. Interestingly. the rate of intraventricular hemorrhage is considerably higher in both colloid groups-albumin and y-globulin---compared with those who were untreated. Before we can recommend this therapy routinely for sick newborn infants. we must be sure that it will not cauzc other problems. such as intraventricular hemorrhage. Dr. Hill. I totally agree with Dr. Stiehm’s words of caution. My understanding was one infant in the Cleveland study had intraventricular hemorrhage, but perhaps you have
more recent information. Dr. Stiehm. This is a preliminary study. Dr. Hill. I think we need to be very cautious about that. Dr. Ballow. Dr. Hill, you show very impressive data on the effects of passive administration of IVIG in your neonatal mouse model. Why do you think there is such a di-
chotomy in the animal model and these human preterm infants‘?
Dr. Hill. It is the difference between treatment and prophylaxis. A newborn infant who is delivered at 1000 gm is placed in the newborn intensive care unit with every known monitor placed on him. He is going to have all types of infections from the surrounding environment. Many of the infections are associated with catheters and probes and other invasive procedures. In my opinion the prophylaxis story is not going to be anything like the treatment story that we have had in animals. 1 feel that the hospital-acquired infection is not so much dependent on the immune system but on the invasive procedures that are carried out on the infant, and that is why we have not had strikingly good results with 1VIG in those patients
624 Hill et al.
Dr. Ballow. Have you considered the variability and specific antibody titers of IVIG preparations relative to the nature of the infections that these newborns receive? Obviously you can control that in your animal models. Dr. Hill. It is critical that one have specific antibody; if not, there is no protection at all in most instances. Unquestionably there are also differences between preparations. Another question is the role of antibody in protecting against some of these nosocomial pathogens. In Scu$zy/ococcus epidermidis, for instance, we do not even know what the antibody and complement requirements are to protect against that organism. White cells can ingest S. epidermidis without antibody in some cases. Dr. Ballow. Thus these clinical studies are really not controlling for the variability between bottle-to-bottle and specific antibodies to these particular nosocomial infections. Also, have you looked at an animal model for prophylaxis or treatment of other bacteria, such as E. coli or S. epidermidis? Dr. Hill. We have looked at E. coli Kl in particular and have seen reasonable protective activity in commercially available IVIG preparations. Others have also looked at this and have shown activity against E. coli Kl. Against Klebsiellu there is good activity, and we have looked at that in the treatment study and found that it works well. I do not know the role of antibody in Candida or S. epidermidis infections, which one would be expected to contract from intravenous devices. Dr. Powell. Most of the work you have presented is focused on the sick premature newborn, particularly the very small infant. Dr. Hill. That is right. Dr. Powell. A small study was done on one of the Indian reservations with a preparation of hyperimmune serum. Adults were immunized with capsular polysaccharide, which showed efficacy in preventing Haemophilus injluenzae disease and a trend toward preventing Streptococcus pneumoniae. Do you know of other work going on? We have a very small population of sick premature infants, but we have a much larger population of normal newborns who have a nadir in their passive immunity a little bit later and are at risk for a host of common diseases. Dr. Hill. The study with the bacterial polysaccharide immunoglobulin was very dramatic. It was an intramuscular preparation and seemed to work very well. I omitted that in the interest of time. In a very high-risk population, such as the children you describe, who have an incredible incidence of pneumococcal and H. injluenzae disease, IVIG therapy may come into play in the relatively near future. Dr. Kim. Regarding the mechanisms responsible for the protective efficacy of commercially available IVIG against E. coli Kl, there are two antibodies important in E. coli Kl. One is anticapsular, and in certain serotypes of 0 antigens, antisomatic antibody is also protective against E. coli Kl. In your study you used a serotype of Kl but O-85 antigen. Am I correct or have you used different serotypes? Would you elaborate on the mechanisms for antisomatic antibody? Dr. Hill. I believe we have used more than one strain, and I do not know the 0 type of all the strains we have used. What is interesting about E. coli is that although we do not see very good opsonic activity in IVIG, we see
J. ALLERGY
CLIN. IMMUNOL. OCTOBER 1989
reasonable protective activity. It may be that with higher concentrations of antibody and high levels of complement in the serum, you can actually lyse the organism directly in some cases. Dr. Schwartz. I am intrigued by the oral study to which you referred. I presume that was not a serum preparation. Was that a colostrum preparation? Do you know anything about it? Dr. Hill. It is a serum preparation from Vienna. It had a little bit of IgG in it and almost no 1gM. I do not believe that it was a colostrum preparation. Dr. Schwartz. With regard to the findings with this preparation, would you speculate on whether there might be a direct antimicrobial effect at the enteric level or might it be immune modulation as has been implied in some of the discussions at this symposium? Second, are you aware of any controls that have been done with oral IVIG in neonates? Dr. Hill. I believe a study was done at Johns Hopkins in which infants were given oral y-globulin and examined for rotavirus levels in the stool. Treated infants showed significant shedding of rotavirus in the stool after oral yglobulin. I am unaware of any other studies in newborn infants presently, and 1 do not know the mechanism of study from Vienna. Maybe it is affecting attachment. We are very interested in IgA as a possible active antibacterial agent, We have shown some protection when you include IgA with fibronectin; it can actually opsonize organisms under these circumstances. Dr. Schwartz. We have used bovine colostrum, as had been described from the National Cancer Institute group, about 10 years ago for patients with selective IgA deficiency. We have been impressed with the reduction of clinical gastrointestinal problems that these patients have and speculated about how that might work-whether it is direct antimicrobial modulation. It does not affect serum IgA. Dr. Stiehm. Another word of caution. Jim Bussel did a study on the use of IVIG during the newborn period and gave an enormous dose of y-globulin in the first week of life. He observed efficacy for the first month of life and showed that there was less infection. However, in the second month of life, the rate of infection increased so that the overall rate of infection was exactly the same. That suggests that there was a feedback inhibition of antibody synthesis so that at the second month of life, infants were actually worse than they would have been had they not been treated. Thus I think another caution in the use of IVIG is how long it should be continued because one may get passive inhibition of antibody production thereafter. Dr. Gardner. In view of some of the case reports on treatment of enteroviral infections with IVIG and some of the data on babies with respiratory syncytial virus, do you think that there might be any role for IVIG in newborns with severe neonatal enteroviral infections or severe respiratory syncytial virus in the nosocomial situation, in addition to ribavirin? Dr. Hill. The answer is yes; there may well be a role for that, but I do not feel that the current studies are really conclusive. Studies by Fischer of respiratory syncytial virus are ongoing. They seem to have some significant benefits in a few animal models, but data on humans are not in as yet.
immune
globulin
use in newborns
Harry R. Hill, MD, Ann 0. Shigeoka, MD, Luis Antonia Robert D. Christensen, MD Srtit L-&Y, Cirv 1‘t& Yn&m~
infirnts.
over~~heltning mtmber
of defects
of ctpsonic studies
in the txwnute’s
host de~~xse
sJ.s/c’tn have
ubnorn~aiities
haw
found
detccruhle adults.
hacteriui
that doses side effects.
Limited
decrease
itmunrc~.s.
ittttrtunotherrtr\-
studies
morhidi~\~
prophyiactic~
have
indicuted itntnune
infec.ti0n.s
as high This
freutment Optimul
giob~riitt in ti~rtntrti
inf~x~riott
inf~tnts
iti prettzttw(~
imtnlrttc
,qtc~btrlitr
What first stimulated our interest in GBS disease was a 22-year-old primipara who had a perfectly normal pregnancy and delivery, resulting in the birth of a healthy male infant. Shortly thereafter. the infant began to have some respiratory distress. By 10 hours of age. the infant had marked respiratory distress. which on x-ray films was interpreted as hyaline membrane disease. Despite maxima1 respirator settings and all the facilities of the intensive care unit, the infant died at 36 hours of age. The fulminant nature of the infection suggested to us the need for more than conventional antimicrobial therapy. This initiated our studies on immunotherapy of neonatal infection. During our early investigations (1974 to 1976). we specifically looked at opsonic antibody in patients who had GBS disease. We found that the vast majority ot persons who became infected lacked opsonins to the specific strain with which they had been infected.’ In contrast, infants who did not become infected often had antibody to their colonizing strain. Thus antibody deficiency appeared to be at least one risk factor for developing disease. From the Division of Clmical Immunology and Allergy and Human Development of Aging. Department of Pediatrics. and the Department of Pathology, University of Utah, Salt Lake City, Utah Supported in part hv LJ.S. Public Health Service grants AIl?iSti and A119094.. Reprint requests: Hay R. Hill, MD, Division of Clinical Immunology and Allergy. Department of Pediatrics. University 01 Utah, SO North Medical Dr.. Salt Lake City. UT X4132 l/O/14663
I.1 :A11 Lw;!
IVIG:
globulin
humun
it@nts
inJitt1t.s anti therap?
tttct\
Ftrrthern~ore. of‘ itlfktiott may
tn
invol\~c~
(‘I I? I.W$f~ ,L’OL 19K9:84:hI;~_‘J.
Abbreviutions
CBS:
immttne
the devc~ioptnetlt
to WC
itzfctnt.s witkw
in tertn
inJ~tttts.
uftemplrd
giohuiin.
rhose
Itutmm
and ,sotne term
animal rhrou yh rhr
have
itntnune
fo ptwnuture
ttirh
tnrrx .set’\‘cJ lo ptwwrt
of pruterm
deftcretq
E.vperitnet~tai
can he ovcrc,ome
itttrm~c~nor4.v
intravcnotrs
A
including
.se~~~rrti in~wtigtrtors with
ir1~~l.s c~otnparuhle .sL(g,qes’ thaf
from
munagetnett~
rl?th intruvenou.t
Kr~c rttfl! tteotiofe.\
md
understood.
of‘~~~~!\‘morphot~uc~iertr
c~otttpotwnts
us 750 mg / kg (‘(ttt br udmini.stc~red
in the literature of‘preterm
and Jirnc~tiott
compl~wettt
MD, and
to wwre
been desr~rihrd,
rhut mutt>’ of’ rhe.\e de/kct.s
re.su1f.s itI I&
und mortoiir\
str.sceptihir~
t11i.s are not cwnpiefel!,
in the number
ie\~eis oj rhe mujor
of intruvenou.s
or ptwent
(ire unique/\. fitr
in our Iuborntor~
treat
otws,
The reuson.s
trtrd depressed
ct~lntit~isrrcrtiotl
premtrrirre
ittfec~tiotts
antibodies,
lerrkocytes,
.swrw
and e.specirril~
hacteriui
Gonzales,
I
used
Group B streptococcus Intravenous immune globulin
At that time, working with Dr. Robert Hall in Kansas City, we began to assessthe effects of transfusion of whole blood into infants who appeared to be septic.’ Fig. 1 shows that if an infant receives whole blood from an adult donor who has antibody, the infant’s level of opsonic activity after transfusion is raised above that in the pretransfusion specimen. Unfortunately. the donor blood might not contain antibody or may have lower levels than that of the patient, and this may actually reduce the level of opsonic activity as shown in the right-hand panels of Fig. 1. Another drawback in trying to use whole blood or plasma in our experience is that the level of antibody in most adults is relatively low so that a very high volume of blood must be given to bring about a change in antibody titer. In our study the transfusion volume needed to effect a significant increase in opsonic activity was at least 40% of the infant’s total blood volume (Fig. 2). Despite that, a retrospective analysis of a very small number of infants showed a mortality rate of 0% in nine infants who demonstrated an antibody rise after transfusion. In contrast, of six infants who had no antibody rise after transfusion, the mortality rate was 50% (p < 0.04’). We next turned to an animal model of neonatal GBS 617
618
Hill et
J. ALLERGY CLIN. IMMUNOL. OCTOBER 1989
al. Infant pe-tmmhsii
f-h
-
Umpsanized
FIG. 1. Effect of fresh, whole neutrophil chemiluminescence.
blood transfusion (From Shigeoka
disease in which bacteria and IVIG were administered to neonatal rats. We believe that this is an excellent model for studying neonatal infection because the infant rat’s complement and antibody levels are similar to those found in the human neonate, and infant rats also have abnormalities of white cell movement and white cell kinetics that are very similar to those in the human neonate. In 1981 we showed that a modified serum globulin preparation resulted in protection against GBS infection in this neonatal rat model.3 The level of protection attained with 100 mg/kg of the modified preparation (40% to 50%) exceeded that attained with an intramuscular -y-globulin preparation made from the same plasma. Subsequently, we showed significant protective activity in IVIG against both GBS and Escherichia co/i Kl, two of the major organisms infecting neonates. Moreover, we found that a pH 4.25 material (Gamimune-N, Cutter Biological, West Haven, Conn.) had significantly greater protective activity against GBS than did a reduced and alkylated preparation .4 In other studies we have looked specifically at the in vitro opsonic activity of IVIG against a variety of bacterial pathogens.5 These studies were performed by adding adult human neutrophils, a source of complement, and IVIG to test tubes, allowing the organisms to be phagocytized, and then plating them out and doing colony count analysis to determine the titer of antibody that results in significant bacterial killing. In addition, we have evaluated the ability of these antibody preparations, combined with the organisms, to consume complement.5 In our studies with GBS,
UnopMnized
on opsonic activity to GBS as measured AO, et al. Lancet 1978;1:636-8.)
by
we found that antibody, complement, and functional white cells are absolutely essential for bacterial killing; we believe that is the case with most organisms. As shown in Figs. 3 and 4, we compared complement consumption with opsonic activity of IVIG against two strains of GBS. The first was a more sensitive strain, type IIIs, for which the IVIG had fairly good opsonic activity, with titers of 1: 160 to 1: 320, which resulted in excellent consumption of complement (Fig. 3). The second was a more resistant group B strain, IIIa, which had a higher amount of sialic acid in its type-specific capsule material.‘j More antibody is required to protect against this organism. As shown, the opsonic activity of the IVIG drops off at lower titers, and the bacteria opsonized in IVIG did not consume as much complement as did the IIIs strain (Fig. 4). Fig. 5 shows the activity of IVIG against Klebsieflu pneumoniae. Both opsonic activity and complement consumption are excellent. In fact these IVIG preparations have excellent activity against most of the gram-negative organisms that we have studied.’ Serratia marcescens is an interesting organism in that either antibody or complement alone results in essentially 100% neutrophil-mediated bacterial killing. Moreover, the organism consumes complement adequately through either the classic or alternative complement pathway.’ A plot of complement consumption versus polymorphonuclear leukocyte-mediated killing results in essentially a straight line graph with an R value of 0.85. Thus with most bacterial pathogens, there is a clear need for both IVIG and a functional level of
VOLUME NUMBER
WIG in rewborns
84 4, PART 2
complement to promote adequate opsonization and killing. When protection versus complement consumption is compared in the animal models, the R value is an even more striking 1.O, which indicates a very significant correlation between consumption of complement and the ability to protect these animals.’ In our opinion, another important function of IVIG is its ability to affect the kinetics of leukocytes in neonates. An adult rat given GBS very rapidly controls bacterial invasion, has few symptoms of infection, and never dies. However, in the neonatal animal, there is a marked dropoff in the neutrophil count after GBS inoculation, which parallels in many cases what happens in the infected human neonate.7 Immature cells are released into the circulation, and the marrow is unable to keep up with the consumption of neutrophils that occurs when infection is established. The storage pool of cells that will develop into mature neutrophils rapidly decrease\ in the neonatal animal, resulting in severe marrow depletion. profound neutropenia, and death. fn contrast, animals given IVIG or monoclonal GBS antibody mobilize cells from the neutrophil storage pool faster, get them to the site of infection sooner. and survive the infection.8. ‘) In an attempt to explain this phenomenon, we recently purified ;I fragment from the third component ofcomplemtent, C3d,g. “’ The effect of giving neonatal rats GBS in association with this fragment is shown in Fig. 6. The fragment alone generates little accumulation of leukocytes in the peritoneal cavities of neonatal rats. Inoculation of GBS alone resulted in a significant accumulation of neutrophils in the peritoneal cavity but only after approximately 4 hours. However, when both C3d,g and the bacteria were introduced simultaneously, there was a marked and very rapid influx of neutrophils. On the basis of these findings. we believe a major activity of IVIG is to combine with the infecting organism. trigger the complement system, and produce these leukocytemobilizing fragments. Such factors cause more prompt release of polymorphonuclear leukocytes from the marrow into the peripheral circulation, from where they are called into the local site of invasion by chemotactic factors. In studies just completed, we have given infusions of IVIG or an albumin solution randomly into 20 premature human infants.” These infants weighed an average of approximately 1000 gm. We administered 750 mgikg of IVIG (pH 4.25) (Gamimune-N) to 10 infants to cletermine whether this has any major effect on a variety of parameters. Heart rate. respiratory rate, blood gases, temperature. and urine output were compared and showed essentially no difference between those who received
619
‘z100% 1 80
i
I I
.
”
WITHO~,A~TIBODY FIG. 2. Association between volume of blood transfused into human neonates and increases in opsonic antibody to GBS. (From Hill RH. Diagnosis and treatment of sepsis in the neonate. In: Root RK, Sande MA, eds. Septic shock: contemporary issues in infectious diseases. New York: Churchill Livingstone, 1985:219-32.)
750 mglkg of IVIG and those who received a similar volume of an albumin control. Numerous chemistries and blood counts were compared, and the only change detected was a major drop in the red cell count 15 minutes after infusion of this volume of IVIG. This was rapidly corrected and can be explained on a dilutional basis. Serum IgG levels were determined in newborns who received infusions of this dose of IVIG or the albumin control. ” Those who received the albumin control showed no increase in IgG levels, whereas there was approximately a threefold increase in recipients of IVIG. This brought the level of IgG up approximately to adult levels, which one expects to see in term neonates. The level of IgG dropped off rapidly initially as equilibration within the body occurred, but there was still significant elevation 8 days after IVIG infusion.” GBS type-specific opsonic antibody levels are shown in Fig. 7. Since the vast majority of neonates are unable to leave the intensive care unit without the administration of antibiotics, we had to devise a tech-
620
Hill et al.
J. ALLERGY
CLIN. IMMUNOL. OCTOBER 1989
80 60
0 10
20
40
80
160 320 6401280
CC
IGIV Titer FIG. 3. Opsonic activity and complement consumption of IVIG (pH 4.25) with an antibodysensitive (S) type Ill GBS strain. (From Yang KD, et al. J Infect Dis 1989;159:701-7, University of Chicago, publisher.)
10
20
40
80
160 320 6401280
CC
IGIV Tlter FIG. 4. Opsonic activity and complement consumption of IVIG (pli 4.25) with resistant (R) type Ill GBS strain. (From Yang KD, et al. J Infect Dis 1989;159:701-7, Chicago, publisher.)
nique to measure opsonins using killed organisms rather than live ones. For this reason, we used an a&dine orange-crystal violet visual assay of phagocytic uptake.” We observed a marked increase in opsonic activity for GBS after IVIG infusion that peaked at 15 minutes and persisted for 1 week or more (Fig. 7). This indicates that IVIG supplies functional antibody to the infants. Very limited data are presently available on the efficacy of treating infected human neonates with IVIG. In 1986 Sidiropoulos et al. I3reported the results of a study of 35 neonates with sepsis who had positive
an antibodyUniversity of
blood culture findings. When antibiotics alone were used, 4 of 15 patients died, for a 27% mortality rate. When IVIG and antibiotics were given, 2 of every 20 infants died, for a 10% mortality rate. However, the differences were not statistically significant. When only premature infants weighing <2500 gm were considered, the mortality rate was 44% in those who received antibiotics alone versus 8% in those given IVIG plus antibiotics (p < 0.04). At present this is the only study in the literature that addresses therapy of established human neonatal infection with IVIG.
VOLUME NUMBER
IVIG jr. newborns
84 4. PAR r 2
621
80 60
10
20
40
80
160 320 6401280
CC
IGIV Titer FIG. 5. Opsonic of K. pneumoniae. publisher.)
activity and complement consumption of WIG (pH 4.25) with a strain (From Yang KD, et al. J Infect Dis 1989;159:701-7, University of Chicago,
; HOURS
;
; POSTINOCULATION
FIG. 6. Effect of C3d,g fragments on mobilization of leukocytes to peritoneal cavity of GBSinfected and noninfected neonatal rats. (From Shigeoka AO, et al. Am J Pathol 1988;133: 623-9.)
Several studies on prophylaxis of neonatal bacterial infections with IVIG have appeared in the literature. A prophylaxis study in premature infants from Saudi Arabia was interesting in that about one-third of the organisms were Salmonella; other etiologic agents included Klebsiella and E. coli.14 Since these are somewhat different organisms from those usually seen in our United States newborn intensive care units, one should be very cautious about transposing
the results of this study to the United States. Nonetheless, in the group (100) given either one or two doses of IVIG (120 mg/kg), the infection rate was 4%. In those who received no IVIG, the rate was 16% (p < 0.005). In another study, Sidiropoulos et a1.15infused large amounts of IVIG into mothers and tried to determine whether the cord blood showed evidence of placental transfer of antibody. Infants <32 weeks’ gestation
622
Hill et al.
J. ALLERGY
CLIN. IMMUNOL. OCTOBER 1989
GBS+IVIG+BBC+PMNs
Ll
“:I!
,
0
GBS+5% Patient’s Serum +PMNs
,
,
15’ 48h 72h Time Postinfusion
,GBS+PMNs
1 week
FIG. 7. Opsonic antibody to type III GBS in the sera of premature of 750 mg/kg of WIG (pH 4.25).
generally had no significant elevation of IgG despite infusion of 24 gm of IVIG to the mother each day for 5 days. In infants >32 weeks’ gestation, supposedly after the placenta had matured, there was significant elevation of IgG in the infant when the mother had received infusions of high-dose IVIG. Whether this type of therapy is warranted remains to be seen, since 60% to 70% of neonatal infections occur in premature infants. Rather than infusing IVIG into the mother, we will probably have to treat the infant directly in most cases. Chirico et al. l6 reported that 500 mgl kg/ wk of IVIG had no significant effect on the prophylaxis of infection in a group of high-risk infants of diverse gestational age. However, in infants who weighed <1500 gm, a significant reduction in total infections (51% vs 77%) was noted. Moreover, a 5% versus a 20% sepsis rate and a 2% versus a 15% mortality rate were noted in infants who received IVIG versus those who did not. These differences were statistically significant (p < 0.05). In a recent report from Rome, Stabile et a1.17used Venogamma Polivalente, presumably an Italian preparation, prophylactically in neonates of cl500 gm. Six doses of 500 mg/kg were administered during the first 28 days of life. Forty infants were studied in the treated and untreated groups. The rates of proved sepsis (12.5% vs 7.5%) and total mortality (17.5% vs 17.5%) did not differ significantly between the groups. The only reported U.S. study on prophylaxis of neonates with IVIG at present is that of Clapp et al.‘* These investigators attempted to maintain an IVIG level in preterm neonates of >700 mg/dl through repeated infusions of IVIG. In infants who received either placebo (3/37) or no IVIG therapy (9/85), the rate of infection ranged from 8% to 10%; no cases of
infants before and after infusion
sepsis occurred among 41 infants given IVIG. Clapp et al.” indicated that approximately 500 persons would have to be studied before the results would reach significance. At least three large, multicenter, IVIG prophylaxis studies are currently underway in human neonates. Preliminary analysis of these studies is expected shortly. It is hoped that these studies may allow us to judge the efficacy of IVIG in the prophylaxis of bacterial infections in high-risk newborns. The role of IVIG in the therapy of established neonatal infection must still be determined. Another interesting potential use for IVIG in neonates was recently reported from Vienna. Eibl et a1.19 administered an oral IVIG preparation containing 73% IgA, 25% IgG, and 2% IgM (Ingalubin) in a dose of 300 mg/day in three divided doses randomly to neonates. The treated group (n = 88) showed a significant reduction in the incidence of necrotizing enterocolitis (6 vs 0; p = 0.01) and a slight reduction in pneumonia (9 vs 5) that was not statistically significant. In conclusion, systemic bacterial infection in neonates is often associated with antibody deficiency and an inadequate acute cellular inflammatory response. The administration of IVIG to neonatal animals corrects many of these abnormalities and markedly improves survival. IVIG in doses of 500 to 700 mg / kg may be administered to premature and term human neonates without significant side effects in most instances. Preliminary studies suggest that IVIG administration may be warranted in the treatment and prophylaxis of bacterial infection in premature and perhaps mature infants. However, further studies are needed to confirm the efficacy of such therapy or prophylaxis and establish the true indications and optimal dose required to improve survival and decrease
VOLUME NUMBER
IVIG in newborns
84 4. PART 2
morbidity in these patients. If we are going to continue to decrease the morbidity and mortality from such serious and overwhelming infections in immunocompromised hosts, such as the human neonate, new antibiotics are not the answer, but rather, we need to find ways to reconstitute the host’s own defense mechanism. The administration of IVIG may be one way to accomplish this goal.
14.
ii.
16. We thank Nancy Augustine for excellent technical support and Jeannette Rejali for secretarial help. Don Morse prepared the illustrations. REFERENCES 1. Hemming VB. Hall RT, Rhodes PG, Shigeoka AO. Hill HR. Assessment of group B streptococcal opsonins in human and rabbit serum by neutrophil chemiluminescence. J Clin Invest 1976;58: 1379-87. 2. Shigeoka AO. Hall RT, Hill HR. Blood transfusion in group B streptococcal sepsis. Lancet 1978;1:636-8. 3. Santos JI, Shigeoka AO, Hill HR. Protective efficacy of a modified immune serum globulin in experimental group B streptococcal infection. J Pediatr 1981;99:873-9. 4. Hill HR, Batbras JM. Protective and opsonic activities of a native, pH 4.25 intravenous immunoglobulin G preparation against common bacterial pathogens. Rev Infect Dis 1986; 8:S396-S400. 5. Yang KD, Bathras JM, Shigeoka AO, James J, Pincus S, Hill HR. Mechanisms of bacterial opsonization by immune globulin intravenous: correlation of complement consumption with opsonic activity and protective efficacy-. J Infect Dis 1989;159:701-7. 6 Shigeoka AO. Rote NS, Santos Jl, Hill HR. Assessment of the virulence factors of group B streptococci: I. Correlation with sialic acid content. J Infect Dis 1983;147:857-63. 1 Christensen RD, MacFarlane JL, Taylor NL, Hill HR, Rothstein G. Blood and marrow neutrophils during experimental group B streptococcal infection: quantification of stem cell. proliferal.ive, storage and circulating pools. Pediatr Res 1982:16:549-53. 8 Christensen RD, Rothstein G, Hill HR, Pincus SH. The effect of hybridoma antibody administration upon neutrophil kinetics during experimental type III group B streptococci sepsis. Pcdiatr Rea 1983;17:795-9. 9. Harper TE, Christensen RD, Rothstein G, Hill HR. Effect of intravenous immunoglobulin G on neutrophil mobilization induced by complement fragments during experimental group B streptococcal infection in rats. Rev Infect Dis 1983;8:S401S308. 10. Shigeoka AO, Gobel RJ, Janata J, Hill HR. Neutrophil mobilization induced by complement fragments during experimental group B streptococcal infection. Am J Pathol 1988;133:623-9. Il. Christensen RD, Hardman T, Thornton J, Hill HR. A rdndomized. double-blind, placebo-controlled investigation of the safety of intravenous immune globulin administration to preterm neonates. Perinatology (in press). 12. Yang KD, Augustine NH, Gonzales LA, Bohnsack JF. Hill HR. Effects of fibronectin on the interaction of polymorphonuclear !eukocytes with unopsonized and antibody-opsonired bacteria. J Infect Dis 1988;185:823-30. 13. Sidiropoulos D, Boehme U, von Murait G, Morel1 A, Bardndun
I?.
IX.
19.
623
S. lmmunoglobuhn supplementation in prevention or treatment of neonatal sepsis. Pediatr Infect Dis 1986;S(supp! ?):Sl93-4. Haque KN, Zaidi MH, Haque SK, Bahakim H. El-Hazmi M. El-Swailam M. Intravenous immunoglobulin for prevention of sepsisin preterm and Iow brrthweight infants. Pcdiatr Infect Dis 1986;5:622-5. Sidiropoulous D, Herrmann U Jr. Morcil I\. van Muralt G. Barandun S. Transplacental passage of intravenous immunoglobulin in the last trimester of prcgnancv J Pcdiatr 1986: 109:505-8. Chirico G. Rondini G. Plebani A, Chiara A. Maasa M. tigario AG Intravenous gamma globulin therapy for prophylaxis of infection in high-risk neonates, J Pediatr 19X7:10:437-42. Stabile A. Miceli Sopo S, Romanelli V, Pastore M. Pescaresi MA. Intravenous immunoglobulin for prophylaxis of neonatal sepsis in premature infants. Arch Dis Child 198X:63:341-3. Clapp DW, Baley JE, Kliegman RM. et al. Randomized trial of IVIG to prevent nosocomial infections m preterm infantspreliminary data [Abstract]. Pediatr Res I988:73:47 1. Eibl MM. Wolf HM. Fumkranz H, Roscnkranz A. Prevention of necrotizing cnterocolitia m low-birth-werght intents by IgAIpG feeding. K Engl I Med 1988:310: l-7
DISCUSSION Dr. Stiehm. I believe that the most significant fact about IVIG in ncwboms is that there is a large double-blind, controlled study underway at Cleveland and several other centers that are using albumin and IVIci infusions. A third group of patients, whose parents did not consent to treatment, arc receiving nothing. Interestingly. the rate of intraventricular hemorrhage is considerably higher in both colloid groups-albumin and y-globulin---compared with those who were untreated. Before we can recommend this therapy routinely for sick newborn infants. we must be sure that it will not cauzc other problems. such as intraventricular hemorrhage. Dr. Hill. I totally agree with Dr. Stiehm’s words of caution. My understanding was one infant in the Cleveland study had intraventricular hemorrhage, but perhaps you have
more recent information. Dr. Stiehm. This is a preliminary study. Dr. Hill. I think we need to be very cautious about that. Dr. Ballow. Dr. Hill, you show very impressive data on the effects of passive administration of IVIG in your neonatal mouse model. Why do you think there is such a di-
chotomy in the animal model and these human preterm infants‘?
Dr. Hill. It is the difference between treatment and prophylaxis. A newborn infant who is delivered at 1000 gm is placed in the newborn intensive care unit with every known monitor placed on him. He is going to have all types of infections from the surrounding environment. Many of the infections are associated with catheters and probes and other invasive procedures. In my opinion the prophylaxis story is not going to be anything like the treatment story that we have had in animals. 1 feel that the hospital-acquired infection is not so much dependent on the immune system but on the invasive procedures that are carried out on the infant, and that is why we have not had strikingly good results with 1VIG in those patients
624 Hill et al.
Dr. Ballow. Have you considered the variability and specific antibody titers of IVIG preparations relative to the nature of the infections that these newborns receive? Obviously you can control that in your animal models. Dr. Hill. It is critical that one have specific antibody; if not, there is no protection at all in most instances. Unquestionably there are also differences between preparations. Another question is the role of antibody in protecting against some of these nosocomial pathogens. In Scu$zy/ococcus epidermidis, for instance, we do not even know what the antibody and complement requirements are to protect against that organism. White cells can ingest S. epidermidis without antibody in some cases. Dr. Ballow. Thus these clinical studies are really not controlling for the variability between bottle-to-bottle and specific antibodies to these particular nosocomial infections. Also, have you looked at an animal model for prophylaxis or treatment of other bacteria, such as E. coli or S. epidermidis? Dr. Hill. We have looked at E. coli Kl in particular and have seen reasonable protective activity in commercially available IVIG preparations. Others have also looked at this and have shown activity against E. coli Kl. Against Klebsiellu there is good activity, and we have looked at that in the treatment study and found that it works well. I do not know the role of antibody in Candida or S. epidermidis infections, which one would be expected to contract from intravenous devices. Dr. Powell. Most of the work you have presented is focused on the sick premature newborn, particularly the very small infant. Dr. Hill. That is right. Dr. Powell. A small study was done on one of the Indian reservations with a preparation of hyperimmune serum. Adults were immunized with capsular polysaccharide, which showed efficacy in preventing Haemophilus injluenzae disease and a trend toward preventing Streptococcus pneumoniae. Do you know of other work going on? We have a very small population of sick premature infants, but we have a much larger population of normal newborns who have a nadir in their passive immunity a little bit later and are at risk for a host of common diseases. Dr. Hill. The study with the bacterial polysaccharide immunoglobulin was very dramatic. It was an intramuscular preparation and seemed to work very well. I omitted that in the interest of time. In a very high-risk population, such as the children you describe, who have an incredible incidence of pneumococcal and H. injluenzae disease, IVIG therapy may come into play in the relatively near future. Dr. Kim. Regarding the mechanisms responsible for the protective efficacy of commercially available IVIG against E. coli Kl, there are two antibodies important in E. coli Kl. One is anticapsular, and in certain serotypes of 0 antigens, antisomatic antibody is also protective against E. coli Kl. In your study you used a serotype of Kl but O-85 antigen. Am I correct or have you used different serotypes? Would you elaborate on the mechanisms for antisomatic antibody? Dr. Hill. I believe we have used more than one strain, and I do not know the 0 type of all the strains we have used. What is interesting about E. coli is that although we do not see very good opsonic activity in IVIG, we see
J. ALLERGY
CLIN. IMMUNOL. OCTOBER 1989
reasonable protective activity. It may be that with higher concentrations of antibody and high levels of complement in the serum, you can actually lyse the organism directly in some cases. Dr. Schwartz. I am intrigued by the oral study to which you referred. I presume that was not a serum preparation. Was that a colostrum preparation? Do you know anything about it? Dr. Hill. It is a serum preparation from Vienna. It had a little bit of IgG in it and almost no 1gM. I do not believe that it was a colostrum preparation. Dr. Schwartz. With regard to the findings with this preparation, would you speculate on whether there might be a direct antimicrobial effect at the enteric level or might it be immune modulation as has been implied in some of the discussions at this symposium? Second, are you aware of any controls that have been done with oral IVIG in neonates? Dr. Hill. I believe a study was done at Johns Hopkins in which infants were given oral y-globulin and examined for rotavirus levels in the stool. Treated infants showed significant shedding of rotavirus in the stool after oral yglobulin. I am unaware of any other studies in newborn infants presently, and 1 do not know the mechanism of study from Vienna. Maybe it is affecting attachment. We are very interested in IgA as a possible active antibacterial agent, We have shown some protection when you include IgA with fibronectin; it can actually opsonize organisms under these circumstances. Dr. Schwartz. We have used bovine colostrum, as had been described from the National Cancer Institute group, about 10 years ago for patients with selective IgA deficiency. We have been impressed with the reduction of clinical gastrointestinal problems that these patients have and speculated about how that might work-whether it is direct antimicrobial modulation. It does not affect serum IgA. Dr. Stiehm. Another word of caution. Jim Bussel did a study on the use of IVIG during the newborn period and gave an enormous dose of y-globulin in the first week of life. He observed efficacy for the first month of life and showed that there was less infection. However, in the second month of life, the rate of infection increased so that the overall rate of infection was exactly the same. That suggests that there was a feedback inhibition of antibody synthesis so that at the second month of life, infants were actually worse than they would have been had they not been treated. Thus I think another caution in the use of IVIG is how long it should be continued because one may get passive inhibition of antibody production thereafter. Dr. Gardner. In view of some of the case reports on treatment of enteroviral infections with IVIG and some of the data on babies with respiratory syncytial virus, do you think that there might be any role for IVIG in newborns with severe neonatal enteroviral infections or severe respiratory syncytial virus in the nosocomial situation, in addition to ribavirin? Dr. Hill. The answer is yes; there may well be a role for that, but I do not feel that the current studies are really conclusive. Studies by Fischer of respiratory syncytial virus are ongoing. They seem to have some significant benefits in a few animal models, but data on humans are not in as yet.