Original Articles

Indian Pediatrics 1999; 36: 1113-1118

Multicenter randomized placebo controlled trial of therapy with intravenous immunoglobulin in decreasing mortality due to neonatal sepsis 

Arvind Shenoi, N. Karthik Nagesh, P.P. Maiya*, Swarna Rekha Bhat** and S.D. Subba Rao**

From the Departments of Neonatology and Pediatrics, Manipal Hospital, M.S. Ramaiah Medical College Hospital* and St. John's Medical College Hospital**, Bangalore, India.
Reprint requests: Dr. Arvind Shenoi, Consultant Neonatologist, Department of Pediatrics, Manipal Hospital, Airport Road, Bangalore 560 017, India. E-mail: perinatal@vsnl.com
Manuscript received: December 7, 1998; Initial review completed: January 12, 1999; Revision accepted: May 4, 1999. 

Objective: To determine whether therapy with intravenous immunoglobulin G (IVIG) would decrease mortality in neonatal sepsis. Setting: Three tertiary care neonatal intensive care units in the city of Bangalore. Methods: All neonates admitted to the Neonatal Intensive Care Units with the clinical diagnosis of sepsis and having atleast C-reactive protein and one other rapid diagnostic criteria positive were enrolled. Neonates with a birth weight of less than 1000 g and those with any major congenital malformation were excluded. The neonates were randomized to receive 1 g/kg of IVIG on three consecutive days or an equivalent amount of placebo. The rest of the treatment including antibiotics and supportive care was as per the treating physician's decision. The main outcome variable was survival. Results: The trial was carried out over a period of 8 months and recruited 58 neonates. Seven neonates who qualified but did not receive either IVIG or placebo were taken into a separate control group, and one baby who received only one dose of IVIG was excluded from the analysis. Twenty-five neonates were enrolled into the IVIG arm and 25 in the placebo arm. The neonates in the therapy and placebo groups were comparable in terms of birth weight (2144±675 g vs. 2072±682 g), gestation (37.0±3.56 vs. 35.8±3.52 weeks), sex distribution, duration of stay, and number requiring ventilation. The placebo group had a significantly higher number of babies with positive blood culture. Seven babies in each group died (p >0.05). There was no significant benefit in using IVIG (OR 1.0; 95% CI 0.25-4.07) (p=0.74). Conclusion: In the sample studied therapy with IVIG did not reduce mortality in neonatal sepsis.

Key words: Intravenous immunoglobulin, Newborn, Sepsis.

Neonates are considered immuno-compromised hosts in view of their  relatively immature immune defence mechanisms(1). Specifically, they have quantitative as well as qualitative deficiency in their humoral immunity. The preterm neonate is at further risk as transplacental transfer of antibodies starts after 32 weeks gestation. It is thus not surprising that one of the proposed methods of treating neonatal sepsis is intravenous immunoglobulin (IVIG).

One of the most common cause of neonatal deaths in our country is neonatal sepsis. Inspite of antibiotic therapy and supportive care the mortality rate remains close to 30-50%(2). Immuno-modulatory therapy has been used as an adjunct in an effort to decrease mortality from neonatal sepsis. Therapy with IVIG is an important immuno-modulatory step which may help decrease the mortality. Various trials have been published(3-5) using IVIG as therapy in neonatal sepsis. These trials have been difficult to compare in view of the different doses, frequency, source of IVIG and infecting organisms. They all have been conducted in centers from advanced countries where the organisms causing neonatal sepsis are different from our setting. Indian units including in Bangalore report a predominance of Gram negative organisms like Klebsiella, Pseudo-monas, etc. and Gram positive bacteria like Staphylococcus aureus. The centers abroad have reported sepsis with Group B Streptococcus and coagulase negative Staphylo-coccus(2). This study aimed to evaluate the role of intravenous immunoglobulins along with antibiotics and supportive therapy in reduc- ing mortality in neonatal sepsis in our circumstances.

Subjects and Methods

Recruitment Criteria

All neonates admitted to the neonatal intensive care units of Manipal Hospital (MHB), M.S. Ramaiah Medical College Hospital (MSR) and St. John's Medical College Hospital (SJMC) and meeting the diagnostic criteria for probable sepsis as defined below, after informed consent were randomized into the trial. A neonate would enter into the trial as soon as informed consent and results of the rapid diagnostic tests were available. Neonates with major congenital malformations and birth-weights of less than 1000 g were excluded from the trial. A neonate eligible to enter the trial but whose family could not afford to pay even subsidised cost of IVIG were enrolled into a separate control group.

Case Definitions

Probable sepsis was defined as a neonate suspected clinically to have sepsis, with C-reactive protein greater than 6 mg/dl and at least one of the following rapid diagnostic tests positive-absolute neutrophil count, thrombocy-topenia, toxic granules in the peripheral smear, or a band count of greater than 500 per cu mm. The blood culture was negative.

Confirmed sepsis was defined as a baby who was clinically suspected to have sepsis, had C-reactive protein and one of the above rapid diagnostic tests positive and had a blood, urine, or cerebrospinal fluid culture yielding an organism.

Sample Size estimates

The study was designed to be a randomized control trial with two limbs. Neonates on qualifying for the trial would be randomized to receive either IVIG or placebo. In order to calculate the sample size we assumed an a = 0.05 and b = 0.20 and that the IVIG would produce a 50% reduction in mortality, with an incidence of 50% blood culture positivity. We would have needed to enroll at least 46 in each limb of the trial.

Outcome variables

The primary outcome variable was survival. The secondary variables like need for ventilation, exchange transfusion, and use of granulocyte monocyte colony stimulating factor (GM-CSF) were also recorded.

Baseline Data

The details of birth weight, gestation, sex, and rapid diagnostic tests were recorded in a pre-determined format. The severity of illness as judged by the score for Neonatal Acute Physiology (SNAP)(6) was recorded in two of the centers MHB and MSR.

The baby was followed up till death or discharge and the duration of hospital stay was noted along with outcome.

Randomization and Drug Administration

Random number allocation was done at co-ordinating center (MHB) and the sealed, numbered envelopes with the allocations were sent to three centers. On entry of a neonate into the trial a sealed, numbered envelope was opened and therapy/placebo instituted. A neonate on entry into the trial was randomized to receive 1g/Kg of intravenous immuno-globulin (IVIG-Sandoglobin) or an equivalent amount of 0.15% saline in 10% dextrose placebo, repeated on three consecutive days.

The use of antibiotics and day to day management was as per the unit protocol and treating physician's judgement. Laboratory samples for the rapid diagnostic tests and microbial cultures were collected and processed as per standard laboratory protocol.

Statistical Analysis

The data was tabulated and analyzed using Minitab statistical software; the Chi-square test for comparing frequencies of dichotomous variables and ANOVA for comparison of variables across centers.

Ethical and Safety Considerations

The study had been cleared by the institutional ethical committees and two inde-pendent referees. Safety of the IVIG preparation was ascertained from the manufacturer and the guidelines of the Drug regulatory authority.

Costs and Funding

The study was partially funded by Sandoz (India) Ltd. who provided the IVIG (Sandoglobin) at a subsidised cost. Other costs were borne by the participating departments.

Results

The trial commenced after discussions between the three collaborating centers in October 1995. The trial was terminated in May  1996 as the company withdrew support. Fifty-eight neonates were enrolled in the study-13 from MHB, 18 from MSR and 27 from SJMC. Seven neonates from SJMC who qualified but did not receive either IVIG or placebo were taken into a separate control group, and one baby who received only one dose of IVIG was excluded from the analysis. All the neonates had early onset sepsis.

The mean birth weight of babies in the three centers were as follows: MHB-2172±229 g, MSR-2186±167g, SJMC-1953±121 g and control-1807±227 g. The mean gestations were MHB-36.3±0.99 weeks. MSR 36.3±0.94 weeks, SJMC-36.3±0.75 weeks and controls-36.3±1.06 weeks. The babies enrolled at SJMC were significantly smaller (p <0.01) than those in the other centers, and included more small for gestational age neonates. Of the 50 neonates who entered the trial 40 were outborn, 38 were male and 25 small for date. The other characteristics were as outlined in Table I.

Table I^__Comparison of Neonates Enrolled in the Three Centers

Values are depicted as either mean ± SEM or number (%).
⁺ p<0.01 ⁺⁺ p <0.05 

The groups were comparable in mean duration of stay, need for ventilation, and mortality. The two groups were comparable in terms of SNAP score in the two centers where this data was collected. MSR had significantly lower positive blood cultures (p = 0.001). None of the neonates were ventilated at SJMC and this was significantly different from the other two centers (p = 0.04).

Twenty-five neonates were enrolled into the IVIG arm (MHB = 6, MSR = 12, SJMC = 7) and 25 in the placebo arm (MHB = 7, MSR = 6, SJMC=12). The neonates in the therapy and placebo groups were comparable in terms of birth weight (2144±675 g vs. 2072±682), gestation (37.0±3.56 weeks vs. 35.8±3.52 weeks), sex distribution, duration of stay, number requiring ventilation (Table II).

Table II^__Comparison of Placebo, IVIG and Control Groups

⁺OR = 1 (95% CI 0.25_4.07)
Values are depicted as either mean ± SEM or number (%).

In the two centers where SNAP score was recorded the two groups were comparable. The placebo group had a higher number of babies with a positive blood culture (definite sepsis) but this did not reach statistical significance (p = 0.14). The organisms isolated in both groups were similar with Enterobacter, Klebsiella pneumoniae and Staphylococcus aureus being the three commonest organisms in both the groups. There was no case of meningitis or urinary tract infection in either of the two groups. No adverse effect was reported to the IVIG or placebo infusions in any of the three centers.

Seven babies in each group died (p=0.74, OR=1, 95% CI 0.25_4.07). These was no significant difference in the survival in either group at the 3 centers_MHB (p=0.46), MSR (p=0.37), SJMC (p=0.18). The mortality in the probable sepsis group (23.3%) was slightly lower than in the definite sepsis group (40.0%) but was not statistically significant and therapy with IVIG did not change this significantly (OR 1.0; 95% CI 0.25-4.07) (p=0.74).

Discussion

This study was initiated because the role of intravenous immunoglobulin in neonatal sepsis was unclear. The only study of relevance to our NICU, available at the time was the trial from Saudi Arabia(3) which randomised 60 neonates with suspected and proven sepsis into a group treated by antibiotics alone or antibiotics and a brand of immunoglobulin which claims to be IgM enriched. This study showed a significantly lower death rate in IVIG group (5%) as compared to the placebo group. The study was subsequently criticised(1) on the basis that this difference would not be statistically significant and the IgM enriched IVIG did not contain physiologically significant amount of IgM. One trial has been published from our country susbequently using this brand of IgM enriched immunoglobulin in neonatal sepsis which suggested that there is no significant advantage from this preparation(7). Another article has discussed the issue in depth and suggested the need for trials from our country(8).

The next consideration was which dosage schedule to follow. We chose a dosage regimen of 1g/kg for three doses after reviews of evidence(9,10) suggesting that multiple doses seem to have a better effect than a single intravenous dose. Ideally one would need to  have monitored the level of the antibody but in the absence of such facilities we gave three sequential doses of 1g/Kg/day.

The need for placebo was essential for study purposes and we administered 0.15% saline in 10% dextrose as this is the common intra-venous fluid in the newborn. The alternatives of using albumin (high cost), or any other fluid (electrolyte imbalance) were also discussed but discarded. Unfortunately our request to the company to provide the placebo in identical vials was not acceded. Hence blinding of physicians to therapy could not be done. This might have introduced an element of bias in the study, but since the outcome was mortality such bias, if any, was negligible.

One of the main strengths of our study is that the trial involved relatively mature low birth weight infants who form the bulk of neonatal admissions in our country. The earlier trials(5,9,10) have concentrated on smaller infants in the 550-1500 g weight group. The neonates enrolled in each of the centers were also similar except that SJMC had significantly more small for gestational age neonates. The other paremeters were comparable. Only a few neonates received adjunctive therapy in the form of exchange transfusions, GM-CSF, and dopamine and it is unlikely that these interventions influenced the outcome.

The IVIG and the placebo groups were also comparable in most respects except for the higher number of neonates with definite sepsis in the placebo group. This difference was not statistically significant and did little to tilt the result in favour of IVIG. In the two centers where this data was collected the two groups were also comparable in terms of their illness as measured by the SNAP at admission. This parameter has not been measured in the studies published earlier. In our study the mean SNAP of the study and placebo groups in two centers were comparable suggesting that babies with relatively equal degree of illness were entered into limbs of the trial. This further strengthens the belief that the role of IVIG is likely to be insignificant.

The organisms isolated in the IVIG and placebo groups reflect the predominant organisms causing neonatal sepsis in our NICUs. The culture positivity rate of 40% and the mortality rate in this study of 28% are in keeping with the data from the units in the previous years and the published data from other units(2).

A major criticism of our study would be the necessity to terminate the trial prematurely. Had we recruited the requisite numbers we could have conclusively defined the value of IVIG. However, the available data suggest that the effect of IVIG in neonatal sepsis could not have been dramatic. It is evident that even if the entire predetermined sample size were to be recruited the results would not have changed, since the differences between the groups were so small. We would thus have required a much larger sample size for providing enough power.

The cost of intravenous immunoglobulin is around Rs. 1000/- per gram or more. Our study suggests that this expense may perhaps be spared when it comes to treatment of neonatal sepsis. This is in view of the fact that not only is the outcome unaltered but also the mean duration of stay was similar in the IVIG and placebo groups. The duration of stay of control group was significantly shorter than the other two groups but the reasons for this may have been extraneous. We conclude that there is no role of intravenous immunoglobulin in the treatment of neonatal sepsis, though a larger study sample would have made it possible to affirm this more confidently.

Acknowledgements

We gratefully acknowledge the support of  Sandoz India Ltd. and the residents and staff at the three participating centers without whose help this study would not have been possible.

References

1. Weisman LE, Cruess DF, Fischer GW. Current status of intravenous immunoglobulin in preventing or treating neonatal bacterial infections. Clin Rev Allerg 1992; 10: 223-235.

2. Bhakoo ON. Neonatal bacterial infections at Chandigarh: A decade of experience. Indian J Pediatr 1980; 47: 419-424.

3. Haque KN, Zaidi MH, Bahakim H. IgM enriched intravenous immunoglobulin therapy in neo- natal sepsis. Am J Dis Child 1988; 142; 1293-1296.

4. Weisman L, Cruess D, Fischer D, Fischer G. Standard versus hyperimmune immunoglobulin for preventing or treating neonatal bacterial infections. Clin Perinatol 1993; 20: 211-224.

5. Lacy J, Ohlsson A. Administration of intravenous immunoglobulins for prophylaxis or treatment of infection in preterm infants: Meta analysis. Arch Dis Child 1995; 72: 151-155.

6. Richardson DK, Gray JE, McCormick MC, Workman K, Goldman DA. Score for neonatal acute physiology: A physiologic severity index for neonatal intensive care. Pediatrics 1993; 91: 617-623.

7. Samatha S, Jalalu MP, Hegde RK, Vishwanath D, Maiya PP. Role of IgM enriched intravenous immunoglobulin as an adjuvant to antibiotics in neonatal sepsis. Karnataka Pediatr J. 1997; 1: 1-6.

8. Irani SF, Wagle SV, Deshpande PG. Role of intravenous immunoglobulin in prevention and treatment of neonatal infection. Indian Pediatr 1991; 28: 443-449.

9. Haque KN, Remo C, Bahakim H. Comparison of two types of intravenous immunoglobulins in the treatment of neonatal sepsis Clin Exp Immunol 1995; 101: 328-333.

10. Chen JY. Intravenous immunoglobulin in the treatment of full-term and premature newborns with sepsis. J Formos Med Assoc 1996; 95: 839-844.