Pediatric Pulmonology 34:262–266 (2002)

Readmission With Respiratory Syncytial Virus (RSV)Infection Among Graduates From a Neonatal

Intensive Care Unit

Jonathan McCormick, MRCPCH* and Richard Tubman, MD, FRCP, FRCPCH

Summary. We evaluated the incidence of readmission with respiratory syncytial virus (RSV)infection among the graduates of a regional Neonatal Intensive Care Unit (NICU), and charac-

terized those who were rehospitalized. These data were used as a predictive tool to estimate the

number of babies likely to suffer readmission with RSV for the year 2000 cohort. Using the pub-

lished efficacies of palivizumab, the costs and benefits of protecting this cohort were assessed.Retrospective analysis of 2,507NICU inpatient records fromJanuary 1, 1994–December 31, 1999

from theRoyalMaternityHospital, Belfast, were comparedwith data on positiveRSVsamples from

1,790 patients between January 1, 1995–December 31, 1999 from the Northern Ireland Regional

Virus Laboratory.The analysis yielded 136 (7.6%) ex-NICU patients among the positive RSV samples over this

5-year period. Characteristic seasonal peaks of RSV infection with interseasonal variability were

observed. Of those readmitted, 86.9% were hospitalized with RSV before their first birthday.A calculated readmission rate of 5.4% for all NICU graduates, and 6.4% for those�35 weeks, was

found, leading to an expectation of 36 readmissions from the 668NICU graduates in the year 2000

over the next 1–2 years, 20 of whom would be�35 weeks and 12 would be�32 weeks. A cost of

£1.3 million would be required to protect the �35-week year 2000 cohort and prevent 11 re-admissions. This equals £120,000 per admission prevented, or 28.2 patients treated to prevent 1

readmission. A readmission rate of 6.4%maydiffer fromother studies, as it represents analysis of a

greater number of RSV seasons. Using economic arguments alone, the cost of routine administra-

tion of Palivizumab to ex-NICU�35-week infants is prohibitive. A selective practice of immunizingthose with chronic lung disease with a background of extreme prematurity over the November to

March RSV season may be more cost-effective. Pediatr Pulmonol. 2002; 34:262–266.� 2002 Wiley-Liss, Inc.

Key words: respiratory syncytial virus; Palivizumab; NICU; bronchiolitis; epidemiology;medical economics; cost-effectiveness.

INTRODUCTION

Respiratory syncytial virus (RSV) targets young infants,especially those left vulnerable to infection due to pre-maturity, chronic lung disease, or congenital heart disease.Lower respiratory tract infection (LRTI) with RSVaccounts for considerable morbidity each winter seasonin the Northern Hemisphere, and is a major factor in winterpediatric ‘‘bed crises.’’ RSV infection can lead to a pedi-atric intensive care unit (PICU) admission, a requirementfor mechanical ventilation, and death in some cases. Itseffects can be long-lasting, with respiratory symptomssuch as wheezing persisting well into childhood.1,2

Graduates from the neonatal intensive care setting areoften admitted with RSV-LRTI due to their predispositionto infection from factors such as chronic lung disease,immunologic immaturity, bronchial hyperresponsiveness,and major congenital abnormalities, including cardiacdefects.

The respiratory syncytial virus has two subtypes, A andB, and the two show variable dominance with each season.

Palivizumab, the humanized RSV monoclonal antibody,is directed against 1 of the 3 major surface glycoproteinsthat are responsible for cell fusion and penetration. Theantibody protects against infection from both subtypes,and was launched in the UK in 1999. Prophylaxis withPalivizumab has been shown to reduce the absolute risk ofhospital readmission in babies�35 weeks old from 10.6%to 4.8%.3

This audit aimed to characterize the graduates of ourNeonatalIntensiveCareUnit(NICU)at theRoyalMaternity

Neonatal Intensive Care Unit, Royal Maternity Hospital, Belfast, Northern

Ireland, UK.

*Correspondence to: J. McCormick, MRCPCH, Neonatal Intensive Care

Unit, Royal Maternity Hospital, Grosvenor Road, Belfast, Northern Ireland

BT12 6BA, UK. E-mail: jdmccormick@btinternet.com

Received 28 September 2001; Accepted 11 July 2002.

DOI 10.1002/ppul.10169

Published online in Wiley InterScience (www.interscience.wiley.com).

� 2002 Wiley-Liss, Inc.

Hospital, Belfast, who were subsequently readmitted withRSV infection. The data can be used to estimate thenumber of inpatients in the year 2000 likely to be reho-spitalized in the future with RSV. Using the currentlyavailable data on the reduction in hospital admission withPalivizumab, the likely impact of this therapy on thispopulation was estimated.

METHODS

Study Design

A retrospective analysis of NICU patients from January1, 1994 to December 31, 1999 from the Royal Mater-nity Hospital, Belfast was carried out. This hospital is theonly tertiary referral center for neonatal intensive care inNorthern Ireland.

Data Sources

NICU admissions from January 1, 1994–December 31,1999, Royal Hospital computer records, and data on allRSV-positive patients from January 1, 1995–December31, 1999 from the Northern and Eastern Health and SocialServices Boards covered by the Regional Virus Labo-ratory, Belfast, were used. Data on actual readmission ofRSV graduates during the 2000–2001 season were col-lected similarly. The laboratory processes samples forRSV from the laboratories of four hospitals caring forinfants with RSV bronchiolitis: the regional Children’sHospital (Royal Belfast Hospital for Sick Children), theUlster Hospital, near Belfast, and the Northern BoardHospitals (Antrim Area Hospital and Coleraine). Samplesare collected on hospitalized infants during winter epi-demics and on high-risk cases with bronchiolitis symp-toms at other times of the year. Infants admitted to fiveother hospitals that look after children, but do not usethe regional virus laboratory, could not be accuratelyquantified for this study, nor could infants who were nottested for RSV infection. RSV infection was diagnosedusing antigen detection by immunofluorescence (Dako)and, since 1996, the laboratory has also checked for in-fluenza A and B, parainfluenza, and adenovirus byimmunofluorescence.

Names and dates of birth of RSV-positive patients werematched with the admission data stored at the RoyalMaternity Hospital and cross-checked on the hospitalcomputer system to ensure that no patients were missed.

Analysis

These datawere analyzed in threegroups:1)>35 weeks;2) �35 weeks; and 3) �32 weeks. All infants in the�32-week group also appear in the �35-week group. Thefollowing variables were sought in all cases: gender;gestational age; birthweight; live siblings; multiple births;maternal smoking; age at first diagnosis of RSV; date ofRSV diagnosis in hospital; and other risk factors (e.g.,major congenital abnormality or congenital heart disease).The cost of immunizing one patient with Palivizumab is£2,544–4,235 (US$3,688–6,140; Euro 3,969–6,607)over the winter season. In the IMpact-RSV Trial,3 prophy-laxis resulted in a 55% relative reduction in the rate ofreadmission in the �35-week age group, and a 47% rela-tive reduction in the rate of readmission in the �32-weekgroup.3 Statistical significance between groups was cal-culated using the w2 test.

RESULTS

During the 5-year period January 1, 1995–December31, 1999, the Regional Virus Laboratory recorded 2,113RSV-positive samples from 1,790 pediatric patients. Ofthese patients, 136 (7.6%) had been NICU inpatients in theRoyal Maternity Hospital. Patients were tracked fromadmissions from January 1, 1994 until December 31, 1999.Over this period, the NICU dealt with a total of 2,507admissions (Table 1), and the RSV seasons showed cha-racteristic winter peaks of admissions for these patients(Fig. 1). The data for the three sets of RSV-positivepatients are shown in Table 2. From a potential 2,507patients, 136 babies were hospitalized with RSV bronch-iolitis in the period January 1, 1995–December 31, 1999,giving a calculated readmission rate of 5.4%, with 119 ofthese patients (86.9%) being hospitalized with RSV beforetheir first birthday (Fig. 2). There were 80 readmissionsin the �35-week group, from a potential 1,249 patients,giving a calculated readmission rate in this age group of6.4%; similarly, there were 56 readmissions from the 762�32-week group, giving a rate of 7.3%. There was nosignificant difference between median ages at first diag-nosis of RSV between groups. Significantly more womensmoked in the >35-week group compared to the othergroups. Three-quarters of all infants came from a house-hold that contained at least one other sibling. The numberof infants with major congenital abnormalities or con-genital heart disease was significantly greater in the>35-week group compared to the more premature groups.Multiple births were significantly more common in thetwo premature groups compared to the >35-week group.

Using the figure of 5.4% for all admissions to projectthe expected number of readmissions from the year 2000NICU graduates, it can be calculated that 36 babies mightbe readmitted over the next 1–2 years. Of these, 20 patients(6.4% of 310 babies) could be expected to be born at

ABBREVIATIONS

RSV Respiratory syncytial virus

LRTI Lower respiratory tract infection

PICU Pediatric intensive care unit

NICU Neonatal intensive care unit

Respiratory Syncytial Virus Readmission to NICU 263

TABLE 1— Neonatal Intensive Care Unit Admissions (1994–2000)

Year Admissions (N)

>35 weeks �35 weeks �32 weeks

N (%) N (%) N (%)

1994 412 221 (53.6) 191 (46.4) 116 (28.2)1995 415 228 (54.9) 187 (45.1) 127 (30.6)1996 441 207 (46.9) 234 (53.1) 124 (28.1)1997 357 176 (49.3) 181 (50.7) 123 (34.5)1998 437 228 (52.2) 209 (47.8) 134 (30.7)1999 445 198 (44.5) 247 (55.5) 138 (31.0)Total 2,507 1,258 (50.2) 1,249 (49.8) 762 (30.0)2000 668 358 (53.6) 310 (46.4) 164 (24.6)

Fig. 1. Seasonal variation in RSV-positive NICU graduates from 1995–1999.

264 McCormick and Tubman

�35 weeks, and 12 born at �32 weeks of gestation (7.3%of 164 babies).

NICU admission rates for the year 2000 are shown inTable 1. To protect the 310 babies in the �35-week cohortfor 2000 would cost a maximum of £1.3 million. Of the20 patients in the �35-week group who could expect to bereadmitted, an admission rate in an immunized populationof 2.88% (after a 55% reduction from the 6.4% read-mission rate) might prevent 11 readmissions. This is cal-culated as a cost per readmission prevented of £120,000(i.e., spending £1.3 million to prevent 11 readmissions).Similarly, 28.2 patients would need to be treated to preventone readmission (i.e., treating 310 patients to prevent 11readmissions).

There were 18 actual readmissions over the 2000–2001season; 11 were �35 weeks, and 4 were �32 weeks.Median age (range) at admission with RSV was 4 (2–15)months. As the tertiary referral center for NorthernIreland, many babies who are potential candidates forPalivizumab are transferred back to their local hospitalsprior to discharge. Therefore, the precise number of these

NICU graduates who subsequently received Palivizumabin the 2000–2001 season could not be quantified.

DISCUSSION

This study characterized the population of babiesadmitted to our NICU who subsequently were readmittedwith confirmed RSVinfection. A readmission rate of 6.4%for all babies �35 weeks was noted and this is consi-derably smaller than the 10.6% quoted in the IMpactTrial,3 but larger than in other studies.4–6 This figureof 6.4% represents analysis of a greater number of RSVseasons than any of the studies above, and it may alsodiffer as it analyzed data from Northern Ireland, a partof the UK that was not included in the 8% of contributorydata to the IMpact Trial.3 Figure 1 shows the familiarseasonal peaks for RSV infection as well as the inter-seasonal variability.

The epidemiological data demonstrate that 86.9% ofpatients with proven RSV infection were infected duringthe first 12 months of life. The three data sets of RSV-positive patients raise a number of interesting pointsconcerning risk factors. Firstly, three-quarters of infectedchildren come from households with at least one othersibling, an important source of infection. The averagehousehold size in Northern Ireland is 2.75 individuals,higher than in the rest of the UK (2.36).7 Secondly,although the female smoking prevalence is higher thanmale smoking prevalence, at 29% in Northern Ireland,8

reported maternal smoking habits at time of hospitaladmission frequently are underestimates of true smokingbehavior. However, significantly more mothers of babiesin the >35-week age groups smoked, compared to the twopremature age groups. Smokers may conceal a smokinghabit at time of hospitalization. Also, this information wasvery poorly recorded, with 18% of admission sheets notreporting whether an expectant mother smoked or not.Also, this does not account for sources of environmentalsmoke exposure, such as from fathers and grandparents.

TABLE 2— Respiratory Syncytial Virus-Positive Admissions by Gestational Age

>35 weeks �35 weeks �32 weeks

N 56 80 56Male:female ratio 28:28 39:41 26:30Median gestational age (weeks) (range) 38 (35–42) 30 (24–35) 29 (24–32)Median birth weight (g) (range) 3,025 (1,480–4,450) 1,378 (496–3,110) 1,121 (496–2,280)Known maternal smokers 26%* 10% 7%Households with >1 sibling 79% 74% 77%Multiple births 4% 22%** 25%**Median age at first RSV diagnosis

(months) (range)4 (1–30) 5 (1–51) 5.5 (1–51)

Major congenital abnormalities 18%** 3% 0%Congenital heart disease (%) 14%** 3% 0%

*P< 0.05.**P< 0.01.

Fig. 2. Age at first diagnosis of RSV infection.

Respiratory Syncytial Virus Readmission to NICU 265

The high numbers of children >35 weeks with majorcongenital abnormalities or congenital heart disease whowere later readmitted with RSV infection confirmed thatthese children are especially at risk, and they should becandidates for prophylaxis.

A predicted 20 babies out of 310 from the year 2000�35-week cohort might be expected to be readmitted inthe following 1–2 years. Extrapolating from the IMpactRSV Trial,3 immunizing the entire at-risk group mightgenerate a relative reduction in readmission of 55%, pro-tecting 11 patients and families from a hospital admission.The estimated cost of immunizing these babies is £1.3million, extrapolating costs from another UK hospital.This is equivalent of £120,000 per admission prevented.In the United States, calculated costs have been reportedas $39,000–420,000 (£26,000–280,000) per hospitali-zation averted.5 As for numbers needed to be treated toprevent one admission, this study finds 28.2 patients;others have found 17.2 patients.9 Ultimately, 18 NICUgraduates were actually readmitted during the 2000–2001season: 11 were�35 weeks, and 4 were�32 weeks. Thesenumbers were smaller than expected and may not accountfor infants admitted to other units not sending samplesto the regional laboratory and the increasing use ofPalivizumab among neonatal units. During this season,the hospital’s Palivizumab administration criteria includ-ed babies born at �32 weeks going home on oxygen or onoxygen within 2 weeks of going home. Babies with othersignificant respiratory problems, such as repaired congeni-tal diaphragmatic hernia, were also considered candidatesfor Palivizumab.

There are now several published studies stating that thisdrug should only be used to protect the most prematureinfants with chronic lung disease. A recent editorial dis-cussed how cost-effectiveness decides clinical practiceabove evidence-base.10 Healthcare economics are unableto put a price on the impact that a serious readmission withRSV-LRTI can have on quality of life, future respiratoryhealth, and parental anxiety. A logical way forward is toimmunize only those most at risk, i.e., those with extremeprematurity, chronic lung disease, or major congenitalcardiac or other abnormalities, such as diaphragmatichernias.

This study has shown that over the 5-year period1995–1999, 6.4% of �35-week NICU graduates were

readmitted with proven RSV infection. Using this as astandard, it is predicted that 36 infants of the NICUyear 2000 cohort might be admitted with RSV infec-tion, and 20 of them would be ex-�35-week gestationinfants. On current economic arguments, the cost ofroutine administration of Palivizumab to most infantsis prohibitive. Current selective practice of immunizingthose with chronic lung disease and a background ofextreme prematurity (generally <28 weeks) over theNovember to March RSV season appears to be the bestway forward.

ACKNOWLEDGMENTS

We are grateful for the assistance of Dr. ConallMcCaughey in the Regional Virus Laboratory, and RuthHutchinson in the Royal Maternity Hospital.

REFERENCES

1. Sigurs N, Bjarnason R, Sigurbergsson F, Kjellman B. Respiratorysyncytial virus bronchiolitis in infancy is an important risk factorfor asthma and allergy at age 7. Am J Respir Crit Care Med2000;161:1501–1507.

2. Noble V, Murray M, Webb MSC, Alexander J, Swarbrick AS,Milner AD. Respiratory status and allergy nine to 10 years afteracute bronchiolitis. Arch Dis Child 1997;76:315–319.

3. IMpact-RSV Study Group. Palivizumab, a humanized respiratorysyncytial virus monoclonal antibody, reduces hospitalisation fromrespiratory syncytial virus infection in high-risk infants. Pedia-trics 1998;102:531–537.

4. Clark SJ, Beresford MW, Subhedar NV, Shaw NJ. Respiratorysyncytial virus infection in high risk infants and the potentialimpact of prophylaxis in a United Kingdom cohort. Arch DisChild 2000;83:313–316.

5. Joffe S, Ray GT, Escobar GJ, Black SB, Lieu TA. Cost-effectiveness of respiratory syncytial virus prophylaxis amongpreterm infants. Pediatrics 1999;104:419–427.

6. Sorrentino M, Powers T. Effectiveness of palivizumab: evaluationof outcomes from the 1998 to 1999 respiratory syncytial virusseason. The Palivizumab Outcomes Study Group. Pediatr InfectDis J 2000;19:1068–1071.

7. Northern Ireland Statistics and Research Agency. 1998. NorthernIreland annual abstract of statistics. National statistics publication.p 18.

8. Northern Ireland Statistics and Research Agency. 2000. NorthernIreland annual abstract of statistics. National statistics publication.p 51.

9. Nicholl RM. RSV prevention. Arch Dis Child 2000;82:88–90.10. Sharland M, Bedford-Russell A. Preventing respiratory syncytial

virus bronchiolitis. Br Med J [Clin Res] 2001;322:62–63.

266 McCormick and Tubman