Author Affiliations: Department of Pediatrics, Hospital of the University of Pennsylvania and The Children's Hospital of Philadelphia, Philadelphia (Dr Schmidt); Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada (Dr Schmidt and Mr Roberts); Department of Pediatrics, University of Manitoba, Winnipeg, Canada (Dr Seshia); Children's Hospital of Michigan, Detroit (Dr Shankaran); Department of Newborn Services, Middlemore Hospital, Auckland, New Zealand (Dr Mildenhall); Department of Pediatrics, The University of Texas, Houston (Dr Tyson); Department of Newborn Care, School of Women's and Children's Health, University of New South Wales, Sydney, Australia (Dr Lui); and Department of Pediatrics, Chinese University, Hong Kong Special Administrative Region (Dr Fok).
Objective To examine whether treatment with antenatal corticosteroids modifies the immediate and long-term effects of prophylactic indomethacin sodium trihydrate in extremely low-birth-weight infants.
Design Post hoc subgroup analysis of data from the Trial of Indomethacin Prophylaxis in Preterms.
Setting Thirty-two neonatal intensive care units in Canada, the United States, Australia, New Zealand, and Hong Kong.
Participants A total of 1195 infants with birth weights of 500 to 999 g and known exposure to antenatal corticosteroids. We defined as adequate any exposure to antenatal corticosteroids that occurred at least 24 hours before delivery.
Intervention Indomethacin or placebo intravenously once daily for the first 3 days.
Outcome Measures Death or survival to 18 months with cerebral palsy, cognitive delay, severe hearing loss, or bilateral blindness; severe periventricular and intraventricular hemorrhage; patent ductus arteriosus; and surgical closure of a patent ductus arteriosus.
Results Of the 1195 infants in this analysis cohort, 670 had adequate and 525 had inadequate exposure to antenatal corticosteroids. There was little statistical evidence of heterogeneity in the effects of prophylactic indomethacin between the subgroups for any of the outcomes. The adjusted P values for interaction were as low as .15 for the outcome of death or impairment at 18 months and as high as .80 for the outcome of surgical duct closure.
Conclusion We find little evidence that the effects of prophylactic indomethacin vary in extremely low-birth-weight infants with and without adequate exposure to antenatal corticosteroids.
Trial Registration clinicaltrials.gov Identifier: NCT00009646
Although prophylactic indomethacin sodium trihydrate reduces the risks of severe periventricular and intraventricular hemorrhage (PIVH, grades 3-4), patent ductus arteriosus (PDA), and the use of surgery to close a PDA in extremely low-birth-weight (ELBW) infants, this therapy has not been shown to improve the longer-term outcome of death or disability at 18 months of age.1,2 Antenatal corticosteroids given to women with threatened preterm birth also reduce the risk of PIVH.3,4 We hypothesized that ELBW infants without adequate exposure to antenatal corticosteroids benefit more from indomethacin prophylaxis than do infants who had adequate exposure to antenatal corticosteroids. We tested this hypothesis in a post hoc subgroup analysis of the Trial of Indomethacin Prophylaxis in Preterms (TIPP) data set.
Infants with birth weights of 500 to 999 g were enrolled in the TIPP from January 29, 1996, through March 30, 1998, and followed up to a corrected age of 18 months.1 The research ethics boards of all 32 participating clinical centers (located in Canada, the United States, Australia, New Zealand, and Hong Kong) approved the trial protocol, and written informed consent was obtained from a parent or a guardian of each infant. The details of the randomization to indomethacin or placebo and the administration of the study drug doses have been previously reported.1 The primary goal of the TIPP was to determine whether prophylactic indomethacin improves survival without neurodevelopmental impairment in ELBW infants.
The TIPP case report form recorded the use of antenatal corticosteroids as 1 of the following 4 mutually exclusive regimens: (1) no antenatal corticosteroids, (2) antenatal corticosteroids less than 24 hours before delivery, (3) antenatal corticosteroids from 24 hours to 7 days before delivery, and (4) antenatal corticosteroids more than 7 days before delivery. For the present analysis, we compared subgroups of study participants with and without adequate exposure to antenatal corticosteroids, in which adequate was defined as any exposure to antenatal corticosteroids that occurred at least 24 hours before delivery. In a secondary analysis, an additional comparison was performed after the cohort was divided further into subgroups of the 4 mutually exclusive regimens of antenatal corticosteroid exposure.
To limit the possibility of type I errors, only the following outcomes were examined for heterogeneity of the indomethacin prophylaxis effect: the primary composite TIPP outcome at 18 months of death or neurodevelopmental impairment; the 5 components of this composite outcome, including death, cerebral palsy, cognitive delay, deafness, and blindness; and the secondary TIPP outcomes that showed significant treatment effects in the overall trial, including PDA, surgical closure of a PDA, and severe PIVH.1
Cerebral palsy was diagnosed if the child had nonprogressive motor impairment characterized by abnormal muscle tone and decreased range or control of movements. Cognitive delay was defined as a Mental Development Index score of less than 70 on the Bayley Scales of Infant Development II. The score was assumed to be less than 70 if the child could not undergo testing because of severe developmental delay. Audiometry was performed to determine the presence or the absence of hearing loss. A central adjudication committee that was unaware of the group assignments reviewed the results of audiologic tests for all infants with potential deafness whose hearing had not been amplified. Blindness was defined as a corrected visual acuity of less than 20/200. Follow-up was targeted for a corrected age of 18 months, but the protocol allowed a window of 18 to 21 months (12-21 months for audiometry).
Patent ductus arteriosus was a prespecified secondary outcome in the TIPP; it was diagnosed by echocardiography, which was requested only when there was a clinical suspicion of the condition. Left-to-right shunting through the PDA had to be confirmed by echocardiography with Doppler flow studies before drug or surgical therapy to close the duct was undertaken.
Cranial ultrasonography was recommended from the 5th to 8th days of life, the 21st to 28th days, and the 34th to 36th weeks of postmenstrual age if the infant was still in the study center at that time. The scan results were read locally, and copies of the written reports were sent to the coordinating center. Periventricular and intraventricular hemorrhages of grades 3 and 4 were considered severe.
The statistical significance of the observed difference in the size of the treatment effect (odds ratio) between subgroups was determined via a test of treatment × subgroup interaction in a logistic regression model. The models also included adjustment for center and birth weight stratum, as for the original TIPP analyses. Additional adjusted analyses were performed with the following prognostically important baseline variables: gestational age, sex, multiple births, and mother's educational level. A significant P value for a test of treatment × subgroup interaction would indicate that the effect of prophylactic indomethacin is different for infants with and without adequate antenatal corticosteroid exposure.
The present subgroup analysis is post hoc and was not considered when we performed our power calculations during the design phase of this trial. The sample size for the original TIPP had been preset at 600 patients per treatment group. This size of study would have yielded more than 80% power to detect a 25% proportional treatment effect (2-sided α = .05) for a control group event rate of at least 30%.
Of the 1202 TIPP participants, 525 had inadequate exposure to antenatal corticosteroids, including 231 infants with no antenatal corticosteroid exposure and 294 whose mothers received corticosteroids less than 24 hours before delivery. A total of 670 infants had adequate exposure to antenatal corticosteroids, including 500 exposed from 24 hours to 7 days before delivery and 170 whose mothers received corticosteroids more than 7 days before delivery. Data concerning the use of antenatal corticosteroids were missing for 7 infants. A total of 1136 children in this analysis cohort had complete data for the composite outcome of death or neurodevelopmental impairment at a corrected age of 18 months. The baseline characteristics before enrollment in the TIPP of these 1136 infants and their mothers in the subgroups with and without adequate use of antenatal corticosteroids are shown in Table 1.
Outcome event rates in the 4 subgroups are shown in Table 2. We found little statistical evidence of heterogeneity in the effects of prophylactic indomethacin between subgroups for any of the outcomes assessed in the unadjusted or adjusted analyses (Table 2). The observed risk of severe PIVH was twice as high in infants with inadequate antenatal corticosteroid exposure compared with adequately exposed TIPP study participants (Table 2). Similarly, we found little evidence of variable beneficial effects of prophylactic indomethacin on the outcomes of death or disability, death, severe PIVH, PDA, and PDA ligation after the study cohort was divided into further subgroups according to the presence or absence as well as the timing of antenatal corticosteroid use. The unadjusted analyses and all analyses that were adjusted for center and birth weight stratum yielded nonsignificant interaction P values (data not shown).
The international TIPP is the single largest trial of prophylactic indomethacin in very preterm infants and, to our knowledge, the only trial to date with a long-term primary outcome of death or neurodevelopmental impairment in survivors.1,2 However, almost a decade after the publication of the main results of the TIPP, neonatal practitioners remain divided into proponents and opponents of indomethacin prophylaxis for ELBW infants.5 Those who prescribe prophylactic indomethacin can claim to practice evidence-based neonatology because this therapy has been shown to reduce the rates of severe PIVH, PDA, and PDA ligation.1,2 Those who do not prescribe prophylactic indomethacin can also claim that their practice is evidence based because this therapy does not increase survival or reduce disability in the longer term1,2 and is not cost-effective.6 In addition, the long-standing conviction that early pharmacologic closure of a PDA is a desirable outcome has recently been questioned.7,8 Last, alternative strategies are available to reduce the incidence of PIVH, including the routine use of antenatal corticosteroids and better regionalization of neonatal intensive care.3,4,9
In the absence of adequate antenatal corticosteroid treatment and thus in a preterm infant at heightened risk of PIVH, one might expect an added benefit of prophylactic indomethacin. However, in the present analysis, we found no statistically significant heterogeneity for the neonatal outcomes or longer-term effects of prophylactic indomethacin by antenatal corticosteroid exposure. This is in contrast to the previously documented weak differential effect of prophylactic indomethacin by sex.10,11 When interpreting these post hoc subgroup analyses, readers should be mindful of the strengths but also of the pitfalls of subgroup analyses.12
The TIPP has been criticized for having had insufficient statistical power to detect a small but clinically important beneficial effect of indomethacin prophylaxis on outcomes at 18 months.5 The converse is equally true: the TIPP had insufficient statistical power to rule out a small but clinically important harmful effect of indomethacin prophylaxis on outcomes at 18 months in the study overall and, in particular, in the present subgroup analysis by antenatal corticosteroid exposure. Clinicians who care for ELBW infants with adequate exposure to antenatal corticosteroids may find it unsettling that the primary outcome of death or disability appeared to occur more often after prophylactic indomethacin than after placebo in the present analysis.
With these caveats, we conclude that there is little evidence that the effects of prophylactic indomethacin vary in ELBW infants with and without adequate exposure to antenatal corticosteroids.
Correspondence: Barbara Schmidt, MD, MSc, Department of Pediatrics, Hospital of the University of Pennsylvania, Ravdin 8, 3400 Spruce St, Philadelphia, PA 19104 (Barbara.Schmidt@uphs.upenn.edu).
Accepted for Publication: December 14, 2010.
Author Contributions: Dr Schmidt and Mr Roberts had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Schmidt, Seshia, Mildenhall, and Roberts. Acquisition of data: Schmidt, Shankaran, Lui, Fok, and Roberts. Analysis and interpretation of data: Schmidt, Seshia, Shankaran, Tyson, Lui, and Roberts. Drafting of the manuscript: Schmidt. Critical revision of the manuscript for important intellectual content: Seshia, Shankaran, Mildenhall, Tyson, Lui, Fok, and Roberts. Statistical analysis: Roberts. Obtained funding: Schmidt, Shankaran, and Roberts. Administrative, technical, and material support: Seshia, Tyson, Lui, Fok, and Roberts. Study supervision: Shankaran and Tyson.
Financial Disclosure: None reported.
Funding/Support: This study was supported by grant MT-13288 from the Medical Research Council of Canada. The US centers were supported in part by grants U10 HD21364, U10 HD27851, U10 HD21373, U10 HD27881, M01 RR 00997, U10 HD27880, M01 RR 00070, U10 HD21385, U10 HD27904, and U10 HD34216 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development. Indomethacin sodium trihydrate (Indocid PDA) was donated by Merck Frosst.
A. Solimano, M. Whitfield, F. Germain, and J. Tomlinson, British Columbia Children's Hospital, Vancouver; A. Peliowski, P. Etches, and B. Young, Royal Alexandra Hospital, Edmonton, Alberta; C. M. T. Robertson, Glenrose Rehabilitation Hospital, Edmonton; D. McMillan, R. Sauve, L. Bourcier, and H. Christianson, Foothills Hospital and Alberta Children's Hospital, Calgary; K. Sankaran and B. Andreychuk, Royal University Hospital, Saskatoon, Saskatchewan; M. Seshia, O. Casiro, V. Debooy, and V. Cook, Health Sciences Centre, Winnipeg, Ontario; C. M. G. Cronin, D. Moddemann, and N. Granke, St Boniface Hospital, Winnipeg; C. Nwaesei and L. St Aubin, The Salvation Army Grace Hospital, Windsor, Ontario; D. Reid, D. Lee, C. Kenyon, L. Whitty, and J. Farrell, St Joseph's Health Centre, London, Ontario; B. Schmidt, S. Saigal, P. Gillie, J. Dix, and B. Zhang, Hamilton Health Sciences Corporation, Hamilton, Ontario; A. Ohlsson, E. Asztalos, and L. Wiley, Women's College Hospital, Toronto, Ontario; A. James, The Hospital for Sick Children, Toronto; K. F. W. Young Tai and M. Clarke, Kingston General Hospital, Kingston, Ontario; M. Vincer and S. Stone, IWK Grace Health Centre, Halifax, Nova Scotia.
R. Kohan, N. French, and H. Benninger, King Edward Memorial Hospital for Women, Perth; C. Barnett, R. Haslam, and J. Ramsay, Women's and Children's Hospital, Adelaide; P. Davis, L. Doyle, B. Faber, and K. Callanan, Royal Women's Hospital, Melbourne; S. Fraser, Mercy Hospital for Women, Melbourne; K. Lui, M. Rochefort, and E. McAvoy, Westmead Hospital, Westmead; P. Colditz, and M. Pritchard, Royal Women's Hospital, Brisbane; P. Steer, D. I. Tudehope, V. Flenady, and J. Hegarty, Mater Mothers' Hospital, Brisbane.
New Zealand Centers
L. Mildenhall, W. Smith, and L. McCarthy, National Women's Hospital and Middlemore Hospital, Auckland.
Hong Kong Center
T. Fok, Prince of Wales Hospital, Shatin.
US Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network Centers
D. K. Stevenson, B. Fleisher, and B. Ball, Stanford University Medical Center, Palo Alto, California; L. A. Papile, G. Laadt, and C. Backstrom, University of New Mexico School of Medicine, Albuquerque; J. Tyson, S. Broyles, and S. Madison, The University of Texas Southwestern Medical Center, Dallas; W. A. Carlo, K. Nelson, M. Collins, and S. Johnson, The University of Alabama at Birmingham; S. Shankaran, V. Delaney-Black, G. Muran, and D. Driscoll, Children's Hospital Michigan, Detroit; B. J. Stoll, N. Simon, and E. Hale, Emory University, Atlanta, Georgia; A. A. Fanaroff, D. Wilson, M. Hack, N. Newman, Case Western Reserve University, Cleveland, Ohio; C. R. Bauer, A. M. Worth, and W. Griffin, University of Miami, Miami, Florida; W. Oh, B. R. Vohr, and A. Hensman, Brown University, Providence, Rhode Island.
Steering Committee: B. Schmidt (chair), P. Davis, D. Moddemann, A. Ohlsson, R. Roberts, S. Saigal, A. Solimano, M. Vincer, and L. Wright. External Safety Monitoring Committee: M. Gent, W. Fraser, and M. Perlman. Bayley Scales of Infant Development II Certification: R. Adkins. Audiology Central Adjudication Committee: L. Elden, C. M. T. Robertson, and B. R. Vohr. Consultant Pharmacist: S. Gray. Coordinating and Methods Center: R. Roberts and K. Thorpe (biostatisticians); N. LaPierre (trial coordinator).
Schmidt B, Seshia M, Shankaran S, Mildenhall L, Tyson J, Lui K, Fok T, Roberts R, Trial of Indomethacin Prophylaxis in Preterms Investigators FT. Effects of Prophylactic Indomethacin in Extremely Low-Birth-Weight Infants With and Without Adequate Exposure to Antenatal Corticosteroids. Arch Pediatr Adolesc Med. 2011;165(7):642-646. doi:10.1001/archpediatrics.2011.95