Secondary Logo

Journal Logo

Original Articles: Gastroenterology: Inflammatory Bowel Disease

Corticosteroid Dosing in Pediatric Acute Severe Ulcerative Colitis

A Propensity Score Analysis

Choshen, Sapir*,†; Finnamore, Helen; Auth, Marcus K.H.; Bdolah-Abram, Tali; Shteyer, Eyal*,†; Mack, David§; Hyams, Jeffrey||; Leleiko, Neal; Griffiths, Anne#; Turner, Dan*,†

Author Information
Journal of Pediatric Gastroenterology and Nutrition: July 2016 - Volume 63 - Issue 1 - p 58-64
doi: 10.1097/MPG.0000000000001079
  • Free


What Is Known

  • It is a common practice to use steroids in acute severe colitis in a dose of 1 to 1.5 mg · kg−1 · day−1 methylprednisolone daily up to 40 to 60 mg daily in 1 to 2 divided doses.
  • There are no clinical trials to support this recommendation.

What Is New

  • There does not seem to be a consistent superiority of high dose (>2 mg · kg−1 · day−1) versus standard (1.25 mg · kg−1 · day−1) or low-dose (1 mg · kg−1 · day−1) methylprednisolone in pediatric acute severe colitis.

See “Steroid Limbo in Acute Severe Ulcerative Colitis: How Low Can You Go?” by Hansen and Russell on page 2.

In comparison with adult-onset disease, ulcerative colitis (UC) occurring in childhood is much more often extensive (1,2) and, as such, more likely to be associated with acute severe exacerbations treated with intravenous corticosteroids (IVCS) as first-line therapy (3,4). Steroid-failure rate was 34% in a systematic review of five pediatric studies including 291 children with acute severe colitis (ASC) (5).

Based on common clinical practice, the recommended steroid dosing in ASC is 1 to 1.5 mg · kg−1 · day−1 methylprednisolone daily up to 40 to 60 mg daily in 1 to 2 divided doses (6–8). No dose of IVCS, however, has been found to be superior to others in ASC. Although some suggest that very high steroid doses may be more effective than standard doses (9–11), this is not evidence based.

It is important to optimize efficacy of steroids, if need for second line salvage therapies is to be reduced. Because randomized controlled dose-ranging trials of IVCS in ASC are not likely to be conducted in children, we aimed to use a robust statistical method of propensity score (PS), to compare high versus low dosing of IVCS in pediatric ASC. The PS method is a powerful statistical tool to control for confounding-by-indication bias in retrospective longitudinal cohorts, but requires large datasets. We therefore established the largest cohort to date of 283 children with ASC treated with IVCS.


This study used existing datasets of the prospective Outcome of Steroid therapy in Colitis Individuals (OSCI) study (n = 128 from 5 centers in North America) and the retrospective OSCI study (n = 99 from Toronto), both of which reported the 1-year outcome of children admitted for ASC (2,12). In addition, 54 children were added for the present study by chart review (from Jerusalem and Liverpool) using the same design and dataset structure as the OSCI studies. The Jerusalem cohort was added as a convenience sample, and Liverpool was chosen because of the local general, but not universal, tendency of prescribing higher IVCS doses.

Data acquisition was based on the same detailed protocol in all cohorts. This includes similar eligibility criteria, case report forms, timing of visits, and outcome. Explicit demographic, clinical, and laboratory data were recorded at admission, at 3 and 5 days thereafter, at introduction of second line therapy, at discharge and 1 year after admission. Disease activity was measured using both subjective physician global assessment (PGA) using 100 mm visual analogue scale, and the Pediatric UC Activity Index (PUCAI), a valid 6-item clinical measure of UC activity. The responsiveness of the PUCAI to rapid change is high, allowing capture of change in disease activity day by day (2,12–14). Children were managed according to the discretion of the physician, including corticosteroids dosing. Therefore, there was a significant dosing variability in the cohorts, an essential asset to be used in the PS matching to allow for dose-effect assessment.

Inclusion criteria were children with confirmed UC by accepted criteria (15) at the age of 2 to 18 years, admitted for IVCS treatment for ASC. In order to avoid repeated measures of the same patients, only the first admission of each child was included. Children with infectious gastroenteritis or other positive bacterial cultures were excluded. Hydrocortisone and other types of corticosteroids were standardized as methylprednisolone-equivalent using established conversion factors (16).

The cohort was divided 3 times into “high” and “low” dose groups, based on 3 different cutoffs, justified from different practices (all of the doses are calculated based on the children weight on admission):

  1. 1 mg · kg−1 · day−1 methylprednisolone to 40 mg · day−1, referred herein as the “low cutoff”;
  2. 1.25 mg · kg−1 · day−1 methylprednisolone to 50 mg · day−1, referred herein as the “standard cutoff”;
  3. 2 mg · kg−1 · day−1 methylprednisolone to 80 mg · day−1, referred herein as the “high cutoff”.

The primary outcome was percentage of children failing IVCS, based on the need for salvage therapy by hospital discharge (colectomy or second-line medical therapy with infliximab or calcineurin inhibitors). Secondary outcomes included percentage of children who had at most mild disease at day 5 (defined by PUCAI < 35 points), length of admission, and the need for salvage therapy during the subsequent year.

Statistical Analysis

A simple comparison of the steroid dosing is not possible because of confounding-by-indication bias. The latter is a term describing a bias in which the indication to allocate subjects to either groups (ie, treating with high- or low-dose steroids) is related to the outcome in ways other than the comparison of interest. In our case, physicians may prefer higher steroids dose in the most severe end of the spectrum, those who previously failed standard dosing, or those with prolonged oral steroid course before admission. These factors, among others, may affect outcome of treatment, independently of the prescribed dose. This bias may be addressed by the PS matching that allows controlling for many background covariates simultaneously by matching on a single scalar variable. The PS in our case is the probability that a child receives at admission either high- or low-IVCS dose, given the observed covariates (17–28). By matching pairs of children with similar probabilities of being treated with high or low IVCS dose, and assuming no hidden bias, we created a quasi-randomized experiment.

Another method for controlling for the confounding-by-indication bias is the “doubly robust” weighted regression (29,30). We used this model to validate our findings using the entire cohort in regression models weighted by the inverse variance of the PS and adjusted for the unbalanced baseline variables: age, prior use of thiopurines, days of bloody stools before admission, percent of weight loss during the month before admission, PUCAI, albumin, C-reactive protein (CRP), platelets, and ESR. In addition, we forced into the models 3 clinically important variables: number of steroid courses in the year before admission, weight, and PGA upon admission.

The explicit stages of matching and analysis with PS is described elsewhere (19). Briefly, a logistic regression model was constructed with high or low-dose IVCS as the dependent variable (repeated using the 3 cutoff values defined above) and the following 17 pretreatment variables as the explanatory variables: sex, age, disease duration, prior 5-aminosalycilic acid treatment, days of oral steroids before admission, days of bloody stools before admission, number of steroid courses in the year before admission, number of admissions in the year before the index admission, disease activity at baseline (PUCAI score and PGA), temperature at baseline, vomiting at baseline, and laboratory values at baseline including CRP, ESR, albumin, platelets, and hemoglobin. Next, according to the beta estimates of each variable, a PS was calculated for each child according to his/her individual parameters. This model is based only on baseline parameters and is simply a means by which to assign a score for estimating the probability that a subject receives either dose within our specific cohort. Overfitting is thus of less concern because the model is not intended to predict outcomes nor to be applied on unseen data.

Each child from the high-dose subgroup was individually matched to a low-dose child according to the nearest value of the logit of the PS, in a blinded fashion to all outcome variables (19), within a caliper size of a quarter of a standard deviation, according to Rosenbaum and Rubin (31). The goodness of the PS matching was evaluated by the degree that these and other baseline covariates were balanced between the groups.

In order to test the hypothesis that children treated with oral prednisone before the admission may require higher steroid doses because of downregulation of receptors, we performed a sensitivity analysis in which only children that were treated with prednisone upon admission were included in the PS-matched analysis.

Data are presented as means ± standard deviation or medians (interquartile range (IQR)), as appropriate. Paired data of matched children were compared using McNemar test (categorical) and paired Student t test (continuous) and the signed-rank test as appropriate for the distribution normality. Unpaired data were compared using χ2 test, student t test, or Wilcoxon rank sum test, as appropriate. Imputation was performed using the hot deck method for parameters with <10% missing data occurring at random (32). No outcome measures were missing. All of the comparisons were made using 2-sided significance levels of P < 0.05. Statistical analyses were performed using SPSS version 22.0 (IBM SPSS Statistics, Armonk, NY). The study was approved by the institutional review board of each participating center.


A total of 283 children were eligible and included in the study (Table 1), of whom 127 (45%) had at most mild disease by day 5 of admission (ie, PUCAI < 35); 89 (31%) had failed IVCS and required salvage therapy during the admission after a median of 12 (8–16) days. The overall colectomy rate by discharge was 20% (n = 57), and the median admission duration was 9 days (IQR 6–19). By 1 year after discharge, 122 children (43%) required salvage medical therapy, of whom 89 (31% of the entire cohort) required colectomy.

Table 1
Table 1:
Baseline characteristics of the included cohort

The median IVCS dose in the entire cohort was 1.0 mg · kg−1 · day−1 (IQR 0.8–1.4) and 44 mg · kg−1 · day−1 (32–60). We split our cohort as per present guidelines that dictate dose limit of 40 to 60 mg per day for children >40 kg, rather than dosing per kilogram (33). The median IVCS dose for children weighing <40 kg was 1.2 mg · kg−1 · day−1 (IQR 1–1.8) and 32 mg · day−1 (24–50) (Fig. 1a). For children weighing >40 kg, the median IVCS dose was 0.95 mg · kg−1 · day−1 (0.77–1.19) and 58.5 mg · day−1 (40–60) (Fig. 1b).

Steroid dosing of 283 children admitted for acute colitis. A, IVCS dose for children weighing 40 kg or less (18 children were treated with doses >2.5 mg · kg−1 · day−1 [2.7–31 mg · kg−1 · day−1]). B, IVCS dose for children >40 kg (14 with doses ≥100 mg · day−1 [100–1000 mg · day−1]). IVCS = intravenous corticosteroids.

Unadjusted, there was a poor but significant correlation between the dose used and total PUCAI score at day 5 when using mg · kg−1 · day−1 in those <40 kg (Spearman rho = −0.19, P = 0.03) (Fig. 2a) but not with mg · kg−1 · day−1 in those weighing >40 kg (Spearman rho = 0.15, P = 0.06) (Fig. 2b). When comparing the baseline disease severity in the high- and low-dose groups (standard cutoff), the low-dose group, however, had significantly lower CRP (1.3 mg/dL (IQR 0.6–2.5) vs 1.8 mg/dL (0.9–3.3); P = 0.047) but similar PUCAI scores (75 points (IQR 65–80) vs 70 points (60–80); P = 0.2). Moreover, 1 center using a local protocol with 2 to 20 mg · kg−1 · day−1 steroids and intravenous antibiotics obtained a lower colectomy rate of 12%. These findings support the following use of PS-based matching for elucidating conclusions.

Scatter plot of IVCS dose versus PUCAI at day 5. A, Children weighing ≤40 kg (18 children were treated with doses >2.5 mg · kg−1 · day−1 [2.7–31 mg · kg−1 · day−1]). B, Children weighing >40 kg (14 with doses ≥100 mg · day−1 [100–1000 mg · day−1]). IVCS = intravenous corticosteroids; PUCAI = Pediatric Ulcerative Colitis Activity Index.

Standard Cutoff (Below or Above a Dose of 1.25 mg · kg−1 or 50 mg · day−1)

A total of 136 (48%) children were treated with doses <1.25 mg · kg−1 50 mg · day−1 (standard dosing), and the other 147 (52%) children above either value. Matching the 2 groups according to the PS yielded 109 matched pairs (Fig. 3). Fifty-two percent of the children in the prospective OSCI cohort, 60% of the retrospective OSCI cohort, 20% of the children in Jerusalem, and 15% of the children in Liverpool were treated with low-dose IVCS. In the lower dose group, the median dose was 0.8 mg · kg−1 · day−1 (0.7–1.0) or 34 mg · day−1 (25–40), and in the higher dose group 1.5 mg · kg−1 · day−1 (1.0–2.0) or 60 mg · day−1 (52–80). No significant differences were found in all of the 25 pretreatment baseline parameters included in the PS calculation (Table 1 and other variables not presented), implying successful matching.

Distribution of the PSs of the low- and high-dose groups in the standard cutoff (ie, 1.25 mg · kg · day−1 to 50 mg · day−1), showing the matching zone. Children with similar PS were treated with different doses and were therefore good candidates for matching.

None of the following outcomes were significantly different between the low- and high-dose groups: the need for salvage therapy during admission (30% vs 31%; P = 0.88), rates of PUCAI < 35 points at day 5 (44% vs 40%; P = 0.68), median admission days (9 [IQR 15–18] vs 10 [6–21]; P = 0.27), and the need for salvage therapy during the year after admission (47% vs 38%; P = 0.22) (Fig. 4).

Outcomes of the low- and high-dose matched groups according to the standard cutoff (1.25 mg · kg · day−1 to 50 mg · day−1 methylprednisolone).

Similarly, the day-5 median laboratory tests values were not different between the low- versus high-dose groups: CRP (0.62 mg/dL (IQR 0.12–1.62) vs 0.82 (0.24–2.42); P = 0.62), ESR (32 mm/hour (12–45) vs 31(15–50); P = 0.92), hemoglobin (10.25 g/dL (8.5–11.3) vs 10.1 (8.7–11.4); P = 0.42), albumin (31 g/L (28–36) vs 31(28–35); P = 0.58), and platelets (450 × 103/μL (383–576) vs 449 (393–575); P = 0.57, respectively.

Low Cutoff (Below or Above a Dose of 1 mg · kg · day−1 or 40 mg · day−1)

A total of 51 (18%) children were treated with IVCS doses lower than the low cutoff, and the other 232 children were treated with higher doses. Matching of the 2 subgroups according to the PS was successful to yield 47 matched pairs. In the low-dose subgroup, the median dose was 0.8 mg · kg−1 · day−1 (IQR 0.6–0.9) or 25 mg · day−1 (20–32), and in the high-dose subgroup it was 1.2 mg · kg−1 · day−1 (0.9–1.9) or 48 mg · day−1 (40–60). No significant differences were found in all 25 baseline parameters (Table 1) nor in the tested outcomes, including the need for salvage therapy during admission (28% in the low-dose subgroup vs 30% in the high-dose; P = 1.0), rates of PUCAI < 35 points at day 5 (46% vs 46%; P = 1.0), median admission days (11 (IQR 5–22) vs 11 (6–21); P = 0.4), and the need for salvage therapy during the year after admission (38% vs 38%; P = 1.0).

High Cutoff (Below or Above a Dose of 2 mg · kg · day−1 or 80 mg · day−1)

A total of 44 (16%) children were treated with IVCS doses higher than the high cutoff. Matching the low and high subgroups according to the PS yielded 43 matched pairs (median dose 1.1 mg · kg−1 · day−1 (IQR 1.0–1.6) or 56 mg · day−1 (30–60) and 4.3 mg · kg−1 · day−1 (2.0–22.4) or 200 mg · day−1 (80–730), respectively). No significant differences were found in all of the 25 baseline parameters (data not shown), implying successful matching. It is, however, noteworthy that 19 children (44%) of the high-dose group were treated in 1 center as part of local practice of commencing high doses, as opposed to only 2 (5%) in the low-dose subgroup. The only significant outcome between the 2 matched groups was the rates of PUCAI < 35 points at day 5 (60% vs 35% favoring the high-dose group; P = 0.03). The other outcomes were numerically better in the high-dose group but did not reach statistical significance: need for salvage therapy during admission (46% vs 26%; P = 0.09), median admission days (12 (IQR 7–24) vs 11 (6–23); P = 0.38), and need for salvage therapy during the year after admission (56% vs 35%; P = 0.11). Given the consideration that the outcomes may have been influenced by local practice, the 19 patients with high-dose steroids from 1 center were excluded in a sensitivity analysis, yielding 21 matched pairs. The median IVCS dose was 1.1 mg · kg−1 · day−1 (IQR 0.9–1.4) or 60 mg · day−1 (IQR 50–60) in the low-dose group and 2.7 mg · kg−1 · day−1 (1.6–3.8) or 100 mg · day−1 (80–190) in the high dose. Of the 25 baseline variables, platelet count (433 ± 172 in the low dose vs 515 ± 111 in the high dose; P = 0.03) and prior 5-aminosalycilic acid treatment (71% in the low dose vs 29% in the high dose; P = 0.02) were significantly different between the matched groups. In this sensitivity analysis, all of the outcomes were numerically better in the “low” dose groups except PUCAI at day 5, but none reached statistical significance—need for salvage therapy during admission (29% in the low dose vs 48% in the high group; P = 0.39), PUCAI < 35 points at day 5 (29% vs 33%; P = 1.0), median admission duration (11 days (IQR 5–29) vs 21 (10–29.5); P = 0.38) and need for salvage therapy during the year after admission (38% vs 62%; P = 0.27). The 34 children from the center excluded for the sensitivity analysis demonstrated the need for salvage therapy by 1 year of 11% (reported in a seperate manuscript in this issue).

Entire Cohort

We used the “doubly robust” PS-weighting technique on the entire cohort, using the standard cutoff, the unbalanced baseline variables, and the 3 clinically important variables. The only significantly different outcome was the length of admission with a clinically irrelevant effect size (median of 9 days (IQR 5–16) in the low-dose group vs 10 days (6–21) in the high-dose group, P = 0.005). The need for salvage therapy during admission (OR = 1.0 (95% CI = 0.6–1.5); P = 0.92), rates of PUCAI < 35 points at day 5 (OR = 1.2 (0.8–1.8); P = 0.42), and the need for salvage therapy during the year after admission (OR = 1.3 (0.9–2.0); P = 0.18) were similar.

In the PS-matched sensitivity analysis, performed on the children that were treated with prednisone upon admission, no significant difference was found in the outcomes of the high- versus low-dose subgroups in all of the 3 aforementioned cutoffs (data not presented), supporting the notion that high dose is not required also in those who failed oral steroids.


We assembled the largest cohort to date of children with ASC to compare the effectiveness of different IVCS doses using a robust statistical method across several outcomes and subgroups. We were successful in balancing 25 possible confounders, approximating random allocation. For most outcomes and analyses, we could not find superiority of the high doses, including the need for salvage therapy during admission, mean admission duration, and the need for salvage therapy during the year following admission. This was evident for matched children treated with doses lower or higher than 1 mg · kg−1 · day−1 and 1.25 mg · kg−1 · day−1. Similarly, in a controlled analysis of the entire cohort, high-dose steroids were not associated with any of the 5 outcomes explored.

Although some outcomes were better in the higher doses when considering a cutoff of 2 mg · kg−1 · day−1 (or 80 mg · day−1), it is likely this was influenced by further differences in local practice protocols in one center. For instance, the more frequent use of antibiotics in that center that provided a large portion of the very high dose to patients. When patients from that center were excluded in a sensitivity analysis, the superiority of the high dose was eliminated, consistent with the other comparisons. Moreover, except for this cutoff, no dose-effect association was noted in the other cutoffs which included the very same patients but with a different seperation.

Unfortunately, there is a surprising paucity of literature regarding the minimal effective dosing of IVCS in ASC. A meta-regression of cohort studies that encompassed 1991 patients did not demonstrate a dose-colectomy association or advantage of doses of methylprednisolone-equivalent >60 mg · day−1(34). A prospective trial assessed 128 children (included also in this report) and found that the severity of disease on admission predicted IVCS responsiveness rather than the administered dose (2,12). Glucocorticoid bioassay did not predict response to corticosteroids in pediatric ASC (35). A randomized controlled trial of 66 adults admitted for ASC showed equivalency between continuous and bolus IVCS dosing (36). A double-blind randomized trial comparing outcomes of 35 adult patients treated with either adrenocorticotropic hormone or hydrocortisone failed to identify significant differences in response (37). Finally, an outpatient randomized trial compared 3 doses of oral prednisolone and found that 40 and 60 mg · day−1 were as effective, and superior to 20 mg · day−1(38).

The anecdotal evidence for the effectiveness of very high-dose pulse steroids of up to 1000 mg · day−1 is limited to case reports and is inconsistent (9–11) (39). A prospective study of 20 patients treated with pulse steroids showed remission rates of only 60%, similar to the rates found in local historical controls (10). A retrospective Japanese report found that pulse dosing in 20 children achieved remission faster than in 17 children treated with standard dosing; however, treatment was successful in all 37 children and none required salvage therapy (9–11). A national survey from Japan showed 55% short-term remission among 21 children treated with pulse steroid therapy; the comparative results of the other 41 children treated with the traditional dosing are not reported (11). Taken together, the present anecdotal evidence does not support superiority of pulse steroids over the known ∼70% response to standard dosing [11].

The nonsignificant differences in the outcomes of prednisone-treated children that received high- or low-dose IVCS may indicate that also these children do not require higher IVCS doses.

Data for this large cohort were collected in a variety of hospitals in North America, Europe, and Israel, including a wide spectrum of patients and health systems. The variety is an essential asset when using PS. This is, however, still a post hoc analysis study, and we cannot exclude the possibility that unknown confounding variables were missed. Given the paucity of children treated with doses <1 mg · kg−1 · day−1, we could not subdivide this group. Common wisdom, however, supports a minimal dose of 1 mg · kg−1 · day−1 up to 40 mg daily. Finally, we did not collect data on adverse events as dictated by the original OSCI datasets. Nonetheless, higher steroid doses are associated with more adverse events (40), although these depend on the duration of treatment, and are mostly reversible. Despite the inherent limitations, this study provides for the first time a comprehensive approach on a large cohort of children to explore the dose-effect relation of IVCS in pediatric ASC. The consistent results across several analyses support the use of lower IVCS doses, as dictated by the present guidelines (1–1.5 mg · kg−1 · day−1 up to 40–60 mg · day−1) (33). Additional studies are, however, needed to determine the best dose with more certainty.


1. Griffiths AM. Specificities of inflammatory bowel disease in childhood. Best Pract Res Clin Gastroenterol 2004; 18:509–523.
2. Turner D, Walsh CM, Benchimol EI, et al Severe paediatric ulcerative colitis: incidence, outcomes and optimal timing for second-line therapy. Gut 2008; 57:331–338.
3. Chakravarty BJ. Predictors and the rate of medical treatment failure in ulcerative colitis. Am J Gastroenterol 1993; 88:852–855.
4. Truelove SC, Jewell DP. Intensive intravenous regimen for severe attacks of ulcerative colitis. Lancet 1974; 1:1067–1070.
5. Turner D, Griffiths AM. Acute severe ulcerative colitis in children: a systematic review. Inflamm Bowel Dis 2011; 17:440–449.
6. Travis SPL, Stange EF, Lémann M, et al European evidence-based consensus on the management of ulcerative colitis: current management. J Crohn Colitis 2008; 2:24–62.
7. Turner D, Travis SP, Griffiths AM, et al Consensus for managing acute severe ulcerative colitis in children: a systematic review and joint statement from ECCO, ESPGHAN, and the Porto IBD Working Group of ESPGHAN. Am J Gastroenterol 2011; 106:574–588.
8. Turner D, Levine A, Escher JC, et al Management of pediatric ulcerative colitis: joint ECCO and ESPGHAN evidence-based consensus guidelines. J Pediatr Gastroenterol Nutr 2012; 55:340–361.
9. Kudo T, Nagata S, Ohtani K, et al Pulse steroids as induction therapy for children with ulcerative colitis. Pediatr Int 2011; 53:974–979.
10. Rosenberg W, Ireland A, Jewell DP. High-dose methylprednisolone in the treatment of active ulcerative colitis. J Clin Gastroenterol 1990; 12:40–41.
11. Nagata S, Shimizu T, Kudo T, et al Efficacy and safety of pulse steroid therapy in Japanese pediatric patients with ulcerative colitis: a survey of the Japanese Society for Pediatric Inflammatory Bowel Disease. Digestion 2010; 81:188–192.
12. Turner D, Mack D, Leleiko N, et al Severe pediatric ulcerative colitis: a prospective multicenter study of outcomes and predictors of response. Gastroenterology 2010; 138:2282–2291.
13. Turner D, Otley AR, Mack D, et al Development, validation, and evaluation of a pediatric ulcerative colitis activity index: a prospective multicenter study. Gastroenterology 2007; 133:423–432.
14. Turner D, Hyams J, Markowitz J, et al Appraisal of the pediatric ulcerative colitis activity index (PUCAI). Inflamm Bowel Dis 2009; 15:1218–1223.
15. Levine A, Koletzko S, Turner D, et al ESPGHAN revised porto criteria for the diagnosis of inflammatory bowel disease in children and adolescents. J Pediatr Gastroenterol Nutr 2014; 58:795–806.
16. Mager DE, Lin SX, Blum RA, et al Dose equivalency evaluation of major corticosteroids: pharmacokinetics and cell trafficking and cortisol dynamics. J Clin Pharmacol 2003; 43:1216–1227.
17. Austin PC. A critical appraisal of propensity-score matching in the medical literature between 1996 and 2003. Stat Med 2008; 27:2037–2049.
18. Brookhart MA, Schneeweiss S, Rothman KJ, et al Variable selection for propensity score models. Am J Epidemiol 2006; 163:1149–1156.
19. D’Agostino RB Jr. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998; 17:2265–2281.
20. Fitzmaurice G. Confounding: propensity score adjustment. Nutrition 2006; 22:1214–1216.
21. Hullsiek KH, Louis TA. Propensity score modeling strategies for the causal analysis of observational data. Biostatistics 2002; 3:179–193.
22. Leon AC, Hedeker D. Quantile stratification based on a misspecified propensity score in longitudinal treatment effectiveness analyses of ordinal doses. Comput Stat Data Anal 2007; 51:6114–6122.
23. Oakes JM, Church TR. Invited commentary: advancing propensity score methods in epidemiology. Am J Epidemiol 2007; 165:1119–1121.
24. Senn S, Graf E, Caputo A. Stratification for the propensity score compared with linear regression techniques to assess the effect of treatment or exposure. Stat Med 2007; 26:5529–5544.
25. Stukel TA, Fisher ES, Wennberg DE, et al Analysis of observational studies in the presence of treatment selection bias: effects of invasive cardiac management on AMI survival using propensity score and instrumental variable methods. JAMA 2007; 297:278–285.
26. Sturmer T, Joshi M, Glynn RJ, et al A review of the application of propensity score methods yielded increasing use, advantages in specific settings, but not substantially different estimates compared with conventional multivariable methods. J Clin Epidemiol 2006; 59:437–447.
27. Sturmer T, Schneeweiss S, Rothman KJ, et al Performance of propensity score calibration—a simulation study. Am J Epidemiol 2007; 165:1110–1118.
28. Yue LQ. Statistical and regulatory issues with the application of propensity score analysis to nonrandomized medical device clinical studies. J Biopharm Stat 2007; 17:1–13.
29. Bang H, Robins JM. Doubly robust estimation in missing data and causal inference models. Biometrics 2005; 61:962–973.
30. Kang JD, Schafer JL. Demystifying double robustness: a comparison of alternative strategies for estimating a population mean from incomplete data. Stat Sci 2007; 523–539.
31. Rosenbaum PR, Rubin DB. Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. Am Stat 1985; 39:33–38.
32. Streiner DL. The case of the missing data: methods of dealing with dropouts and other research vagaries. Can J Psychiatry 2002; 47:68–75.
33. Turner D, Travis SP, Griffiths AM, et al Consensus for managing acute severe ulcerative colitis in children: a systematic review and joint statement from ECCO, ESPGHAN, and the Porto IBD Working Group of ESPGHAN. Am J Gastroenterol 2011; 106:574–588.
34. Turner D, Walsh CM, Steinhart AH, et al Response to corticosteroids in severe ulcerative colitis: a systematic review of the literature and a meta-regression. Clin Gastroenterol Hepatol 2007; 5:103–110.
35. Turner D, Kolho KL, Mack DR, et al Glucocorticoid bioactivity does not predict response to steroid therapy in severe pediatric ulcerative colitis. Inflamm Bowel Dis 2010; 16:469–473.
36. Bossa F, Fiorella S, Caruso N, et al Continuous infusion versus bolus administration of steroids in severe attacks of ulcerative colitis: a randomized, double-blind trial. Am J Gastroenterol 2007; 102:601–608.
37. Meyers S, Lerer PK, Feuer EJ, et al Predicting the outcome of corticoid therapy for acute ulcerative colitis: results of a prospective, randomized, double-blind clinical trial. J Clin Gastroenterol 1987; 9:50–54.
38. Baron J, Connell A, Kanaghinis T, et al Out-patient treatment of ulcerative colitis. Br Med J 1962; 2:441.
39. Sachar DB. Pulse steroids for ulcerative colitis: good news, bad news, and no news. J Clin Gastroenterol 2002; 35:291–292.
40. Lichtenstein GR, Abreu MT, Cohen R, et al American Gastroenterological Association Institute technical review on corticosteroids, immunomodulators, and infliximab in inflammatory bowel disease. Gastroenterology 2006; 130:940–987.

corticosteroids; pediatric; propensity score; ulcerative colitis

© 2016 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,