Adjunctive Corticosteroids in Acute Pyelonephritis to Reduce Scarring
A three-year-old girl with a high-risk early DMSA finding, and parents who found the steroid literature before the team raised it, sits directly on top of a real disagreement between an older positive meta-analysis and the largest, most rigorous trial to date.
Freya S., a three-year-old girl, was admitted two days ago with her first documented urinary tract infection — fever to 39.6°C, flank tenderness, and a urine culture growing more than 100,000 colony-forming units of E. coli, treated since admission with intravenous ceftriaxone. An early DMSA scan obtained on admission showed an inflammatory volume of 6.1mL, above the 4.6mL threshold some trials have used to flag children at meaningfully higher risk of ending up with a permanent kidney scar. She was previously healthy, met all her developmental milestones on schedule, and her parents — a physician assistant and a data analyst who read the discharge-planning materials more thoroughly than most families do — arrived at rounds this morning having already found a study online asking whether a short course of steroids alongside the antibiotic could lower her odds of scarring, a question they asked before the team had raised it themselves.
The literature genuinely disagrees, not just in emphasis. Huang and colleagues' 2011 randomized trial of adjunctive oral methylprednisolone in children selected for high scarring risk — by an elevated early DMSA volume much like Freya's — found scarring in 33.3% of the steroid group versus 60.0% of placebo, and Meena and colleagues' subsequent 2021 meta-analysis, pooling three trials and 529 children, found a real benefit, a risk ratio of 0.57 favoring corticosteroids. But the largest and most methodologically careful trial to date, Rius-Gordillo and colleagues' DEXCAR trial, a multicenter, placebo-controlled study of ninety-one children given three days of intravenous dexamethasone, found essentially no difference — 22% scarring with dexamethasone against 21% with placebo — and its own regression analysis found early DMSA severity, not treatment arm, was what actually predicted who scarred.
Two trials, two answers, one three-year-old
I'd add methylprednisolone. Huang and colleagues' trial, which used almost exactly Freya's own risk criteria — an elevated early DMSA volume — found scarring cut nearly in half, from 60% to 33%, and Meena and colleagues' subsequent meta-analysis of three trials and 529 children found a real pooled benefit, risk ratio 0.57. She fits the population that trial was built around about as closely as a real patient ever does.
That trial's numbers are real, and I won't dismiss them — but Rius-Gordillo and colleagues' DEXCAR trial, the largest, most rigorously designed trial on this exact question, ninety-one children, multicenter, placebo-controlled, found essentially no difference at all: 22% scarring with dexamethasone versus 21% with placebo.
Weighting an 84-patient single-center trial and a meta-analysis that includes it over a larger, later, better-controlled negative trial gets the usual evidence hierarchy backwards — bigger and more rigorous should update our confidence more than smaller and earlier, not less, and that later trial's own regression analysis found her DMSA severity predicts scarring on its own, without any credit going to the drug.
I don't think either of you is wrong about your own trial — I think the trials may not be answering quite the same question. Methylprednisolone by mouth and dexamethasone by vein aren't interchangeable regimens, and neither group has been tested head-to-head. Rather than resolve that today, I'd hold off on steroids for Freya specifically because the larger, more recent trial is the closer match to how we'd actually deliver this — IV, inpatient — and put our energy into what both trials agree matters regardless: a formal repeat DMSA at six months to know, for her specifically, whether a scar actually forms.
No corticosteroid was added to Freya's regimen; ceftriaxone continued per her culture sensitivities, and a six-month DMSA scan was scheduled before discharge.
Not resolved, and stated as such to the family rather than smoothed over: the pediatric literature on this question currently contains a positive meta-analysis and a negative large trial that weren't designed to be compared head-to-head, and this team's choice to follow the larger trial is a considered judgment call, not a settled answer the parents' own reading was wrong to question.