Steven M. Willi
VerifiedUniversity of Pennsylvania · Rehabilitation Medicine
Active 1980–2025
About
Steven M. Willi, BA, MD, ML, is a Professor of Pediatrics specializing in Endocrinology and Diabetes at the Children's Hospital of Philadelphia. He is a full-time faculty member at the University of Pennsylvania and has been involved in designing clinical diabetes programming and conducting research for over 25 years. His research interests focus on pediatric diabetes and obesity, including efforts to delay the progression of type 1 diabetes, assess new treatments for type 1 and type 2 diabetes in children, and examine the safety and pharmacokinetics of various medications. Dr. Willi has contributed significantly to optimizing type 1 diabetes therapy, investigating risk factors for complications, and demonstrating the benefits of continuous insulin pump therapy. He has also been involved in immunomodulatory trials aimed at preserving β-cell function in new-onset T1DM, and in studies related to the course and treatment of type 2 diabetes in youth, including the influential TODAY study. Additionally, his work explores the role of pediatric obesity in cardiovascular disease, emphasizing clinical and lifestyle interventions to improve cardiovascular health. With a background that includes degrees from Johns Hopkins University and a healthcare law degree from the University of Pennsylvania Carey Law School, Dr. Willi has held leadership roles such as Director of the Diabetes Center for Children at CHOP and has served on various NIH and national trial committees. His clinical expertise encompasses intensive management of diabetes, novel therapies, and the use of advanced insulin delivery systems, with a focus on improving outcomes for pediatric patients with diabetes and obesity.
Research topics
- Endocrinology
- Medicine
- Internal medicine
- Pediatrics
Selected publications
2025-11-12
articleOpen access<p dir="ltr">Objective: To evaluate inhaled technosphere insulin (TI) in children with diabetes.</p><p dir="ltr">Research Design and Methods: Youth 4 to 17 years old with type 1 (98%) or type 2 (2%) diabetes treated with multiple daily injections of insulin were randomly assigned 1:1 to TI or rapid-acting analogue (RAA) insulin plus continuation of long-acting basal insulin and continuous glucose monitoring (CGM) for 26 weeks. The primary outcome was change in HbA1c, tested for non-inferiority with margin of 0.4%.</p><p dir="ltr">Results: In intent-to-treat analysis, mean HbA1c was 8.22±0.87% at baseline and 8.41±1.38% at 26 weeks with TI and 8.21±0.96% and 8.21±1.10%, respectively, with RAA (adjusted difference = 0.18%, 95% CI -0.07 to 0.43, non-inferiority p-value = 0.091). CGM-measured time-in-range 70-180 mg/dL was not significantly different between groups (adjusted difference -2.2%, 95% CI -7.0% to 2.7%, p=0.38). Two severe hypoglycemic events occurred in the TI group and one in the RAA group. Change in forced expiration volume in one second (FEV1) from baseline to 26 weeks did not differ comparing TI and RAA (p=0.53). The TI group reported greater treatment satisfaction (p=0.004) and had less gain in weight and body mass index percentile (p=0.009) than the RAA group.</p><p dir="ltr">Conclusions and Relevance: The primary analysis did not meet the pre-specified criteria for HbA1c non-inferiority. However, TI use was safe over 26 weeks without impacting pulmonary function and was associated with greater treatment satisfaction and less weight gain compared with RAA, supporting TI as a treatment option for some pediatric patients with type 1 diabetes.</p><p><br></p>
Impact of Empagliflozin on Glomerular Hyperfiltration and Albuminuria in Youth with Type 2 Diabetes
Clinical Journal of the American Society of Nephrology · 2025-11-11
articleOpen accessKey Points We investigated the effects of empagliflozin on hyperfiltration and albuminuria in youth with type 2 diabetes. Hyperfiltration and albuminuria were significantly attenuated compared with placebo in adolescents aged 10–17 years with type 2 diabetes on empagliflozin. Additional studies are required to evaluate the effect of sodium-glucose cotransporter-2 inhibition on diabetic kidney disease risk in youth. Background Hyperfiltration, common in youth with type 2 diabetes (T2D), may increase the risk of early diabetic kidney disease. The Diabetes Study of Linagliptin and Empagliflozin in Children and Adolescents and Diabetes Study of Linagliptin and Empagliflozin in Children and Adolescents MONO trials showed that compared with placebo, empagliflozin improved glycemic control in youth with T2D. This post hoc analysis evaluated empagliflozin versus placebo on selected parameters in participants from both trials according to their baseline (BL) hyperfiltration/normofiltration status. Methods We calculated eGFR using the Zappitelli equation (combined serum creatinine and cystatin C), with hyperfiltration defined as 2 SDs above average for healthy youth per National Health and Nutrition Examination Survey (>126.8 ml/min per 1.73 m 2 ). After randomization, 116 participants received empagliflozin (10 mg or 25 mg) or placebo. We compared responses to therapy for eGFR and urine albumin-creatinine ratio (UACR) at week 26 by BL hyperfiltration and normofiltration (≤126.8 ml/min per 1.73 m 2 ), with empagliflozin groups pooled. Results Empagliflozin treatment led to more significant reductions in eGFR among participants with BL hyperfiltration compared with normofiltration ( P interaction = 0.01). At week 26, eGFR decreased significantly with empagliflozin versus placebo in those with hyperfiltration (adjusted mean difference [95% confidence interval], −11.67 ml/min per 1.73 m 2 [−19.90 to −3.43]; P = 0.006); 15% of those with hyperfiltration on empagliflozin shifted to normofiltration at week 26 versus 6% for placebo. At week 26, the geometric mean ratio of UACR was 55% lower with empagliflozin versus placebo in participants with UACR ≥30 mg/g at BL (0.45 [0.22 to 0.93]; P = 0.03). In the normofiltration subgroup, eGFR and UACR changes over time were similar between treatment groups. Conclusions Empagliflozin was associated with attenuated hyperfiltration and albuminuria in youth with T2D. Future longitudinal evaluations can assess potential long-term benefits of treatment. Clinical Trial registry name and registration number: ClinicalTrials.gov, NCT03429543.
SUN-797 Management of Hypercalcemia in an Infant With Fungemia and Its Challenges
Journal of the Endocrine Society · 2025-10-01
articleOpen accessSenior authorAbstract Disclosure: S. Muneer: None. M. Munoz: None. J. Iyer: None. D.R. Weber: None. S.M. Willi: None. Background: Hypercalcemia in infants is rare but potentially serious. While often asymptomatic, it may present with failure to thrive, constipation, dehydration, poor feeding, and in severe cases, hypotonia or seizures. Differential diagnosis includes hypervitaminosis, malignancy, endocrinopathies, granulomatous disease, and fat necrosis. Case: A full-term male infant with congenital lymphatic malformation and neonatal chylothorax developed candidal fungemia on day of life (DOL) 30. Before this diagnosis, he had been noted to have hypercalcemia, 11.5mg/dL (ref 7.8-11.3) on DOL 28. His fungemia was treated with amphotericin B on DOL 32-35 and then fluconazole DOL 36-47. During this time his Ca remained elevated and did not improve. Initial evaluation for hypercalcemia commenced on DOL 48 which showed hypercalcemia 12.7mg/dL, normal 25-OH Vitamin D, low 1,25-dihydroxy Vitamin D, appropriately suppressed PTH, and elevated urinary Ca/creatinine. Bone turnover markers, including c-telopeptide and bone-specific alkaline phosphatase, were normal. The patient initially responded to hyperhydration and removal of calcium from feeds, but hypercalcemia recurred. On DOL 63, he required calcitonin at 2 IU/kg BID. Calcitonin was discontinued after 17 days as calcium normalized, and 1,25-dihydroxy Vitamin D returned to normal. Discussion: Infant hypercalcemia is rare, and its association with fungemia is even more unusual. Hypercalcemia secondary to fungemia is suspected to result from extra-renal 1,25OHD2 production by activated macrophages expressing CYP27B1 (1-alpha hydroxylase). Excess 1,25OHD2 causes hypercalcemia via increased intestinal absorption and increased bone resorption of calcium. In this case, hypercalcemia in the setting of elevated 1,25OHD2 and suppressed PTH is consistent with extra-renal production by activated macrophages. Fluconazole is an inhibitor of CYP27B1, and therefore exerts a beneficial effect on hypercalcemia by reducing 1,25OHD2 synthesis, in addition to treating the fungemia. Calcitonin, an inhibitor of osteoclast activity, is effective in the acute management of hypercalcemia driven in whole or in part by excess bone resorption. Steroids are also an effective treatment of hypercalcemia secondary to activated macrophages, however, are typically not an option with a fungal infection because of the risk of reactivating a systemic infection. This case adds to the limited literature on hypercalcemia secondary to systemic infections in infants. It was noted that the patient started to develop hypercalcemia shortly after the diagnosis of the fungemia and failed to improve while receiving fluconazole. Calcitonin proved effective in rapidly reducing serum Ca levels, highlighting its utility in acute management. Further studies are warranted to understand the pathophysiological mechanisms linking infection and hypercalcemia in infants. Presentation: Sunday, July 13, 2025
2025-11-12
articleOpen access<p dir="ltr">Objective: To evaluate inhaled technosphere insulin (TI) in children with diabetes.</p><p dir="ltr">Research Design and Methods: Youth 4 to 17 years old with type 1 (98%) or type 2 (2%) diabetes treated with multiple daily injections of insulin were randomly assigned 1:1 to TI or rapid-acting analogue (RAA) insulin plus continuation of long-acting basal insulin and continuous glucose monitoring (CGM) for 26 weeks. The primary outcome was change in HbA1c, tested for non-inferiority with margin of 0.4%.</p><p dir="ltr">Results: In intent-to-treat analysis, mean HbA1c was 8.22±0.87% at baseline and 8.41±1.38% at 26 weeks with TI and 8.21±0.96% and 8.21±1.10%, respectively, with RAA (adjusted difference = 0.18%, 95% CI -0.07 to 0.43, non-inferiority p-value = 0.091). CGM-measured time-in-range 70-180 mg/dL was not significantly different between groups (adjusted difference -2.2%, 95% CI -7.0% to 2.7%, p=0.38). Two severe hypoglycemic events occurred in the TI group and one in the RAA group. Change in forced expiration volume in one second (FEV1) from baseline to 26 weeks did not differ comparing TI and RAA (p=0.53). The TI group reported greater treatment satisfaction (p=0.004) and had less gain in weight and body mass index percentile (p=0.009) than the RAA group.</p><p dir="ltr">Conclusions and Relevance: The primary analysis did not meet the pre-specified criteria for HbA1c non-inferiority. However, TI use was safe over 26 weeks without impacting pulmonary function and was associated with greater treatment satisfaction and less weight gain compared with RAA, supporting TI as a treatment option for some pediatric patients with type 1 diabetes.</p><p><br></p>
Endocrinology and Metabolism Clinics of North America · 2025-04-11 · 1 citations
reviewSenior author2025-12-01
articleOpen access<p dir="ltr">Objective: To evaluate inhaled technosphere insulin (TI) in children with diabetes.</p><p dir="ltr">Research Design and Methods: Youth 4 to 17 years old with type 1 (98%) or type 2 (2%) diabetes treated with multiple daily injections of insulin were randomly assigned 1:1 to TI or rapid-acting analogue (RAA) insulin plus continuation of long-acting basal insulin and continuous glucose monitoring (CGM) for 26 weeks. The primary outcome was change in HbA1c, tested for non-inferiority with margin of 0.4%.</p><p dir="ltr">Results: In intent-to-treat analysis, mean HbA1c was 8.22±0.87% at baseline and 8.41±1.38% at 26 weeks with TI and 8.21±0.96% and 8.21±1.10%, respectively, with RAA (adjusted difference = 0.18%, 95% CI -0.07 to 0.43, non-inferiority p-value = 0.091). CGM-measured time-in-range 70-180 mg/dL was not significantly different between groups (adjusted difference -2.2%, 95% CI -7.0% to 2.7%, p=0.38). Two severe hypoglycemic events occurred in the TI group and one in the RAA group. Change in forced expiration volume in one second (FEV1) from baseline to 26 weeks did not differ comparing TI and RAA (p=0.53). The TI group reported greater treatment satisfaction (p=0.004) and had less gain in weight and body mass index percentile (p=0.009) than the RAA group.</p><p dir="ltr">Conclusions and Relevance: The primary analysis did not meet the pre-specified criteria for HbA1c non-inferiority. However, TI use was safe over 26 weeks without impacting pulmonary function and was associated with greater treatment satisfaction and less weight gain compared with RAA, supporting TI as a treatment option for some pediatric patients with type 1 diabetes.</p><p><br></p>
Journal of the Endocrine Society · 2025-10-01
articleOpen accessAbstract Disclosure: M. Munoz: None. T. Maxwell: None. S.M. Willi: None. M.R. Rickels: None. L. Katz: None. Background: Individuals with type 1 diabetes (T1D) face ongoing challenges to achieve recommended targets for glycemic control, further complicated by obesity and insulin resistance. Approximately 20% of individuals with T1D are obese. Glucagon-like peptide-1 receptor agonists (GLP1-RA), approved for type 2 diabetes and obesity, have emerged as potential adjunct therapies in T1D. Recent studies in adults with T1D and obesity demonstrate a reduction in HbA1c, weight, and total insulin dose (TDD). Hypothesis: We hypothesize that GLP1-RA are safe and effective in reducing weight and improving glycemic control in adolescents and young adults with T1D and obesity without increasing hypoglycemia. Methods: This retrospective cohort study was conducted at a large, urban academic children’s and adult hospitals. Patients were identified by an electronic health record search for GLP1-RA prescriptions between 2014 and 2024. Inclusion criteria were age 10-35 years with a diagnosis of T1D requiring insulin, obesity, and GLP1-RA treatment for 3 months.Primary outcomes included changes in BMI, and glucose time in range (TIR, 70-180 mg/dL) and time below range (TBR, &lt;70 mg/dL) on continuous glucose monitor (CGM). Secondary outcomes included TDD, HbA1c, and weight change. Outcomes were analyzed by paired t-test two sample for Means. Results: We identified 24 individuals (14 F, 10 M; mean age 19.4 yrs and diabetes duration 10.7 yrs). Charts were reviewed on the date of GLP1-RA initiation and two subsequent follow up visits: 3-5 months and 6-10 months after GLP1-RA initiation.Pre-treatment values: HbA1c 8.0% (n=22), weight 105.3 kg (n=23), BMI 37.1 (n=24), TDD 1.04 units/kg (n=18), TIR 55% (n=20), TBR 3.2% (n=20). At 1st follow-up: HbA1c 7.5% (n=16, p=0.05*), weight -4.4% (n=22, p=0.04*), BMI -4.8% (n=21, p=0.04*), TDD 0.95 units/kg (n=18, p=0.06), TIR 62.9% (n=19, p=0.06), TBR 2.1% (n=19, p=0.56). At 2nd follow-up: HbA1c 7.4% (n=16, p=0.21), weight -5.6% (n=16, p=0.03*), BMI -6.1% (n=16, p=0.01*), TDD 1.0 units/kg (n=12, p=0.07), TIR 63.3% (n=15, p=0.06), TBR 2.8% (n=15, p=0.35). Values of statistical significance (p&lt;0.05) marked by *.Side effects were predominantly gastrointestinal that were self-reported and without hospitalization. No episodes of severe hypoglycemia requiring assistance were documented and only 1 episode of DKA was observed in a patient with a history of insulin non-compliance. Conclusion: In this case series, patients experienced improvements in BMI and initial improvement in HbA1C. This study is limited due to size, documentation inconsistencies, and incomplete follow up from patients. Further studies should focus on larger populations powered to detect statistical differences and long-term effects. GLP1-RA offers a promising and safe adjunctive therapy for patients with T1D and obesity due to its effects on weight, BMI, and HbA1c. Presentation: Saturday, July 12, 2025
Diabetes Care · 2025-11-12 · 5 citations
articleOpen accessOBJECTIVE: To evaluate inhaled technosphere insulin (TI) in children with diabetes. RESEARCH DESIGN AND METHODS: A total of 230 youth 4-17 years old with type 1 (98%) or type 2 (2%) diabetes treated with multiple daily injections of insulin were randomly assigned 1:1 to TI or rapid-acting analog (RAA) insulin plus continuation of long-acting basal insulin and continuous glucose monitoring (CGM) for 26 weeks. The primary outcome was change in HbA1c, tested for noninferiority with margin of 0.4%. RESULTS: In intent-to-treat analysis, mean HbA1c (% ± SD) was 8.22 ± 0.87 at baseline and 8.41 ± 1.38 at 26 weeks with TI and 8.21 ± 0.96 and 8.21 ± 1.10, respectively, with RAA (adjusted difference = 0.18; 95% CI -0.07, 0.43; noninferiority P = 0.091). CGM-measured time in range 70-180 mg/dL was not significantly different between groups (adjusted difference -2.2%; 95% CI -7.0, 2.7; P = 0.38). Two severe hypoglycemic events occurred in the TI group and one in the RAA group. Change in forced expiration volume in 1 s from baseline to 26 weeks did not differ comparing TI and RAA (P = 0.53). The TI group reported greater treatment satisfaction (P = 0.004) and had less gain in weight and BMI percentile (P = 0.009) than did the RAA group. CONCLUSIONS: The primary analysis did not meet the prespecified criteria for HbA1c noninferiority. However, TI use was safe over 26 weeks without affecting pulmonary function and was associated with greater treatment satisfaction and less weight gain compared with RAA, supporting TI as a treatment option for some pediatric patients with type 1 diabetes.
Journal of clinical lipidology · 2025-02-06 · 1 citations
articleOpen accessJournal of the Endocrine Society · 2025-06-09 · 1 citations
articleOpen accessSenior authorContext: There is an unmet need for type 2 diabetes (T2D) treatments in addition to metformin and insulin for adolescents. This is due to the challenges of monotherapy in youth with T2D and need for treatment escalation to maintain glycemic control in youth generally more so than in young adults. Objective: We assessed the efficacy and safety of sodium-glucose co-transporter-2 (SGLT2) or dipeptidyl peptidase-4 (DPP-4) inhibitor monotherapies in adolescents and young adults with T2D not on active therapy. Methods: Drug-naïve adolescents and those not on active therapy received the SGLT2 inhibitor empagliflozin, the DPP-4 inhibitor linagliptin, or placebo for 26 weeks; young adults with no antidiabetic background therapy received empagliflozin, the DPP-4 inhibitor sitagliptin, or placebo for 24 weeks. The primary endpoint was treatment failure occurrence. Secondary outcomes assessed glycated hemoglobin A1c (HbA1c), fasting plasma glucose, and weight. Results: = .017). Empagliflozin modestly reduced mean HbA1c vs placebo in adolescents (-0.35% vs 0.41%) compared with greater reductions in young adults (-1.01% vs -0.30%). No new safety signals were identified. Conclusion: Empagliflozin reduced HbA1c in adolescents and young adults; however, these results highlight the challenges of monotherapy for youth with T2D and need for further studies.
Recent grants
NIH · $66k
Career Development Program in Pediatric Diabetes Research (K12) at CHOP
NIH · $3.6M · 2011–2022
Frequent coauthors
- 184 shared
Linda A. DiMeglio
Indiana University – Purdue University Indianapolis
- 154 shared
Roy W. Beck
Jaeb Center for Health Research
- 150 shared
Stéphanie Woerner
- 143 shared
Ruth S. Weinstock
SUNY Upstate Medical University
- 139 shared
Jamie Wood
Icahn School of Medicine at Mount Sinai
- 135 shared
Richard M. Bergenstal
Park Nicollet Clinic
- 133 shared
Paul McGuigan
- 130 shared
Michael R. Rickels
University of Pennsylvania
Awards & honors
- Clinical Associate Physician Award from NIH
- Resume-aware match score
- Save to shortlist
- AI-drafted outreach
See your match with Steven M. Willi
PhdFit ranks faculty by your research interests, methods, and publications — grounded in their actual work, not templates.
- Free to start
- No credit card
- 30-second signup