Scholarlink

Background <p>Sulfonylureas (SU) are widely used for diabetes management in older adults but can cause hypoglycemia, which may be worsened by drug interactions. We applied high-throughput data mining to identify medications that could increase hypoglycemia risk when taken with SU.</p> Methods <p>Using Medicare, MarketScan, and Optum Clinformatics (2003–2022), we identified patients aged ≥ 65 years who experienced a severe hypoglycemic event after at least 90 days on SU. We evaluated all medications dispensed in the 90 days before the event using a case-crossover (CCO) design. We adjusted for time-varying confounding and direct effect of the evaluated medications (precipitant) using a case-case time-control (CCTC) approach and metformin as control. We computed odds ratios (ORs) for its association with hypoglycemia. The false discovery rate (FDR) was controlled at 0.05 to adjust for multiple testing. To reduce confounding from other diabetes medications, we analyzed non-diabetes and diabetes medications separately.</p> Results <p>Among 1607 candidate drugs received before experiencing hypoglycemia, 86 non-diabetes medications showed a CCO OR ≥ 1.00. With metformin as control, sulfamethoxazole/trimethoprim (CCTC OR 1.76, <i>p</i> &lt; 0.01, FDR <i>q</i> &lt; 0.01) and metronidazole (CCTC OR 2.17, <i>p</i> &lt; 0.01, FDR <i>q =</i> 0.04) were associated with severe hypoglycemia. Among 10 diabetes medications, insulin showed increased association (CCO OR 1.22, <i>p</i> &lt; 0.01); however, once adjusted for the drug’s direct effects, CCTC OR was 1.03 (<i>p =</i> 0.47, FDR <i>q =</i> 0.47).</p> Conclusions <p>Using a high-throughput data mining approach, we identified two antibiotics (sulfamethoxazole/trimethoprim and metronidazole) that may increase hypoglycemia risk in older adults on sulfonylureas. Given the exploratory nature of this study, these findings warrant further investigation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Medications Associated with Increased Risk of Hypoglycemia in Older Adults on Sulfonylureas: A High-Throughput Case-Crossover-Based Screening Study

  • 期刊论文

  • 发表日期：

    2025年11月27日

查看全文
  • Sungho Bea,
  • Elisabetta Patorno,
  • Georg Hahn,
  • Julie M. Paik,
  • Deborah J. Wexler,
  • Katsiaryna Bykov

摘要

Background

Sulfonylureas (SU) are widely used for diabetes management in older adults but can cause hypoglycemia, which may be worsened by drug interactions. We applied high-throughput data mining to identify medications that could increase hypoglycemia risk when taken with SU.

Methods

Using Medicare, MarketScan, and Optum Clinformatics (2003–2022), we identified patients aged ≥ 65 years who experienced a severe hypoglycemic event after at least 90 days on SU. We evaluated all medications dispensed in the 90 days before the event using a case-crossover (CCO) design. We adjusted for time-varying confounding and direct effect of the evaluated medications (precipitant) using a case-case time-control (CCTC) approach and metformin as control. We computed odds ratios (ORs) for its association with hypoglycemia. The false discovery rate (FDR) was controlled at 0.05 to adjust for multiple testing. To reduce confounding from other diabetes medications, we analyzed non-diabetes and diabetes medications separately.

Results

Among 1607 candidate drugs received before experiencing hypoglycemia, 86 non-diabetes medications showed a CCO OR ≥ 1.00. With metformin as control, sulfamethoxazole/trimethoprim (CCTC OR 1.76, p < 0.01, FDR q < 0.01) and metronidazole (CCTC OR 2.17, p < 0.01, FDR q = 0.04) were associated with severe hypoglycemia. Among 10 diabetes medications, insulin showed increased association (CCO OR 1.22, p < 0.01); however, once adjusted for the drug’s direct effects, CCTC OR was 1.03 (p = 0.47, FDR q = 0.47).

Conclusions

Using a high-throughput data mining approach, we identified two antibiotics (sulfamethoxazole/trimethoprim and metronidazole) that may increase hypoglycemia risk in older adults on sulfonylureas. Given the exploratory nature of this study, these findings warrant further investigation.