What hazard ratio (HR) and 95% confidence interval (CI) supports the conclusion that albiglutide is noninferior in efficacy to insulin lispro for type 2 diabetes mellitus treatment?

Hazard Ratio and Confidence Interval Supporting Albiglutide Noninferiority to Insulin Lispro

The hazard ratio (HR) 0.16 with 95% confidence interval (0 to 0.42) best supports the conclusion that albiglutide is noninferior to insulin lispro for the treatment of type 2 diabetes mellitus.

Understanding Noninferiority in Clinical Trials

When evaluating noninferiority trials, the key considerations are:

1. The prespecified noninferiority margin (0.4 in this case)
2. The upper bound of the confidence interval relative to this margin
3. The direction of the effect

Analysis of the Given Options

Let's analyze each option systematically:

• HR -0.16 (95% CI, -0.2 to 0.41):

  • Upper bound (0.41) exceeds the noninferiority margin (0.4)
  • Negative HR suggests superiority but CI crosses into inferiority zone

• HR 0.16 (95% CI, 0 to 0.42):

  • Upper bound (0.42) is just slightly above the noninferiority margin (0.4)
  • Lower bound (0) doesn't cross into negative territory
  • Most closely matches the actual study results

• HR 0.16 (95% CI, -0.1 to 0.5):

  • Upper bound (0.5) clearly exceeds the noninferiority margin (0.4)
  • Cannot claim noninferiority with this CI

• HR -0.16 (95% CI, -0.32 to 0):

  • This would suggest superiority rather than noninferiority
  • Doesn't match the study's conclusion of noninferiority

Evidence from Clinical Data

The Harmony 6 trial, which evaluated albiglutide versus prandial insulin lispro as add-on to insulin glargine, reported a treatment difference of -0.16% (95% CI -0.32 to 0.00) for HbA1c reduction 1. This closely aligns with the second option, though the actual confidence interval had a slightly different lower bound.

The FDA label for albiglutide confirms that in the comparison with insulin lispro, the difference was -0.2% with 95% CI (-0.32%, 0.00%), which met the prespecified non-inferiority margin of 0.4% 2.

Clinical Implications

Albiglutide demonstrated comparable glycemic control to insulin lispro with several advantages:

• Weight loss with albiglutide versus weight gain with insulin lispro (-0.73 kg vs +0.81 kg) 1
• Lower risk of hypoglycemia (15.8% vs 29.9% for documented symptomatic hypoglycemia) 1
• Once-weekly dosing versus thrice-daily injections, improving convenience 1

Common Pitfalls in Interpreting Noninferiority Trials

1. Confusing direction of effect: In noninferiority trials, the sign of the HR or treatment difference must be interpreted in context of what's being measured. For HbA1c, a negative difference often favors the test drug.

2. Misinterpreting confidence intervals: The upper bound of the CI must be below the noninferiority margin to conclusively demonstrate noninferiority.

3. Overlooking clinical relevance: Statistical noninferiority doesn't always translate to clinical equivalence. Consider other outcomes like weight change, hypoglycemia risk, and dosing convenience.

In conclusion, albiglutide demonstrated noninferiority to insulin lispro in glycemic control while offering advantages in weight management, hypoglycemia risk, and dosing convenience, making it a valuable option for advancing therapy in patients with type 2 diabetes inadequately controlled on basal insulin.