Skip to main content

Table 2 Differences in proportions of patients with abnormally low values at the endpoint of the SIRIUS study

From: Improving analysis practice of continuous adverse event outcomes in randomised controlled trials - a distributional approach

Outcome

Placebo, n/N (proportion)

Mepolizumab, n/N (proportion)

Linear regression

Automated procedure

Empirical estimates

Adjusted¥ mean difference [95% CI], p-value

Adjusted¥ difference in proportion [95% DCI], p-valueΩ

Fisher’s exact test p-value

Difference in proportion [95% CI]

Alanine Aminotransferase(IU/L)

0/61 (0.00)

0/66 (0.00)

− 0.71 [− 3.97, 2.56], 0.67

0.003 [− 0.01, 0.01], 0.67

NE

0.00 [0.00, 0.00]

Calcium (mmol/L)

0/61 (0.00)

1/66 (0.02)

0.005 [− 0.03, 0.04], 0.85

− 0.0008 [− 0.005, 0.004], 0.85

> 0.99

0.02 [− 0.01, 0.05]

Eosinophils (109/L)

2/62 (0.03)

27/65 (0.42)

− 0.36 [− 0.43, − 0.28], < 0.01

0.48 [0.38, 0.59], < 0.01

< 0.01

0.39 [0.26, 0.52]

Glucose (mmol/L)

1/61 (0.02)

2/66 (0.03)

0.01 [− 0.34, 0.36], 0.96

− 0.002 [− 0.07, 0.06], 0.96

> 0.99

0.01 [− 0.04, 0.06]

Haematocrit (fraction of 1)

1/62 (0.02)

4/66 (0.06)

− 0.03 [− 0.04, − 0.01], < 0.01

0.02 [0.01, 0.03], < 0.01

0.37

0.04 [− 0.03, 0.11]

Haemoglobin (g/L)

4/62 (0.06)

8/66 (0.12)

− 8.18 [− 12.73, − 3.62], < 0.01

0.06 [0.03, 0.09], < 0.01

0.37

0.06 [− 0.04, 0.16]

Lymphocytes/Leukocytes (%)

12/62 (0.19)

10/65 (0.15)

0.69 [− 2.85, 4.22], 0.70

− 0.02 [− 0.09, 0.05], 0.70

0.64

− 0.04 [− 0.17, 0.09]

Platelets (109/L)

0/61 (0.00)

0/65 (0.00)

18.75 [− 3.86, 41.37], 0.10

− 0.004 [− 0.01, 0.002], 0.10

NE

0.00 [0.00, 0.00]

Potassium (mmol/L)

4/61 (0.07)

0/66 (0.00)

0.10 [− 0.003, 0.20], 0.06

− 0.02 [− 0.03, − 0.001], 0.06

0.06

− 0.07 [− 0.13, − 0.01]

Sodium (mmol/L)

0/61 (0.00)

1/66 (0.02)

− 0.43 [− 1.03, 0.17], 0.16

1.41e−15 [− 1.23e−15, 4.04e−15], 0.16

> 0.99

0.02 [− 0.01, 0.05]

  1. Ωp-value from the adjusted linear regression model
  2. CI confidence interval, DCI distributional confidence interval, SE standard error, NE non-estimable
  3. N.B: Positive estimates correspond to a greater proportion in the placebo arm; shaded results are significant at p < 0.05
  4. ¥Adjusted for country and duration of previous use of oral glucocorticoids (< 5 years vs. ≥5 years)