Sample Size Determination Example

AREA: Means | Cardiology

nQuery In Use Background Web  Opacity 35

Bayer Pulmonary Hypertension Research

TitleRiociguat for the Treatment of Pulmonary Arterial Hypertension.

Objective: In this phase 3, double-blind study, we randomly assigned 443 patients with symptomatic pulmonary arterial hypertension to receive placebo, riociguat in individually adjusted doses of up to 2.5 mg three times daily (2.5 mg–maximum group), or riociguat in individually adjusted doses that were capped at 1.5 mg three times daily (1.5 mg–maximum group). The 1.5 mg–maximum group was included for exploratory purposes, and the data from that group were analyzed descriptively. Patients who were receiving no other treatment for pulmonary arterial hypertension and patients who were receiving endothelin-receptor antagonists or (nonintravenous) prostanoids were eligible. The primary end point was the change from baseline to the end of week 12 in the distance walked in 6 minutes. Secondary end points included the change in pulmonary vascular resistance, N-terminal pro–brain natriuretic peptide (NT-proBNP) levels, World Health Organization (WHO) functional class, time to clinical worsening, score on the Borg dyspnea scale, quality-of-life variables, and safety.

Year: 2013

Source: New England Journal of Medicine

Link: www.nejm.org/doi/full/10.1056/NEJMoa1209655

Clinical Area: Cardiology

Sample Size Section in Paper/Protocol:

“The primary efficacy outcome is change from baseline to end of study in 6 minute walking distance test. Assuming a standard deviation of 70 m, a power of 90 % and a two-sided significance level of 5 %, with a 4:2 randomization, then to detect a placebo-adjusted difference of 25 m, 250 and 125 patients would be required valid for ITT in the Bay 63-2521 Individual Dose Titration Arm and placebo group respectively, for a total of 375 patients. In addition, the exploratory 1.5 mg Dose Arm would have a sample size of one-half that of the placebo arm, approximately 63 patients. Hence, the total number of valid for ITT patients should be 438. Allowing for an invalidity rate of 5%, a total of 462 randomized patients would be required.”

 

Note: Final analysis was conducted using ANCOVA model using baseline as a covariate.This study or resultant future studies could use effect of baseline covariate to adjust sample size downwards due to between-subjects variance decrease


Summary of Necessary Parameter Estimates for Sample Size Calculation:

 Parameter Value
 Significance Level 0.05
 Expected Placebo Mean (From Baseline) 363
 Expected Treatment Mean 363 + 25 = 388
 Standard Deviation (Common) 70
 Power 90%

Email me this example in PDF format

Step 1:
Search for the Two Sample Student’s t-test  (equal variances) Unequal n’s from the Select Test Design & Goal window.

This can be done using the radio buttons or alternatively, you can use the search bar at the end of the Select Test Design & Goal window.

Sample Size Calculator Example- nQuery- Example 07- Img 01- Two Sample Student’s t-test  (equal variances)  Unequal n’s

Step 2:
Enter the parameter values for sample size calculation taken from the study protocol.

Sample Size Calculator Example- nQuery- Example 07- Img 02- Two Sample Student’s t-test  (equal variances)  Unequal n’s

Step 3:
Once Power is entered, sample size is calculated automatically.

This gives the sample sizes of 125 and 250 as per the protocol

 

Sample Size Calculator Example- nQuery- Example 07- Img 03- Two Sample Student’s t-test  (equal variances)  Unequal n’s

Note:  In the case of ANCOVA analysis, the only additional information necessary would be an estimate of R2 between the response and the baseline covariate (or average R2 if other covariates were used).

For example, using the ANCOVA sample size table available in nQuery for the above study and assuming a 10% R2 between baseline and the response would have yielded an adjusted total sample size of 336.

Step 4:
Once the calculation is completed, nQuery Advanced provides an output statement summarizing the results. It States:

Output Statement:

A two group t-test with a 5% two-sided significance level will have 90% power to detect the difference between a Group 1 mean, µ₁, of 363 and a Group 2 mean, µ₂, of 338 a difference in means of 25 assuming that the common standard deviation is 70, when the sample sizes in the two groups are 124 and 250, respectively (a total sample size of 374).

 

Sample Size Calculator Example- nQuery- Example 07- Img 04- Two Sample Student’s t-test  (equal variances)  Unequal n’s

Email me this example in PDF format

Free 14 Day Trial

Demo Icon white
No download required