0
Correspondence |

Composite End Points of Death and Hospitalization Are the Only Appropriate Option for Most TrialsComposite End Points in Clinical Trials FREE TO VIEW

Ichraq Latar, BSc; Nicolas Girerd, MD, PhD
Author and Funding Information

From INSERM, Centre d’Investigations Cliniques 9501, Université de Lorraine, CHU de Nancy, Institut Lorrain du Cœur et des Vaisseaux.

CORRESPONDENCE TO: Nicolas Girerd, MD, PhD, Centre d’Investigations Cliniques 9501, Rue du Morvan, Institut Lorrain du Coeur et des Vaisseaux, CHU de Nancy - site Brabois, 54500 Vandoeuvre les Nancy, France; e-mail: nicolas_girerd@yahoo.com


Editor’s Note: Authors are invited to respond to Correspondence that cites their previously published work. Those responses appear after the related letter. In cases where there is no response, the author of the original article declined to respond or did not reply to our invitation.

FINANCIAL/NONFINANCIAL DISCLOSURES: The authors have reported to CHEST that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details.


Chest. 2015;147(5):e189. doi:10.1378/chest.14-3246
Text Size: A A A
Published online
To the Editor:

We read with great interest the article by Collard et al1 in a recent issue of CHEST (November 2014) regarding the usefulness of composite end points of death or hospitalization in idiopathic pulmonary fibrosis. The authors highlight the large impact of the use of composite end points on statistical power and subsequent trial sample size.

Aside from statistical power considerations, we would like to comment on other methodologic aspects that deserve attention. Indeed, the use of hospitalization rather than hospitalization or death would result in very similar power because of the much higher hospitalization rates compared with mortality rates. In addition, the use of composite end points remains a matter of controversy. For instance, Montori et al2 suggested certain criteria required for their appropriate use, including the clinical relevance of individual components, the frequency of each component, as well as treatment effects on each component. Hence, why not use hospitalizations rather than composite end point, which would result in very similar power and would facilitate conclusions regarding treatment effect?

First, in keeping with previous reports from the clinical trial literature,3 since the conclusion of a trial is solely based on the primary end point, the latter should capture a treatment effect that is relevant to clinical practice. Importantly, trialists should fear treatment effects in the opposite direction. Since the decrease in hospitalization rates would hardly indicate clinical benefit if their reduction is gained at the expense of higher mortality rates, only the assessment of hospitalization and death is clinically relevant.3

Second, the high methodologic value of composite end points is related to competing risks. Indeed, when hospitalization alone is considered, noninformative censoring of death is hypothesized, which is an unverifiable and probably dangerous assumption. As highlighted in previous reports,4 traditional survival analysis overestimates the risk of morbid event rates when the mortality rate is high. In this setting, the appropriate strategy is to model both events at the same time—a survival analysis strategy called the competing risk approach, such as the competing risk regression of Fine and Gray.4 Nonetheless, this strategy is more technically challenging and does not overcome the problem that arises from treatment effect in the opposite direction.

As a result, we strongly advocate for the use of composite end points in clinical trials when targeting all-cause death is not feasible. From our viewpoint, composite end points are the only appropriate option for most trials, which eventually prompted their use in several fields of medicine, including oncology and cardiology.3,5

References

Collard HR, Brown KK, Martinez FJ, Raghu G, Roberts RS, Anstrom KJ; for the IPFnet Investigators. Study design implications of death and hospitalization as end points in idiopathic pulmonary fibrosis. Chest. 2014;146(5):1256-1262. [CrossRef] [PubMed]
 
Montori VM, Permanyer-Miralda G, Ferreira-González I, et al. Validity of composite end points in clinical trials. BMJ. 2005;330(7491):594-596. [CrossRef] [PubMed]
 
Green S, Benedetti J, Smith A, Crowley J. Clinical Trials in Oncology.3rd ed. Boca Raton, FL: CRC Press; 2012.
 
Berry SD, Ngo L, Samelson EJ, Kiel DP. Competing risk of death: an important consideration in studies of older adults. J Am Geriatr Soc. 2010;58(4):783-787. [CrossRef] [PubMed]
 
Skali H, Pfeffer MA, Lubsen J, Solomon SD. Variable impact of combining fatal and nonfatal end points in heart failure trials. Circulation. 2006;114(21):2298-2303. [CrossRef] [PubMed]
 

Figures

Tables

References

Collard HR, Brown KK, Martinez FJ, Raghu G, Roberts RS, Anstrom KJ; for the IPFnet Investigators. Study design implications of death and hospitalization as end points in idiopathic pulmonary fibrosis. Chest. 2014;146(5):1256-1262. [CrossRef] [PubMed]
 
Montori VM, Permanyer-Miralda G, Ferreira-González I, et al. Validity of composite end points in clinical trials. BMJ. 2005;330(7491):594-596. [CrossRef] [PubMed]
 
Green S, Benedetti J, Smith A, Crowley J. Clinical Trials in Oncology.3rd ed. Boca Raton, FL: CRC Press; 2012.
 
Berry SD, Ngo L, Samelson EJ, Kiel DP. Competing risk of death: an important consideration in studies of older adults. J Am Geriatr Soc. 2010;58(4):783-787. [CrossRef] [PubMed]
 
Skali H, Pfeffer MA, Lubsen J, Solomon SD. Variable impact of combining fatal and nonfatal end points in heart failure trials. Circulation. 2006;114(21):2298-2303. [CrossRef] [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543