Originally published as JCO Early Release 10.1200/JCO.2009.23.9483 on August 31 2009

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Google Scholar Articles by Mehta, J. PubMed PubMed Citation Articles by Mehta, J. Related Articles Related Article Social Bookmarking                            What's this?

Sirolimus-Containing Graft-Versus-Host Disease Prophylaxis in Patients With Lymphoma

The Robert H Lurie Comprehensive Cancer Center, Northwestern University, Chicago, IL

To the Editor:

In a retrospective analysis, Armand et al¹ found high relapse rates and poor overall survival (OS) and progression-free survival (PFS) in a small number (N = 23) of lymphoma patients who underwent reduced-intensity conditioning allogeneic hematopoietic stem-cell transplantation (RIC HSCT). They concluded that lack of sirolimus-containing graft-versus-host disease (GVHD) prophylaxis was responsible for this phenomenon, on the basis of a comparison with 103 patients who had received sirolimus-based GVHD prophylaxis and whose outcome was superior. However, this observation may be flawed as a result of potential analytic errors.

Some of the concerns are relatively minor: Performance status, donor age, CD34+ cell dose, and lactate dehydrogenase levels are significant determinants of outcome after RIC HSCT.^2–5 The influence of these was not analyzed in the study.¹ Armand et al's¹ Figure 3B shows that approximately half of the 18 relapses in the no-sirolimus group occurred within 6 months of HSCT, suggesting a rapid rate of disease progression. That, coupled with the fact that only two of 23 no-sirolimus patients were in complete remission compared with 25 of 103 sirolimus patients (P = .08; Fisher's exact test), suggests that patients in the no-sirolimus arm may have had higher risk disease to begin with.

The data inconsistencies and possible statistical errors are of much greater concern. In Table 1, the P values reflecting categorical comparisons are difficult to understand.¹ One would expect a single P value for each category, not a P value for each line. How were these nonsignificant P values obtained? There also seems to be an error in the number of donors seropositive for cytomegalovirus (CMV) in the no-sirolimus group. Either the absolute number or the percentage figure is incorrect. If the absolute number is correct, a significantly greater proportion of no-sirolimus patients at were risk of CMV infection and possibly poorer outcome.

In Table 2 of Armand et al,¹ some of the parameters have a reference category specified and some do not. Under "Prior therapy," the only two categories specified are "more than 2 lines" and "Prior ASCT," and no reference category has been specified. Obviously, "more than 2 lines" and "Prior ASCT" cannot be covariates in the same category. The hazard ratio (HR) for OS for patients in the "Prior ASCT" category, presumably compared with the unnamed reference category "No prior ASCT," is stated to be 1.9 with a nonsignificant (P = .16). The HR of 0.5 (corresponding to and inverse of 2.0) for OS for "GVHD prophylaxis" is stated to be significant (P = .042), in favor of sirolimus-containing regimens. The ratio of the number of patients in the two "GVHD prophylaxis" categories is approximately 1:4, whereas that in the two "Prior ASCT" categories is exactly 1:1. If the GVHD prophylaxis HR and P values stated in the paper are accurate, the P value for an HR of 1.9 with an even distribution of patients into two groups should be highly significant. The same question arises for PFS, for which the HR for "Prior ASCT" was shown as 1.6, with a P value of .2.

In Table 2, the HR for OS for "Age more than 50 years" is 1.5, whereas it is 0.8 for PFS.¹ OS and PFS are usually closely linked. It is unusual to see the same risk factor influence OS and PFS so disparately, and indeed in opposite directions (on opposite sides of unity, implying an unfavorable effect on OS and a favorable effect on PFS). The same question arises for the donor type, where the HR is 1.6 for OS and 0.9 for PFS.

The last sentence under "Comparison With Nonlymphoma Patients" is, "This held true even if we only considered the 154 patients receiving RIC HSCT (3-year OS, 35% for sirolimus v 34% for no sirolimus, P = .06; and 3-year PFS, 26% for sirolimus v 20% for no sirolimus, P = .3)."¹ Is the P value of .06 accurate with the relatively small number of patients and virtually identical 3-year OS? In Figure 3A, is the P value of .07 accurate considering the curves are so close? Similarly, in Figure 3D, is the P value of .06 accurate considering the curves are largely superimposed?

Given that the main conclusion of the paper hinges on the outcome of 23 patients with a number of different types of lymphomas and the analysis shows evidence of possible errors, a careful review of the data for accuracy and repeat analysis, including other potential factors mentioned earlier, is warranted.

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest.

REFERENCES

1. Armand P, Gannamaneni S, Kim HT, et al: Improved survival in lymphoma patients receiving sirolimus for graft-versus-host disease prophylaxis after allogeneic hematopoietic stem-cell transplantation with reduced-intensity conditioning. J Clin Oncol 26:5767–5774, 2008.[Abstract/Free Full Text]

2. Mehta J, Gordon LI, Tallman MS, et al: Does younger donor age affect the outcome of reduced-intensity allogeneic hematopoietic stem cell transplantation for hematologic malignancies beneficially? Bone Marrow Transplant 38:95–100, 2006.[CrossRef][Medline]

3. Artz AS, Pollyea DA, Kocherginsky M, et al: Performance status and comorbidity predict transplant-related mortality after allogeneic hematopoietic cell transplantation. Biol Blood Marrow Transplant 12:954–964, 2006.[CrossRef][Medline]

4. Anasetti C: What are the most important donor and recipient factors affecting the outcome of related and unrelated allogeneic transplantation? Best Pract Res Clin Haematol 21:691–697, 2008.[Medline]

5. Guilfoyle R, Demers A, Bredeson C, et al: Performance status, but not the hematopoietic cell transplantation comorbidity index (HCT-CI), predicts mortality at a Canadian transplant center. Bone Marrow Transplant 43:133–139, 2009.[CrossRef][Medline]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

Related Article

• Reply to J. Mehta
  Philippe Armand, Haesook T. Kim, Corey S. Cutler, Vincent T. Ho, John Koreth, Edwin P. Alyea, Ann S. LaCasce, Eric D. Jacobsen, David C. Fisher, Jennifer R. Brown, George P. Canellos, Arnold S. Freedman, Robert J. Soiffer, and Joseph H. Antin
  JCO 2009 27: 139-140 [Full Text]

This article has been cited by other articles:

				P. Armand, H. T. Kim, C. S. Cutler, V. T. Ho, J. Koreth, E. P. Alyea, A. S. LaCasce, E. D. Jacobsen, D. C. Fisher, J. R. Brown, et al. Reply to J. Mehta J. Clin. Oncol., October 1, 2009; 27(28): e139 - e140. [Full Text] [PDF]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Google Scholar Articles by Mehta, J. PubMed PubMed Citation Articles by Mehta, J. Related Articles Related Article Social Bookmarking                            What's this?