We would like to thank Manfredi  for his interest in our recently published article , in which we reported an independent association between thrombocytopenia and subsequent risk of non-AIDS-defining cancer.
Manfredi  states, correctly, that platelet counts were reported by the study centers and were not centrally reexamined under a microscope to exclude pseudothrombocytopenia. This is a relatively uncommon artifact with an estimated incidence of 0.1%  and thus an unlikely confounder in our study. Furthermore, in order to account for laboratory errors, we repeated analyses requiring a confirmed low platelet count to define thrombocytopenia and the association with cancer remained unaltered. Manfredi  criticizes the lack of data on hepatitis C virus (HCV) load in our report. In analyses removed from the originally submitted manuscript at the request of the editor, we investigated, among 2160 HCV-antibody-positive participants [of whom 406 (18.8%) developed thrombocytopenia], the contribution of liver fibrosis and HCV RNA levels to the development of thrombocytopenia during follow-up. Liver fibrosis, but not HCV RNA levels, was a strong predictor of thrombocytopenia in adjusted analyses (Supplementary Table; http://links.lww.com/QAD/A714). Manfredi  also comments that we did not investigate several specific causes of thrombocytopenia, an important limitation already fully acknowledged in our article, and that our observational study had limited ability to account for ‘multiple and often related confounding factors’. This criticism is indisputable but refers to an inherent weakness of all observational studies.
More seriously, Manfredi  writes that we deliberately set out to identify an significant association between thrombocytopenia and cancer and that this finding was only possible given the large sample size of EuroSIDA , an international HIV cohort that follows participants not only from across Europe but also from Argentine and Israel. Our analyses were based on clear, a-priori hypotheses, based on a clear scientific rationale. Investigations assessing the potential surrogacy of a given parameter, such as platelet counts, for cancer events can only be conclusive with a long follow-up of a large number of participants so that a sufficiently large number of rare clinical events can accrue. The fact that smaller studies with insufficient follow-up and small number of events had failed to find such an association  called for, not against, a subsequent investigation in a larger dataset.
Manfredi  says that he did not find an association between thrombocytopenia and non-AIDS cancer among the 2000 patients followed up in his center. Unfortunately, we cannot identify potential reasons for this discrepancy because Manfredi  refers to unpublished data and articles that did not investigate this. Manfredi  also raises doubts about the clinical significance or usefulness of our findings. We are extremely wary of overinterpreting our data and believe that it would be speculative to discuss, at this point in time, whether monitoring platelet counts would be helpful to inform care in a clinical setting. This was not the primary research question that we sought to address in our article and confirmation of our findings by other research groups is warranted.
To conclude, our recent findings  pose an important research question as to whether the pathophysiological mechanisms underlying thrombocytopenia contribute to cancer development during treated HIV infection. We believe this to be an interesting avenue for investigation into carcinogenesis during HIV infection. Activated platelets release a variety of inflammatory mediators such as chemokines and molecules from the tumor necrosis factor-α ligand superfamily into the circulation [6,7]. The plasma levels of platelet-derived inflammatory mediators have been shown to persist raised or even increase despite suppressive antiretroviral therapy  and may contribute to chronic immune activation . It is thus enticing to hypothesize that, in the setting of treated HIV infection, persistent thrombocytopenia reflecting the removal from circulation of activated platelets  may be an epiphenomenon as a consequence of activated inflammatory and coagulation pathways , which may be causally linked to cancer. We intend to perform studies to investigate this in the near future.
Primary support for EuroSIDA is provided by the European Commission BIOMED 1 (CT94-1637), BIOMED 2 (CT97-2713), the 5th Framework (QLK2-2000-00773), the 6th Framework (LSHP-CT-2006-018632), and the 7th Framework (FP7/2007-2013, EuroCoord no. 260694) programs. Current support also includes unrestricted grants from Bristol Myers Squibb, Janssen R&D, Merck and Co. Inc., Pfizer Inc., and GlaxoSmithKline LLC. The participation of centers from Switzerland was supported by the Swiss National Science Foundation (Grant 108787).
Conflicts of interest
There are no conflicts of interest.
1. Manfredi R. Thrombocytopenia and risk of cancer during treated HIV disease
2. Borges ÁH, Lundgren JD, Ridolfo A, Katlama C, Antunes F, Grzeszczuk A, et al. Thrombocytopenia is associated with an increased risk of cancer during treated HIV disease
3. Bizzaro N. EDTA-dependent pseudothrombocytopenia: a clinical and epidemiological study of 112 cases, with 10-year follow-up
. Am J Hematol
4. Mocroft A, Ledergerber B, Katlama C, Kirk O, Reiss P, d’Arminio Monforte A, et al. Decline in the AIDS and death rates in the EuroSIDA study: an observational study
5. Zetterberg E, Neuhaus J, Baker JV, Somboonwit C, Llibre JM, Palfreeman A, et al. Platelet count kinetics following interruption of antiretroviral treatment
6. Landrø L, Ueland T, Otterdal K, Frøland SS, Aukrust P. Persistently raised plasma levels of platelet-derived inflammatory mediators in HIV-infected patients during highly active antiretroviral therapy
. J Thromb Haemost
7. Damien P, Cognasse F, Lucht F, Suy F, Pozzetto B, Garraud O, et al. Highly active antiretroviral therapy alters inflammation linked to platelet cytokines in HIV-1-infected patients
. J Infect Dis
8. Donhauser N, Pritschet K, Helm M, Harrer T, Schuster P, Ries M, et al. Chronic immune activation in HIV-1 infection contributes to reduced interferon alpha production via enhanced CD40:CD40 ligand interaction
. PLoS One
9. Metcalf-Pate KA, Lyons CE, Dorsey JL, Shirk EN, Queen SE, Adams RJ, et al. Platelet activation and platelet-monocyte aggregate formation contribute to decreased platelet count during acute simian immunodeficiency virus infection in pig-tailed macaques
. J Infect Dis