Abstract
To ascertain which mononuclear cell subset deficiency plays a role in the marrow transplant recipient's susceptibility to infections, mononuclear cell subset counts were prospectively determined in 108 patients on day 80. Infections occurring between day 100 and 365 were recorded by an investigator blinded to the subset counts. In univariate analyses, the counts of the following subsets showed a significant inverse correlation with infection rates: total B cells, IgD+ B cells, IgD− B cells, total CD4 T cells, CD28+ CD4 T cells, CD28− CD4 T cells, CD45RAlow/− CD4 T cells and monocytes. In multivariate analyses, the counts of the following subsets remained significantly inversely correlated with the infection rates: total B cells (P = .0004) and monocytes (P = .009). CD28− CD8 T-cell counts showed no correlation with infection rates. In conclusion, the susceptibility of patients to infections late posttransplant may be due in part to the slow reconstitution of B cells and monocytes.
Introduction
Infections are frequent after marrow transplantation, even after neutrophil engraftment.1-6Ochs et al5 noted that the occurrence of postengraftment infection(s) was the dominant independent factor associated with increased nonrelapse mortality (RR = 5.5,P = .0001).
Quantitative deficiencies of lymphocytes and their subsets have been described in transplant recipients surviving past engraftment.6 For example, low memory B-cell counts are frequently detected in the first year and low CD4 T-cell counts (both naive and memory) are frequently detected in the first 5 years after transplant.7 8 However, which cell subset deficiency plays the most crucial role in the susceptibility of patients to late infections is unknown. Therefore, we evaluated mononuclear cell (MNC) subset counts and determined their correlations with the rate of late infections (occurring between day 100 and day 365).
Study design
MNC subset counts were determined in 108 allogeneic marrow recipients transplanted between May 1996 and August 1997, who were outpatients and had no signs of relapse around day 80. Three of the 108 patients subsequently died before day 100 and were not evaluable for day 100-365 infections. The demographic and clinical information on the evaluable 105 patients is given in Table1. Prophylaxis of late infections included sulfamethoxazole/trimethoprim until day 180. A longer course of sulfamethoxazole/trimethoprim, usually with penicillin, was given to patients in whom clinical extensive chronic graft-vs-host disease (GVHD) developed. Six patients received intravenous immunoglobulin between day 100 and 365.
Blood was drawn on approximately day 80 under an IRB-approved protocol and MNC subsets were enumerated as described.9 Day 100-365 infections were counted by a chart reviewer blinded to the MNC subset counts (G.E.) and associations between the MNC subset counts and the infection rate were statistically tested (Table2, footnotes).
Results and discussion
On day 80, the counts of all MNC subsets studied (Table 2) were significantly lower compared with the MNC subset counts in 103 healthy adult volunteers (P < .001, Mann-Whitney test). Between day 100 and day 365 or the day of relapse or death (whichever occurred first), a total of 168 infections developed (average, 1.6 per patient).
In univariate analyses, the counts of the following MNC subsets inversely correlated with the rate of all infections (P < .05): total B cells, IgD+ B cells, IgD− B cells, total CD4 T cells, CD45RAlow/−CD4 T cells, CD28+ CD4 T cells, CD28− CD4 T cells, CD28+ CD8 T cells, and monocytes (Table 2).
In multivariate analyses, factors other than the MNC cell subsets possibly influencing the rate of infections were considered as confounders (Table 2, footnote). Of all the potential confounders, only the use of corticosteroids in the first 3 months after transplant was associated with significantly increased rate of all day 100-365 infections in univariate analyses (P = .04). However, in the multivariate analyses, only low total B-cell count and low monocyte count remained significantly associated with increased rate of infections (P = .0004 for B cells, P = .009 for monocytes).
Both the univariate and the multivariate analyses were also performed separately for severe infections (defined as infections requiring hospitalization), viral infections, bacterial infections, and fungal infections. In the univariate analyses, the rates of severe, viral, bacterial, and fungal infections tended to inversely correlate with the same MNC subset counts, as the rates of all infections (Table 2). In the multivariate analyses, the low IgD− B-cell count remained significantly associated with increased rates of severe infections (P = .003) and bacterial infections (P = .03), the low monocyte count remained significantly associated with increased rate of viral infections (P = .004), and the low total B-cell count remained significantly associated with increased rate of fungal infections (P = .0003).
We also evaluated whether a day 80 MNC subset count was associated with relapse rate or nonrelapse mortality before day 365, using a Mann-Whitney test. No MNC subset count was significantly different in the patients who relapsed compared with those who did not. The total B-cell count was lower in the patients who died without relapse (median, 0.9 × 106/L) than in the patients who did not (median, 2.5 × 106/L), (2-sided P = .092, 1-sided P = .046); whereas the other MNC subset counts were not different.
These data are insufficient to resolve the question of whether the association between B cell and monocyte counts and infections reflects the importance of B cells and monocytes for decreasing susceptibility to infections or that the B cell and monocyte counts may be merely surrogate markers for the quality of the marrow graft. We attempted to determine this by evaluating absolute neutrophil count as a measure of graft function. Day 80 neutrophil counts showed no correlation with infections (Spearman rank correlation coefficientr = −0.05) and no correlation with B-cell counts. In contrast, day 80 IgG level showed a trend toward inverse correlation with infections (Spearman rank correlation coefficientr = −0.40), suggesting that the B cells themselves may be important. As the IgG level was available for only 34 patients who did not receive intravenous immunoglobulin between day 0 and day 80, this was not statistically significant. Other data also support B-cell function as important. Riches et al10 described 3.5 times greater incidence of late infections in patients with low compared with those with normal serum IgG2 and IgG4 levels. Among marrow transplant recipients participating in our randomized trial of placebo versus intravenous immunoglobulin that did not contain IgA,11 there was a significant inverse correlation between day 100 serum IgA level and the rate of late infections (P = .01) (unpublished data). Sheridan et al12 described significant association between low serum IgG2 and IgG4 levels and the incidence of late pneumococcal infections. Thus, given the lack of an association between day 80 neutrophil counts and late infections and given the probable association between day 80 immunoglobulin levels and late infections, we favor the explanation that the slow reconstitution of B cells themselves plays an important role in susceptiblity to infections. Interestingly, low monocyte counts were also associated with infections. However, monocyte deficiency was not as dramatic as B-cell deficiency (median monocyte count was 291 vs 391 × 106/L in patients and normals, respectively; whereas median B-cell count was 2 vs 253 × 106/L). It is possible that the association of monocyte counts with infections could be due to the importance of monocyte/macrophage-mediated phagocytosis, antigen presentation, or secretion of monokines such as interleukin-1.
The primary goal of this study was to identify an MNC subset associated with infections. However, the data also provide the opportunity to ask which MNC subsets are not associated with infections. Of the subsets listed in Table 2, the count of the CD28− CD8 T cells was the least associated with the rate of all infections (Spearman rank correlation coefficient r = −0.01, P = .84). This is not surprising, as CD28− CD8 T cells in vitro do not readily proliferate on CD3 cross-linking and are considered anergic.13 14 Thus, the lack of association between CD28− CD8 T-cell counts and infections suggests that the CD28− CD8 T cells are also anergic in vivo.
We conclude that the reconstitution of B cells and monocytes may play an important role in marrow transplant recipient defense against microorganisms. This may be tested by adoptively transferring donor B cells and/or monocytes. Alternatively, the use of donor peripheral blood stem cells instead of marrow may provide a larger B-cell/monocyte inoculum, or these cells may be augmented by the use of growth factors.
Acknowledgments
We are indebted to the hard work of the staff of the Fred Hutchinson Cancer Research Center Long-Term Follow-Up Department and the staff of the Fred Hutchinson Cancer Research Center Outpatient Department.
Supported by National Institutes of Health grants Nos. CA68496 and AI46108.
The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 U.S.C. section 1734.
References
Author notes
Jan Storek, FHCRC, D1-100, 1100 Fairview Ave N, Seattle, WA 98109-1024; e-mail: jstorek@fhcrc.org.
