Although observational studies and case series support the use of either conventional PP or DFPP as the treatment of severe and refractory TEN, no attempts have been made to compare the efficacy of DFPP with that of conventional PP. Based on our results presented here, conventional PP appears to be more efficacious in TEN than DFPP. Treatment with conventional PP is beneficial during the first 2 weeks, but the benefit is likely to be greatest when treatment is given early. Thus, conventional PP can replace DFPP as the preferred treatment for severe TEN. However, concern has also been expressed for many years that FFP used as the replacement solution in conventional PP is a blood product and poses a risk of contamination with infectious organisms despite testing and antiseptic measures to kill or filter them. A catheter-related line infection is a more common complication than is generally realized. However, the risk of a catheter-related line infection in conventional PP is comparable to that in DFPP. Indeed, case 2 developed sepsis despite its remarkable efficacy, after conventional PP. This type of infection was not observed in other cases. The major disadvantage, and the primary reason why DFPP has not been employed in a widespread manner in TEN, are that many pro-inflammatory cytokines responsible for the development of TEN cannot be sufficiently depleted by DFPP because their molecular weights are smaller than the pore size of membrane filters used for DFPP; this is in sharp contrast to pemphigus vulgaris, in which the removal of pemphigus antibodies by DFPP is a reasonable therapeutic approach.19,20 Therefore, pro-inflammatory cytokines sufficiently removed by DFPP include monocyte colony-stimulating factor (M-CSF) but not IL-1, IL-6 and TNF-α. In support of this possibility, Kodama et al.21 reported that serum levels of IFN-γ, TNF-α, IL-1β, IL-6, and GM-CSF did not decrease after DFPP was started while G-CSF and M-CSF gradually decreased. Indeed, after the DFPP treatment in case 1, no decrease in any cytokine level was discovered, while slight clinical improvement were found. On the contrary, the highest levels of IL-8 and TNF-α occurred immediately after the DFPP, suggesting the rebound production or release of these cytokines. In view of the fact that either IL-8 or TNF-α can be secreted by neutrophils that accumulate in the circulation after DFPP, the unexpected increase in these cytokine levels might reflect ongoing in vivo activation of neutrophils during the process of DFPP. Interestingly, similar, although less remarkable, increase in leukocytes have been reported to occur immediately after blood transfusion but not after plasma infusion.22
Because it has been shown that direct contact of blood mononuclear cells with the filter membrane and adherence of platelet to the membrane leads to not only cytokine production but also complement activation,23 the unexpected increase in white blood cell number after DFPP would result from serum complement activation. Given the ability of neutrophils to respond to these cytokines, the unexpected increase in these cytokine levels after the DFPP appears to result from an in vivo positive feedback mediated by the rebound synthesis of these cytokines by activated neutrophils. Alternatively, macrophages would be the major source of these cytokines. However, because these cytokine levels eventually decreased 2 days after the DFPP in patients with bullous pemphigoid (unpubl. data, T. Shiohara, 2010), the rebound of cytokine production would be short-lived. The rebound in cytokine levels after DFPP, but not conventional PP, indicates that the use of FFP during the process of conventional PP would serve to prevent neutrophil or macrophages from secreting more cytokines. Because, in DFPP, original blood cell components initially separated are reconstituted with the processed plasma and replacement solution and returned to the circulation, this process may stimulate neutrophils and macrophages to additionally release these cytokines. Thus, the rebound of cytokine production after DFPP is likely to be the reason why its onset of action is usually delayed after the DFPP and why DFPP is less efficacious in TEN than conventional PP. These results explain why the therapeutic efficacy of DFPP is not high as compared with conventional PP, although theoretically it can selectively reduce toxic substances from the circulation.