Bleeding in acquired hemophilia: have we figured it out?

In this issue of Blood, Holstein et al show the extent to which the risk of bleeding after a diagnosis of acquired hemophilia A (AHA) remains significant until near-normal factor VIII (FVIII) level is attained.¹ 

All consultant hematologists, whether from the benign or malignant side of our specialty, have a duty to recognize the cardinal signs of AHA, a rare but severe autoimmune condition where diagnosis is frequently and dangerously delayed.²  Cardinal signs are unexplained excessive bleeding, usually in an elderly person who has an isolated newly prolonged activated partial thromboplastin time. Simple diagnostic algorithms have been published,³  but the first crucial step in making the diagnosis is clinical suspicion. The type of bleeding can be very suggestive (see figure). Plasma FVIII level should be measured without delay. Few patients with AHA now die as a result of uncontrolled bleeding, but delayed diagnosis and treatment continue to lead to debilitating morbidity in these often-frail patients. However, long-term prognosis is good for patients who survive the initial acute episode.⁴ 

The prospective GTH-AH 01/2010 study has already greatly advanced the field with landmark reports on immunosuppression⁵  and serologic predictors of remission in AHA.⁶  In the accompanying report, our German and Austrian colleagues look more closely at bleeding events and response to hemostatic therapy in their cohort of 102 patients. Patients were recruited early during their treatment, with day 1 defined as the first day of immunosuppressive treatment (IST). They were monitored for all clinically relevant instances of bleeding, with the analysis focusing on the first 12 weeks of observation. A total of 148 bleeds in 80 patients were documented at presentation, but more importantly, 141 new bleeds occurred in 59% of the patients during the subsequent observation period at a mean rate of 0.27 bleeds per patient-week. Severe bleeds before day 1 or at presentation lasted a median of 9 days compared with 2 days for those occurring after day 1, emphasizing again the importance of early recognition and treatment.

Hemostatic treatments (mostly bypassing agents) were rated as effective in 96% of cases, confirming what other series and registries have previously reported. More interesting and unique to this study is the multivariate analysis of clinical features, including FVIII levels, predicting the risk of new bleeding after day 1. The pathophysiology of bleeding in AHA is puzzling in more ways than one. In contrast to congenital hemophilia, the measured plasma level of FVIII at diagnosis is not predictive of the risk or severity of bleeding in AHA.^2,7  Spontaneous, catastrophic bleeding may occur in patients with FVIII levels that would be considered of mild severity in congenital hemophilia. This phenomenon is generally attributed to the weak interaction of the inhibitory antibodies with FVIII in vitro (so called second-order kinetics) seen in AHA.³  Presumably, this interaction is different and more powerful in vivo or in specific vascular beds, but to my knowledge, this has not been clearly elucidated. In this publication, the authors confirm that FVIII level and inhibitor titer at baseline do not predict the cumulative incidence of new bleeds after day 1. However, they do demonstrate that subsequent weekly FVIII levels are highly statistically associated with new bleeding during the 12 weeks after initiation of IST. It is important to point out that the differences in risk observed between FVIII levels of <1%, 1% to <5%, and 5% to 20% are minimal. Indeed, one starts feeling somewhat safer at levels >20%, but the authors justifiably insist that only achieving an FVIII level ≥50% abolishes the risk of bleeding. Again, this is consistent with the poorly understood physiopathology mentioned. Another intriguing finding is the marked decrease in the risk of new bleeding after 4 to 6 weeks of IST, irrespective of the residual plasma FVIII level. Might the antiinflammatory effects of corticosteroids mitigate bleeding risk or lead to other compensatory procoagulant effects?⁸ 

Although important consensus has emerged in the management of active bleeding in patients with AHA,^3,9  the prophylactic use of bypassing agents in AHA patients is controversial, given the attendant thrombotic risk of these agents and the high prevalence of cardiovascular comorbidities in these patients. Promptly instituting IST to hasten inhibitor eradication has long been viewed as the best prophylaxis. The authors rightly point out that intensive IST presents its own risks and may also result in fatal complications. Defining the appropriate role of prophylactic hemostatic treatment is becoming even more important with the availability of new and emerging non–coagulation factor–based substitution therapies such as emicizumab.¹⁰  These new agents are therapeutically active in patients with congenital hemophilia and inhibitors, but their risk-benefit profile has not been defined in patients with AHA. These new agents are likely to soon be the focus of exciting clinical trials in AHA. Although we have not yet completely determined why bleeding is more unpredictable in AHA, the findings of Holstein et al contribute significantly to quantifying the risk and lay the groundwork for future prophylaxis trials in AHA.