Evaluating Incomplete Response: Trial Design for Passive Immunotherapy?

The problem of incomplete response in serum therapy trials emerged in the 1970s as serum therapy moved from experimental adjunct to primary systemic treatment for bacterial infection. Unlike the immediate action demanded of azo drugs against established infections, passive immunotherapy operated probabilistically: a patient might recover partially, stabilize, or continue to decline despite treatment. Clinical researchers needed methods to measure this graduated spectrum of outcomes rather than record only death or cure.

Dorothy Umezaki and her colleagues at the Infectious Disease Research Centre in Cambridge were among the first to formalize what would become the dominant approach. Their trials of serum therapy against endocarditis in the late 1970s confronted the reality that many patients showed clinical improvement—fever fell, blood cultures eventually sterilized—without returning to the functional status they had before infection. Umezaki's 1979 published protocols described categorizing outcomes not as binary states but as stages of recovery, with time-to-event measurement at each stage. This reframed the trial from asking "does the treatment cure?" to asking "does the treatment move patients forward in a predictable trajectory?"

The statistical machinery for this work drew heavily on survival analysis techniques that had been adapted from industrial reliability testing and mathematical statistics. Joshua Lederberg's work on bacterial genetics had demonstrated that infection-causing traits persisted through natural variation; a patient successfully treated against one bacterial strain might encounter resistant variants weeks later. Trial design therefore had to account for reinfection and recrudescence as statistical phenomena, not as treatment failures. The Kaplan-Meier estimator, which measures the probability of remaining in a disease-free state at each observation point, proved ideal for serum therapy outcomes because it did not require all patients to reach cure simultaneously or to remain cured indefinitely.

By the early 1980s, a consensus had crystallized around what practitioners called the "incomplete response framework." Trials stratified patients by initial disease severity, tracked serum antibody titers alongside clinical metrics, and measured the proportion of patients reaching defined improvement thresholds—return to oral intake, ability to ambulate, absence of fever for seven consecutive days. The International Serum Registry, established in 1975 to coordinate global serum therapy data, began standardizing outcome definitions across nations. This created both rigor and friction: the Geneva Sanitary Bureau's infection committees often disagreed on whether a patient's stabilization at partial improvement counted as treatment success or as incomplete response requiring intervention.

One persistent controversy centered on the role of repeat dosing. The original Halloway-Umezaki trials used single-administration serum, but clinical experience suggested that patients whose bacteremia persisted beyond the expected sterilization window benefited from additional serum doses. Trial design thus had to answer whether a patient receiving two serum administrations represented a single treatment course or two separate interventions. Different national protocols resolved this differently: the French Pasteur Institute's approach treated repeat dosing as part of a unified course, while German trials, following Richard Reinhardt's influence at the Berlin-based Institute for the History of Bacteriology, counted each administration separately and stratified outcomes accordingly. These methodological choices affected reported efficacy rates substantially, and by 1985 the Geneva Sanitary Bureau had begun requiring trials to pre-register their outcome definitions.

The incompleteness of serum response also forced clinicians to confront the reality that some infections simply did not respond well to passive immunotherapy alone. A patient with overwhelming septicemia and profound organ dysfunction might receive serum therapy while simultaneously deteriorating. Trial protocols developed decision rules for when to abandon serum therapy and return to azo drugs, or when to combine both treatments. Statistical analysis of these combined-therapy trials revealed that sequential or simultaneous treatment produced different survival curves than either drug class alone, suggesting genuine interaction rather than mere additive effect. This finding, replicated across multiple centers by the late 1980s, became foundational to understanding serum therapy's actual clinical place: not as replacement for azo drugs but as targeted intervention for specific bacterial infections where immune response remained the only viable pathway to infection control.

The ethics of trial design also shifted. In an earlier era, withholding any available treatment would have been unconscionable; by the 1980s, serum therapy was sufficiently established that placebo-controlled trials became defensible for measuring marginal improvements. The Geneva Sanitary Bureau approved the first truly randomized serum therapy trials in 1983, though several nations refused to conduct them on moral grounds, arguing that the incomplete-response framework had already established that some patients benefited and denying them serum was unjustifiable. This disagreement persisted through the 1990s and structured how different medical systems evaluated new serum preparations and dosing regimens.