The possible effects of iron loss from bloodletting on mortality from pneumonia in the nineteenth century.

OBJECTIVE: To estimate iron losses and disease severity following 19th century bloodletting in patients with pneumonia. STUDY DESIGN AND SETTING: Benefits of bloodletting in pneumonia patients were contested during the 19th century. Although large blood volumes during infection were removed there was no systematic data collection assessing efficacy and knowledge of iron composition of blood was rudimentary. This observational analysis of historical data quantifies iron losses in pneumonia cases in relation to disease severity. RESULTS: Based on one detailed case series average blood volume removed for survivors was 830 mL (range 114-2272 mL), and mean recovery times were shorter in patients bled within 2 days of illness (P < 0.001). Average iron removed was 446 mg with phlebotomy done ≤2 days of illness presentation and 347 mg after >2 days of illness (P = 0.012). Across several European hospitals average case fatality in pneumonia patients receiving phlebotomy was higher than in those treated without phlebotomy (19.9% vs. 12.8%, OR 1.55, 95% CI 1.38-1.74, P < 0.001). CONCLUSION: Variable efficacy for bloodletting could at least in part be explained by altered iron status.

In defense of Pierre Louis who pioneered the epidemiological approach to good medicine.

Pierre Louis, in his 1836 publication, "Researches on the effects of bloodletting in some inflammatory diseases," concluded that there were "narrow limits to the utility" of bloodletting for the treatment of pneumonia. Louis's conclusions were based on an observed excess mortality of patients bled rapidly post disease onset compared with patients bled later on. Doctors had used bloodletting for at least 2,000 years for a wide variety of purposes. Louis's approach was a first-step toward what we would call today an evidence-based medicine. A book denouncing "Bad Medicine" throughout history suggests that Louis inappropriately analyzed his data, overestimated the alleged benefits of venesection on the evolution of pneumonia and therefore contributed to promote this dangerous treatment. I show in this article that these criticisms are inaccurate. In particular, Louis was aware of the potential confounding effect of age and interpreted his data accordingly. Accusing Louis of "bad medicine" misses the historical importance of his struggle to improve clinical medicine using population thinking and group comparison.