Things that "just make sense" make excellent research topics because our intuition turns out quite frequently to be wrong. The New England Journal published a paper by Maitland et al. which had some rather counterintuitive findings. The FEAST trial randomly assigned children with a febrile illness to an infusion of saline, albumin, or no infusion. Statistically, more children died with either saline or albumin as compared to no infusion.
How could this be? After all, the accompanying editorial opens with the point that "Fluid resuscitation is a fundamental intervention in the treatment of critically ill patients". But that statement is just an assertion of the standard of care, and has nothing to do with scientific proof of efficacy. In the US, I think it would be quite hard to conduct a study like this, because the IRBs that review if research is ethical would not be likely to approve a protocol where patients would randomly be assigned to a treatment considered less than the standard of care (ie: no fluid). In Africa however, since access to resources like IV fluids is not guaranteed, it is not necessarily the standard of care, and one could conduct this study.
The study did not collect much information which sheds light on why the findings came out the way they did. Do IV fluids interfere with the body's protective mechanisms for dealing with infection? Do they somehow strengthen the infectious agents against their host? Do they alter acid/base or electrolyte concentrations to the detriment of the patients? These and many other are covered in the NEJM blog. One very important question in my mind is while this study was done in Africa, are the findings limited to that patient care setting? Can we reasonably expect that care in industrialized countries is different enough that regular infusion of IV fluids for infected patients is beneficial and not detrimental? Does this only apply to febrile children? This study has counterintuitive findings and raises more questions than it answers, both of which are hallmarks of good science.