Why a Doctor Would Be Relieved When a New Study Fails To Reduce Deaths


Sometimes the New England Journal of Medicine reads like a journal of failed bright ideas. I do not remember ever having read of so many failed trials of treatment as I have recently, but perhaps that is a sign of increasing scientific honesty. After all, it is as important to know what does not work as what does, especially when what does not work is very expensive to administer.

Septic shock is a condition of dangerously low blood pressure brought about by serious infection. About 750,000 cases a year are treated in the United States alone, with a death rate above 20 percent, that is to say at least 150,000 people die of it each year. This is a number well worth reducing.

More than a decade ago the results of a trial were published in which it was shown that aggressive treatment according to a pre-arrange protocol could reduce the death rate from septic shock by about a half. In those days (medicine 10 years ago seems that of a bygone era), the death rate in septic shock was much higher than it is today, which may in part explain the success of that trial compared with the failure of a more recent trial published recently in the NEJM. 


Patients who had septic shock from 31 centers (which had to have more than 40,000 emergency room attendances per year to be eligible for entry to the trial) were allocated randomly to one of three groups, with one of two different protocols of treatment, or standard treatment according to the usual practice of the hospital. The protocol group received more “aggressive,” that is to say invasive, treatment, with central venous line measurements from which automatically triggered certain treatments with blood pressure-raising drugs or packed red cell transfusions.

The end point to be compared was death within 60 days. The authors found that there was no statistically significant difference in the death rates of the three groups. If anything, the death rate in the group with the most ‘aggressive’ protocol was the highest, though the difference was small and could have arisen by chance: 21 percent against 18.9 percent for the group that received usual care.


Of course I want death rates to be lowered, but I was at the same time rather relieved by the results. I don’t really want a world in which all our actions are predetermined by written protocol, and in which we are obliged to act like bureaucrats without ever exercising our individual judgment. Had the protocols been shown to reduce mortality, we should have been ethically advised to follow them. However, the victory for the continued exercise of judgment is only temporary: even now it is probable that someone is trying to devise a better protocol. It is probable that one day a protocol will be found that works.

One little statistic in the paper intrigued me. Initially 600 patients were excluded from the trial because they had been given “not for resuscitation” orders. That seemed quite a lot to me, nearly half as many – 1331 – as the patients who were actually included. (Of course, septic shock is much more common among the very old and already infirm than among the young and fit.) The paper gave no details as to how or why the 600 were chosen for no-treatment; they far exceeded in number, however, those who died among the treated, who totaled 259. Were the 600, the vast majority of whom, presumably, died, chosen, that is to say excluded from treatment, by protocol, or was it a matter of judgment for the treating physicians?


images via shutterstock / wandee007 / Lisa S.