Three trials back to back in the PNEJM (that’s the prestigious New England Journal of Medicine for any new readers) in adults with septic shock are disappointing:
Early Goal Directed Therapy has become a dogma in recent years; it even appeared on an episode of ER, as I recall. It promotes a protocol of early insertion of central catheters, fluid administration to certain hemodynamic measurements, and monitoring of central venous saturation. The original trial was a single center RCT of moderate size (260 adults with severe sepsis, septic shock, or sepsis syndrome) and had significantly more survivors in the intervention group than the usual treatment controls. I think the study was well-done, but includes individual items that could be, and were, challenged. In total there were about 20 more deaths in the control group than the intervention. Considering that this modest size single center trial, with a significant result, but a relatively small numerical difference in survival led to changes in approach that have affected many 10’s, maybe 100’s of thousands of patients, makes me pause. Smaller trials often tend to have results that are more impressive than larger ones, and the control group mortality in Rivers’ trial was very high, at 46.5%.
The new trial in the PNEJM, is a larger multicenter comparison of early goal directed therapy (a very similar protocol to Rivers et al) to 2 other approaches, one was another protocol which did not mandate a central catheter, and even if you needed a central venous catheter for venous access you weren’t supposed to measure CVP or central venous oxygen saturation, transfusions were given at much lower hemoglobin levels, management was based on clinical evaluation of the patient, including the ‘shock index’ (heart rate divided by systolic BP), and also the clinical signs of poor perfusion. The third arm was usual care, which has of course changed over the years since the original Rivers study. 1341 adults were randomized, they all were in shock, and the study found very little difference in outcomes. The primary outcome, which was whether you survived for 60 days, was not different. There were some minor differences between the groups, but nothing earth shattering. It seems that careful clinical evaluation and judicious standard care is as good as anything else. Of note the mortalities in the 3 groups were around 20%, much lower than the control group in the earlier study which was 44%.
The second study was a comparison of different blood pressure targets. In a multi-center trial from France of adults in septic shock, the 776 subjects were randomized to a target mean BP of 65 to 70mmHg, as recommended by the surviving sepsis campaign, or 80 to 85 mmHg. BP was raised for the most part using norepinephrine, if the patients were below target after fluid resuscitation. The high target group had more norepinephrine used, and some more atrial fibrillation, but no differences in survival to 28 days were seen, which was about 65% in each group. They also had less renal impairment in the high BP group. These patients seem to have been sicker at baseline than the first study, with shock refractory to fluid boluses and already receiving norepinephrine at 0.1 microg/kg/min or more when randomized. Observational data had previously suggested that you did better if the blood pressure was kept higher, which shows the importance, once again, of doing these sorts of trials.
The third trial was an Italian multi-center trial of adults with severe sepsis or septic shock, 1818 patients were randomized to either get crystalloid solution, or 20% albumin and crystalloids, with a goal of achieving a serum albumin of 30 g per liter. 28 day mortality was just over 30% in both groups. The only effect of giving albumin was that the serum albumin was higher!
As I said above, disappointing, as we don’t find clear advantages of one intervention (or bundle) over the other. The really important message from these trials is that we need to keep doing large pragmatic trials in septic shock, they are feasible, and they point out that septic shock is common, still has a high mortality, and we don’t have a lot of evidence based methods to treat it.
There is a message here for neonatology: trials in critically ill unstable patients are feasible, and are absolutely essential if we are to move forward.