Alpha-2 adrenoceptor agonists and sepsis: improved survival?

Dr Pandharipande and colleagues should be commended for segregating the effect of dexmedetomidine, an alpha-2 adrenoceptor agonist, in sepsis [1]. However, three questions arise from their study: what are the P-values for the data reported in Table 1? Why was there such a discrepancy between the fentanyl dose given to patients on dexmedetomidine and those on lorazepam (1,114 versus 117 μg/day, P = 0.01) considering the 50 to 80% reduction in the use of opiates commonly observed in the literature when alpha-2 adrenoceptor agonists are administered? And how many days did the patients spend on spontaneous (for example, pressure support) versus controlled/assisted ventilation?

In their study, survival was better (a 70% reduction in risk of dying at 28 days) in patients on dexmedetomidine (n = 31) than in those on lorazepam (n = 32). Improved survival was observed earlier in tetanus patients [2] (rate of death of 50% versus 11% in control (n = 10) versus clonidine-treated (n = 17) patients; P = 0.04; this 1998 reference is not cited in the bibliography). In the study of Dr Pandharipande and colleagues, baseline characteristics were slightly different (Table 1 in [1]): temperature, heart/respiratory rate, incidence of vasopressors (dexmedetomidine, 32%; lorazepam, 56%) and drotrecogin alpha (activated; P = 0.20) were higher and systolic pressure lower in the lorazepam group despite 'similar severity of illness'. Could bias explain partially improved survival? As concluded by the authors [1], a larger trial should demonstrate improved survival (for example, upon septic shock [3]).

Secondly, the dexmedetomidine patients received ten times more fentanyl and had more ventilator-free days. Usually, alpha-2 adrenoceptor agonists reduce the need for opiates by 50 to 80% and preserve spontaneous ventilation. So why this discrepancy?

Thirdly, vasopressor requirements were reduced in the dexmedetomidine group (Table 3 in [1]). A 2003 reference [4] showed previously a reduced vasopressor requirement and was not cited in the bibliography. Could more ventilator-free days lead to less infections [5,6], improved survival, lowered intrathoracic pressure and reduced vasopressor requirements?

Authors' response

Pratik P Pandharipande, Robert D Sanders, Timothy D Girard, Mervyn Maze and E Wesley Ely

Dr Quintin raises interesting questions regarding our analyses of the septic subgroup in the MENDS trial [7], which found improved outcomes and survival in septic patients treated with dexmedetomidine versus lorazepam [1]. In Table 1, we did not report P-values to avoid misleading readers into believing the groups were perfectly balanced. Indeed, we advised caution when interpreting these results since subgroup analyses are prone to type II errors; that is, due to the reduced sample sizes, some imbalances could have occurred that - though not statistically significant - could have been clinically important. We attempted, therefore, to reduce the impact of potential imbalances by adjusting for age, severity of illness and use of drotrecogin alfa.

Fentanyl was used both as an analgesic and supplemental sedative when a deeper level of sedation (than that achieved with the study drug) was ordered by the medical team. Higher doses of fentanyl were noted in the dexmedetomidine group primarily when patients were deeply sedated [7], suggesting the increased fentanyl use reflected a need for additional sedation rather than analgesia. Previous studies reporting opioidsparing effects of dexmedetomidine have examined intraoperative use [8] or short-term use after surgery [9], both of which involve different populations than that studied in MENDS.

We did not evaluate modes of ventilation. We did find an increase in ventilator-free days in septic patients sedated with dexmedetomidine versus lorazepam, but did not find a reduction in secondary infections as seen in SEDCOM [6]. Thus, our results do not support the hypothesis that a reduced ventilator time in the septic dexmedetomidine group resulted in lower secondary infections and thereby improved survival.

Competing interests

PPP, TGD, MME and EWE received research grants and honoraria from Hospira Inc. This is an investigator initiated study and Hospira Inc did not have a role in the generation of the hypothesis, conduct of the trial, data analysis or financing of the manuscript.

PPP and EWE received honoraria from GSK; EWE received honoraria from Aspect Medical and Eli Lily. None of these have any relevance to this manuscript.