Early versus delayed administration of norepinephrine in patients with septic shock

Septic shock is the most challenging problem in critical care medicine and has a very high mortality [1]-[3]. Owing to the complex pathophysiology, the outcomes for septic shock patients remain disappointing. Rivers et al. demonstrated that early goal-directed therapy provided significant benefits with respect to outcomes for septic shock patients [4]. However, massive fluid resuscitation may increase extravascular edema, aggravate pulmonary dysfunction and compromise tissue oxygenation [5],[6]. Therefore, investigating the rational use of vasopressors in septic shock is very important. Thus far, most studies have focused on the rational use of different types of vasopressors [7]-[9], and the third edition of the guidelines for management of severe sepsis and septic shock also concentrates on the choice of vasopressors [10]. However, it is the timing of vasopressor therapy, rather than the specific agent, that appears to be crucial [11]. The current guidelines recommend that vasopressors (norepinephrine as the first choice) be administered for hypotension refractory to initial fluid resuscitation and to maintain a mean arterial pressure (MAP) ≥65 mm Hg, and this should be completed within six hours [10]. According to the guidelines, physicians currently prefer fluid resuscitation without vasopressors until a lack of hypotension correction is confirmed. However, this may result in prolonged hypotension, and valuable time may have passed. For example, some vital organs could be damaged irreversibly because of low perfusion. In addition, some researchers have argued that early initiation of norepinephrine administration, especially when hypovolemia has not been resolved, can adversely affect vital organ microcirculatory blood flow and perfusion [12],[13]. Hence, this retrospective study examined the relationship between delay in initial norepinephrine administration and hospital mortality and investigated the effects of early norepinephrine administration on septic shock.

This retrospective cohort study was conducted in two general surgical intensive care units (ICUs) of a tertiary care hospital, Jinling Hospital in Nanjing, China, which mainly admit patients with severe complications after surgery and severe trauma. In these ICUs, vasopressors were administered according to the Surviving Sepsis Campaign (SSC) guidelines [14]. First, initial fluids resuscitation termed early goal-directed therapy and early antimicrobial treatment were administered as soon as sepsis-induced tissue hypoperfusion was recognized [4]. Then, when the physicians judged that hypotension did not respond to initial fluids resuscitation, vasopressor therapy was initiated with the aim of maintaining MAP at least 65 mm Hg and norepinephrine was the first-choice vasopressor.

The study was approved by the Institutional Review Board of Jinling Hospital. Because all data were anonymous and collected retrospectively, the Institutional Review Board of Jinling Hospital waived the need for informed consent.

These analyses included 213 patients. All patients received norepinephrine as the initial vasopressor agent. The overall 28-day mortality was 37.6%.

Initial serum lactate level was 4.7 ± 1.5 mmol/L and the APACHE II score was 28.4 ± 4.2. Infections were documented in 191 cases (89.7%); the remaining cases represented suspected infections without a positive culture or definitive surgical, radiologic, biopsy, or autopsy evidence of infection. Positive cultures were identified in 175 cases (82.2%). A plausible microbial pathogen was isolated from the blood in 132 cases (62.0%).

There were no significant differences between survivors and non-survivors with respect to age, gender or the frequencies of documented infections and positive cultures or source of infection. Initial serum lactate levels and APACHE II scores were significantly higher in the non-survivors than in the survivors (Table 1). Survivors received more intravenous fluids within 6 hours (3.4 ± 0.9 L versus 3.0 ± 0.9 L; P =0.003) but less intravenous fluids within 24 hours (6.5 ± 0.8 L versus 6.9 ± 0.5 L; P <0.001) than the non-survivors. They also received less norepinephrine within 24 hours (29.4 ± 9.7 mg versus 34.8 ± 9.6 mg; P <0.001). Survivors also had a shorter time to initial norepinephrine administration (2.7 ± 2.1 hours versus 3.8 ± 2.9 hours; P =0.002), had a shorter time to initial antimicrobial treatment (1.4 ± 1.2 hours versus 2.2 ± 1.8 hours; P =0.001) and received effective antimicrobial therapy more frequently (72.9% versus 56.3%; P =0.012) (Table 2). The survivors were treated with norepinephrine for a shorter duration (2.4 ± 0.6 days versus 3.4 ± 0.9 days; P <0.001; Table 2). The frequency of corticosteroid treatment was similar in the survivors and the non-survivors. Overall, the average time to initial norepinephrine administration was 3.1 ± 2.5 hours. A total of 18.8% of all patients received norepinephrine administration within 1 hour after the onset of septic shock, 40.4% within 2 hours and 9.9% of patients ≥6 hours after the first occurrence of septic shock (Figure 2). The relationship between hospital mortality and the time to initial norepinephrine administration is shown in Figure 3. Mortality was 27.5% if norepinephrine administration was initiated <1 hour after the onset of septic shock, 30.4% if initiated from 1 to 2 hours after the onset and 65.2% if initiated ≥6 hours after the onset. Twenty-eight-day mortality was significantly higher in patients who received norepinephrine ≥2 hours after septic shock onset and increased when compared to those who received it <2 hours after onset (OR for death =1.86; 1.04 to 3.34; P =0.035; Figure 4). The OR of death was 2.16 (1.23 to 3.81, P =0.007) when norepinephrine was administered after three hours and 3.61 (1.45 to 8.95, P =0.004) when given after six hours (Figure 4).

The 28-day mortality was 29.1% in the Early-NE group (within two hours) and 43.3% in the Late-NE group (after two hours). There were no significant differences with respect to age, gender, weight, APACHE II score, documented infections, positive cultures and primary infection sites between the Early-NE and the Late-NE group (Table 3). Serum lactate levels at the onset of septic shock were even significantly higher in the Early-NE group than in the Late-NE group (Table 3).

Patients in the Early-NE group received less intravenous fluid within 24 hours (6.2 ± 0.6 L versus 6.9 ± 0.7 L; P <0.001) and less norepinephrine within 24 hours (29.4 ± 9.7 mg versus 32.8 ± 10.0 mg; P =0.013) (Table 4). There was no significant difference between the two groups in the onset of intravenous fluid therapy, the time to initial antimicrobial treatment, the frequency of effective antimicrobial therapy and the frequency of corticosteroid treatment (Table 4). The duration of hypotension was shorter in the Early-NE group (4.6 ± 1.2 hours versus 6.1 ± 1.0 hours; P <0.001), and the duration of norepinephrine was also statistically shorter in these patients (2.6 ± 0.6 days versus 2.9 ± 1.0 days; P =0.001). There was no significant difference in the duration of ICU stay. Mean arterial pressure was higher at one, two, four and six hours after the onset of septic shock and serum lactate levels at two, four, six and eight hours after the onset of septic shock were significantly lower in the Early-NE group than in the Late-NE group (P <0.05, Figure 5). Multivariate logistic regression analysis identified the time to initial norepinephrine administration, the time to initial antimicrobial treatment, serum lactate level at the onset of septic shock, APACHE II score, intravenous fluid therapy within six hours and effective antimicrobial therapy as independent determinants of 28-day mortality (Table 5).

The present data show that the time from the onset of septic shock to initial norepinephrine administration is an important determinant of survival. The risk of death was significantly increased with longer delay in norepinephrine initiation after the onset of septic shock. For every one-hour delay in norepinephrine initiation within six hours after the onset of septic shock, the mortality increased by 5.3%. This could not be attributed to the more severe septic shock in the delayed norepinephrine initiation group. Initial blood pressure levels were similar. On the contrary, patients in the Early-NE group had higher serum lactate levels than those in the Late-NE group.

Early norepinephrine administration reduced the duration and the total dose of norepinephrine treatment and shortened the duration of shock with less intravenous fluid requirements (a 10% decrease in the first 24 hours for treatment). These data suggest that in the case of severe septic shock, early norepinephrine administration should not be delayed. According to the SSC guidelines [10],[14], fluids and antimicrobials should be administered as early as possible in patients with septic shock, but the optimal timing of vasopressor administration is less clear. The patients with septic shock receiving vasopressors early or late depend on the patient’s response to initial fluid resuscitation and the judgment of the physicians. It remains difficult to determine if late administration of norepinephrine reflects a poor evaluation of the disease severity of the patient and if the higher mortality is related to a globally less appropriate management of the patients.

Our results also indicate that although the time to initial antimicrobial treatment was not significantly different between the Early-NE group and the Late-NE group, it was significantly longer in the non-survivor group than the survivor group and it is a critical determinant of outcome. In the multivariate logistic regression analyses, time to initial antimicrobial treatment was a significant independent risk factor for 28-day mortality, a result consistent with prior observations [16],[17]. Kumar et al. reported a linear increase in the risk of mortality for each hour delay in antibiotic therapy [16]. The percentage of patients with positive blood cultures was 62% in our study compared to around 40% in other studies [16]. Some factors may have influenced these results. First, our ICUs were general surgical ICUs and most patients admitted to them had severe intra-abdominal infections or intestinal fistulae, whereas most of the other studies were conducted in medical ICUs and mixed ICUs. The bias between the different patient characteristics might be one reason for the different results. Second, most patients in our study underwent multiple surgical interventions, including surgical puncture, which would increase the opportunity for pathogens to enter the blood. Other factors could have influenced the results of positive blood cultures as well (for example, nosocomial infection).

Additionally, our data show that the time to initial norepinephrine administration, the time to initial antimicrobial treatment, serum lactate level at the onset of septic shock, APACHE II score, intravenous fluids within six hours and effective antimicrobial therapy were independently associated with hospital mortality.

The results of this study are consistent with those reported in some previous studies [18],[19]. In a retrospective study, Morimatsu et al. reported that survival of septic shock patients treated with early and exclusive norepinephrine administration (median time 1.3 hours) compared favorably with that predicted by severity scores [19]. In rats with septic shock, Sennoun et al. compared the effects of early versus delayed norepinephrine administration, and showed that the mesenteric/aortic blood flow ratio was higher and blood lactate concentrations lower in an early compared to a late norepinephrine group [18]. Additionally, early norepinephrine administration spared approximately 30% of the fluid volume required in the late or non-norepinephrine groups [18]. In another retrospective study, Subramanian et al. found that liberal vasopressor use was not associated with less progression to organ failure in septic shock [20]. However, the volume of fluids received within six hours was significantly greater in the liberal than in the conservative group and they used different types of vasopressors, including dopamine, phenylephrine, norepinephrine and vasopressin. Their study was also conducted in a medical ICU, so that differences in patient characteristics may have influenced the results as well.

Septic shock is characterized by hypotension, related in part to absolute and relative hypovolemia. The former is the result of vascular leakage caused by endothelial injury and the latter is the result of systemic vasodilation [21]. In addition, sepsis can result in down-regulation of norepinephrine receptors [22]. Increased MAP with norepinephrine may improve organ perfusion and this may result in lower serum lactate levels. Therefore, expedited norepinephrine administration can be a rational approach to rapidly restore organ perfusion. One may, therefore, be recommended to initiate norepinephrine administration simultaneously with fluid resuscitation at the onset of septic shock. Obviously, this strategy should not prevent adequate fluid resuscitation, as one must pay attention to a so-called `vasoconstrictor-masked hypovolemia’ [23].

Our study has several important limitations. First, the retrospective trial design limits our ability to identify precisely the causes for delay in initial vasopressor administration and to determine a causal relationship between the delay in norepinephrine administration and the septic-shock mortality. In addition, the sample size was relatively small. Therefore, large multicenter prospective cohort studies are needed to validate whether early norepinephrine administration is associated with decreased mortality in patients with septic shock.

This study indicates that early administration of norepinephrine for septic shock is associated with improved survival. Mortality increases when initial norepinephrine administration is delayed. Early norepinephrine initiation can increase MAP, shorten the duration of hypotension and, thereby, may improve vital organ perfusion and decrease serum lactate levels. These data suggest that more prompt and aggressive norepinephrine administration should be considered as part of initial resuscitation therapy for septic shock.

XB created the study protocol, performed the statistical analyses and wrote the first manuscript draft. NL conceived the study and critically revised the manuscript. WY assisted with the study design and performed data collection. WJ assisted with data collection and manuscript editing. ZL confirmed the data and assisted with the statistical analyses. ST assisted with study coordination and helped draft the manuscript. KD assisted with manuscript revision and data confirmation. YD contributed to data interpretation and manuscript revision. LX assisted with study design and data analyses. All authors read and approved the final manuscript.