7.1 Introduction

Recommendations for care in patients with sepsis or septic shock are largely based on evidence originating from resource-rich settings [1]. It is increasingly appreciated that these recommendations cannot be directly generalized to resource-limited settings for several reasons, including restrictions in human and material resources but also concerns regarding costs and safety [2, 3]. It is even possible that the efficacy and effectiveness of certain strategies differ between resource-rich and resource-limited settings. Indeed, efficacy and effectiveness could depend on the type of sepsis, and it is well known that non-bacterial sepsis is much more common in resource-limited than in resource-rich settings [3].

In this chapter, we aim to answer five practical questions regarding hemodynamic assessment and support in sepsis and septic shock in resource-limited settings. As recognition of hypoperfusion and return to normal perfusion, and detection of fluid responsiveness, could avoid under- and over-resuscitation as well as under- and overuse of vasoactive agents, (1) there is need for affordable bedside tools for tissue perfusion monitoring and (2) a better understanding of practicalities of passive leg raise tests in these settings; as costs and availability of, but also indications for, intravenous fluids could be different in resource-limited settings, (3) advises regarding the preferable type of intravenous fluid to be used during fluid resuscitation, as well as (4) amounts and timing of intravenous fluids for sepsis shock in resource-limited ICUs, are essential. Finally, seen the limited availability of vasopressors and inotropes, and the risks associated with their use, (5) recommendations on their indications, titrations, and ways of administration in settings with limited resources are highly necessary. Recommendations and suggestions are summarized in Table 7.1.

Table 7.1 Recommendations for fluid management and hemodynamic support in patients with sepsis or septic shock in resource-limited settings (with grading)
Full size table

7.2 Simple Bedside Tools to Assess Tissue Perfusion

Timely detection of tissue hypoperfusion is one crucial aspect of hemodynamic assessment in patients with sepsis or septic shock. Several studies showed that capillary refill times >5 s following initial hemodynamic optimization are associated with worsening organ failures [4,5,6]. Normalization of capillary refill time was prognostic of survival in septic shock patients [7]. During early septic shock, capillary refill time was found to be a good predictor of short-term mortality [8] and related to perfusion of the liver, spleen, kidneys, and intestines in adults [9]. There was noticeable variation, though, in how capillary refill times were checked, at least in investigations involving children (Table 7.2), and several factors may affect the accuracy capillary refill time, like the ambient temperature and light, the site of measurement, and the amount of pressure applied to the capillary bed [10]. There was debate on whether capillary refill time is subject to interobserver variability [10, 11]. One study in India suggests capillary refill time to be insensitive to detect tissue hypoperfusion in patients with malaria [12].

Table 7.2 Different methods of measuring and interpreting capillary refill time in children
Full size table

Mottling, patchy skin discolorations due to heterogenic small-vessel vasoconstriction that usually start around the knees and elbows in patients with shock could also reflect abnormal skin perfusion. A simply to apply at the bedside score, using a scale from 0 (“no mottling”) to 5 (“grave mottling”) (Table 7.3 and Fig. 7.1) related well to plasma lactate levels, urine output, degree of organ dysfunctions, and even mortality in patients with septic shock [13]. Patients whose mottling score decreased during the resuscitation period had a better prognosis [13]. The prognostic value of this score was confirmed in other cohorts of critically ill patients [14, 15]. The mottling score had a good reproducibility and did not suffer from interobserver variability [13].

Table 7.3 Skin mottling score after initial fluid resuscitation
Full size table
Fig. 7.1
figure 1

Skin mottling score; from Ait-Oufella et al. [13]

Full size image

Skin temperature gradients, the difference between two different measurement points, such as between the forearm and fingertip, or the central and toe, could be useful in detecting changes in skin perfusions in sepsis and septic shock [16, 17]. The advantage of using skin temperature gradients between the forearm and fingertip, instead of a single skin temperature, is that both spots are similarly affected by ambient temperature. The normal skin temperature gradient between forearm and fingertip is 0 °C. Skin temperature gradients between the forearm and fingertip of >4 °C were associated with severe vasoconstriction. Increased skin temperature gradient was related to the outcome of sepsis [18].

We suggest using capillary refill time, skin mottling scores, and, if affordable, skin temperature gradients to assess adequacy of tissue perfusion in pediatric and adult sepsis and septic shock, either alone or in combination (UG). It remains uncertain whether these tools are effective in malaria. These tools are noninvasive and safe and come at no additional or low costs, though costs of temperature probes could still be too high for certain resource-limited settings. This recommendation remains weak, mainly because of the absence of evidence that these bedside tools can adequately guide important decisions in hemodynamic support.

7.3 The Passive Leg Raise Test and Other Simple Tools to Replace Direct Measurements of Cardiac Output

If it is decided that a patient is hypovolemic, it should also be determined whether that patients is “fluid responsive.” The method for performing passive leg raise test is important because it fundamentally affects its hemodynamic effects and reliability [19]. The test needs to be executed so that it does not result in pain and anxiety as this may influence the results. Furthermore, a proper passive leg raise test consists lifting the bed at the foot end and not lifting the legs (Fig. 7.2). The latter could be a challenge in resource-limited settings where beds are usually not easy adjustable. While it is best to use a direct measure of cardiac output or stroke volume, this is frequently impossible in settings where resources are low. A less accurate but still acceptable approach is to detect changes in pulse pressure. The test then starts with an initial (noninvasive) blood pressure measurement—after 60–90 s of passively raising the legs, the blood pressure measurement is repeated—and a change in the difference between systolic and diastolic pressure >15% could indicate that the patient is “fluid responsive” [20].

Fig. 7.2
figure 2

For maximal reliability, a passive leg raise test should be performed following some rules. One possible variation of test starts from the semi-recumbent position. The second step comprises to go down the trunk and raise legs maintaining the angle between them using the automatic motion of the bed for avoiding artifacts. Finally the third step goes back to the semi-recumbent position to ensure that the subject recovers the previous hemodynamic parameters

Full size image

It remains uncertain whether the passive leg raise test has comparable predictive values in various types of sepsis and septic shock, e.g., in severe malaria or severe dengue, as literature is lacking. This could actually be seen as one major objection against widespread use of passive leg raise tests in resource-limited settings. This is also true for young children. So far, only one preliminary study suggests that a passive leg raise test is helpful in predicting fluid responsiveness in children but not in those under 5 years of age [21].

We suggest using the passive leg raise test to guide fluid resuscitation in sepsis or septic shock in resource-limited settings (2A). It is uncertain whether the passive leg raise test has predictive values in all types of sepsis and septic shock, like in severe malaria or severe dengue. We suggest using the passive leg raise test and in children but only in those above the age of 5 (2C). We recommend direct measurement of changes in cardiac output when performing a passive leg raise test (1C) and suggest using changes in pulse pressure if the former is not possible (2C).

7.4 Fluid Strategies

There is a large body of literature from resource-rich settings on the choice of fluids in severe sepsis and septic shock, with a strong focus on sepsis caused by bacterial pathogens. The theoretical benefits of colloid solutions over crystalloids, with better retention in the intravascular compartment, have not translated to better outcomes with colloids for the treatment of severe sepsis of septic shock in randomized clinical trials performed in resource-rich settings. In addition, synthetic colloid solutions have shown important adverse effects, in particular nephrotoxicity with the use of starch solutions. Consequently, the Surviving Sepsis Campaign makes a strong recommendation for the use of crystalloid solutions over colloids for fluid resuscitation [1].

The “Fluid Expansion as Supportive Therapy” trial in children in sub-Saharan Africa with compensated septic shock, of which 57% had severe falciparum malaria, showed a detrimental effect of saline bolus as well as albumin bolus therapy compared to a more conservative fluid therapy [22]. The study supersedes earlier small studies suggesting a survival benefit of albumin infusion over crystalloids in children with severe falciparum malaria and severe sepsis [23, 24].

Three randomized trials in patients with dengue shock syndrome did not show better outcome parameters with (more expensive) colloids over crystalloid fluids [25,26,27]. A quasi-randomized study from the Philippines alternating allocation of colloids with crystalloids also did not show an additional benefit of colloids [28].

From the task force members’ experience, it is important that in wards caring for critically ill patients, intravenous fluids are stockpiled so that they are immediately available for emergency treatment, to save time and to prevent incurring additional costs for the patient’s family.

We recommend crystalloid solutions as the initial fluid of choice in patients with bacterial severe sepsis or septic shock (1B) and recommend against the use of synthetic colloid solutions (1B). We recommend the same for patients with severe falciparum malaria (1B). We also recommend using crystalloids and not colloids in severe dengue with compensated shock for initial fluid resuscitation (1B), but there is insufficient evidence to recommend fluid choices in severe dengue with hypotensive shock. In order to avoid delays in initial resuscitation, it is advisable that wards carrying for patients with sepsis or septic shock stockpile crystalloid solutions for their immediate availability, to avoid delaying initial fluid resuscitation (UG).

7.5 Amounts and Timing of IV Fluids

A landmark study from an emergency department in a resource-rich setting found that so-called early goal-directed therapy, in which intravenous fluids were given to swiftly have physiological parameters return to pre-defined levels, reduced mortality by as much as a third [29]. Early goal-directed therapy has since become mainstream practice in the treatment of critically ill patients. The Surviving Sepsis Campaign recommends that, in the resuscitation from sepsis-induced hypoperfusion, at least 30 ml/kg of intravenous crystalloid fluid be given within the first 3 h [1].

The largest fluid trial performed in resource-limited settings is the above-cited FEAST trial in children [22]. This trial showed an alarming increase in mortality with bolus intravenous infusion in critically ill children. There is an ongoing debate whether mortality increased because of development of pulmonary fluid overload, which could not be compensated for by mechanical ventilation; a secondary analysis of FEAST exploring whether boluses may have caused excess deaths from fluid overload actually suggested cardiovascular collapse rather than fluid overload appeared to contribute most to excess deaths with rapid fluid resuscitation [30]. Nevertheless, similar alarming findings come from several studies in adult patients in resource-limited settings [31,32,33,34]. The most recent trial clearly showed a protocol for early resuscitation with administration of intravenous fluids and vasopressors to increase mortality [34]. The absolute or relative absence of vasopressors and maybe mechanical ventilation could make fluid loading too dangerous.

We recommend that fluid resuscitation is initiated in patients with sepsis and suspected hypovolemia as early as possible, ideally within the first 30 min after recognition, and to start with 30 ml/kg over the first 3 h (1A). Larger amounts of fluid may be needed in patient that remains fluid responsive (e.g., according to the results of a passive leg raise test) and still shows signs of tissue hypoperfusion (e.g., according to the capillary refill time, the skin mottling score, or skin temperature gradients) (1C). We recommend being extremely cautious and thus more conservative in patients in settings with no or limited access to vasopressors and mechanical ventilation, where consideration should be given to stopping fluid administration if the patient develops signs of respiratory distress or lung crepitations on chest auscultation (1A). This also applies for fluid resuscitation in children (1A). Patients with severe malaria or severe dengue without hypotension should not receive fluid bolus therapy (see Chap. 9).

7.6 Vasopressors and Inotropes

The Surviving Sepsis Campaign recommends norepinephrine as the first-choice vasopressor and adding epinephrine to norepinephrine with the intent of raising MAP to target, to decrease norepinephrine dosage. The Surviving Sepsis Campaign also suggests using dopamine as an alternative vasopressor only in selected patients and using dobutamine in patients who show evidence of persistent hypoperfusion despite adequate fluid loading and the use of vasopressors [1].

Extravasation of vasopressors causes skin necrosis, and extravasation is more likely with administration through a peripheral infusion line compared to central venous administration. Central venous catheters, however, are frequently not available, expensive, and inserted too late and frequently require extra payments by family members of the patient further delaying its use. Administration of vasopressors is thus frequently done through a peripheral line. We consider it reasonable to await the effect of initial fluid resuscitation before starting infusion of vasopressors through a peripheral infusion line, but in patients with extreme low blood pressure, and in those not immediately responding to initial fluid loading, it could be necessary to continue without a central venous catheter. Additional advantages of a central venous line are that it can also be used for repeated blood sampling, measurement of static hemodynamic measures, and where possible follow-up of central venous oxygenation.

Vasopressors and inotropes have a narrow therapeutic window, necessitating accurate dosing. Continuous administration at exact doses is safeguarded preferably by automatic infusion with a syringe or infusion pump. Although less accurate, when syringe pumps are not available, these drugs can be diluted in normal saline and administered using a mechanical drop counter.

Norepinephrine is not generally available in hospitals with limited resources. Dopamine is more widely available, but reported best access in resource-limited settings is to epinephrine. We prefer dopamine to epinephrine as the latter may cause lactate acidosis [35, 36]. In resource-limited settings, dobutamine is only available in selected regions, and stockouts of the drug are very common.

Titration of inotropes in resource-limited ICUs is a challenge, as assessed by means of plasma lactate levels is expensive and frequently not possible. Capillary refill time (<3 s) and the skin mottling score can be used to evaluate the effect of infusion of vasopressors and inotropes, but there is no documented evidence regarding efficacy or safety. Of note, vasopressors can affect capillary refill time and skin mottling scores.

We recommend against the start of a vasopressor before initial fluid resuscitation, especially when a central line cannot be used (1C). We suggest starting a vasopressor in patients with persistent arterial hypotension (2C) and recommend targeting a mean arterial blood pressure ≥65 mmHg (1B). We recommend using norepinephrine (noradrenaline) as first-line vasopressor (1B) and suggest using dopamine if norepinephrine is not available (2B). The target for titration of inotropic drugs could be normalization of plasma lactate levels (<2 mmol/L) or normalization of capillary refill time (<3 s) or reduction in skin mottling (UG) if plasma lactate levels cannot be measured. We suggest using dobutamine as first-line inotrope (2B) and epinephrine (adrenaline) if dobutamine is not available (2B). We recommend administering vasopressors via a central venous line (1C) and suggest titrations of vasopressors and inotropes using a syringe or infusion pump when available (2D).

7.7 Conclusions

The paucity of evidence from resource-limited settings and in specific types of sepsis and septic shock underscores the urgent need for rigorous trials, since efficacy and effectiveness of commonly used interventions in resource-rich settings could differ importantly in resource-limited settings.