Trying To Improve Sepsis Care In Low-Resource Settings.
Published 2017 · Medicine
Earlier this year, the World Health Organization recognized sepsis as a global health problem, responsible for millions of preventable deaths every year, and adopted a resolution targeting the prevention, diagnosis, and treatment of sepsis, especially in lowand middle-income countries.1 Although most sepsis cases are assumed to occur in lowand middle-income countries, nearly all research on both the epidemiology of sepsis and optimal treatment comes from high-income countries. High-income countries can claim some success with regard to declining rates of mortality related to sepsis in recent years, whereas data from lowand middle-income countries suggest sepsis still carries very high fatality rates.2,3 In other words, data about which patients develop sepsis and how they should be treated are lacking from locations where sepsis is thought to be most common and associated with the poorest outcomes. If sepsis was a homogenous condition and the effectiveness of therapy was uniform, such paucity of data would not be troubling. However, neither is the case. Sepsis is a broad syndrome of acute organ dysfunction arising due to a complex host response to a wide variety of inciting pathogens, with the result that both the presenting symptoms and subsequent course are highly variable. Furthermore, the effectiveness of each element of the antimicrobial, resuscitation, and organ support therapies used to treat sepsis appears to be highly conditional on these host and pathogen factors, as well as on what other therapies are deployed, and on the capacity of the health care setting to deliver therapies optimally and monitor, prevent, and treat complications. In many lowand middle-income countries, patients who develop sepsis may be much more likely to have malnutrition, HIV, or malarial infection. These patients may also develop sepsis secondary to a spectrum of infections different from that seen in high-income countries, such as highly resistant gram-negative infections, dengue, melioidosis, or viral hemorrhagic fevers. Patients may also incur extreme delays before receiving definitive care, which may be lacking basic resources, such as vasopressors or oxygen, as well as artificial respiratory or kidney support. It is therefore extremely valuable to determine how well even the most basic therapeutic strategies work in these settings. In this issue of JAMA, Andrews and colleagues4 report the results of an unblinded randomized clinical trial testing whether a protocol approach to resuscitation for patients with presumed sepsis-induced hypotension presenting to the emergency department of a 1500-bed referral hospital in Zambia would improve survival compared with usual care. The overarching premise was that patients who received usual care were at high risk of underresuscitation, and therefore a time-based protocol to promote aggressive use of fluids, blood, and vasopressors would improve the quality of resuscitation and hopefully lower mortality. Based on earlier work, which suggested potential harm from pulmonary edema due to fluid overload, the authors excluded patients who presented with an arterial oxygen saturation of less than 90% and a respiratory rate of greater than 40 breaths per minute. The intervention was modeled on the early goal-directed therapy (EGDT) protocol tested by Rivers et al5 at Henry Ford Hospital in Detroit, Michigan, but with a number of important alterations. Similar to EGDT, the intervention consisted of a 6-hour resuscitation protocol initiated at the time of enrollment and beginning with a bolus of intravenous saline, but provided 2 L, which is potentially a greater volume than the 20 to 30 mL/kg typically used in EGDT. After the initial bolus, the EGDT protocol titrates fluid, blood, and vasoactive agents based on blood pressure, central venous pressure, and central venous oxygen saturation. Because such monitoring was unavailable in Zambia, the protocol tested by Andrews et al4 followed the initial bolus with an additional 2 L administered over the next 4 hours regardless of an improvement in hypotension and only interrupted fluid administration if arterial oxygen saturation decreased by 3%, the respiratory rate increased by 5 breaths per minute, or the jugular venous pressure was 3 cm or greater above the sternal angle, all of which had to be determined hourly. Dopamine was initiated if the mean arterial pressure remained below 65 mm Hg after completion of the initial saline bolus and blood transfusion was given if the hemoglobin concentration decreased to less than 7 g/dL or if the patient had severe pallor. The usual care was described as variable but generally lacked aggressive intravenous fluid administration. The primary outcome was in-hospital mortality and the secondary outcomes included 28-day mortality, time to death, and safety outcomes such as worsening hypoxemia or tachypnea. Contrary to the authors’ hypothesis, the modified intention-to-treat analysis showed a higher mortality rate among patients who received the early resuscitation protocol (51/106 patients [48.1%] in the sepsis protocol group) than in usual care group (34/103 patients [33.0%]) with a between-group difference of 15.1% (95% CI, 2.0%-28.3%; relative risk, 1.46 [95% CI, 1.04-2.05]; P = .03). The between-group difference remained significant after 28 days (67.0% with the sepsis protocol vs 45.3% with usual care; P = .002) and persisted in the multivariate analysis adjusting for severity of illness and across prespecified subgroups defined by preexisting HIV infection and baseline Glasgow Coma Scale score, hemoglobin level, lactate level, severity of illness, and jugular venous pressure. Related article page 1233 Opinion