Sepsis and the Hematology Laboratory

An affordable, widely available test can impact today`s biggest healthcare challenge.

Sepsis, the inflammatory response to infection, is quickly becoming one of the biggest healthcare problems worldwide. No matter the perspective one takes, the numbers are staggering. Currently the number of diagnosed cases per year in the United States is at least 750,000; some estimates surpass one million. Worldwide mortality estimates are as high as 20 percent, and thus we are dealing with one of the biggest drivers of mortality in modern medicine. Sepsis kills nearly as many people as heart attack, HIV, and breast cancer combined.

Viewed from the perspective of health economics, the average in-hospital cost per case is approximately $20,000 dollars, and yearly estimates of sepsis-related expenses in the U.S. alone exceed $20 billion.

To compound an already dire situation, all predictive models indicate steady increases in sepsis prevalence over the next years and decades, as the risk factors typically linked to sepsis become ever more prevalent: The proportion of elderly patients in the general population is increasing; the number of patients living with chronic health problems such as diabetes, cancer, and chronic renal failure, among others, is increasing; and the number of patients taking immunosuppressant therapies for a variety of reasons is also on the rise. Put these together and there is a perfect storm brewing, and healthcare systems and providers, among them pathologists and laboratories, must be ready to respond.

In order to prepare, it is critical to understand how sepsis patients enter healthcare pathways, and where key opportunities to improve outcomes lie. Literature on this topic yield some key findings which can guide healthcare institutions and providers in designing optimal strategies to deliver fast and effective care to their sepsis patients:

• Sepsis patients typically enter healthcare pathways via the emergency department (ED) (approximately 70 percent of cases), or become septic during an intensive care unit (ICU) admission (approximately 25 percent). These two locations are critical target areas for optimizing early sepsis detection protocols.
• The majority of patients who ultimately die of sepsis already are septic upon presentation at the ED.
• The majority of patients who die of sepsis do not have more severe forms of sepsis (with documentation of end-organ failure) at presentation.

Taken together, these findings highlight the need for well-established early warning mechanisms to raise the suspicion of sepsis even in the earliest stages of the disease. The ideal location for the deployment of such mechanisms is the ED (including the laboratory services that target this patient population).

Improvements in sepsis outcomes

In the last decade, while sepsis incidences were sharply increasing, the good news was that patient outcomes were actually improving. Some studies demonstrated decreases from 18 percent to 12 percent in sepsis-associated hospital mortalities between 2004 and 2013. This success was due in great part to efforts such as those of the Surviving Sepsis Campaign, which raised awareness among treating physicians of the importance of having a high index of suspicion for the disease. In this effort, researchers documented that the most critical factor leading to improved outcomes in sepsis was the time to start effective antibiotics. One study showed that each hour of delay was associated with a 7.6 percent increase in mortality. With this data in mind, many institutions now have early sepsis detection protocols to ensure that these patients are recognized and treated quickly, and such protocols and increased awareness have been credited with bringing about the significant improvements in sepsis mortality.

However, now that sepsis is already being treated quickly and effectively once clinical suspicion arises, the challenge has moved elsewhere; in order to further improve outcomes, the goal now is to assist clinicians in suspecting sepsis sooner, ideally before the more obvious clinical signs and symptoms are present. And the hematology laboratory may play a critical role in addressing this new challenge in the sepsis arena.

How hematology can help

Today the laboratory plays only a limited role in the early detection of sepsis and the improvement of patient outcomes, for several reasons:

• Lab tests for sepsis tend to have low specificity, as most current laboratory biomarkers of sepsis are also elevated in other inflammatory conditions, for example, systemic inflammatory response syndrome (SIRS).
• The majority of recently proposed tests for sepsis (procalcitonin, lactate) are performed only when the clinician is already suspecting it (and thus is ordering the test). At this point, the sepsis protocol, including broad spectrum antibiotics, will be implemented anyway, and thus further improvements in patient outcomes are unlikely.
• Cost constraints prevent the widespread utilization of these recently proposed biomarkers in the broad ED population, thus making them inadequately suited as early markers of sepsis (prior to clinical suspicion).

Sepsis researchers have recognized that any new biomarkers they propose will likely face similar constraints, especially if they are additional tests not routinely performed during the ED visit or ICU admission. Therefore, a new focus has emerged in studies trying to identify mechanisms for earlier suspicion of sepsis using data that is already available. In particular, laboratory data such as CBC-diff results, analyzed in new multi-parametric algorithms or using novel data not previously available for clinical use, are especially valuable, because they are readily available at a very low cost for the majority of patients visiting the ED or admitted at the ICU. Thus these “upgraded” CBC-diff results can be used as early warning biomarkers even before clinical suspicion would justify the ordering of more costly tests such as procalcitonin.

For these reasons, researchers have recently published numerous studies on new strategies to leverage already available CBC-diff data for early detection of sepsis. These strategies fall in two broad categories: multi-parametric algorithms, and utilization of cellular morphologic data.

Multi-parametric algorithms

Multi-parametric algorithms are logistic regression models which identify various parameters differing in two groups (in this case, sepsis versus non-sepsis), giving weight to each individual parameter based on the magnitude of the difference seen for that parameter, and using a mathematical formula including those weights to identify a final “factor” –an index result which predicts the likelihood of sepsis. The key challenge with this approach is reproducibility across patient populations; the more parameters one includes in the model, the less likely it is that results will be replicated at a different institution.

Another common mistake seen in this line of research has been the inclusion of non-hematological data in the models to increase the discriminating power (for example, other biomarkers such as procalcitonin, CRP, interleukins, and sometimes even patient clinical and demographic data). This is a mistake because these are not tests that are routinely available to most patients at the same time the CBC-diff is being performed, so an algorithm using them will have the same problem of multiple recently proposed biomarkers for sepsis which have not been widely adopted by clinicians; if you need to suspect sepsis first to order the test, the window of opportunity to further improve sepsis outcomes has already passed.

Cellular morphologic data

The utilization of cellular morphologic data is based on the recognition that leucocytes, when primed to fight infection, undergo key changes in various morphologic features such as their shape, size, cytoplasmic granularity, and even their pliability. In fact, these features have already been used for decades by pathologists and technologists while reviewing cells under the microscope, and findings such as toxic granulation, Dohle bodies, and cytoplasmic vacuolization are currently used to raise the suspicion that a patient is undergoing an infectious process regardless of the total and differential leucocyte counts. The key challenge in this approach is that only a small proportion of CBC-diff test orders ever lead to a microscopic review, and given current staff limitations, cost containment efforts, and increasing pressure for a very fast turnaround time of results, it is not feasible for laboratories to manually review all CBC-diff tests coming from ED patients in order to search for these morphologic features, no matter how diagnostically relevant they may be.

A key solution currently under investigation is the utilization of hematological analyzers which automatically quantify these morphologic features, so that they can be reported to clinicians as numerical parameters included in the routine CBC-diff results. This approach is already widely accepted in the red blood cell arena, where the mean cell volume (MCV) and the red blood cell distribution width (RDW) are part of the traditional CBC-diff results. It can offer clinicians critical insight into morphological features of red cells, so that they can further guide their diagnostic work-up, especially in anemic patients.

In summary, the hematology laboratory is perfectly positioned in the sepsis diagnostic pathway to become a key solution to further improve clinical outcomes in this disease. This is true because the CBC-diff is a test routinely ordered at very low cost for the majority of patients entering the ED or admitted at the ICU, the two main locations where early sepsis detection is critical. In order for this use of the CBC-diff to be realized, however, it is critical that researchers find better ways to use hematological data, because the traditional parameters reported today as part of the CBC-diff do not have the level of sensitivity and specificity needed to properly distinguish sepsis from the myriad of mimicking conditions that clinicians must include in their differential diagnoses.

The impact of new diagnostic criteria

A word of caution is needed, however, as the official criteria for sepsis diagnosis have recently changed. A task force of 19 leaders in the field of sepsis was convened by the Society for Critical Care Medicine and the European Society of Intensive Care Medicine to put forth new guidelines for sepsis diagnosis. While a full description of these changes is beyond the scope of this article, it is important to highlight how they impact the role of the laboratory in sepsis diagnosis.

In previous guidelines, sepsis was diagnosed when patients had clinical evidence of systemic inflammatory response, and a documented or suspected infection. “Severe sepsis” was diagnosed when end organ failure was also documented based on a set of clinical and laboratory tests for organ function (i.e., creatinine, bilirubin, platelet counts, among others).

In the new guidelines, the category “severe sepsis” was removed, and the diagnosis of sepsis itself now requires the recognition of a life-threatening organ dysfunction due to a disregulated host response to infection. This organ dysfunction is based on the “sequential organ failure score” (SOFA), which uses several laboratory parameters.

The key challenge to the hematology laboratory is that its greatest potential value is in the identification of early sepsis patients; those who don’t yet have end organ failure and thus could benefit most from early antibiotic therapy. Under the new guidelines, these patients will no longer be formally diagnosed as “septic.” Having said that, it is well recognized in the literature that these patients do have higher mortality rates compared with patients with simple infections not associated with a disregulated inflammatory response, and thus early identification of these patients remains critical to improve outcomes, regardless of the formal diagnosis they ultimately receive.

It is critical that laboratory medicine researchers and practicing pathologists consider the differences between the old and new sepsis guidelines in their decisions, and in particular that future studies continue to distinguish patient populations with simple infections from those who have associated disregulated inflammation, because prompt identification of the latter population is needed for early initiation of antibiotic therapy and overall improvements in sepsis-related outcomes.