Abstract
Abstract
Background:
Sepsis is an excessive systemic inflammatory response activated by invasive infection. There has been substantial epidemiologic literature addressing perceived disparities in sepsis by demographic factors such as gender and race. There also have been multiple examinations of the disparities of sepsis with regard to environmental and socioeconomic factors. This paper reviews the current epidemiologic literature evaluating the association of race with the development of sepsis and its associated outcomes.
Methods:
Review of pertinent English-language literature.
Results:
Race is a marker of poverty, preexisting conditions, increased allostatic loads, and decreased access to health systems. Racial disparities and the incidence of sepsis likely are explained by a multiplicity of environmental factors that are not captured by administrative data.
Conclusion:
Race is a surrogate for many intangible factors that lead to the development of sepsis and inferior outcomes.
Beyond the pathophysiological considerations, a plethora of studies have assessed the association of the development of sepsis with sociodemographic factors. Epidemiologic studies of as many as six million people have described an incidence of 3/1,000 population per year or approximately 750,000 cases a year in the United States [2–5,8]. The care of patients with sepsis costs as much as $50,000 per episode, amounting, as noted above, to a total of nearly $17 billion annually the U.S. Several risk factors for sepsis have been identified, including male gender, race, age, certain co-morbid conditions, alcohol abuse, and lower socioeconomic status [1]. From these studies has emerged a growing association between race and the development of sepsis [9,10]. In particular, non-whites are nearly twice as likely as whites to develop sepsis, and men are more likely than women to develop sepsis [11]. These race and gender disparities likely are explained by differences in a variety of environmental factors and co-morbid conditions [11]. Racial differences in severe sepsis also have been described, with higher infection rates and a higher risk of acute organ dysfunction in black than in white individuals [10]. On the other hand, some studies suggest that black patients have a higher survival rate for post-traumatic sepsis than do others [12].
Furthermore, there is consensus among studies describing the considerable variability of the criteria utilized to identify patients with sepsis [13]. This variability is increased by the use of administrative data, which have been utilized for the majority of published epidemiologic studies. This paper reviews studies that address the association between race and sepsis, evaluates the data utilized, and describes the current racial disparities reported in the literature with regard to the development of sepsis.
Administrative Data and Its Flaws
No discussion would be complete without the reader being apprised of the type of data utilized in these studies and the inherent limitations of these administrative data. Administrative data are obtained primarily from inpatient data based on hospital billing. These data typically become uniform billing data to be reported to agencies such as the Agency for Healthcare Research and Quality (AHRQ) or Medicare. The data are standardized to create sets such as the State Inpatient Data (SID) and the Nationwide Inpatient Sample (NIS).
The benefits of the utilization of administrative data for research include the large sample size, regular collection, the relative ease of access, and inclusion of an all-payor in the data set. This being stated, there are several limitations that must be considered to have an appropriate appreciation of the published literature evaluating race and sepsis. As administrative data are based on billing codes, they are subject to variation in reporting at the hospital level. Coding error rates differ widely among states, hospitals, geographic location, and hospital characteristics. Coding errors vary widely among demographic groups and according to whether the state used billing versus abstracted data. Coding differences also occur because of variable systems and manual coding processes for chart abstraction [14].
These analyses are limited by the accuracy of coding and may be subject to differences in regional and institutional practice. In addition, administrative data often lack patient-specific information such as physiology scores, hemodynamic parameters, and re-admission rates, which could be instrumental in assessing the uniformity of the sample and minimizing confounding. Although these methods have been well validated, one cannot attribute death conclusively to sepsis when another acute illness is present.
Other limitations of the data used in the majority of epidemiologic studies evaluating sepsis include the fact that the data cannot show the complete episode of care, may contain various data elements, have sparse clinical detail, are dependent on the state for reporting, and usually do not allow longitudinal assessments. There also have been concerns regarding the adequacy of reporting of co-morbidities for risk adjustment. It has been suggested that the greater observance of sepsis in blacks is secondary to the differential distribution of specific chronic co-morbid medical conditions [15].
As well, some authors have suggested different methods for evaluation of sepsis to enhance the ability of administrative data to predict death in patients with sepsis, including multi-level mixed-effects logistic regression models. These models may be a useful alternative method of severity adjustment for benchmarking purposes or for conducting large, retrospective epidemiologic studies of sepsis patients [16].
Published Epidemiologic Studies Addressing Race and Sepsis
Epidemiologic studies consistently report a higher incidence of sepsis and severe sepsis among black patients than in white patients in medical and surgical populations [2,3,5]. The reasons for these disparities remain unknown; suggested explanations include differences in susceptibility and in the risk of acute organ dysfunction [10]. That being stated, there are no administrative data to support these hypotheses.
One of the first epidemiologic analyses examining the association of race and sepsis was published by Martin et al. [3]. This analysis specifically examined sepsis in relation to race and gender, causative organisms, disposition of patients, and incidence and outcome. The analysis was performed using a nationally representative sample of all non-federal acute care hospitals in the United States from 1979 through 2000. The review of discharge data encompassed approximately 750 million hospitalizations over that 22-year period, and identified 10,319,418 cases of sepsis. Sepsis was more common among males than females and among non-white persons than white persons. The authors concluded that the incidence of sepsis and the number of sepsis-related deaths were increasing, although the overall mortality rate was decreasing.
Other epidemiologic evaluations have supported these findings. Dombrovskiy et al. [9] evaluated pre-morbid conditions and sociodemographic characteristics associated with differences in hospitalization and mortality rates for sepsis in black and white patients utilizing the New Jersey State Inpatient Database for 2002 [9]. A total of 5,466 black and 19,373 white adult patients with sepsis were identified. Blacks had greater hospitalization rates than whites, with the greatest disparity between the ages of 35 and 44 years. Compared with whites, blacks had higher age-adjusted rates of hospitalization and mortality, but similar case fatality rates. Germane to this study was the subset analysis demonstrating that black patients with sepsis had a greater likelihood of having human immunodeficiency virus (HIV) infection, diabetes mellitus, obesity, burns, and chronic kidney disease. The authors concluded that age-adjusted case fatality rates were similar for hospitalized white and black patients with sepsis. These data do not suggest systematic disparities between blacks and whites in the quality of treatment of sepsis. However, blacks had higher rates of hospitalization for and population-based mortality from sepsis. The authors speculated that disparities in disease prevention and care of pre-existing conditions before sepsis onset explained these differences [9].
Further analysis has been performed of sepsis in the surgical population utilizing administrative data. Two epidemiologic studies using different sources described a higher incidence in post-operative sepsis in the black population [5,17]. The first utilized state inpatient data for New Jersey. Disparities were identified in age, gender, and ethnicity and post-operative sepsis. The lowest rate of surgical sepsis was noted in white patients (2.77%), whereas black patients had the highest rate (3.82%; p<0.0002). A similar distribution was found after both elective and non-elective procedures. The authors concluded that the development of post-operative sepsis is multi-factorial, and procedures most likely to be complicated by sepsis did not demonstrate the highest mortality rate after sepsis developed. Factors associated with sepsis included race, age, hospital size, hospital location, and patient income.
Another epidemiologic study evaluating post-operative sepsis that was published in the Annals of Surgery utilized the NIS. This analysis of 6,512,921 elective surgical cases demonstrated significant independent sepsis predictors: Octogenarians compared with patients ages 50 years and younger, and blacks and Hispanics compared with whites. The reasons for these disparities are unclear at an epidemiologic level.
Following these publications, there was a retrospective cohort epidemiologic study published in the Journal of the American Medical Association by Mayr et al. [10]. This analysis utilizing administrative data found a higher rate of hospitalization for infections among black patients than white patients in the U.S. The difference in the rates of invasive pneumococcal disease between black and white patients was significant, particularly among patients younger than 65 years. The investigators concluded that racial differences in severe sepsis are explained by both a higher infection rate and a higher risk of acute organ dysfunction in black individuals. Other authors have questioned the validity of this study secondary to the effect of many unmeasured factors such as smoking status, alcohol consumption, and others that might have contributed to the findings [18]. These factors are addressed below.
Other studies have explored racial differences in the incidence and associated case fatality of severe sepsis, accounting for clinical, social, healthcare service delivery, and geographic characteristics. Barnato et al. [19] evaluated 71,102,655 subjects utilizing hospital discharge and U.S. census data. They noted that blacks had the highest age- and gender-standardized population-based incidence of sepsis and intensive care unit (ICU) case fatality. After adjusting for differences in poverty in their region of residence, blacks still had a higher population-based incidence of severe sepsis than whites, but Hispanics had a lower incidence. The authors concluded that the higher adjusted black incidence and the lower Hispanic incidence could reflect residual confounding or signal biologic differences in susceptibility [19].
Agreeing that there is likely confounding in many of these studies, Esnaola et al. published “Race and Surgical Outcomes: It Is Not All Black and White” [20]. These authors utilized another administrative data source, the National Surgery Quality Improvement Program (NSQIP). They concluded that, after controlling for co-morbidity, African American race had no independent effect on, but was associated with a higher risk of, post-operative cardiac arrest and acute kidney injury [20]. The reality is that all of these data sets are confounded and often do not represent co-morbidities well.
Issues With Race as a Marker
With the aforementioned epidemiologic studies and their findings, it is important to realize that race likely is a poor marker for the development of sepsis. O'Brien described it best in his editorial, “Why Is It Always About Race with You Americans?” [21]. That author suggests that from a study design standpoint, exploring differences in biological pathways would best be accomplished by “controlling” all exposures except the variable of interest, and concludes that a health system that provides equal access to care comes closest to providing such a control [21]. As well, it has been demonstrated on an epidemiologic level that black patients with sepsis had a greater likelihood of having HIV infection, diabetes mellitus, obesity, burns, or chronic kidney disease [9].
It is likely that race is a marker for many secondary issues that cannot be detected through exploration of administrative data. The studies that have been presented utilizing these epidemiologic and administrative data sets are each flawed, as there are many unmeasured factors, and race is likely a surrogate for these unidentified factors.
Environmental and Co-Morbidity Disparities
Factors associated with race that likely are environmental include smoking status, alcohol consumption, nutritional status, and the prevalence rate of undiagnosed HIV infection. Wong et al. [22] estimated cause-specific risks of death from the National Health Interview Survey data collected from 1986 through 1994 and from linked vital statistics. They concluded that persons without a high school education lost 12.8 potential life-years per person in the population, compared with 3.6 for persons who graduated from high school. The pattern of disparities according to income was similar to that according to level of education. Blacks and whites lost 7.0 and 5.2 potential life-years per person, respectively, as a result of deaths from any cause. The investigators concluded that many conditions contribute to socioeconomic and racial disparities in potential life-years lost, but that a few conditions account for most of these disparities: Smoking-related diseases, hypertension, HIV, diabetes mellitus, and trauma as the cause of death among black persons [22]. Therefore, environment accounts for many of the disparities seen, as well as the increased association of race with co-morbidities.
Disparities in critical illness are evident in a variety of racial and ethnic groups. Most data available in the literature reflect variations in the incidence, presentation, diagnosis, treatment, and outcomes between African Americans and whites. Most research in critical care concerning disparities relates to cardiovascular illnesses. Substantially less information is available regarding disparities in common ICU diagnoses. Data also are lacking to delineate the reasons for disparities among critically ill patients. Further research is required to elucidate the root causes of racial or ethnic differences, provide adequate education for healthcare providers, and develop and implement evidence-based interventions targeted to specific patient groups [23].
Specific Conditions of Possible Importance
Human immunodeficiency virus infection
Regarding sepsis specifically in these studies, HIV is a major confounder. Although administrative studies exclude patients with known HIV infection from analysis, racial differences for rates of undiagnosed HIV infection are significant. Campsmith et al. described adults and adolescents living with undiagnosed HIV infection in the United States at the end of 2006 [24]. The authors concluded that whites had the lowest percentage (18.8%) compared with Hispanics/Latinos (21.6%), blacks/African Americans (22.2%), American Indians/Alaska Natives (25.8%), and Asians/Pacific Islanders (29.5%; all p<0.001). They demonstrated significant differences in undiagnosed HIV, which can assist in explaining the higher rates of sepsis observed in these populations. Furthermore, the unmeasured effect of undiagnosed HIV infection on racial differences in rates of infection, particularly for invasive pneumococcal disease, should be recognized. Black persons of all ages are affected disproportionately by invasive pneumococcal disease [24]. These undiagnosed infectious diseases can account for the higher sepsis rates and infectious complications described for different races.
Alcohol abuse
Chronic alcohol abuse has been associated with delayed resolution of symptoms, higher rates of bacteremia, increased use of intensive care, prolonged hospital stays, and greater cost of hospitalization for infected patients [11].
The association of alcohol and sepsis has been described. For example, a history of chronic alcohol abuse significantly increases the risk of acute respiratory distress syndrome in critically ill patients with an identified at-risk diagnosis [25]. It has been suggested that the drinking patterns of racial/ethnic groups differ considerably. Keyes et al. [26] reviewed evidence from publications regarding racial/ethnic differences in overall alcohol-attributable injury and found that Native Americans had higher rates of alcohol-attributable motor vehicle crash fatality, suicide, and falls than other racial/ethnic groups. Black subgroups also evidence higher rates of alcohol positivity than would be expected given the estimates of alcohol use, including rates of alcohol positivity among drivers who fatally injured black children and homicide victims.
Finally, Esper et al. [15] utilized the National Hospital Discharge Survey to demonstrate that chronic co-morbid conditions that alter immune function (chronic kidney disease, diabetes mellitus, HIV, alcohol abuse) were more common in non-white patients with sepsis, and cumulative co-morbidities were associated with greater acute organ dysfunction [15].
Access to healthcare and insurance
It has long been said that there is differential access to healthcare for minorities that minimizes the opportunities for prevention and health education. Authors have stated that this would be manifested by decreased access to health information, participation in health promotion and disease prevention activities, nutritious foods, convenient exercise spaces, freedom from ambient violence, adequate social support, communities with social capital, and high-quality healthcare. Access to healthcare is facilitated by health insurance, a regular source of care, and a usual primary care provider. Various mechanisms through which access to health promotion and care are mediated by race and ethnicity are discussed; these include the built environment, social environment, residential segregation, stress, racism, and discrimination [27].
Other influential factors in the development of sepsis and survival therefrom include access to care; this often is associated with having health insurance. The risks of sepsis-associated hospitalization and death differ significantly by insurance status [28]. O'Brien et al. stated that socio-demographic and clinical factors associated with increased sepsis risk, including older age, non-white race, and specific co-morbidities, are more common among patients on Medicare or Medicaid or having no health insurance of any kind. Those authors demonstrated that Medicaid, Medicare, and uninsured patients had significantly higher risk-adjusted odds of a sepsis-associated admission than those with private insurance [28].
Intensity of care
There has been a suggestion that black patients receive inferior care, such as lower quality and intensity of care. Studies in the surgical literature have suggested inferior intensity for minorities. For example, for breast surgery, black race is an independent predictor of underuse of surgery among urban patients [29]. Therefore, it is possible that a low intensity of care confounds other issues of race and access to care. These factors do not receive suitable attention in current epidemiologic analyses attempting to explain the disparities of sepsis and mortality rates.
Mayr et al. [30] evaluated the treatment of pneumonia in black patients in 28 U.S. hospitals in an attempt to confirm this hypothesis. This analysis compared care quality on the basis of receipt of antibiotic within 4 h, adherence to American Thoracic Society antibiotic guidelines, and care intensity based on ICU admission and mechanical ventilation use. Black patients were less likely to receive antibiotics within 4 h and less likely to receive guideline-adherent antibiotics. The authors concluded that according to crude analyses, black patients receive lower-quality and higher-intensity care. The authors also suggest that these differences were explained by different case mixes and variations in care across hospitals [30].
Marital status
Many environmental factors have been associated with sepsis and found to affect its outcome. Seymour et al. published a study entitled “Marital Status and the Epidemiology and Outcomes of Sepsis” [31]. This study analyzed 1,113,581 hospitalizations in New Jersey and estimated risk-adjusted incidence rate ratios for sepsis among divorced, widowed, legally separated, single, and married subjects using population data from the American Community Survey. Hospitalization for sepsis was more common among single, widowed, and legally separated individuals, independent of other demographic factors. Among patients hospitalized for sepsis, single and divorced men and single women experienced higher hospital mortality rates [31]. This study highlights the importance of environmental factors to sepsis and suggests potentially modifiable mechanisms linking marital status to a greater burden of critical illness.
Genetic variation
Several authors have attempted to assess genetic variation and its association with outcome disparity in critical illness. Analysis of single nucleotide polymorphisms among various genes, such as those coding for pro- and anti-inflammatory mediators, in the sepsis response has the potential to develop into prediction tools. However, unraveling the genetic variation in sepsis is complicated, and close attention must be paid to study design [32].
Evidence from family based studies and recent gene-association studies suggest that a significant portion of the apparent variability in susceptibility is attributable to genetic factors. Common sequence variations in genes coding for innate immune effectors, inflammatory mediators, and modulators of coagulation have received particular attention [33]. This being theorized, there is little tangible evidence that genetic variation by race is linked to sepsis.
Allostatic load
Following the theme of environmental factors and the outcomes of sepsis, some authors have stated that marginalized populations of lower socioeconomic position exposed disproportionately to irritants (e.g., tobacco smoke), pollutants (e.g., diesel-related particles), and indoor allergens may also live in neighborhoods that increasingly are socially toxic [34]. A recent study attempted to account for environmental stressors and immune function in relation to sepsis.
Stressors are theorized to be associated with a higher allostatic load (AL), a concept of physiologic wear measured as a composite of physical biomarkers. The risk of a high AL may differ by gender and may be intensified in places with substantial environmental risks, otherwise known as “environmental riskscapes.” Mair et al., using primary data collected in a sample of patients exposed to various environmental and social stressors and residential proximity to petrochemical plants, perceived poor neighborhood conditions and daily “hassles” to be associated with higher AL components and variation, depending on gender and race [35]. Those investigators found that non-Hispanic black women had higher cardiovascular- and immune-related AL.
The theory is that cumulative levels of “stress” have deleterious effects on health and longevity. Adrenal hormones, neurotransmitters, and immuno-cytokines may be elevated chronically in environments with high AL. Many of these factors are adaptive in the short term but become damaging long term [36]. Factors associated with greater environmental stress are more prevalent in minorities, and include smoking, alcohol consumption, social conflict, and chronic stressor experience, leading in theory to lower functional immune activity. The theory of allostasis supports the contention that low socio-economic status can lead to “downstream” peripheral biology changes. Furthermore, AL has been associated with a higher mortality rate. Borrell et al. investigated the association between cumulative AL and all-cause mortality risk and concluded that a high allostatic score increased the risk of all-cause mortality, and suggested that the conditions reflected in a high allostatic score may contribute to premature death in the United States [37].
Brame at al. [38] asked the question, “Stressing the Obvious? An Allostatic Look at Critical Illness.” The authors concluded that a clear analogy may be made to critical care, where excess stress affects metabolic, hormonal, and immune-inflammatory responses and contributes to the development of organ failure. Ongoing stress also compromises recovery, so it is incumbent on caregivers to reduce stress, be it induced by tissue hypoxia, catecholamine infusion, sleep deprivation, pain, anxiety, or excessive noise [38].
Racial and ethnic patterns of AL among adult women in the United States were evaluated utilizing the National Health and Nutrition Examination Survey [39]. Measures of AL using 10 biomarkers representing cardiovascular, inflammatory, and metabolic system function were identified. Black females had the highest predicted AL scores, and a marked black/white gap in AL persisted across all age groups. This study provides one of the first descriptive profiles of AL among a nationally representative sample of adult females in the United States and presents racial/ethnic trends in AL across age groups that are useful for identifying demographically and clinically important subgroups at risk of having high cumulative physiologic dysregulation [39]. Analysis of the National Health and Nutrition Examination Survey, 1988–1994, also found a high AL score among Mexican immigrants and suggested that repeated or chronic physiologic adaptation to stressors is one contributor to the “unhealthy assimilation” effect observed in Mexican immigrants [40].
Social scientists have examined the effects of residence in poor neighborhoods on child health for more than 50 years. These marginalized individuals may live in communities that are increasingly toxic socially, which, in turn, may be related to a greater experience of psychosocial stress that may influence asthma morbidity. Epidemiologic trends suggest that asthma provides an excellent paradigm for understanding the role of community-level contextual factors in disease. Specifically, a multi-level approach that includes an ecological perspective may help to explain heterogeneities in asthma expression across socioeconomic and geographic boundaries that, to date, remain largely unexplained. Far less attention has been given to the broader social context in which individuals live. A multi-level approach that recognizes explicitly the embedding of asthma in its biological, psychosocioeconomic, environmental, and community contexts is likely to provide a better understanding of asthma disparities at different stages in the life course. Is it simply asthma disparities, or is it social disparities in asthma? [41]
It has even been suggested that allostatic load may be responsible for perceived differences in genetic endowment [42]. Such differences may have consequences for patterns of physiologic activity [42].
Conclusions
Racial disparities and sepsis are multi-factorial and are not explained easily by race alone. Disparities in sepsis likely reflect differences in vulnerability to disease as a result of social resources, environmental factors, and healthcare interventions [11]. Race is a likely marker of poverty, high stress levels, less education, and decreased access to healthcare. More than 44 million Americans lack health insurance, and hospitalized blacks are more than three times as likely as whites to be uninsured. This disparity increased to nearly fourfold for black sepsis patients [43]. The reality is that lack of insurance limits access to preventive health services and thus not only contributes to greater pre-hospitalization co-morbidity but often delays decisions to seek treatment.
Current epidemiologic studies describe the broad finding of race, as it is easy to define in current administrative data, but these analyses have only begun to elucidate the problems. Further investigation into areas such as AL and environment may help to unravel the complex issues of sepsis. Identifying race as a reason for the development of sepsis is myopic at best; race is a surrogate for other, more basic and systemic environmental issues that need to be addressed.
Footnotes
Acknowledgments
The author thanks Donald E. Fry for his persistence, review, and comments. As well, the author is grateful to Viktor Y. Dombrovskiy for his review and continued collaboration.
Author Disclosure Statement
No competing financial interests exist.