Chronic kidney disease is associated with a higher 90-day mortality than other chronic medical conditions in patients with sepsis

According to previous studies, the clinical course of sepsis could be affected by preexisting medical conditions, which are very common among patients with sepsis. This observational study aimed at investigating whether common chronic medical conditions affect the 90-day mortality risk in adult Caucasian patients with sepsis. A total of 482 patients with sepsis were enrolled in this study. The ninety-day mortality was the primary outcome; organ failure was the secondary outcome. Sepsis-related organ failure assessment (SOFA) scores and the requirements for organ support were evaluated to assess organ failure. A multivariate Cox regression model for the association between the 90-day mortality risk and chronic preexisting medical conditions adjusted for all relevant confounders and mortality predictors revealed the highest hazard ratio for patients with chronic kidney disease (CKD) (hazard ratio, 2.25; 95% CI, 1.46-3.46; p = 0.0002). Patients with CKD had higher SOFA scores than patients without CKD (8.9 ± 4.0 and 6.5 ± 3.4, respectively; p < 0.0001). Additionally, an analysis of organ-specific SOFA scores revealed higher scores in three organ systems (kidney, cardiovascular and coagulation). Patients with CKD have the highest 90-day mortality risk compared with patients without CKD or with other chronic medical conditions.

conditions were retrospective in nature 12,13 . Additionally, the major outcome of previous investigations was short-term mortality (28-day, ICU and hospital mortality) [12][13][14][15] . Because many sepsis patients remain hospitalized at day 28, and because strong evidence suggests that many late sequelae from sepsis are not captured by this time point 16 , many experts suggest that sepsis studies consider a larger window of time, such as 60 or 90 days 17 .
Moreover, as several improved treatments for sepsis, especially for sepsis-associated organ dysfunction, have been developed in recent years 18 , reevaluating the effect of common chronic medical conditions on the clinical course of patients with sepsis is important.
This prospective observational study aimed at investigating whether and to what extent common chronic medical conditions (arterial hypertension, coronary heart disease, chronic obstructive pulmonary disease (COPD), chronic kidney disease (CKD), insulin-dependent diabetes mellitus (IDDM), non-insulin-dependent diabetes mellitus (NIDDM), chronic liver disease, and history of stroke) affect the 90-day mortality risk in Caucasian patients with sepsis.

Results
Baseline characteristics. A total of 482 adult Caucasian patients with sepsis were enrolled in this observational investigation. All of the patients were successfully followed for a maximum of 90 days. The patients' ages ranged from 19 to 92 years (median, 64 years) ( Table 1); 35% were women and 65% were men. The distribution of sepsis/severe sepsis and septic shock was 40% and 60%, respectively. An assessment of the frequency of the eight chronic medical conditions revealed arterial hypertension to be the most common preexisting disease (55%) and history of stroke to be the least common (6%). Thirty percent of patients had a history of recent elective surgery, and 52% had a history of recent emergency surgery. The most common site of infection was the lung (55%).
At baseline, the mean sepsis-related organ failure assessment (SOFA) and acute physiology and chronic health evaluation II (APACHE II) morbidity scores were 9.1 ± 4.0 and 21.3 ± 7.0, respectively ( Table 1). The frequency of organ support therapy (mechanical ventilation, vasopressor therapy and renal replacement therapy) at the time of sepsis onset was 84%, 60% and 9%, respectively (Table 1).

Outcomes. Mortality.
A multivariate Cox regression model for the association between the 90-day mortality risk and chronic preexisting medical conditions adjusted for all of the relevant confounders and mortality predictors revealed the highest hazard ratio for patients with CKD (hazard ratio, 2.25; 95% CI, 1.46-3.46; p = 0.0002) ( Table 2) followed by those with diabetes mellitus (NIDDM: hazard ratio, 1.65; 95% CI, 0.96-2.83; p = 0.0684 and IDDM: hazard ratio, 1.62; 95% CI, 0.99-2.64; p = 0.0527) and a history of cancer (hazard ratio, 1.63; 95% CI, 1.09-2.34; p = 0.0182) ( Table 2). This finding indicates that despite potential baseline confounders (age, gender, initial APACHE II and SOFA scores, septic shock, type of infection, recent surgical history), pre-existing CKD is an independent and significant prognostic variable for the 90-day mortality risk (Table 2).
Similarly, a Kaplan-Meier survival analysis of the 90-day mortality risk among patients with CKD and patients without CKD revealed a significantly higher mortality risk among patients with CKD compared with patients without CKD (p < 0.0001, log-rank test) (Fig. 1). Similarly, patients with CKD had a significantly higher 28-day mortality rate compared to patients with no history of CKD (Table 3).
Disease severity. Over the 28-day observational period in the ICU, the mean SOFA score was 6.8 ± 3.6 ( Table 3). Analyses of organ-specific SOFA scores revealed the highest scores in the respiratory, cardiovascular and central nervous systems (1.9 ± 0.8, 1.5 ± 1.0 and 1.9 ± 1.1, respectively).
To explore the effect of CKD (as the most predictive variable for mortality) on disease severity and the extent of organ dysfunction over the course of the ICU stay, we calculated organ-specific SOFA scores. Patients with CKD had higher SOFA scores compared with patients without CKD (8.9 ± 4.0 and 6.5 ± 3.4, respectively; p < 0.0001) ( Table 3). Additionally, analyses of organ-specific SOFA scores revealed higher scores in three organ systems (kidney, cardiovascular and coagulation). Compared with patients without CKD, patients with CKD had higher SOFA-Renal scores (1.8 ± 1.4 and 0.7 ± 1.7, respectively; p < 0.0001), SOFA-Cardiovascular scores (1.9 ± 1.1 and 1.4 ± 0.9, respectively; p = 0.0012) and SOFA-Coagulation scores (0.6 ± 0.8 and 0.3 ± 0.5, respectively; p = 0.0242). The remaining two SOFA scores (respiratory and central nervous system) did not differ between the groups (Table 3).
Regarding organ support-free days, patients with CKD had significantly fewer vasopressor-free days compared with patients without CKD (7 ± 6 and 11 ± 7, respectively; p = 0.0002). CKD patients also had significantly fewer dialysis-free days compared with patients with no CKD history (11 ± 8 and 14 ± 8, respectively; p = 0.0030). The groups did not differ with regard to ventilator-free days.

Discussion
This observational clinical investigation assessed the effect of the most common chronic medical conditions on the 90-day survival among patients with sepsis. According to our main findings, patients with CKD had the highest 90-day mortality risk compared with patients without CKD or patients with other chronic medical conditions. This observation underscores the negative consequences of pre-existing CKD for the clinical course of sepsis and confirms previous investigations on the role of CKD in sepsis 19 Efforts are needed to reduce the incidence and control the negative effects of infections in patients with CKD.
A major advantage of our study is the fact that we investigated, for the first time, organ-specific dysfunctions over the clinical course of disease with regard to CKD using organ-specific SOFA scores. Patients with CKD showed three higher organ-specific SOFA scores compared with patients without CKD. This result indicates more pronounced organ dysfunction in these three organ systems (cardiovascular, renal and coagulation). The higher SOFA-Cardiovascular score and the more frequent use of vasopressor therapy in patients with CKD compared with patients without CKD are consistent with previous observations demonstrating a higher demand for vasopressor administration in septic patients with renal failure 31 . The higher SOFA-Renal scores and the more frequent renal-replacement therapy Organ support, %

Mechanical ventilation 84
Vasopressor therapy 60 Renal-replacement therapy 9 Table 1. Patient baseline characteristics at the onset of sepsis.
in CKD patients compared with patients without CKD are intuitive, because sepsis patients with CKD are more likely to develop renal failure of acute or chronic kidney disease 2,19,31,32 . The observed higher SOFA-Coagulation scores, indicating severe thrombocytopenia, in the CKD group compared with the non-CKD group is in agreement with previous investigations showing that renal disease is associated with platelet dysfunction and thrombocytopenia 33 . Additionally, the pronounced thrombocytopenia observed in this patient group is consistent with the higher recorded mortality risk in these patients, because thrombocytopenia was shown to be a prognostic variable for mortality in patients with sepsis 7 . The observed high mortality risk among patients with diabetes mellitus is in accordance with several previous studies showing that DM is associated with higher mortality caused by sepsis in several populations 11 . These findings are consistent with the fact that several aspects of immunity and host defense  are altered in patients with DM 34 . Similarly, the high risk of mortality among patients with a history of cancer in our cohort is consistent with the results from previous studies 9 and could be attributed to the fact that patients with cancer are at a high risk for developing a state of immunosuppression resulting from cancer therapy or the malignancy itself, thus leading to severe infection and sepsis, which is a is major cause of mortality in this group 35,36 .
Our study confirms the prognostic value of the initial morbidity scores, SOFA and APACHE II 37 . Furthermore, the significant association between age over 65 years and higher mortality risk is plausible.
Our study provides an important update of the prognostic value of the most common chronic medical conditions on the 90-day mortality risk among patients with sepsis. We found the highest mortality risk among patients with CKD, and much effort must be made to minimize the mortality risk in this group. Researchers and clinicians need to develop new treatment strategies, both preventive and curative, that are specially adapted for this patient group.   , 2008), and the study protocol was approved by the institutional ethics committee of the University of Goettingen in Goettingen, Germany. The study was performed in accordance with relevant guidelines and regulations. The methods were performed in accordance with the approved guidelines. Written informed consent was obtained either from the patients or from their legal representatives.
Exclusion criteria. As described previously 39 , the exclusion criteria were: (1) age younger than 18; (2) pregnancy or breastfeeding (3) receipt of immunosuppressive therapy; (4) documented myocardial infarction within the previous 6 weeks; (5) New York Heart Association functional class IV chronic heart failure; (6) human immunodeficiency virus infection; (7) a do not resuscitate or do not treat order; (8) expected death within 28 days due to uncorrectable medical condition (e.g., poorly controlled neoplasm); (9) chronic vegetative state with pronounced neurological impairment; (10) current participation in any clinical trial (of a drug or device); (11) inability to be fully evaluated during the study period; and (12) study-site employee or a family member of a study-site employee.
Data collection and clinical endpoints. Upon enrollment, the patient's demographic characteristics, type of sepsis (sepsis/severe sepsis and septic shock), chronic comorbidities, recent surgical history (elective surgery, emergency surgery), site of infection and organ support were recorded. All of the patients were followed for 90 days, and the mortality risk within this observational period was recorded as the primary outcome. Sequential Organ Failure Assessment (SOFA) 40 and Acute Physiology and Chronic Health Evaluation (APACHE) II 41 scores were evaluated at the onset of sepsis. Organ dysfunction was reassessed over 28 days in the ICU using organ-specific SOFA scores to monitor morbidity. Organ failure (as assessed by SOFA scores), organ support requirements (mechanical ventilation, vasopressor therapy and renal-replacement therapy) and the length of ICU stay were recorded as secondary outcomes. Several relevant laboratory values were recorded as secondary variables. All of the relevant clinical data were obtained from the electronic patient record system (IntelliSpace Critical Care and Anesthesia (ICCA); Philips Healthcare, Andover, Massachusetts, USA); all medical records could be found in this system. Information regarding medical history and preexisting medical history were completed by examining previous physicians' notes, through anamnestic questionnaires and by consulting each patient's family doctor.
Statistical analyses. The statistical analyses were performed using Statistica software (version 10; StatSoft, Tulsa, Oklahoma, USA). The significance of categorical variables was calculated using two-sided Fisher's exact or chi-square tests, as appropriate. Two continuous variables were compared using the Mann-Whitney test. We performed a multivariate Cox regression analysis to examine the impact of common medical conditions on survival; several covariates, including mortality predictors (age, SOFA, APACHE II) and potential confounders (gender, BMI, septic shock, infection type, recent surgical history) were included in this model. Time-to-event data were compared using the log-rank test from the Statistica package for Kaplan-Meier survival analysis. A power calculation was performed using the Statistica package for power analysis. A p-value of < 0.05 was considered statistically significant.