Prognostic determinants of community-acquired bloodstream infection in type 2 diabetic patients in ED

a b s t r a c t

Objectives: The objective of the study is to describe the epidemiology and outcome of community-acquired Bloodstream infection in type 2 Diabetic patients in emergency department (ED).

Methods: All patients admitted to the ED of the university hospital from June 2010 to June 2011 with a history of type 2 diabetes mellitus and microbiologically documented BSI were retrospectively enrolled. Demographic characteristics, Charlson Comorbidity Index, antibiotic therapy, clinical severity, microbiological etiology, and diabetes-related complications were recorded in a standardized form. The major outcome measure was 30-day survival. ?² Or Student t test was used for univariate analysis, and Cox proportional hazards models were used for multivariate analysis.

Results: Among 250 enrolled emergency patients with BSI, the overall 30-day mortality rate was 15.5%. Twenty-seven patients (10.7%) developed diabetic ketoacidosis (DKA), and 22 patients (8.8%) developed hyperosmolar hyperglycemic state. On univariate analysis, DKA rather than hyperosmolar hyperglycemic state was associated with adverse outcome. Other risk factors include higher mean glycated hemoglobin level, presence of underlying malignancy, long-term use of steroids, lower respiratory tract infection, and higher Charlson scores. Multivariate analysis identified 3 independent risk factors for early mortality when severity, comorbidity, age, and sex were under control: DKA (hazard ratio, 3.89; 95% confidence interval, 1.6-8.9), inappropriate antibiotics (2.25, 1.05-4.82), and chronic use of steroid (3.89, 1.1-13.2). Conclusion: In type 2 diabetic patients with BSI, a substantial proportion of patients developed DKA. This condition was probably underrecognized by clinicians and constituted an independent risk factor for short-term mortality. Other identified risk factors are potentially correctable and may allow preventive efforts to individuals at greatest potential benefit.

Introduction

Patients with diabetes may carry a higher risk of bloodstream infection (BSI) [1-3]. Despite promising advances in the care of critically ill patients with severe sepsis, case fatality for diabetic patients rushed to the emergency department (ED) is 10% to 40% [1,2]. Diabetic patients with BSI are generally thought to carry a higher case fatality rate than nondiabetic patients due to risk of metabolic derangement, decreased immunity, or microangiopathic changes of vital organs.

? Financial disclosure and conflict interest: None declare.

* Corresponding author. Department of Emergency Medicine, National Taiwan Uni- versity Hospital Yunlin Branch, No. 579, Yunlin Road, Douliou 640, Taiwan. Tel.: + 886 2 23565926; fax: +886 2 23223150.

E-mail address: [email protected] (C.-C. Lee).

To date, most epidemiological studies were largely comparative in design concerning the epidemiology and outcome of BSI between diabetic and nondiabetic patients [4-10]. There are few studies that examine the outcome predictors specifically among the diabetic patients in the ED. Efforts to identify the comorbid conditions that are associated with an increased risk of mortality are important because high-risk characteristics could be used as screening criteria by clinicians to focus their attention on preventing mortality in the ED.

Research design and methods

Study design and setting

From June 1, 2010, to June 1, 2011, we conducted a retrospective observational study in the ED at a medical center. The medical center

http://dx.doi.org/10.1016/j.ajem.2014.08.071

was a primary and tertiary care center and has 2100 beds with annual ED census of more than 110000 visits.

Selection of participants and data collection

The decision of obtaining blood culture was based on the clinical manifestations of Systemic Inflammatory Response Syndrome or clinical indication of severe infection such as pneumonia, cellulitis, abscess formation, cholecystitis, or pyelonephritis in the absence of SIRS. Patients were included in the study if they had clinically significant true bacteremia, which was defined as positive blood cultures for at least 2 sets at separate sites or 1 set for Gram-negative bacterial pathogen or 1 set for gram-positive pathogen in a patient with intravascular device and clinical compatibility. Coagulase-negative staphylococci and other common skin flora isolated in single blood culture without clinical risk and compatible disease course were viewed as contamination and excluded. Patients younger than 15 years were not included in the study. All eligible patients were reevaluated by study physicians using a structured recording form. The admission course and medical records were also followed. For all eligible patients, the following data were collected retrospectively: demographic characteristics, preexisting comorbidities, exposure to indwelling catheters, initial vital signs and laboratory tests results, admission and final discharge diagnoses, and microorganisms isolated from the blood cultures. Patient outcomes were defined as 30-day mortality acquired either by hospital records or telephone interviews.

Identification of patients with diabetes

We ascertained the presence of diabetes in patients with a concurrent medical history of diabetes by searching the hospital-based medical records database for earlier hospitalizations with diabetes or earlier pre- scriptions for insulin or an oral antidiabetic drug. Those without history of diabetes but had diabetes newly diagnosed during the admission course were also eligible for analysis. Hospital records were reviewed for these patients using a detailed standardized form for diabetes- related complications.

We classified a diabetic patient as having type 1 diabetes if they were younger than 40 years at diagnosis and were treated with insulin in monotherapy. Having type 2 diabetes was classified if they were treated by diet alone or had been treated with oral antidiabetics or if they were older than 40 years at diagnosis, regardless of treatment.

Glycated hemoglobin (HbA1c) was evaluated from the value closest to the day of ED admission [10]. Diabetic ketoacidosis (DKA) was defined as presence of hyperglycemia, ketonuria or ketonemia, and metabolic acidosis with blood bicarbonate less than 18 mEq/L. Hyperosmolar hyperglycemic state (HHS) was defined as presence of hyperglycemia 600 mg/dL, hyperosmolarity, and absence of significant amounts of ketone bodies in the urine. Diabetic microvascular complications included nephropathy, retinopathy, or neuropathy, as documented in the hospital record and with no other suspected etiology than diabetes. Macrovascular complications included history of transient ischemic attack, stroke, angina pectoris, myocardial infarction, coronary or peripheral artery revascularization, claudication, and ulcer or amputa- tion on the lower extremities as a result of ischemia.

Patients with history of type 1 diabetes, incomplete data recording, or hospitalization history within 48 hours of ED admission were excluded for analysis. We used the first episodes of BSI for analysis in those with repeated BSI episodes during the study period. Finally, a total of 250 patients met the inclusion criteria.

Definitions

Hypothermia was defined as 35.5?C, hypotension as systolic blood pressure less than 90 mm Hg, altered Conscious levels as Glasgow Coma Scale less than or equal to 12, hypoxemia as pulse oxygenation less

than 90%, acute renal failure as creatinine level greater than 2 mg/dL in the absence of preexisting renal disease, or creatinine greater than double of the previous creatinine level in the presence of chronic renal disease. We used modified versions of previously defined consensus criteria for defining the sepsis syndromes. Systemic inflammatory response syndrome was defined as the presence of 2 or more of the following: (a) tachycardia (heart rate, 90); (b) tachypnea (respiratory rate, 20) or hypoxia (oxygen saturation b 95% or need for oxygen supplementation); (c) hyperthermia 38?C or hypothermia 35.5?C; and (d) leukocytosis (white blood cell count, 15000 cells/mm³ or bands 10%) or leukopenia (white blood cell count, b 4000 cells/mm³). Severe sepsis was defined as 2 or more criteria for SIRS plus organ dysfunction (altered mental status; pulmonary, oxygen saturation b 90%, acute renal failure, and abnormal liver enzyme in the absence of explainable liver disease). Septic shock was defined as severe sepsis plus hypotension (systolic blood pressure, b 90 mm Hg). The presence and source of a focal infection were classified by final discharge diagnosis as lower respiratory tract infection, urinary tract infection, biliary tract infection, Liver abscess, skin and musculoskeletal infection, other intraAbdominal infection, and miscellaneous. Those without a localized bacteremic source after extensive admission workup were classified as primary bacteremia. Patients were followed up until discharge. Long-term use of steroid was defined as use of steroid longer than 1 month in the past 3 months. The hospital’s microbiology laboratory determined antimicrobial susceptibility of isolates. Therapy was considered adequate when at least 1 effective drug was included in the empirical antibiotic treatment within the first 24 hours of the admission to the ED and the dose and pattern of administration were in accordance with current medical standards. The surveillance strategy definitions did not change over the study period.

Statistical analysis

For univariate analysis, the comparison between groups was done using the ?² test or Fisher exact test for category variables and Student t or Mann-Whitney U test for continuous variables. Significant factors (P b .20) identified on univariate analysis were considered as potential covariates and further introduced into Cox proportional hazard regres- sion models using backward elimination Selection process. Considering the effect of sex and age on survival, these 2 factors were forced into the final model as independent variables. On Cox Regression modeling, the outcome was time to death, and the censoring event was survival at the end of the 30-day follow-up. Log-log survival plots were constructed to assess the proportionality assumption underlying the Cox models. Effects of each independent predictor on outcome are expressed as adjusted Hazard ratios (HRs) with 95% confidence inter- vals (CIs). To determine the survival difference between patients with and without DKA or HHS, survival analysis was performed using Kaplan-Meier survival curves, and the significance was tested using the log-rank tests. All tests were 2 tailed, and P b .05 was considered statistically significant. All data were analyzed with SPSS software for Windows (Release 13.0; SPSS, Inc, Chicago, IL).

Results

Patient population

During the 1-year study, 95630 patients were admitted to the adult ED. A total of 964 episodes had clinical significant positive results for blood culture, of which 285 had a confirmed history of diabetes. Of the 285 episodes, 14 episodes were excluded as repeated infection, 9 excluded as Nosocomial infection, 5 excluded as type 1 diabetes, and

7 excluded as incomplete medical records of diabetes-associated comorbidities. Finally, a total of 250 patients entered the analysis. The distribution of age, sex, and 30-day mortality of the excluded 35 cases with incomplete data were comparable with those of the included patients.

Table 1

Diabetes-related complications and conditions of Glycemic control of 250 patients with type 2 diabetes and community-acquired BSI, a comparison between 30-day survivor

and nonsurvivor

we found patients with DKA had a lower survival probability than those without the condition (Fig. 1A; log-rank test, P = .007), whereas patients with HHS did not have a statistically significant survival differ-

ence as compared with the rest of the patients (Fig. 1B; log-rank test,

Characteristics	Total	Survivor	Nonsurvivor	P
	(n = 250)	(n = 220)	(n = 30)
HbA1c	8.18 +- 1.91	8.02 +- 1.92	9.11 +- 1.58^a	.021
Any microvascular	63 (30.4%)	77 (35.0%)	10 (33.3%)	.857
complication
Any macrovascular	99 (39.6%)	87 (39.5%)	12 (40.0%)	.962
complication
Blood glucose	268.6 +- 197. 7	263.0 +- 195.8	301.3 +- 209.3	.342
DKA on admission	27 (10.8%)	19 (8.6%)	8 (26.7%)^a	.007
HHS on admission	5 (2.0%)	4 (1.8%)	1 (3.3%)	.581

P = .559). We further compared the treatment modalities for these 2 Hyperglycemic crises in our patients. In patients with DKA, all were treated with intermittent insulin, base buffer, and fluid resuscitation. In patients with HHS, 4 (80%) received continuous insulin because of extremely high Serum glucose concentrations (N 800 mg/dL), and others received intermittent insulin and fluid therapy.

5.3. Comorbidity and treatment

^a Significant higher value (P b .05).

Diabetes-associated comorbidities

Table 1 shows the characteristics of the 250 patients with type

2 diabetes. Microvascular and macrovascular complications were present in 30% and 34%, respectively. In 32 patients (15.8%), ketoacidosis was present (mean bicarbonate, 8.1 +- 1.16 mEq/L; Reference interval, 21.5-26.5 mEq/L). Compared with survivors, nonsurvivor group has a higher mean HbA1c level. The proportion of patients with microvascular or macrovascular complications and the level of blood glucose on admis- sion were comparable between survivors and nonsurvivors. We further analyzed the effect of the 2 major acute metabolic complications, DKA and HHS, on short-term survival. On Kaplan-Meier survival analysis,

Fig. 1. Kaplan-Meier survival curves comparing survival in patients with DKA (A) or HHS

(B) with survival in those without those complications. The dotted line represents DKA

(A) or HHS (B), and the solid line represents patients without such complication.

Baseline Patient demographic characteristics and organ dysfunctions significantly associated with 30-day mortality by univariable comparison are given in Table 2. Comparison between survivors and nonsurvivors showed significant differences for Charlson comorbidity index greater than 6, long-term use of steroid, underlying liver cirrhosis, and nonhematogenous malignancy. Of 250 patients, 33 (13.2%) received no appropriate antimicrobial therapy within 24 hours of ED admission.

Microbiologic findings and Infection sites

Major infection sites and univariate analysis of their effects on 30-day survival were summarized in Table 3. On the frequency of the source of infection, urinary tract infection was the most common source (36.4%), followed by biliary tract infection (12.8%), liver abscess (11.1%), primary bacteremia (10.4%), and skin and musculoskeletal infection (10.1%). Patients with lower respiratory tract infection (30-day case fatality rate, 30%; P b .001) were associated with significant poorer outcome on univariate analysis, whereas patients with urinary tract infection were associated with better outcome (0; P b .001). Liver abscess tended to be associated with better outcome (P = .057), although the significance was borderline.

On the microbiology of the culture results, 230 patients (92.0%) had monomicrobial, and 20 (8.0%) had polymicrobial BSIs. Gram-negative bacteria accounted for most of the isolates (189/250, 75.6%), whereas Gram-positive bacteria represented only one-third of Gram-negative pathogens (64/250, 25.9%). Anaerobes of all microorganisms were identi- fied in a minor proportion of patients (9/250, 3.6%). There was no statis- tically survival difference among these different patterns of blood cultures isolates.

Leading pathogens were Escherichia coli (n = 104, 41.6%), Klebsiella pneumoniae (n = 53, 21.2%), Staphylococcus aureus (n = 20, 8.0%), and viridans Streptococcus (n = 14, 5.6%).

Table 2

Demographic characteristics and underlying comorbid condition of type 2 diabetic patients with community-acquired BSI, a comparison between 30-day survivor and nonsurvivor

	Total	Survivor	Nonsurvivor	P
	(n = 250)	(n = 220)	(n = 30)
Age (mean +- SD)	68.03 +- 11.26	68.07 +- 10.99	67.08 +- 13.29		.904
Age, >=65 y	153 (61.2%)	133 (60.5%)	20 (66.7%)		.512
Sex (male %)	119 (47.6%)	103 (46.8%)	16 (53.3%)		.503
Comorbidities Liver cirrhosis	24 (9.6%)	18 (8.2%)	6 (20.0%)^a		.050
Old myocardial infarction	41 (16.4%)	36 (16.4%)	5 (16.7%)		.966
End-stage renal disease	16 (6.4%)	11 (6.4%)	5 (16.7%)^a		.03
Hemiplegic stroke	48 (19.2%)	44 (20.0%)	4 (13.3%)		.384
Nonhematologic	32 (12.8%)	23 (10.5%)	9 (30.0%)^a		.006
malignancies
Hematologic malignancies	7 (4.0%)	7 (100%)	0 (0%)	1.000
intravenous catheters	23 (9.2%)	19 (8.6%)	4 (13.3%)	.496
Prior corticosteroid use	8 (3.4%)	4 (1.7%)	4 (12.5%)^a	.008
Charlson score >=6	13 (4.3%)	8 (3.6%)	5 (16.7%)^a	.012
Inappropriate antibiotic use	33 (13.2%)	22 (10.0%)	11 (36.7%)	b .001

^a Significant higher incidence (P b .05).

outcome prediction“>Table 3

Major sources of infection of type 2 diabetic patients with community-acquired BSI, a comparison between 30-day survivor and nonsurvivor

Table 5

Independent risk factors associated with 30-day mortality in type 2 diabetic patients with community-acquired BSI

Variables	Total	Survivor	Nonsurvivor	P	Predictors	HR (95% CI)
	(n = 250)	(n = 220)	(n = 30)		Steroid	3.89 (1.1-13.2)
Urinary tract infection	91 (36.4%)	91 (41.4%)^a	0 (0.0%)	b .001	DKA	3.80 (1.6-8.9)
Primary bacteremia^b	26 (10.4%)	22 (10.0%)	4 (13.3%)	.530	Inappropriate antibiotics	2.25 (1.05-4.82)
Biliary tract infection	32 (12.8%)	29 (13.2%)	3 (10.0%)	.777	Charlson comorbidity score >=6	2.46 (1.0-6.1)
Liver abscess	24 (11.1%)	24 (10.9%)	0 (0%)	.057	Classification of severity of sepsis	2.32 (1.6-3.4)
Skin and musculoskeletal infection	22 (10.1%)	19 (8.6%)	3 (10.0%)	.735
Lower respiratory tract infection	20 (9.2%)	11 (5.0%)	9 (30.0%)^a	b .001

^a Significant higher incidence (P b .05).

^b Sources of unknown origin.

Organ dysfunctions

On ED presentation, the most frequent organ dysfunctions were Neurologic dysfunction (18.0%), septic shock (16.0%), acute renal failure (11.1%), and respiratory distress (8.1%). Patients with any type of organ failure were associated with increased 30-day mortality. The overall 30-day all-cause mortality was 15.5%. Related results were summarized in Table 4.

Outcome prediction

Using Cox proportional hazards regression model analysis, 3 variables were identified as being independently associated with death, adjusting for clinical severity of sepsis, comorbidity, age, and sex. Age (HR, 1.01; 95% CI, 0.98-1.04) and sex (1.21, 0.57-2.56) did not significantly affect the Short-term outcome in either univariate or multivariate analysis. Long-term use of steroid was associated with the highest risk of early death, followed by DKA, inappropriate initial antibiotic treatment. Hazard ratio with 95% CI of each significant independent predictor was presented in Table 5. Source of infection significantly associated with 30-day mortality on univariate analysis was explained by the severity of sepsis and underlying comorbidity and thus became insignifi- cant in the multivariate model.

Discussion

In this analysis of 250 diabetic patients with community-acquired BSI, we identified long-term use of steroid, presence of DKA, and inappropriate antibiotics use, controlling for sex, age, clinical severity, and underlying comorbidity, as the independent predictors for 30-day mortality. Compared with limited published data on diabetic patients with BSI, our results agree with previous findings on the trend of more acquisition in the community, more urinary tract or soft tissue infection. Our results also agreed with previous observations that a

Table 4

Clinical and laboratory manifestations of type 2 diabetic patients with community-acquired BSI, a comparison between 30-day survivor and nonsurvivor

Variables Total Survivor Nonsurvivor P

(n = 250) (n = 220) (n = 30)

Altered consciousness (GCS <=12) 45 (18.0%) 30 (13.6%) 15 (50.0%)^a b .001

substantial proportion of type 2 diabetic patients may develop DKA as well as HHS state [7].

Previous studies showed several aspects of immunity are altered in patients with diabetes including antioxidant systems, polymorphonuclear leukocyte function, leukocyte adherence, chemotaxis, and phagocytosis [11-13]. These in vitro findings, however, have not been fully confirmed in clinical studies. Our results showed that patients with poorer outcome had a significantly higher serum HbA1c levels, providing the clinical evidence that glycemic exposure is associated with poorer outcome among type 2 diabetic patients with community-acquired BSI.

In population studies, the estimated mortality rate for DKA (4%-10%) is lower than that for HHS (10%-50%) [14]. In this study, however, we found that DKA rather than HHS was independently associated with adverse outcome. The contrast is most likely due to the exclusion of young type I diabetic patients in this study. Diabetic ketoacidosis used to be considered as a hallmark feature that would differentiate individuals with type 1 diabetes mellitus from those with type 2 DM. However, many clinical studies have pointed out that DKA also occurs in type 2 DM patients, and half of DKA episodes are accompanied with identifiable precipitating factors [15-17]. Infection was shown to be a major triggering event (40%) in a previous series [17]. We further clarify that community-acquired bacteremia may trigger DKA in approximately 10% of type 2 diabetic patients. When DKA occurs, it was an independent risk factor for early mortality. Previous series reported an overall case fatality of 3.9% to 6.8% among both type 1 and type 2 DM patients with DKA, of which most deaths arose from severe infection or myocardial infarction [15-19]. Our data showed a high case fatality rate of 30%, which may be explained by the facts that BSI is the more Severe form of infection and the relatively young type 1 DM patients were excluded in this analysis. Besides, undertreatment of DKA was another factor contributing to the poor prognosis in this cohort. Although the diagnostic criteria for diabetic acidosis were all met in these patients, most of the primary care physicians responded with administration of fluid, base, and intermittent insulin, rather than continuous insulin therapy. This reflects the clinicians’ underrecognition of DKA in type 2 diabetic patients. The clinicians may ascribe the hyper- glycemia to the stress response and metabolic acidosis to the results of sepsis-related lactic acidemia. Typically, HHS was associated with higher morbidity because of age and associated comorbidity. In our Cox regres- sion model, severity of sepsis and age-related comorbidity have been controlled, which might explain HHS was not identified as an indepen- dent risk factor for mortality in our study.

The infectious complications associated with glucocorticoids use are

well known. A meta-analysis of 71 trials involving more than 2000 glucocorticoid-treated patients with nonmalignant pathology showed

Hypoxemia on presentation

(SpO2 b 90%)

Acute renal failure (creatinine N 2 mg/dL)

Severity

17 (8.1%) 10 (5.7%) 7 (21.9%)^a .007

22 (11.1%) 12 (5.5%) 10 (33.3%)^a b .001

that the relative risk for infection was 1.5 times that of controls [19]. However, previous series of BSI comprising both diabetic and nondiabetic patients did not identify long-term use of steroids as an independent risk factor for death [20,21]. The significant adverse effects of long-term

Simple BSI 13 (5.2%) 11 (5.9%)
SIRS 143 (57.2%) 137 (62.3%)^a

3. Severe sepsis	54 (21.6%)	43 (19.5%)	15 (36.7%)^a	.033
4. Septic shock	40 (16.0%)	27 (12.3%)	13 (43.3%)^a	b .001

0 (0.00%) .171

6 (20.0%) b .001

steroid use shown in our study may be due to the stricter criteria for patient selection, which required longer than 1 month’s duration within the 3 months of ED admission. Besides, 2 of 4 fatal cases of long-term

Abbreviations: GCS, Glasgow Coma Scale; SpO2, oxygen saturation as measured by pulse oximetry.

^a Significant higher incidence (P b .05).

steroid users had signs of Adrenal insufficiency, and 1 had DKA. Relative adrenal insufficiency has recently been identified as an important prog- nostic factor in patients with severe sepsis and septic shock [22]. Long-

term use of steroid in diabetic patients may precipitate not only the adrenal insufficiency but insulin resistance as well and, therefore, with worse outcome as compared with nondiabetic patients. Of the 8 patients with long-term use of steroids, 4 had chronic obstructive pulmonary disease or asthma, 2 had chronic arthritis, 1 had autoimmune disease, and1 had myasthenia gravis.

Inappropriate antibiotic treatment time has been identified as important predictors for BSI-related mortality in some previous series [20,23,24]. Our findings, again, provide another evidence of the impor- tance of timely Antimicrobial treatment on survival for the diabetic patient population.

Finally, all 3 independent risk factors identified in this study are potentially correctable. It may be worthwhile to further study the benefits of aggressive screening and treatment of DKA among diabetic patients with severe infection and adrenal insufficiency and insulin resistance among diabetic patient with a history of long-term steroid use. Most inappropriate empirical use of antibiotics use arose from the community-acquired resistance strain or Polymicrobial infection; there- fore, the development of a clinical prediction rule for these 2 situations may improve the choice of Empirical antibiotics treatment and improve the survival.

There are both strengths and limitations for our study. We used retrospective ED-based registries with complete follow-up data. We used multivariate analysis, which limit the potential confounding effects. Nevertheless, generalizing the results to other settings should be considered in the context of several limitations. First, we did not include our patients based on strictly defined SIRS criteria. systemic inflammatory response syndrome criteria are not exactly the working criteria for clinicians to identify severe infection in the real world. There- fore, we modified the SIRS criteria to include all possible severe infec- tions in our study. Second, it is possible that patients with diabetes have a lower threshold of Bacterial translocation into the bloodstream during less severe BSI that we did not include in our study. Third, the patients with long-term use of steroid was identified as a group suscepti- ble to high mortality, but this result should be interpreted with more caution due to limited patient number. Lastly, we did not use 2 abstractors to abstract the data. However, all data were abstracted from electronic medical records, and most abstracted variables were objective laboratory data or Diagnostic codes.

In conclusion, we found that DKA, long-term use of steroid, and inappropriate empirical antimicrobial treatment are independently associated with early mortality among type 2 diabetic patients with community-acquired BSI. The poor prognosis of patients with DKA may be closely related to the underrecognition and undertreatment of meta- bolic derangement. Besides, we provided the clinical evidence that glyce- mic exposure is associated with poorer outcome among type 2 diabetic patients with community-acquired BSI. All these poor prognostic factors are potentially correctable and warrant the awareness of clinicians.

References

AJEM

Prognostic determinants of community-acquired bloodstream infection in type 2 diabetic patients in ED

a b s t r a c t

Introduction

Research design and methods

Study design and setting

Selection of participants and data collection

Identification of patients with diabetes

Definitions

Statistical analysis

Results

Patient population

Diabetes-associated comorbidities

Microbiologic findings and Infection sites

Organ dysfunctions

Outcome prediction

Discussion

Simple BSI 13 (5.2%) 11 (5.9%)
SIRS 143 (57.2%) 137 (62.3%)^a

Bryan CS, Reynolds KL, Metzger WT. Bacteremia in diabetic patients: comparison of incidence and mortality with nondiabetic patients. Diabetes Care 1985;8:244-9.

Leave a Reply Cancel reply

Prognostic determinants of community-acquired bloodstream infection in type 2 diabetic patients in ED

a b s t r a c t

Introduction

Research design and methods

Study design and setting

Selection of participants and data collection

Identification of patients with diabetes

Definitions

Statistical analysis

Results

Patient population

Diabetes-associated comorbidities

Microbiologic findings and Infection sites

Organ dysfunctions

Outcome prediction

Discussion

Simple BSI 13 (5.2%) 11 (5.9%) SIRS 143 (57.2%) 137 (62.3%)a

Bryan CS, Reynolds KL, Metzger WT. Bacteremia in diabetic patients: comparison of incidence and mortality with nondiabetic patients. Diabetes Care 1985;8:244-9.

Leave a Reply Cancel reply

Simple BSI 13 (5.2%) 11 (5.9%)
SIRS 143 (57.2%) 137 (62.3%)^a