Antibiotic Use and Antimicrobial Resistance of Surgical Patients with Peritonitis at a Tertiary Referral Hospital in Rwanda.

BACKGROUND: There is growing recognition of the worsening problem of antibiotic resistance and the need for antibiotic stewardship in low-resource settings. The aim of this study was to describe antibiotic use and antimicrobial resistance in patients undergoing surgery for peritonitis at a Rwandan referral hospital. PATIENTS AND METHODS: All surgical patients with peritonitis at a Rwandan referral hospital were enrolled. Prospective data were collected on epidemiology, clinical features, interventions, and outcomes. Antibiotic agents were prescribed and cultures were collected according to surgeon discretion. High risk for antibiotic treatment failure or death was defined as patients with severe sepsis, older than 70 years of age, tumor, or operating room delay more than 24 hours from hospital admission. Logistic regression was used to determine factors associated with high risk of antibiotic treatment failure or death. RESULTS: Over a six-month period, 280 patients underwent operation for peritonitis; 79 patients were excluded because no infectious etiology was identified at operation. Data on antibiotic usage were available for 165 patients. The most common diagnoses were intestinal obstruction (n = 43) and appendicitis (n = 36). Most patients received antibiotic agents, the most of of which being third-generation cephalosporins (n = 149; 90%) and metronidazole (n = 140; 85%). The mean duration of antibiotics was 5.1 days (range: 0-14). Overall, 80 (54%) patients were high-risk for antibiotic treatment failure or death. Risk for antibiotic treatment failure or death was associated with localized peritonitis (p = 0.001) and high American Society of Anesthesiologist score (p = 0.003). Cultures were collected from 33 patients and seven patients had an organism isolated. Escherichia coli was identified in in five surgical specimens and two 2 urine cultures. All Escherichia coli specimens showed resistance to cephalosporins. CONCLUSIONS: Broad antibiotic coverage with third-generation cephalosporins and metronidazole is common in Rwandan surgical patients with peritonitis. Areas for improvement should focus on choice and duration of antibiotic agents, tailored to underlying diagnosis and risk factors for antibiotic treatment failure or death. More data are needed on antibiotic resistance patterns to guide antimicrobial therapy.

Critical Care Management of Peritonitis in a Low-Resource Setting.

BACKGROUND: Management of critically ill patients is challenging in a low-resource setting. In Rwanda, peritonitis is a common surgical condition where patients often present late, with advanced disease. We aim to describe critical care management of patients with peritonitis in Rwanda. METHODS: Data were collected at a tertiary referral hospital in Rwanda on patients undergoing operation for peritonitis over a 6-month period. Data included epidemiology, hospital course and outcomes. Patients requiring admission to the intensive care unit (ICU) were compared with those not requiring ICU admission using Chi-square and Wilcoxon rank-sum test. RESULTS: Over a 6-month period, 280 patients were operated for peritonitis. Of these, 46 (16.4%) were admitted to the ICU. The most common diagnoses were intestinal obstruction (N = 17, 37.0%) and typhoid intestinal perforation (N = 6, 13.0%). Thirty-nine (89%) patients had sepsis. The median American Society of Anesthesiologist score was 3 (range 2-4), and the median Surgical Apgar Score was 4 (range 0-6). Twenty-four (52.2%) patients required vasopressors, with dopamine and adrenaline being the only vasopressors available. Patients admitted to the ICU, compared with non-critically ill patients, were more likely to have major complications (80.4 vs. 14%, p < 0.001), unplanned reoperation (28 vs. 10%, p < 0.001) and death (72 vs. 8%, p < 0.001). CONCLUSION: Patients with peritonitis admitted to the ICU commonly presented with features of sepsis. Due to limited resources in this setting, interventions are primarily supportive with intravenous fluids, intravenous antibiotics, ventilator support and vasopressors. Morbidity and mortality remain high in this patient population.

Conscious status is associated with the likelihood of trauma centre care and mortality in patients with moderate-to-severe traumatic brain injury.

OBJECTIVE: To assess the relationship between The International Classification of Diseases, Ninth Revision, Clinical Modification-derived conscious status and mortality rates in trauma centres (TC) vs. non-trauma centres (NTC). METHODS: Patients in the 2006-2011 Nationwide Emergency Department Sample meeting, The Centers for Disease Control and Prevention criteria for traumatic brain injury (TBI), with head/neck Abbreviated Injury Scale (AIS) scores ≥3 were included. Loss of consciousness (LOC) was computed for each patient. Primary outcomes included treatment at a level I/II TC vs. NTC and in-hospital mortality. We compared logistic regression models controlling for patient demographics, injury characteristics, and AIS score with identical models that also included LOC. RESULTS: Of 66,636 patients with isolated TBI identified, 15,761 (23.6%) had missing LOC status. Among the remaining 50,875 patients, 59.0% were male, 54.0% were ≥65 years old, 56.7% were treated in TCs, and 27.3% had extended LOC. Patients with extended LOC were more likely to be treated in TCs vs. those with no/brief LOC (71.1% vs. 51.4%, p < 0.001). Among patients aged <65, TC treatment was associated with increased odds of mortality [Adjusted Odds Ratio (AOR) 1.79]; accounting for LOC substantially mitigated this relationship [AOR 1.27]. Similar findings were observed among older patients, with reduced effect size. CONCLUSION: Extended LOC was associated with TC treatment and mortality. Accounting for patient LOC reduced the differential odds of mortality comparing TCs vs. NTCs by 60%. Research assessing TBI outcomes using administrative data should include measures of consciousness.

Validation of the Surgical Apgar Score After Laparotomy at a Tertiary Referral Hospital in Rwanda.

BACKGROUND: The surgical Apgar score (SAS) has demonstrated utility in predicting postoperative outcomes in a variety of surgical disciplines. However, there has not been a study validating the utility of the SAS in surgical patients in low-income countries. We conducted a prospective, observational study of patients undergoing laparotomy at a tertiary referral hospital in Rwanda and determined the ability of SAS to predict inpatient major complications and mortality. METHODS: All adult patients undergoing laparotomy in a tertiary referral hospital in Rwanda from October 2014 to January 2015 were included. Data were collected on patient and operative characteristics. SAS was calculated and patients were divided into four SAS categories. Primary outcomes were in-hospital mortality and major complications. Rates and odds of in-hospital mortality and major complications were examined across the four SAS categories. Logistic regression modeling and calculation of c-statistics was used to determine the discriminative ability of SAS. RESULTS: 218 patients underwent laparotomy during the study period. One hundred and forty-three (65.6%) were male, and the median age was 34 years (IQR 27-51 years). The most common diagnosis was intestinal obstruction (97 [44.5%]). A high proportion of patients (170 [78%]) underwent emergency surgery. Thirty-nine (18.3%) patients died, and 61 (28.6%) patients had a major complication. In-hospital mortality occurred in 25 (50%) patients in the high-risk group, 12 (16%) in the moderate-risk group, 2 (3%) in the mild-risk group and there were no deaths in the low-risk group. Major complications occurred in 32 (64%) patients in the high-risk group, 22 (29%) in the moderate-risk group, 7 (11%) in the mild-risk group and there were no complications in the low-risk group. SAS was a good predictor of postoperative mortality (c-statistic 0.79) and major complications (c-statistic 0.75). CONCLUSIONS: SAS can be used to predict in-hospital mortality and major complications after laparotomy in a Rwandan tertiary referral hospital.

Conscious status predicts mortality among patients with isolated traumatic brain injury in administrative data.

BACKGROUND: Outcome studies in trauma using administrative data traditionally employ anatomy-based definitions of injury severity; however, physiologic factors, including consciousness, may correlate with outcomes. We examined whether accounting for conscious status in administrative data improved mortality prediction among patients with moderate to severe TBI. METHODS: Patients meeting Centers for Disease Control and Prevention (CDC) guidelines for TBI in the 2006 to 2011 Nationwide Emergency Department Sample were identified. Patients were dichotomized as having no/brief loss of consciousness (LOC) vs extended LOC greater than 1 hour using International Classification of Diseases, Ninth Revision (ICD-9) fifth digit modifiers. Receiver operating curves compared the ability of logistic regression to predict mortality in models that included LOC vs models that did not. RESULTS: Overall, 98,397 individuals met criteria, of whom 25.8% had extended LOC. In univariate analysis, AIS alone predicted mortality in 69.6% of patients (area under receiver operating characteristic curve .696, 95% CI .689 to .702), extended LOC alone predicted mortality in 76.8% (AUROC .768, 95% CI .764 to .773), and a combination of AIS and extended LOC predicted mortality in 82.6% of cases (AUROC .826, 95% CI .821 to .830). Similar differences were observed in best-fit models. CONCLUSIONS: Accounting for LOC along with anatomical measures of injury severity improves mortality prediction among patients with moderate/severe TBI in administrative datasets. Further work is warranted to determine whether other physiological measures may also improve prediction across a variety of injury types.

Peritonitis in Rwanda: Epidemiology and risk factors for morbidity and mortality.

BACKGROUND: Few studies discuss causes and outcomes of peritonitis in low-income settings. This study describes epidemiology of patients with peritonitis at a Rwandan referral hospital. Identification of risk factors associated with mortality and unplanned reoperation could improve management of peritonitis. METHODS: Data were collected on demographics, clinical presentation, operative findings, and outcomes for all patients with peritonitis. Multivariate regression analysis identified factors associated with in-hospital mortality and unplanned reoperation. RESULTS: A total of 280 patients presented with peritonitis over a 6-month period. Causes of peritonitis were complications of intestinal obstruction (39%) and appendicitis (17%). Thirty-six (13%) patients required unplanned reoperation, and in-hospital mortality was 17%. Factors associated with increased odds of in-hospital mortality were unplanned reoperation (adjusted odds ratio 34.12), vasopressor use (adjusted odds ratio 24.91), abnormal white blood cell count (adjusted odds ratio 12.6), intensive care unit admission (adjusted odds ratio 9.06), and American Society of Anesthesiologist score ≥3 (adjusted odds ratio 7.80). Factors associated with increased odds of unplanned reoperation included typhoid perforation (adjusted odds ratio 5.92) and hypoxia on admission (adjusted odds ratio 3.82). CONCLUSION: Peritonitis in Rwanda presents with high morbidity and mortality. Minimizing delays in care is important, as many patients with intestinal obstruction present with features of peritonitis. A better understanding of patient care and management prior to arrival at the referral hospital is needed to identify areas for improvement at the health center and district hospital.

Tackling causes and costs of ED presentation for American football injuries: a population-level study.

BACKGROUND: American tackle football is the most popular high-energy impact sport in the United States, with approximately 9 million participants competing annually. Previous epidemiologic studies of football-related injuries have generally focused on specific geographic areas or pediatric age groups. Our study sought to examine patient characteristics and outcomes, including hospital charges, among athletes presenting for emergency department (ED) treatment of football-related injury across all age groups in a large nationally representative data set. METHODS: Patients presenting for ED treatment of injuries sustained playing American tackle football (identified using International Classification of Diseases, Ninth Revision, Clinical Modification code E007.0) from 2010 to 2011 were studied in the Nationwide Emergency Department Sample. Patient-specific injuries were identified using the primary International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis code and categorized by type and anatomical region. Standard descriptive methods examined patient demographics, diagnosis categories, and ED and inpatient outcomes and charges. RESULTS: During the study period 397363 football players presented for ED treatment, 95.8% of whom were male. Sprains/strains (25.6%), limb fractures (20.7%), and head injuries (including traumatic brain injury; 17.5%) represented the most presenting injuries. Overall, 97.9% of patients underwent routine ED discharge with 1.1% admitted directly and fewer than 11 patients in the 2-year study period dying prior to discharge. The proportion of admitted patients who required surgical interventions was 15.7%, of which 89.9% were orthopedic, 4.7% neurologic, and 2.6% abdominal. Among individuals admitted to inpatient care, mean hospital length of stay was 2.4days (95% confidence interval, 2.2-2.6) and 95.6% underwent routine discharge home. The mean total charge for all patients was $1941 (95% confidence interval, $1890-$1992) with substantial injury type-specific variability. Overall, at the US population, estimated total charges of $771299862 were incurred over the 2-year period. CONCLUSION: In this nationally representative sample, most ED-treated injuries associated with football were not acutely life threatening and very few required major therapeutic intervention. This study provides a cross-sectional overview of ED presentation for acute football-related injury across age groups at the population level in recent years. Longitudinal studies may be warranted to examine associations between the patterns of injury observed in this study and long-term outcomes among American tackle football players.

The use of an essay examination in evaluating medical students during the surgical clerkship.

BACKGROUND: Third-year medical students are graded according to subjective performance evaluations and standardized tests written by the National Board of Medical Examiners (NBME). Many "poor" standardized test takers believe the heavily weighted NBME does not evaluate their true fund of knowledge and would prefer a more open-ended forum to display their individualized learning experiences. Our study examined the use of an essay examination as part of the surgical clerkship evaluation. METHODS: We retrospectively examined the final surgical clerkship grades of 781 consecutive medical students enrolled in a large urban academic medical center from 2005 to 2011. We examined final grades with and without the inclusion of the essay examination for all students using a paired t test and then sought any relationship between the essay and NBME using Pearson correlations. RESULTS: Final average with and without the essay examination was 72.2% vs 71.3% (P < .001), with the essay examination increasing average scores by .4, 1.8, and 2.5 for those receiving high pass, pass, and fail, respectively. The essay decreased the average score for those earning an honors by .4. Essay scores were found to overall positively correlate with the NBME (r = .32, P < .001). CONCLUSIONS: The inclusion of an essay examination as part of the third-year surgical core clerkship final did increase the final grade a modest degree, especially for those with lower scores who may identify themselves as "poor" standardized test takers. A more open-ended forum may allow these students an opportunity to overcome this deficiency and reveal their true fund of surgical knowledge.

United States trends in thrombolysis for older adults with acute ischemic stroke.

OBJECTIVE: Thrombolysis for ischemic stroke has been increasing in the United States. We sought to investigate recent trends in thrombolysis use in older adults. METHODS: A retrospective, observational analysis of hospitalization data from the Nationwide Inpatient Sample (NIS) in 2005-2010 was performed. Older adults (≥65 years) admitted with a primary diagnosis of acute ischemic stroke were included. Trends in the population-based rates of thrombolysis and outcomes from the NIS were evaluated using the Cochran-Armitage test. RESULTS: Thrombolysis in older adult stroke patients increased from 1.7% to 5.4% (2005-2010; trend P<0.001). Large increases were observed among urban patients, urban hospitals, and high volume facilities. Individuals ≥85 years were less likely to receive thrombolysis than younger ages throughout the study period, although there was an increase from an odds ratio of 0.50 (95% CI: 0.44-0.57) to 0.75 (95% CI: 0.69-0.81) from 2005-2006 to 2009-2010 when compared to 65-74 year olds. For those receiving thrombolysis, no change was observed in intracerebral hemorrhage over time. In-hospital mortality rates did not change significantly over the study period for age subgroups and length of stay declined from 2005 to 2010 for the thrombolysis group (7.6 vs 7.0 days; trend P<0.001). CONCLUSIONS: Rates of thrombolysis in older adults progressively increased, especially in the oldest old. Increases were largely driven by urban and high volume hospitals.