Hypotension, bedridden, leukocytosis, thrombocytopenia and elevated serum creatinine predict mortality in geriatric patients with fever

Transcription

1 bs_bs_banner Geriatr Gerontol Int 2015; 15: ORIGINAL ARTICLE: EPIDEMIOLOGY, CLINICAL PRACTICE AND HEALTH Hypotension, bedridden, leukocytosis, thrombocytopenia and elevated serum creatinine predict mortality in geriatric patients with fever Min-Hsien Chung, 1,2, Feng-Yuan Chu, 1,3 Tzu-Meng Yang, 1 Hung-Jung Lin, 1,4,5 Jiann-Hwa Chen, 6,7 How-Ran Guo, 8,9 Si-Chon Vong, 1,2 Shih-Bin Su, 10,11,12, Chien-Cheng Huang 1,8,13,14 and Chien-Chin Hsu 1,4 1 Department of Emergency Medicine, Chi-Mei Medical Center, Tainan, 2 Department of Emergency Medicine, Chi-Mei Medical Center, Liouying, Tainan, 3 Department of Emergency Medicine, Chi-Mei Medical Center, Chiali, Tainan 10 Occupational Medicine and 12 Medical Research, Chi-Mei Medical Center, Departments of 4 Biotechnology, 11 Leisure, Recreation and Tourism Management and 13 Child Care and Education, Southern Taiwan University of Science and Technology, Tainan, 5 Department of Emergency Medicine, Taipei Medical University, 6 Department of Emergency Medicine, Cathay General Hospital, 7 Fu Jen Catholic University School of Medicine, Taipei, 8 Department of Environmental and Occupational Health, College of Medicine, National Cheng Kung University, 9 Department of Occupational and Environmental Medicine, National Cheng Kung University Hospital, and 14 Department of Emergency Medicine, Kuo General Hospital, Tainan, Taiwan Aim: The geriatric population (aged 65 years) accounts for 12 24% of all emergency department (ED) visits. Of them, 10% have a fever, 70 90% will be admitted and 7 10% of will die within a month. Therefore, mortality prediction and appropriate disposition after ED treatment are of great concern for geriatric patients with fever. We tried to identify independent mortality predictors of geriatric patients with fever, and combine these predictors to predict their mortality. Methods: We enrolled consecutive geriatric patients visiting the ED between 1 June and 21 July 2010 with the following criteria of fever: a tympanic temperature 37.2 C or a baseline temperature elevated 1.3 C. We used 30-day mortality as the primary end-point. Results: A total of 330 patients were enrolled. Hypotension, bedridden, leukocytosis, thrombocytopenia and serum creatinine >2 mg/dl, but not age, were independently associated with 30-day mortality. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) ranged from 18.2% to 90.9%, 34.7% to 100%, 9.0% to 100% and 94.5% to 98.2%, respectively, depending on how many predictors there were. Conclusions: The 30-day mortality increased with the number of independent mortality predictors. With at least four predictors, 100% of the patients died within 30 days. With none of the predictors, just 1.8% died. These findings might help physicians make decisions about geriatric patients with fever. Geriatr Gerontol Int 2015; 15: Keywords: elderly, emergency, fever, geriatric, mortality, prediction. Introduction Accepted for publication 15 June Correspondence: Dr Chien-Cheng Huang MD, Department of Emergency Medicine, Chi-Mei Medical Center, 901 Zhonghua Road, Yongkang Dist., Tainan City 710, Taiwan. chienchenghuang@yahoo.com.tw These authors contributed equally to this work as first authors. The geriatric population (aged 65 years) constituted 6.2% of the world population in 1992, and is expected to reach 20% by In the USA, 20% of the population is estimated to be aged more than 65 years by the year In Taiwan, because of a decrease in the fertility rate, medical advances and our comprehensive national healthcare system, the geriatric population (5.96% in 1989) reached 11.39% in September The elderly-to-child ratio in Taiwan also increased from 21.67% in 1989 to 79.15% in September As the elderly population steadily increases, the need for medical and healthcare resources also increases, especially emergency medical care. The elderly account for 12 24% of all emergency department (ED) patients: 4 10% have a fever, 70 90% are admitted and 7 10% die within a month doi: /ggi Japan Geriatrics Society

2 Mortality prediction for geriatric fever Evaluating geriatric patients with fever is challenging and can be time-consuming. 5 The fever could be attributable to infections, drug effects or multiple comorbidities. 5 However, 20 30% of older adults with an infection will present to the ED with a blunted fever response, in part, perhaps, because of a lower basal body temperature. 5 Other causes are changes in thermal homeostasis, decreased response to endogenous and exogenous pyrogens, decreased production and conservation of body heat, comorbidities, and drugs. 5 Many studies have documented the sensitivity and specificity of temperature for the presence of acute infection. Although a higher temperature cut-point is more specific for infection, a lower cut-point with a higher sensitivity is more important to prevent missing potentially serious infections. 5 The current consensus is that using a tympanic temperature 37.2 C or a baseline temperature elevated 1.3 C is most appropriate for defining fever threshold. 5,6 Mortality prediction and appropriate disposition (namely, discharge; treatment in the ED, a general ward or the intensive care unit [ICU]) are of great concern when managing geriatric patients with fever. Some studies have proposed decision rules based on mortality predictors to help emergency physicians make optimum management decisions for geriatric patients with fever. Most of the reported predictors, however, are impractical for the ED. For example, one study proposed that altered mental status, vomiting and a white blood cell count (WBC) band form >6% were independent predictors of bacteremia; 7 another study proposed that all patients had bacterial infection when they had fever ( 37.5 C), leukocytosis (WBC cells/mm 3 ) and bandemia (band form >6%); 8 and a third study proposed that an oral temperature 39.4 C, a respiratory rate 30/ min, leukocytosis (WBC cells/mm 3 ), infiltration on a chest radiograph and a pulse rate >120/min were associated with serious illness. 9 However, in clinical practice with geriatric patients, vomiting and infiltration on a chest radiograph are sometimes difficult to verify. Furthermore, because these three studies are now 20-years-old, many predicted factors they considered might be different because of changes in demography and geriatric care. We thus conceived a research question for developing a feasible and applicable contemporary prediction for decision-making in an ED. Materials and methods Study design, setting, population and selection of participants The present study was carried out in a 700-bed university-affiliated medical center in Taipei with a 40-bed ED staffed with board-certified emergency physicians; the center provides care for approximately patients per year. Approximately 33% of the ED patients are elderly. Consecutive geriatric patients who visited the ED between 1 June and 21 July 2010 were enrolled when they met one of the following criteria of fever: 5,6 a tympanic temperature 37.2 C or a baseline temperature elevated 1.3 C. The baseline temperature information came from the previous medical record, patient, caregiver or nursing home staff. Data collection and definition of variables Patients were prospectively selected in the ED. After the patient had been discharged, reviewers retrospectively collected missing information from the medical record or a telephone follow up in compliance with the policies approved by the hospital s Human Investigation Committee, which also approved the study protocol. The reviewers were blinded to knowledge of the patient s hospital course and outcomes. Information for a number of variables for each patient was recorded (Table 1). Any variable not included in the patient s medical history or physical examination reports was considered missing. The categorical variables used are generally acceptable in emergency care, critical care and geriatric care. The older adults were divided into three groups: young elderly (aged years), moderately elderly (aged years) and old elderly (aged 85 years). Altered mental status was defined as an acute mental status change (confusion, lethargy or coma). 7 Bedridden was defined as an Eastern Cooperative Oncology Group (ECOG) Performance Status (also called the World Health Organization or Zubrod score) Score of Hypotension was defined as a systolic blood pressure <90 mmhg. 11 Leukocytosis was defined as a white blood cell (WBC) count > cells/mm 3 and bandemia as >10% immature band forms. 11 Anemia was defined as hemoglobin <10 g/dl or hematocrit <30%. 12 Serum creatinine >2 mg/dl was a criterion of severe sepsis. 11 Thrombocytopenia was defined as a platelet count < /mm The infections diagnosed in the ED included urinary tract infection, lower respiratory tract infection, fever without a significant focus, intraabdominal infection, upper respiratory tract infection and skin or soft tissue infection. The clinical diagnosis was based on the attending physician s documentation, and on laboratory and image results (such as pneumonia on a chest radiograph, pyuria on a urinary analysis, abscess or intracranial hemorrhage on computed tomography, etc.). Overall, 350 geriatric patients from the ED met the criterion of fever. A total of 330 patients were enrolled after excluding 20 patients with insufficient data or transferred patients who had been treated in other hospitals. The enrolled patients were divided into two groups: (i) survival; and (ii) 30-day mortality based on 2014 Japan Geriatrics Society 835

3 M-H Chung et al. Table 1 Univariate analysis of variables of 330 elderly patients presenting with fever in the emergency department Variable Survival (n = 308) 30-Day mortality (n = 22) All (n = 330) P-value Mean age (years) 78.3 ± ± ± Age subgroup (%) Young elderly (65 74 years) Moderately elderly (75 84 years) Old elderly ( 85 years) Sex, male (%) Altered mental status, % Hypotension, SBP < 90 mmhg (%) Tachycardia, HR > 100/min (%) >0.950 Tachypnea, RR > 20/min (%) Tympanic temperature, >38.3 C (%) Medical history (%) Hypertension Diabetes Stroke Congestive heart failure Cancer Nursing home resident Bedridden Nasogastric feeding Leukocytosis, WBC > cells/mm 3 (%) Anemia, Hb < 10 g/dl or Hct < 30% (%) Thrombocytopenia, platelet < /mm 3 (%) Bandemia, >10% band (%) Serum creatinine >2 mg/dl (%) Mean C-reactive protein (mg/dl) 6.0 ± ± ± ED diagnosis (%) Urinary tract infection Low respiratory tract infection Fever without significant focus Intra-abdominal infection Upper respiratory tract infection >0.95 Skin or soft tissue infection >0.95 Not all the emergency department (ED) diagnoses are listed in the table. GCS, Glasgow Coma Scale; Hb, hemoglobin; Hct, hematocrit; HR, heart rate; RR, respiratory rate; SBP, systolic blood pressure; SD, standard deviation; WBC, white blood cell count. their 30-day outcome. All the study variables were used for comparisons between groups. Definition of end-point We used 30-day mortality as the primary end-point. People who survived at least 30 days whether or not they were still hospitalized were considered survivors for this analysis. Data analysis All analyses were carried out using SPSS 16.0 for Windows (SPSS, Chicago, IL, USA). Continuous data are means ± SD. Comparisons between two groups were made using either an independent-samples t-test (assuming normal distribution) or Mann Whitney/ Wilcoxon tests (assuming non-normality) for the continuous variables. Either a χ 2 -test or a Fisher s exact test was used for categorical variables. One-way ANOVA was used to compare the subgroups of age. The significant α level was set at 0.1 for univariate variables that are included in a multiple logistic regression analysis of risk for 30-day mortality. Significance was set at P < 0.05 (two tailed). The present study was organized as follows: (i) identify univariate correlates of death in geriatric ED Japan Geriatrics Society

4 Mortality prediction for geriatric fever Table 2 Multivariate logistic regression modeling using univariate comparison P < 0.1 of 330 elderly patients presenting with fever in the emergency department Mortality predictors Odds ratio (95% CI) P-value Hypotension (SBP < 90 mmhg) 5.74 ( ) Bedridden (ECOG 4) 3.49 ( ) Leukocytosis (WBC > ( ) cells/mm 3 ) Thrombocytopenia (platelets 4.21 ( ) < /mm 3 ) Serum creatinine >2 mg/dl 3.22 ( ) CI, confidence interval; ECOG, Eastern Cooperative Oncology Group Performance Status; ED, emergency department; SBP, systolic blood pressure; WBC, white blood cell count. Table 3 Sensitivity, specificity, positive predictive value, and negative predictive value of the number of independent mortality predictors for 30-day mortality No. mortality predictors present Sensitivity Specificity PPV NPV All data are %. NPV, negative predictive value; PPV, positive predictive value. patients with fever; (ii) use multivariate analyses to investigate independent mortality predictors; and (iii) combine the independent mortality predictors to predict the prognosis. Results We enrolled 330 geriatric patients (151 men [45.8%] and 179 women [54.2%]; mean age 78.5 ± 7.7 years; age range years; Table 1). The most common causes of fever were urinary tract infection (29.7%), lower respiratory tract infection (22.1%), fever without significant focus (7.6%), intra-abdominal infection (7.0%), upper respiratory tract infection (3.6%) and skin or soft tissue infection (3.0%). The 30-day mortality rate was 22 (6.7%). Univariate analysis showed that age was not associated with mortality (mean ages: survival group vs 30-day mortality group 78.3 ± 7.6 years vs 81.4 ± 4.5 years, P = 0.066; Table 1), and that there were no significant differences between age subgroups (young elderly vs moderately elderly vs old elderly, P = 0.244). Patients with the following variables had a higher mortality risk (P < 0.05): altered mental status, hypotension, leukocytosis, thrombocytopenia, bandemia, serum creatinine level <2 mg/dl, medical history of stroke, congestive heart failure, bedridden and nasogastric feeding. Other variables were not significantly different between groups. Multiple logistic regression modeling, using a univariate comparison with P < 0.1 (Table 1), showed that the presenting variables independently associated with 30-day mortality were hypotension, bedridden, leukocytosis, thrombocytopenia and serum creatinine level >2 mg/dl (Table 2). Sensitivity was highest (90.9%) for patients with the fewest predictors and lowest (18.2%) for patients with the most predictors (Table 3). Specificity, however, was just the opposite: 34.7% for the fewest and 100% for the most. Of the patients with all four predictors, 100% died during their first 30 days. Conversely, of the patients without any predictors, just 1.8% died. A total of 16 (72.7%) of the 22 patients who died within 30 days succumbed to sepsis, three (13.6%) to sepsis with end-stage cancer, one (4.5%) to intracranial hemorrhage, one (4.5%) to upper gastrointestinal bleeding and one (4.5%) to an acute coronary syndrome. Three (12.5%) geriatric patients with fever died from non-infection related illness Japan Geriatrics Society 837

5 M-H Chung et al. Discussion We found that for geriatric patients with fever who visited the ED, the independent mortality predictors were hypotension, bedridden, leukocytosis, thrombocytopenia and elevated serum creatinine >2 mg/dl. Multiple logistic regression analysis identified these five independent correlates of mortality. With at least one of these predictors present, the sensitivity was 90.9% and the NPV was 98.2%. With at least four of them present, the specificity was 100% and the PPV was 100%. Readily available to physicians, these predictors can provide important references for the treatment and disposition of geriatric patients with fever who visit the ED. In patients with a higher mortality risk, aggressive intervention, including admission to the intensive care unit, should be considered. Hypotension is a warning sign in sepsis, and needs aggressive resuscitation for surviving sepsis campaign bundles. 11,13 More than one-half of elderly patients have hypertension. Blood pressure 140/90 mmhg indicates a pathophysiological manifestation of altered cardiovascular physiology and structure. 14 The poor response from the cardiovascular reserve makes shock even more difficult to detect and diagnose. 15 Therefore, elderly patients with normal blood pressure might not actually indicate a normal state. Furthermore, hypotension (<90/ 60 mmhg) might indicate that a patient is in shock. Baseline functional dependence, such as being bedridden, is the most prevalent risk factor predicting various adverse outcomes in elderly patients in the ED. 7,8,16 In the present study, bedridden was defined as an ECOG score of 4: completely disabled, cannot carry on any self-care and totally confined to bed or chair. 10 The ECOG score has the advantage of simplicity over other performance status scores, such as the Karnofsky scale 17 and the Barthel scale. 18 Leukocytosis, a criterion of sepsis, was identified as an independent predictor of mortality in the present study. 11 Different studies have proposed different definitions of leukocytosis in geriatric patients. Marco et al., who surveyed geriatric patients presenting with fever in the ED, defined leukocytosis as WBC > cells/mm 3, and said that it predicted serious illness. 9 Wasserman et al., however, found that leukocyte counts of , and /mm 3 were associated with increasing specificity, but decreasing sensitivity. 8 We used WBC > cells/mm 3, which is more acceptable and more commonly used. Thrombocytopenia (platelets < /mm 3 ), another criterion of sepsis, was also an independent predictor of mortality in the present study. Serum creatinine >2 mg/dl, which is more likely caused by renal impairment rather than infection, also indicates a poor prognosis. 11 In geriatric patients who present with fever in the ED, mortality increased with the number of independent mortality predictors: hypotension, bedridden, leukocytosis, thrombocytopenia and serum creatinine >2 mg/ dl. All patients who presented with at least four of these predictors died within 30 days. Just 1.8% of the patients who presented without any of these predictors died within 30 days. These findings could help physicians make management and disposition decisions about geriatric patients who present with fever in the ED. The present study had several limitations. First, some data were collected from a retrospective chart review. The clinical presentations and records might not have been completely documented. Second, this was a singlecenter study. Findings from our database might not be generalizable to other cohorts in Taiwan or to cohorts in other nations. Third, the sample size might not be large enough to make conclusions with good statistical power. Additional studies with larger samples are necessary. Fourth, we did not validate this study. Internal validation for the same population or external validation for other populations is necessary. Fifth, we did not carry out comprehensive geriatric assessment on each participant, which is essential for managing geriatric patients, such as cognitive function, physical function, and instrumental activities of daily living. These potential predictors should be included when assessing mortality in further study. Acknowledgements This study was supported in part by a grant from CMFHR10311 from the Chi-Mei Medical Center. We thank Bill Franke for his invaluable advice and editorial assistance. MHC, SBS, CC Huang and CC Hsu conceived the study concept and design, acquired data, carried out statistical analysis, analyzed and interpreted the data, wrote the manuscript, and reviewed and edited the manuscript. FYC, TMY, HJL, SCV and HRG reviewed and edited the manuscript. JHC acquired, analyzed and interpreted data. CC Huang takes responsibility for the paper as a whole. All authors have read and approved the final manuscript. Conflict of interests All the authors declare that they have no potential conflicts of interest related to the writing or publication of this article. References 1 Cagatay AA, Tufan F, Hindilerden F et al. The causes of acute fever requiring hospitalization in geriatric patients: comparison of infectious and noninfectious etiology. J Aging Res 2010; 2010: Japan Geriatrics Society

6 Mortality prediction for geriatric fever 2 High KP, Bradley SF, Gravenstein S et al. Clinical practice guideline for the evaluation of fever and infection in older adult residents of long-term care facilities: 2008 update by the Infectious Diseases Society of America. Clin Infect Dis 2009; 48: Department of Statistics, Ministry of Interior, Taiwan. Table on the Elderly Population Count in Recent Years in Taiwan. [Cited 17 Oct 2013.] Available from URLs: [in Chinese]; [in Chinese] & [in English]. 4 Samaras N, Chevalley T, Samaras D et al. Older patients in the emergency department: a review. Ann Emerg Med 2010; 56: Meldon SW, Ma OJ, Woolard R. Geriatric Emergency Medicine. New York: McGraw-Hill, 2004; Norman DC. Fever in the elderly. Clin Infect Dis 2000; 31: Fontanarosa PB, Kaeberlein FJ, Gerson LW et al. Difficulty in predicting bacteremia in elderly emergency patients. Ann Emerg Med 1992; 21: Wasserman M, Levinstein M, Keller E et al. Utility of fever, white blood cells, and differential count in predicting bacterial infections in the elderly. J Am Geriatr Soc 1989; 37: Marco CA, Schoenfeld CN, Hansen KN et al. Fever in geriatric emergency patients: clinical features associated with serious illness. Ann Emerg Med 1995; 26: Oken MM, Creech RH, Tormey DC et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol 1982; 5: Dellinger RP, Levy MM, Rhodes A et al. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock, Intensive Care Med 2013; 39: Knaus WA, Draper EA, Wagner DP et al. APACHE II: a severity of disease classification system. Crit Care Med 1985; 13: Barochia AV, Cui X, Eichacker PQ. The surviving sepsis campaign s revised sepsis bundles. Curr Infect Dis Rep 2013; 15: Sander GE. High blood pressure in the geriatric population: treatment considerations. Am J Geriatr Cardiol 2002; 11: Oyetunji TA, Chang DC, Crompton JG et al. Redefining hypotension in the elderly: normotension is not reassuring. Arch Surg 2011; 146 (7): Aminzadeh F, Dalziel WB. Older adults in the emergency department: a systematic review of patterns of use, adverse outcomes, and effectiveness of interventions. Ann Emerg Med 2002; 39: Karnofsky DA, Burchenal JH. The clinical evaluation of chemotherapeutic agents in cancer. In: MacLeod CM, ed. Evaluation of Chemotherapeutic Agents. New York: Columbia University Press, 1949; O Sullivan SB, Schmitz TJ. Physical Rehabilitation, Fifth Edition. Philadelphia, PA: F.A. Davis Company, 2007; Japan Geriatrics Society 839