CHEST Delayed Administration of Antibiotics and Atypical Presentation in Community-Acquired Pneumonia* Grant W. Waterer, MD, FCCP; Lori A. Kessler, PharmD; and Richard G. Wunderink, MD, FCCP Original Research PNEUMONIA Objectives: The time to the first antibiotic dose (TFAD) has been adopted as a measure of quality of care in patients with community-acquired pneumonia (CAP) based on two retrospective studies of large Medicare databases. The mechanism by which a difference of a few hours in receiving antibiotics can be deleterious is difficult to understand given the historical data regarding how long it takes for antibiotics to influence outcome. We investigated the factors that predict a prolonged TFAD and their association with mortality. Design: Prospective cohort study. Setting: A large tertiary hospital. Patients: Immunocompetent adults admitted to the hospital with CAP. Results: A total of 451 patients with CAP were studied. A TFAD of > 4 h was associated with increased mortality (p 0.017). Altered mental state (p 0.001), absence of fever (p 0.02), absence of hypoxia (p 0.025), and increasing age (p 0.038) were significant predictors of a TFAD of > 4 h. After adjusting for these factors, the association between TFAD and mortality was not statistically significant (p 0.131). Similar findings were observed in patients who were > 65 years. Conclusions: A delay in administering antibiotics in patients with CAP is more common in patients who present with an altered mental state or minimal signs of sepsis. TFAD is likely to be a marker of comorbidities driving both an atypical presentation and mortality rather than directly contributing to outcome. Using TFAD as an indicator of quality of care in patients with CAP without significant additional clinical information is potentially misleading as the relationships among TFAD, comorbidities, and outcome are complex. (CHEST 2006; 130:11 15) Key words: age; altered mental state; antibiotics; community-acquired pneumonia; presentation; time Abbreviations: CAP community-acquired pneumonia; CI confidence interval; OR odds ratio; PSI pneumonia severity index; TFAD time to first antibiotic dose Community-acquired pneumonia (CAP) is a major health problem in the United States. As well as being the seventh leading cause of death, the estimated financial cost of treating CAP in the United States exceeds 12 billion dollars per year. 1 *From the School of Medicine and Pharmacology (Dr. Waterer), University of Western Australia, Perth, WA, Australia; Physicians Research Network (Dr. Kessler), Methodist Le Bonheur Healthcare, Memphis, TN; and Division of Pulmonary and Critical Care (Dr. Wunderink), Feinberg School of Medicine, Northwestern University, Evanston, IL. No author has any conflict of interest with respect to this article. Dr. Waterer is supported by the National Health and Medical Research Council of Australia. Two retrospective analyses 2,3 of large Medicare databases identified the time between presentation to the hospital and the time to the first antibiotic dose (TFAD) as a significant predictor of outcome. Manuscript received February 16, 2006; revision accepted February 25, 2006. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml). Correspondence to: Grant W. Waterer, FCCP, University of Western Australia, School of Medicine and Pharmacology, 4th Floor MRF Building, Royal Perth Hospital, GPO Box X2213, Perth, WA, Australia 6847; e-mail waterer@cyllene.uwa.edu.au DOI: 10.1378/chest.130.1.11 www.chestjournal.org CHEST / 130 / 1/ JULY, 2006 11
Both of these studies have inconsistencies, particularly the higher mortality rate among those receiving antibiotics within 2 h. However, the findings of these studies have convinced national regulatory bodies in the United States to make TFAD a benchmark for quality of care in patients with CAP. 4 A reasonable biological explanation for why a difference of a few hours in antibiotic administration should lead to a better outcome has not been established. Historical data strongly indicate that For editorial comment see page 1 antibiotics take several days to impact on outcome from pneumococcal pneumonia. 5 Possible explanations include confounding factors inadequately accounted for in the retrospective database reviews and TFAD being a surrogate marker for other quality-of-care factors. While the use of large Medicare databases has advantages with respect to statistical power and the ability to generalize findings across broader populations, clear disadvantages occur with respect to the loss of detailed individual clinical data and the loss of surety of diagnosis inherent in prospective clinical trials. We analyzed our prospectively collected cohort of patients with CAP specifically examining the clinical factors influencing TFAD. Study Design Materials and Methods A prospective cohort of patients admitted to the Methodist Healthcare-Memphis Hospitals with CAP between November 1998 and July 2001 was recruited. Informed consent was obtained from all patients. The Methodist Healthcare Institutional review board approved the study. for 7 days), nonambulatory nursing home residents, and patients who have been hospitalized within the past 30 days. Definitions The TFAD was defined as the time difference between the recorded time on presentation to triage in the emergency department and the recorded time of administration of the first dose of antibiotics by the nursing staff. Pneumonia severity index (PSI) scores 7 were calculated at the time of the patient s admission to the hospital. Septic shock was defined using American College of Chest Physicians/Society of Critical Care Medicine criteria. 8 Hypoxia was defined as an oxygen saturation of 90% while breathing room air. An altered mental state was defined as any disorientation in time, place, or person including any Glasgow coma scale score of 15. Statistical Analysis Logistic regression analysis was used to assess the impact of clinical factors on TFAD. Significant interactions were included in all models at a threshold of p 0.1. Analysis was performed using a statistical software package (SPSS, version 11.5.0; SPPS Inc; Chicago, IL). A p value of 0.05 was considered to be significant. Results A total of 451 patients were enrolled into the study. The cohort had a mean ( SD) age of 58.2 19.2 years, 53.3% were women, and the severity breakdown by PSI grade was as follows: grade I, 11.3%; grade II, 22.2%; grade III, 18.9%; grade IV, 20.4%; and grade V, 8.7%. The mean TFAD was 285 202 min, with 50.7% of patients not receiving their first dose within 4 h. Of those patients with a TFAD of 4 h, 32.0% were given antibiotics within 2 h of presentation. Death occurred in 36 patients (8.0%). We next analyzed the data points on the PSI predicting a TFAD of 4 h. Table 1 summarizes the significant findings. Various age cutoffs were as- Inclusion Criteria Consistent with published guidelines, 6 CAP was defined as an acute illness ( 14 days of symptoms) in which the presence of a new chest radiographic infiltrate was confirmed by either a radiologist or a pulmonary/critical care physician, and having clinical features suggestive of acute pneumonia. The clinical features required for diagnosis were one feature from group A (fever [temperature of 37.8 C], hypothermia [temperature of 36 C], cough, and sputum production), or two features from group B (dyspnea, pleuritic pain, physical findings of lung consolidation, and leukocyte count of 12 10 cells/l or 4.5 10 cells/l). Exclusion Criteria Exclusion criteria included patients with HIV infection, those receiving chemotherapy or who had received immunosuppressive therapy in the past 60 days (including prednisolone, 20mg/d Table 1 Factors Predicting a Delay in Antibiotic Delivery of > 4h Altered mental state 2.89 (1.53 5.45) 0.001 Absence of hypoxia 1.82 (1.09 3.04) 0.022 Absence of fever 1.59 (1.06 2.40) 0.025 Age 1.01 (1.00 1.06) 0.038 Cerebrovascular disease 1.09 (0.56 2.14) 0.351 Congestive cardiac failure 0.79 (0.47 1.31) 0.791 Renal disease 0.87 (0.43 1.74) 0.406 Liver disease 0.89 (0.37 1.91) 0.851 Nursing home resident 3.24 (0.32 32.60) 0.294 Absence of shock 1.76 (0.46 6.72) 0.405 Tachycardia 125 beats/min 1.30 (0.84 2.02) 0.245 Respiratory rate 30 breaths/min 1.25 (0.71 1.86) 0.288 12 Original Research
Table 2 Univariate Analysis of Predictors of Mortality Altered mental state 5.88 (2.59 13.33) 0.0001 Age 1.03 (1.01 1.05) 0.027 TFAD 4 h 2.82 (1.21 6.58) 0.017 Absence of hypoxia 2.21 (0.94 5.06) 0.07 Absence of fever 4.00 (1.73 9.26) 0.001 Table 4 Univariate Analysis of the Association Between Clinical Factors and Mortality in Patients > 65 Years of Age Altered mental state 3.94 (1.39 11.11) 0.01 Age 1.03 (0.96 1.10) 0.40 TFAD 4 h 2.94 (0.92 9.43) 0.07 Absence of hypoxia 1.90 (0.67 5.35) 0.22 Absence of fever 1.36 (0.41 4.50) 0.16 sessed, but none performed better than treating age as a continuous variable. We also analyzed the predictors of antibiotic delivery in 2 h. Significant predictors were the presence of shock (odds ratio [OR], 3.63; 95% confidence interval [CI], 1.63 to 8.09), fever with a temperature of 101.0 F (OR, 2.20; 95% CI, 1.31 to 5.43), and the presence of hypoxia (OR, 1.69; 95% CI, 1.04 to 2.75). The effect on mortality was analyzed for a TFAD of 4 h and the factors in Table 1 predicting a delay in antibiotic administration. As shown in Table 2, a TFAD of 4 h was associated with excess mortality (p 0.017); however, altered mental state and absence of fever were even stronger predictors. Logistic regression was used to assess the interactions among mortality, TFAD 4 h, and the factors in Table 1. As shown in Table 3, after adjustment for altered mental state, a TFAD of 4 h was not a significant predictor of mortality. As the principal association between mortality and TFAD has been in patients aged 65 years, 2,3 we analyzed this subset of patients separately. A total of 158 patients were 65 years of age. In this cohort, there were 17 deaths (10.8%), and 71 patients (44.9%) received antibiotics after 4 h. In those patients aged 65, an altered mental state remained a significant predictor of TFAD of 4 h (OR, 3.2; 95% CI, 1.4 to 6.1). As seen in the univariate analysis presented in Table 4, a TFAD of 4 h approached statistical significance as a predictor of mortality, while the only statistically significant predictor was an altered mental state. In a multivariate analysis, a TFAD of 4 h was not an independent predictor of mortality (p 0.158) once an altered mental state (p 0.02) was entered into the model. Table 3 Predictors of Mortality in the Multivariate Model Altered mental state 3.33 (1.28 8.77) 0.009 Age 1.01 (0.99 1.03) 0.370 TFAD 4 h 1.85 (0.84 5.00) 0.117 Absence of hypoxia 2.08 (1.15 6.35) 0.204 Absence of fever 2.55 (1.02 6.37) 0.045 Discussion Using detailed clinical data from a prospectively collected cohort of patients with CAP, we have confirmed that a delay of 4 h in receiving antibiotics is associated with adverse outcome. A delay of 4 h is predicted by increasing age, the presence of an altered mental state, and the absence of significant fever or hypoxia. However, the association between TFAD and mortality is not straightforward as a delay in antibiotic administration is also associated with comorbid diseases that are very likely to be contributing to excess mortality. Our data also suggest that in elderly patients, and especially those presenting with confusion, physicians need to be more vigilant in excluding a diagnosis of pneumonia. The association of a TFAD of 4 h with an altered mental state and the absence of obvious signs of pneumonia, such as fever and hypoxia, have significant implications for the current recommendations for assessing the quality of care of patients with pneumonia. The Joint Commission of Health Care Organizations, the Centers for Medicare and Medicaid Services, and the National Quality Forum have adopted TFAD based on the findings from two studies of large Medicare databases that a delay of 4h 2 or 8h 3 was associated with an increased mortality. However, the fact that an altered mental state and the absence of fever were stronger predictors of mortality than a TFAD of 4 h, combined with the multivariate analysis, suggest that a prolonged TFAD is more a marker for comorbidities that are responsible for both the increased mortality and atypical presentation. Particularly pertinent is the presence of an altered mental state, as this is associated not only with increased in-hospital mortality, 9 but also with increased mortality after hospital discharge. 10 An altered mental state may not only make it difficult for physicians to make a diagnosis, it may also delay presentation to hospital. It is possible that an atypical presentation leads to a delay in the administration of antibiotics and that it is the delay that leads to increased mortality. However, historical data on the length of time that it takes for antibiotics to make an impact on outcome makes www.chestjournal.org CHEST / 130 / 1/ JULY, 2006 13
it unlikely that a difference of a few hours in administration will impact adversely on mortality, 5 particularly as most patients have had symptomatic disease for several days. Furthermore, an altered mental state is associated with adverse in-hospital outcome 11 and after-hospital outcome 12 across a diverse range of medical conditions including noninfectious diseases. The adverse implications of an altered mental state are probably due to its association with increasing age and recognized or unrecognized comorbid illnesses, such as dementia. 11,12 Therefore, while there is no clear biological basis for why a difference of a few hours in administering antibiotics should have a direct affect on outcome, there is a cogent argument for comorbidities causing both an increased mortality and an atypical presentation of pneumonia, which lead to delayed recognition of pneumonia and a prolonged TFAD. As TFAD is still a predictor of mortality regardless of whether it is causative or not, it may still be a useful quality-of-care marker as a surrogate marker for clinical acumen (with higher acumen being associated with an earlier diagnosis of pneumonia and therefore earlier treatment). Prolonged TFAD was associated with an atypical clinical presentation, suggesting that physicians do not consider pneumonia initially in patients who are unable to give a good history of respiratory tract illness or when signs of sepsis are absent. The observation that florid signs of sepsis were associated with the early administration of antibiotics further supports a prolonged TFAD as being a failure of recognition rather than a problem such as excess workload and failure to assess patients within a reasonable period of time. If TFAD is used as a quality-of-care marker, it cannot be compared across institutions without good information on the proportion of patients who present with an altered mental state or an otherwise atypical presentation. Although captured by the PSI, this information is lost when only the total score is factored into the analysis. The Medicare database studies 2,3 were also limited by the fact that data were collected retrospectively and that key data points were not always available. As our study was prospective and patients were seen by an investigator within 24 h of presentation, we can be much more confident of the diagnosis of pneumonia and the relevant features of the clinical presentation. This may be particularly pertinent with respect to the recognition and documentation of an altered mental state in elderly patients. As our study was conducted at a single site, the findings may not be able to be generalized to other institutions, which was the strength of the large Medicare database studies. 2,3 However, evidence that altered mental state (or delirium) is underrecognized in emergency departments 13 and that it leads to the misdiagnosis of the primary problem 14 has been documented in other studies. An atypical presentation of pneumonia, particularly in the elderly, has also been reported as a factor contributing to delayed diagnosis, 15,16 suggesting that the issues we have identified are common ones. Although we did not show a significant association between TFAD and mortality in the subset of patients 65 years of age, this is most likely due to the low numbers of patients in the study. The fact that an altered mental state remained a significant predictor despite the low numbers adds further weight to the hypothesis that this is the key factor driving both mortality and a prolonged TFAD. In conclusion, in a detailed prospective cohort of patients admitted to the hospital with CAP, we have found that the presence of an altered mental state and an atypical clinical presentation predict a delay of 4 h in administering the first dose of antibiotics. While this delay was associated with an increased likelihood of adverse outcome, the most likely explanation is the impact of comorbidities on mortality rather than differences in the process or quality of medical care. While physicians need to be more alert to the possibility of pneumonia in patients presenting with an altered mental state, the major implication of our findings is that the widespread adoption of TFAD by regulatory bodies as a quality measure in treating patients with CAP is based on an incomplete understanding of the complexity of the relationships among TFAD, comorbidities, and outcome. References 1 Colice GL, Morley MA, Asche C, et al. Treatment costs of community-acquired pneumonia in an employed population. Chest 2004; 125:2140 2145 2 Meehan TP, Fine MJ, Krumholz HM, et al. Quality of care, process, and outcomes in elderly patients with pneumonia. JAMA 1997; 278:2080 2084 3 Houck PM, Bratzler DW, Niederman M, et al. Pneumonia treatment process and quality. Arch Intern Med 2002; 162: 843 844 4 Joint Commission on Accreditation of Healthcare Organizations. Specification manual for national implementation of hospital core measures, version 2.0. Oakbrook Terrace, IL: Joint Commission on Accreditation of Healthcare Organizations, 2004; 1 11 5 Austrian R, Gold J. Pneumococcal bacteremia with special reference to bacteremic pneumococcal pneumonia. Ann Intern Med 1964; 60:759 776 6 Chow AW, Hall CB, Klein JO, et al. General guidelines for the evaluation of new anti-infective drugs for the treatment of respiratory tract infections. Clin Infect Dis 1992; 15:S62 S88 7 Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify low-risk patients with community-acquired pneumonia. N Engl J Med 1997; 336:243 250 8 Bone RC, Balk, Cerra, et al. Definitions for sepsis and organ 14 Original Research
failure and guidelines for the use of innovative therapies in sepsis: the ACCP/SCCM Consensus Conference Committee; American College of Chest Physicians/Society of Critical Care Medicine. Chest 1992; 101:1644 1655 9 Fine MJ, Smith MA, Carson CA, et al. Prognosis and outcomes of patients with community-acquired pneumonia: a meta-analysis. JAMA 1996; 275:134 141 10 Waterer GW, Kessler LA, Wunderink RG. Medium-term survival after hospitalization with community-acquired pneumonia. Am J Respir Crit Care Med 2004; 169:910 914 11 Marcantonio ER, Kiely DK, Simon SE, et al. Outcomes of older people admitted to postacute facilities with delirium. J Am Geriatr Soc 2005; 53:963 969 12 Kakuma R, dufort GG, Arsenault L, et al. Delerium in older emergency department patients discharged home: effect on survival. J Am Geriatr Soc 2003; 51:443 450 13 Hustey FM, Meldon SW. The prevalence and documentation of impaired mental status in elderly emergency department patients. Ann Emerg Med 2002; 39:248 253 14 Reeves RR, Pendarvis EJ, Kimble R. Unrecognized medical emergencies admitted to psychiatric units. Am J Emerg Med 2000; 18:390 393 15 Metlay JP, Shulz R, Li YH, et al. Influence of age on symptoms at presentation in patients with communityacquired pneumonia. Arch Intern Med 1997; 157:1453 1459 16 Emmett KR. Nonspecific and atypical presentation of disease in the older patient. Geriatrics 1998; 53:50 52 www.chestjournal.org CHEST / 130 / 1/ JULY, 2006 15