• 
• 

Three Part Question

In [adult patients presenting to the hospital with suspected community-acquired pneumonia (CAP) requiring admission], does [antibiotic administration within a specific time window] [reduce patient morbidity and mortality]?

Clinical Scenario

A 70 year-old lady presents to the emergency department(ED) with 2 days of fever, shortness of breath, and cough productive of green sputum. CXR confirms right basal pneumonia. She requires oxygen therapy and admission. It is busy in the ED. This patient has been waiting for 3.5 hours. Her bed is ready in the ward. You wonder if giving her the antibiotics now would affect her clinical outcome in terms of time to clinical stability, length of hospital stay, and mortality.

Search Strategy

Medline 1990 to September week 1 2010 using OVID interface.
[exp Community-Acquired Infections/ OR exp Pneumonia/ OR community acquired pneumonia.mp.] AND [antibiotics.mp. OR exp Anti-Bacterial Agents/ OR antimicrobials.mp.] AND [early administration.mp. OR timely administration.mp. OR rapid delivery.mp. OR timing.mp. OR delayed administration.mp. OR within 4 hours.mp. OR within 6 hours.mp. OR within 8 hours.mp.] AND [mortality.mp. OR exp Hospital Mortality/ OR exp Mortality/ OR exp “Outcome and Process Assessment (Health Care)”/ OR outcome.mp. OR morbidity.mp. OR exp Morbidity/ OR length of stay.mp. OR clinical stability.mp.]

Limit to abstracts, english language, humans, and adults.

Reviews and case reports were excluded.

Search Outcome

74 papers were found of which only 13 were relevant. 4 of these were review articles and 1 was a case report. 8 remaining papers were reviewed.

6 additional papers found via citation reference search on the Web of Science® and from bibliographic references were also reviewed.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Meehan et al 1997 USA	14 069 patients. ≥65 yr. Randomly selected from 3555 Medicare hospitals with: International Classification of Disease (ICD)-9 codes of pneumonia on discharge, admission diagnosis of pneumonia, and new (48h) CXR changes.	Multicentre retrospective cohort	Adjusted association between process of care performance including time to first antibiotic dose (TFAD) and 30-day mortality	Significantly reduced 30-day mortality with TFAD within 8h (OR, 0.85; 95% CI, 0.75-0.96; p<0.001). Stronger association when limited to no pre-hospital antibiotics (OR, 0.78; 95%CI, 0.67-0.89). Blood culture collection within 24h was also associated with reduced mortality	Retrospective. Observational. Sample based on claims data. Data based on discharge diagnosis. Focused on ≥65yr. Did not assess if the antibiotics were appropriate. Unmeasured confounding factors. Unexplained reason for 8h cutoff: TFAD within 9h and 10h have similar OR and CI.
Dedier et al 2001 USA	1062 patients. ≥18 yr. From 38 academic hospitals with: ICD-9 codes of pneumonia on discharge, new (24h) CXR changes, and no pre-hospital antibiotics.	Multicentre retrospective cohort	Adjusted association between process of care performance including TFAD within 8h and:		Retrospective. Observational. Data based on discharge diagnosis. Non-randomized. Sickest patients received more timely antibiotics. 29% were low risk Pneumonia Severity Index score(PSI) I-II patients. High process maker achievement limited study power. Unmeasured confounding factors.
			48h clinical stability	No association found
			Length of stay (LOS)	No association found
			Inpatient mortality	No association found
Battleman et al 2001 USA	609 patients. ≥18 yr. Randomly selected from 7 hospitals. ED admission with: Diagnosis-Related Group (DRG) codes of pneumonia on discharge, admission diagnosis of pneumonia, and no pre- hospital antibiotics.	Multicentre retrospective cohort	Adjusted association between prolonged LOS (≥9 days) and:		Retrospective. Observational. Data based on discharge diagnosis. Unmeasured confounding factors. Data for TFAD not shown.
			Site of initial antibiotic administration (ED versus ward)	Reduced LOS when antibiotics given in the ED (OR, 0.31; 95% CI, 0.19- 0.48; p <0.001)
			TFAD	Prolong LOS with delayed TFAD (OR, 1.75 per 8h; 95% CI, 1.34-2.29; p <0.001)
			Choice of antibiotics	Reduced LOS with appropriate antibiotic selection (OR, 0.55; 95% CI, 0.35-0.88; p<0.05)
Silber et al 2003 USA	409 patients. ≥21 yr. From 1 teaching hospital. ED admission with moderate to severe pneumonia (PSI III- V). Placed into 3 groups (group 1 TFAD within 4h, group2 >4-8h, group 3 >8h).	Prospective cohort	Differences between the three groups in:		Small sample size. Single setting. Observational. Inclusion of patients with pre-hospital antibiotics. Exclusion of patients never reached clinical stability.
			Mean time to clinical stability	No statistically significant differences
			LOS	No statistically significant differences
			Mortality	No statistically significant differences
Houck et al 2004 USA	13 771 patients. ≥65 yr. Randomly selected from 3732 Medicare hospitals with: ICD-9 codes of pneumonia on discharge, admission diagnosis of pneumonia, CXR changes, and no pre-hospital antibiotics.	Multicentre retrospective cohort	Adjusted association between TFAD within 4h and:		Retrospective. Observational. Sample based on claims data. Data based on discharge diagnosis. Focused on ≥65yr. Did not assess if the antibiotics were appropriate. Unmeasured confounding factors. Unexplained reason for 4h cutoff: OR of 4h and 8h cutoffs were identical.
			In-hospital mortality	Reduced in-hospital mortality (OR, 0.85; 95% CI, 0.74-0.98; p= 0.03)
			30-day mortality	Reduced 30-day mortality (OR, 0.85; 95% CI, 0.76-0.95; p= 0.005)
			LOS	Reduced LOS (OR, 0.90; 95% CI, 0.83-0.96; p = 0.003)
			Re-admission	No reduction in re-admission
Marrie et al 2005 Canada	3034 patients. From 6 hospitals. ED admission with: ≥2 symptoms or signs of pneumonia and CXR changes.	Multicentre prospective cohort	Adjusted association between a list of predictors including TFAD and in-hospital mortality	No significant difference in TFAD between those who survived and those who died (p=0.48). PSI score, age, site of care, functional status, and specialist involvement were independent predictors for mortality	Observational. Inconsistent exclusion criteria. Inclusion of patients with recent hospital admission. Exclusion of the most critically unwell patients requiring ICU admission from the ED.
Waterer et al 2006 Australia	451 patients. From 1 tertiary hospital. ED admission with: symptoms or signs of pneumonia or laboratory signs of infection, and new CXR changes.	Prospective cohort	Identify clinical factors predictive of delay in TFAD	Altered mental state, absence of fever, hypoxia, and old age were predictive of a TFAD >4h	Small sample size. Single setting. Observational. Non-randomized. Exclusion of non-ambulatory nursing home residents. Unclear if patients had pre- hospital antibiotics. Limited severity adjustments.
			Adjusted association between TFAD >4h and mortality	No association between TFAD >4h and mortality (OR, 1.85; 95% CI, 0.84-5.0; p= 0.117)
Schaaf et al 2007 Germany	105 patients. From 1 University and 2 community hospitals with: symptoms or signs of pneumonia, new CXR changes, laboratory signs of infection, and isolation of Streptococcus pneumoniae from blood, cerebrospinal fluid, respiratory secretions or other sterile sites.	Prospective cohort	The effect of TFAD on in-hospital mortality	Higher mortality in patients with TFAD within 8h (15.9%) than patients with TFAD >8h (0%) but of no significant difference (p = 0.1)	Small sample size. Observational. Did not analyze the association between TFAD and mortality. Bacteraemia represented the sicker subgroup of CAP. 40.6% patients with TFAD within 8h were in severe sepsis or septic shock i.e. already high predictors for mortality.
Kanwar et al 2007 USA	518 patients. ≥21 yr. From 1 teaching hospital. ED admission with initial diagnosis of pneumonia.	Retrospective cohort	Impact of reducing TFAD guidance from 8h to 4h on:		Small sample size. Single setting. Retrospective. Non-randomized. More patients in TFAD within 4h group. Accurate diagnosis was based on the subjective opinion of the attending physician.
			The accuracy of CAP diagnosis	17% more inaccurate CAP diagnosis (p< 0.001)
			The use of antibiotics <4h	12% increase in use (p= 0.007)
			ICU requirement, LOS, mortality	No significant difference
				No change in mean TFAD
Welker et al 2008 USA	548 patients. ≥18 yr. From 1 teaching hospital. ED admission with: ≥2 symptoms or signs of pneumonia or hypoxia, new CXR changes, and temperature >38ºC or <35.1ºC or WCC>10/µL or <4.5/µL.	Retrospective cohort	Impact of reducing TFAD guidance from 8h to 4h on the accuracy of CAP diagnosis	39% less likely to meet pre-defined pneumonia criteria on admission (p= 0.004) and agree with physician diagnosis at discharge (p= 0.05)	Retrospective. Small sample size. Single setting. Data were abstracted by the authors, not randomized, not blinded, and with potential inter-observer error. p value of 0.05 was of borderline significance.
				No statistical increase in antibiotic use
				No statistical difference in mean TFAD
Berjohn et al 2008 USA	363 patients. ≥18 yr. From 43 hospitals. ED admission with: clinical and radiological diagnosis of pneumonia, and at least 1 blood culture positive for Streptococcus pneumoniae.	Multicentre retrospective cohort	Adjusted association between TFAD and:	Receipt of at least 1 appropriate antibiotics within 4h was associated with:	Large number of patients excluded. Small remaining sample size. Retrospective. Bacteraemia represented the sicker subgroup of CAP. Majority (66%) of the patients received TFAD within 4h.
			30-day in-hospital mortality	Reduced mortality (OR, 0.47; 95% CI, 0.2-1.0; p= 0.04)
			Complication rates	No change in complication rates
			LOS	Shortened LOS (OR , 0.77; CI, 0.6-1.0; p= 0.03)
			Time to vital sign stability	No change to time to stability
Bruns et al 2009 Netherlands	152 patients. ≥18 yr. From University Medical Centres and their affiliated teaching hospitals with: ≥2 acute lower respiratory tract symptoms, new CXR changes, moderate to severe pneumonia (PSI score >90), and no pre-hospital antibiotics.	Prospective cohort using data derived from a multicentre prospective randomized controlled trial	Difference between TFAD within or >4h in the rate of developing early clinical failure (Day 3 clinical instability, mortality and ICU admission)	No statistically significant difference (p= 0.28)	Small sample size considering data derived from a 3 year study. Observational. Secondary data. Non-randomized. Exclusion of patients requiring ICU admission from the ED. More patients in TFAD within 4h group. Median for TFAD >4h was only 5h17min. Absence of early clinical failure did not imply a favourable long-term outcome.
			Identify factors predictive of early clinical failure	PSI score, confusion, Staphylococcus aureus infection and multilobar pneumonia, but not TFAD, were independently associated with early clinical failure
Cheng et al 2009 Australia	501 patients. ≥18 yr. From 1 teaching hospital with: acute respiratory tract symptoms, new CXR changes, admission diagnosis of pneumonia, and no pre-hospital antibiotics.	Prospective cohort	Effect of TFAD on:		Small sample size. Single setting. Observational. Non-randomized. No statistical calculations. TFAD within 8h in 91% patients limited study power.
			In-hospital mortality	Shorter TFAD in patients who died (median 1.5h; IQR, 0.9- 2.7h) than those who survived (median 2.9h; IQR, 1.7- 4.8h). 98% patients who died were PSI IV or V
			LOS	TFAD similar in patients with (median 2.6h; IQR 1.2-6.5h) or without prolonged LOS (median 2.9h; IQR 1.8-4.8h)
Garnacho- Montero et al 2010 Spain	125 patients. From 1 tertiary hospital with Streptococcus pneumoniae positive blood culture secondary to CAP	Prospective cohort	Adjusted association between a list of predictors including TFAD and:		Small sample size. Single setting. Observational. Non-randomized. Bacteraemia represented the sicker subgroup of CAP. Time to first appropriate antibiotic >4h did not fulfill the stated inclusion criteria of p <0.1 in the unadjusted model (p= 0.103) to be entered into the multivariate model.
			In-hospital mortality	Delayed time to first appropriate antibiotic >4h (HR, 2.62; 95% CI, 1.06- 6.45; p= 0.037) and severe sepsis or septic shock on admission, were independent predictors for in-hospital mortality
			90-day mortality	Delayed time to first appropriate antibiotic >4h (HR, 2.21; 95% CI, 1.01-4.86; p=0.048), severe sepsis or septic shock, and Charlson comorbidity index were independent predictors for 90-day mortality

Comment(s)

Studies on the association between time to first antibiotic administration and clinical outcomes of patients with community-acquired pneumonia have been conflicting and controversial.

Firstly, there are no randomized control trials, nor are there likely to be. Secondly, there are no universally accepted criteria for CAP in the studies. Thirdly, the cutoff time for first antibiotic dose, sample demographics, pneumonia severity, setting, inclusion/exclusion criteria, and predictors used for adjustments all differ between papers, making it difficult to come to a clear conclusion. Fourthly, the pre-hospital phase of illness, which could have contributed to days of treatment delay, was not accounted for in many studies.

Clinical Bottom Line

Early administration of guideline-recommended antibiotics in septic patients with a likely diagnosis of CAP is good ED practice according to the Surviving Sepsis guidelines. However, there is no consistent evidence that for CAP alone, antibiotics given within a specific time window shortens time to clinical stability, length of hospital stay, or reduces mortality.

References

1. Meehan TP, Fine MJ, Krumholz HM, et al. Quality of care, process, and outcomes in elderly patients with pneumonia. JAMA. 1997;278:2080-2084.
2. Dedier J, Singer DE, Chang Y, et al. Process of care, illness severity, and outcomes in the management of community-acquired pneumonia at academic hospitals. Arch Intern Med. 2001;161:2099-2104.
3. Battleman DS, Callahan M, Thaler HT. Rapid antibiotic delivery and appropriate antibiotic selection reduce length of hospital stay of patients with community-acquired pneumonia. Arch Intern Med. 2002;162:682-688.
4. Silber SH, Garrett C, Singh R, et al. Early administration of antibiotics does not shorten time to clinical stability in patients with moderate-to-severe community-acquired pneumonia. Chest. 2003;124:1798-1804.
5. Huock PM, Bratzler DW, Nsa W, et al. Timing of antibiotic administration and outcomes for Medicare patients hospitalized with community-acquired pneumonia. Arch Intern Med. 2004;164:637-644.
6. Marrie TJ, Wu L. Factors influencing in-hospital mortality in community-acquired pneumonia: a prospective study of patients not initially admitted to the ICU. Chest. 2005;127:1260-1270.
7. Waterer GW, Kessler LA, Wunderink RG. Delayed administration of antibiotics and atypical presentation in community-acquired pneumonia. Chest. 2006;130:11-15.
8. Schaff B, Kruse J, Rupp J, et al. Sepsis severity predicts outcome in community-acquired pneumococcal pneumonia. Eur Respir J. 2007;30:517-524.
9. Kanwar M, Brar N, Khatib R, et al. Misdiagnosis of community-acquired pneumonia and inappropriate utilization of antibiotics: side effects of the 4-h antibiotic administration rule. Chest. 200;131:1865-1869.
10. Welker JA, Huston M, McCue JD. Antibiotic timing and errors in diagnosing pneumonia. Arch Intern Med. 2008;168:351-356.
11. Berjohn CM, Fishman NO, Joffe MM, et al. Treatment and outcomes for patients with bacteremic pneumococcal pneumonia. Medicine. 2008;87:160-166.
12. Bruns AHW, Oosterheert JJ, Hustinx WNM, et al. Time for first antibiotic dose is not predictive for the early clinical failure of moderate-severer community-acquired pneumonia. Eur J Clin Microbiol Infect Dis. 2009;28:913-919.
13. Cheng AC, Buising KL. Delayed administration of antibiotics and mortality in patients with community-acquired pneumonia. Ann Emerg Med. 2009;53:618-624.
14. Garnacho-Montero J, Garcia-Cabrera E, Diaz-Martin A, et al. Determinants of outcome in patients with bacteraemic pneumococcal pneumonia: importance of early adequate treatment. Scand J Infect Dis. 2010;42:185-192.