- Email: [email protected]
Analysis of emergency department management of suspected bacterial meningitis
ORIGINAL CONTRIBUTION bacterial meningitis, management
Analysis of Emergency Department Management of Suspected Bacterial Meningitis Previous studies of emergency department management of bacterial meningitis have indicated that there are often long delays before initiation of antibiotics. The purpose of our study was to determine whether these delays were related to specific aspects of patient management. From 1981 through 1988, we retrospectively reviewed the medical records of 122 patients primarily evaluated in the ED and admitted for suspected bacterial meningitis at a university (55) and a c o m m u n i t y (67) hospital. The median time (interquartile range) from ED registration until initiation of antibiotics (time to antibiotics) was 3.0 hours (1.6 and 4.3 hours, respectively) (total range, 0.5 to 18 hours). The time to antibiotics was not significantly related to the time of ED registration. Ninety percent of the total time to antibiotics occurred after the initial physician encounter. Time to antibiotics was significantly (P < .00005) longer for patients in whom computed tomography scan and/or laboratory analysis of cerebrospinaI fluid preceded initiation of antibiotics compared w~th patients in whom antibiotic administration was not contingent on the results of these procedures (4.3 [3.2 and 6.0] versus 1.9 [1.2 and 3.4] hours, respectively). Also, time to antibiotics was significantly (P < .00005) longer for patients in whom antibiotics were initiated on the ward as compared with in the ED (4.5 [3.5 and 6.8] versus 2.2 [1.4 and 3.5] hours, respectively). We conclude that long delays exist in the ED before initiation of antibiotics for cases of suspected bacterial meningitis, and that in general these delays appear to be physician generated and to a great extent potentially avoidable. [Talan DA, Guterman JJ, Overturf GD, Singer C, Hoffman JR, Lambert B: Analysis of emergency department m a n a g e m e n t of suspected bacterial meningitis. Ann Emerg Med August 1989;18:856-862.]
David A Talan, MD* Jeffrey J Guterman, MD* Gary D Overturf, MDt Craig Singer, MD*¢ Jerome R Hoffman, MD§ Bret Lambert, MD§ Los Angeles, California From the Departments of Emergency Medicine* and Pediatrics,t Olive View/ UCLA Medical Center; UCLA/Olive View Emergency Medicine Residency;¢ and the Department of Medicine, Division of Emergency Medicine, UCLA Medical Center,§ Los Angeles, California. Received for publication December 13, 1988. Revision received March 27, 1989. Accepted for publication April 4, 1989. Presented at the 2rid International Conference on Emergency Medicine in Brisbane, Australia, October 1988. Address for reprints: David A I-alan, MD, Department of Emergency Medicine, Olive View/UCLA Medical Center, 14445 Olive View Drive, Sylmar, California 91342.
INTRODUCTION Bacterial meningitis is typically a rapidly progressive, life-threatening illness that may be amenable to cure by the prompt administration of appropriate antibiotics. As such, bacterial meningitis constitutes an infectious disease emergency. Textbooks of infectious diseases emphasize the need for rapid diagnosis and treatment and state that antibiotics should be administered within 30 minutes of encountering the patient. 1 However,
recent studies indicate that, in actual practice, therapy is typically not given for several hours.2, 3 Bryan et al z reported a series of 135 patients with bacterial meningitis in whom the mean time from emergency department patient registration to administration of antibiotics for pediatric and adult cases was 2.1 and 4.9 hours, respectively. Delayed diagnosis and therapy may have serious prognostic implications and may also account in part for the fact that bacterial meningitis ranks second in total malpractice dollars paid per case originating from EDs. 4 The factors that contribute to longer than expected times until therapy remain to be elucidated. We report a retrospective review of the ED management of 122 children and adults admitted with the diagnosis of suspected bacterial meningitis. The purpose of our study was to determine which aspects of ED management of these patients were associated with delays of antibiotic therapy. We hypothesized that delays were associated with care during the busiest time of the day, the practice of withholding antibiotics until the results of
18:8 August 1989
Annals of Emergency Medicine
856/105
MENINGITIS MANAGEMENT Talan et al
20
150
15 ¢o
"b
100 10
5O
0 1 |
2
3
4
5
6
7
9
10
18
2
Reg-Room
Room-MD
M D - AB
Time to Antibiotics (hr)
FIGURE 1. Time to antibiotics for 122 patients with suspected bacterial meningitis. FIGURE 2. Median time of managem e n t steps for patients w i t h susp e c t e d bacterial m e n i n g i t i s (120). Reg-Roorn ED, time from registration to p l a c e m e n t in an e x a m i n a t i o n room; Room-MD, time from placement in an examination room to initial physician encounter; MD-AB, time from initial physician encounter to initiation of antibiotics. computed tomography (CT) and/or cerebrospinal fluid (CSF) laboratory analysis were available, and initiation of therapy only after patient admission to an inpatient hospital bed compared with initiation of therapy in the ED.
METHODS The medical records of all patients discharged from two study hospitals with the discharge diagnosis of bacterial meningitis from January 1, 1981, through February 29, 1988, were reviewed. One hospital (A) is a 700bed, urban university teaching hospital with an ED census of approximately 44,000 patients a year. Patients were seen primarily by emergency medicine, internal medicine, or pediatric housestaff who were directly supervised by an attending physician (or occasionally by a senior emergency medicine resident). The o t h e r h o s p i t a l (B) is a 1,000-bed, university-affiliated urban c o m m u nity hospital with an ED census of approximately 50,000 patients a year. Patients were initially evaluated exclusively by attending e m e r g e n c y physicians. Both hospitals had routine laboratory services and CT scanning available 24 hours a day. 106/857
8
Cases were included for analysis if patients had the hospital discharge diagnosis of bacterial meningitis, were primarily evaluated in the ED, were admitted to the hospital with the ED diagnosis of meningitis, and were subsequently treated with antibiotic therapy. Patients were thus included on the basis of clinical suspicion of bacterial meningitis in the ED that was confirmed during hospitalization; we did not attempt to independently validate the diagnosis of bacterial m e n i n g i t i s by reviewing CSF laboratory parameters or CSF culture. Cases were excluded from analysis if the diagnosis of meningitis was considered before ED presentation. Thus, patients referred to the ED for w o r k u p or t r e a t m e n t of possible meningitis from either another outpatient facility or another hospital were not included. The medical records of each patient were reviewed for demographic, historical, physical examination, laboratory, and time data. The location of the patient (hospital bed or ED) when antibiotics were initiated was also noted. Data were extracted from triage, nurse's and physician's notes, registration and informed consent documents, and laboratory reports. The time of patient registration at the ED, placement in an examination room, first encounter with a physician, performance of various procedures (eg, lumbar puncture [LP] and transport to CT scan), reporting of CSF laboratory results, and time of initial parenteral antibiotic administration were recorded. At each hospital, the time of patient registration is routinely recorded by an ED triage nurse. The time that a patient is placed in a room and first seen by a physician is also generally noted in Annals of Emergency Medicine
the nursing records. The time of LP and transport to CT scan is routinely noted in either the physician's or nurse's record or on the informed consent form. The time that laboratory tests are returned or called to the physician is noted in the physician's or nurse's notes or on the laboratory requisition. The time that antibiotics are star'ted is r o u t i n e l y noted in the nurse's or physician's notes.
The charts were reviewed by one of three authors, and all were secondarily reviewed by the first author. When data were not recorded for a particular variable, we did not include the case in question in the analysis of that variable. The following definitions were used. An immunocompromised state was defined as the preexistence of cancer, diabetes, or primary immunocompromising disease; concurrent use of corticosteroids; or known infection with h u m a n i m m u n o d e f i ciency virus. Mental status was considered either normal, irritable, confused, lethargic, or coma based on the medical record description. Gross appearance of the CSF was defined as the description recorded by the physician or, when this was not available, as the description recorded on the laboratory report. The busiest time of the day was defined as the period from 3:00 PM tO 11:00 PM. This was based on a previous analysis of ED census of approximately 2,000 patients at each hospital (unpublished data, DA Talan, 1988). Cases were placed in two groups a c c o r d i n g to w h e t h e r a n t i b i o t i c s were withheld until the results of CT scan, CSF laboratory analysis, or both were available to the clinician. Of six separate management scenarios identified, there were three in which anti18:8 August 1989
TABLE 1. Relationship of time to antibiotics (median, interquartile range) to registration time for 122 patients with
suspected bacterial meningitis Hospital A (55) % Daily Census Time to Antibiotics (hr) Registration Time
Expected
Actual
7:00 AM to 3:00 PM
41
53
3:00 PM to 11:00 PM 11:00 PM to 7:00 AM
46 12
38 9
Hospital B (67) % Daily Census Time to Antibiotics (hr) Expected
Actual
2.8 (1.2 and 4.1)
35
31
2.1 (1.5 and 4.0)
3.5 (2.1 and 5.2) 3.0 (2.0 and 5.5)
46 19
45 24
3.3 (1,6 and 5.3) 3.1 (1,5 and 3.7)
P > .5/3 (Bonferroni correction) for all pairwise comparisons by Mann-Whitney U test.
biotics were not delayed to obtain these test results: performance of LP followed directly by initiation of antibiotics (LP/antibiotics), initiation of a n t i b i o t i c s w i t h s u b s e q u e n t LP (antibiotics/LP), and initiation of antibiotics followed by the performance of CT scan and then LP (antibiotics/ CT/LP). Patients managed by any of these three approaches comprised group 1. There were three management scenarios in which antibiotics were delayed until test results of CT scan, CSF laboratory analysis, or both were available to the clinician: performance of LP followed by report of laboratory analysis of CSF and subsequent initiation of antibiotics (LP/LP results/antibiotics), performance of CT scan followed by LP with subsequent initiation of antibiotics (CT/ LP/antibiotics), and performance of CT scan followed by the initiation of antibiotics with subsequent LP (CT/ antibiotics/LP). Patients managed by any of these three approaches comprised group 2. All data analyses were performed using the SPSS/PC+ statistical software package. Data entry into the computer was subject to two independent verification techniques to eliminate errors. The distribution of values of the dependent variable (the time from ED registration until initiation of parenteral antibiotics, or time to antibiotics) was skewed to the right. Therefore, for all comparisons i n v o l v i n g this variable, the Mann-Whitney U test was used. For comparisons involving nominal data, the Xz test was used. All data are summarized as the median followed by the 25th to 75th percentile interquartile range (in parentheses). Statistical significance was defined as P < .05 unless otherwise indicated. 18:8 August 1989
RESULTS Of 507 cases reviewed with the discharge diagnosis of bacterial meningitis, 122 met the necessary criteria and were analyzed; 55 cases (45%) were seen at hospital A and 67 (55%) were seen at hospital B. The following demographic, clinical, and laboratory data are provided to characterize the study group. Eleven percent of patients were less than 2 months old, 45% were between 2 months and 2 years old, 22% were between 2 and 18 years old, 17% were between 18 and 60 years old, and 5% were more than 60 years old. Fifty-seven percent of study patients were men; 40% were Caucasian, 26% were black, 29% were Hispanic, and 5% were Asian. Recorded patient symptoms were fever (93%), stiff neck (19%), vomiting (55%), headache (24%), oral antibiotics used within 48 hours before presentation (24%), and history of an i m m u n o c o m p r o m i s e d state
I8%). The following signs on physical examination were recorded: 0ra] temperature of 38.0 C or higher (81%); m e n t a l status recorded as n o r m a l (14%), irritable (29%), confused (5%), lethargic (47%), or comatose (5%); nuchal rigidity (62%); and a bulging anterior fontanelle in 13 of 49 patients (27%) with an open fontanelle. Fifty-nine percent of patients had a peripheral white blood cell count of m o r e t h a n 1 0 , 0 0 0 c e l l s / m m 3, Seventy-seven percent of patients were noted to have cloudy CSF. The CSF laboratory findings were CSF white blood cells more than 500 c e l l s / m m 3, 7 3 % ; CSF p o l y morphonuclear leukocytes more than 50%, 90%; CSF/serum glucose less than 50%, 79%; CSF Grampositive stain, 61%; and positive CSF bacterial culture, 80%. At least one
Annals of Emergency Medicine
of the above CSF findings was present in 94% of the patients. The etiologic bacteria were 65 cases of Haemophilus influenzae (53%), 15 Streptococcus pneumoniae (12%), eight other Streptococcus species (7%), three Listeria monocytogenes (3%), two Pseudomonas species (2%), two Staphylococcus aureus (2%), and one Mycoplasma pneumoniae (1%). For the total study group, the time from patient registration until initiation of antibiotic therapy was 3.0 (1.6 and 4.3 hours, respectively) hours. The d i s t r i b u t i o n of these times is shown (Figure 1). There was no significant difference in the time to antibiotics between hospital A (3.0 [1.7 and 4.2] hours) and hospital B (3.1 [1.6 and 4.3] hours). The percentage of patients registered between 3:00 PM and lh00 PM, 7:00 AM and 3:00 PM, and 11:00 PM and 7:00 AM was 42%, 41%, and 17%, respectively. Time to antibiotics was not significantly related to time of registration in either hospital ED (Table 1). The times from patient registration to placement in an examination room, from placement in an examination room until first encounter with a physician, and from first physician encounter to initiation of antibiotics are shown (Figure 2). The majority (90%) of the time before init i a t i o n of a n t i b i o t i c s ( m e d i a n percentage of time to antibiotics) occurred after the initial physician encounter. Time to antibiotics for each management scenario is shown (Table 2). Time to antibiotics was significantly longer for patients in group 2 as compared with those in group 1 (P < .00005). (All comparisons of managem e n t scenario groups were made with a Bonferroni correction with a divisor of 4 lie, P = .05/4].) 858/107
MENINGITIS MANAGEMENT Talan et al
FIGURES 3A, B, and C. Median time of m a n a g e m e n t steps after initial p h y s i c i a n e n c o u n t e r for p a t i e n t s with suspected bacterial meningitis (by s c e n a r i o ) . A: S c e n a r i o LP/ antibiotics (66); B: scenario LP/LP results/antibiotics (32); C: scenario CT/LP/antibiotics (11). MD-LP, time from initial physician encounter to performance of LP; LP-AB, time from performance of LP to initiation of antit~otics; LP-LP res, time from performance of LP to return of laboratory CSF analysis; LP res-AB, time from return of laboratory CSF analysis to initiation of antibiotics; MD-CT, time from initial physician encounter to transport to CT scan; CT-LP, time from transport to CT scan to performance of LP. The three m o s t c o m m o n l y used m a n a g e m e n t s c e n a r i o s w e r e LP/ antibiotics, LP/LP resuhs/antibiotics, and C T / L P / a n t i b i o t i c s , a c c o u n t i n g for 90% of total cases. The times of management steps after initial physician encounter for each of these scenarios are shown (Figures 3A, 3B, and 3C). For the m a n a g e m e n t scenario LP/antibiotics, time from initial physician e n c o u n t e r to performance of LP was 38 (14 and 86) minutes, and time from performance of LP to initiation of antibiotics was 40 (20 and 75) minutes (Figure 3A). T i m e to antibiotics was signific a n t l y longer for the m a n a g e m e n t s c e n a r i o LP/LP r e s u l t s / a n t i b i o t i c s than for the scenario LP/antibiotics (4.3 [3.2 and 5.9] versus 2.1 [1.3 and 3.5] hours, respectively; P < .00005). For this scenario, the time from initial physician encounter to LP was 30 (15 and 64) minutes; from LP to receipt of LP results, 95 (61 and 109) minutes; and from receipt of LP results to initiation of antibiotics, 58 {20 and 124) m i n u t e s (Figure 3B). These delays occurred despite the fact that the gross appearance of the CSF was noted to be cloudy in 19 of the 32 cases (59%) managed by this scenario. T i m e to antibiotics was significantly longer for cases in which CT scan preceded the initiation of antibiotics compared with those in which CT scan was not used or followed antibiotics (3.9 [3.1 and 6.6] versus 2.8 [1.5 and 4.0] hours, respectively; P = .005). Although the n u m b e r of cases was small, the time to antibiotics was also longer for cases managed by 108/859
120
80 ..m t--
40
m
MD-LP
3A
LP-AB
120
8O 09 :3 ¢-
40
3B
MD-LP
L P - L P res
LP r e s - A B
MD-CT
CT-LP
LP-AB
I
120
8O (,9
*5 c" 40
3C
the scenario CT/LP/antibiotics compared with the scenario antibiotics/ CT/LP (3.9 [3.0 and 6.3] versus 1.5 [0.8 and 3.1] hours, respectively; P .02, NS). For the CT/LP/antibiotics Annals of Emergency Medicine
m a n a g e m e n t scenario, the time from initial physician encounter to transport to CT scan was 70 (50 and 155) minutes; from transport to CT scan to LP, 92 (65 and 175) minutes; and 18:8 August 1989
TABLE 2. Relationship of time to antibiotics (median, interquartile range) to
management scenarios of 122 patients with suspected bacterial meningitis Group 1
N
%
Time to Antibiotics (hr)
LP/antibiotics Antibiotics/LP
67 6
55 5
2.1 (1.3 and 3.5) 1.2 (0.9 and 2.9)
Antibiotics/CT/LP
4
3
1.5 (0.8 and 3.1)
77
63
1.9 (1.2 and 3.4)*
LP/LP results/antibiotics
32
26
4.3 (3.2 and 5.9)
CT/LP/antibiotics CT/antibiotics/LP
11 2
9 2
3.9 (3.0 and 6.3) 6.4
All group 2 45 *P < .00005 by Mann-Whitney U test.
37
All group 1 Group 2
from LP to antibiotic initiation, 40 (15 and 58) minutes (Figure 3C). Time to antibiotics was significantly longer if the patient received initial antibiotic treatment in the hospital inpatient ward compared w i t h in the ED (4.5 [3.5 and 6.8] versus 2.2 [1.4 and 3.5] hours, respectively; P < .00005). The practice of the initiation of antibiotics on the ward occurred in 32% of the cases overall and was significantly more frequent (P = .017) at the community hospital (42% of patients) than at the university hospital (20% of patients). Because both the location of initial treatment and the performance of CT scan, CSF laboratory analysis, or both before initiation of antibiotics {group 2) were significantly associated with the time to antibiotics, the independent contribution of each of these factors to time to antibiotics was evaluated. Time to antibiotics rem a i n e d significantly.longer w h e n treatment was initiated on the ward compared with in the ED, regardless of whether patients were in group 1 or group 2 (P ~< .0003 for each comparison). Alternately, for patients given antibiotics in the ED, the time to antibiotics was significantly (P < .00005) longer for cases in which CT scan, CSF laboratory analysis, or both preceded antibiotics {group 2) compared with cases in which antibiotics were not delayed for these test results {group 1); this was not true, however, for patients initially treated on the ward (P = NS). 18:8 August 1989
4.3 (3.2 and 6.0)*
DISCUSSION We examined the relationship of various aspects of m a n a g e m e n t to the time period from ED registration until the initiation of antibiotics. These cases were representative of those managed in modern busy emergency facilities with full laboratory and radiologic services. Although we did not attempt to confirm the diagnosis of bacterial meningitis by the usual gold standards such as CSF parameters or culture results, the age distribution, clinical presentation, and laboratory findings of these cases were consistent with previous series of p a t i e n t s w i t h b a c t e r i a l m e n ingitis.S s In addition, 94% of the cases had CSF parameters consistent with bacterial meningitis. The bacterial etiology of meningitis was predominantly H influenzae (53%) and S pneumoniae (12%)and was remarkable only for the lack of cases secondary to Neisseria meningitidis. Consistent with previous reports, the time from patient registration until the administration of parenteral antibiotics was significantly longer than the recommended "30 minute" textbook standard, with a median time of 3.0 hours and a range from 30 m i n u t e s to 18 hours {Figure 1). 1-3 Most patients were seen at the times of the day for which the ED census is greatest. We suspected that patients presenting at the busiest times might require longer to be seen and evaluated; however, there was no significant relationship between the time to antibiotics to the time of presentaAnnals of Emergency Medicine
tion {Table l). In fact, regardless of the time of arrival, patients generally were triaged into examination rooms and seen by physicians relatively quickly. These two steps accounted for only 10% of the total time to antibiotics {Figure 2). Therefore, the prolonged time to antibiotics does not appear to be due to long stays in the waiting room, mistriage, or inability of physicians to attend to these patients even during the busiest times of the day. The greatest component of management time until initiation of antibiotics followed the initial physician encounter with the patient (Figure 2). Six m a n a g e m e n t s c e n a r i o s were identified in which certain critical diagnostic tests (ie, LP, CT scan, or CSF analysis) were employed after the physician's initial evaluation. As we hypothesized, scenarios in which CT scan and/or CSF laboratory analysis were used before initiation of antibiotics were associated with a significant delay of therapy {Table 1). Time to antibiotics was significantly shorter when antibiotics were administered immediately after the performance of LP than when antibiotics were withheld pending laboratory analysis of CSF (P < .00005). Although it is impossible to be certain when the physician first suspected bacterial meningitis in each case, a delay of this magnitude could potentially have been avoided if antibiotics had been administered immediately after the LP. Surprisingly, we found that 19 of the 32 cases (59%) in which CSF analysis was obtained before therapy had grossly cloudy CSF. Cloudy CSF generally correlates with at least 500 WBC/mm 3 or high bacterial concentrations and is consistent with a high likelihood of bacterial meningitis. Certainly if suspicion of bacterial meningitis is great enough to p r o m p t performance of LP and cloudy CSF is discovered, it seems unreasonable to delay antibiotics until laboratory CSF analysis is complete. The contraindications to LP and the coinciding indications for CT scan before this procedure have been debated. In our study, antibiotics were significantly delayed for cases in which CT scan was employed before the initiation of antibiotics compared with all other cases (P = .005). This practice was not infrequent, occurring in 11% of the cases, and the 860/109
MENINGITIS MANAGEMENT Talan et al
median time from transport to CT scan to the performance of LP was substantial (92 minutes). A recent review concluded that if LP is delayed to obtain a CT scan, immediate empiric antibiotics before both of these procedures are indicated. 9 In the four cases in which this strategy was used (antibiotics/CT/LP), the median time to antibiotics was 1.5 hours, which was significantly less than that of the scenario CT/LP/antibiotics. Because of selection bias inherent in this study, it is impossible to know the frequency with which CT scanning detected other intracranial diagnoses in patients with clinical findings initially c o n s i s t e n t with meningitis. However, if the diagnosis of bacterial meningitis is considered and a CT scan is believed to be necessary before LP, the magnitude of the delay involved in withholding antibiotics until LP is performed suggests that antibiotics should be initiated before CT scanning. In view of the fact that the performance of LP and the interpretation of diagnostic tests results do not generally require a significant amount of time, it was surprising to find that after these events, antibiotics were not administered for as long as 40 to 60 minutes (Figures 3A, 3B, and 3C). This was not because antibiotics were unavailable, as these EDs were stocked with appropriate drugs. The most striking association with the time to antibiotics was the location of the patient when antibiotics were initiated. Patients who received their initial therapy in the ED were treated within a median time of 2.2 compared with 4.5 hours for those patients who were not treated until admission to an inpatient hospital bed (P < .00005). A substantial prop o r t i o n of p a t i e n t s in this s t u d y (32%) were not treated until they were admitted to the ~vard, despite the fact that all were admitted with the diagnosis of suspected bacterial meningitis. This practice was significantly more common at the private community hospital than at the university hospital (42% versus 20%, respectively) and may reflect the emergency physician's deference of the choice of antibiotics to the admitting pediatrician or internist. It is clear, however, that admission to the ward and reintroduction to a new nursing staff and physician requires a significant amount of time, which is a po110/861
tentially avoidable delay in the therapy of a patient with suspected bacterial meningitis. The location of antibiotic initiation was s i g n i f i c a n t l y a s s o c i a t e d with the time to antibiotics regardless of whether therapy was delayed to obtain CT scan or CSF laboratory results. On the other hand, the performance of these diagnostic procedures before antibiotics was associated with a significant delay of therapy only for patients treated in the ED. This suggests that when patients are not treated until after hospital admission, this practice is the overriding factor associated with delay of antibiotic therapy. Any conclusions and recommendations to be drawn from our data depend on the assumptive premise that more rapid treatment will improve the outcome of patients with bacterial meningitis. There are no ethical means to study this question directly, and no consistent relationship between a longer duration of symptoms and a worsening outcome can be derived from available indirect evidence. 9 Nevertheless, the overall mortality of bacterial meningitis has been reduced from virtually 100% to 5% to 10% since the advent of effective antibiotics. Thus, it is certainly logical to conclude that more rapid diagnosis and treatment will only improve outcome. Whether a delay of several hours before initiating antibiotic therapy is significantly deleterious must remain a matter of speculation, but attempts to diminish delay can only be beneficial and in certain cases may well be critical. A prospective study of the management of bacterial meningitis, while usually an ideal design for accurate data collection, is problematic for several reasons. A prospective study of c o n s e c u t i v e patients with suspected bacterial meningitis, being a relatively uncommon problem, would require expensive and prolonged "silent" surveillance. Even if this could be done, the prospective observation of patient management would very likely cause a change in nurse and physician behavior, which could invalidate the study. Although a retrospective study is the most practical method to evaluate the m a n a g e m e n t of suspected bacterial meningitis, this design also has i n h e r e n t l i m i t a t i o n s . In our study, the only cases selected were Annals of Emergency Medicine
those with the discharge diagnosis of bacterial meningitis; therefore, other cases initially suspected of having bacterial meningitis but not subsequently discharged with this diagnosis would have been missed. To the extent that diagnostic procedures such as computed tomography scan and CSF laboratory analysis led to other diagnoses, this might justify withholding antibiotics pending the results of the tests. However, in light of the small risk associated with a single dose of parenteral antibiotics and the potential consequences of a substantial delay of effective therapy for bacterial meningitis, this strategy would not appear logical. It was also not possible to know the physician's initial clinical impression of each patient and the extent to which this affected the decision to withhold antibiotics pending t i m e - c o n s u m i n g d i a g n o s t i c procedures. Thus, delays may have occurred because some patients presented with subtle of confusing findings for w h i c h the d i a g n o s i s of bacterial meningitis was not initially obvious until further diagnostic evaluation was completed. However, the demographic profile and the clinical and laboratory findings of the study patients were typical of those commonly appreciated to be suggestive of bacterial meningitis. Additionally, on review of clinical signs noted in the medical record, in general we found no significant relationship between time to antibiotics and overt signs of meningitis. For example, time to antibiotics was not significantly related to the presence or absence of a temperature of 39.0 C or higher in combination with either of the findings of nuchal rigidity, abnormal mental status, or a bulging fontanelle {unpublished data, DA Talan, 1989). The fact that dramatic delays were related to the location of antibiotic initiation regardless of the m a n a g e m e n t strategy employed is also consistent with the analysis that clinical presentation was not a prominent factor accounting for the prolonged time before therapy. CONCLUSION The textbook "standard of care" suggesting administration of antibiotics within 30 minutes of encountering patients with suspected bacterial meningitis may not be reasonable in modern busy EDs. Generally, 18:8 August 1989
it takes approximately 20 minutes for a p a t i e n t to b e r e g i s t e r e d , e v a l u a t e d b y a n u r s e , t r i a g e d to a r o o m , and first seen by a physician. However, t h i s s t a g e of t h e e v a l u a t i o n acc o u n t s for o n l y a m i n o r p o r t i o n of the total time until antibiotic thera p y is t y p i c a l l y i n i t i a t e d to s u c h pat i e n t s ; t h e g r e a t e s t a m o u n t of m a n a g e m e n t t i m e a p p e a r s to b e p h y s i c i a n g e n e r a t e d a n d to a g r e a t e x t e n t represents potentially avoidable delay. I n i t i a t i o n of a n t i b i o t i c s i n t h e E D i n all p a t i e n t s a d m i t t e d w i t h s u s p e c t e d b a c t e r i a l m e n i n g i t i s w o u l d result in substantially more rapid treatment. Similarly, antibiotics should n o t be r o u t i n e l y w i t h h e l d p e n d i n g dia g n o s t i c t e s t r e s u l t s . W h e n LP is del a y e d to o b t a i n a C T s c a n a n d b a c t e rial m e n i n g i t i s is a d i a g n o s t i c possibility, therapy should be started
18:8 August 1989
i m m e d i a t e l y . E v e n w h e n LP is n o t d e l a y e d , a n a l y s i s of C S F r e q u i r e s a s i g n i f i c a n t a m o u n t of t i m e , a n d init i a t i o n of a n t i b i o t i c s b e f o r e a v a i l a b i l i t y of l a b o r a t o r y r e s u l t s s h o u l d b e s t r o n g l y c o n s i d e r e d , p a r t i c u l a r l y if t h e g r o s s a p p e a r a n c e of t h e CSF is cloudy.
agement of pediatric patients with bacterial meningitis in the emergency department. Drug Intell Clin Pharm 1986;20:796-798.
We acknowledge the expert m a n u s c r i p t preparation of Ms Imelda Angelo and Ms Cynthia Heyer as well as the consultation of Dr Larry Baraff.
6. Carpenter RR, Petersdorf RG: The clinical spectrum of bacterial meningitis. A m } Med 1962;33:262-275.
REFERENCES 1. McGee ZA, Kaiser AB: Acute meningitis, in Mandell GL, Douglas RG Jr, Bennett JE {eds): Principles and Practice of Infectious Diseases,
ed 2. New York, John Wiley and Sons, 1985, p 566. 2. Bryan CS, Reynolds KL, Crout L: Promptness of antibiotic therapy in acute bacterial meningitis. Ann Emerg Med 1986;15:544-547. 3. Nahata MC, Arnoto RT, PoweI1 DA: Man-
Annals of Emergency Medicine
4. Rogers JT: Risk ManagemelTt in Emergency Medicine, ed 1. Dallas, American College of Emergency Physicians, 1985, p 1-36. 5. Geisler PJ, Nelson KE, Levin S, et ah Community-acquired purulent meningitis: A review of 1,316 cases during the antibiotic era, 1954-1976. Rev Infect Dis 1980;2:725W45.
7. Finland M~ Barnes MW: Acute bacterial meningitis at Boston City Hospital during 12 selected years, 1935 1972. J Infect Dis 1977;136: 400-415. 8. Gorse GJ, Thrupp LD, Nudleman KL, et al: Bacterial meningitis in the elderly. Arch Intern Med I984;144:1603-1607. 9. Talan DA, Hoffman JR, Yoshikawa TT, et al: Role of empiric parenteral antibiotics prior to lumbar puncture in suspected bacterial meningitis: State of the art. Rev Infect Dis 1988;10: 365-376.
862/111
Analysis of Emergency Department Management of Suspected Bacterial Meningitis Previous studies of emergency department management of bacterial meningitis have indicated that there are often long delays before initiation of antibiotics. The purpose of our study was to determine whether these delays were related to specific aspects of patient management. From 1981 through 1988, we retrospectively reviewed the medical records of 122 patients primarily evaluated in the ED and admitted for suspected bacterial meningitis at a university (55) and a c o m m u n i t y (67) hospital. The median time (interquartile range) from ED registration until initiation of antibiotics (time to antibiotics) was 3.0 hours (1.6 and 4.3 hours, respectively) (total range, 0.5 to 18 hours). The time to antibiotics was not significantly related to the time of ED registration. Ninety percent of the total time to antibiotics occurred after the initial physician encounter. Time to antibiotics was significantly (P < .00005) longer for patients in whom computed tomography scan and/or laboratory analysis of cerebrospinaI fluid preceded initiation of antibiotics compared w~th patients in whom antibiotic administration was not contingent on the results of these procedures (4.3 [3.2 and 6.0] versus 1.9 [1.2 and 3.4] hours, respectively). Also, time to antibiotics was significantly (P < .00005) longer for patients in whom antibiotics were initiated on the ward as compared with in the ED (4.5 [3.5 and 6.8] versus 2.2 [1.4 and 3.5] hours, respectively). We conclude that long delays exist in the ED before initiation of antibiotics for cases of suspected bacterial meningitis, and that in general these delays appear to be physician generated and to a great extent potentially avoidable. [Talan DA, Guterman JJ, Overturf GD, Singer C, Hoffman JR, Lambert B: Analysis of emergency department m a n a g e m e n t of suspected bacterial meningitis. Ann Emerg Med August 1989;18:856-862.]
David A Talan, MD* Jeffrey J Guterman, MD* Gary D Overturf, MDt Craig Singer, MD*¢ Jerome R Hoffman, MD§ Bret Lambert, MD§ Los Angeles, California From the Departments of Emergency Medicine* and Pediatrics,t Olive View/ UCLA Medical Center; UCLA/Olive View Emergency Medicine Residency;¢ and the Department of Medicine, Division of Emergency Medicine, UCLA Medical Center,§ Los Angeles, California. Received for publication December 13, 1988. Revision received March 27, 1989. Accepted for publication April 4, 1989. Presented at the 2rid International Conference on Emergency Medicine in Brisbane, Australia, October 1988. Address for reprints: David A I-alan, MD, Department of Emergency Medicine, Olive View/UCLA Medical Center, 14445 Olive View Drive, Sylmar, California 91342.
INTRODUCTION Bacterial meningitis is typically a rapidly progressive, life-threatening illness that may be amenable to cure by the prompt administration of appropriate antibiotics. As such, bacterial meningitis constitutes an infectious disease emergency. Textbooks of infectious diseases emphasize the need for rapid diagnosis and treatment and state that antibiotics should be administered within 30 minutes of encountering the patient. 1 However,
recent studies indicate that, in actual practice, therapy is typically not given for several hours.2, 3 Bryan et al z reported a series of 135 patients with bacterial meningitis in whom the mean time from emergency department patient registration to administration of antibiotics for pediatric and adult cases was 2.1 and 4.9 hours, respectively. Delayed diagnosis and therapy may have serious prognostic implications and may also account in part for the fact that bacterial meningitis ranks second in total malpractice dollars paid per case originating from EDs. 4 The factors that contribute to longer than expected times until therapy remain to be elucidated. We report a retrospective review of the ED management of 122 children and adults admitted with the diagnosis of suspected bacterial meningitis. The purpose of our study was to determine which aspects of ED management of these patients were associated with delays of antibiotic therapy. We hypothesized that delays were associated with care during the busiest time of the day, the practice of withholding antibiotics until the results of
18:8 August 1989
Annals of Emergency Medicine
856/105
MENINGITIS MANAGEMENT Talan et al
20
150
15 ¢o
"b
100 10
5O
0 1 |
2
3
4
5
6
7
9
10
18
2
Reg-Room
Room-MD
M D - AB
Time to Antibiotics (hr)
FIGURE 1. Time to antibiotics for 122 patients with suspected bacterial meningitis. FIGURE 2. Median time of managem e n t steps for patients w i t h susp e c t e d bacterial m e n i n g i t i s (120). Reg-Roorn ED, time from registration to p l a c e m e n t in an e x a m i n a t i o n room; Room-MD, time from placement in an examination room to initial physician encounter; MD-AB, time from initial physician encounter to initiation of antibiotics. computed tomography (CT) and/or cerebrospinal fluid (CSF) laboratory analysis were available, and initiation of therapy only after patient admission to an inpatient hospital bed compared with initiation of therapy in the ED.
METHODS The medical records of all patients discharged from two study hospitals with the discharge diagnosis of bacterial meningitis from January 1, 1981, through February 29, 1988, were reviewed. One hospital (A) is a 700bed, urban university teaching hospital with an ED census of approximately 44,000 patients a year. Patients were seen primarily by emergency medicine, internal medicine, or pediatric housestaff who were directly supervised by an attending physician (or occasionally by a senior emergency medicine resident). The o t h e r h o s p i t a l (B) is a 1,000-bed, university-affiliated urban c o m m u nity hospital with an ED census of approximately 50,000 patients a year. Patients were initially evaluated exclusively by attending e m e r g e n c y physicians. Both hospitals had routine laboratory services and CT scanning available 24 hours a day. 106/857
8
Cases were included for analysis if patients had the hospital discharge diagnosis of bacterial meningitis, were primarily evaluated in the ED, were admitted to the hospital with the ED diagnosis of meningitis, and were subsequently treated with antibiotic therapy. Patients were thus included on the basis of clinical suspicion of bacterial meningitis in the ED that was confirmed during hospitalization; we did not attempt to independently validate the diagnosis of bacterial m e n i n g i t i s by reviewing CSF laboratory parameters or CSF culture. Cases were excluded from analysis if the diagnosis of meningitis was considered before ED presentation. Thus, patients referred to the ED for w o r k u p or t r e a t m e n t of possible meningitis from either another outpatient facility or another hospital were not included. The medical records of each patient were reviewed for demographic, historical, physical examination, laboratory, and time data. The location of the patient (hospital bed or ED) when antibiotics were initiated was also noted. Data were extracted from triage, nurse's and physician's notes, registration and informed consent documents, and laboratory reports. The time of patient registration at the ED, placement in an examination room, first encounter with a physician, performance of various procedures (eg, lumbar puncture [LP] and transport to CT scan), reporting of CSF laboratory results, and time of initial parenteral antibiotic administration were recorded. At each hospital, the time of patient registration is routinely recorded by an ED triage nurse. The time that a patient is placed in a room and first seen by a physician is also generally noted in Annals of Emergency Medicine
the nursing records. The time of LP and transport to CT scan is routinely noted in either the physician's or nurse's record or on the informed consent form. The time that laboratory tests are returned or called to the physician is noted in the physician's or nurse's notes or on the laboratory requisition. The time that antibiotics are star'ted is r o u t i n e l y noted in the nurse's or physician's notes.
The charts were reviewed by one of three authors, and all were secondarily reviewed by the first author. When data were not recorded for a particular variable, we did not include the case in question in the analysis of that variable. The following definitions were used. An immunocompromised state was defined as the preexistence of cancer, diabetes, or primary immunocompromising disease; concurrent use of corticosteroids; or known infection with h u m a n i m m u n o d e f i ciency virus. Mental status was considered either normal, irritable, confused, lethargic, or coma based on the medical record description. Gross appearance of the CSF was defined as the description recorded by the physician or, when this was not available, as the description recorded on the laboratory report. The busiest time of the day was defined as the period from 3:00 PM tO 11:00 PM. This was based on a previous analysis of ED census of approximately 2,000 patients at each hospital (unpublished data, DA Talan, 1988). Cases were placed in two groups a c c o r d i n g to w h e t h e r a n t i b i o t i c s were withheld until the results of CT scan, CSF laboratory analysis, or both were available to the clinician. Of six separate management scenarios identified, there were three in which anti18:8 August 1989
TABLE 1. Relationship of time to antibiotics (median, interquartile range) to registration time for 122 patients with
suspected bacterial meningitis Hospital A (55) % Daily Census Time to Antibiotics (hr) Registration Time
Expected
Actual
7:00 AM to 3:00 PM
41
53
3:00 PM to 11:00 PM 11:00 PM to 7:00 AM
46 12
38 9
Hospital B (67) % Daily Census Time to Antibiotics (hr) Expected
Actual
2.8 (1.2 and 4.1)
35
31
2.1 (1.5 and 4.0)
3.5 (2.1 and 5.2) 3.0 (2.0 and 5.5)
46 19
45 24
3.3 (1,6 and 5.3) 3.1 (1,5 and 3.7)
P > .5/3 (Bonferroni correction) for all pairwise comparisons by Mann-Whitney U test.
biotics were not delayed to obtain these test results: performance of LP followed directly by initiation of antibiotics (LP/antibiotics), initiation of a n t i b i o t i c s w i t h s u b s e q u e n t LP (antibiotics/LP), and initiation of antibiotics followed by the performance of CT scan and then LP (antibiotics/ CT/LP). Patients managed by any of these three approaches comprised group 1. There were three management scenarios in which antibiotics were delayed until test results of CT scan, CSF laboratory analysis, or both were available to the clinician: performance of LP followed by report of laboratory analysis of CSF and subsequent initiation of antibiotics (LP/LP results/antibiotics), performance of CT scan followed by LP with subsequent initiation of antibiotics (CT/ LP/antibiotics), and performance of CT scan followed by the initiation of antibiotics with subsequent LP (CT/ antibiotics/LP). Patients managed by any of these three approaches comprised group 2. All data analyses were performed using the SPSS/PC+ statistical software package. Data entry into the computer was subject to two independent verification techniques to eliminate errors. The distribution of values of the dependent variable (the time from ED registration until initiation of parenteral antibiotics, or time to antibiotics) was skewed to the right. Therefore, for all comparisons i n v o l v i n g this variable, the Mann-Whitney U test was used. For comparisons involving nominal data, the Xz test was used. All data are summarized as the median followed by the 25th to 75th percentile interquartile range (in parentheses). Statistical significance was defined as P < .05 unless otherwise indicated. 18:8 August 1989
RESULTS Of 507 cases reviewed with the discharge diagnosis of bacterial meningitis, 122 met the necessary criteria and were analyzed; 55 cases (45%) were seen at hospital A and 67 (55%) were seen at hospital B. The following demographic, clinical, and laboratory data are provided to characterize the study group. Eleven percent of patients were less than 2 months old, 45% were between 2 months and 2 years old, 22% were between 2 and 18 years old, 17% were between 18 and 60 years old, and 5% were more than 60 years old. Fifty-seven percent of study patients were men; 40% were Caucasian, 26% were black, 29% were Hispanic, and 5% were Asian. Recorded patient symptoms were fever (93%), stiff neck (19%), vomiting (55%), headache (24%), oral antibiotics used within 48 hours before presentation (24%), and history of an i m m u n o c o m p r o m i s e d state
I8%). The following signs on physical examination were recorded: 0ra] temperature of 38.0 C or higher (81%); m e n t a l status recorded as n o r m a l (14%), irritable (29%), confused (5%), lethargic (47%), or comatose (5%); nuchal rigidity (62%); and a bulging anterior fontanelle in 13 of 49 patients (27%) with an open fontanelle. Fifty-nine percent of patients had a peripheral white blood cell count of m o r e t h a n 1 0 , 0 0 0 c e l l s / m m 3, Seventy-seven percent of patients were noted to have cloudy CSF. The CSF laboratory findings were CSF white blood cells more than 500 c e l l s / m m 3, 7 3 % ; CSF p o l y morphonuclear leukocytes more than 50%, 90%; CSF/serum glucose less than 50%, 79%; CSF Grampositive stain, 61%; and positive CSF bacterial culture, 80%. At least one
Annals of Emergency Medicine
of the above CSF findings was present in 94% of the patients. The etiologic bacteria were 65 cases of Haemophilus influenzae (53%), 15 Streptococcus pneumoniae (12%), eight other Streptococcus species (7%), three Listeria monocytogenes (3%), two Pseudomonas species (2%), two Staphylococcus aureus (2%), and one Mycoplasma pneumoniae (1%). For the total study group, the time from patient registration until initiation of antibiotic therapy was 3.0 (1.6 and 4.3 hours, respectively) hours. The d i s t r i b u t i o n of these times is shown (Figure 1). There was no significant difference in the time to antibiotics between hospital A (3.0 [1.7 and 4.2] hours) and hospital B (3.1 [1.6 and 4.3] hours). The percentage of patients registered between 3:00 PM and lh00 PM, 7:00 AM and 3:00 PM, and 11:00 PM and 7:00 AM was 42%, 41%, and 17%, respectively. Time to antibiotics was not significantly related to time of registration in either hospital ED (Table 1). The times from patient registration to placement in an examination room, from placement in an examination room until first encounter with a physician, and from first physician encounter to initiation of antibiotics are shown (Figure 2). The majority (90%) of the time before init i a t i o n of a n t i b i o t i c s ( m e d i a n percentage of time to antibiotics) occurred after the initial physician encounter. Time to antibiotics for each management scenario is shown (Table 2). Time to antibiotics was significantly longer for patients in group 2 as compared with those in group 1 (P < .00005). (All comparisons of managem e n t scenario groups were made with a Bonferroni correction with a divisor of 4 lie, P = .05/4].) 858/107
MENINGITIS MANAGEMENT Talan et al
FIGURES 3A, B, and C. Median time of m a n a g e m e n t steps after initial p h y s i c i a n e n c o u n t e r for p a t i e n t s with suspected bacterial meningitis (by s c e n a r i o ) . A: S c e n a r i o LP/ antibiotics (66); B: scenario LP/LP results/antibiotics (32); C: scenario CT/LP/antibiotics (11). MD-LP, time from initial physician encounter to performance of LP; LP-AB, time from performance of LP to initiation of antit~otics; LP-LP res, time from performance of LP to return of laboratory CSF analysis; LP res-AB, time from return of laboratory CSF analysis to initiation of antibiotics; MD-CT, time from initial physician encounter to transport to CT scan; CT-LP, time from transport to CT scan to performance of LP. The three m o s t c o m m o n l y used m a n a g e m e n t s c e n a r i o s w e r e LP/ antibiotics, LP/LP resuhs/antibiotics, and C T / L P / a n t i b i o t i c s , a c c o u n t i n g for 90% of total cases. The times of management steps after initial physician encounter for each of these scenarios are shown (Figures 3A, 3B, and 3C). For the m a n a g e m e n t scenario LP/antibiotics, time from initial physician e n c o u n t e r to performance of LP was 38 (14 and 86) minutes, and time from performance of LP to initiation of antibiotics was 40 (20 and 75) minutes (Figure 3A). T i m e to antibiotics was signific a n t l y longer for the m a n a g e m e n t s c e n a r i o LP/LP r e s u l t s / a n t i b i o t i c s than for the scenario LP/antibiotics (4.3 [3.2 and 5.9] versus 2.1 [1.3 and 3.5] hours, respectively; P < .00005). For this scenario, the time from initial physician encounter to LP was 30 (15 and 64) minutes; from LP to receipt of LP results, 95 (61 and 109) minutes; and from receipt of LP results to initiation of antibiotics, 58 {20 and 124) m i n u t e s (Figure 3B). These delays occurred despite the fact that the gross appearance of the CSF was noted to be cloudy in 19 of the 32 cases (59%) managed by this scenario. T i m e to antibiotics was significantly longer for cases in which CT scan preceded the initiation of antibiotics compared with those in which CT scan was not used or followed antibiotics (3.9 [3.1 and 6.6] versus 2.8 [1.5 and 4.0] hours, respectively; P = .005). Although the n u m b e r of cases was small, the time to antibiotics was also longer for cases managed by 108/859
120
80 ..m t--
40
m
MD-LP
3A
LP-AB
120
8O 09 :3 ¢-
40
3B
MD-LP
L P - L P res
LP r e s - A B
MD-CT
CT-LP
LP-AB
I
120
8O (,9
*5 c" 40
3C
the scenario CT/LP/antibiotics compared with the scenario antibiotics/ CT/LP (3.9 [3.0 and 6.3] versus 1.5 [0.8 and 3.1] hours, respectively; P .02, NS). For the CT/LP/antibiotics Annals of Emergency Medicine
m a n a g e m e n t scenario, the time from initial physician encounter to transport to CT scan was 70 (50 and 155) minutes; from transport to CT scan to LP, 92 (65 and 175) minutes; and 18:8 August 1989
TABLE 2. Relationship of time to antibiotics (median, interquartile range) to
management scenarios of 122 patients with suspected bacterial meningitis Group 1
N
%
Time to Antibiotics (hr)
LP/antibiotics Antibiotics/LP
67 6
55 5
2.1 (1.3 and 3.5) 1.2 (0.9 and 2.9)
Antibiotics/CT/LP
4
3
1.5 (0.8 and 3.1)
77
63
1.9 (1.2 and 3.4)*
LP/LP results/antibiotics
32
26
4.3 (3.2 and 5.9)
CT/LP/antibiotics CT/antibiotics/LP
11 2
9 2
3.9 (3.0 and 6.3) 6.4
All group 2 45 *P < .00005 by Mann-Whitney U test.
37
All group 1 Group 2
from LP to antibiotic initiation, 40 (15 and 58) minutes (Figure 3C). Time to antibiotics was significantly longer if the patient received initial antibiotic treatment in the hospital inpatient ward compared w i t h in the ED (4.5 [3.5 and 6.8] versus 2.2 [1.4 and 3.5] hours, respectively; P < .00005). The practice of the initiation of antibiotics on the ward occurred in 32% of the cases overall and was significantly more frequent (P = .017) at the community hospital (42% of patients) than at the university hospital (20% of patients). Because both the location of initial treatment and the performance of CT scan, CSF laboratory analysis, or both before initiation of antibiotics {group 2) were significantly associated with the time to antibiotics, the independent contribution of each of these factors to time to antibiotics was evaluated. Time to antibiotics rem a i n e d significantly.longer w h e n treatment was initiated on the ward compared with in the ED, regardless of whether patients were in group 1 or group 2 (P ~< .0003 for each comparison). Alternately, for patients given antibiotics in the ED, the time to antibiotics was significantly (P < .00005) longer for cases in which CT scan, CSF laboratory analysis, or both preceded antibiotics {group 2) compared with cases in which antibiotics were not delayed for these test results {group 1); this was not true, however, for patients initially treated on the ward (P = NS). 18:8 August 1989
4.3 (3.2 and 6.0)*
DISCUSSION We examined the relationship of various aspects of m a n a g e m e n t to the time period from ED registration until the initiation of antibiotics. These cases were representative of those managed in modern busy emergency facilities with full laboratory and radiologic services. Although we did not attempt to confirm the diagnosis of bacterial meningitis by the usual gold standards such as CSF parameters or culture results, the age distribution, clinical presentation, and laboratory findings of these cases were consistent with previous series of p a t i e n t s w i t h b a c t e r i a l m e n ingitis.S s In addition, 94% of the cases had CSF parameters consistent with bacterial meningitis. The bacterial etiology of meningitis was predominantly H influenzae (53%) and S pneumoniae (12%)and was remarkable only for the lack of cases secondary to Neisseria meningitidis. Consistent with previous reports, the time from patient registration until the administration of parenteral antibiotics was significantly longer than the recommended "30 minute" textbook standard, with a median time of 3.0 hours and a range from 30 m i n u t e s to 18 hours {Figure 1). 1-3 Most patients were seen at the times of the day for which the ED census is greatest. We suspected that patients presenting at the busiest times might require longer to be seen and evaluated; however, there was no significant relationship between the time to antibiotics to the time of presentaAnnals of Emergency Medicine
tion {Table l). In fact, regardless of the time of arrival, patients generally were triaged into examination rooms and seen by physicians relatively quickly. These two steps accounted for only 10% of the total time to antibiotics {Figure 2). Therefore, the prolonged time to antibiotics does not appear to be due to long stays in the waiting room, mistriage, or inability of physicians to attend to these patients even during the busiest times of the day. The greatest component of management time until initiation of antibiotics followed the initial physician encounter with the patient (Figure 2). Six m a n a g e m e n t s c e n a r i o s were identified in which certain critical diagnostic tests (ie, LP, CT scan, or CSF analysis) were employed after the physician's initial evaluation. As we hypothesized, scenarios in which CT scan and/or CSF laboratory analysis were used before initiation of antibiotics were associated with a significant delay of therapy {Table 1). Time to antibiotics was significantly shorter when antibiotics were administered immediately after the performance of LP than when antibiotics were withheld pending laboratory analysis of CSF (P < .00005). Although it is impossible to be certain when the physician first suspected bacterial meningitis in each case, a delay of this magnitude could potentially have been avoided if antibiotics had been administered immediately after the LP. Surprisingly, we found that 19 of the 32 cases (59%) in which CSF analysis was obtained before therapy had grossly cloudy CSF. Cloudy CSF generally correlates with at least 500 WBC/mm 3 or high bacterial concentrations and is consistent with a high likelihood of bacterial meningitis. Certainly if suspicion of bacterial meningitis is great enough to p r o m p t performance of LP and cloudy CSF is discovered, it seems unreasonable to delay antibiotics until laboratory CSF analysis is complete. The contraindications to LP and the coinciding indications for CT scan before this procedure have been debated. In our study, antibiotics were significantly delayed for cases in which CT scan was employed before the initiation of antibiotics compared with all other cases (P = .005). This practice was not infrequent, occurring in 11% of the cases, and the 860/109
MENINGITIS MANAGEMENT Talan et al
median time from transport to CT scan to the performance of LP was substantial (92 minutes). A recent review concluded that if LP is delayed to obtain a CT scan, immediate empiric antibiotics before both of these procedures are indicated. 9 In the four cases in which this strategy was used (antibiotics/CT/LP), the median time to antibiotics was 1.5 hours, which was significantly less than that of the scenario CT/LP/antibiotics. Because of selection bias inherent in this study, it is impossible to know the frequency with which CT scanning detected other intracranial diagnoses in patients with clinical findings initially c o n s i s t e n t with meningitis. However, if the diagnosis of bacterial meningitis is considered and a CT scan is believed to be necessary before LP, the magnitude of the delay involved in withholding antibiotics until LP is performed suggests that antibiotics should be initiated before CT scanning. In view of the fact that the performance of LP and the interpretation of diagnostic tests results do not generally require a significant amount of time, it was surprising to find that after these events, antibiotics were not administered for as long as 40 to 60 minutes (Figures 3A, 3B, and 3C). This was not because antibiotics were unavailable, as these EDs were stocked with appropriate drugs. The most striking association with the time to antibiotics was the location of the patient when antibiotics were initiated. Patients who received their initial therapy in the ED were treated within a median time of 2.2 compared with 4.5 hours for those patients who were not treated until admission to an inpatient hospital bed (P < .00005). A substantial prop o r t i o n of p a t i e n t s in this s t u d y (32%) were not treated until they were admitted to the ~vard, despite the fact that all were admitted with the diagnosis of suspected bacterial meningitis. This practice was significantly more common at the private community hospital than at the university hospital (42% versus 20%, respectively) and may reflect the emergency physician's deference of the choice of antibiotics to the admitting pediatrician or internist. It is clear, however, that admission to the ward and reintroduction to a new nursing staff and physician requires a significant amount of time, which is a po110/861
tentially avoidable delay in the therapy of a patient with suspected bacterial meningitis. The location of antibiotic initiation was s i g n i f i c a n t l y a s s o c i a t e d with the time to antibiotics regardless of whether therapy was delayed to obtain CT scan or CSF laboratory results. On the other hand, the performance of these diagnostic procedures before antibiotics was associated with a significant delay of therapy only for patients treated in the ED. This suggests that when patients are not treated until after hospital admission, this practice is the overriding factor associated with delay of antibiotic therapy. Any conclusions and recommendations to be drawn from our data depend on the assumptive premise that more rapid treatment will improve the outcome of patients with bacterial meningitis. There are no ethical means to study this question directly, and no consistent relationship between a longer duration of symptoms and a worsening outcome can be derived from available indirect evidence. 9 Nevertheless, the overall mortality of bacterial meningitis has been reduced from virtually 100% to 5% to 10% since the advent of effective antibiotics. Thus, it is certainly logical to conclude that more rapid diagnosis and treatment will only improve outcome. Whether a delay of several hours before initiating antibiotic therapy is significantly deleterious must remain a matter of speculation, but attempts to diminish delay can only be beneficial and in certain cases may well be critical. A prospective study of the management of bacterial meningitis, while usually an ideal design for accurate data collection, is problematic for several reasons. A prospective study of c o n s e c u t i v e patients with suspected bacterial meningitis, being a relatively uncommon problem, would require expensive and prolonged "silent" surveillance. Even if this could be done, the prospective observation of patient management would very likely cause a change in nurse and physician behavior, which could invalidate the study. Although a retrospective study is the most practical method to evaluate the m a n a g e m e n t of suspected bacterial meningitis, this design also has i n h e r e n t l i m i t a t i o n s . In our study, the only cases selected were Annals of Emergency Medicine
those with the discharge diagnosis of bacterial meningitis; therefore, other cases initially suspected of having bacterial meningitis but not subsequently discharged with this diagnosis would have been missed. To the extent that diagnostic procedures such as computed tomography scan and CSF laboratory analysis led to other diagnoses, this might justify withholding antibiotics pending the results of the tests. However, in light of the small risk associated with a single dose of parenteral antibiotics and the potential consequences of a substantial delay of effective therapy for bacterial meningitis, this strategy would not appear logical. It was also not possible to know the physician's initial clinical impression of each patient and the extent to which this affected the decision to withhold antibiotics pending t i m e - c o n s u m i n g d i a g n o s t i c procedures. Thus, delays may have occurred because some patients presented with subtle of confusing findings for w h i c h the d i a g n o s i s of bacterial meningitis was not initially obvious until further diagnostic evaluation was completed. However, the demographic profile and the clinical and laboratory findings of the study patients were typical of those commonly appreciated to be suggestive of bacterial meningitis. Additionally, on review of clinical signs noted in the medical record, in general we found no significant relationship between time to antibiotics and overt signs of meningitis. For example, time to antibiotics was not significantly related to the presence or absence of a temperature of 39.0 C or higher in combination with either of the findings of nuchal rigidity, abnormal mental status, or a bulging fontanelle {unpublished data, DA Talan, 1989). The fact that dramatic delays were related to the location of antibiotic initiation regardless of the m a n a g e m e n t strategy employed is also consistent with the analysis that clinical presentation was not a prominent factor accounting for the prolonged time before therapy. CONCLUSION The textbook "standard of care" suggesting administration of antibiotics within 30 minutes of encountering patients with suspected bacterial meningitis may not be reasonable in modern busy EDs. Generally, 18:8 August 1989
it takes approximately 20 minutes for a p a t i e n t to b e r e g i s t e r e d , e v a l u a t e d b y a n u r s e , t r i a g e d to a r o o m , and first seen by a physician. However, t h i s s t a g e of t h e e v a l u a t i o n acc o u n t s for o n l y a m i n o r p o r t i o n of the total time until antibiotic thera p y is t y p i c a l l y i n i t i a t e d to s u c h pat i e n t s ; t h e g r e a t e s t a m o u n t of m a n a g e m e n t t i m e a p p e a r s to b e p h y s i c i a n g e n e r a t e d a n d to a g r e a t e x t e n t represents potentially avoidable delay. I n i t i a t i o n of a n t i b i o t i c s i n t h e E D i n all p a t i e n t s a d m i t t e d w i t h s u s p e c t e d b a c t e r i a l m e n i n g i t i s w o u l d result in substantially more rapid treatment. Similarly, antibiotics should n o t be r o u t i n e l y w i t h h e l d p e n d i n g dia g n o s t i c t e s t r e s u l t s . W h e n LP is del a y e d to o b t a i n a C T s c a n a n d b a c t e rial m e n i n g i t i s is a d i a g n o s t i c possibility, therapy should be started
18:8 August 1989
i m m e d i a t e l y . E v e n w h e n LP is n o t d e l a y e d , a n a l y s i s of C S F r e q u i r e s a s i g n i f i c a n t a m o u n t of t i m e , a n d init i a t i o n of a n t i b i o t i c s b e f o r e a v a i l a b i l i t y of l a b o r a t o r y r e s u l t s s h o u l d b e s t r o n g l y c o n s i d e r e d , p a r t i c u l a r l y if t h e g r o s s a p p e a r a n c e of t h e CSF is cloudy.
agement of pediatric patients with bacterial meningitis in the emergency department. Drug Intell Clin Pharm 1986;20:796-798.
We acknowledge the expert m a n u s c r i p t preparation of Ms Imelda Angelo and Ms Cynthia Heyer as well as the consultation of Dr Larry Baraff.
6. Carpenter RR, Petersdorf RG: The clinical spectrum of bacterial meningitis. A m } Med 1962;33:262-275.
REFERENCES 1. McGee ZA, Kaiser AB: Acute meningitis, in Mandell GL, Douglas RG Jr, Bennett JE {eds): Principles and Practice of Infectious Diseases,
ed 2. New York, John Wiley and Sons, 1985, p 566. 2. Bryan CS, Reynolds KL, Crout L: Promptness of antibiotic therapy in acute bacterial meningitis. Ann Emerg Med 1986;15:544-547. 3. Nahata MC, Arnoto RT, PoweI1 DA: Man-
Annals of Emergency Medicine
4. Rogers JT: Risk ManagemelTt in Emergency Medicine, ed 1. Dallas, American College of Emergency Physicians, 1985, p 1-36. 5. Geisler PJ, Nelson KE, Levin S, et ah Community-acquired purulent meningitis: A review of 1,316 cases during the antibiotic era, 1954-1976. Rev Infect Dis 1980;2:725W45.
7. Finland M~ Barnes MW: Acute bacterial meningitis at Boston City Hospital during 12 selected years, 1935 1972. J Infect Dis 1977;136: 400-415. 8. Gorse GJ, Thrupp LD, Nudleman KL, et al: Bacterial meningitis in the elderly. Arch Intern Med I984;144:1603-1607. 9. Talan DA, Hoffman JR, Yoshikawa TT, et al: Role of empiric parenteral antibiotics prior to lumbar puncture in suspected bacterial meningitis: State of the art. Rev Infect Dis 1988;10: 365-376.
862/111