Brief Observations| Volume 106, ISSUE 1, P114-116, January 1999

Ok

# Is an observation period necessary after intravenous antibiotics are changed to oral administration?

Open AccessPublished:August 16, 2004
After a change is made from intravenous to oral antibiotics, patients are frequently observed in the hospital for 24 hours before discharge (
• Farber B.F
• Cooper D.J
Observing patients after antibiotics are discontinued.
). This practice arises from concerns about intolerance to the oral antibiotic therapy, side effects, or worsening of the infection upon oral therapy.
We believed that adverse events were infrequent during the 24 hours after the route of administration of an antibiotic is changed from intravenous to oral in hospitalized patients. We also postulated that the likelihood of identifying a therapeutic failure in such a short time would be low, even if there is inadequate absorption of the oral form of the antibiotic drug.
Other studies have addressed the utility of inpatient observation of patients after cessation or change of antibiotics (
• Birken G.A
• Schropp K.P
• Boles E.T
• King D.R
Discharge planning for children with perforated appendicitis.
,
• Nelson J.D
• McCracken Jr, G.H
Recrudescence and relapse in bacterial meningitis of childhood.
). A retrospective study by Caceres et al (
• Caceres V.M
• Stange K.C
• Kikano G.E
• Zyzanski S.J
The clinical utility of a day of hospital observation after switching from intravenous to oral antibiotic therapy in the treatment of pyelonephritis.
) found limited utility of observation after a change from intravenous to oral antibiotic therapy in 138 patients admitted with acute pyelonephritis. Another retrospective study (
• Rhew D.C
• Hackner D
• Henderson L
• Ellrodt A.G
• Weingarten S.R
The clinical benefit on in-hospital observation in low-risk pneumonia patients after conversion from parenteral to oral antimicrobial therapy.
) analyzed the validity of a 24-hour observation period after changing from intravenous to oral antibiotics in 144 low-risk patients with community-acquired pneumonia. They also concluded that this practice is of limited benefit.
We conducted a prospective observational study at the University of Pennsylvania Medical Center to determine the frequency of adverse events during the 24-hour period after antibiotic therapy is changed from intravenous to oral. Additionally, we attempted to identify risk factors that might predispose patients to adverse events and to estimate the annual cost of this practice at our institution.

## Subjects and methods

A prospective observational study was performed from October 1, 1993, to February 1, 1994, on adult in-patient units at the University of Pennsylvania Medical Center, a 750-bed tertiary care hospital. All hospital pharmacy orders were screened daily, and we identified any patient who had an antibiotic order changed from intravenous to oral route of administration of the same. The patient’s medical record was reviewed, and the patient’s physician was interviewed within 48 hours. Patients were included in the study only if their hospitalization was prolonged solely for the purpose of observation after the change from intravenous to oral antibiotic. This reason was verified by discussion with the patient’s physician.
We collected data on the patient’s age, sex, race, and length of hospitalization; type, indication, and length of time on antibiotics; vital signs on the day of the switch from intravenous to oral antibiotic; and any adverse event that occurred after the antibiotic route of administration was changed. Adverse events were defined as temperature ≥38°, nausea, vomiting, diarrhea, rash, relapse or worsening of the infectious process, or any event that resulted in prolongation of hospitalization. After the patients were discharged, we contacted the patient’s physician to obtain follow-up on the course of the adverse events for those patients that developed them. Re-admission to our hospital during the subsequent 30 days was obtained from medical records. Additionally, a record was made of diagnostic tests that were performed during the observation period.
We compared categorical variables using chi-square tests and continuous variables using t tests. We estimated 95% confidence intervals for rates of adverse event. All analyses were performed using EpiInfo (Centers for Disease Control and Prevention). Statistical significance was set at P <0.05, two tailed.

## Results

A total of 204 consecutive patients were reviewed. Ninety-four met the case definition and were included in the study (Table 1). Of the 110 patients who were not included, 99 had an additional reason for the hospital stay; for 11 patients the physician or medical record was unavailable.
Table 1Characteristics of 94 Patients Observed following Discontinuation of Antibiotic Therapy
Plus-minus values are mean ± SD.
CharacteristicNo. (%)
Male gender38 (40)
Race
Black54 (57)
White37 (39)
Other3 (4)
Service
Medicine50 (53)
Surgery17 (18)
Obstetrics14 (15)
Oncology7 (7)
Urology6 (6)
Age, years48.7 ± 20
Plus-minus values are mean ± SD.
The most common reasons for prescribing antibiotics are displayed in Table 2. In 55% of the patients, the antibiotic therapy was empirical. Ninety-four percent of the patients had a comorbid condition, including cancer (n = 19), diabetes (n = 11), cardiovascular disease (n = 27), asthma (n = 7), alcohol abuse (n = 6), pregnancy (n = 5), and human immunodeficiency virus infection (n = 4).
Table 2Reasons for Antibiotic Therapy
Reason for AntibioticNo. of PatientsNosocomial InfectionCommunity-acquired Infection
Urinary tract infection27621
Soft tissue infection19613
Pneumonia17710
Intra-abdominal infection624
Bone/joint infection404
Sinusitis303
Wound infection101
Fever of unknown origin101
Other918
Prophylaxis7
Total942265
The mean duration of intravenous antibiotic treatment was 5.6 days (range, 1 to 26 days). In all patients the antimicrobial spectrum of action of the intravenous and oral antibiotics were similar. Seven patients had a temperature >38°C on the day that intravenous antibiotics were changed to oral therapy, but only one of these patients remained febrile during the 24-hour observation period. None of these seven patients developed any other adverse event. Five patients (5%) developed an adverse event during the 24-hour observation period (Table 3). An adverse event that could potentially be attributed to the oral antibiotic occurred in two patients (2%; 95% CI, 0.3% to 7%). In four patients with an adverse event, no change in the oral antibiotic therapy was made, and the adverse event resolved. In only one patient (patient 3), a change in oral antibiotic treatment was made during the observation period (95% CI, 0.3% to 6%). No therapeutic failures were identified during the 24-hour observation period (95% CI, 0% to 4%). There were no readmissions during the next 30 days.
Table 3Outcomes of Patients with Adverse Events
No change in antibiotic
CellulitisNauseaOral antibiotic therapyDischarge delayed for observation
Resolved without change of antibiotic
CellulitisRashOral antibiotic therapyAntibiotic changed
Discharged
PneumoniaChest painSickle cell crisisDischarge delayed
No change in antibiotic
AbscessFeverMetastatic cancerDischarged
No change in antibiotic
We were unable to identify any risk factors for the development of an adverse event, including the indication for antibiotic therapy, length of intravenous therapy, type of antibiotic drug, fever on the day of the switch, age, or underlying disease.
The average per diem hospital nursing cost during the study period was $294 for direct costs related to patient care delivery and$590 for full costs including indirect costs, such as overhead. (In 1998 the average per diem full nursing cost was $741.) Based on the number of patients observed during the 4-month study period, we extrapolated that this practice cost approximately$166,000 in 1993 to 1994, and approximately $209,000 in 1998. Thirty-seven patients in the study (40%) had blood drawn on the day that followed the change in antibiotic administration. The most commonly ordered tests were complete blood counts and serum chemistry tests. ## Discussion Physicians often prolong a patient’s hospitalization by 24 hours following a change from intravenous to oral antibiotic administration with the intention of monitoring for intolerance to the oral formulation as well as the therapeutic failure. We extrapolated that at our institution about 280 hospital days per year are required by this practice. In this study, the 95% confidence interval for adverse events potentially related to an oral antibiotic during the observation period was 0.3% to 7%. No therapeutic failures were identified (upper 95% confidence limit of 4%), and no important sequelae occurred. A change in the oral antibiotic therapy was made in only one patient. Additionally, each adverse event could have been amenable to out-patient management. The current estimated cost of the observation practice at our institution is more than$200,000 per year.
Interestingly, diagnostic blood tests were performed on 40% of the patients in our study, although their hospitalization was prolonged solely for observation during the 24 hours after the change from intravenous to oral antibiotic treatment. Most of the tests were ordered before the decision to observe the patient (eg, regularly scheduled blood tests), and we were not able to discern any clear indication for them.
Due to the small sample size and infrequent occurrence of adverse events, we were not able to detect any potential risk factors. Another limitation of our study is that only patients who were observed in the hospital were followed. It is possible that patients who are discharged immediately after a change in the route of administration or type of antibiotic therapy have a different incidence of adverse events. Factors that influenced individual decisions to observe patients rather than discharge them were not known to us. We believe, however, that this bias could result in the selection of sicker patients who are more likely to develop an adverse event, as prolonged hospitalization for observation was one of the criteria for selecting study participants. The design of this study is also limited by reliance on unblinded physician reporting at a single institution. As we collected data during a 4-month period and extrapolated to estimate the rate of adverse events during 1 year, seasonal variations may limit the accuracy of our estimate.
The upper limit of 95% confidence interval for all adverse events in our study was 12%. The fairly broad confidence interval is a reflection of the small sample size. Nevertheless, our results provide important information for clinicians. At present, given the infrequent occurrence of adverse events in addition to their benign nature, routine observation of a patient in-hospital following a change from intravenous to oral antibiotic administration does not appear justified.

## References

• Farber B.F
• Cooper D.J
Observing patients after antibiotics are discontinued.
NEJM. 1995; 333 ([Letter]): 1083-1084
• Birken G.A
• Schropp K.P
• Boles E.T
• King D.R
Discharge planning for children with perforated appendicitis.
J Pediatr Surg. 1986; 67: 592-595
• Nelson J.D
• McCracken Jr, G.H
Recrudescence and relapse in bacterial meningitis of childhood.
Pediatrics. 1981; 67: 188-195
• Caceres V.M
• Stange K.C
• Kikano G.E
• Zyzanski S.J
The clinical utility of a day of hospital observation after switching from intravenous to oral antibiotic therapy in the treatment of pyelonephritis.
J Fam Prac. 1994; 39: 337-339
• Rhew D.C
• Hackner D
• Henderson L
• Ellrodt A.G
• Weingarten S.R
The clinical benefit on in-hospital observation in low-risk pneumonia patients after conversion from parenteral to oral antimicrobial therapy.
Chest. 1998; 113: 142-146