Journal of Antimicrobial Chemotherapy (1998) 42, 107111

JAC

Antimicrobial practice The effect of intravenous-to-oral switch guidelines on the use of parenteral antimicrobials in medical wards
R. B. S. Lainga*, A. R. Mackenziea, H. Shawb, I. M. Gouldc and J. G. Douglasa
a

The Infection Unit and bDepartment of Pharmacy, Aberdeen Royal Inrmary, Foresterhill, Aberdeen AB25 2ZB; cDepartment of Medical Microbiology, University of Aberdeen, Aberdeen, UK
The effect of an intravenous (iv)-to-oral switch policy on antibiotic prescribing in general medical wards was examined. Three audits, each of 2 months duration, were carried out to examine the duration of iv therapy and length of patient stay. The rst audit (S1) was performed before the introduction of switch guidelines, the second (S2) after guidelines had been placed in patient case-notes and the third (S3) after the guidelines had been introduced into the drug charts. The duration of iv therapy was signicantly shorter in the S3 group (mean = 3.7 days) than in the S2 or S1 groups (mean 4.4 and 4.35 days, respectively) (P < 0.05). There was no signicant difference in the length of patient stay between the three audit periods but the stay was signicantly shorter in 81 switched patients (mean duration = 8.9 days) than in matched controls (mean duration = 12.6 days) (P = 0.01). Fewer patients with respiratory infection were treated for >24 h with iv antimicrobials in the S3 audit period (75/549) than in the S2 (85/372) and S1 audits (83/326) (P < 0.01). The introduction of switch guidelines to drug charts reduces the length of iv therapy. Switched patients spend less time in hospital than their matched controls.

Introduction
The use of oral antimicrobials for seriously ill patients is rare, even when appropriate oral agents are tolerated, since drug absorbtion may be impaired.1 As a result, many patients hospitalized with infection are initially given intravenous (iv) antimicrobials. Switching from iv to oral therapy has many potential benets, including reduction in drug cost, patient stay and hospital-related morbidity. Guidelines for switching from iv to oral antimicrobials have already been published.2,3 In this study we have examined the effect of introducing, in two different ways, guidelines for iv-to-oral switch in patients admitted to general medical wards.

Materials and methods

Three audits were carried out on patients admitted to general medical wards within Aberdeen Royal Inrmary and treated with at least 24 h of iv antimicrobials for

suspected or proven community-acquired infections. Patients given oral antimicrobials from the outset or treated for 24 h with iv antimicrobials were not included in the audits. Each audit covered a period of 2 months and there were 5 months between the rst and second audits and between the second and third audits. During each audit, the following information was gathered: the patients clinical features before starting treatment, the reason for and choice of iv antimicrobial, the incidence of iv line complications, the duration of iv antimicrobials and, where applicable, the choice and timing of oral switch therapy. Fever was considered present where the temperature was 37.5C. Neutrophilia was taken as an absolute neutrophil cell count of 7 109/L. The rst audit (S1) was carried out before the introduction of any guidelines. The second audit (S2) was performed after the introduction of ward guidelines (Figure). These advised on a suitable time to switch from iv to oral treatment and suggested appropriate antimicrobials for use in these circumstances. These guidelines were placed in the front cover of the case-notes of those

*Tel: +44-1224-681818; Fax: +44-1224-685307.

107
1998 The British Society for Antimicrobial Chemotherapy

R. B. S. Laing et al.

Figure. Selection criteria for intravenous to oral switch and suggested oral antimicrobials.

patients treated for 24 h with iv antimicrobials and completed by the attending doctor caring for the patient. The third audit (S3) was performed during a period when the same guidelines were placed alongside the patients drug prescription sheets and accompanied by a high-prole poster campaign to alert clinicians to switch guidelines. The S1 audit was carried out in September and October, S2 in April and May and S3 in December and January. Patients were said to have been switched according to the guidelines if the relevant form (Figure) had been completed and one of the suggested oral antimicrobials commenced. For the purpose of comparison, patients who had not been switched (unswitched controls) were matched to switched patients by age (within 2 years), gender, diagnosis, indication for treatment, fever and

neutrophilia. As with the switched patients, those in the unswitched control group had no exclusion criteria for oral switch. The number of patients admitted to medical wards with pneumonia, chronic obstructive pulmonary disease (COPD), urinary tract infections (UTI) or soft tissue infections (STI) were calculated for each audit period from the hospital records of coded diagnoses. Statistical analysis of categorical data was made using the 2 test with Yates correction for small values. Population means were compared by the t-test.

Results
One hundred and eleven patients were audited on antimicrobial treatment before switch guidelines were

108

Effect of switch guidelines in medical wards Table I. Comparison of the pre- (S1) and post-switch (S2 and S3) groups S1 Total number of patients No. changed from iv to oral therapy Number (%) switched according to guidelines Mean age (S.D.) Male:female ratio Duration (days)of iv therapy (S.D.) Length (days) of patient stay (S.D.) Pathogen isolated (%) Iv line complications Readmitted within 28 days Died
NA, not applicable; NS, not signicant. a Signicant change.

S2 113 85 26 (23) 64 (17) 55:59 4.4 (2.5) 10.9 (10.6) 13 (12) 35 2 3

S3 103 83 77 (75)a 65 (16) 42:61 3.7 (2.1)a 10.8 (7.8) 7 (7) 32 2 5

P value NS NS 0.01a NS NS 0.05a NS NS NS NS NS

111 79 NA 65 (17) 50:61 4.35 (3.0) 11.6 (9.9) 11 (10) 25 3 4

Table II. Timing of iv-to-oral switch in audit groups S1 to S3 Number of patients switched Time of switch (days) 2 3 4 5 6 6 Total S1 5 10 22 19 8 15 79 S2 3 12 20 16 10 29 85 S3 7 12 29 12 10 13 83

Table III. Number of patients from the major diagnostic categories treated with iv antimicrobials for 24 h in audit groups S1 to S3; the total number of admissions to medical wards is given in parentheses Diagnostic category S1 S2 S3 P value

Respiratory infection 84 (326) 83 (372) 75 (549)a 0.001a UTI 6 (64) 5 (100) 9 (99) NS Soft tissue infection 5 (15) 4 (10) 5 (16) NS
a

Signicant change after introduction of switch guidelines to drug charts (S3).

introduced (S1), 113 patients following the introduction of guidelines in the case-notes (S2) and 103 patients following the introduction of guidelines in the drug charts (S3). Table I shows the mean age, sex ratio and clinical

features in each group. The age range and diagnoses were similar in all three audits. A similar proportion of patients in each audit group received oral antimicrobials as followon from iv therapy. After the introduction of guidelines (S2 and S3) a signicantly higher proportion of patients was switched in accordance with the guideline recommendations during the S3 audit than the S2 audit. There was no signicant difference in the duration of patient stay between the three groups. The length of iv treatment was signicantly shorter in the S3 audit than in the other two groups (P 0.05). Patient outcome (death or readmission) did not differ signicantly between the three audits. The timing of iv-to-oral change for the three audits is shown in Table II. Compared with S1 or S2, a higher proportion of patients in the S3 audit were changed to an oral agent on or before day 4 of treatment. The difference between the S3 proportion (48/83) and S2 proportion (35/85) of those changed on or before day 4 reached statistical signicance (P 0.05). The numbers of patients from the principal diagnostic groups who had been given 24 h of iv antimicrobials are shown in Table III. The total number admitted to medical wards with these diagnoses is shown in the same table. Patients with respiratory infection (pneumonia or COPD) were signicantly less likely to be treated with 24 h of iv antimicrobials after the introduction of switch guidelines to drug charts (S3). The choice of oral agents given as follow-on treatment for respiratory infection did not change signicantly following the introduction of switch guidelines. During S1, 57/65 patients given oral antimicrobials received follow-on treatment with amoxycillin, erythromycin or co-amoxiclav, i.e. those agents recommended on the guidelines sheet. The proportion during S2 was 50/62 and during S3 the proportion was 54/59.

109

R. B. S. Laing et al. During the S2 and S3 audits, guideline forms were completed by medical staff for 103 patients; 81 of these were switched to suitable oral antimicrobials at the stage of illness suggested by the guidelines and 22 failed to meet the necessary inclusion criteria. The 81 patients who were switched to oral antimicrobials after completing switch guidelines were matched to unswitched controls. The duration of iv therapy was signicantly shorter in the switched group (mean 2.8 days, S.D. 1.5) than in the controls (mean 4.4 days, S.D. 2.4; P < 0.001). Patient stay was also shorter in the switched patients (mean 8.9 days, S.D. 6) than the controls (mean 12.6 days, S.D. 11.3; P = 0.01). The incidence of line complications was lower in the switched patients (17/81) than in the controls (26/81) although this difference did not reach statistical signicance (P 0.077). The duration of oral antimicrobials used as follow-on therapy in the switched patients (mean 6.1 days, S.D. 3.2) was similar to that in the control group (mean 5.4 days, S.D. 11.5; P 0.6). of parenteral antimicrobial treatment for respiratory infection. This would suggest that the switch guidelines had affected clinical practice by reducing the number of patients treated with iv antimicrobials or prompting a switch to oral therapy within 24 h of admission. It is notable that, amongst patients with respiratory infection, the use of those oral agents suggested in the guidelines (amoxycillin, co-amoxiclav or erythromycin) did not differ signicantly before and after the introduction of switch guidelines. This would suggest that the guidelines inuenced the timing, but not the choice, of oral switch. These results might be further improved by a higher uptake of switch guidelines. It is known that the use of coloured stickers attached to drug charts has proven effective in encouraging clinicians to use specic oral agents in preference to their iv formulation8 and a similar approach to switch guidelines may give better results than were achieved by the approach used in the S2 audit. The effect of guidelines on reducing patient stay was evident in the analysis comparing switched patients with controls but not in the comparison of the three audit groups. It seems probable that a signicant proportion of patients were kept in hospital following oral switch for reasons other than the treatment of their infection. It has been previously reported that concomitant medical conditions are the commonest reasons for patients to remain in hospital after switching to oral antimicrobials9 and it may be that such conditions diluted the effect of oral switch on length of stay. Our experience suggests that signicant reductions in the duration of iv therapy can follow the introduction of guidelines to the drug charts of suitable patients. The latter also has the effect of reducing the number of patients receiving 24 h of iv therapypresumably as a consequence of heightened awareness amongst clinicians of selection criteria for oral therapy.

Discussion
The term switch therapy is best used to describe the conversion of parenteral therapy to oral therapy as soon as the patient is judged suitable and without loss of antimicrobial potency.4 In this study the term was used to encompass all forms of iv-to-oral conversion regardless of the relative potency of each agent. Selection for oral switch is based on clinical improvement, tolerance of oral medication, ability to absorb oral medication and the absence of any clinical indication (e.g. endocarditis) to continue iv therapy. 5 It has been estimated that, based on these criteria, about 75% of patients who are hospitalized with an infection are eligible for switch.6 In this study the selection criteria for switch therapy were in keeping with the aforementioned principles and were similar to those previously employed elsewhere.7 Choosing a suitable oral agent based on bacterial isolates is often difcult since, as in this study, the diagnosis is frequently clinical with relatively few patients having positive microbiology. The reasons for this, though not formally assessed here, include previous antimicrobial treatment in the community and empirical therapy being given in hospital before collection of samples suitable for culture. The introduction of guidelines to the front of patient case-notes resulted in a completion rate of only 23% of forms and it was therefore unsurprising that the S2 patients did not receive iv therapy for a shorter period than the S1 group. A much higher response was seen when the guidelines were introduced in the patients drug charts along with posters reminding medical and nursing staff to consider switch. Not only was the latter associated with a reduction in the length of iv therapy but there was also a signicant fall in the number of patients receiving 24 h

References
1. DEspine, M., Bellido, F., Pechere, J. C., Auckenthaler, R., Rohner, P., Lew, D. et al. (1989). Serum levels of ciprooxacin after single oral doses in patients with septicaemia. European Journal of Clinical Microbiology and Infectious Diseases 8, 101923. 2. Ramirez, J. A. (1994). Advances in antibiotic use: switch therapy. Current Therapeutic Research: Clinical and Experimental 55, Suppl. A, 304. 3. Allen, B., Naismith, N. W., Mauser, A. J. & Moulds, R. F. W. (1992). A campaign to improve the timing of conversion from intravenous to oral administration of antibiotics. Australian Journal of Hospital Pharmacy 22, 4349. 4. Nathwani, D., Tillotson, G. & Davey, P. (1997). Sequential antimicrobial therapythe role of quinolones. Journal of Antimicrobial Chemotherapy 39, 4416. 5. Hamilton-Miller, J. M. T. (1996). Switch therapy: the theory and practice of early change from parenteral to non-parenteral antibiotic administration. Clinical Microbiology and Infection 2, 129.

110

Effect of switch guidelines in medical wards

6. Quintiliani, R., Nightingale, C. H., Crowe, H. M., Cooper, B. W., Bartlett, R. C. & Gousse, G. (1991). Strategic antibiotic decisionmaking at the formulary level. Reviews of Infectious Diseases 13, Suppl. 9, S7707. 7. Brooks, M. (1995). IV to oral antibiotic switch therapythe Beaumont experience. Irish Pharmacy Journal, 2256. 8. Frighetto, L., Nickoloff, D., Martinusen, S. M., Mamdani, F. S. & Jewesson, P. J. (1992). Intravenous to oral stepdown program: four years experience in a large teaching hospital. Annals of Pharmacotherapy 26, 144751. 9. Grasela, T. H., Paladino, J. A., Schentag, J. J., Huepenbecker, D., Rybacki, J., Purcell, J. B. et al . (1991). Clinical and economic impact of oral ciprooxacin as follow-up to parenteral antibiotics. Annals of Pharmacotherapy 25, 85762. Received 28 July 1997; returned 17 September 1997; revised 27 October 1997; accepted 5 February 1998

111