556813
research-article2014
AOPXXX10.1177/1060028014556813Annals of PharmacotherapyRosini et al
Research Report
A Randomized Trial of Loading Vancomycin in the Emergency Department
Annals of Pharmacotherapy 1–8 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1060028014556813 aop.sagepub.com
Jamie M. Rosini, PharmD1, Julie Laughner, MD1, Brian J. Levine, MD1,2, Mia A. Papas, PhD3, John F. Reinhardt, MD1, and Neil B. Jasani, MD1,2
Abstract Background: Optimizing vancomycin dosing may help eradicate bacteria while avoiding resistance. The guidelines recommend loading doses; however, there are no data to demonstrate that this may result in a more rapid achievement of therapeutic troughs. Objective: To evaluate the percentage of troughs reaching therapeutic levels at 12, 24, and 36 hours following an initial vancomycin dose of 30 mg/kg compared with 15 mg/kg. Methods: This prospective, randomized study was performed in a community academic medical center. Patients who were to receive vancomycin in the emergency department were randomized to an initial traditional dose of 15 mg/kg or a 30-mg/kg loading dose followed by 15 mg/kg every 12 hours for 3 doses. Patients weighing >120 kg or with creatinine clearances 10 mg/L and > 15 mg/L in the traditional-dose group compared with the odds in the loading-dose group.13,14 Treatment group, time after initial dosing, and the interaction between treatment group and time were included in all models. A statistically significant group by time interaction is interpreted as a significant difference in the vancomycin levels over time between the 2 groups. β Coefficients and 95% CIs were computed, and all effect estimates were adjusted for age, gender, weight, baseline creatinine, and creatinine clearance. Model-based standard errors (Wald statistics) were used to assess statistical significance. Statistical analyses were conducted using Stata version 12.0.15
Results Patient Enrollment and Characteristics A total of 99 patients were enrolled in the study, with 49 patients in the traditional-dose group and 50 patients in the loading-dose group. Patient enrollment is described in Figure 1. There were no statistically significant differences
Downloaded from aop.sagepub.com at UNIV OF KANSAS MEDICAL CENTER on October 31, 2014
4
Annals of Pharmacotherapy
Table 1. Baseline Characteristics of Study Participants by Treatment Group.a Age (SD) Gender, male (%) Weight, kg (SD) CrCl, mL/min (SD) Average length of stay, days (SD) Admitting diagnoses, n (%) SSSI Pneumonia Bacteremia Meningitis Osteomyelitis Neutropenic fever Empiric therapy Sepsis
15-mg/kg Traditional Dose, n = 49
30-mg/kg Loading Dose n = 50
58.4 (15.6) 25 (51.0) 86.7 (20.3) 104.1 (55.8) 7.08 (7.6)
57.5 (14.5) 28 (56.0) 81.6 (19.0) 93.4 (28.1) 5.87 (4.9)
14 (28.6) 12 (24.5) 3 (6.1) 3 (6.1) 3 (6.1) 1 (2.0) 13 (26.5) 6 (12.2)
16 (32.0) 14 (28.0) 6 (12.0) 0 (0) 0 (0) 2 (4.0) 12 (24.0) 6 (12.0)
Abbreviations: CrCl, creatinine clearance; SSSI, skin soft structure infection. a P was not significant for each baseline characteristic.
in any baseline characteristics between the 2 groups (Table 1). The most common admitting diagnosis was skin and skin structure infection (SSSI) followed by pneumonia. In all, 12 patients met criteria for a sepsis alert, and they were distributed equally between both groups. A sepsis alert was initiated if the initial lactate was above 4.0 mmol/L or the patient had 2 or more systemic inflammatory response syndrome criteria with a suspected infectious source. Patients in the traditional-dose group received an average initial vancomycin dose of 1308 mg (SD = 307), in comparison to 2433 mg (SD = 546) in those receiving a loading dose. Trough levels when available and appropriately drawn were included as a data point despite the number of doses they received. In the traditional-dose group, 40, 33, and 24 trough levels were obtained at 12, 24, and 36 hours, respectively. In the loading-dose group, 35, 33, and 27 trough levels were obtained at 12, 24, and 36 hours, respectively. Two trough levels (same patient) in the traditional-dose group were above 30 mg/L, whereas no levels in the loading-dose group were above 30 mg/L.
Primary End Point Data The proportion of patients in the 2 groups reaching a minimum target level of both 10 and 15 mg/L was compared (Figure 2). The proportion of patients reaching these target levels in the traditional dose group steadily increased over a 36-hour period, whereas the proportion of patients reaching these levels in the loading-dose group decreased. When a 15-mg/L level was applied as the target, the 12-hour trough displayed a significantly greater proportion of patients
reaching the goal among those who received a 30-mg/kg dose when compared with a 15-mg/kg dose (34% vs 3%, P < 0.01). This trend continued at 24 hours but was not significantly different. When presumed steady state (36 hours or before the fourth dose) was reached, there was no difference in the percentage of patients reaching target levels between the 2 groups. When 10 mg/L was used as the target cutoff, statistically significant differences in the proportion of patients reaching this level comparing the loading-dose group with the traditional-dose group were found both at 12 hours (80% vs 15%, P < 0.01) and 24 hours (67% vs 39%, P < 0.05) after the initial dose. The mean trough values for both groups at times 12, 24, and 36 hours post–initial dose are depicted in Table 2. A significantly higher average trough level was found 12 — hours after the initial dose in the loading-dose group (X = 13.9 mg/L; SD = 4.8) compared with the traditional-dose — group (X = 7.7 mg/L; SD = 3.1; P < 0.001) and 24 hours — after the initial dose in the loading-dose group (X = 12.7 mg/L; SD = 5.7) compared with the traditional-dose group — (X = 10.2 mg/L; SD = 5.4; P value 10 mg/L
> 15 mg/L
Figure 2. Percentage of troughs >10 mg/L and >15 mg/L at times 12, 24, and 36 hours after initial dose.
Table 2. Average Trough Levels Over Time by Treatment Group.a 15 mg/kg Traditional Dose Time since initial dose 12 Hours, mean (SD)** 24 Hours, mean (SD)* 36 Hours, mean (SD) Mean difference across time 24 and 12 Hours, mean (95% CI) 36 and 24 Hours, mean (95% CI) 36 and 12 Hours, mean (95% CI)
30 mg/kg Loading Dose
7.7 (3.1) n = 40 10.2 (5.4) n = 33 13.6 (7.8) n = 24 2.7 (1.7, 3.7)* 3.2 (1.9, 4.5)* 6.2 (4.0, 8.3)*
13.9 (4.8) n = 35 12.7 (5.7) n = 33 12.3 (5.1) n = 27 −1.0 (−2.4, 0.4) 0.3 (−1.8, 1.0) −0.4 (−1.6, 0.8)
a P values based on log-transformation of vancomycin trough levels. *P < 0.05; **P < 0.001.
Paired t tests were conducted to examine pairwise comparisons of mean trough levels over time within the 2 groups. For the traditional-dose group, there were statistically significant differences in the mean trough levels between 12 and 24 hours (mean difference = 2.7 mg/L; 95% CI = 1.7, 3.7), 24 and 36 hours (mean difference = 3.2 mg/L; 95% CI = 1.9, 4.5), and 12 and 36 hours (mean difference = 6.2 mg/L; 95% CI = 4.0, 8.3). There were no differences across time for the loading-dose group (Table 2). The percentage of therapeutic troughs between non–critically ill (SSSIs) and critically ill (all other diagnoses) patients were compared. In the traditional-dose group, the percentage of levels >15 mg/L in non–critically ill patients
at 12, 24, and 36 hours were 8.3%, 16.7%, and 33.3%, respectively. There was no significant difference when compared with the critically ill patients in the traditionaldose group. The percentage of levels >15 mg/L at 12, 24, and 36 hours were 0.0%, 9.5%, and 16.7%, respectively. Similarly, there was no difference between the percentage of levels greater than 15 mg/L between the non–critically ill patients and the critically ill patients in the loading-dose group at 12, 24, and 36 hours (36.4% vs 37.5%, 20.0% vs 21.7%, and 25.0% vs 16.7%). The results from the GEE analysis examining log vancomycin trough levels as the primary outcome indicated a statistically significant difference in the log trough levels over time by dosing group (β = −0.19; 95% CI = −0.28, −0.10; P < 0.01). For the loading-dose group, log trough levels remained constant over time (β = 0.03; 95% CI = −0.03, 0.10; P = 0.27), whereas for the traditional-dose group, they steadily increased (β = 0.25; 95% CI = 0.20, 0.29; P < 0.01). When trough levels were categorized into >10 mg/L and 15 mg/L, there were statistically significant group by time interactions (β = −1.85; 95%CI = −2.67, −1.03; P < 0.01 and β = −2.82; 95% CI = −4.54, −1.11; P < 0.01, respectively). On average, the odds of reaching minimal levels to prevent resistance (10 mg/L) in the loading-dose group were 7 times greater compared with that in the traditionaldose group (OR = 7.35; 95% CI = 2.89, 18.60). This increase was slightly greater when trough levels were categorized using a 15-mg/L minimum level. Results indicated an 11-fold increase in the odds of reaching therapeutic levels (15 mg/L) in the loading-dose group compared with the
Downloaded from aop.sagepub.com at UNIV OF KANSAS MEDICAL CENTER on October 31, 2014
6
Annals of Pharmacotherapy
traditional-dose group (OR = 11.75; 95% CI = 3.0, 45.9). All models controlled for age, gender, weight, baseline creatinine, and creatinine clearance.
decreased. This patient completed the study doses, and no further reaction to subsequent doses was reported.
Hospital Length of Stay and Mortality
Adverse Events Nephrotoxicity occurred overall in 5.1% of study patients, with 3 (6.1%) in the traditional- and 2 (4%) in the loadingdose groups (P nonsignificant). In the traditional-dose group, all 3 patients met criteria for nephrotoxicity within 24 hours of the initial dose. In the loading-dose group, 1 patient met criteria on day 1 and the other on day 9. Among the traditional-dose group, only 1 patient experienced a clinically significant rise in serum creatinine from 0.7 to 3.2 mg/dL. This occurred in a bedridden quadriplegic patient being treated with vancomycin for possible skin and softtissue infection. The patient was also thought to have intraabdominal septic shock secondary to a perforated diverticulitis or ischemic bowel that may have contributed to the nephrotoxicity. The patient was discharged and did not require dialysis. The second patient’s creatinine returned to baseline prior to discharge. The third patient had a transient creatinine increase from 0.2 to 0.3 mg/dL. Although this meets criteria for nephrotoxicity with an increase of 50%, the clinical significance is minimal. Among the 2 patients in the loading dose group, there was a transient increase in creatinine in both patients that later resolved without clinical intervention. Because of the low incidence of nephrotoxicity in this study, additional adjustment by potential confounders was not conducted. Evaluation of the creatinine change among all the patients during hospitalization showed that there was no difference between a loading and traditional dose (−11% vs +11%, P > 0.05). It should be noted that 6 patients, 3 in each group, were found to have vancomycin trough levels >25 mg/L. No patient had a level >40 mg/L. Two of these patients met the criteria for nephrotoxicity, yet all returned to baseline creatinine at study completion. Review of the medical record 30 days after discharge did not reveal any readmissions with evidence of nephrotoxicity in any study patient. In the traditional-dose group, 2 patients (4.1%) developed pruritus, flushing, and/or rash. One of the 2 patients developed pain at the IV site, and the other patient experienced pruritus that was thought to be attributed to hydromorphone, which resolved after antihistamine administration. Vancomycin was discontinued in these patients. In the loading-dose group, 3 patients (6.0%) developed pruritus, flushing, and/or rash. Vancomycin was immediately discontinued in 2 of these patients because of suspected allergic reaction by the treating nurse in consultation with the admitting team. The third patient experienced flushing, which was speculated to be red man syndrome, and the symptoms resolved when the infusion rate was
There was no statistically significant difference in hospital length of stay (traditional dose, 6.33 days [SD = 5.6]; loading dose, 5.84 days [SD = 4.8]) or mortality between the 2 groups. One study patient died during their hospital visit. The patient presented to the ED and was thought to be septic. The patient was randomized to the loading-dose group and received 2 doses of vancomycin. Vancomycin was discontinued when the blood culture results revealed a Gramnegative organism.
Discussion Optimization of vancomycin dosing is required to achieve appropriate therapeutic levels and minimize resistance; however, the literature on loading doses for vancomycin is scarce. The current recommendation for loading doses is based on expert opinion and warrants the need for wellcontrolled studies. The primary outcome of this study was to determine if therapeutic vancomycin levels were attained more often by administering the consensus-recommended initial loading dose of 30 mg/kg as opposed to the more traditional 15-mg/kg dose. In our study, a significantly higher percentage of therapeutic levels were attained at 12 hours post–initial dose in those who received a loading dose compared with those who received 15 mg/kg. A small study of 34 patients demonstrated that the majority of patients who received vancomycin doses of 500 mg every 6 hours were subtherapeutic at 48 hours.16 Although there were significant differences initially with loading doses, our study demonstrated similar results, with approximately 20% (10/51) of levels attaining concentrations of greater than 15 mg/L in both groups at 36 hours. This inability to maintain the percentage of levels above 15 mg/L in the loading-dose group suggests that the maintenance dosing strategy may not have been optimized for the renal function of the patients studied. The guidelines recommend the consideration of loading doses in seriously ill patients; however, these patients are not clearly defined. In our study, we designated patients with SSSIs as non–critically ill. Those with other diagnoses (pneumonia, bacteremia, meningitis, osteomyelitis, neutropenic fever, and empiric therapy) were defined as critically ill patients. There were no differences in the percentage of levels that were >15 mg/dL in the traditional-dose group or the loading-dose group for patients who were deemed critically ill or those designated non–critically ill. Although the total number of levels in each group was small, the severity of illness did not appear to play a role in the percentage of troughs reaching therapeutic levels.
Downloaded from aop.sagepub.com at UNIV OF KANSAS MEDICAL CENTER on October 31, 2014
7
Rosini et al The average vancomycin levels at 12 and 24 hours in the traditional-dose group were artificially elevated because of the serum assay used; levels
research-article2014
AOPXXX10.1177/1060028014556813Annals of PharmacotherapyRosini et al
Research Report
A Randomized Trial of Loading Vancomycin in the Emergency Department
Annals of Pharmacotherapy 1–8 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1060028014556813 aop.sagepub.com
Jamie M. Rosini, PharmD1, Julie Laughner, MD1, Brian J. Levine, MD1,2, Mia A. Papas, PhD3, John F. Reinhardt, MD1, and Neil B. Jasani, MD1,2
Abstract Background: Optimizing vancomycin dosing may help eradicate bacteria while avoiding resistance. The guidelines recommend loading doses; however, there are no data to demonstrate that this may result in a more rapid achievement of therapeutic troughs. Objective: To evaluate the percentage of troughs reaching therapeutic levels at 12, 24, and 36 hours following an initial vancomycin dose of 30 mg/kg compared with 15 mg/kg. Methods: This prospective, randomized study was performed in a community academic medical center. Patients who were to receive vancomycin in the emergency department were randomized to an initial traditional dose of 15 mg/kg or a 30-mg/kg loading dose followed by 15 mg/kg every 12 hours for 3 doses. Patients weighing >120 kg or with creatinine clearances 10 mg/L and > 15 mg/L in the traditional-dose group compared with the odds in the loading-dose group.13,14 Treatment group, time after initial dosing, and the interaction between treatment group and time were included in all models. A statistically significant group by time interaction is interpreted as a significant difference in the vancomycin levels over time between the 2 groups. β Coefficients and 95% CIs were computed, and all effect estimates were adjusted for age, gender, weight, baseline creatinine, and creatinine clearance. Model-based standard errors (Wald statistics) were used to assess statistical significance. Statistical analyses were conducted using Stata version 12.0.15
Results Patient Enrollment and Characteristics A total of 99 patients were enrolled in the study, with 49 patients in the traditional-dose group and 50 patients in the loading-dose group. Patient enrollment is described in Figure 1. There were no statistically significant differences
Downloaded from aop.sagepub.com at UNIV OF KANSAS MEDICAL CENTER on October 31, 2014
4
Annals of Pharmacotherapy
Table 1. Baseline Characteristics of Study Participants by Treatment Group.a Age (SD) Gender, male (%) Weight, kg (SD) CrCl, mL/min (SD) Average length of stay, days (SD) Admitting diagnoses, n (%) SSSI Pneumonia Bacteremia Meningitis Osteomyelitis Neutropenic fever Empiric therapy Sepsis
15-mg/kg Traditional Dose, n = 49
30-mg/kg Loading Dose n = 50
58.4 (15.6) 25 (51.0) 86.7 (20.3) 104.1 (55.8) 7.08 (7.6)
57.5 (14.5) 28 (56.0) 81.6 (19.0) 93.4 (28.1) 5.87 (4.9)
14 (28.6) 12 (24.5) 3 (6.1) 3 (6.1) 3 (6.1) 1 (2.0) 13 (26.5) 6 (12.2)
16 (32.0) 14 (28.0) 6 (12.0) 0 (0) 0 (0) 2 (4.0) 12 (24.0) 6 (12.0)
Abbreviations: CrCl, creatinine clearance; SSSI, skin soft structure infection. a P was not significant for each baseline characteristic.
in any baseline characteristics between the 2 groups (Table 1). The most common admitting diagnosis was skin and skin structure infection (SSSI) followed by pneumonia. In all, 12 patients met criteria for a sepsis alert, and they were distributed equally between both groups. A sepsis alert was initiated if the initial lactate was above 4.0 mmol/L or the patient had 2 or more systemic inflammatory response syndrome criteria with a suspected infectious source. Patients in the traditional-dose group received an average initial vancomycin dose of 1308 mg (SD = 307), in comparison to 2433 mg (SD = 546) in those receiving a loading dose. Trough levels when available and appropriately drawn were included as a data point despite the number of doses they received. In the traditional-dose group, 40, 33, and 24 trough levels were obtained at 12, 24, and 36 hours, respectively. In the loading-dose group, 35, 33, and 27 trough levels were obtained at 12, 24, and 36 hours, respectively. Two trough levels (same patient) in the traditional-dose group were above 30 mg/L, whereas no levels in the loading-dose group were above 30 mg/L.
Primary End Point Data The proportion of patients in the 2 groups reaching a minimum target level of both 10 and 15 mg/L was compared (Figure 2). The proportion of patients reaching these target levels in the traditional dose group steadily increased over a 36-hour period, whereas the proportion of patients reaching these levels in the loading-dose group decreased. When a 15-mg/L level was applied as the target, the 12-hour trough displayed a significantly greater proportion of patients
reaching the goal among those who received a 30-mg/kg dose when compared with a 15-mg/kg dose (34% vs 3%, P < 0.01). This trend continued at 24 hours but was not significantly different. When presumed steady state (36 hours or before the fourth dose) was reached, there was no difference in the percentage of patients reaching target levels between the 2 groups. When 10 mg/L was used as the target cutoff, statistically significant differences in the proportion of patients reaching this level comparing the loading-dose group with the traditional-dose group were found both at 12 hours (80% vs 15%, P < 0.01) and 24 hours (67% vs 39%, P < 0.05) after the initial dose. The mean trough values for both groups at times 12, 24, and 36 hours post–initial dose are depicted in Table 2. A significantly higher average trough level was found 12 — hours after the initial dose in the loading-dose group (X = 13.9 mg/L; SD = 4.8) compared with the traditional-dose — group (X = 7.7 mg/L; SD = 3.1; P < 0.001) and 24 hours — after the initial dose in the loading-dose group (X = 12.7 mg/L; SD = 5.7) compared with the traditional-dose group — (X = 10.2 mg/L; SD = 5.4; P value 10 mg/L
> 15 mg/L
Figure 2. Percentage of troughs >10 mg/L and >15 mg/L at times 12, 24, and 36 hours after initial dose.
Table 2. Average Trough Levels Over Time by Treatment Group.a 15 mg/kg Traditional Dose Time since initial dose 12 Hours, mean (SD)** 24 Hours, mean (SD)* 36 Hours, mean (SD) Mean difference across time 24 and 12 Hours, mean (95% CI) 36 and 24 Hours, mean (95% CI) 36 and 12 Hours, mean (95% CI)
30 mg/kg Loading Dose
7.7 (3.1) n = 40 10.2 (5.4) n = 33 13.6 (7.8) n = 24 2.7 (1.7, 3.7)* 3.2 (1.9, 4.5)* 6.2 (4.0, 8.3)*
13.9 (4.8) n = 35 12.7 (5.7) n = 33 12.3 (5.1) n = 27 −1.0 (−2.4, 0.4) 0.3 (−1.8, 1.0) −0.4 (−1.6, 0.8)
a P values based on log-transformation of vancomycin trough levels. *P < 0.05; **P < 0.001.
Paired t tests were conducted to examine pairwise comparisons of mean trough levels over time within the 2 groups. For the traditional-dose group, there were statistically significant differences in the mean trough levels between 12 and 24 hours (mean difference = 2.7 mg/L; 95% CI = 1.7, 3.7), 24 and 36 hours (mean difference = 3.2 mg/L; 95% CI = 1.9, 4.5), and 12 and 36 hours (mean difference = 6.2 mg/L; 95% CI = 4.0, 8.3). There were no differences across time for the loading-dose group (Table 2). The percentage of therapeutic troughs between non–critically ill (SSSIs) and critically ill (all other diagnoses) patients were compared. In the traditional-dose group, the percentage of levels >15 mg/L in non–critically ill patients
at 12, 24, and 36 hours were 8.3%, 16.7%, and 33.3%, respectively. There was no significant difference when compared with the critically ill patients in the traditionaldose group. The percentage of levels >15 mg/L at 12, 24, and 36 hours were 0.0%, 9.5%, and 16.7%, respectively. Similarly, there was no difference between the percentage of levels greater than 15 mg/L between the non–critically ill patients and the critically ill patients in the loading-dose group at 12, 24, and 36 hours (36.4% vs 37.5%, 20.0% vs 21.7%, and 25.0% vs 16.7%). The results from the GEE analysis examining log vancomycin trough levels as the primary outcome indicated a statistically significant difference in the log trough levels over time by dosing group (β = −0.19; 95% CI = −0.28, −0.10; P < 0.01). For the loading-dose group, log trough levels remained constant over time (β = 0.03; 95% CI = −0.03, 0.10; P = 0.27), whereas for the traditional-dose group, they steadily increased (β = 0.25; 95% CI = 0.20, 0.29; P < 0.01). When trough levels were categorized into >10 mg/L and 15 mg/L, there were statistically significant group by time interactions (β = −1.85; 95%CI = −2.67, −1.03; P < 0.01 and β = −2.82; 95% CI = −4.54, −1.11; P < 0.01, respectively). On average, the odds of reaching minimal levels to prevent resistance (10 mg/L) in the loading-dose group were 7 times greater compared with that in the traditionaldose group (OR = 7.35; 95% CI = 2.89, 18.60). This increase was slightly greater when trough levels were categorized using a 15-mg/L minimum level. Results indicated an 11-fold increase in the odds of reaching therapeutic levels (15 mg/L) in the loading-dose group compared with the
Downloaded from aop.sagepub.com at UNIV OF KANSAS MEDICAL CENTER on October 31, 2014
6
Annals of Pharmacotherapy
traditional-dose group (OR = 11.75; 95% CI = 3.0, 45.9). All models controlled for age, gender, weight, baseline creatinine, and creatinine clearance.
decreased. This patient completed the study doses, and no further reaction to subsequent doses was reported.
Hospital Length of Stay and Mortality
Adverse Events Nephrotoxicity occurred overall in 5.1% of study patients, with 3 (6.1%) in the traditional- and 2 (4%) in the loadingdose groups (P nonsignificant). In the traditional-dose group, all 3 patients met criteria for nephrotoxicity within 24 hours of the initial dose. In the loading-dose group, 1 patient met criteria on day 1 and the other on day 9. Among the traditional-dose group, only 1 patient experienced a clinically significant rise in serum creatinine from 0.7 to 3.2 mg/dL. This occurred in a bedridden quadriplegic patient being treated with vancomycin for possible skin and softtissue infection. The patient was also thought to have intraabdominal septic shock secondary to a perforated diverticulitis or ischemic bowel that may have contributed to the nephrotoxicity. The patient was discharged and did not require dialysis. The second patient’s creatinine returned to baseline prior to discharge. The third patient had a transient creatinine increase from 0.2 to 0.3 mg/dL. Although this meets criteria for nephrotoxicity with an increase of 50%, the clinical significance is minimal. Among the 2 patients in the loading dose group, there was a transient increase in creatinine in both patients that later resolved without clinical intervention. Because of the low incidence of nephrotoxicity in this study, additional adjustment by potential confounders was not conducted. Evaluation of the creatinine change among all the patients during hospitalization showed that there was no difference between a loading and traditional dose (−11% vs +11%, P > 0.05). It should be noted that 6 patients, 3 in each group, were found to have vancomycin trough levels >25 mg/L. No patient had a level >40 mg/L. Two of these patients met the criteria for nephrotoxicity, yet all returned to baseline creatinine at study completion. Review of the medical record 30 days after discharge did not reveal any readmissions with evidence of nephrotoxicity in any study patient. In the traditional-dose group, 2 patients (4.1%) developed pruritus, flushing, and/or rash. One of the 2 patients developed pain at the IV site, and the other patient experienced pruritus that was thought to be attributed to hydromorphone, which resolved after antihistamine administration. Vancomycin was discontinued in these patients. In the loading-dose group, 3 patients (6.0%) developed pruritus, flushing, and/or rash. Vancomycin was immediately discontinued in 2 of these patients because of suspected allergic reaction by the treating nurse in consultation with the admitting team. The third patient experienced flushing, which was speculated to be red man syndrome, and the symptoms resolved when the infusion rate was
There was no statistically significant difference in hospital length of stay (traditional dose, 6.33 days [SD = 5.6]; loading dose, 5.84 days [SD = 4.8]) or mortality between the 2 groups. One study patient died during their hospital visit. The patient presented to the ED and was thought to be septic. The patient was randomized to the loading-dose group and received 2 doses of vancomycin. Vancomycin was discontinued when the blood culture results revealed a Gramnegative organism.
Discussion Optimization of vancomycin dosing is required to achieve appropriate therapeutic levels and minimize resistance; however, the literature on loading doses for vancomycin is scarce. The current recommendation for loading doses is based on expert opinion and warrants the need for wellcontrolled studies. The primary outcome of this study was to determine if therapeutic vancomycin levels were attained more often by administering the consensus-recommended initial loading dose of 30 mg/kg as opposed to the more traditional 15-mg/kg dose. In our study, a significantly higher percentage of therapeutic levels were attained at 12 hours post–initial dose in those who received a loading dose compared with those who received 15 mg/kg. A small study of 34 patients demonstrated that the majority of patients who received vancomycin doses of 500 mg every 6 hours were subtherapeutic at 48 hours.16 Although there were significant differences initially with loading doses, our study demonstrated similar results, with approximately 20% (10/51) of levels attaining concentrations of greater than 15 mg/L in both groups at 36 hours. This inability to maintain the percentage of levels above 15 mg/L in the loading-dose group suggests that the maintenance dosing strategy may not have been optimized for the renal function of the patients studied. The guidelines recommend the consideration of loading doses in seriously ill patients; however, these patients are not clearly defined. In our study, we designated patients with SSSIs as non–critically ill. Those with other diagnoses (pneumonia, bacteremia, meningitis, osteomyelitis, neutropenic fever, and empiric therapy) were defined as critically ill patients. There were no differences in the percentage of levels that were >15 mg/dL in the traditional-dose group or the loading-dose group for patients who were deemed critically ill or those designated non–critically ill. Although the total number of levels in each group was small, the severity of illness did not appear to play a role in the percentage of troughs reaching therapeutic levels.
Downloaded from aop.sagepub.com at UNIV OF KANSAS MEDICAL CENTER on October 31, 2014
7
Rosini et al The average vancomycin levels at 12 and 24 hours in the traditional-dose group were artificially elevated because of the serum assay used; levels
