ABSTRACT
The use of very high doses of polymyxin B (PMB) against carbapenem-resistant Gram-negative bacilli has been addressed in in vitro experiments as a strategy to improve bacterial killing and suppress resistance emergence. However, the toxicities of very high doses in patients are unknown. We conducted a retrospective cohort study assessing patients receiving PMB at >3 mg/kg of body weight/day or a total dose of ≥250 mg/day. The main outcomes were severe infusion-related adverse events according to the Common Terminology Criteria for Adverse Events and the renal failure category of RIFLE criteria for acute kidney injury (AKI) during treatment. A total of 222 patients were included for analysis of infusion-related events. The mean PMB dose was 3.61 ± 0.97 mg/kg/day (median total dose/day = 268 mg). Severe infusion-related adverse events occurred in two patients, resulting in an incidence of 0.9% (95% confidence interval, 0.2 to 3.2%); one was classified as a life-threatening adverse event, and one was classified as a severe adverse event. Renal failure was analyzed in 115 patients, and 25 (21.7%) patients presented renal failure (54 [47.0%] developed any degree of AKI, categorized as risk [27.8%], injury [25.9%], and failure [46.3%]). Treatment with a vasoactive drug, concomitant treatment with nephrotoxic drugs, and baseline creatinine clearance were independent risk factors for renal failure. Neither the PMB daily dose scaled by body weight nor the total daily dose was associated with renal failure. The in-hospital mortality rate was 60% (134 patients): 26% of deaths (57 patients) occurred during treatment, and none occurred during infusion. Our data suggest that high-dose schemes have an acceptable safety profile and could be further tested in clinical trials assessing strategies to improve patient outcomes and minimize the emergence of PMB resistance.
KEYWORDS: adverse event, nephrotoxicity, neurotoxicity, polymyxin B, polymyxins
INTRODUCTION
The old polymyxin antibiotics colistin and polymyxin B (PMB) have increasingly been used against extensively drug-resistant Gram-negative bacteria, notably, against carbapenem-resistant Enterobacteriaceae, Acinetobacter baumannii, and Pseudomonas aeruginosa (1, 2). Although colistin, administered as the inactive prodrug colistimethate, has traditionally been the most commonly used polymyxin worldwide, there has been an increasing interest in the use of PMB in recent years owing to some pharmacokinetic advantages over colistin, such as administration as the active compound and lower interindividual variability (3, 4).
The free (protein-unbound) 24-h area under the concentration-time curve (fAUC24) has been shown to drive the antibacterial activity of polymyxins (5), and although the target fAUC24 associated with clinical outcomes has not been evaluated in humans, experimental thigh infection and hollow-fiber infection models have shown that an fAUC24/MIC varying between 9 and 14 is associated with a 1- to 2-log decline in the number of CFU in 24 h (6). However, this target may not be suitable for many other types of infections, such as intra-abdominal infections and, in particular, pneumonia. Additionally, this target is unlikely attained with the usually recommended doses of PMB, especially when the MIC of polymyxin is at the current susceptibility breakpoint (2 mg/liter) (7).
Some clinical studies with PMB have shown that higher doses, most within the upper limit of the recommended range, have been associated with lower rates of mortality (8–10). Despite this, unfavorable clinical outcome rates are still considered unacceptably high (8–10). In addition, resistance to both polymyxins has increasingly been reported, and in vitro studies have shown that the emergence of resistance is relatively common even by isolates exposed to polymyxin concentrations similar to those expected in vivo after the administration of an adequate dose of polymyxins (11–13). To counteract this occurrence, in vitro exposures to very high concentrations of PMB through the use of distinct strategies, such as administration of a loading dose or the use of high doses in combination schemes, have been evaluated, with promising results (14–17). However, the toxicity of such regimens, in particular, the toxicity related to the infusion of high doses, is largely unknown. In this study, we aimed to evaluate severe infusion-related adverse events and renal failure after the intravenous administration of very high doses of PMB.
RESULTS
A total of 244 patients were eligible for the study. Twenty-two were excluded because the prescribed PMB was not administered to the patients either because the antibiotic was changed or because the patient died before administration, resulting in 222 patients receiving a high dose of PMB of ≥250 mg/day (n = 53, 23.8%), >3 mg/kg of body weight/day (n = 52, 23.4%), or both (n = 117, 52.7%). Only one patient received a second high-dose PMB treatment (not analyzed). The characteristics of the patients are described in Table 1.
TABLE 1.
Characteristics of patients in entire cohort and according to ICU statusa
| Characteristic | Value(s) for: |
||
|---|---|---|---|
| Entire cohort (n = 222) | ICU patients (n = 148) | Non-ICU patients (n = 74) | |
| No. (%) of female patients | 96 (43.2) | 68 (45.9) | 28 (37.8) |
| Mean ± SD age (yr) | 57 ± 17 | 58 ± 17 | 55 ± 18 |
| Wt (kg) | |||
| Median (IQR) | 75 (62–90) | 80 (64–95) | 69 (59–87) |
| Range | 32–190 | 32–190 | 33–151 |
| No. (%) of patients in hospital: | |||
| A | 58 (26.1) | 30 (20.3) | 28 (37.8) |
| B | 164 (73.9) | 118 (79.7) | 46 (62.2) |
| No. (%) of patients with the following comorbidities: | |||
| Diabetes | 55 (24.8) | 39 (26.3) | 16 (21.6) |
| Congestive heart failure | 49 (22.1) | 34 (23.0) | 15 (20.3) |
| Chronic kidney disease | 43 (19.4) | 29 (19.6) | 14 (18.9) |
| Solid neoplasia | 38 (17.1) | 26 (17.6) | 12 (16.2) |
| Chronic obstructive pulmonary disease | 32 (14.4) | 23 (15.5) | 9 (12.2) |
| Stroke | 29 (13) | 17 (11.5) | 12 (16.2) |
| Leukemia/lymphoma | 22 (9.9) | 11 (7.4) | 11 (14.7) |
| AIDS/HIV infection | 22 (9.9) | 12 (8.1) | 10 (13.5) |
| Corticosteroid use | 18 (8.1) | 7 (4.7) | 11 (14.9) |
| Transplant | 9 (4.1) | 4 (2.7) | 5 (6.8) |
| Cirrhosis | 4 (1.8) | 4 (2.7) | 0 |
| Median (IQR) Charlson comorbidity index | 4 (2–6) | 4 (2–6) | 4 (2–6) |
| PMB dose (mg/kg/day) | |||
| Mean ± SD | 3.61 ± 0.97 | 3.41 ± 0.75 | 4.02 ± 1.2 |
| Range | 1.7–8.7 | 1.7–7.0 | 2.4–8.7 |
| No. (%) of patients receiving a loading dose | 23 (10.4) | 14 (9.5) | 9 (12.2) |
| No. (%) of patients in whom the dose was changed during treatment | 29 (13.1) | 15 (10.1) | 14 (18.9) |
| Total daily dose (mg) | |||
| Median (IQR) | 268 (240–300) | 250 (244–300) | 276 (219–300) |
| Range | 114–475 | 164–475 | 114–450 |
| Duration of treatment (days) | |||
| Median (IQR) | 7 (4–12) | 7 (4–12) | 7 (3–11) |
| Range | 1–42 | 1–42 | 1–42 |
| No. (%) of patients with infection at the following site: | |||
| Respiratory tract | 123 (55.4) | 90 (60.8) | 33 (44.6) |
| Primary bloodstream | 24 (10.8) | 14 (9.5) | 10 (13.5) |
| Sepsis without a defined primary site | 24 (10.8) | 18 (12.2) | 6 (8.11) |
| Urinary tract | 14 (6.3) | 4 (2.7) | 10 (13.5) |
| Intra-abdominal | 13 (5.8) | 9 (6.1) | 4 (5.4) |
| Others (neutropenia, meningitis, endocarditis) | 9 (4.1) | 5 (3.3) | 4 (5.4) |
| Skin and soft tissue | 5 (2.3) | 2 (1.35) | 3 (4) |
| No. (%) of patients with microbiologically confirmed infection | 136 (60) | 87 (58.8) | 49 (66.2) |
| No. (%) of patients with the following etiology: | |||
| Acinetobacter baumannii | 67 (30.1) | 44 (29.8) | 23 (31) |
| Enterobacteriaceae | 50 (22.5) | 32 (21.6) | 18 (24.3) |
| Pseudomonas aeruginosa | 15 (6.7) | 7 (4.7) | 8 (10.8) |
| Others | 4 (1.8) | 4 (2.7) | 0 |
| Unknown | 86 (38.7) | 61 (41.2) | 25 (33.8) |
| No. (%) of patients with bacteremia | 46 (20.7) | 28 (18.9) | 18 (24.3) |
| No. (%) of patients receiving a concomitant antibiotic of study interest | 210 (94.6) | 4 (2.7) | 8 (10.8) |
| Aminoglycoside | 41 (18.5) | 24 (16.2) | 17 (23) |
| Carbapenem | 156 (70.3) | 113 (76.3) | 43 (58.1) |
| Cefepime | 7 (3.1) | 3 (2) | 4 (5.4) |
| Amphotericin B | 11 (5.0) | 7 (4.7) | 4 (5.4) |
| Isoniazid | 2 (0.9) | 1 (0.7) | 1 (1.3) |
| No. (%) of patients receiving other drugs potentially of interest | |||
| Sedatives | 145 (65.3) | 119 (80.4) | 26 (35.1) |
| Vasoactive drugs | 137 (61.7) | 114 (77.1) | 23 (31.1) |
| Neuromuscular blockers | 26 (11.7) | 23 (15.5) | 3 (4.1) |
| Chemotherapy | 1 (0.4) | 1 (0.7) | 0 |
| Antiseizure medication | 19 (8.5) | 7 (4.7) | 12 (16.2) |
| Median (IQR) creatinine concn (mg/dl) at day 1 of PMB treatment | 0.91 (0.54–1.39) | 1 (0.61–1.56) | 0.75 (0.42–1.36) |
| Median (IQR) estimated creatinine clearance (ml/min) | 91 (53–142) | 85 (50–139) | 96 (62–147) |
| No. (%) of patients receiving hemodialysis before PMB treatment | 95 (42.8) | 87 (58.8) | 8 (10.8) |
| No. (%) of patients with in-hospital mortality | 134 (60.3) | 106 (71.6) | 28 (37.8) |
| 30-day mortality | 112 (50.4) | 87 (58.8) | 25 (33.8) |
| Mortality during treatment | 57 (25.6) | 45 (30.4) | 12 (16.2) |
| Mortality during infusion | 0 (0) | ||
ICU, intensive care unit; PMB, polymyxin B; IQR, interquartile range.
Severe infusion-related adverse events.
Two of 222 patients presented a severe infusion-related adverse event during PMB infusion, resulting in a crude incidence of 0.9% (95% confidence interval [CI], 0.2 to 3.2%); one was classified as life-threatening, and one was classified as severe (crude incidence of each adverse event, 0.45%; 95% CI, 0.08 to 2.5%). The life-threatening adverse event occurred in an intensive care unit (ICU) patient (crude incidence among ICU patients, 0.67%; 95% CI, 0.12 to 3.7%), a 40-year-old male with cystic fibrosis who used 3.3 mg/kg/day of PMB and developed sudden thoracic pain, dyspnea, and hypoxemia on the 4th day of treatment. The severe adverse event occurred in a non-ICU patient (crude incidence among non-ICU patients, 1.3%; 95% CI, 0.2 to 7.2%), a 23-year-old male with lymphoma exposed to 3.6 mg/kg/day of PMB who presented perioral paresthesia, dizziness, and dyspnea on the 1st day of treatment. Four patients presented mild to moderate neurotoxicity events during infusion (overall crude incidence, 1.8%; 95% CI, 0.7 to 4.5%; crude incidence among non-ICU patients, 5.4%; 95% CI, 2.1 to 13.1%). Considering only patients not receiving sedatives, three patients presented mild to moderate neurotoxicity (crude incidence, 3.9%; 95% CI, 1.3 to 10.8%). One patient presented a moderate adverse event during treatment but not during infusion. The characteristics of patients with severe infusion-related and other neurotoxic adverse events are described in Table 2.
TABLE 2.
Description of severe infusion-related adverse events and other neurotoxic eventsa
| Patient | Age (yr) | Gender | Wt (kg) | Charlson comorbidity index score | Comorbidity(ies) | Site(s) of infection | Receipt of loading dose | PMB dose (mg/kg/day) | Total daily dose (mg) | Adverse event(s) | CTCAE grade(s), symptoms | Day of occurrence of adverse event | Total treatment duration | Other drug(s) received | In-hospital mortality |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 38 | F | 71 | 2 | Acute leukemia | Primary bloodstream | Yes (2.4 mg/kg) | 4.2 | 294 | Perioral paresthesia, oculomotor nerve disorder and peripheral sensory and motor neuropathy | 1, mild symptoms; 2, moderate symptoms limiting instrumental ADL | 2 (CTCAE grade 1) and 4 (CTCAE grade 2) | 7 | None | Yes (day 59 after first dose of PMB) |
| 2 | 71 | F | 151 | 2 | Congestive heart failure, COPD, CKD stage 2 | Intra-abdominal | No | 3.0 | 450 | Perioral paresthesia | 1, mild symptoms | 3 | 9 | Meropenem | No |
| 3 | 37 | M | 70 | 0 | None | Pneumonia | No | 5.7 | 400 | Paresthesia and lethargy | 1, mild symptoms, reduced alertness and awareness | 1 | 5 | Piperacillin-tazobactam | No |
| 4 | 23 | M | 74 | 2 | Lymphoma | Sepsis without a defined primary site | No | 3.6 | 270 | Perioral paresthesia and dizziness, dyspnea | 1, mild symptoms; 3, severe shortness of breath at rest limiting self-care ADL | 1 | 2 | Meropenem | No |
| 5 (patient in ICU) | 40 | M | 60 | 1 | Cystic fibrosis | Pneumonia | No | 3.3 | 200 | Noncardiac chest pain, dyspnea and hypoxemia | 2, moderate pain limiting instrumental ADL; 3, decreased oxygen saturation at rest; 4, life-threatening | 4 | 14 | Amikacin | No |
| 6 | 34 | F | 57 | 0 | None | Intra-abdominal and primary bloodstream | No | 5.6 | 300 | Dizziness and confusion | 1, mild disorientation | 1 | 9 | Ceftazidime | No |
| 7 (not infusion related) | 28 | M | 66 | 0 | Cystic fibrosis | Pneumonia | Yes (2.27 mg/kg) | 3.4 | 246 | Peripheral sensory and motor neuropathy and paresthesia | 2, moderate symptoms limiting instrumental ADL | 1 | 14 | Cefepime and vancomycin | No |
CTCAE, Common Terminology Criteria for Adverse Events; PMB, polymyxin B; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; ADL, activities of daily living, ICU, intensive care unit; F, female; M, male.
In comparison with patients who did not present any neurotoxicity, non-ICU patients who presented any neurotoxicity were older (mean ages, 38 ± 17 and 57 ± 17 years, respectively; P = 0.044), the two groups had similar mean weights (81 ± 35 and 72 ± 21 kg, respectively; P = 0.55) and received similar doses (mean dose scaled by weight, 4.2 ± 1.5 and 4.0 ± 1.2 mg/kg, respectively [P = 0.67]; median total dose, 297 mg/day [interquartile range {IQR}, 264 to 412 mg/day] and 272 mg/day [IQR, 203 to 300 mg/day], respectively [P = 0.14]), patients with neurotoxicity had a lower Charlson comorbidity index score (median, 1 [IQR, 0 to 2] and 4 [IQR, 2 to 6], respectively; P = 0.003) and a higher median baseline creatinine clearance (232 ml/min [IQR,160 to 288 ml/min] and 96 ml/min [IQR, 60 to 141 ml/min], respectively; P = 0.01), and there was no difference in the proportion of patients in the two groups receiving loading doses (2 [33.3%] and 7 [10.3%], respectively; P = 0.15). There was no significant difference between these patients regarding the other variables presented in Table 1, including the concomitant use of other potentially neurotoxic drugs.
Nephrotoxicity.
Renal failure was evaluated in 115 patients who received PMB for ≥48 h (107 patients were excluded from this analysis because 95 patients were under dialysis when PMB was initiated and 12 patients received PMB for <48 h). During PMB treatment, a total of 54 (46.9%) patients developed any degree of acute kidney injury (AKI) according to the RIFLE criteria, categorized as follows: risk, 15 (27.8%) patients; injury, 14 (25.9%) patients; and failure, 25 (46.3%) patients. The incidence of renal failure was 21.7% (25 of the 115 patients in the evaluable cohort). The mean time from PMB initiation to renal failure was 4 days (IQR, 3.5 to 6 days). Ten patients needed hemodialysis during PMB therapy. The results of the bivariate analysis of risk factors for renal failure are shown in Table 3.
TABLE 3.
Risk factors for renal failurea
| Characteristic | Value(s) for: |
P | ||
|---|---|---|---|---|
| Entire cohort (n = 115) | Patients with renal failure |
|||
| Yes (n = 25) | No (n = 90) | |||
| No. (%) of female patients | 49 (42.6) | 14 (56) | 35 (38.9) | 0.12 |
| Mean ± SD age (yr) | 57 ± 17 | 55 ± 16 | 57 ± 18 | 0.51 |
| Median (IQR) wt (kg) | 71 (60–90) | 73 (65–90) | 70 (60–90) | 0.66 |
| Median (IQR) PMB dose (mg/kg/day) | 3.5 (3.1–4.2) | 3.61 (3.2–4) | 3.5 (3.1–4.3) | 0.74 |
| Median (IQR) total daily dose (mg) | 271 (246–300) | 283 (200–300) | 270 (249–300) | 0.68 |
| No. (%) of patients in ICU | 54 (46.9) | 14 (56) | 40 (44.4) | 0.30 |
| Median (IQR) duration of treatment (days) | 8 (5–12) | 10 (6–14.5) | 7 (4–11) | 0.06 |
| No. (%) of patients with bacteremia | 22 (19.1) | 3 (12) | 19 (21.1) | 0.39 |
| Median (IQR) Charlson comorbidity index | 4 (2–6) | 4 (2–6) | 4 (2–6) | 0.98 |
| No. (%) of patients receiving: | ||||
| A vasoactive drug | 50 (43.5) | 16 (64) | 34 (37.8) | 0.02 |
| PMB alone | 10 (8.6) | 0 | 10 (11.1) | 0.11 |
| Other nephrotoxic drugsb | 61 (53) | 18 (72) | 43 (47.8) | 0.04 |
| Median (IQR) creatinine concn on day 1 (mg/dl) | 0.87 (0.56–1.35) | 0.74 (0.51–0.97) | 0.91 (0.59–1.55) | 0.04 |
| Median (IQR) creatinine clearance on day 1 (ml/min) | 88 (50–143) | 129 (71–175) | 80 (49–136) | 0.02 |
ICU, intensive care unit; PMB, polymyxin B; IQR, interquartile range.
Vancomycin (n = 51 patients), aminoglycosides (n = 21 patients), and amphotericin B (n = 5 patients).
In the multivariate analysis, receipt of a vasoactive drug and other nephrotoxic drugs and baseline creatinine clearance were independent risk factors for renal failure (Table 4). Addition of either the PMB daily dose scaled by body weight or the median total daily dose with and without total body weight to the model did not significantly change the final results.
TABLE 4.
Cox proportional hazards regression models for renal failurea
| Variable | Final model |
Alternative model 1 |
Alternative model 2 |
Alternative model 3 |
||||
|---|---|---|---|---|---|---|---|---|
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | |
| Treatment with another nephrotoxic drugb | 2.50 (1.02–6.14) | 0.046 | 2.50 (1.01–6.22) | 0.048 | 2.31 (0.95–5.65) | 0.065 | 2.47 (0.99–6.19) | 0.053 |
| Treatment with a vasoactive drug | 2.75 (1.17–6.43) | 0.020 | 2.75 (1.17–6.44) | 0.020 | 2.82 (1.20–6.63) | 0.018 | 1.78 (1.18–6.57) | 0.019 |
| Baseline creatinine clearance (ml/min) | 1.01 (1.00–1.01) | 0.035 | 1.01 (1.00–1.01) | 0.035 | 1.01 (1.00–1.01) | 0.030 | 1.01 (1.00–1.01) | 0.034 |
| PMB dose (mg/kg/day) | 0. 78 (0.49–1.27) | 0.324 | 0.78 (0.43–1.41) | 0.422 | — | — | — | — |
| Total daily dose (mg) | — | — | — | — | 1.00 (0.99–1.01) | 0.986 | 0.99 (0.99–1.01) | 0.792 |
| Wt (kg) | — | — | 1.00 (0.98–1.02) | 0.999 | — | — | 1.01 (0.98–1.03) | 0.485 |
PMB, polymyxin B; HR, hazard ratio; CI, confidence interval; —, not included or not performed.
Vancomycin, aminoglycosides, and amphotericin B.
Post hoc analysis.
Patients with a creatinine clearance of ≥90 ml/min presented an increased risk of renal failure compared with that for patients with lower creatinine clearance values (30.4% [17/56 patients] and 13.6% [8/59 patients], respectively; relative risk, 1.56; confidence interval, 1.09 to 2.24; P = 0.04).
Mortality.
The in-hospital mortality rate, the 30-day mortality rate, and the mortality rate during treatment were 60.3% (134 of 222 patients), 50.4%, and 25.6%, respectively. In the 115 patients eligible for assessment of AKI, the overall in-hospital mortality rate was 53% (61 patients); in patients who developed AKI, it was 57.4% (31 of 54), while it was 49.2% (30 of 61) in patients who did not develop AKI (P = 0.45). In-hospital mortality rates among the 61 patients who presented any degree of AKI according to the RIFLE categories were as follows: risk, 8.2% (5 patients); injury, 11.5% (7 patients); and failure 31.1% (19 patients) (P = 0.04).
DISCUSSION
Schemes of high-dose PMB to overcome the emergence of resistance to this antibiotic during exposure have been evaluated in in vitro infection models, with promising results (14–17). Moreover, few observational studies have suggested that increased doses might be associated with improved clinical outcomes (8–10, 18). Nonetheless, toxicity, particularly that related to the infusion of such high doses, has not been addressed so far.
Our study assessed patients receiving high doses of PMB, defined as a total daily dose of ≥250 mg and/or a dose of >3.0 mg/kg daily, and found a low incidence of severe infusion-related adverse events in the entire cohort. In fact, there was only one life-threatening adverse event, which occurred in an ICU patient, resulting in an incidence of 0.67% among these patients, and one severe adverse event, which occurred in a non-ICU patient, resulting in an incidence among patients outside the ICU a bit higher than that among ICU patients (1.3%). Although severe adverse events may be underestimated in sedated patients (like most patients admitted to an ICU), we believe that the high-dose regimens showed an acceptable safety profile. Our findings support the possible eventual assessment of such a dosing strategy in prospective trials to evaluate its role in PMB resistance emergence and in clinical and other microbiological outcomes. It must be noted, however, that owing to the retrospective design of the study, we could not reliably assess the duration of infusion of each dose. Thus, we could not characterize whether the severe adverse events may occasionally have been associated with shorter-duration infusions. In fact, institutional policies recommend PMB infusions of 1 to 4 h, but the possibility of inadvertent administrations for a period shorter than 1 h could not be fully ruled out.
Severe infusion-related adverse events during PMB and colistin infusion have been reported anecdotally (19–23); however, the incidence of such adverse events was not addressed before this study. So, we were also unable to determine whether the incidence might be higher in patients receiving higher doses or whether it may not be a dose-dependent adverse event because of both the absence of other studies and the lack of a control group receiving usual doses in our study.
Other mild to moderate neurotoxic adverse events were found in five other patients (during infusion in four patients), including perioral paresthesia, dizziness, peripheral neuropathy, lethargy, and mental confusion. We consider that the incidence of such events may also be considered relatively low, but we believe that it is likely also underestimated owing to the retrospective design of the study. The occurrence of neurotoxicity during treatment with polymyxins has been reported in approximately 0 to 27% of patients (21, 24, 25), with higher rates being found in older studies (21, 25, 26). However, owing to the design of most of these studies, we postulate that these rates were likely underestimated as well. Notably, in contrast to what might be expected, the patients in our cohort who had neurotoxicity were relatively young, had low Charlson comorbidity index scores, and had better renal function than the patients without neurotoxicity. It might be argued that these patients probably present a better capacity to perceive the symptoms of neurotoxicity, and this might explain such a finding, but further investigation is needed. Furthermore, no patients with neurotoxicity had renal failure or were on hemodialysis.
An important finding of our study is that renal failure rates were not impacted either by the dose scaled by total body weight or by the total daily dose, regardless of the inclusion or not of total body weight in the multivariate model. In fact, overall AKI rates were similar to the average rate of 40% found in previous studies of PMB (range, 20% to 60%) (27–34). The proportion of patients with renal failure in our study might be considered higher (46.3% of patients with AKI) than that described in reports of previous studies, which presented an average proportion of patients with renal failure of approximately 22% (but the average proportion ranged widely from 4% to 50%) (27–34), and the onset of failure may have occurred earlier in our sample than patients in previous studies. The number of patients requiring renal replacement therapy in our cohort may also be considered high, potentially indicating a higher risk for more severe kidney damage in patients receiving these high doses, although a definitive conclusion requires a proper control group exposed to lower doses. A noteworthy finding is that the mean/median dose in previous studies was approximately 150 mg, while in our study the median daily dose was 268 mg. This observation is consistent with the findings of Rigatto et al., who found a lack of a significantly increased risk for renal failure in 410 patients treated with PMB with total daily doses above 150 mg regardless of total body weight (35). In addition, it is noteworthy that patients with fluctuating renal function before PMB initiation were not evaluated for exclusion, and this might have occasionally increased the AKI rates in our study.
Interestingly, patients with higher creatinine clearances (≥90 ml/min) presented an approximately 100% increased risk for developing renal failure. This finding is consistent with the increased tubular reabsorption that is expected in these patients, and this higher reabsorption likely results in more extensive kidney damage (36).
The in-hospital rate of mortality for the overall cohort was higher than but similar to that in previous studies, especially those from Brazil (10, 33, 35, 38–40). However, the rate of mortality during treatment was substantially lower than the in-hospital rate of mortality, a finding that might represent the severity of the patients' conditions, which would determine a high risk of death. Finally, among patients addressed for renal toxicity, a significantly higher proportion of those who developed renal failure presented in-hospital mortality. We believe that although the renal failure itself contributed to this higher rate of mortality, it must be noted that patients who developed renal failure were significantly more frequently receiving vasoactive drugs at the beginning of PMB treatment (64.0%) than those who did not develop renal failure (39.2%) (P = 0.025) (data not shown). So, reverse causality bias cannot be ruled out.
A strength of our study is that we used objective and standardized criteria for severe adverse events, which will allow comparisons with future studies. Additionally, the doses of PMB used in this study have not been reported to have been used before, particularly with no dose adjustment for renal function. The major limitation, i.e., the possible underestimation of some neurotoxic adverse events, particularly in patients ICU, was acknowledged earlier, although underreporting of such events in the medical records might also have occurred. However, we believe that severe infusion-related adverse reactions have been much less affected by both shortcomings, that is, the retrospective design and the potential for underreporting. We also could not evaluate the reasons for the prescription of these high-dosage regimens. We believe that our patients might represent a sample of more severely ill patients in which physicians tend to prescribe higher doses owing to a possible greater severity of infections. Nevertheless, some facts may at least partially explain the prescription of high doses of PMB in our cohort. The first is the absence of a dose-capping recommendation, which means that patients weighing just above 80 kg will receive 250 mg if a dose of 3.0 mg/kg/day was chosen by the physician. Second, some observational studies from our group have suggested lower mortality rates with total PMB daily doses of ≥200 mg/day (8, 10), and it is usual for many prescribers in our institutions to round up doses to better fit into the vials (50 mg each). For example, a dose of 3.0 mg/kg in a patient would total 225 mg for a 75-kg patient or 240 mg for a 80-kg patient. These total doses are likely to be rounded up to 250 mg. In contrast, patients with lower weights tend to receive more than 3.0 mg/kg/day to reach a total dose near 200 mg/day. Finally, a considerable number of isolates in Brazil (some data have been published previously [41]) have an MIC at the polymyxin B/colistin breakpoint of 2 mg/liter. The proposed PK/PD target is unlikely attained with the doses recommended for the treatment of infections caused by isolates with these MICs (5–7). These likely pushed the doses up in many Brazilian hospitals.
In conclusion, in this retrospective cohort study, patients treated with very high doses of intravenous PMB presented a relatively low incidence of severe infusion-related adverse events and rates of AKI comparable to that found in previous studies with usual doses, even though the rates of renal failure among patients with AKI may be considered higher than those reported in the majority of previous studies. Our data suggest that high-dose schemes have an overall safety profile similar to that of usual doses and could be further tested in clinical trials evaluating strategies to improve patient outcomes and minimize the emergence of PMB resistance.
MATERIALS AND METHODS
Study design, settings, and participants.
This was a retrospective cohort study performed at two tertiary-care teaching hospitals in Porto Alegre, Brazil. The study was approved by the ethical committee of each hospital, under project numbers 15-0638 (Hospital de Clinicas de Porto Alegre) and 5236541580000527 (Hospital Nossa Senhora da Conceicao). From January 2013 to December 2015, all patients who received an intravenous dose of PMB of >3 mg/kg/day (or 30,000 IU/kg/day) or a total daily dose of ≥250 mg (or 2,500,000 IU/day), which were defined as high-dose regimens, were included in the study. For analysis of AKI, patients under hemodialysis at the beginning of PMB therapy and those treated for less than 48 h were excluded. Patients with a fluctuating renal function before polymyxin B initiation were not evaluated for exclusion. The doses administered were at the discretion of the attending physician. Dose recommendations in each hospital range from 1.5 to 3.0 mg/kg/day divided into two administrations every 12 h. No dose capping is recommended. Only the first treatment with high-dose PMB was included.
Variables and definitions.
The primary outcomes were severe infusion-related adverse events, which were classified as severe or medically significant or life-threatening adverse events or death according to Common Terminology Criteria for Adverse Events (42), and renal failure during treatment with PMB according to the RIFLE criteria for AKI (43). Secondary outcomes were either mild or moderate neurotoxic adverse events, according to the same criteria mentioned above, that occurred during infusion; any neurotoxic event during PMB treatment but not during infusion; the mortality rate during PMB treatment; and the in-hospital and 30-day mortality rates.
The variables potentially associated with the primary outcomes are described to be the following: demographics, actual body weight, Charlson comorbidity index score (44), estimated creatinine clearance according to the Cockcroft-Gault equation, dialysis at the beginning of therapy (except for renal failure, in which case these patients were excluded), ICU admission at the time of PMB initiation, treatment with vasoactive drugs during PMB treatment, site of infection (assigned according to the diagnosis of the attending physician), the etiology of the infection (the specific organisms that grew from a culture of a sample from the site of infection and unknown when cultures yielded negative results or were not required by the attending physician), the presence of bacteremia, concomitant antibiotics, polymicrobial infections, length of hospital stay prior to PMB initiation, dialysis after the onset of PMB, and PMB doses. Dosages were evaluated as the average daily dose of PMB (the sum of the total daily dose each day divided by the number of days until the end of therapy or death) and the daily dose scaled by total body weight.
Statistical analysis.
All statistical analyses were carried out using SPSS for Windows (version 18.0). Bivariate analysis was performed separately for each of the variables. P values were calculated using the χ2 or Fisher's exact test for categorical variables and the Student t test or the Wilcoxon rank-sum test for continuous variables. Covariates with a P value of ≤0.2 were included in a Cox proportional hazards model in a forward stepwise regression. Variables were checked for confounding and colinearity. Variables with a P value of ≤0.10 were maintained in the model. The proportional hazards assumption was graphically checked by inspecting the log[−log(S)] plot, where S is survival. Tests for interactions were not performed. All tests were two-tailed, and a P value of ≤0.05 was considered significant.
ACKNOWLEDGMENTS
This work was supported by the Fundo de Incentivo à Pesquisa do Hospital de Clínicas de Porto Alegre and the Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (FAPERGS). A.P.Z. is a research fellow of the National Council for Scientific and Technological Development (CNPq), Ministry of Science and Technology, Brazil.
The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.
A.P.Z. has received honoraria for speaking engagements and consultancies from AstraZeneca, Cipla, MSD, Pfizer, and United Pharmaceuticals. D.R.F. has given lectures and received research/educational grants from MSD, Pfizer, Gilead Sciences, and United Medical.
REFERENCES
- 1.Zavascki AP, Goldani LZ, Li J, Nation RL. 2007. Polymyxin B for the treatment of multidrug-resistant pathogens: a critical review. J Antimicrob Chemother 60:1206–1215. doi: 10.1093/jac/dkm357. [DOI] [PubMed] [Google Scholar]
- 2.Li J, Nation RL, Turnidge JD, Milne RW, Coulthard K, Rayner CR, Paterson DL. 2006. Colistin: the re-emerging antibiotic for multidrug-resistant Gram-negative bacterial infections. Lancet Infect Dis 6:589–601. doi: 10.1016/S1473-3099(06)70580-1. [DOI] [PubMed] [Google Scholar]
- 3.Nation RL, Li J, Cars O, Couet W, Dudley MN, Kaye KS, Mouton JW, Paterson DL, Tam VH, Theuretzbacher U, Tsuji BT, Turnidge JD. 2015. Framework for optimisation of the clinical use of colistin and polymyxin B: the Prato polymyxin consensus. Lancet Infect Dis 15:225–234. doi: 10.1016/S1473-3099(14)70850-3. [DOI] [PubMed] [Google Scholar]
- 4.Nation RL, Velkov T, Li J. 2014. Colistin and polymyxin B: peas in a pod, or chalk and cheese? Clin Infect Dis 59:88–94. doi: 10.1093/cid/ciu213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Tran TB, Velkov T, Nation RL, Forrest A, Tsuji BT, Bergen PJ, Li J. 2016. Pharmacokinetics/pharmacodynamics of colistin and polymyxin B: are we there yet? Int J Antimicrob Agents 48:592–597. doi: 10.1016/j.ijantimicag.2016.09.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Cheah SE, Wang J, Nguyen VT, Turnidge JD, Li J, Nation RL. 2015. New pharmacokinetic/pharmacodynamic studies of systemically administered colistin against Pseudomonas aeruginosa and Acinetobacter baumannii in mouse thigh and lung infection models: smaller response in lung infection. J Antimicrob Chemother 70:3291–3297. doi: 10.1093/jac/dkv267. [DOI] [PubMed] [Google Scholar]
- 7.Sandri AM, Landersdorfer CB, Jacob J, Boniatti MM, Dalarosa MG, Falci DR, Behle TF, Bordinhao RC, Wang J, Forrest A, Nation RL, Li J, Zavascki AP. 2013. Population pharmacokinetics of intravenous polymyxin B in critically ill patients: implications for selection of dosage regimens. Clin Infect Dis 57:524–531. doi: 10.1093/cid/cit334. [DOI] [PubMed] [Google Scholar]
- 8.Rigatto MH, Falci DR, Lopes NT, Zavascki AP. 2016. Clinical features and mortality of patients on renal replacement therapy receiving polymyxin B. Int J Antimicrob Agents 47:146–150. doi: 10.1016/j.ijantimicag.2015.11.007. [DOI] [PubMed] [Google Scholar]
- 9.Nelson BC, Eiras DP, Gomez-Simmonds A, Loo AS, Satlin MJ, Jenkins SG, Whittier S, Calfee DP, Furuya EY, Kubin CJ. 2015. Clinical outcomes associated with polymyxin B dose in patients with bloodstream infections due to carbapenem-resistant Gram-negative rods. Antimicrob Agents Chemother 59:7000–7006. doi: 10.1128/AAC.00844-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Elias LS, Konzen D, Krebs JM, Zavascki AP. 2010. The impact of polymyxin B dosage on in-hospital mortality of patients treated with this antibiotic. J Antimicrob Chemother 65:2231–2237. doi: 10.1093/jac/dkq285. [DOI] [PubMed] [Google Scholar]
- 11.Poirel L, Jayol A, Nordmann P. 2017. Polymyxins: antibacterial activity, susceptibility testing, and resistance mechanisms encoded by plasmids or chromosomes. Clin Microbiol Rev 30:557–596. doi: 10.1128/CMR.00064-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Giamarellou H. 2016. Epidemiology of infections caused by polymyxin-resistant pathogens. Int J Antimicrob Agents 48:614–621. doi: 10.1016/j.ijantimicag.2016.09.025. [DOI] [PubMed] [Google Scholar]
- 13.Kaye KS, Pogue JM, Tran TB, Nation RL, Li J. 2016. Agents of last resort: polymyxin resistance. Infect Dis Clin North Am 30:391–414. doi: 10.1016/j.idc.2016.02.005. [DOI] [PubMed] [Google Scholar]
- 14.Bulman ZP, Satlin MJ, Chen L, Kreiswirth BN, Shin BS, Walsh TJ, Holden PN, Forrest A, Nation RL, Li J, Tsuji BT. 2017. New polymyxin B dosing strategies to fortify old allies in the war against KPC-2-producing Klebsiella pneumoniae. Antimicrob Agents Chemother 61:e02023-16. doi: 10.1128/AAC.02023-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lenhard JR, Bulitta JB, Connell TD, King-Lyons N, Landersdorfer CB, Cheah SE, Thamlikitkul V, Shin BS, Rao G, Holden PN, Walsh TJ, Forrest A, Nation RL, Li J, Tsuji BT. 2017. High-intensity meropenem combinations with polymyxin B: new strategies to overcome carbapenem resistance in Acinetobacter baumannii. J Antimicrob Chemother 72:153–165. doi: 10.1093/jac/dkw355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Rao GG, Ly NS, Bulitta JB, Soon RL, San Roman MD, Holden PN, Landersdorfer CB, Nation RL, Li J, Forrest A, Tsuji BT. 2016. Polymyxin B in combination with doripenem against heteroresistant Acinetobacter baumannii: pharmacodynamics of new dosing strategies. J Antimicrob Chemother 71:3148–3156. doi: 10.1093/jac/dkw293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Tam VH, Schilling AN, Vo G, Kabbara S, Kwa AL, Wiederhold NP, Lewis RE. 2005. Pharmacodynamics of polymyxin B against Pseudomonas aeruginosa. Antimicrob Agents Chemother 49:3624–3630. doi: 10.1128/AAC.49.9.3624-3630.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Trifi A, Abdellatif S, Daly F, Mahjoub K, Nasri R, Oueslati M, Mannai R, Bouzidi M, Ben Lakhal S. 2016. Efficacy and toxicity of high-dose colistin in multidrug-resistant gram-negative bacilli infections: a comparative study of a matched series. Chemotherapy 61:190–196. doi: 10.1159/000442786. [DOI] [PubMed] [Google Scholar]
- 19.Berie MB, King MS, Thomsen IP. 2015. Polymyxin B infusion leading to cardiac arrest: a case report and literature review. Infection 43:121–124. doi: 10.1007/s15010-014-0704-3. [DOI] [PubMed] [Google Scholar]
- 20.Wunsch H, Moitra VK, Patel M, Dzierba AL. 2012. Polymyxin use associated with respiratory arrest. Chest 141:515–517. doi: 10.1378/chest.11-1483. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Koch-Weser J, Sidel VW, Federman EB, Kanarek P, Finer DC, Eaton AE. 1970. Adverse effects of sodium colistimethate. Manifestations and specific reaction rates during 317 courses of therapy. Ann Intern Med 72:857–868. [DOI] [PubMed] [Google Scholar]
- 22.Lindesmith LA, Baines RD Jr, Bigelow DB, Petty TL. 1968. Reversible respiratory paralysis associated with polymyxin therapy. Ann Intern Med 68:318–327. doi: 10.7326/0003-4819-68-2-318. [DOI] [PubMed] [Google Scholar]
- 23.Pohlmann G. 1966. Respiratory arrest associated with intravenous administration of polymyxin B sulfate. JAMA 196:181–183. [PubMed] [Google Scholar]
- 24.Kelesidis T, Falagas ME. 2015. The safety of polymyxin antibiotics. Expert Opin Drug Saf 14:1687–1701. doi: 10.1517/14740338.2015.1088520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Fekety FR Jr, Norman PS, Cluff LE. 1962. The treatment of gram-negative bacillary infections with colistin. The toxicity and efficacy of large doses in forty-eight patients. Ann Intern Med 57:214–229. [DOI] [PubMed] [Google Scholar]
- 26.Falagas ME, Kasiakou SK. 2006. Toxicity of polymyxins: a systematic review of the evidence from old and recent studies. Crit Care 10:R27. doi: 10.1186/cc3995. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Crass RL, Rutter WC, Burgess DR, Martin CA, Burgess DS. 2017. Nephrotoxicity in patients with or without cystic fibrosis treated with polymyxin B compared to colistin. Antimicrob Agents Chemother 61:e02329-16. doi: 10.1128/AAC.02329-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Osorio J, Barreto J, Samboni CF, Candelo LA, Alvarez LC, Benavidez S, Tellez RP, Santofimio D, Ramos JA, Gomez CA. 2017. Risk factors for acute kidney injury in patients treated with polymyxin B experience from 139 cases at a tertiary university hospital in Colombia. Rev Chilena Infectol 34:7–13. (In Spanish.) doi: 10.4067/S0716-10182017000100001. [DOI] [PubMed] [Google Scholar]
- 29.Soares DS, Reis ADF, Silva Junior GBD, Leite TT, Parente Filho SLA, Rocha CVO, Daher EF. 2017. Polymyxin-B and vancomycin-associated acute kidney injury in critically ill patients. Pathog Glob Health 111:137–142. doi: 10.1080/20477724.2017.1309338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Rigatto MH, Oliveira MS, Perdigao-Neto LV, Levin AS, Carrilho CM, Tanita MT, Tuon FF, Cardoso DE, Lopes NT, Falci DR, Zavascki AP. 2016. Multicenter prospective cohort study of renal failure in patients treated with colistin versus polymyxin B. Antimicrob Agents Chemother 60:2443–2449. doi: 10.1128/AAC.02634-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Dubrovskaya Y, Prasad N, Lee Y, Esaian D, Figueroa DA, Tam VH. 2015. Risk factors for nephrotoxicity onset associated with polymyxin B therapy. J Antimicrob Chemother 70:1903–1907. doi: 10.1093/jac/dkv014. [DOI] [PubMed] [Google Scholar]
- 32.Phe K, Lee Y, McDaneld PM, Prasad N, Yin T, Figueroa DA, Musick WL, Cottreau JM, Hu M, Tam VH. 2014. In vitro assessment and multicenter cohort study of comparative nephrotoxicity rates associated with colistimethate versus polymyxin B therapy. Antimicrob Agents Chemother 58:2740–2746. doi: 10.1128/AAC.02476-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Tuon FF, Rigatto MH, Lopes CK, Kamei LK, Rocha JL, Zavascki AP. 2014. Risk factors for acute kidney injury in patients treated with polymyxin B or colistin methanesulfonate sodium. Int J Antimicrob Agents 43:349–352. doi: 10.1016/j.ijantimicag.2013.12.002. [DOI] [PubMed] [Google Scholar]
- 34.Kubin CJ, Ellman TM, Phadke V, Haynes LJ, Calfee DP, Yin MT. 2012. Incidence and predictors of acute kidney injury associated with intravenous polymyxin B therapy. J Infect 65:80–87. doi: 10.1016/j.jinf.2012.01.015. [DOI] [PubMed] [Google Scholar]
- 35.Rigatto MH, Behle TF, Falci DR, Freitas T, Lopes NT, Nunes M, Costa LW, Zavascki AP. 2015. Risk factors for acute kidney injury (AKI) in patients treated with polymyxin B and influence of AKI on mortality: a multicentre prospective cohort study. J Antimicrob Chemother 70:1552–1557. doi: 10.1093/jac/dku561. [DOI] [PubMed] [Google Scholar]
- 36.Zavascki AP, Nation RL. 2017. Nephrotoxicity of polymyxins: is there any difference between colistimethate and polymyxin B? Antimicrob Agents Chemother 61:e02319-16. doi: 10.1128/AAC.02319-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Reference deleted.
- 38.de Oliveira MS, de Assis DB, Freire MP, Boas do Prado GV, Machado AS, Abdala E, Pierrotti LC, Mangini C, Campos L, Caiaffa Filho HH, Levin AS. 2015. Treatment of KPC-producing Enterobacteriaceae: suboptimal efficacy of polymyxins. Clin Microbiol Infect 21:179.e1–e7. doi: 10.1016/j.cmi.2014.07.010. [DOI] [PubMed] [Google Scholar]
- 39.Rigatto MH, Ribeiro VB, Konzen D, Zavascki AP. 2013. Comparison of polymyxin B with other antimicrobials in the treatment of ventilator-associated pneumonia and tracheobronchitis caused by Pseudomonas aeruginosa or Acinetobacter baumannii. Infection 41:321–328. doi: 10.1007/s15010-012-0349-z. [DOI] [PubMed] [Google Scholar]
- 40.Kvitko CH, Rigatto MH, Moro AL, Zavascki AP. 2011. Polymyxin B versus other antimicrobials for the treatment of Pseudomonas aeruginosa bacteraemia. J Antimicrob Chemother 66:175–179. doi: 10.1093/jac/dkq390. [DOI] [PubMed] [Google Scholar]
- 41.Bartolleti F, Seco BM, Capuzzo dos Santos C, Felipe CB, Lemo ME, Alves TDS, Passadore LF, Mimica MJ, Sampaio SC, Zavascki AP, Sampaio JL. 2016. Polymyxin B resistance in carbapenem-resistant Klebsiella pneumoniae, Sao Paulo, Brazil. Emerg Infect Dis 22:1849–1851. doi: 10.3201/eid2210.160695. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.National Cancer Institute. 2009. Common terminology criteria for adverse events, version 4.0. NIH publication 09-7473. National Cancer Institute, Bethesda, MD. [Google Scholar]
- 43.Lopes JA, Jorge S. 2013. The RIFLE and AKIN classifications for acute kidney injury: a critical and comprehensive review. Clin Kidney J 6:8–14. doi: 10.1093/ckj/sfs160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Charlson ME, Pompei P, Ales KL, MacKenzie CR. 1987. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 40:373–383. doi: 10.1016/0021-9681(87)90171-8. [DOI] [PubMed] [Google Scholar]