Piperacillin-Tazobactam: The Antibiotic You’re Not Administering Correctly
“Half of the time we’re gone but we don’t know where, and we don’t know where ...”
A 42 year old man is admitted with fever, right upper-quadrant pain and jaundice. Over the last few hours he has become progressively confused and hypotensive. He has normal renal function with adequate urine output, but the microbiology department alerts your house officer of multiple blood culture bottles containing non-fermenting gram negative rods. The patient is given 2 litres of lactated ringers and a vasopressin infusion is initiated. The patient is also immediately begun on piperacillin-tazobactam [pip-tazo] 4.5 grams every 6 hours.
There is surprisingly little data pertaining to the pharmacokinetics of pip-tazo in the critically-ill patient population. In one small investigation, trough levels of pip-tazo demonstrated significant variability in those with normal renal function. Further, in critically-ill patients with normal renal function and moderately-impaired renal function, the administration of pip-tazo 4.5 grams every 6 hours and every 8 hours, respectively, led to insufficient plasma levels.
Pip-tazo, like other beta-lactam antibiotics, exhibit time-dependent antibacterial activity; that is, efficacy is related to the absolute duration that the antibiotic remains above the minimum inhibitory concentration [MIC] of the pathological organism of interest.
In the critically-ill, there are multiple mechanisms by which pip-tazo [and beta-lactams, in general] can fall below the MIC when dosing is intermittent. These alterations include an increase in both the apparent volume of distribution, as well as, clearance of the antibiotic. Consequently, administering beta-lactams as a continuous infusion has been shown to increase time above the MIC and amplify bacterial annihilation. Nevertheless, in a meta-analysis which included patients with non-severe sepsis [i.e. patients likely not labeled as septic by the new definition], no significant difference in mortality was found between patients who received continuous infusion versus intermittent beta-lactam dosing.
Two investigations published in the last 6 months shed some light on pip-tazo dosing in the critically-ill. Firstly, Zander and colleagues evaluated piperacillin concentration using multiple plasma assessments in patients receiving pip-tazo 4.5 grams either twice or thrice daily based on creatinine clearance. Interestingly, they found that – in the first day – no patients who received pip-tazo thrice daily and who had a creatinine clearance [CrCl] more than 65 mL/min achieved adequate trough levels of piperacillin. That is, in all patients with a CrCl more than 65 mL/min, piperacillin levels were found to be sub-therapeutic. Further, 55% of those who received pip-tazo thrice daily and who had a CrCl of 30 – 65 mL/min had sub-therapeutic piperacillin trough levels. The authors also noted a negative relationship between piperacillin trough levels and C-reactive protein raising the concern that in those with mild to moderate renal insufficiency, pip-tazo was under-dosed using the commonly suggested dosing scheme. The percentage of insufficient piperacillin levels remained fairly stable over many days of therapy. In patients with severely-impaired renal function, however, the likelihood of sub-therapeutic piperacillin levels was low even when dosed – as suggested – as 4.5 grams twice daily. This was also true of patients receiving renal replacement therapy. Overall, there was no relationship between piperacillin levels and mortality, though the investigation included only 60 patients.
Last month, Roberts and colleagues published a fascinating meta-analysis of beta-lactam infusions in patients with severe sepsis and septic shock. That is, those with ‘sepsis’ and ‘septic shock’ based on the new definition. This combined analysis included three trials and a total of 632 patients – 312 of whom received a continuous beta-lactam infusion versus 320 who received standard intermittent dosing. Hospital mortality censored at day 30 revealed an absolute risk reduction in mortality by 6.7% [i.e. from 26.3% mortality with intermittent dosing compared to 19.6% mortality in those who received a continuous infusion]. Additionally, while there was a trend towards improved clinical cure and ICU-free days at day 28 in those who received a continuous infusion of beta-lactam, neither were statistically significant as depicted in figure 2 within the meta-analysis.
With regards to the subgroup of patients who received pip-tazo as the beta-lactam of choice, 30 day hospital mortality was 25.8% in those who received intermittent dosing compared to 16.3% who received continuous infusion, which is both statistically and clinically significant. With respect to all beta-lactams, mortality was not reduced by continuous infusion when patients received renal replacement therapy [RRT]; this echoes the Zander paper which found that piperacillin concentration was more likely to be therapeutic in patients with the most severe degree of renal impairment, including those who received RRT.
As Zander et al. found, there is a rather tight correlation between CrCl and piperacillin trough; the significance being that patients with greater glomerular filtration rate run the highest risk of having inadequate trough levels – at least in those dosed 4.5 grams every 8 hours. These pharmacokinetic implications reverberate within the paper by Roberts et al; that is, continuous infusion of beta-lactams – and pip-tazo in subgroup analysis – holds mortality benefit. Indeed, in an accompanying commentary, the call for prolonged pip-tazo infusion in the critically-ill was made and/or monitoring trough concentrations at the steady-state [i.e. after 3-4 doses].
But should we be swayed by the meta-analysis data described above? Certainly, an absolute risk reduction in mortality of 6.7% for beta-lactam infusion and 9.5% absolute risk reduction in mortality for pip-tazo infusion compared to intermittent dosing is quite profound. Do we remember when everyone jumped aboard the Xigris train 15 years ago – for an absolute risk reduction in mortality of 6.1% [30.8% to 24.7%]?
But if we compare the patient population of the Roberts meta-analysis [i.e. those with ‘sepsis’ by the new definition: life-threatening organ dysfunction caused by a dysregulated host response to infection] to accepted baseline mortality rates, it appears that the patients receiving intermittent dosing had slightly higher than expected mortality [i.e. ~ 26%]. The Roberts meta-analysis contained patients with 'severe sepsis' and 'septic shock' with an APACHE II score of 20-21. A very similar patient population was studied in the ProCESS trial – in which all three arms carried a lower morality rate of 18-21%. Notably, these mortality rates are more similar to the continuous infusion group in the Robert’s analysis [i.e. ~ 20% mortality] and, presumably, the patients in ProCESS did not receive continuous beta-lactam infusions.
Nevertheless, the results of both Zander et al. and Roberts et al. should provide an interesting inter-professional learning opportunity for those ICUs fortunate enough to be staffed by dedicated pharmacists and pharmacologists. Should your ICU offer continuous beta-lactam infusions to patients with sepsis 3.0? For some additional reading on pip-tazo renal toxicity, please refer to this terrific analysis by Josh Farkas.
Return to the Case
As it happens, the patient is a television celebrity with multiple friends and family in the medical profession. An outside pharmacologist calls your house-officer overnight on the family’s behalf to inform her that she is “not administering the pip-tazo correctly.” She acquiesces and initiates a 24 hour infusion of pip-tazo with the help of the on-call ICU pharmacologist. 72 hours later, the patient is quite well and off vasoactive medications. He receives an ERCP followed by a cholecystectomy and is transferred to the surgical floor.