This study had several limitations. The short time frame (6 months) did not allow analysis of temporal trends in milrinone use. The use of logistic regression, which focuses on the presence or absence of an event, prevented us from capturing the time course (i.e., onset, frequency, duration) of outcomes. The time course of independent variables was not analyzed, which may further limit the ability to establish a temporal rela­tionship between changes in milrinone dosing and outcomes, including the ability to evaluate initial response to milrinone. Data collection was confined to the period during which milrinone was being administered, and it is therefore unknown whether patients had pre-existing low cardiac output syndrome, renal failure, and/or thrombocytopenia (i.e., before use of milrinone). We did not follow up on outcomes after milrinone was stopped, potentially missing important milrinone- related outcomes that occurred after the drug was discontinued.

The results of this single-centre study may not reflect prac­tice in other institutions.

The observational, retrospective nature of the data meant that we were unable to determine causality, i.e., whether the milrinone caused the adverse events described or if the milri- none prevented more severe low cardiac output syndrome. It was difficult to control for confounders that might predispose patients to low cardiac output syndrome, arrhythmias, or thrombocytopenia. Also, the reasons for changes in dose rate (e.g., adverse effects or lack of efficacy) were unknown. In addition, it was not possible to obtain milrinone serum concentrations for these patients; only the dose could be ascer­tained. Hypotension is likely multifactorial and fluctuates with the clinical status of a critically ill patient, so it was difficult to associate this potential adverse effect of milrinone with the specific dosing parameters. A prospective study would allow temporal association between development of hypotension and loading doses or accumulation of milrinone. Also, although the results of this study suggest that sicker patients require more milrinone, this is only a hypothesis; a prospective study design would be helpful in exploring severity of illness (e.g., type of cardiac lesion or repair) and dosing as predictors of outcome. canadian cialis online

The use of a lactate difference of greater than 2 mmol/L and AVO2 difference of 30% or more as markers of low cardiac output syndrome had limitations. The incidence of these 2 measures was quite different: 76.1% of admissions with AVO2 difference of 30% or more and 15.3% of admissions with lactate difference greater than 2 mmol/L (Table 1). The PRIMACORP study similarly found that the incidence of AVO2 difference of 30% or more was greater than the incidence of metabolic acidosis (45.5% versus 22.7%). Thus, lactate difference may underdetect the incidence of low cardiac output syndrome. The AVO2 difference is thought to be the more sensitive marker in our study.

Using maximum creatinine levels as a marker for renal impairment has limitations, including the fact that elevations in creatinine lag behind renal insult. Creatinine clearance would have been preferable but was not measured for the patients studied here. Data on urine output and fluid balance were also collected, but these measures can be confounded by use of diuretics. Therefore, creatinine was the best available marker of renal function.

Finally, some patients were included more than once in our study (12 patients had 2 admissions, and 2 patients had 3 admissions). It is difficult to tell what impact these repeat patients had on the data, as they represented a minority (7.1%) of the study population. Also, the condition of patients in the CCU tends to be unstable, and the same patient may have presented very differently during separate admissions.

Implications for Practice

This study highlights the paucity of information to guide optimal milrinone dosing for children, particularly in more complex cases (e.g., single ventricle, septic shock) and for patients with renal impairment. The results provide a basis for further pharmacokinetic and pharmacodynamic studies. One of the co-investigators on the current study (C.P.) is now work­ing on a pilot prospective pharmacokinetic and pharmaco- dynamic study at another pediatric centre, based on issues identified here. In addition, results from this study have increased awareness among practitioners at The Hospital for Sick Children that duration of therapy for postoperative cardiac patients should be limited and that maximum doses should be de-escalated as soon as possible. In addition, along with other clinical practice improvement initiatives, milrinone is now being discontinued for selected postoperative cardiac patients (those who have undergone uncomplicated cardiac repair) after the initial loading dose given in the operating room. buy levitra 20 mg


This report of a retrospective cohort study has described the use of milrinone in critically ill pediatric patients in the CCU of The Hospital for Sick Children. Most of the patients had undergone cardiovascular surgery. The initial loading dose, usually given in the operating room, was greater than the recommended 50 pg/kg. Maintenance infusion was not adjusted for renal impairment. Dose tapering and infusion interruptions did not occur for the majority of patients.

There was an association between the milrinone dosing regimen and outcomes. In particular, there appeared to be a significant association between average dose rate and duration of milrinone therapy and low cardiac output syndrome. A prospective pharmacokinetic and pharmacodynamic study of this drug is warranted to define the relationship between the efficacy and toxicity of the drug and its plasma concentrations. The results of this study support the suggestion that efficacy should be explored as an end point. Up to 78% of patients experienced low cardiac output syndrome while receiving milrinone, which raises the question of whether current dosing is adequate. A greater understanding of the concentration— efficacy relationship for milrinone is necessary to improve therapeutic outcomes. Children with renal impairment require further study because of the risks associated with milrinone use and the paucity of prospective dose-finding literature in this group.