Implementation of a Pharmacist Monitoring Process for Patients on QTc Prolonging Antibiotics: A Pilot Study

Abstract

Purpose: The purpose of this pilot study was to implement a pharmacist monitoring process for 4 antimicrobials; azithromycin, ciprofloxacin, levofloxacin, and fluconazole. This pilot study was a patient safety initiative to screen patients and engage providers about therapies at risk for QT prolongation. Methodology: A concurrent chart review was performed at a single center from January 6, to February 22, 2020, of adult patients ≥ 18 years of age initiated on azithromycin, ciprofloxacin, levofloxacin, and fluconazole. Patient risk factors assessed: age, female sex, loop diuretic use, potassium ≤ 3.5 mEq/L, QTc ≥ 450 ms, acute myocardial infarction (MI) or heart failure, 1 or more QTc prolonging agents, and sepsis. The primary endpoint was successful implementation of the QTc monitoring process by pharmacists. Secondary endpoints were the interventions made by pharmacists. Results: From January 6, to February 22, 2020, there were a total of 412 orders for one of the target antimicrobials that resulted in 157 documented pharmacist reviews (38.1%). Of the 157 evaluations, 100 of these represented patients in our high risk group (84 moderate, 16 high risk). Successful implementation was observed through documentation of assessment on all patients with moderate or high risk scores in the 100 person cohort. Conclusion: The pilot study demonstrated a successful implementation of a QTc monitoring process by pharmacists since all patients had documented reviews. Further steps include investigating how to improve efficiency, as well as ways for continued success in monitoring.

Background

The American Heart Association (AHA) and the American College of Cardiology Foundation (ACCF) released a scientific statement in 2010 to raise awareness among healthcare professionals about the risk of QT interval drug interactions. The statement included the need for electrocardiogram (ECG) monitoring, and prevention of drug-induced QT interval prolongation and Torsade de Pointes (TdP) in hospitalized patients. The American College of Cardiology (ACC) recommended the upper limit for a normal QTc be 460 ms for women and 450 ms for men. ¹ Clinically, QT prolongation that can potentially lead to fatal dysrhythmias have an unadjusted interval value longer than 500 ms. ¹ Many investigators have explored characteristics that may be associated with a higher risk for patients developing prolonged QTc. However, the variability among different screening tools disrupts the ability for a unified scoring system. It is estimated that with each 10-ms increase in QT prolongation there is approximately 5% to 7% cumulative increase in risk for TdP.^1-3

Tisdale et al ⁴ created a risk score development group (n = 900) at a single tertiary care institution and a validation group (n = 300) to identify risk factors associated with QTc prolongation. Risk factors were assigned a point value for a total of 21 points. Low risk was scored 0-6, moderate risk 7 to 10, and high risk is ≥11 with criteria listed in Table 1 with QTc risk associated with medications in List 1 of CredibleMeds^®. ⁴ A high risk was associated with sensitivity of 0.74, specificity of 0.77, positive predictive value of 0.79 and negative predictive value of 0.76. The incidence of QTc prolongation was 15% for low risk, 37% moderate risk, and 73% for high risk.

Table 1.

Risk Factor Calculator.

Risk factor	Point value
Age ≥68 y old	1
Female sex	1
Loop diuretic	1
Serum potassium ≤3.5 mEq/L	2
Admission QTc ≥450 ms	2
Acute myocardial infarction	2
≥2 QTc prolonging drugs	3
Sepsis	3
Acute heart failure	3
One QTc prolonging drug	3
Maximum risk score	21

Buss et al investigated a similar scoring system called the RISQ-PATH score. The RISQ-PATH score incorporated risk factors for QTc prolongation with a maximum score of 40.5 points and high risk was determined to be ≥10 points (Table 2). ⁵

Table 2.

RISQ-PATH Score.

Risk factor	Score
Age ≥65 y old	3
Female gender	3
Smoking status	3
Body Mass Index ≥30 kg/m2	1
Cardiomyopathy	3
Hypertension	3
Arrhythmia	3
Existing prolonged QT interval	6
Thyroid disturbances	3
Liver failure	1
Neurological disorders (stroke, tumor, infection, trauma)	0.5
Diabetes	0.5
Hypokalemia (≤3.5 mmol/L)	6
Hypocalcaemia (<2.15 mmol/L)	3
Inflammation (CRP >5 mg/L)	1
Renal impairment (eGFR ≤30 mL/min/1.73 m2)	0.5
Known risk QT interval-prolonging drug	3 per drug
Possible risk QT interval-prolonging drug	0.5 per drug
Conditional risk QT interval-prolonging drug	0.25 per drug
Maximum score	40.5

Vandael et al ⁶ conducted a meta-analysis that reviewed 89 532 patients and assessed their evidence for different risk factors for QTc-prolongation. The investigators determined “very strong” evidence was found for hypokalemia, use of diuretics, antiarrhythmic drugs and QTc-prolonging drugs in List 1 of CredibleMeds^®. ⁷ CredibleMeds^® is an online database with information regarding the degree of QT prolongation found in mediations. ⁷ Little or no evidence was found for hyperlipidemia, the use of digoxin or statins, neurological disorders, diabetes, renal failure, depression, alcohol abuse, heart rate, pulmonary disorders, hormone replacement therapy, hypomagnesaemia, history of a prolonged QTc-interval/Torsade de Pointes, familial history of cardiovascular disease, and the use of only QTc-prolonging drugs of List 2 or 3 of CredibleMeds. ⁶

Additional research into risk factors for patients taking these medication is important, to help implement a screening system to identify patients at highest risk of TdP. A conservative measure may include enhanced monitoring, or a more aggressive method may suggest switching to an alternative therapy. Currently there is minimal published data that details an effective screening and monitoring method in patients with higher risk and more guidance would be beneficial. Especially since it has been suggested that regardless of risk category the medication falls into, the overall risk increases when administered to patients with certain co-morbidities.^4-6,8-10 To note, Daniel et al implemented a pharmacist driven ECG ordering protocol in a psychiatric hospital utilizing the aspects from Tisdale and colleagues. Their goal was to improve the appropriateness of ECGs and QTc-interval monitoring of at-risk psychiatric inpatients. From their study 1 center saw appropriate ECG utilization increased by 25.5% after implementation of a standardized protocol (P = .0172) and appropriate omission of ECG utilization improved by 26% (P < .00001). While their study aimed to focus on antipsychotic medications, this study will focus on how to implement the process while piloting 4 antimicrobials that are commonly used and to solidify a pharmacist driven monitoring program. These 4 antimicrobials were selected because they are on formulary, used frequently in pneumonia, urinary tract infections, and fungal infections without significant follow-up from providers. From this study, future studies at our institution will be used to assess the appropriateness of ECGs like in Daniel et al. ¹¹

The importance of pharmacist monitoring of this process can be seen through Tien and colleagues where pharmacist driven QTc monitoring was feasible and reduced the risk of QTc interval prolongation when compared to traditional physician team monitoring. QTc interval prolongation occurred less frequently in the pharmacist group compared with the standard care group (19% vs. 39%, respectively; P = .006). These results are why this study attempted to implement a monitoring process in order to display pharmacist interventions and monitoring abilities to other healthcare providers. ¹²

The purpose of the pilot study was to create a QTc monitoring process based on previous literature and determine if implementing a risk scoring system was feasible to identify patients at higher risk. Based on the ability to implement this scoring system the natural progression would be to implement the system into the electronic medical record so pharmacists could actively monitor and communicate with providers on risks of QTc prolonging agents.

Methods

Trial Design

This study was a single center, concurrent chart review at a 500-bed community hospital in Cleveland, Ohio, with a goal of piloting a patient safety initiative that involved pharmacist monitoring, and interventions through recommendations when necessary. Preliminary data derived from electronic medical records (EMR) reviews suggested approximately 50 patients per week were initiated on azithromycin, ciprofloxacin, levofloxacin, or fluconazole with 34 percent in the high risk group.

Patients were screened through an EMR report if they were at least 18 years of age and received any dose of azithromycin, ciprofloxacin, levofloxacin, or fluconazole between January 6, and February 22, 2020. These medications were used for the pilot because they are on formulary, frequently prescribed, and limited follow-up by the medical team. Data collected included: age, sex, antimicrobial, risk factor score, loop diuretic use, potassium level, admission QTc, concurrent myocardial infarction, number of QTc prolonging medications, concurrent sepsis, and acute heart failure exacerbation which were used to tabulate a risk score. Normally electrolytes were drawn in the morning or on admission. Risk scores were categorized into low (0-6), moderate (7-10) and high risk (≥11). A daily review of patients prescribed these medications would occur, and a risk score would be calculated; if they were found to be in a moderate or high risk group, a follow-up note would be documented in a communication screen accessible to all pharmacists. A list in the EMR with all adult patients on azithromycin, ciprofloxacin, levofloxacin, and fluconazole was shared to all pharmacists. This was a real-time list as patients were added and removed based on prescribed therapies. Clinical pharmacists were notified which patients on their floors were on these therapies through the pharmacist communication screen. The risk score was typically calculated by the author to help decrease variability among scores, and the score was used throughout the admission and remained on the communication screen until the patient completed or discontinued therapy. Once reviewed, an intervention within the EMR was documented with the phrase “.QTcMonitoring”. Within the intervention documentation it stated what occurred with therapy and potential future plans such as; needs EKG for QTc monitoring, therapy is not recommended due to prolonged QTc, therapy is appropriate, and the response of the provider. Pharmacists were given a template of what to include in their intervention such as what was the initial QTc, if they notified the prescriber, what their intervention was, and the result of the intervention. This helped to streamline any previous processes into a singular standard process. This process added a step by step approach to be utilized. Data collected were pharmacist documentation of review, if an EKG was ordered, and result of requested intervention. The data was collected and analyzed using REDCAP^®. ⁷ The study was approved by the Institutional Review Board.

Primacy Outcome

The primary endpoint was successful implementation of the QTc monitoring process by pharmacists. Successful implementation was defined by a documented review of one of the pilot medications and a potential pharmacist driven intervention when therapy was adjusted, discontinued, or continued monitoring.

Secondary Outcomes

Secondary outcomes assessed include risk scores and the values used to calculate them: age, sex, antimicrobial, risk factor score, loop diuretic use, potassium level, admission QTc, concurrent myocardial infarction, number of QTc prolonging medications, concurrent sepsis, acute heart failure, pharmacist intervention, if an EKG was ordered, and results of any interventions. Secondary outcomes also included descriptions of any recommendations based on reviews generated through the process.

Results

From January 6, 2020 to February 22, 2020, there were a total of 412 orders for one of the target antimicrobials that resulted in 157 pharmacist evaluations (38.1%). Patients identified to be at the highest risk, based on scores calculated (16 high risk, and 84 moderate risk) resulted in 100 pharmacist documented reviews. For those reviews, 27/100 (27%) had an EKG repeated with an average of 30.6 hours between pharmacist intervention and repeat EKG. Retrospective review identified 15 patients detailed in the high risk target group that had therapy modified or discontinued to an alternate therapy within 5 hours of communication.

Primary Outcome

Successful implementation of QTc monitoring by pharmacists was documented in 100/100 (100%) of the high risk target group with a majority of the reviews occurring in the moderate risk group 84/100 (84%) compared to high risk 16/100 (16%).

Secondary Outcomes

Patient characteristics and pharmacist intervention details are presented in Tables 3 and 4. The most common antibiotic prescribed was azithromycin (50%), and a median risk score of 8 (range 7-16). The most common risk factor observed was age ≥ 68 (76%). Recommendations were offered 47/100 (47%) with 42/47 (89%) of the recommendations accepted. Interventions included: ordering an EKG, implementing stop dates to therapy, changing antibiotic therapy, discussion of continued monitoring, or discontinuing antibiotic therapy. For the 27 repeat EKGs, 7/27 (25.9%) were patients identified to be in the high risk group. On average they had a QTc of 477 ms on assessment, and an increase of 12 ms on average to 489 ms after repeat on targeted therapy. Evaluation review identified 3/16 (18.7%) of patients in the high risk group and 14/84 (16.6%) in the moderate group had therapy stopped due to pharmacist communication with the provider. The other 13/16 (81.2%) of the high risk score group had documented communication from the provider as to why therapy was not discontinued but would receive continued monitoring. Continued monitoring consisted of additional EKGs and placing patients on telemetry. The average time to discontinuation was 5 hours but for the 3 high risk interventions it was on average less than 1 hour to discontinue therapy. The high risk data can be found in Table 5.

Table 3.

Baseline Characteristics.

Number of patients	N = 100
Mean age (years)	75
Male (%)	31 (31)
Antimicrobial choice	N = 100
Azithromycin	50
Ciprofloxacin	13
Levofloxacin	34
Fluconazole	4
Median risk score	8
Risk factors	N = 100
One QTc prolonging drug	100
Age ≥68 y	76
Female	69
Admission QTc ≥450 ms	56
≥2 QTc prolonging drugs	52
Loop diuretic	33
Serum potassium ≤3.5 mEq/L	22
Sepsis	17
Acute heart failure	8
Acute myocardial infarction	8

Table 4.

Interventions.

Assessments	N = 100
Average QTc assessed	463 ms
Average repeat QTc	466 ms (n = 27)
Intervention offered	47
Intervention accepted	42
Continue therapy	83
Discontinue therapy	15
Electrocardiogram ordered	27
Initiate alternative therapy	11

Table 5.

High Risk Group.

High risk group EKG	N = 16
Average QTc assessed	477 ms
Average repeat QTc	489 ms
Repeat EKG	7 (25.9%)
Discontinuation data	N = 16
Discontinued therapy	3 (18.7%)
Discontinuation time	<1 h

Discussion

Established literature has suggested that patient specific risk factors may be more predictive of QT prolongation risk leading to TdP than medications alone.^6-8 These specific risk factors when reviewed and combined in a scoring system by Tisdale et al ⁴ have the potential to identify those patients at higher risk. Risk score calculators have been developed by Tisdale et al^4,5 and investigators of the RISQ-PATH score to further classify those at the highest risk. Due to the simple risk score calculator by Tisdale it was determined to calculate scores based on Tisdale’s system and not calculate scores by the RISQ-PATH since it is more cumbersome to manually calculate and the process needed to be efficient to coincide with daily workflows. Previous literature, and clinical recommendations (AHA/ACCF) suggest that reviewing and monitoring these patients can provide additional patient safety. ¹ With the current COVID-19 outbreak and the use of azithromycin and hydroxychloroquine, there were safety considerations for ventricular arrhythmia. The ACC released an article referencing the risk score calculator of Tisdale et al which provided additional support for its use in our pilot study. ¹³ CredibleMeds^® and MedSafety Scan^® (MSS) also released a decision algorithm for initiating medications with the potential for TdP which involves assessing risk factors. ⁷ This calculator along with this designed plan for implementation will assist in improving patient safety. ¹³ The next step will be to incorporate the scoring tool into the EMR and expanding on the amount of medications reviewed.

Based on the lack of a current, formalized process at our institution, we sought to explore implementing a pharmacist monitoring process that may improve the care and safety of our patients. We chose acutely prescribed agents that were meant for short term therapy to see if a successful program could be put in place. Therefore, an intervention or documentation was determined to be a successful implementation and reviews were recorded for those at the highest risk since those patients were most likely to experience a prolonged QT interval leading to Torsades. We did not observe any incidence of TdP in our patient population, which increased our validity of the monitoring process.

Limitations of this study included the inability to compare the results to previous data at our institution because there was no previous monitoring process. There may have been individual processes for monitoring, however nothing standardized to follow or track. Additional limitations to the process could be the daily intravariability of a patient risk score due to different interpretations of the scoring system, with the potential for the scoring to change daily. For example, a patient’s potassium may be ≤ 3.5 mEq/L on initial scoring and then corrected but the score was not updated. Score calculation and human error, as well as time to calculate risks could have impact on scoring efficiency and variability. These limitations can potentially be corrected with an automated scoring system built into the electronic medical record, where the score is continuously updated based on patient risk factors.

Another limitation is the timing of the EKG and repeat EKG since they may not have been timed based on dose but rather when the provider is rounding on the patient or when the nurse is available to check the EKG. If the EKG is not checked when the antimicrobials are at their peak concentrations then we are unable to check when the QT interval may be the most prolonged.

Future studies will need to determine an appropriate amount of documentations and interventions to effectively impact patient safety. This pilot study has successfully implemented the process but there is need to analyze how many interventions are possible and what other medications fit this monitoring parameter. The cohort has a small sample of high risk patients and a majority continued therapy post pharmacist interventions, however there was increased monitoring of those patients through follow-up EKGs. In the interventions that resulted in discontinuation, the time to discontinuation was immediate so this did demonstrate the impact the recommendation could make. The therapy may not always be discontinued but increased monitoring did result from a pharmacist review.

Conclusion

The QTc monitoring process was a successful implementation. The results demonstrated a successful implementation as observed through high percentage of pharmacy evaluations on identified higher risk patients and their outcomes observed through communications. Thus, improvements and knowledge of the process will increase monitoring and may lead from a pilot to an established process which will allow for increased safety for our patients.