 |
 |
 |
 |
 |
 |
 |
|
|
|
|||||||||
|
|||||||||||
Head of Pharmacy, VieCuri Medical Centre, Venlo, Netherlands; Department of Medical Informatics, Radboud University Nijmegen, Nijmegen, Netherlands
Hospital Pharmacist, VieCuri Medical Centre
SIR Institute for Pharmacy Practice Research, Leiden, Netherlands
Department of Clinical Pharmacy, Radboud University Nijmegen; Scientific Institute Dutch Pharmacists, The Hague, Netherlands
Reprints: Dr. van Dijk, VieCuri Medical Centre, PO Box 1926, 5900 BX Venlo, Netherlands, fax 31-77-320 6001, lvdijk{at}viecuri.nl
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
OBJECTIVE: To perform a retrospective analysis of the incidence of required versus implemented dosage adjustments according to guidelines in patients with renal insufficiency at discharge and evaluate specific determinants responsible for the percentage of overlooked dosage adjustments.
METHODS: Medication and laboratory data were collected from patients at discharge during February and November 2004. For patients with a calculated creatinine clearance less than 51 mL/min/1.73 m2, the necessity for dosage adjustments was evaluated by pharmacists. All data were collected for further research and subsequent statistical analysis.
RESULTS: At discharge, 237 of 647 (36.6%) patients had a calculated creatinine clearance less than 51 mL/min/1.73 m2. Dosage adjustment based on renal function was necessary in 411 of 1718 (23.9%) of prescriptions. These adjustments were performed in 242 (58.9%) prescriptions and not performed in 169 (41.1%) cases. The risk of not adjusting the dosage was significantly associated with serum creatinine levels greater than 1.71 mg/dL and creatinine clearance less than 35 mL/min/1.73 m2 (p < 0.05). The risk of not adjusting the dosage was also significantly associated with drugs producing severe consequences when dosing guidelines were overlooked (p < 0.05).
CONCLUSIONS: In patients with a calculated creatinine clearance less than 51 mL/min/1.73 m2, dosing according to their renal function can be improved. Because of the need and the association found in this study, an alert system could help prescribers and pharmacists to adapt drug dosage in patients with renal impairment.
Key Words: dosage guidelines, medication safety, renal impairment, serum creatinine
Published Online, June 27, 2006. www.theannals.com, DOI 10.1345/aph.1G742
Automated systems providing alerts and dose checks for medication taken by patients with impaired renal function are not yet common in daily practice. To assess their practical role, we set out to determine whether a reminder system would be useful in our setting. Therefore, a retrospective analysis was performed on drug dosage adjustments in patients with impaired renal function at hospital discharge, a time at which the patient's medication profile is evaluated by the physician. Most patients leave the hospital in stable condition with medication for chronic use, at least until their next consultation. This makes the discharge counseling an important moment for a check and subsequent intervention.
Our analysis focused on the following objectives: (1) determining the percentage of people with impaired renal function at discharge, (2) establishing the frequency of inappropriate dosing at the time of discharge in accordance with renal function, and (3) gaining insight into the circumstances or factors that give rise to inappropriate drug dosing, such as patient characteristics, (groups of) drugs involved, and physician practice type. Furthermore, other potential determinants were assessed, including an estimate of the clinical relevance of the necessary dosage adjustment.
Unlike previous reports, our study focused on required dosage adjustments for patients with renal impairment at hospital discharge. These patients, with chronic drug therapy, require optimized pharmaceutical care.9,10
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
Subsequently, these prescriptions are sent to the patient's pharmacy or nursing home. At the same time, the general practitioner or nursing home physician receives the medication profile at discharge. The service has been implemented in regular wards for patients who are on several drugs. Intensive care units, pediatrics, and maternity and psychiatric wards are excluded from the discharge service. This service results in 5000 consultations at discharge out of approximately 18 000 hospital admissions annually.
DESIGN
All medication records of the patients who had been discharged from the
hospital through the discharge counseling service during 2 randomly chosen
months, February 2004 and November 2004, were studied. The following
characteristics were collected: sex, age, duration of hospital stay,
creatinine values, drugs prescribed at discharge, and physician specialty.
Creatinine values were extracted from the laboratory system of the hospital. Renal function was estimated by calculating the average creatinine clearance based on the average of the last 2 serum creatinine levels during admission. If only one creatinine value was available, this one was used to calculate creatinine clearance.
Creatinine clearance (Clcr) in mL/min/1.73 m2
was calculated by using the Jelliffe-I
formula.11
 |
 |
To assess the need for dosage adjustment, we referred to the following handbooks: National Drug Compendium from the Royal Dutch Pharmaceutical Society,14 Renal Drug Handbook,12 Drug Prescribing in Renal Failure,13 and online information from Micromedex Healthcare service.15 If guidelines differed among these references, either the lowest percentage of the usual dose or the maximum interval prolongation was applied.
All of the prescriptions for each patient were compared with those that needed dosage adjustment according to these handbooks. Prescriptions for which dosage guidelines were followed were defined as controls. Prescriptions were defined as cases when a contraindicated drug was prescribed or dosage adjustments were not performed.
To assess the clinical relevance of each requirement for dosage adjustment, a list of the involved drugs was assessed by a panel of experts: 2 clinical pharmacologists and a nephrologist. Each drug received code 1, 2, or 3, respectively, for clinically irrelevant, moderate, or severe consequences in the event of not using the guidelines for dosage adjustments in patients with a creatinine clearance less than 51 mL/min. In the analysis, relevance codes 2 and 3 are combined because both lead to similar interventions in daily clinical practice.
All data were collected by 2 clinical pharmacists and recorded in an Excel (MS Windows XP) spreadsheet for classification and submitted to statistical analysis using SPSS for Windows version 10.1 (SPSS, Chicago).
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
|
|
In 411 prescriptions for 188 patients (79.3%), dosage adjustment was required according to renal function. This adjustment was done in 242 prescriptions. The other 169 medications, representing the cases, were not adjusted, involving 48.3% of the patients.
In the control group, 57% of the performed dosage adjustments were considered to be of minor clinical importance (relevance code 1). The other 43% of prescriptions (23 drugs) were deemed to be of more clinical importance (relevance code 2 or 3).
In cases where no dosage adjustments were performed, 67 of the prescriptions were of minor importance (relevance code 1) and 102 were moderately serious to serious (relevance code 2 or 3) (Table 3). The latter category consisted of 32 drugs. In this category, 7 drugs were contraindicated based on renal function: acitretin, alendronate, amiloride/hydrochlorothiazide, benzbromarone, hydrochlorothiazide, nitrofurantoin, and risedronate.
|
Most (53.6%) of the patients were discharged by the cardiology and internal medicine services. Together, these departments were responsible for 60% of the patients with one or more drugs without appropriate dosage adjustment. Patients were discharged without dosage adjustment in one or more medications in the cardiology department (41 of 72; 56.9%), internal medicine (28 of 55; 50.9%), surgery (19 of 41; 46.3%), neurology (7 of 17; 41.1%), orthopedic surgery (7 of 18; 38.9%), and lung diseases (11 of 32; 34.4%). Overall, all departments were not adjusting dosages in about 50% of their discharged patients.
STATISTICAL ANALYSIS
Possible influential factors were traced by means of a determinant
analysis. In this analysis, controls were defined as prescriptions adjusted
according to recommendations in the handbooks previously mentioned.
Prescriptions for which dosage adjustments were not performed were defined as
cases.
There were no significant differences between cases and controls with respect to the variables of age, sex, duration of hospital stay, number of prescriptions, or physician specialty. Significant associations were found for the variables serum creatinine, creatinine clearance, and relevance (all p < 0.05). Because serum creatinine and creatinine clearance are inversely proportional to each other, 2 models were used to calculate adjusted odds ratios.
In these models, the association and significance remain almost the same as the crude results, suggesting that high serum creatinine values (>1.71 mg/dL) and low creatinine clearances (<35 mL/min/1.73 m2) significantly increased the risk that a dosage adjustment was not performed. The same association was found when the clinical relevance of the dosage adjustment was thought to be moderately severe or severe (relevance code 2 plus 3) (Table 4).
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
Other studies have also shown inappropriate dosing with respect to renal function.16,17 Chertow et al.16 found that, in 7490 patients with renal insufficiency, 15% of the medication orders needed adjustment in at least 1 dosage parameter. This is far below the 23.9% we found in our population. However, these results cannot be compared. Chertow et al. performed an intervention study among inpatients, including women on gynecology and obstetrics wards, who used less medication. In addition, the authors defined impaired renal function as a calculated creatinine clearance less than 80 mL/min.
Salomon et al.18 also performed a study among inpatients. They identified 886 (60.3%) prescriptions that needed dosage adjustment according to guidelines. Of these, 585 (66.0%) complied with guidelines, whereas 301 (34.0%) did not. In the latter category, 124 (14.0%) contraindicated medications were involved. We found 58.9% compliance with prescribing guidelines. In the noncompliant category, we found only 2.4% contraindicated drugs. The patients of Salomon et al. were frequently prescribed antibiotics for which hospital guidelines are readily available, whereas our study focused on patients at discharge, who are primarily on chronic medications. Furthermore, Salomon et al. studied all patients regardless of their creatinine clearance.
Our results show that dosage adjustments are mostly overlooked in commonly prescribed drugs such as oral antidiabetics, serum lipid-reducing agents, antibacterials, and drugs for treatment of bone diseases. The cause of concern is that drugs that are contraindicated in patients with impaired renal function are often prescribed. Some of these agents were administered only once; however, even when use is limited, not adjusting doses based on renal function can be considered a serious medication error. For example, when digoxin is prescribed, renal function is usually taken into consideration (38 of 44; 86.4%). Nevertheless, in 13.6% of the prescriptions in our study, the dosage of digoxin was too high.
Statistical analysis showed some unexpected associations. A significant association was found between worsening renal function (<35 mL/min/1.73 m2) and the incidence of nonperformed dosage adjustments. Furthermore, the more serious the consequences of nonadjustment could be, the more likely it was that a guideline for dosage adjustment had not been followed. Both associations seem illogical, as one would expect physicians to be more accurate in drug dosing in patients with severe renal impairment or in prescribing drugs for which dosage adjustment is more relevant. Our results do not indicate that physicians practicing in surgical departments make more dosing errors than those in nonsurgical departments (Table 4). Perhaps many of the prescriptions may have been written by inexperienced residents; this possibility has not been explored.19
|
Limitations |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
The time of discharge was chosen for the analysis because these patients are leaving the hospital with medication prescribed primarily for chronic use. Because patients with impaired renal function are vulnerable for an adverse drug event after hospitalization, drugs should be prescribed judiciously. Monitoring medication records at discharge could be a useful tool to prevent rehospitalization.20
In our population, renal function was roughly estimated by calculating creatinine clearance with the Jelliffe-I equation, based on serum creatinine values. This equation is not completely reliable. Because neither body weight nor height data are easily accessible in our hospital and the ranges of creatinine clearance in handbooks for dosage adjustments are very broad (eg, >50, 10-50, <10 mL/min), a more accurate calculation of creatinine clearance is unnecessary. Since creatinine clearance based on serum creatinine levels overestimates renal function in the lower region and correlation with renal function decreases with increasing age, the percentage of patients with creatinine clearance less than or equal to 50 mL/min might even be higher.21
By calculating creatinine clearance, we also wanted to identify patients with impaired renal function and lower serum creatinine levels (eg, the elderly) due to reduced muscle mass.
Examining the clinical relevance of noncompliance with dosing guidelines in terms of risk for readmission and comorbidity was not the aim of our study. We wanted to get an impression of the compliance to present dosage guidelines and possible influencing factors. In our opinion, the number and administration frequency of the drugs prescribed is not so relevant when they are dosed according to renal function. Rather, the impact of a single inappropriate drug dosage, especially with contraindicated agents, can produce a negative effect for an individual patient.
As in other studies, we relied on the dosage guidelines presented in frequently used handbooks and expert opinions. The remarkable variations in definitions and recommendations, based on scarce data, make the available sources of these guidelines less reliable.22 Physicians may refrain from dosage adjustments because of these confusing and often conflicting sources.
The unexpected finding that worsening renal function and use of a prescribed drug with relevance code 2 or 3 increase the risk of not using dosage guidelines suggests physician unawareness in drug prescribing according to renal function. This counterintuitive result justifies further research, because an unadjusted dose for a specific patient may be the physician's intention.
The question of why some dosages are adjusted and others are not could not be answered by this study. Most physicians use a selected group of drugs that they know well. One of the reasons why the dosage of some drugs is not adjusted may be a lack of awareness of worsening renal function in the aging patient. By asking physicians the reasons for their decision, this question can probably be answered.23
Extrapolating the results from our study, it can be implied that, for our hospital where 5000 patients are discharged annually through the counseling service, 900 patients could be traced. An intervention might be necessary in this population, leading to 3600 dosage adjustments yearly or at least 15-20 interventions per workday. A dynamic alert system would be useful to monitor the dosage of certain drugs in relation to a patient's renal function. In this matter, we agree with the conclusions of other authors that it is better to focus on improving medication systems than to target patients who are prone to developing adverse drug effects.8,24
Linking medication data to renal function parameters offers the clinical pharmacist the ability to check medication profiles more quickly and accurately, which seems to be the solution to reducing medication errors and improving pharmaceutical care.
Further research is necessary before the specifications of such a system can be developed according to the needs and wishes of pharmacists and physicians. We have started validating and implementing such a system in our hospital.
|
Conclusions |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
Because of these findings, and especially the unexpected association found in this study, we strongly recommend uniformity in dosing guidelines. Subsequently, these guidelines can be implemented in a dynamic alert system that could help physicians and pharmacists to adapt drug dosing in patients with renal impairment. Such a computerized knowledge system, which would select patients at risk and link laboratory data to pharmacy data, can provide uniformity in dosing advice regardless of the healthcare professional. This would lead to continuous and comprehensive screening of all contraindicated drugs and inadequate dosages. We believe that the items mentioned above are essential to optimize pharmaceutical care for all patients.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionLimitationsConclusionsReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. A van Dijk, N. R. Drabbe, M. Kruijtbosch, and P. A. De Smet Dosage Adjustments According to Renal Function at Discharge: A Comparison of 3 Hospitals Ann. Pharmacother., December 1, 2006; 40(12): 2276 - 2277. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
 |
 |
 |
 |
 |
 |
