Evaluation of Dose Adjustment in Patients With Renal Impairment at Muhimbili National Hospital in Tanzania

Abstract

Background: The burden of renal diseases is increasing in developing countries like Tanzania. Drug accumulation exposes patients with renal impairment to drug toxicity that may lead to adverse drug reactions, poor adherence to treatment, and increased healthcare costs. There is limited information on the appropriateness of dosage regimen adjustment for patients with renal impairment, particularly in developing countries such as Tanzania. This study aimed to investigate the appropriateness of drug dosing in hospitalized patients with renal impairment in Tanzania. Methods: This was a retrospective cross-sectional study. It was conducted between November 2019 and April 2020 amongst hospitalized patients at Muhimbili National Hospital. All enrolled patients had serum creatinine levels ≥1.2 mg/dL and taking at least one drug requiring dosage regimen adjustment. Creatinine clearance was calculated from patient serum creatinine using the Cockcroft-Gault equation. Drug dosing appropriateness was determined by comparing the current practice with tertiary references. The relationship between the patient’s baseline characteristics and the rate of dosage regimen adjustment was determined using the X² test. Univariate and multivariate logistic regression analysis evaluated the predictors of dosing adjustment. Results: Most of the enrolled patients, 269 (98.9%) had comorbidities. Of the medication orders included in the final analysis, 372 (27%) needed dosage regimen adjustment. Out of the 372 medication orders, not adjusted were 168 (45.2%), inappropriately adjusted 105 (28.2%), and appropriately adjusted were only 99 (26.6%). In this study, 212 (77.9%) patients received at least one drug with an incorrect dosage regimen. Females and those with level 4 renal impairment patients were more likely to have their doses appropriately adjusted compared to their counterparts. Conclusions: In this study, about three-quarters of the patients received at least one drug with an incorrect dosage regimen. Thus, appropriate measures such as the availability of national guidelines and clinical decision support systems for drug dosing adjustment in patients’ renal impairment should be in place.

Background

Worldwide, the burden of acute and chronic kidney diseases (CKD) is estimated to be about 13.5% and 20% among pediatric and adult populations, respectively. ¹ The number of patients with kidney disease who progress to end-stage renal disease (ESRD) requiring renal replacement therapy is increasing. Worldwide, renal impairment contributes to 5-10 million deaths annually. ² The burden of renal impairment is at an increase in Sub-Saharan Africa (SSA) countries such as Tanzania, owing to similar risk factors like obesity, diabetes mellitus, hypertension, and heart disease.^3-5 Chronic use of nephrotoxic medications such as antiretroviral drugs like tenofovir fumarate, non-steroid anti-inflammatory drugs (NSAIDs), and uncontrolled herbal and traditional medicines contribute to this growing tragedy.^3-5

Renal impairment results in catastrophic consequences on the clearance of natural and xenobiotics substances like medications. ⁶ Alteration of drug clearance may lead to accumulation in the body, and the risk of drug toxicity and/or adverse drug reactions (ADRs) ³ may increase. Patients taking dialyzable drugs on hemodialysis are at risk of subtherapeutic drug concentration due to increased drug clearance which may lead to treatment failure. Although drugs that are not dialyzable like acyclovir can accumulate during renal dialysis leading to supratherapeutic drug concentration. ⁴

For patients with renal impairments, healthcare professionals must be knowledgeable about drug disposition and the extent of the patient’s altered physiology before prescribing medicines. ⁵ However, previous studies indicated that about 25%-77% of patients with CKD were not properly dosed.^1,2,7,8 So far, limited information on prescribing patterns among patients with renal impairment is available in Tanzania. The study evaluated the appropriateness of drug dosing adjustment among renal-impaired patients hospitalized at Muhimbili National Hospital (MNH)-in Tanzania.

Methods

Study Design

This was a retrospective cross-sectional study among hospitalized patients with renal impairments at Muhimbili National Hospital (MNH) in Tanzania.

Study Setting and Duration

The study was conducted between November 2019 and April 2020 at MNH, a national referral hospital and University teaching hospital. Approximately, 1000 to 1200 patients are hospitalized per day. As a national referral hospital, MNH receives almost all renal impairment patients from the 26 regions of Tanzania’s Mainland. Even though the hospital has a specialized dialysis clinic, this study targeted the patients who are admitted to the general ward. The ward where medication records were taken has a pharmacy that issues medication orders to fill prescriptions given by clinicians. The pharmacist at the pharmacy department reviews the prescriptions before being issued.

Study Population

Patients included in this study had acute kidney injury (AKI) or CKD based on the clinical evaluation and laboratory data obtained from patients’ files. AKI for this study was considered as an abrupt (within hours) decrease in kidney function, which encompasses both injury (structural damage) and impairment (loss of function). ⁹ CKD was the presence of an abnormality in kidney structure or function persisting for more than 3 months. ⁶

Study Eligibility Criteria

Patients over 18 years of age receiving at least one pharmacological agent requiring renal dose adjustment with initial serum creatinine (SCr) levels ≥1.2 mg/dL were included in the study.

Patients not receiving any pharmacological agent(s) and/or not requiring dosing adjustment and pregnant were excluded from the study.

Sample Size Calculation

The estimated sample size was estimated using the cross-sectional formula. ¹⁰ The sample size calculated was 384 patients’ medical files.

A Sampling of Study Participants

A systematic random sampling technique was used to obtain the sample size of 384 from 669 patients’ files. Yet only 272 hospitalized patients met the inclusion criteria.

Creatinine Clearance Calculation

Cockcroft Gault (CG) equation was used to calculate creatinine clearance (CrCl) for each patient from serum creatinine levels recorded before drug prescribing. ¹¹ The Cockcroft-Gault equation is shown below for men and women.

Men

$$CrCl\left(\frac{ml}{min}\right)=\frac{[(140-age)xweight(kg)]}{SCr\left(\frac{mg}{dl}\right)x72}$$

Women:

$$CrCl\left(\frac{ml}{min}\right)=\frac{[(140-age)xweight(kg)]x0.85}{SCr\left(\frac{mg}{dl}\right)x72}$$

Adjustment of CrCl for obese patients was by a factor of 0.4. ¹² The corresponding GFR category was assigned for each patient based on laboratory results at the time of data collection. Patients included, in the final analysis, were those with the following GFR level 3 (30-59 mL/min), level 4 (15-29 mL/min), and level 5 (<15 mL/min).

Data Collection

Data were collected from patients with renal impairment in the internal medicine ward (nephrology unit). Individual patient data such as date of admission, diagnosis, age, gender, weight, serum creatinine, dialysis status, comorbidities, and medications and dosage regimen information were obtained from the patient physical medical file maintained by the MNH. Each patient was assigned a unique study ID. The data were collected through a pre-prepared standard data collection form.

Data Management and Statistical Analysis

Data collected were edited, cleaned, and analyzed using Statistical Package for Social Sciences (SPSS), version 24 (IBM Corp, New York USA). All figures were plotted using the R statistical software version 3.5.2. Data were summarized and described as tables, graphs frequencies, and percentages. Prescribed medicines were grouped as antihypertensive, hypoglycemic, antimicrobial, and miscellaneous.

Drug dosing appropriateness was determined by comparing the current national practice with the tertiary references.^13,14 The review of patient dosage regiment appropriateness was done by the study team and audited by 2 senior clinical pharmacy pharmacists from MNH.

The Chi-Square test compared drug dosing appropriateness and patients’ socio-demographic characteristics. Univariate and backward stepwise elimination multivariate logistic regression computed predictors of the appropriateness of drug dosing adjustment. Pearson’s correlation coefficient test was used to determine the linear correlation between drug dosing adjustment, socio-demographics, and participants’ clinical characteristics. Descriptive statistics were used wherever appropriate. Two-tailed P-value < .05 was considered statistically significant at a 95% confidence interval. Appropriate dose adjustment as per guidelines was assigned a “yes” and inappropriate dose adjustment contrary to what was in the guidelines was assigned a “no”

Results

In total 384 patients’ files randomly were sampled and screened from 669 patients’ medical files. Out of 384 files screened, only 272 patients met the inclusion criteria. More than half of the participants were male. The median age in years and median weight in kg with interquartile range (IQR) were 50 (35-61) and 68 (60-76), respectively. The median serum creatinine (IQR) was 9.3 mg/dL (5.7-14.3) and the median GFR (IQR) was 8.3 mL/min (5.3-13.8). In this study, 214 (78.7%) and 58 (21.3%) patients had CKD and AKI, respectively. Patients on hemodialysis were 116 (54.2%) of that 19 (16.4%) and 97 (83.6%) had AKI and CKD, respectively. In addition, 208 (76.8%), 43 (15.8%), and 20 (7.4%) of the studied patients were in levels 5, 4, and 3 of renal impairment, respectively. In totality, patients with GFR levels below 30 mL/min were 251 (92%).

Almost, all studied patients 269 (98.9%) had comorbidities where 190 (69.9%) had hypertension, 76 (27.9%) had diabetes, 33 (12.1%) had sepsis and 30 (11.0%) had HIV infection. Tenofovir-induced nephrotoxicity (TIN) was the cause of renal impairment in patients with HIV infection. On average, each patient took 5 medications per day. Patients’ socio-demographic and clinical characteristics are summarized in Table 1.

Table 1.

Patients’ Social Demographic and Clinical Characteristics.

Demographic and clinical data	Frequency (%)
Number of Patients with Renal Impairment	272 (100)
Region of residence
Dar es Salaam	182 (66.9)
Upcountry	90 (33.1)
Gender of patient
Male	145 (53.3)
Female	127 (46.7)
Referral status
Hospital Referral	145 (53.3)
Self-Referral	127 (46.7)
Age category (years)
18-29	45 (16.5)
30-39	44 (16.2)
40-49	45 (16.5)
50-59	56 (20.6)
60 +	82 (30.1)
Type of renal impairment
CKD	214 (78.7)
AKI	58 (21.3)
Patient on hemodialysis (HD)
Yes	116 (42.6)
No	156 (57.4)
Renal impairments level
Level3 (30-59)	20 (7.4)
Level4 (15-29)	43 (15.8)
Level5 (<15)	208 (76.8)
Number of drugs per patient
1-3	65 (23.9)
4-6	140 (51.5)
7-9	59 (21.7)
10 +	8 (2.9)
Comorbidity present
Yes	269 (98.9)
No	3 (1.1)
Specific comorbidities
Hypertension	190 (69.9)
Diabetes	76 (27.9)
Heart failure	17 (6.3)
Malaria	17 (6.3)
HIV	30 (11.0)
Sepsis	33 (12.1)
Pneumonia	18 (6.6)
Cancer	21 (7.7)
Other	137 (50.4)
Patients with HIV Having TIN	15/30 (50.0)
Median Weight (kg); IQR	68 (60-76)
Median Serum Creatinine (mg/dL); IQR	9.3 (5.7-14.3)
Median GFR (mL/min); IQR	8.3 (5.3-13.4)

Note. TIN = Tenofovir Induced Nephrotoxicity; IQR = interquartile range.

Evaluation of Dosage Regimen Adjustment in Patients With Renal Impairment

In this study, of the prescribed 1396 medications 21 (1.5%) were contraindicated in patients with renal impairment. The final analysis included 1375 medication orders of that 372 (27%) required dosage regimen adjustment. Out of the 372 medication orders that needed dosage regimen adjustment, 168 (45.2%) were not adjusted at all, 105 (28.2%) were inappropriately adjusted, and 99 (26.6%) were dosed appropriately (Figure 1). In total, 212 (77.9%) patients received at least one drug with an incorrect dose.

Figure 1.

Drug dosing adjustment medication profile.

Evaluation of Appropriateness of Drug Dosing Adjustment for Individual Drug

Of the 372 medication orders that required dose adjustment concerning individual drugs, hydralazine orders were the least appropriately dosed, of which 79 (21.2%) of its order were inappropriately dosed, followed by a combination of amoxicillin and clavulanic acid 54 (14.5%) orders, bisoprolol, and metoclopramide each 19 (5.1%) orders, meropenem 13 (3.5%) orders, ciprofloxacin 12 (3.2%) orders, and pregabalin 10 (2.7%) orders (Table 2).

Table 2.

Specific Medication and Appropriateness of Drug Dosing Adjustment.

Drug name	The dose was adjusted appropriately?
	No	Yes	Total
Hydralazine	79	31	110
Amoxicillin/clavulanic acid	54	10	64
Paracetamol	8	30	38
Metoclopramide	19	2	21
Bisoprolol	19	2	21
Ciprofloxacin	12	1	13
Meropenem	13	1	14
Glimepiride	5	4	9
Glibenclamide	8	0	8
Clonidine	5	0	5

Evaluation of Appropriateness of Drug Dosing Adjustment for Medication Categories

Dose adjustment based on different medication classes depicted in Figure 2, analgesics had the highest proportion of medication orders adjusted appropriately (71.3%, 95% CI: 57.5-82.7) compared to other classes. In addition, antiemetics had the lowest proportion of medication orders (9.5%. 95% CI: 2.7-28.9) adjusted appropriately. Overall, drugs were significantly prescribed inappropriately (P < .005) (Figure 2)

Figure 2.

Appropriateness of dose adjustment among patients with renal impairment for different drug categories.

Association of Social Demographic and Clinical Characteristics of Participants With Drug Dosing Adjustment

In this study, female patients were more appropriately dosed (33.5%, 95% CI: 26.7-41.1) compared to male patients (21.3%, 95% CI: 16.3-27.3) and the difference was statistically significant (P = .008) (Figure 3).

Figure 3.

Appropriateness of dose adjustment between males and females’ patients with renal impairment.

In this study, the GFR levels and age distribution were almost the same for both male and female study participants, for both male and female study participants, the GFR levels and age distribution were almost the same.

A correlation between inappropriate dosing and GFR level was not found (r² = .0049, P = .1) Even so, many patients who were inappropriately dosed had GFR levels below 20 mL/min.

Patients not on hemodialysis were appropriately dosed (28.1%, 95% CI: 17.1-36.7) compared with those on hemodialysis (25%, 95% CI: 22.1-32.1), yet, the differences were not statistically significant (P = .496).

Furthermore, there was no correlation between GFR levels and the age of patients on hemodialysis who were appropriately dosed compared to those not dosed appropriately (r² = 0.0036, P = .2802). For most of the patients on hemodialysis, 89.7% (104) had GFR levels below 20 mL/min. Patients on hemodialysis were across all age groups.

Additionally, drug dosing in patients with renal level 4 was more appropriately adjusted (42.5%, 95% CI: 31.6-54.5) compared with those in level 5 (23%, 95% CI: 18.5-28.1) (P = .004).

Univariate and Multivariate Regression Analysis

On univariate and multivariate analysis, crude odd ratio (COR) and adjusted odd ratio (AOR), respectively, revealed that age category, renal impairment type, SCr, and dialysis status had no significant association with the appropriateness of dosage regimen adjustment in the studied patients (Table 3). Although, sex and level 4 renal impairment were significantly associated with appropriate dose adjustment on univariate analysis. Female patients (COR = 1.86, 95% CI: 1.17-2.96) were more likely to have their doses appropriately adjusted compared to male patients. This association was also, observed in multivariate analysis; female patients had higher odds (AOR = 2.10, 95% CI: 1.27-3.45) of having their doses appropriately adjusted compared to men. For patients with level 4 renal impairment doses were more appropriately dosed than any level of renal impairment (COR = 2.49, 95% CI = 1.44-4.33). Similar results were seen in multivariate analysis (AOR = 2.88, 95% CI = 1.41-5.87) (Table 1).

Table 3.

Univariate and Multivariate Logistic Regression of factors Associated with Dose Adjustment in Patients with Renal Impairment.

		Univariate regression			Multivariate regression
Variable	Categories	cOR	95% CI	P-value	aOR	95% CI	P-value
1. Sex	Male	1			1		.004
	Female	1.86	1.17-2.96	.009	2.10	1.27-3.45
2. Renal Impairment Type	CKD	1			1		.990
	AKI	0.94	0.53-1.68	.839	1.00	0.52-1.94
3. Age Category	18-29	0.67	0.32-1.41	.289	0.63	0.27-1.47	.284
	30-39	0.54	0.25-1.14	.105	0.58	0.26-1.29	.575
	40-49	0.64	0.32-1.25	.19	0.54	0.26-1.11	.094
	50-59	0.75	0.40-1.42	.38	0.77	0.39-1.51	.451
	60 +	1			1
4. Renal Impairment levels	G3	1.12	0.22-5.67	.893	1.48	0.25-8.64	.664
	G4	2.49	1.44-4.33	.001	2.88	1.41-5.87	.004
	G5	1
5. Number of medications	1-3	1.57	0.44-5.65	.488	1.66	0.42-6.57	.468
	4-6	1.33	0.42-4.20	.627	1.24	0.37-4.19	.731
	7-9	1.40	0.43-4.52	1.395	1.28	0.37-4.49	.698
	10 +	1
6. Serum creatinine (mg/dL)	3.21	0.97	0.93-1.00	.075	1.01	0.97-1.06	.677

Discussion

Several studies have reported data regarding medication dosing in patients with renal impairment ^2,15-17 However, to our best knowledge, this is the first study to report medication dosing among renal-impaired patients in Tanzania. In this study, the majority of patients with renal impairment were inappropriately dosed. Also, female patients and renal impairment level 4 were significant predictors of appropriate dosing.

Findings from this study demonstrate that 77.9% of the renal impaired patients received at least one medication with an incorrect dose. Among the 372 medication orders issued to these patients that required dosage regimen adjustment, only 99 (26.6%) were adjusted correctly. Of the incorrectly dosed orders, 168 (61.5%) were not adjusted, and 105 (28.2%) were inappropriately adjusted. These findings are comparable to others conducted in Palestine where 73.6% of 193 prescription orders were inappropriately dosed. ¹⁸

Several previous studies have reported various rates of inappropriate dosing in renal-impaired patients. Ethiopia’s study reported a rate of 51%, ¹⁵ Lebanon’s study of 63%, and South Africa’s study of 68%.^7,16 The reported rates are lower than the 73.4% in this study. On the other hand, higher rates of about 81.1% of inappropriate dose adjustments have been reported. ² A recent study conducted in Ethiopia reported inappropriate dose adjustment in patients with renal impairment to be as high as 42.6%. ¹⁹

The possible explanation for the highly inappropriate dosing adjustment in the current study could include but not limited to be the large number and continuously increasing list of drugs that requires dose adjustment, making it hard for medical personnel to stay up to date; underestimation or lack of knowledge on the importance of using CrCl in the determination of appropriate dose. ²⁰ Also, clinicians did not review renal function tests before prescribing, as explained by a high number of patients excluded in this study due to missing serum creatinine levels in the medical records and poor involvement of clinical pharmacists in reviewing prescriptions for patients with renal impairment.^6,17 Other reasons could be the lack of national dosage regimen adjustment guidelines for patients with a renal impairment that is validated and approved for use in local settings and creatinine clearance is not routinely calculated by a laboratory technician when requested by the clinician, thus, requiring self-calculation. Comparable explanations have been given in previous studies.^7,15,16 Furthermore, patients with CKD and the elderly taking several drugs (polypharmacy) due to comorbidities as found in this study require dosage regimen adjustment. Thus, appropriate dose adjustment requires sufficient knowledge of medication-medication and medication-disease interaction. Encouraging teamwork involving physicians, pharmacists, and nurses during patient management could help reduce the burden of inappropriate dosing among patients with renal impairment. A recent systematic review has reported a positive impact on the reduction of inappropriate prescribing in patients with renal dysfunction when physicians received immediate concurrent feedback from a clinical pharmacist. ²¹ Also, among the important strategies for improving patient care is the involvement of clinical pharmacists and all clinical care team members.^17,22

Females significantly predicted appropriate dose adjustment compared to males. The possible reason for this finding could be the social behaviors of women in seeking health services. Usually, women may easily ask questions about their disease and medications that might influence their appropriate management compared to men not easily ask questions. ²³

The findings of this study indicate that analgesics were significantly appropriately adjusted compared to other groups of medications. Analgesics of particular NSAIDs are known to cause renal complications and could account for this finding.

In this study, surprisingly patients with renal impairment level 4 were dosed more appropriately compared to those with level 5 renal impairment. Even so, patients with renal impairment level 5 were expected to be dosed more appropriately than those with lower renal impairment levels. The observed finding could be by chance that prescribers do not concentrate on the dosing adjustment requirements among patients with renal impairment.

The adoption of some methods suggested for improving drug dose adjustment would reduce the impact of inappropriate dosing among renal-impaired patients. Activities like education, via standardization of prescription sources, updated prescription protocols, and pocketbooks or tablets with dosing guidelines, as a complete interventional program, could substantially reduce prescribing errors. ²³ A previous study showed that computer-aided applications are helpful in this area. ²⁴ In patients with renal impairment, computerized physician order entry and clinical decision support systems can improve the quality of prescribing decisions.^24,25

This study had some limitations. The cut-off SCr ≥ 1.2 mg/dL might have missed some patients with renal impairment as some patients can have renal impairment despite having an SCr value of below 1.2 mg/dL. But, setting up a cut-off point was necessary to ensure that patients included in the study had renal impairment. Again, practitioners might have been using different guidelines from the one used in this study as there are no standardized national guidelines for dosing patients with renal impairment. Also, an appropriate decision on whether drug dosing adjustment is needed or not for a patient on dialysis often is challenging. The estimation of GFR becomes problematic and unreliable for patients who experience rapid changes in renal function, like AKI patients.

Conclusion

This study found that about three-quarters of the patients received at least one drug with an incorrect dosage regimen. Appropriate measures like the availability of national guidelines and clinical decision support systems for drug dosing adjustment for patients with renal impairment should be in place. This will be useful in preventing poor clinical treatment outcomes and toxicity resulting from dosing errors.