SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References

J Clin Hypertens (Greenwich). 2011;13:473–478.©2011 Wiley Periodicals, Inc.

Renin-angiotensin system inhibitor (RASi) agents improve renal and cardiovascular outcomes in patients with diabetes mellitus (DM) and chronic kidney disease (CKD). Studies examining conformity to guidelines have relied on pharmacy claims or filled prescriptions rather than provider-based information. The main outcome of RASi use was examined in 2889 patients with CKD and DM for its association with patient characteristics and specialty clinic visits. RASi use was 77% during the 2-year study period, and declined with worsening CKD stages (79%, 59%, and 48% in stages III, IV, and V, respectively; P<.0001). By multivariate analysis, hypertension (odds ratio [OR], 5.49, 95% confidence interval [CI], 4.16–7.25); older age (OR, 0.85; 95% CI, 0.78–0.93), and higher glomerular filtration rate (OR, 1.42; 95% CI, 1.31–1.53) were associated with RASi use. In a model examining the effect of each specialty, RASi use was greater in patients attending cardiology (OR, 3.52; 95% CI, 2.63–4.71), pharmacy (OR, 3.15; 95% CI, 2.49–3.98), endocrine (OR, 3.39; 95% CI, 2.22–5.16), and renal clinic visits (OR, 2.04; 95% CI, 1.54–2.71). Diagnosis of hypertension increases RASi usage, whereas older age and lower glomerular filtration rate reduce that likelihood. Appropriate specialty referrals improve conformity to guidelines in practice.

Chronic kidney disease (CKD) affects 16% of the US population and is a precursor to end-stage renal disease (ESRD).1,2 Diabetes mellitus is the single largest contributor to the growing prevalence of CKD and accounts for up to 45% of new cases of ESRD. According to the National Diabetes Fact Sheet (2007) (http://www.cdc.gov), 23.1% of all US adults 60 years and older have diabetes. It is well recognized that patients with CKD and diabetes mellitus are among the highest-risk groups for developing cardiovascular (CV) events and experience relentless progression of their kidney disease to ESRD.3–5

Renin-angiotensin system inhibitor (RASi) agents (angiotensin-converting enzyme [ACE] inhibitors and angiotensin receptor blockers [ARBs]) have revolutionized the treatment of patients with CKD and diabetes. Grade A evidence supports the use of RASi agents to reduce the risk of CV events and delay the onset and progression of kidney disease in patients with diabetes, independent of blood pressure control.6–9 Based on accumulated evidence, the clinical practices guidelines suggest that: (1) patients with CKD represent a “high-risk” group when considering pharmacologic therapy to reduce CV events, irrespective of the cause of CKD; and (2) patients with diabetic kidney disease, with or without hypertension, should be treated with an ACE inhibitor or ARB. Implementation of guidelines into practice is necessary to achieve the desired outcomes, yet there are wide gaps between evidence derived from clinical trials and the treatment goals achieved in practice settings. Effective utilization of evidence-based guidelines has been identified as a national high-priority research area, because identifying barriers to implementation bears the promise to reduce the second translational gap.10,11

Studies examining the use of RASi agents in diabetes or CKD are largely derived from administrative or pharmacy-claims databases, which lack patient-level information, and depend on prescriptions filled by a patient as an indicator of RASi usage rather than provider-based information.12,13 There are pitfalls in interpreting the existing literature, because such estimates do not account for primary nonadherence of a patient to fill a prescription, which is known to occur in up to a third of patients.14 We conducted a meticulous assessment of conformity to RASi agent usage in the setting of CKD and diabetes mellitus within a Veterans Affairs (VA) health care system. One of our objectives was to determine the effect of patient-related factors and specialty clinic referrals within a health care system on RASi usage.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References

Patient Population

The study includes a cohort of all patients with CKD and diabetes mellitus seen at an ambulatory care setting at the Cincinnati VA Medical Center during a 2-year period between January 1, 2006, and December 31, 2007. There were 28,295 patients who had at least one creatinine value available during the study period. We excluded 862 patients with only in-patient creatinine values or missing data, yielding 27,433 patients with outpatient creatinine values, 8932 of which had an estimated glomerular filtration rate (GFR) of <60 mL/min/1.73 m2 on at least one occasion. We then excluded 873 patients who had a documented allergy or adverse reaction to a RASi agent. Of the remaining 8059 patients, 2889 patients had diabetes mellitus and were available for analysis.

Definition of Outcomes and Risk Factors

The main outcome measure was RASi usage, defined as at least one outpatient RASi prescription dispensed during the study period. RASi usage included classes of ACE inhibitors, ARBs, and combination pills that include either an ACE inhibitor or ARB with a diuretic. CKD was defined as ≥1 estimation of GFR during the study period of <60 mL/min/1.73 m2 and classified in stages based on Kidney Disease Quality Initiative (K/DOQI) guidelines. The date of first available GFR estimate of <60 mL/min/1.73 m2 was used for the extraction of other laboratory and comorbid information. We examined the association between the following variables and RASi use: demographic (age, sex, race) and comorbid conditions (CKD stage, hypertension, hyperlipidemia, peripheral vascular disease, cardiovascular disease, obesity, viral hepatitis [B and C], and human immunodeficiency virus infection), and laboratory variables including serum creatinine, hemoglobin, albumin, hemoglobin A1c (HbA1c), and total cholesterol. Comorbid conditions, including diabetes mellitus status, were defined by International Classification of Diseases, Ninth Revision, Clinical Modification codes and chosen based on their plausible association with CKD or diabetes. Types of outpatient specialty clinic visits (renal clinic, endocrine clinic, cardiology clinic, and clinical pharmacy clinics) attended by patients during the study period were examined for their association with RASi usage. To examine medication adherence (MA) of patients to the prescribed drug classes of interest, we extracted initial and refill dates of prescriptions for RASi agents. Medication possession ratio (MPR) was calculated, defined as actual treatment days divided by total possible treatment days (truncated for the duration of study period or death). Patients with an MPR of ≥0.8 were considered to have good MA, vs those with MPR <0.80 who were considered to have poor MA. The definitions of MA and methods of calculating MPR were based on prior literature, including our prior publication data.15–17 All of the data were extracted from electronic records from the Veterans Integrated Service Network 10 database. The institutional review board at the University of Cincinnati and Cincinnati VA Research and Development Committee approved the analyses.

Statistical Analysis

Patients who were prescribed a RASi agent were compared with those who were not prescribed a RASi agent for univariate differences in their demographic, comorbid, and laboratory characteristics by chi-square test and t test following tests of normality. Frequency of patients who attended specialty clinics was compared between the groups with and without RASi prescriptions. To examine the effect of the level of renal function on RASi usage, patients were classified according to CKD stages and categorized by 15-mL/min/1.73 m2 increments in their GFR, ranging between 0 and 60, and compared by chi-square test. Similar categories of GFR were used to compare the frequency of patients with or without RASi use, stratified by whether they attended a specialty clinic during the study period. Multivariate logistic regression models were developed where RASi use was the dependent variable, with the following independent predictors including age, sex, race, comorbid conditions, level of GFR, and any specialty clinic visits. In a separate model, logistic regression was used to examine the effect of each specialty clinic visit on RASi usage after adjusting for age and GFR. Since patients could attend more than one clinic, each specific clinic was individually compared with categories of no clinic visits or other specialty clinic visits. Average MPR and proportions of good vs poor MA were calculated for ACE inhibitor and ARB agents. We compared average MPR and frequency of good MA for ACE inhibitor use by specialty clinics by chi-square tests. Model performance was judged by receiver operator characteristic (ROC) curve analysis. Risk estimates were expressed as odds ratios (ORs) and 95% confidence intervals (CIs). Two-tailed P values <.05 were considered significant.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References

Univariate Analysis

The sample of 2889 patients was predominantly comprised of men (98%), with a mean age of 70.4 years (standard deviation [SD], 10.3). Overall, during a 2-year period, 2227 of 2889 patients (77%) with CKD and diabetes received a RASi agent, whereas 662 of 2889 (23%) did not. Table I shows comparison of demographic, comorbid, and laboratory characteristics between those who were prescribed a RASi agent vs those who were not. Female sex, older age, and race were associated with a lower likelihood of being in the RASi-prescribed group. Patients with hypertension, hyperlipidemia, obesity, and heart disease were more likely to be prescribed RASi agents. Patients with RASi usage had higher levels of total cholesterol, HbA1c, and albumin. The mean GFR in the no RASi group was significantly lower compared with that in the RASi group, and RASi agent usage declined with worsening GFR (79% in stage III, 59% in stage IV, and 48% in stage V; P<.0001) (Figure 1).

Table I.   Baseline Demographic and Comorbid Characteristics of Diabetic Patients With CKD by RASi Medication Usage
CharacteristicsOverall (N=2889)No RASi (N=662)RASi Use (N=2227)P Value
  1. Abbreviations: CKD, chronic kidney disease; GFR, estimated glomerular filtration rate; HIV, human immunodeficiency virus; RASi, renin-angiotensin system inhibitor (includes angiotensin-converting enzyme inhibitors and angiotensin receptor blockers). aInformation on race was available in 976 patients. bContinuous variables are expressed as mean and standard deviation.

Sex, %
 Male9896.898.3.02
 Female23.21.7
Race, %a
 Black19.813.221.6.008
 Non-black80.286.878.4
GFR, mL/min/1.73 m2, %
 30–59 (CKD III)9284.694.2<.0001
 15–29 (CKD IV)5.710.14.4
 <15 (CKD V)2.35.31.4
Ageb70.4 (10.3)72.5 (10.2)69.8 (10.2)<.0001
Hypertension, %90.474.295.2<.0001
Hyperlipidemia, %73.860.177.8<.0001
Obesity, %24.114.726.9<.0001
Hepatitis C, %2.11.72.3<.0001
HIV, %0.10.20.1.62
Heart disease, %60.357.961.1.57
Prior antihypertensive treatment, %8078.481.1.0009
Specialty visits, %533757<.0001
Serum creatinine, mg/dL1.7 (1.1)2.0 (1.5)1.7 (1.0)<.0001
GFR, mL/min/1.73 m2)48.2 (11.2)44.7 (13.6)49.2 (10.2)<.0001
Hemoglobin, g (n=1966)13.0 (1.8)12.9 (2.0)13.0 (1.8).79
Albumin, g (n=1005)3.9 (0.6)3.8 (0.6)3.9 (0.6).02
Hemoglobin A1c (n=2087)7.1 (1.6)6.9 (1.3)7.2 (1.6)<.0001
Total cholesterol (n=1878)162.9 (47.6)157.6 (40.6)164.3 (49.2).006
image

Figure 1.  Renin-angiotensin system (RAS) inhibitor usage changes by level of glomerular filtration rate. CKD indicates chronic kidney disease; GFR, estimated glomerular filtration rate. (A) Patients who attended specialty clinics. (B) Patients who did not attend specialty clinics. Patients attending at least one renal, cardiology, endocrine, or pharmacy clinic were considered to have met the subspecialty clinic visit criteria.

Download figure to PowerPoint

Of the specialty clinic visits, 15.6% (n=451) had at least one renal visit, 19.9% (n=574) had at least one cardiology visit, and 8.3% (n=241) had at least one endocrine visit, whereas 31.9% (n=922) had at least one pharmacy clinic visit. Overall, 1365 of 2889 patients did not attend any of the examined specialty clinics during a 2-year period. In patients who did not attend specialty clinics, RASi usage was 70%, compared with 84% of patients who attended at least one of the specialty clinics (P<.0001). Figure 2 shows that the frequency of patients prescribed RASi agents significantly decreased with worsening GFR; however, the proportion of patients prescribed a RASi agent remained consistently higher in patients who attended specialty clinics.

image

Figure 2.  Patterns of renin-angiotensin system (RAS) inhibitor usage based on glomerular filtration rate (GFR), stratified by specialty clinics. (A) Patients who attended specialty clinics. (B) Patients who did not attend specialty clinics. Patients attending at least one renal, cardiology, endocrine, or pharmacy clinic were considered to have met the subspecialty clinic visit criteria.

Download figure to PowerPoint

Multivariate Analysis

Logistic regression model examined the factors associated with RASi use after adjusting for diverse patient characteristics. Patients were 5 times more likely to be taking a RASi agent if they had hypertension (OR, 5.49; 95% CI, 4.16–7.25) and almost twice as likely to be prescribed a RASi agent if they had hyperlipidemia (OR, 1.59; 95% CI, 1.30–1.95), or obesity (OR, 1.44; 95% CI, 1.12–1.86). Older patients (per 10-year increments) had lower odds of taking a RASi agent (OR, 0.85; 95% CI, 0.78–0.93), whereas patients with higher GFRs had greater odds of being prescribed a RASi agent (OR, 1.42; 95% CI, 1.31–1.53). In addition, attending a specialty clinic improved the odds of conformity to RASi use (OR, 1.92; 95% CI, 1.58–2.34). The model performance was good, with an ROC value of 0.72.

Table II indicates the results of separate logistic regression models examining the adjusted odds of RASi use associated with a particular specialty clinic. Among the 4 specialties, the magnitude of impact associated with RASi use was greater in patients who attended a cardiology clinic (OR, 3.52; 95% CI, 2.63–4.71), pharmacy clinic (OR, 3.15; 95% CI, 2.49–3.98), and endocrine clinic (OR, 3.39; 95% CI, 2.22–5.16). Renal visits were 2.04 times (95% CI, 1.54–2.71) more likely to be associated with RASi use when compared with no specialty visits.

Table II.   Effect of Specialty Visits on RASi Use in CKD With Diabetes Mellitus
Clinic VisitsUnadjusted OR (95% CI)Adjusted OR (95% CI)
  1. Abbreviations: CI, confidence interval; CKD, chronic kidney disease; OR, odds ratio; RASi, renin-angiotensin system inhibitor (includes angiotensin-converting enzyme inhibitors and angiotensin receptor blockers). Four separate logistic regression models based on subgroups attending each specialty clinic, each adjusted for age and estimated glomerular filtration rate. aA total of 1365 patients did not attend any of the specialty clinics examined and therefore formed the reference group.

Renal clinic
 Renal clinic (n=451) vs nonea1.44 (1.13–1.85)2.04 (1.54–2.71)
 Other specialties (n=1073) vs none2.90 (2.35–3.58)2.68 (2.16–3.33)
Pharmacy clinics
 Pharmacy clinic (n=922) vs none3.20 (2.55–4.03)3.15 (2.49–3.98)
 Other specialties (n=602) vs none1.52 (1.21–1.90)1.79 (1.41–2.28)
Cardiology clinics
 Cardiology clinic (n=574) vs none3.45 (2.63–4.64)3.52 (2.63–4.71)
 Other specialty (n=950) vs none1.85 (1.52–2.26)2.07 (1.68–2.55)
Endocrine clinic
 Endocrine clinic (n=241) vs none3.20 (2.14–4.80)3.39 (2.22–5.16)
 Other specialties (n=1283) vs none2.15 (1.79–2.59)2.35 (1.94–2.86)

Patient Adherence to RASi Agents

Of the 2227 patients who were prescribed RASi agents, MA data were available on 1948 patients. A total of 1842 patients received an ACE inhibitor and 184 patients received an ARB. The average MPR for ACE inhibitor use was 0.86 (SD, 0.20) and was similar to ARB use (0.86; SD, 0.20). The proportion of patients with good MA was 75.2% for the ACE inhibitor group and 74.4% for the ARB group. Considering that the majority of our sample received ACE inhibitors, and that the overall MPR was similar to those who received ARBs, further analyses were conducted using MPR for ACE inhibitors. Mean MPR was not significantly different when compared across patients who attended any of the 4 specialty clinics vs those who did not (0.86 vs 0.85; P=.51). Results were qualitatively similar when comparing frequencies of patients with good vs poor MA. After examining the effect of each specialty clinic, only pharmacy clinics visits were associated with a significantly better patient adherence (MPR for patients attending pharmacy clinics=0.88; other specialty clinics except pharmacy=0.84; no specialty clinic=0.86, P=.01).

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References

This study shows that within a VA health care system, conformity to the use of RASi agents in patients with CKD and diabetes mellitus was 77%. Among several factors, RASi usage was 5 times more likely in patients with hypertension, whereas it was inversely related to the level of GFR and age. Approximately half of the patients attended relevant specialty clinics, and this was associated with an increased likelihood of conformity to RASi use. These findings highlight the importance of implementation research within a health care system and allow for the development of strategies to improve delivered care to the patients with diabetes and CKD.

Several studies have examined RASi use in patients with diabetes and hypertension. In one such report, Winkelmayer and colleagues12 examined Medicare beneficiaries in a state Pharmaceutical Assistance Contract for the Elderly (PACE) program. Only 50.7% (95% CI, 50.0–51.4) of patients with diabetes and hypertension filled a prescription for an ACE inhibitor or ARB during the quarter studied. In another study (National Health and Nutrition Examination Survey III 1999–2002 cross-sectional survey) that examined 742 diabetics, RASi use (43%) was determined by asking the survey respondents to present the containers of all prescription medications taken in the past month.13 Clearly, the reports, which rely on pharmacy claims for filled prescriptions or patient self-reporting, are unable to distinguish between providers’ conformity to a particular treatment guideline vs patients’ nonadherence to filling a prescription. Up to 30% of patients who are prescribed a medication do not get it filled.14

This is one of the first reports to assess RASi usage by using provider-level prescription information in patients with CKD and diabetes during a 2-year period. Overall, 77% of patients were prescribed a RASi agent. This proportion was higher than a prior VA observation, which reported facility variation in RASi usage across different centers.18 In that study, Tiwari and colleagues examined a period of 365 days after index GFR for an active prescription of RASi, whereas our study examined a period spanning 2 years, which may have captured additional patients who were prescribed these drugs. The RASi prescription rate was higher than prior non–VA studies, and this may represent differences across health care systems. Another explanation may be that the reported “conformity” in prior studies actually reflects a combination of prescribing practices and patient’s primary nonadherence to fill a prescribed drug.

The patient characteristics of interest that affected RASi usage included hypertension, age, and GFR. For example, presence of hypertension was associated with a 5-fold increase in the odds of RASi usage, whereas older patients or those with lower levels of GFR were significantly less likely to be taking RASi agents. It can be speculated that physicians may not use these agents in patients with lower GFR secondary to the risk of worsening renal function or lack of clear recommendations for their use in older patients. We point out that the average age of patients in our sample was 71 years, and the average GFR was 48 mL/min/1.73 m2. This outlines the gap, by historical comparisons, between patients who are typically enrolled in clinical trials vs those who are cared for in practice settings (eg, average age of patients enrolled in the Irbesartan in Diabetic Nephropathy Trial [IDNT] and the Reduction of Endpoints in NIDDM With the Angiotensin II Antagonist Losartan [RENAAL] investigations was 59 and 60 years, respectively, with a maximum age of 74 years).7,8 This underscores the need for future investigations that should assess the impact of implementing these guidelines into practice settings on cardiovascular and renal outcomes.

Overall, 48% of patients attended at least one of the specialty clinics that were examined. Specialty clinic visits had a positive impact on RASi usage in CKD patients with diabetes, independent of other patient characteristics. The effect was similar in the subgroup of patients with new RASi prescriptions. Of all the clinics, pharmacy clinic visits had among the highest impact. In addition, when examining patient adherence to prescribed use, pharmacy clinic visits were also associated with significantly higher adherence. In contrast, the relative magnitude of impact of renal clinic on conformity was lowest; however, this observation is not totally unexpected, since patient referrals to renal clinics typically occur in more advanced stages of kidney disease, by which time most patients are either already taking a RASi agent or the advanced kidney disease may limit new initiation of a RASi agent. Taken together, our observations suggest that both patient-related factors and early involvement of relevant specialty clinics may lead to further improvement in RASi use, when indicated. Specifically, it outlines a potentially important role that clinical pharmacists can play within a health care system in improving conformity to evidence-based guidelines as well as patient adherence.

Limitations

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References

We examined a single-center cohort of predominantly male patients. The cohort, however, represents one of the largest VA health care systems in Ohio and used a large sample size to conduct an adequately powered analysis, even after accounting for differences in multiple patient characteristics. Most importantly, the VA health care system provides a unique advantage of extracting electronic prescription information at a provider level, during a 2-year period, to determine conformity more accurately than previously published literature. Conformity to RASi usage was examined in a CKD cohort with diabetes, and hypertension was not used as a comorbidity to define the cohort. The rationale was that there are proven benefits to RASi treatment in CKD associated with diabetes with or without hypertension and, as such, 90% of our cohort had a diagnosis of hypertension. Another limitation to our analysis was that accurate information on proteinuria was not available for extraction. Hence, we could not determine the independent effect of proteinuria on RASi use in CKD patients with diabetes or define earlier stages of CKD.

Conclusions

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References

In a VA cohort of CKD patients with diabetes, the conformity to RASi usage was 77%. Diagnosis of hypertension was associated with higher odds of RASi usage, but older age and lower GFR were associated with a lower likelihood of RASi prescription. Specialty clinic visits improved conformity to RASi usage. These observations suggest that targeting certain patient characteristics and appropriate specialty clinic referrals may further improve implementation of evidence-based guidelines into practice in this high-risk population. Future studies are needed to examine the effect of adherence to guidelines on cardiovascular and renal outcomes. In addition, prospective implementation of processes of care to improve adherence to guidelines needs to be studied across different health care systems.

Acknowledgment and disclosure:  The study received support from a Federal Services Research Grant from the American Society of Health System Pharmacists (ASHP) Foundation. The authors do not report any conflicts of interest.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Limitations
  7. Conclusions
  8. References