Changing Patterns of Initial Drug Therapy for the Treatment of Hypertension in a Medicaid Population, 2001–2005

Abstract

Thiazide diuretics have been recommended as one preferred choice for the initial treatment of hypertension. This study was undertaken to determine whether Maine physicians initiating monotherapy for newly diagnosed hypertensive patients from 2001–2005 used this guideline. The Maine Medicaid database was searched for the drug classes used to initiate monotherapy for patients followed for at least 6 months. A total of 5373 patients were included. In 2001, the use of β‐blockers was 23.5%, diuretics 17.5%, angiotensin‐converting enzyme inhibitors 37.5%, calcium channel blockers 9.5%, angiotensin receptor blockers 3.8%, and others 8.2%. By 2005, the use of β‐blockers was 27.8%, diuretics 25.5%, angiotensin‐converting enzyme inhibitors 30.9%, calcium channel blockers 6.4%, angiotensin receptor blockers 1.6%, and others 7.7%. There was an increase in the use of angiotensin‐converting enzyme inhibitors or angiotensin receptor blockers in diabetics but no other condition affected drug choice. Although there was an increase in the use of diuretics as initial therapy in 2003 and 2004, this decreased in 2005. The increase in initial diuretic use was not reflected in patterns of ongoing antihy per tensive use from 1997 to 2005. There appears to have been limited impact from the guidelines on initial drug choice and even less so on ongoing drug therapy.

Although hypertension affects more than 50 million people in the United States, ¹ only 34% of these have been treated to a guideline‐defined target of 140/90 mm Hg. ² Widely published and distributed guidelines have suggested algorithms to initiate therapy. In 1997 the Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC VI) ³ suggested the use of thiazide diuretics followed by β‐blockers (BBs) as initial therapy. This followed a period throughout the 1980s where the use of diuretics had declined. ⁴

It was hoped that physician behavior would be modified by the adoption of these guidelines. Physician adherence to initial therapy prescribing guidelines before and after JNC VI for the state of Maine's Medicaid population was examined in 2002. ⁵ Treatment of 6246 patients during the period 1997–2000 showed little change in the use of the various classes of drugs to initiate therapy. In 1997, use of BBs was 22.6%; diuretics, 23.9%; angiotensin‐converting enzyme inhibitors (ACEIs), 22.9%; and calcium channel blockers (CCBs), 18.3%. By 2000, BBs were used in 27.2%; diuretics, 25%; ACEIs, 26.5%; and CCBs, 10.9%. Fewer than half of the patients were started on evidence‐based initial therapy of diuretics or BBs during this period. There was no correlation between the drug chosen and comorbidities. The only change over this period was the decrease in the use of CCBs, which may have followed a change in marketing patterns by the CCB manufacturers. This has been noted previously. ⁶ Physician behavior may also have been modified by the ongoing misperception that diuretics produced clinically significant adverse metabolic effects on glucose and potassium. ⁷

Based on this apparent lack of influence of JNC VI, the Drug Utilization Review (DUR) Committee of Maine Medicaid began an educational process in early 2002 to influence physician behavior. Mailings restating the guidelines were sent to all physicians in the state on two separate occasions. The Preferred Drug List was annotated to suggest the use of diuretics as initial therapy, to suggest the use of once‐daily BBs next, and to not allow initial therapy with angiotensin receptor blockers (ARBs) unless a written prior authorization documented that the patient was allergic to ACEIs or had type 2 diabetes. During this process, the Antihypertensive and Lipid‐Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) ⁸ was published and widely disseminated to local physicians by the DUR Committee. This study again reinforced the benefits of diuretics by showing no overall difference among the 3 arms of treatment with diuretics, ACEIs, and CCBs on the primary end point of cardiovascular events. Shortly thereafter, the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7) ² was released, which yet again reiterated the appropriateness of diuretics as initial therapy. “Diuretics should be used as initial therapy for most patients with hypertension.” ²

The new guideline, local educational efforts, and some formulary restrictions suggested a review of initial drug utilization for hypertensive patients newly diagnosed in the period 2001–2005 and overall antihypertensive drug utilization from 2001 to 2005 to determine whether physician behavior had been modified.

METHODS

Historical Maine Medicaid pharmacy claims data were reviewed by searching for National Drug Codes (NDC) for all known antihypertensive agents and for patients who had received only one antihypertensive agent. The claims data were reviewed using Clinical Classification Software Diagnostic Groups, as developed by the Agency for Healthcare Research and Quality (AHRQ) (Rockville, MD) to include only patients with a diagnosis of hypertension. The patients were then further categorized to identify patients in whom hypertension had been diagnosed for the first time. Records for each patient were reviewed retrospectively to confirm that hypertension was a new diagnosis for that patient. The data were then analyzed as to the continuation of the initial drug or its replacement during the 6 months after beginning therapy. This event was reported in the year in which the 6‐month review occurred, since the initiating event sometimes occurred in the previous year. Each patient was counted only once. The patients were then divided into subgroups based on prevalent comorbid conditions that might have affected the selection of the antihypertensive agent. These were uncomplicated hypertension, hypertension with congestive heart failure (CHF), hypertension with chronic obstructive pulmonary disease, hypertension with hyperlipidemia, and hypertension with diabetes. Data from January 1, 2001 through December 31, 2005 were reviewed and also compared with data previously collected from January 2,1997 through December 30, 2000, as previously reported. ⁵ All patients with pregnancy‐related hypertension were excluded from this study. Claims data were also used to ensure that a diagnosis of hypertension was not made or antihypertensive drug therapy initiated during the previous 60 days. All eligible patients were required to have undergone at least 6 months of drug therapy.

Based on guidelines provided by the US Food and Drug Administration, ⁹ institutional review board approval was not required for the use of Medicaid data without links to patients.

The chi‐square test was used for statistical purposes.

RESULTS

During the period 2001–2005, 5373 patients with a new diagnosis of hypertension were identified who were given monotherapy and were followed for at least 6 months. Most of these had uncomplicated hypertension (n=4357; 84.4%). Of patients with comorbidities, 53 (0.9%) had CHF, 330 (6.1%) had chronic obstructive pulmonary disease, 417 (7.8%) had hyperlipidemia, and 182 (3.4%) had type 2 diabetes.

In 2001, the use of BBs was 23.5%; diuretics, 17.5%; ACEIs, 37.5%; CCBs, 9.5%; ARBs, 3.8%; and others, 8.2% (Figure 1). In 2005, the use of BBs was 27.8%; diuretics, 25.5%; ACEIs, 30.9%; CCBs, 6.4%; ARBs, 1.6%; and others, 7.7%. There was a marked increase in the use of diuretics coupled with a significant decrease in the use of ACEIs, CCBs, and ARBs. Diuretics were the only class where percentage use increased, although use in 2005 was actually less than in 2003 and 2004. The decrease in CCBs continued from 1997 (18.3%) through 2000 (10.9%). The use of ARBs dropped between 2003 (4.8%) and 2004 (1.6%), presumably due to the mandated prior authorization policy that was initiated after the utilization of this class had risen markedly between 2002 (3.7%) and 2003 (4.8%).

Figure 1.

Hypertensive drug therapy based on monotherapy in patients with initial diagnosis of hypertension. ACE indicates angiotensin‐converting enzyme inhibitor; Beta, β‐blocker; CCB, calcium channel blocker; and ARB, angiotensin receptor blocker.

The potential influence of sex and/or age on drug utilization was examined by the chi‐square test (Table I). No statistically significant differences were found, but there was a trend toward more diuretics being prescribed in women 45 years of age or older than in younger women or in men of any age. This trend was seen throughout the 4‐year period of the study. There were no other differences by age or sex.

Table I.

Age and Sex Distribution of Initial Antihypertensive Therapy, 2001 and 2005

Initial Therapy/Age Group, y	Women	2001 Men	Total	Women	2005 Men	Total
β‐Blockers
0–44	9.1	11.6	10.2	9.1	13.6	11.0
≥45	13.8	12.7	13.3	19.0	15.4	17.4
Total	22.8	24.3	23.5	28.1	29.0	28.5
Diuretics
0–44	8.1	3.6	6.0	9.5	10.8	10.0
≥45	15.4	6.8	11.5	19.6	10.8	15.8
Total	23.5	10.4	17.5	29.1	21.5	25.8
ACE inhibitors
0–44	12.1	19.5	15.5	9.5	16.9	12.7
≥45	19.5	25.1	22.0	19.2	17.9	18.7
Total	31.5	44.6	37.5	28.7	34.9	31.3
Calcium blockers
0–44	2.3	4.0	3.1	2.1	2.1	2.1
≥45	8.4	4.0	6.4	5.4	3.1	4.4
Total	10.7	8.0	9.5	7.6	5.1	6.5
Other
0–44	3.4	6.0	4.6	3.1	4.9	3.9
≥45	8.1	6.8	7.5	3.5	4.6	4.0
Total	11.4	12.7	12.0	6.6	9.5	7.8
Grand total	100.0	100.0	100.0	100.0	100.0	100.0
Values are given as % of patients taking each class of drugs. No statistically significant differences were found by the chi‐square test. ACE indicates angiotensin‐converting enzyme.

The presence of comorbidities appeared to have some influence on the initial choice of antihypertensive agent (Table II). In patients with diabetes, 57% received an ACEI as first choice. When an ARB was included, it rose to 64%; however, 36% of hypertensive diabetics received neither an ACEI nor an ARB. There did not appear to be much difference in the other comorbid groups. In patients with chronic obstructive pulmonary disease, 27% still received a BB. In patients with CHF, 6% still received monotherapy with CCB while 40% received BBs, 23% ACEIs, and 28% diuretics. There were no between‐year or group differences.

Table II.

Initial Therapy for Hypertensive Patients by Specific Comorbidity Categories

Period/	CHF		COPD		Diabetes		Total
Medication Class	n	%	n	%	n	%	n	%
2001–2005
β‐Blockers	21	39.6	89	27.0	31	17.0	132	24.9
Diuretics	15	28.3	92	27.9	18	9.9	119	22.5
ACE inhibitors	12	22.6	89	27.0	104	57.1	187	35.3
ARBs	1	1.9	10	3.0	12	6.6	34	6.4
Calcium blockers	3	5.7	29	8.8	7	3.8	44	8.3
Other	1	1.9	21	6.4	10	5.5	14	2.6
Total	53	100.0	330	100.0	182	100.0	530	100.0
2001
β‐Blockers	3	27.3	8	21.1	1	6.3	11	18.0
Diuretics	6	54.5	13	34.2	2	12.5	19	31.1
ACE inhibitors	2	18.2	11	28.9	11	68.8	23	37.7
ARBs	0	0.0	0	0.0	0	0.0	0	0.0
Calcium blockers	0	0.0	4	10.5	1	6.3	5	8.2
Other	0	0.0	2	5.3	1	6.3	3	4.9
Total	11	100.0	38	100.0	16	100.0	61	100.0
2005
β‐Blockers	7	43.8	30	31.6	11	15.9	43	25.4
Diuretics	6	37.5	22	23.2	9	13.0	35	20.7
ACE inhibitors	1	6.3	28	29.5	39	56.5	64	37.9
ARBs	1	6.3	2	2.1	3	4.3	6	3.6
Calcium blockers	0	0.0	4	4.2	4	5.8	8	4.7
Other	1	6.3	9	9.5	3	4.3	13	7.7
Total	16	100.0	95	100.0	69	100.0	169	100.0
CHF indicates congestive heart failure; COPD, chronic obstructive pulmonary disease; ACE, angiotensin‐converting enzyme; and ARB, angiotensin receptor blocker.

The continued use of initial therapy was reviewed after 6 months (Table 3). Only 67.1% of all patients remained solely on the initial therapy. This was the same across all comorbidities except CHF, where it was only 45.3%. Some patients were maintained on the initial drug but had other agents added (12.9%). This was similar across all groups except with CHF, where it was 37.8%. Patients were switched to a new monotherapy in 17.3% and were switched off the initial agent to two new agents in 3.2%. The changes or addition of another drug showed no trend or statistically significant difference based on what class of drug was initially used. Patients started on BBs were predominantly changed to diuretics (7.9%) or ACEIs (24.29%). Those on diuretics were predominantly changed to BBs (25.38%) or ACEIs (30.9%). Those on ACEIs were changed to BBs (24.8%), diuretics (23.45%), or others (almost always an ARB) (44.83%). No pattern could be determined.

Table III.

Six‐Month Follow‐Up of Initial Drug Class Use Based on Comorbidities in Hypertensive Patients

			Single Rx				Multiple Rx
	Aggregate		Same Therapy		Different Therapy		Same Therapy		Different Therapy
Comorbidity	n	%	n	%	n	%	n	%	n	%
None	4537	84.4	3072	67.7	749	16.5	585	12.9	131	2.9
CHF	53	1.0	24	45.3	9	17.0	17	32.1	3	5.7
COPD	330	6.1	196	59.4	68	20.6	53	16.1	13	3.9
Diabetes	182	3.4	128	70.3	19	10.4	29	15.9	6	3.3
Hyperlipidemia	417	7.8	266	63.8	73	17.5	67	16.1	11	2.6
Summary	5373	102.7	3603	67.1	918	17.1	751	14.0	164	3.1
Rx indicates drug therapy; CHF, congestive heart failure; and COPD, chronic obstructive pulmonary disease.

These data from the period 2001–2005 were then compared with the previously collected data from the period 1997–2000. BB use trended up throughout this 9‐year period. It is now 29% higher than it was in 1997. ACEI use remained at approximately 20% in each of the 9 years. CCB use has fallen every year, from a high of 18.9% to 10.8% (a 42.9% decrease). ARB use rose markedly each year from 1997 to 2003 and then fell 22%, as might be expected in connection with the state's authorization policy, over 2004–2005. Use of diuretics, the therapy most often recommended by guidelines, decreased every year from 1997 to 2004. It was only in 2005 when there was a slight trend upward. Diuretic use in 2005 was still below that in 1997 (28.0% vs 30.6%).

In addition to reviewing the use of initial therapy in newly diagnosed hypertensive patients, we reviewed the overall treatment of hypertension during 2001–2005 (Figure 2). Diuretic use trended downward throughout the period, with a small rise only in the last 12 months was considered in the study. BB and ACEI use was unchanged overall. CCB use decreased yearly and was 13% lower in 2005 than in 2001. ARB use climbed 36% from 2001 to 2003 and then decreased 22%, presumably reflecting the initiation of a rigorous prior authorization program. This was also seen at the same time in overall ARB use in the Maine Medicaid population.

Figure 2.

Drug use by class among all patients with hypertension. Beta indicates β‐blocker; ACE, angiotensin‐converting enzyme inhibitor; CCB, calcium channel blocker; and ARB, angiotensin receptor blocker.

DISCUSSION

JNC 7 ² recommended thiazide‐type diuretics as the preferred initial agent to treat hypertension in most patients. Based on meta‐analyses, the National Institute for Clinical Excellence (NICE) ¹⁰ also issued guidelines in 2004 that suggested a thiazide diuretic as initial therapy. In this review, thiazides were shown to decrease the incidence of stroke by 31%, fatal and nonfatal myocardial infarction by 22%, and all‐cause mortality by 9%. BBs decreased stroke by 19% but did not significantly decrease the incidence of death or myocardial infarction. The ALLHAT study ⁸ of 40,000 patients showed no differences in the primary cardiac outcome or mortality among thiazide‐type diuretics and an ACEI or CCB. Meta‐analysis has been used for comparisons between diuretics and other agents. In 1997, a review of health outcomes associated with antihypertensive therapies used as initial agents ¹¹ found no difference in stroke and CHF between diuretics and BBs, with only diuretics reducing coronary disease. The same group repeated this review in 2003. ¹² None of the available treatments (BB, CCB, ACEI, ARB) were demonstrated to be better than low‐dose diuretics for any outcome when 42 trials encompassing 192,478 patients were reviewed. BBs, also recommended by JNC 7 to be used early for the treatment of new hypertensives, were found to be inferior to diuretics for all outcomes, leading to a recommendation by the JNC 7 that they be used as alternate agents. Another meta‐analysis in 2005 suggested that BBs reduced the incidence of stroke by about half that expected from other hypertension trials using other agents. ¹³ The Cardiovascular Health Study ¹⁴ reviewed the use of antihypertensive medications in 5888 older adults over 10 years (1989–1999). As in the present study, the use of diuretics fell steadily over the 10‐year period. BB use was stable, with an increased use of ACEIs and CCBs.

After the publication of ALLHAT, there appeared to be an increase in the use of thiazide‐type diuretics. ¹⁵ This was also reported by a large managed care organization. ¹⁶ In this study, over 9 years there was a large increase in the use of diuretics to initiate antihypertensive therapy (from 17.5% to 30.3%). This suggested some attention to guidelines and national and local education. The continued drop‐off in the use of CCBs could also be due to the guidelines, although there is no way to be certain that changes in marketing and sampling did not contribute.

A major limitation of this study was that it reviewed drug usage in the Maine Medicaid population only. It is possible that a review of global prescriptions throughout the state or in other locations would give different results. The preferred drug list did not affect the use or order of antihypertensive drug class, but only individual branded agents. In the context of the treatment of hypertension, the prior authorization program limited the use only of ARBs.

CONCLUSIONS

The findings of this study are consistent with those of previous studies suggesting a relative lack of adherence to published and well‐promulgated guidelines. JNC VI, JNC 7, and the ALLHAT study have only slightly influenced initial diuretic use for the newly diagnosed hypertensive patient in the Maine Medicaid program. Although there was some increase in the use of diuretics during the study period, use was still lower than in 1997. The use of diuretics in the ongoing treatment of established hypertensives has not been greatly influenced. The presence of important comorbidities such as diabetes has influenced the choice of antihypertensive agent—but in only 66% of such patients. Other comorbidities did not appear to affect drug choice. The DUR Committee of Maine Medicaid will again use this analysis to attempt direct physician education and to create a tiered system of prior authorizations to control antihypertensive drug use.