2010 Drug Utilization and Cost Trends for Antidiabetic Agents

, , ,
The American Journal of Pharmacy Benefits, January/February 2011, Volume 3, Issue 1

The antihyperglycemic therapeutic class will continue to be a major contributor to prescription drug trends in 2011 and beyond.

In 2009, the treatment of diabetes was one of the top contributors to prescription drug trend. With the continued aging of the Baby Boomers and obesity in the United States, this trend will continue beyond 2011. The diabetes population in the United States will almost double over the next 25 years, and annual medical spending on the disease i

s projected to hit $336 billion

(up from $113 billion today), according to a recently

published study.1 According to the forecast, the number

of Americans living with diabetes will rise from 23.7

million in 2009 to 44.1 million in 2034. For the Medicare

program, the projected increases over the next 25

years are even more dramatic: the number of Americans

living with diabetes and covered by Medicare will rise

from 6.5 million to 14.1 million, and Medicare spending

on diabetes will almost quadruple, skyrocketing from

$45 billion this year to $171 billion in 2034.1 Diabetes

disproportionately affects minorities and low-income

patients.2 Greater numbers of patients with diabetes

will likely be covered under Medicaid and individual

exchange programs due to mandated changes of the

recent healthcare reform bill.3

BACKGROUND

Diabetes is a disease characterized by abnormally high glucose levels in the blood, which are caused by the body’s inability to efficiently convert glucose into energy. The metabolic disorder develops because of either diminished production of the hormone insulin or insulin resistance. There are 3 major types of diabetes: type 1, type 2, and gestational. Type 1 is an autoimmune disease that causes insulin deficiency. Type 2 occurs when either the pancreas fails to produce enough insulin or the body becomes insensitive to insulin. Gestational diabetes occurs during pregnancy. Rarely, diabetes is due to genetic defects or pancreatic diseases, or is drug/chemical induced. Type 2 diabetes (T2DM) is the most common form, representing more than 90% of cases affecting both adults and children. In the United States, 23.6 million people (8.0% of the population, or 1 in every 13 Americans) have diabetes.4 The growing population of persons with prediabetes or impaired fasting glucose is a major concern. It was estimated that 57 million American adults 20 years or older had impaired fasting glucose in 2007.5

Patients newly diagnosed with diabetes have several therapy options to control their blood sugar. Lifestyle (dietary and exercise) changes are essential for all diabetic patients, although the impact of these may not be sufficient to delay the initiation of pharmacotherapy. The spectrum includes multiple classes of oral and injectable drugs in addition to insulin treatment.6 Each of the drugs has several indications for glycemic control in patients, and many are used in conjunction with insulin. The noninsulin injectables Symlin and Byetta commonly are used as adjunctive therapy with oral antidiabetic drugs to improve glycemic control (

Table 1

).

As a baseline objective of therapy, a recent (2009) consensus statement published by the American Association of Clinical Endocrinologists/American College of Endocrinology on glycemic control in diabetes mellitus gives as the stated goal the achievement of a glycosylated hemoglobin (A1C) level of 6.5% or less, with recognition of the need for individualization to minimize the risks of hypoglycemia (low blood sugar).6 The A1C targetof 6.5% may vary based on an individual’s age, useof other medications, number of comorbid conditions,duration of diabetes, history of hypoglycemia, and otherfactors. The organization has developed an algorithmfor a therapeutic approach to glycemic control basedon the A1C level, which can be broadly summarized asfollows:

• When the A1C level is 7.5% or lower, it may be possible to achieve a level of 6.5% with the use of monotherapy.

• If the A1C level falls to between 7.6% and 9.0%, treatment should begin with a dual therapy.

• Where the A1C level is in excess of 9.0%, triple therapy is appropriate.

• When triple drug therapy fails or is contraindicated, it is appropriate to initiate insulin treatment, either with or without additional orally administered agents.

• Other options within this framework include titrating medication dosages and changing regimens (adding or discontinuing medications); when combination therapy is prescribed, it is important to use drug classes that have complementary mechanisms of action.6

Increasingly, there are multiple and complex regimen options for T2DM that consider postprandial hyperglycemia, insulin resistance, and weight changes. Updates to prescribing guidelines, research findings, and the spectrum of A1C levels of T2DM patients at diagnosis will drive increasing use of double and triple therapy as initial therapy.

In recent years, the thiazolidinediones (TZDs) Actos (pioglitazone) and Avandia (rosiglitazone), also referred to as insulin-sensitizing agents, were used as both monotherapy and adjunctively with other antidiabetic drugs. Utilization of TZDs peaked in 2007 when the class was ranked 12th by gross cost in the overall CVS Caremark Book of Business. Their class ranking has fallen dramatically, to 17th overall in August 2010, as a result of several published reports questioning their safety, beginning in May 2007.

On September 23, 2010, the Food and Drug Administration (FDA) said that it would significantly restrict the use of Avandia to patients with T2DM who are unable to control the disease with other drugs. The new restrictions are in response to data suggesting an elevated risk of cardiovascular events such as heart attack and stroke in patients treated with Avandia. The FDA also stated it will require a restricted access program for Avandia under which it will be available to new patients only if they are unable to achieve glucose control on other medications and are unable to take Actos (the only other currently marketed TZD). Current Avandia users benefiting from the drug will be able to continue using the medication if they so choose. Doctors will have to attest to and document their patients’ eligibility; patients will have to review statements describing the cardiovascular safety concerns associated with this drug and acknowledge that they understand the risks. The FDA anticipates that these requirements will significantly limit use of Avandia.7

METHODOLOGY

The analysis is a retrospective study of 23.2 million CVS Caremark members who had more than 240 million prescription claims. The population identified for the study was a representative sample of members across a variety of plan sponsors including health plans, managed care organizations, Medicaid, unions, national and local employers, and government agencies (excluding Medicare) located throughout the United States. Only clients with stable membership (limited to a ±15% change in eligibility over 24 months) that had prescription claims in the entire 24-month study period were included. The CVS Caremark computerized, de-identified database was used to study the utilization of the following antidiabetic drug classes: alpha-glucosidase inhibitors, amylin analogs, biguanides, dipeptidyl peptidase-4 (DPP-4) inhibitors, glucagon-like peptide-1 receptor agonists, meglitinide analogs, sulfonylureas, TZDs, and insulin. The time frame was January 1, 2006, through August 31, 2010. Utilization was based on the change in the number of unique utilizers, average days of supply, and total prescriptions compared year over year (YOY). All metrics increased—including the number of utilizers, total number of scripts, and the volume based on total days of supply—in 2010 compared with 2007.

GENERATIONAL CHARACTERISTICS

In addition, a continuously eligible cohort of 6.5 million CVS Caremark members who had an active prescription drug benefit from 2006 through August 2010 was identified (the CE cohort). The members were then categorized by generation in recognition of the fact that these generations share certain common characteristics and have been shaped in a lasting way by the influences of their youth and early adulthood.8,9 The generational divisions identified for the study include the GI Generation (born 1900-1925); the Silent Generation (born 1926- 1945); Baby Boomers (born 1946-1964); Generation X (born 1965-1978); Generation Y (born 1979-1990); and Generation Z (born 1991-present). Table 2 shows how the generations are represented in the CE cohort. The majority of CE cohort members were Baby Boomers, who comprised approximately one-third of the population and represented nearly half (42.8%) of the gross cost. Generation Z was the second largest population but represented just 5.6% of the gross cost.

Gross cost included discounts, member and plan sponsor contributions. Manufacturer rebates were excluded. Gross cost per day was determined by the total gross cost divided by the total days of supply for the time frame. The gross cost per member per month (PMPM) was determined by the total gross spending on all drug therapies divided by the total number of member months for eligible plan participants (members). Utilization was based on days of supply and the number of unadjusted prescriptions PMPM.

Generic dispensing for glycemic control primarily includes prescription claims for several generic oral antidiabetic agents—metformin, glipizide, glyburide, and glimepiride—which have been available as generics since February 1995, August 1995, January 2002, and October 2005, respectively.10-12

FINDINGS

Cost

Table 2

provides generational detail for the CE cohort, with a comparison of PMPM gross spending in 2007 versus 2010 by both non-diabetic and diabetic members. The overall gross cost PMPM in 2010 was $100.34 for a continuously enrolled member, compared with $380.95 for a continuously enrolled diabetic utilizer. The Silent Generation, members who were born between 1926 and 1945, had the highest cost at $422.55, compared with Generation X members, whose PMPM cost in 2010 was $256.53.

Single-source price increases acted as the primary driver of trend as the gross cost per day increased 2.2%. For the period, brand-name drugs had an 8.4% increase in gross costs PMPM, while generics moderated costs with a 4.3% decrease in gross costs PMPM. Januvia was the primary brand contributor to increased gross costs, and metformin was the primary generic cost moderator. Additionally, declining Avandia prescriptions helped offset growing class costs, while Actos price inflation outpaced its own decline in prescription volume, contributing to higher class costs.

Table 3

compares the cost per day YTD June 2009 versus YTD June 2010 for the top 5 oral diabetic agents. The generics metformin and glimepiride had a decrease in cost per day of nearly 14% and 18%, respectively. The single-source brand product Actos had a 7.7% increase and Januvia had a 4.2% increase in gross cost per day.

Drug Mix

In August 2010, the generic dispensing rate for the combined diabetes class increased by 1.1% to 71.2%, driven by increased metformin (both immediate-release and extended-release) claims. The single-source brand dispensing rate declined 0.9%, driven downward by Avandia, Actos, and Avandamet claims but offset by increased volumes of Januvia and Janumet. Little change in the generic dispensing rate is expected until a generic form of Actos reaches the market.

Market Share

The oral antidiabetic market is highly fragmented from the perspective of the number of currently marketed therapies; however, these are dominated by 2 classes: biguanides and sulfonylureas. These 2 classes combine to account for approximately two-thirds of the market by share of claims (

Figure

). The highest volume oral antidiabetic prescription is generic metformin, which continues to steadily gain market share YOY.

The TZDs Actos and Avandia continued to lose market share YOY at -0.5% and -0.6%, respectively, as the continued fallout from safety concerns have affected class sales, those of Avandia in particular, since 2007.13

GENERATIONAL TRENDS

Prevalence of diabetes varies among the generations.

Table 4

summarizes the results of the CE cohort analysis that compared the total population by generation, existing members on diabeticagents, and new starts in 2007versus 2010. Overall, the totalnumber of diabetic patients byage band increased for all 6generational groups between 2007 and 2010. In 2007, 4.96%of the total CE population wasutilizing glycemic-loweringagents compared with 6.27%in 2010.

In 2007, roughly 5.0% of the CE cohort was on insulin or an oral antidiabetic therapy; the largest generational segment in this group was the Baby Boomers at approximately 44.0%, followed by the Silent Generation at 40.1%. By 2010, the percentage of the total CE cohort population receiving antidiabetic drug therapy had grown to 6.27%. The Silent, Baby Boomer, and GI generations had a greater percentage of their populations on insulin, an oral antidiabetic, or a combination of insulin and oral medications than other generational segments. Among those receiving antidiabetic therapy, 9 out of 10 (89.9%) were older than 45 years, 46.7% were Baby Boomers, 37.9% belonged to the Silent Generation, and 5.4% were members of the GI Generation.

NEW DIABETIC PATIENT TREATMENT TRENDS

New patients requiring glycemic therapy have multiple options available. The majority of treatment-naïve patients using agents for glycemic control are initially started on monotherapy. Table 4 compares the percentages of patients started on monotherapy, dual therapy, triple therapy, or 4+ therapies in 2007 versus 2010.

Table 4 also shows the most common diabetic therapy for newly treated patients on monotherapy, dual therapy, and triple therapy. In 2010, nearly 69% of the newly treated patients on monotherapy were on a biguanide (metformin) compared with 60% in 2007. In 2010, 10.17% of the new diabetic monotherapy was a sulfonylurea compared with 11.95% in 2007. In 2007, the insulinsensitizing agents were used in more than 7% of the new diabetic patients on monotherapy compared with 2.68% in 2010. In 2010, more than 77% of the newly treated patients on dual therapy were on a biguanide compared with nearly 71% in 2007. In 2010, 11.36% were on insulin and a biguanide, which was a 4.7% increase from 2007. The vast majority of triple-therapy utilizers were on a sulfonylurea, a biguanide, and a third oral agent. In 2010 the proportion of patients on insulin and 2 other agents was 24.2% compared with 17.4% in 2007. As in monotherapy and dual therapy, the overall use of TZDs in triple therapy has decreased.

THIAZOLIDINEDIONES (INSULIN-SENSITIZING AGENTS)

On May 21, 2007, a meta-analysis of Avandia conducted by Steve Nissen and Kathy Wolski of the Cleveland Clinic was published in the New England Journal of Medicine, showing a statistically significant increase in the risk of myocardial infarction and an increased risk of death from cardiovascular causes with borderline significance.14 Although the meta-analysis did not include Actos—the only other drug in this class—both drugs received a black box warning.15 GlaxoSmithKline subsequently acknowledged the impact of the published analysis in disclosing that sales of the entire Avandia franchise plummeted in thewake its release.16

Both Avandia and Actos had significant declines in utilization by diabetic patients beginning therapy: -86.3% and -35.1%, respectively. Among existing diabetic patients, the number of Avandia utilizers declined significantly (-77.0%), while those who used Actos increased (5.7%). By the end of August 2010, only 230 members remained on Avandia compared with 1675 members in 2007, an 86.3% reduction. Though use of Actos increased 5.7% in 2010 compared with 2007, fewer new diabetic patients were prescribed an insulin-sensitizing agent, whether Avandia or Actos due to Avandia users moving to Actos.

ANALYSIS AND DISCUSSION

Both Avandia and Actos have had a decrease in utilization since 2007. With the recent FDA ruling and the recent report that stated that both Avandia and Actos may increase a patient’s risk of bone fractures by 43%, the decline in utilization will continue.17 Avandia has been linked to fracture concerns for 2 years, but the new study found that patients using Actos are equally at risk. Women who use either of these 2 diabetes drugs are twice as likely to suffer fractures as men, the research showed. Separately, a study by GlaxoSmithKline also demonstrated an increase in bone fractures with Avandia use, particularly among women, and Takeda said it too has found an increased fracture risk for Actos in middleaged women. It has been suggested that the fracture issue, perhaps more than concerns about cardiovascular events, could impact how doctors use these drugs in the future.17

More new diabetic patients were identified and treated in 2010 compared with 2007. This increase was expected as the population aged and obesity continued to be aproblem in the United States, and as public health effortsto increase awareness of diabetes, hypertension, and lipiddisorders succeeded. Interestingly, new patients withsymptoms related to diabetes generally had a minimalannual cost associated with treatment, mainly due to firstlinetreatment using generic metformin. The percentageof new diabetic patients being placed on monotherapyincreased in 2010 compared with 2007. Use of metforminas monotherapy continued to increase. There has been anincrease in the number of patients receiving monotherapyinitially compared with those members who begintreatment with a combination of antidiabetic drugs. Theincrease of new patients on monotherapy in 2010 couldbe attributable to a greater awareness of the need forglycemic control in prediabetic patients and to concernsabout multiple therapies that included insulin-sensitizingagents (Table 4). There also remain well-recognized challengesof convincing new patients that they need multipledrugs at the initiation of therapy. Unfortunately, manyof these factors contribute to the fact that a majority ofpatients on monotherapy, even when optimally adherent,are not “at goal” and are not experiencing the full benefitof pharmacologic therapy.

New dual-therapy patients were most likely to use a biguanide and a sulfonylurea, both available as inexpensive generics (Table 4). However, cost grew exponentially as the intensity of the therapy increased. In 2010, 36.8% of new triple-therapy utilizers combined a sulfonylurea, a biguanide, and either a DPP-4 inhibitor (Januvia) or insulin. The addition of a DPP-4 inhibitor or insulin to a drug regimen added approximately $2000 a year in gross costs compared with most dual-therapy regimens. Between 2007 and 2010, the number of new triple-therapy diabetic patients using a biguanide, a sulfonylurea, and Januvia decreased from 25.8% to 17.9%, while the number of new triple-therapy diabetic patients who used insulin instead of Januvia increased from 11.2% to 18.9%.

The challenge for plans that want to manage costs of medication therapy yet also manage the medical costs of poorly treated diabetes is the judicious use of utilization review (prior authorization and step therapy) for expensive branded therapy. Clinicians, especially diabetes specialists, trying to aggressively manage patients with T2DM increasingly comment that they encounter challenges regarding access to drugs used in triple-therapy combinations. Plan sponsors and insurers will need to continue to be creative in balancing cost control and access to therapy beyond the basic generic metformin and sulfonylureas.

The analysis found that those patients whose diabetes was most severe and who were using 4 or more therapies were most likely to add insulin, which could further increase the annual gross cost by $2000. Insulin utilization also increased in dual and triple therapy, which may reflect more comfort with the use of injectables.

CONCLUSION

The long-term impact on utilization of a black box safety warning imposed by the FDA varied from product to product and had to do with the severity of the warning, the underlying health of the treatment population, and the comparative safety and efficacy of alternative treatments. Avandia market share and utilization plummeted following a widely publicized study in May 2007 questioning the cardiovascular safety of the drug. In fact, prescribers responded more dramatically to the media blitz that followed than to the black box warning about the risk of heart failure added to labeling for both Avandia and Actos in August 2007. With the recent FDA restrictions, new prescriptions for Avandia are less likely to be written and Avandia utilization will continue to decline.

Although the generic dispensing rate among diabetic patients is high, the overall cost trends will increase as more patients are moved to insulin or have insulin added to their drug therapy regimen to maintain glycemic control. Oral agents and insulin likely will continue to have higher utilization rates due to aging population, more convenient delivery mechanisms for insulin, and adherence programs. Brand-name cost increases are projected over the next 12 months, moderated by increased generic utilization and generic price deflation. Januvia, Janumet, and metformin prescription claims and days of supply are projected to increase at a higher rate than those of other drugs in the class over the next 12 months. An Actos generic launch is anticipated in the third quarter of 2012, but because of declining use of insulin-sensitizing agents (or TZDs) at a class level, the impact on trend is expected to be minimal.

With the anticipated increase of diabetes among the younger generations due to continued obesity and the disease progression among the older generations, the antihyperglycemic therapeutic class will continue to be a major contributor to prescription drug trends in 2011 and beyond. Additionally, though not addressed in this article, a significant number of adult diabetic patients will be on antihypertensive therapy (often combination) and lipid-lowering therapy. Many aggressively managed adult diabetic patients are on more than 4 or 5 drugs, which should make a compelling case for urgent and aggressive analysis of the most effective intervention for diabetes and its associated comorbidities.