O r i g i n a l P a p e r

Barriers to Blood Pressure Control:

A STITCH Substudy
Sigrid A. E. Nelson, MSc;1 George K. Dresser, MD;2,3
Margaret K. Vandervoort, MSc;1 Cindy J. Wong, MSc;1
Brian G. Feagan, MD;1,2,3,4 Jeffrey L. Mahon, MD;1,3,4
Ross D. Feldman, MD1,2,3,5,6

Despite improvements in blood pressure (BP) predictor of lesser BP reduction. Of patients

control, a substantial percentage of patients do unable to reach target after 6 months, only 25%
not achieve target. The relative importance of were prescribed 3 drugs. Patients with diabetes
determinants of poor BP control is unclear. were significantly less likely to reach target than
Therefore, the authors conducted a post hoc those without (26% vs 64%, P<.001).
exploratory analysis to assess determinants of BP Antihypertensive therapy prescribed to patients
control. Data were collected in 45 general with diabetes was only marginally more intensive
practices, which enrolled patients with than to those without. In patients with
uncontrolled hypertension. Antihypertensive hypertension, whether with or without coexisting
medication changes throughout the 6-month diabetes, poor BP control appears to be at least
follow-up period were documented. Baseline and partially due to failure to uptitrate
6-month BPs were recorded. Of the 2030 antihypertensive therapy. Clinical inertia is
patients analyzed, 320 had diabetes. Overall, likely an important barrier to BP control. J Clin
42% of patients did not achieve BP control. In Hypertens (Greenwich). 2011;13:73–80.
multivariate analysis, failure to intensify therapy ª
2010 Wiley Periodicals, Inc.
was identified as a significant independent

From the Robarts Clinical Trials, Robarts Research

Institute;1 the Department of Medicine, Schulich
S ignificant advances have been made in blood
pressure (BP) control. Notwithstanding these
improvements, there remains a significant gap
School of Medicine and Dentistry, the University of
Western Ontario;2 the London Health Sciences between optimal BP control rates, such as those
Centre;3 the Department of Epidemiology and achievable in clinical trials driven by pre-set
Biostatistics, the University of Western Ontario;4 the protocols for antihypertensive drug ⁄ dose escala-
Vascular Biology Research Group, Robarts Research tion,1,2 and those achievable in the general
Institute;5 and the Department of Physiology & community.3
Pharmacology, the University of Western Ontario,6
London, ON, Canada A broad range of factors have been identified
Address for correspondence: that contribute to poor BP control. These include
Ross D. Feldman, MD, Departments of Medicine and systematic health care delivery factors, financial
of Physiology & Pharmacology, Schulich School of considerations (both for the patient as well as for
Medicine & Dentistry, University of Western Ontario, the health care system), but probably most impor-
and Vascular Biology Research Group, Robarts tantly, a number of behavioral factors.4–6 Patient-
Research Institute, 100 Perth Drive, PO Box 5015, Stn based behavioral factors, mostly reflected by mea-
B, London, ON N6A 5K8, Canada sures of adherence to antihypertensive prescription,
E-mail: [email protected] remain important predictors of BP control.7,8 How-
Manuscript received July 14, 2010; revised August 24,
ever, there has been increasing appreciation of
2010; accepted September 8, 2010
the importance of health care professional–
doi: 10.1111/j.1751-7176.2010.00392.x based behavioral factors as determinants of BP

VOL. 13 NO. 2 FEBRUARY 2011 THE JOURNAL OF CLINICAL HYPERTENSION 73

 17517176, 2011, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1751-7176.2010.00392.x by CAPES, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
control.9–11 Poor BP control linked to health care stage 4 or 5 chronic kidney disease. Patients could
professional behavior has been most extensively not be participating in other hypertension studies.
studied in the context of so-called therapeutic (or Preintervention and postintervention BP measure-
clinical) inertia.10,12 In general hypertensive popula- ments were taken approximately 6 months apart
tions, the presence of clinical inertia (at the and were based on an average of 5 readings using
extremes) reduces the probability of BP control by a standardized procedure and an automated BP
>80%.13 device (BpTRU; VSM Med Tech, Coquitlam, Brit-
Although patients with hypertension and diabetes ish Columbia, Canada). Changes in antihyperten-
are among those at highest risk for hypertension- sive medication prescription throughout this follow-
related complications, they demonstrate the lowest up period were documented.
control rates. As demonstrated in a 2006 commu- In this post hoc analysis, STITCH patients were
nity-based cross-sectional survey from Ontario, Can- grouped based on (1) whether they achieved BP
ada (ON-BP), for the subset of patients with control at 6-month follow-up and (2) whether they
hypertension and diabetes, BP control rates were had a documented diagnosis of diabetes at baseline.
only approximately half of those without diabetes.14 Patients whose diabetes status was unknown were
In the ON-BP survey, only 35% of patients with dia- excluded from this analysis. Outcomes for this
betes and hypertension had BPs below the recom- exploratory analysis were BP reduction and the
mended target of 130 ⁄ 80 mm Hg, as compared with proportion of patients achieving BP target at the 6-
66% control in the general hypertensive population month follow-up visit. Antihypertensive medication
whose target was 140 ⁄ 90 mm Hg.14 Notably, in prescription outcomes explored included the sum of
those patients with hypertension and diabetes, 59% standard daily doses of any antihypertensive medi-
achieved a BP <140 ⁄ 90 mm Hg—a control rate cations, sum of standard daily doses of individual
comparable with that reported for patients without classes of medication, number of individual mole-
diabetes.14 Whether the disparity between target-spe- cules prescribed, and use of fixed-dose combination
cific control rates for those with or without diabetes therapies. Interactions in multivariate regression
solely reflects the lower BP targets for those with dia- models were explored to assess whether diabetes
betes (<130 ⁄ 80 mm Hg15,16), resistance to aggres- status moderated the BP-lowering effects of increas-
sive therapy ⁄ use of diuretics by health care ing doses of antihypertensive therapy.
professionals and ⁄ or a greater prevalence of refrac-
tory hypertension in patients with diabetes and Statistical Methods
hypertension is unclear. Therefore, we conducted a Descriptive statistics are presented for baseline char-
post hoc exploratory analysis to assess possible barri- acteristics. Two-sample t tests (unadjusted and
ers to BP control, in the context of prescriptions, in adjusted for clustering) were used to compare indi-
the overall patient population as well as in the subset vidual-level characteristics of patients with and
of patients with diabetes who participated in the without diabetes. Estimates of intraclass correlation
Simplified Therapeutic Intervention To Control coefficients were calculated for each outcome vari-
Hypertension (STITCH) study. able using methods presented by Donner and Klar
(2000) by using the mean square values from a
METHODS one-way analysis of variance.18,19 We report t tests
STITCH was a cluster randomized controlled trial adjusted by the design effect since the data collected
of hypertension management conducted in south- were clustered by family practice in the STITCH
western Ontario between February 2005 and Janu- study. Two-tailed P values were reported. To adjust
ary 2007. The study design was previously for the effects of confounding and clustering by
described in detail.17 Briefly, data were collected in family physician for BP reduction outcomes, the
45 general practices, each enrolling up to 50 SAS PROC MIXED procedure (restricted maxi-
patients with uncontrolled hypertension. Eligible mum likelihood REML method) was used to fit
patients were men or women, 18 years or older multilevel models that incorporated random effects.
who, on entry to the study, had uncontrolled Variables with a multivariate P value >.05 were
hypertension (systolic BP [SBP] 140 mm Hg or removed from the model.
diastolic BP [DBP] 90 mm Hg for patients with- Since the nature of this study was exploratory,
out diabetes mellitus or SBP 130 mm Hg or DBP there was no adjustment for multiple testing. All
80 mm Hg for patients with diabetes mellitus), statistical analyses were performed using SAS soft-
and with no history of ischemic heart disease, atrial ware, version 9.1 for Windows (SAS Institute Inc,
fibrillation, peripheral vascular disease, stroke, or Cary, NC).

74 THE JOURNAL OF CLINICAL HYPERTENSION VOL. 13 NO. 2 FEBRUARY 2011

 17517176, 2011, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1751-7176.2010.00392.x by CAPES, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table I. Characteristics of STITCH Patients Based on study, compared with those who did reach BP tar-
Achieving Blood Pressure Target at 6-Month Follow-Up get. Notably, the intensity of antihypertensive drug
therapy for those patients who failed to reach BP
Did Not
targets was not greater than those who reached tar-
Achieved Achieve
Target Target
gets in terms of the number of drugs prescribed
Characteristic (n=1170) (n=860) (mean of 1.7 vs 1.9 drugs, respectively), the inten-
sity of drug dosing (mean of 1.7 standard doses
Blood pressure characteristics
regardless of whether target achieved), and the pro-
Baseline SBP, mean 152 (13) 157 (16)
(SD), mm Hg portion of those taking 3 antihypertensive drugs
SBP change, mean )27 (15) )10 (16) (25% vs 26%, respectively). This finding suggests
(SD), mm Hg that even for patients with inadequate BP control,
Baseline DBP, mean 87 (11) 89 (11) therapy was not advanced to any greater extent
(SD), mm Hg than for those who achieved BP control, suggesting
DBP change, mean )13 (9) )4 (11) a ‘‘ceiling’’ phenomenon for the prescription of >2
(SD), mm Hg antihypertensive drugs.
Antihypertensive prescription characteristics
No. of standard doses, 1.7 (1.3) 1.7 (1.5) Barriers to BP Reduction: Multivariate Analyses
mean (SD)
Since BP control for the STITCH population was
Increase in standard 0.8 (1.0) 0.6 (1.2)
primarily driven by the extent of SBP control, a
doses from baseline,
mean (SD) multivariate model is presented in Table II with
No. of drugs, mean 1.9 (1.1) 1.7 (1.2) predictors of reduction in SBP. The STITCH-Care
(SD) treatment algorithm (featuring the use of single-pill
Increase in No. of 1.0 (1.0) 0.7 (1.1) combinations), diabetes status and an increase in
drugs from baseline, the number of drugs prescribed prior to the 6-
mean (SD) month visit were identified as independent predic-
No. of patients taking 302 (26%) 214 (25%) tors of reduction in SBP after adjustment for base-
3 drugs, % line SBP.
No. of patients taking 822 (70%) 502 (58%)
2 drugs, %
BP Control in Patients With Diabetes
No. of patients 779 (64%) 545 (49%)
Of the 2030 patients analyzed, 320 (16%) had a
uptitrated over
6 months, % diagnosis of type I or type II diabetes. Patient-level
characteristics are presented in Table III and show
Abbreviations: DBP, diastolic blood pressure; SBP, that patients in the two groups were relatively simi-
systolic blood pressure; SD, standard deviation; STITCH,
lar in terms of demographic and baseline character-
Simplified Therapeutic Intervention To Control Hypertension.
istics. However, mean DBP was 4 mm Hg lower in
patients with diabetes, likely due to the different
definition of control for patients with diabetes in
RESULTS the study entrance criteria. Notably, in this compar-
Of the 2104 hypertensive patients who were ison, baseline SBPs were similar in patients with or
analyzed in the STITCH study, 74 were excluded without diabetes. At baseline, a substantially
from the analysis, 56 patients withdrew during the higher proportion of patients with diabetes had
study so follow-up data were not available, and 18 SBP and DBP above target (67%) compared with
patients had incomplete diabetes status data. Thus, patients without diabetes (43%), as shown in Fig-
this exploratory analysis was based on 2030 ure 1. In essentially all patients with diabetes
patients. (99%), SBP was >130 mm Hg, and in the vast
majority of those without diabetes (91%) SBP was
Predictors of Reaching BP Target >140 mm Hg.
Overall, 42% of patients in the STITCH study BP outcomes after 6 months of follow-up are
failed to reach BP targets through the course of the reported in Table IV by diabetes status. At 6 months,
trial (none of whom were at target at baseline). As patients with diabetes were significantly less likely to
shown in Table I, patients who did not achieve BP reach their recommended BP target than those with-
control through the study tended to have higher out diabetes (26% vs 64%, P-adjusted <.001).
baseline BPs and demonstrated a much less signifi- Notably, the mean BP reduction during the 6-month
cant reduction in BP through the course of the follow-up was comparable between patients with

VOL. 13 NO. 2 FEBRUARY 2011 THE JOURNAL OF CLINICAL HYPERTENSION 75

 17517176, 2011, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1751-7176.2010.00392.x by CAPES, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table II. Predictors of Reduction in SBP During 6 Months
Univariate Analysis Multivariate Model
D SBP, Standard D SBP, Standard
Determinant mm Hg Error P Valuea mm Hg Error P Valuea
STITCH-Care )4.94 1.60 .004 )3.02 1.45 .04
Age (per 10-y increase) )1.87 0.32 <.001
Female )2.52 0.82 .004
Not diabetic )1.25 1.11 .263 )2.26 0.89 .01
Increase in No. of standard dosesb )1.84 0.38 <.001
Increase in No. of drugsb )3.05 0.40 <.001 )2.06 0.33 <.001
Baseline SBP (per 10-mm Hg increase) )6.87 0.24 <.001 )6.76 0.23 <.001
Abbreviations: D, change over time; SBP, systolic blood pressure; STITCH, Simplified Therapeutic Intervention To Control
Hypertension. aP values were derived by adjustment for clustering in the model. bIncrease based on any antihypertensive therapy
prescription change made to baseline medications prior to 6-month follow-up.

confidence interval [CI], 2.12–2.39; no diabetes,

Table III. Baseline Characteristics of STITCH Patients
1.86; 95% CI, 1.82–1.92; P-adjusted <.048), and
by Diabetes Status
only marginally more intensive antihypertensive ther-
Characteristics Diabetes No Diabetes apy, measured as standard daily doses (mean stan-
No. of patients (%) 320 (15.8) 1710 (84.2) dard doses: diabetes, 2.35; 95% CI, 2.17–2.54; no
Age, mean (range), y 61.84 61.2 diabetes, 1.65; 95% CI, 1.59–1.71; P-adjusted
(20.8–88.7) (18.6–93.0) <.001). Interestingly, the apparent resistance to
Women, % 50.0 54.6 advancing antihypertensive therapy in patients with
Baseline SBP, mean (SD), 155.5 (14.5) 154.0 (14.3)
diabetes (despite not achieving BP targets) was not
mm Hg
due to resistance to the use of diuretics. In fact, there
Baseline DBP, mean (SD), 84.5 (10.8) 88.6 (10.8)
mm Hg was a trend toward patients with diabetes being pre-
Randomized to 37.5 39.7 scribed slightly more intensive diuretic therapy
STITCH-Care arm, % (mean standard doses of diuretic if prescribed: diabe-
tes, 0.87; 95% CI, 0.78–0.97; no diabetes, 0.74;
Abbreviations: DBP, diastolic blood pressure; SBP,
systolic blood pressure; SD, standard deviation, STITCH,
95% CI, 0.71–0.78; P-adjusted=.12). Although
Simplified Therapeutic Intervention To Control Hypertension. patients with diabetes were as likely to be prescribed
a diuretic as those without diabetes, in both sub-
groups, patients who achieved target were more
diabetes (change in SBP [DSBP], 18.718.2 mm Hg; likely to be prescribed a diuretic than those who did
change in DBP [DDBP], 8.211.9 mm Hg) and not achieve target (diabetes, 75.9% vs 61.6%,
patients without diabetes (DSPB, 19.918.0 mm P=.02; no diabetes, 69.9% vs 61.3%, P<.001). As
Hg; DDBP, 9.210.5 mm Hg) as was the propor- would be expected, patients with diabetes were more
tion of patients who reached a common target of likely to be prescribed an angiotensin-converting
140 ⁄ 90 mm Hg (diabetes, 60.6%; no diabetes, enzyme [ACE] inhibitor than patients without diabe-
63.6%; P-adjusted=.692). tes (55% vs 35%). The proportion of patients using
any fixed-dose combination therapy (all of which
Prescription-Based Barriers to Poor BP Control in contained a diuretic) at the follow-up visit was not
Patients With Diabetes significantly different between the two groups.
Antihypertensive drug therapy prescriptions are sum- As seen in Figure 2, patients who achieved their
marized by diabetes status in Table V. Based on clin- BP target at 6 months were significantly more likely
ical trial findings,4 we would have expected that to be taking 2+ therapies than those who did not
patients with diabetes receive 1 or 2 drugs more than reach target (diabetes, 81% vs 68%, P=.03; no dia-
those without diabetes in order to achieve compara- betes, 70% vs 55%, P<.001).
ble BP control rates. However, despite being much
less likely to achieve BP targets, patients with diabetes DISCUSSION
were only taking a slightly higher number of drugs Overall, the STITCH trial demonstrated that intro-
(mean number of drugs: diabetes, 2.26; 95% duction to health care practitioners of a simplified

76 THE JOURNAL OF CLINICAL HYPERTENSION VOL. 13 NO. 2 FEBRUARY 2011

 17517176, 2011, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1751-7176.2010.00392.x by CAPES, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Figure 1. Breakdown of uncontrolled blood pressure based on systolic blood pressure (SBP), diastolic blood pressure
(DBP), or both uncontrolled at baseline, by diabetes status.

Table IV. Blood Pressure Outcomes at 6-Month Follow-Up

No Diabetes Adjusted
Outcome Diabetes (n=320) (n=1710) P Value ICC P Valuea
Blood pressure at target, % 25.9 63.6 <.001 0.101 <.001
Blood pressure at target 140 ⁄ 90 mm Hg, % 60.6 63.6 .317 0.122 .692
SBP reduction, mean (95% CI), mm Hg 18.7 (16.7–20.7) 19.9 (19.0–20.7) .284 0.082 .618
DBP reduction, mean (95% CI), mm Hg 8.2 (6.9–9.5) 9.2 (8.7–9.7) .156 0.083 .511
Abbreviations: CI, confidence interval; DBP, diastolic blood pressure; ICC, intraclass correlation coefficient; SBP, systolic
blood pressure. aAdjusted for clustering.

treatment algorithm for the management of hyper- those seen in a general population of patients with
tension improved BP control. Notwithstanding, the hypertension, considering that only patients with
BP control rates achieved through the course of the uncontrolled hypertension at entry could participate
STITCH study were similar to those seen in a com- in the STITCH study. The 6-month control rate of
munity setting. These control rates fall far short of 58% among the population of STITCH patients
those seen in large clinical trials where patients (presumably) aware of their diagnosis is less than
were treated according to a fixed treatment proto- the control rates among ‘‘aware patients’’ reported
col (and where compliance to adherence to the pro- in several recent North American population-based
tocol was closely monitored). The data collected surveys (eg, 62% in National Health and Nutrition
from the STITCH trial offered an opportunity to Examination Survey [NHANES] 20083 and 76%
explore the drug-related determinants of poor BP in ON-BP14). On the other hand, the STITCH pop-
control. This analysis demonstrated that patients ulation would have included proportionally more
who did not achieve target BPs were not prescribed patients with newly diagnosed ⁄ uncontrolled hyper-
more antihypertensive therapy. This analysis also tension and ⁄ or more refractory hypertension, con-
confirmed the observation that patients with diabe- siderations that would be expected to result in a
tes have significantly lower BP control rates. somewhat lower control rate than that seen in a
Despite their poor BP control rates, patients with general population of patients with hypertension.
diabetes were prescribed only marginally greater Notably, these control rates were almost identical
intensity of antihypertensive therapy. Lastly, our to those reported during a comparable follow-up
analysis demonstrated that this resistance to more period for Canadian sites in the Antihypertensive
aggressive treatment in patients with diabetes was and Lipid-Lowering Treatment to Prevent Heart
not due to an obvious aversion to the use of diuret- Attack Trial (ALLHAT) (55% at 6 months20).
ics. In the general population of patients studied in
Overall, the BP control rates achieved in the the STITCH trial, those who failed to reach target
STITCH trial would appear to be representative of BPs started with higher baseline pressures and

VOL. 13 NO. 2 FEBRUARY 2011 THE JOURNAL OF CLINICAL HYPERTENSION 77

 17517176, 2011, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1751-7176.2010.00392.x by CAPES, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Table V. Antihypertensive Medication Use at 6-Month Follow-Up
No Diabetes Adjusted
Medication Use Diabetes (n=320) (n=1710) P Value ICC P Valuea
All antihypertensive drugs
No. of drugs, mean (95% CI) 2.26 (2.12–2.39) 1.86 (1.82–1.92) <.001 0.150 .048
Drug intensity, mean standard doses (95% CI) 2.35 (2.17–2.54) 1.65 (1.59–1.71) <.001 0.071 <.001
Patients using specific drug classes, No. (%)
Diuretic 209 (65) 1141 (67)
ACE inhibitor 176 (55) 589 (34)
ARB 116 (36) 646 (38)
CCB 131 (41) 463 (27)
BB 70 (22) 271 (16)
Drug intensityb by individual class, overall,
mean (95% CI)c
Diuretic 0.57 (0.49–0.65) 0.50 (0.47–0.52) .08 0.058 .35
ACE inhibitor 0.83 (0.72–0.94) 0.42 (0.39–0.46) <.001 0.087 .002
ARB 0.39 (0.33–0.46) 0.40 (0.37–0.42) .93 0.192 .93
CCB 0.36 (0.30–0.41) 0.22 (0.20–0.24) <.001 0.053 .01
BB 0.18 (0.13–0.23) 0.10 (0.09–0.11) .005 0.028 .06
Drug intensityb by individual class, if prescribed,
mean (95% CI)
Diuretic 0.87 (0.78–0.97) 0.74 (0.71–0.78) .01 0.058 .12
ACE inhibitor 1.51 (1.37–1.65) 1.22 (1.16–1.29) <.001 0.087 .02
ARB 1.09 (0.99–1.18) 1.05 (1.01–1.09) .48 0.192 .73
CCB 0.87 (0.79–0.95) 0.81 (0.76–0.85) .16 0.053 .27
BB 0.81 (0.64–0.97) 0.64 (0.59–0.68) .049 0.028 .07
Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; BB, b-blocker; BP, blood pressure;
CCB, calcium channel blocker; CI, confidence interval; ICC, intraclass correlation coefficient. aAdjusted for clustering. bDrug
intensity measured as mean standard doses. cMean of all patients including those with dose of 0.

Figure 2. Proportion of patients taking 0 (h), 1 ( ), 2 ( ), or 3+ ( ) antihypertensive medications by diabetes status

and achieving blood pressure target at 6-month follow-up.

demonstrated a lesser response to antihypertensive Further, the prevalence of a number of secondary

therapy. Multiple factors contribute to resistance to forms of hypertension was not systematically
antihypertensive therapy, including noncompliance assessed. However, regardless of the etiology of the
and secondary forms of hypertension. Since the resistance to therapy, it is most notable that the
assessment of antihypertensive drug intensity was response to poor BP control in these patients did not
based on physician records and not by prescription include any significant increase in their antihyperten-
refills, persistence to therapy and compliance could sive drug prescription. For the majority of these
not be determined from the STITCH trial data. patients, a prescription ceiling was reached at 2

78 THE JOURNAL OF CLINICAL HYPERTENSION VOL. 13 NO. 2 FEBRUARY 2011

 17517176, 2011, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1751-7176.2010.00392.x by CAPES, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
antihypertensive medications with resistance to esca- difficult to achieve for patients with diabetes, but
lating therapy beyond that ceiling despite failure to rather that failure to achieve lower treatment tar-
achieve BP target. Thus, regardless of the presence or gets was primarily behavioral and included failure
absence of any other biologic or patient behavioral to intensify therapy. It has been speculated that
characteristics leading to poor BP control, therapeu- part of the difficulty in achieving better BP control
tic inertia is seen as a key factor. Notably, at the time rates in patients with diabetes was related to a lack
of the study, only 2-drug single-pill combinations of ‘‘buy-in’’ by health care professionals about the
were available in Canada. With the re-emergence of merits of more aggressive treatment targets (ie,
3-drug single-pill combinations in the near future <130 ⁄ 80 mm Hg) for patients with diabetes.
and the clear evidence of the utility of single-pill Whether the recent findings from the Action to
combinations to ameliorate therapeutic inertia, we Control Cardiovascular Risk in Diabetes
would speculate that the greater availability of these (ACCORD) study, ie, failing to show benefit of
formulations may raise the ceiling in antihyperten- adopting an SBP target <140 mm Hg in regards to
sive prescription resistance. reduction in major cardiovascular events,23 will
Patients with hypertension and diabetes are much further erode the extent of buy-in to the benefit of
less likely to reach BP targets. Notwithstanding their lower targets remains to be determined.
lack of BP control, patients with diabetes were Our data support the hypothesis that resistance
receiving only slightly more intensive antihyperten- to advancing antihypertensive therapy past a ‘‘ceil-
sive drug therapy than those without diabetes. Sur- ing’’ of 2 antihypertensive drugs remains a signifi-
prisingly, patients with diabetes were equally likely cant barrier to better BP control in patients with
to be treated with a diuretic and were taking a mar- diabetes. In patients without diabetes, studies such
ginally higher dose of diuretic, suggesting that the as ALLHAT have reported that, on average, 2 or 3
long-standing concern over diuretic-induced insulin antihypertensive drugs are required to achieve the
resistance21 has not outweighed the appreciation of >75% BP control rates,20,24 characteristic of large,
the BP-lowering benefits of the drug in the opinion of regimented clinical trials using a 140 ⁄ 90 mm Hg
many primary physician prescribers. In the ON-BP target. In contrast, to achieve lower BP targets,
study, McInnis and colleagues reported significantly such as in patients with additional cardiovascular
lower use of diuretics among patients with diabe- risk factors such as diabetes, or to control patients
tes.22 They suggested that the aversion to the use of with refractory hypertension, 3 antihypertensive
diuretics might constitute a significant factor in their agents are generally required.4 In the STITCH
lower BP control rates. In the STITCH study, the use study, after 6 months of follow-up, patients with
of diuretics was not reduced in patients with diabe- diabetes were receiving, on average, only slightly
tes. The reason for this discrepancy is unclear but more than 2 antihypertensive drugs. Notwithstand-
could relate to the emphasis on the use of single-pill ing, differences in length of follow-up (the
combinations, which, at the time of the study, uni- ALLHAT study’s BP control rates were reported
versally included thiazide diuretics. However, regard- for up to 5 years of follow-up), in our comparison
less of the explanation for this discrepancy, our data of those prescribing factors differentiating patients
support the hypothesis that even in the setting where who did or did not achieve BP targets, we would
diuretic use is not a limiting factor, BP control rates suggest that resistance to the prescription of the
in patients with diabetes remain significantly lower third (and potentially fourth) drug required to con-
than in those without diabetes, and drug prescription trol BP in patients with diabetes appears to be a
intensity remains suboptimal. Thus, aversion to major barrier to adequate BP control.
diuretic use cannot be viewed as the only factor Potential causes of resistance to escalation of
responsible for suboptimal management of hyperten- antihypertensive therapy in patients with diabetes
sion in this high-risk population. that could not be explored in our data include: (1)
Increased biologic resistance to antihypertensive health care professional resistance to escalation
treatment for patients with diabetes has been sug- of dosing ⁄ addition of additional antihypertensive
gested as an important component of the challenge drugs, and (2) increasing patient resistance to being
in managing BP for this subpopulation. Interest- prescribed additional antihypertensive drugs. How-
ingly, when modified BP targets were explored, ever, we have observed that in hypertensive patients
patients without diabetes displayed parallel losses with diabetes, similar to in the larger population of
in control rate as BP target was lowered compared hypertensive patients without diabetes, insufficient
with patients with diabetes. Thus, our data do not uptitration (clinical inertia) appears to be associated
suggest that lower BP targets are differentially more with poor BP control.

VOL. 13 NO. 2 FEBRUARY 2011 THE JOURNAL OF CLINICAL HYPERTENSION 79

 17517176, 2011, 2, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1751-7176.2010.00392.x by CAPES, Wiley Online Library on [25/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
LIMITATIONS AND CONCLUSIONS pressure: the JNC 7 report. JAMA. 2003;289(19):2560–
2572.
Several limitations to this study should be noted. 5 Ogedegbe G. Barriers to optimal hypertension control.
First, since this was a post hoc analysis and J Clin Hypertens (Greenwich). 2008;10(8):644–646.
intended to be exploratory, confirmatory studies 6 Schmittdiel JA, Uratsu CS, Karter AJ, et al. Why don’t
diabetes patients achieve recommended risk factor targets?
would be required to support any conclusion. There Poor adherence versus lack of treatment intensification.
was no adjustment made for multiple testing. Sec- J Gen Intern Med. 2008;23(5):588–594.
ond, diagnosis of diabetes was only captured 7 Bramley TJ, Gerbino PP, Nightengale BS, et al. Relation-
ship of blood pressure control to adherence with antihyper-
at baseline. There may have been patients diag- tensive monotherapy in 13 managed care organizations.
nosed with diabetes within the 6 months of study J Manag Care Pharm. 2006;12(3):239–245.
follow-up who were classified as not having 8 Elliott WJ. What factors contribute to the inadequate con-
trol of elevated blood pressure? J Clin Hypertens (Green-
diabetes. Finally, antihypertensive medication data wich). 2008;10(suppl 1):20–26.
were based on reported prescriptions rather than 9 Moser M. Physician or clinical inertia: what is it? Is it
pharmacy fill records or patient-reported compli- really a problem? And what can be done about it? J Clin
Hypertens (Greenwich). 2009;11(1):1–4.
ance. This method may result in a higher estimate 10 Oliveria SA, Lapuerta P, McCarthy BD, et al. Physician-
of antihypertensive medication use. However, it related barriers to the effective management of uncon-
should accurately reflect physician willingness to trolled hypertension. Arch Intern Med. 2002;162(4):413–
420.
intensify treatment. 11 Ho PM, Magid DJ, Shetterly SM, et al. Importance of
Notwithstanding these limitations, this post hoc therapy intensification and medication nonadherence for
analysis suggests that poor BP control is associated blood pressure control in patients with coronary disease.
Arch Intern Med. 2008;168(3):271–276.
with clinical inertia or failure to uptitrate antihyper- 12 Phillips LS, Branch WT, Cook CB, et al. Clinical inertia.
tensive therapy in patients both with and without Ann Intern Med. 2001;135(9):825–834.
diabetes. In both subgroups, the increased number 13 Okonofua EC, Simpson KN, Jesri A, et al. Therapeutic
inertia is an impediment to achieving the Healthy People
of drugs required for control to target could be an 2010 blood pressure control goals. Hypertension.
important limiting factor impeding improved BP 2006;47(3):345–351.
control rates. Remedies to clinical inertia in pri- 14 Leenen FH, Dumais J, McInnis NH, et al. Results of the
Ontario survey on the prevalence and control of hyperten-
mary care may include a focus on the patient’s role sion. CMAJ. 2008;178(11):1441–1449.
in therapy intensification, improved BP monitoring, 15 American Diabetes Association. Standards of medical care
effective research dissemination to physicians in diabetes–2009. Diabetes Care. 2009;32(suppl 1):S13–
S61.
regarding appropriate aggressiveness of antihyper- 16 Canadian Diabetes Association Clinical Practice Guide-
tensive treatment, and potentially the use of 3-drug lines Expert Committee. Canadian Diabetes Association
single-pill combination therapies. 2008 clinical practice guidelines for the prevention and
management of diabetes in Canada. Can J Diabetes.
2008;32(suppl 1):S1–S201.
Disclosures: Dr Feldman and Sigrid Nelson had full access to 17 Feldman RD, Zou GY, Vandervoort MK, et al. A simpli-
all of the data in the study and take responsibility for the fied approach to the treatment of uncomplicated hyperten-
integrity of the data and the accuracy of the data analysis. sion: a cluster randomized, controlled trial. Hypertension.
The sponsor had no role in the design and conduct of the 2009;53(4):646–653.
study; collection, management, analysis, and interpretation of 18 Donner A, Klar N. Design and Analysis of Cluster Ran-
the data; and preparation, review, or approval of the manu- domization Trials in Health Research. London, England:
script. The STITCH study was supported by an unrestricted Arnold; 2000.
grant-in-aid from Pfizer Canada, Kirkland, Quebec. 19 Donner A, Koval JJ Design considerations in the estima-
tion of intraclass correlation. Ann Hum Genet. 1982;
46(Pt 3):271–277.
REFERENCES 20 Cushman WC, Ford CE, Cutler JA, et al. Success and pre-
1 ALLHAT Officers and Coordinators for the ALLHAT dictors of blood pressure control in diverse North American
Collaborative Research Group. Major outcomes in high- settings: the Antihypertensive and Lipid-Lowering Treat-
risk hypertensive patients randomized to angiotensin-con- ment to prevent Heart Attack Trial (ALLHAT). J Clin Hy-
verting enzyme inhibitor or calcium channel blocker vs pertens (Greenwich). 2002;4(6):393–404.
diuretic: the antihypertensive and lipid-lowering treatment 21 Dronavalli S, Bakris GL. Mechanistic insights into diure-
to prevent heart attack trial (ALLHAT). JAMA. 2002; tic-induced insulin resistance. Hypertension. 2008;52(6):
288(23):2981–2997. 1009–1011.
2 Julius S, Kjeldsen SE, Weber M, et al. Outcomes in hyper- 22 McInnis NH, Fodor G, Lum-Kwong MM, et al. Antihy-
tensive patients at high cardiovascular risk treated with pertensive medication use and blood pressure control: a
regimens based on valsartan or amlodipine: the VALUE community-based cross-sectional survey (ON-BP). Am J
randomised trial. Lancet. 2004;363(9426):2022–2031. Hypertens. 2008;21(11):1210–1215.
3 Egan BM, Zhao Y, Axon RN. US trends in prevalence, 23 The ACCORD Study Group. Effects of intensive blood-
awareness, treatment, and control of hypertension, 1988– pressure control in type 2 diabetes mellitus. N Engl J
2008. JAMA. 2010;303(20):2043–2050. Med. 2010;362(17):1575–1585.
4 Chobanian AV, Bakris GL, Black HR, et al. The seventh 24 Bakris GL. A practical approach to achieving recom-
report of the Joint National Committee on prevention, mended blood pressure goals in diabetic patients. Arch
detection, evaluation, and treatment of high blood Intern Med. 2001;161(22):2661–2667.

80 THE JOURNAL OF CLINICAL HYPERTENSION VOL. 13 NO. 2 FEBRUARY 2011