new onset diabetes during antihypertensive therapy

AMERICAN JOURNAL OF HYPERTENSION | VOLUME 21 NUMBER 5 | 493-499 | May 2008 493

STATE OF THE ARTnature publishing group

New Onset Diabetes During Antihypertensive TherapyMichael H. Alderman1

Cardiovascular disease (CVD) accounts for the majority of deaths and most of the morbidity among persons with diabe-tes.1 Given the large numbers of hypertensive persons, the age-related character of rising blood pressure, and the coincidence of obesity with hypertension, it is not surprising that diabetes—both at onset and during its treatment—is so common among treated hypertensive persons. With the waist circumferences among populations tending to increase worldwide, the age-related incidence of diabetes has increased to 2–3% annually in persons older than 65 years.2 Indeed, the increasing inci-dence of adult-onset diabetes is now threatening to disrupt the long-standing pattern of declining cardiovascular mortality.2,3 New onset diabetes (NOD) has become an outcome of major interest in clinical trials of antihypertension therapy, as well as a matter of practical concern for both patients and their physi-cians. At the same time, despite the considerable clinical and epidemiological information now available, uncertainty con-tinues to prevail regarding the causes, consequences, preven-tion, and treatment of NOD.

1Department of Epidemiology and Population Health, albert Einstein College of Medicine, Bronx, New york, USa. Correspondence: Michael H. alderman ([email protected])

Received 30 December 2007; first decision 8 January 2008; accepted 20 January 2008; advance online publication 20 March 2008. doi:10.1038/ajh.2008.17

© 2008 American Journal of Hypertension, Ltd.

The epidemiology of NodThe natural incidence of diabetes in untreated hypertensive persons has been demonstrated in the placebo-controlled SHEP (Systolic Hypertension in Elderly Program) Trial, reported in 1988.4 SHEP participants were older and tended to have more evidence of CVD than the general hypertensive population, and 43% of control subjects eventually received active treatment, thus tending to mute the true magnitude of differences between treated and “untreated” subjects. Chlorthalidone was the active drug, but almost 50% of the actively treated participants were also exposed to β-blockers during the trial. The annual incidence of NOD was 2.7% in the placebo group, and nearly 50% higher (3.9%) in the treated group. The precision and generalizability of these findings are limited by participant selection, study characteristics, and the secular trends that have been in place since the study was completed. Nevertheless, the data do provide strong evidence that diuretics and β-blockers increase the risk of hyperglyce-mia and diabetes, while also establishing a benchmark for the natural incidence of NOD in hypertensive subjects.

Subsequent trials confirmed the adverse impact of diuret-ics and β-blockers.5 It was hoped that newer antihypertensive agents, with more favorable metabolic mechanisms, would not increase the natural hyperglycemia associated with aging and hypertension. For the most part, but not invariably, newer agents have indeed realized that hope. For example, in the

New onset diabetes (NOD) is common among hypertensive patients, whether they are being treated for hypertension or not, and is associated with subsequently increased cardiovascular disease (CVD). Thiazide-like diuretics and β-blockers are more likely to provoke hyperglycemia when compared with drugs that block the renin–angiotensin system, and calcium channel blockers. However, in contrast to the NOD arising during treatment with other antihypertensive drugs, the NOD that occurs during diuretic treatment, has not been shown to increase CVD, either in clinical trials, or during longer observational studies. In fact, blood pressure reduction achieved by diuretic treatment may avert the expected increase of CVD in NOD. Conventional blood pressure reduction (along with lipid lowering) is the proven approach to preventing CVD in diabetes, in whatever circumstances the diabetes occurs. apprehensions relating to the potential onset of NOD should not influence the choice of the initial antihypertensive treatment choice,

nor should it invariably lead to discontinuation of diuretics (although such a step may reverse hyperglycemia). NOD can also sometimes be eliminated by correcting hypokalemia with a potassium-sparing diuretic, and/or potassium supplementation, or by adding a potassium-conserving antihypertensive drug such as an aCEI, aRB, or an anti-aldosterone agent. If all these stratagems fail (or are unsuitable), and the diuretic is essential to blood pressure control, then hypoglycemic therapy is indicated. NOD does adversely affect quality of life, and is not to be accepted lightly. However, diuretic-induced hypergycemia can be managed, and should be tolerated if a diuretic is essential for blood pressure control. In summary, the potential for occurrence of NOD certainly needs consideration, but it is not an insurmountable challenge, and must not compromise aggressive blood pressure control, which remains the primary tool for antihypertensive care.Am J Hypertens 2008; 21:493-499 © 2008 American Journal of Hypertension, Ltd.

at Uppsala U

niversitetsbibliotek on Novem

ber 16, 2014http://ajh.oxfordjournals.org/

Dow

nloaded from

http://www.nature.com/doifinder/10.1038/ajh.2008.17

mailto:[email protected]

http://ajh.oxfordjournals.org/

494 May 2008 | VOLUME 21 NUMBER 5 | AMERICAN JOURNAL OF HYPERTENSION

New Onset Diabetes During Antihypertensive TherapySTATE OF THE ART

Losartan Intervention for Endpoint reduction (LIFE) trial, patients randomized to losartan had a 25% lower rate of NOD than those randomized to atenolol.6 The actual percentages of NOD were 6 and 8%, respectively, and this was statisti-cally significant. However, the absolute increment of NOD (241/4,605 vs. 319/4,588) was 1.7%. That is, only 78 out of almost 4,600 treated subjects were affected. This occurred in a setting wherein most participants in both groups also received diuretics at some point in the study.

The incidence of NOD with ACEI/ARB treatment as com-pared to the incidence with other treatments has been further assessed in a meta-analysis of 11 studies in cohorts with hyper-tension, heart failure, or CVD. These 11 studies employed a variety of agents as control intervention, although diuret-ics represented the control agent in the two largest studies. It should also be noted that exposure to diuretics was substan-tial in both arms of most of the studies. Nevertheless, despite these limitations, the meta-analysis revealed that the ACEI/ARB groups had a 1.7% lower incidence of NOD than the con-trol groups did. This excess, albeit significant, was modest, and meant that 58 subjects had to be treated with an ACEI and/or an ARB to prevent one occurrence of NOD.7

Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT),8 the largest antihypertensive trial to date, compared the experiences of patients random-ized to chlorthalidone, amlodipine, or lisinopril. The subjects in each randomized group received multiple additional drugs. In ALLHAT, by design, the second agent was atenolol, which was prescribed for ~25% of all the subjects, and equally across groups. The annual incidence of NOD was 2.45%—within the range in SHEP. But there were substantial differences between groups. As shown in Figure 1, lisinopril-treated subjects fared best, with an annual incidence of 2.03%; for the amlodipine group it was 2.45% and the chlorthalidone group 2.9%. At 4 years, the cumulative incidence of NOD was 11.6% for the chlorthalidone group as compared with 8.1% for lisinopril sub-jects. Although this difference was measurable and significant,

it is important to note that the net absolute incidence of NOD attributable to chlorthalidone was <1 case per 100 treated subjects per year.9

From these multiple sets of comparative data it is not possi-ble to determine whether it was the effect of a particular study drug, or any combination of the opposing effects of multiple drugs on glucose homeostasis that accounted for the observed differences in the incidence of hyperglycemia.

In an attempt to unravel that dilemma, the DREAM (Diabetes REduction Assessment with ramipril and rosiglita-zone Medication) Trial was undertaken to find out whether a renin system blocking agent, ramipril, could prevent NOD.10 Its 5,000 participants each had impired fasting blood glucose, were free of CVD, and included 43% with treated hypertension. Neither the incidence of NOD (18.1% vs. 19.5%), nor the final fasting blood glucose differed between the groups, but fasting blood glucose was more likely to regress in those randomized to ramipril than in the placebo group (Table 1). The hyper-tensive subgroup had the same experience as did the group as a whole. The fact that these trials involved a large number of subjects provides assurance that the estimate is robust, and permits one to draw the conclusion that a meaningful, albeit small advantage in glycemic control could be expected if new antihypertensive agents were to replace older ones.

It would therefore appear that agents that block the renin system (and probably calcium channel blockers as well) nei-ther impair nor significantly improve glycemic control. The modestly higher incidence of NOD during antihypertensive drug treatment is most likely caused by the use of diuretics. The larger question is, however, whether the chlorthalidone-induced rise in NOD also increases CVD incidence.

CVd aNd oTher CoNsequeNCes of NodA major concern is that drug-induced NOD amplifies the CVD risk in the hypertensive population that is already at

11.69.8

8.1

0.0

10.0

20.0

30.0

Dia

bete

s in

cide

nce

(%)

Chlor Amlod Lisin

**

figure 1 | Diabetes incidence—4 years (follow-up fasting blood sugar ≥126 mg/dl for those with values of <126 mg/dl at baseline).9 *P < 0.05 compared to chlorthalidone (Chlor). amlod, amlodipine; Lisin, lisinopril.

Table 1 | hazard ratios for primary outcome in dream trial investigators study10

OutcomeRamipril

(N = 2,623)Placebo

(N = 2,646)Hazard ratio

(95% CI) P value

Diabetes 499 (17.1) 489 (18.5) 0.91 (0.80–1.03) 0.15

Diagnosed on the basis of fasting plasma glucose level and 2-hpost-load glucose level

375 (14.3) 411 (15.5) 0.91 (0.79–1.04)

Diagnosed by physician

74 (2.8) 78 (2.9) 0.95 (0.69–1.30)

Values are expressed as number (percentage).DREAM, Diabetes REduction Assessment with ramipril and rosiglitazone Medication.

at Uppsala U



Dow

nloaded from


AMERICAN JOURNAL OF HYPERTENSION | VOLUME 21 NUMBER 5 | May 2008 495

New Onset Diabetes During Antihypertensive Therapy STATE OF THE ART

increased risk. This possibility was first raised in epidemio-logical studies. A systematic follow-up of 6,886 treated hyper-tensive subjects showed a link between NOD and subsequent CVD events.11 The mean duration of follow-up was 6.3 years, with some subjects being studied for almost 20 years. Blood pressure during treatment averaged 138/83 mm Hg, and 70% of the patients maintained pressures <140/90 mm Hg. The risk of major CVD morbidity or mortality in patients with a history of diabetes was twofold greater than in those with-out (hazard ratio (HR) = 2.37, 95% confidence interval (CI): 1.80–3.11). Stratification by hyperglycemic status as (i) never; (ii) baseline only; (iii) follow-up only; and (iv) baseline and follow-up revealed a step-wise increase in the incidence of NOD (Figure 2). The greater incidence of CVD in the pres-ence of NOD was significant only in a subgroup that received diuretics >90% of the time.11

Subsequently, analysis of a 6-year follow-up study of nearly 800 treated hypertensive Italian subjects revealed that the rel-ative risk of CVD for those with prevalent diabetes was 3.57 (95% CI: 1.65–7.73) and for those with NOD it was 2.92 (95% CI: 1.33–6.41), relative to those who had never had diabetes.12 The use of diuretics was associated with NOD, but only 12% of the subjects received diuretics alone.

Several more recent long-term studies have yielded results consistent with the hypothesis that NOD increases the risk of CVD.13,14 In an 18-year follow-up, among the >11,000 par-ticipants in The Multiple Risk Factor Intervention Trial, who had initially been free of diabetes and CVD, there were 1,846 deaths due to CVD (1,277 of which were due to coronary heart disease (CHD)).13 HRs for CVD and CHD were 50% higher for those with NOD than for those without. Almgren et al. later identified 143 cases of NOD in a 25–28-year follow-up of 754

hypertensive men.14 The incidence of stroke, heart attack, and total mortality were higher in those with NOD than in those without diabetes, but less than in those with prevalent diabe-tes (Figure 3). Finally, in 15,000 high-CVD-risk participants in the Valsartan Antihypertensive Long-Term Use Evaluation (VALUE) trial, stratified according to diabetes status at entry, CHD morbidity was greatest for those with established diabe-tes (HR = 2.20, CI: 0.95–2.49, P = 0.0008), but NOD subjects also had a significantly higher risk of CHD (HR = 1.43, CI: 1.16–1.77, P < 0.0001) than drug-treated subjects who did not develop NOD. Heart failure was particularly common among those with NOD.15

That NOD would increase CVD risk would be expected, given that diabetes is well known to increase CVD inci-dence in hypertension. Moreover, because CVD in diabetes is duration-dependent, and NOD duration falls somewhere between the duration for those who have never had diabetes and for those with prevalent diabetes, the increase in CVD inci-dence in NOD should be somewhere between the correspond-ing CVD increases in the other two categories. Aside from the modest increase in NOD provoked by diuretics, NOD during treatment probably reflects the natural influence of hyperten-sion on the occurrence of hyperglycemia. However, it was only in the VALUE trial that increased CVD was seen during the 4 years of study. It is surprising that vascular complications of diabetes appeared in so short a time frame. This shows that it is possible that NOD in treated hypertensive patients may actu-ally advance the occurrence of CVD. NOD did not, however, show association with increased CVD risk during the course of ALLHAT.

In part because of the concern about heightened CVD risk, testing for hyperglycemia is a routine feature of antihyper-tensive care. NOD is thereby readily identified, and compels

12.59

11

17.57

2.91 3.282.14

1.06

8.42

0

2

4

6

8

10

12

14

16

18

20

A B C D5,447 130 911 398

CVD

Non-CVD

Newerelevated

At baselineonly

Elevation only infollow-up

At baselineis follow-up

Rat

e pe

r 1,

000

pers

on-y

ears

Blood sugar elevation (>7.75 mmol/l) status

Group

figure 2 | age- and gender-adjusted cardiovascular disease (CVD) and non-CVD incidence rates as related to blood sugar elevation in treated hypertensive patients. For CVD rate, *B vs. a: P = 0.041; **D vs. a: P = <0.001.11

100%a b

c

Str

oke

MI

Tot

al m

orta

lity

80%

60%

40%

20%

0%

100%

80%

60%

40%

20%

0%

100%

80%

60%

40%

20%

0%10 13

No diabetes n = 542New-onset diabetes n = 64Diabetes at entry n = 20

16

Years after entry Years after entry

19 22 25

10 13 16

Years after entry

19 22 25

10 13 16 19 22 25

• 754 patients• 25–28 years• 148 (20.4%) incident DM

figure 3 | Risk for (a) stroke, (b) myocardial infarction (MI), and (c) total mortality in hypertensive men as related to diabetes mellitus status.14

at Uppsala U



Dow

nloaded from




the attention of both patients and physicians. Disease labeling can adversely affect patient mood and behavior.16 Patients may be required to check their blood glucose with painful finger sticks. Physicians will invariably initiate further labo-ratory testing, both to confirm the diagnosis and to identify a potentially treatable cause—especially hypokalemia which is known to complicate diuretic use.17 There is good evi-dence that glucose metabolism is impaired by hypokalemia.18 Discontinuation of thiazide often returns the potassium bal-ance to normal and corrects hyperglycemia. In this way NOD can stimulate a change in antihypertensive therapy and this intervention may have its own potential hazards. In addition, NOD might call for hypoglycemic therapy to avoid the micro-vascular consequences of diabetes. These personal, clinical, and economic burdens of NOD have not been quantified, but must be considered when assessing the importance of NOD.

does drug seleCTioN alTer CVd ouTComes?Perhaps the critically important clinical question about NOD is whether its greater incidence in diuretic-treated subjects translates into increased CVD when compared with those who develop NOD during treatment with other antihypertensive agents.

ALLHAT, the largest randomized clinical trial to com-pare antihypertensive agents, assessed both the incidence of NOD and its CVD consequences.9 Because randomiza-tion to chlorthalidone was associated with increased NOD, one might have expected an adverse CVD effect. In the event, during 5 years of study, the primary endpoint and most of the secondary endpoints, as well as blood pressure, were actually either less frequent in the group randomized to the diuretic, or indistinguishable from the other groups. Of course, failure to detect any CVD difference in a rather short-term clinical trial does not rule out the possibility that harm would ultimately emerge. It is also possible that there were adverse effects, but in numbers insufficient to make a qualitative change in the outcome, and that it merely reduced the positive effect of the diuretic. Even the latter seems unlikely, because subjects randomized to chlorthalidone and who developed NOD actually had a lower incidence of CHD than did either the lisinopril or the amlodipine subjects who developed NOD. Most convincing was the finding that a 10% increase in fasting blood glucose at year 2 of the study was associated with an increase in CHD and total CVD exclu-sively in subjects randomized to lisinopril—and not in those randomized to chlorthalidone (Figure 4).

It is also noteworthy that, in ALLHAT, randomization to doxazosin, an agent that has no adverse glycemic and other metabolic effects, was not associated with improved out-comes—quite the contrary.19 The doxazosin arm was termi-nated prematurely, so it could be that a decrease in CVD and

NOD risk were not detectable because of insufficient dura-tion of treatment and follow-up. These data suggest that the higher incidence of NOD in patients on chlorthalidone treat-ment may reflect a different mechanism of hyperglycemia than conventional diabetes.14 Barzilay et al. have suggested, but could not confirm, that the excess hyperglycemia associ-ated with chorthalidone is most likely caused by hypokalemia. Hypokalemia-induced hyperglycemia is not associated with decreased insulin sensitivity and/or hyperinsulinemia, and thereby reflects a pathophysiology different from conven-tional type 2 diabetes.14 Correction of hypokalemia can reverse hyperglycemia.18 While this has not specifically been tested in thiazide-induced NOD, it is reasonable to assume that it would apply. Most compelling were the results of a meta-analysis of 11 trials, involving >100,000 subjects, the results of which showed that ACE/ARB-treated subjects had no greater protection against CVD mortality than did those treated with diuretics or with CCBs20 (Table 2).

Total

CHD

1.02 (0.97–1.07)

1.01 (0.94–1.07)

0.99 (0.89–1.10)

1.09 (1.01–1.17)

1.00 (0.92–1.08)

1.01 (0.91–1.11)

1.02 (0.88–1.18)

0.95 (0.77–1.17)

1.02 (0.92–1.08)

0.99 (0.91–1.09)

1.02 (0.93–1.12)

1.08 (0.93–1.26)

1.07 (0.96–1.19)

1.03 (0.88–1.21)

1.08 (0.71–1.65)

1.18 (0.87–1.58)

1.00 (0.97–1.04)

0.99 (0.94–1.03)

1.00 (0.94–1.07)

1.06 (1.00–1.13)

1.01 (0.97–1.05)

0.99 (0.93–1.05)

1.01 (0.93–1.10)

1.07 (0.98–1.17)

0.5 1 2P = 0.36 for interaction

0.5 1 2

0.5 1 2

P = 0.89 for interaction

P = 0.60 for interaction

Total mortality

CCVD

ESRD




Stroke

Heart failure

Chlorthalidone

Amlodipine

Lisinopril

Total

Chlorthalidone

Amlodipine

Lisinopril

Total

Chlorthalidone

Amlodipine

Lisinopril

figure 4 | Hazard ratios (HRs) of a 10-mg/dl (0.560 mm) rise in fasting blood glucose at 2 years for subsequent CVD and renal disease (The antihypertensive and Lipid-Lowering Treatment to Prevent Heart attack Trial (aLLHaT)).19 CCVD, combined cardiovascular disease; CHD, coronary heart disease; ESRD, end-stage renal disease.

Table 2 | effect of aCeis and arBs on cardiovascular disease mortality7

Source No. of studies No. of subjects OR (95% CI)

all 11 109,052 0.96 (0.91–1.01)

aCEI 6 74,626 0.93 (0.81–1.06)

aRB 5 34,426 0.93 (0.81–1.06)

HTN 7 86,414 0.99 (0.93–1.06)

ACEI, angiotensin converting enzyme inhibitor; ARB, angiotensin II receptor blocker; HTN, hypertension; OR, odds ratio.

at Uppsala U



Dow

nloaded from




From these findings it is seen that clinical trials have failed to detect an adverse CVD consequence on account of NOD that is associated with diuretic use. While these short-term trials do not rule out the possibility of ultimate disadvantage, avail-able evidence suggests that apprehension about the potential occurrence of NOD is not sufficient reason to tilt the balance in the matter of selecting the initial antihypertensive agent.

It is not likely that clinical trials will be mounted to address the issue of long-term consequences. The best available guidance for action may be found in the 18-year mortality follow-up of the participants in SHEP.21 Although no informa-tion about the patients is available after the study terminated in 1988, there was complete and unbiased ascertainment of deaths through 2002 (14.3 years after the trial ended). Almost half of the “placebo” group had begun antihypertensive ther-apy before the end of the study, thus tending to minimize dif-ferences between the groups. CVD was more common in those with prevalent diabetes than in those with NOD. Nevertheless, among those who developed NOD, randomization to chlo-rthalidone led to better outcomes than those in the control subjects. Impressively, chlorthalidone-treated subjects who developed NOD had the same incidence of CVD events as did those without diabetes—control or treated (Figure 5)! This finding, although post hoc and the result of comparing groups not protected by randomization, suggests that NOD related

to diuretic treatment may not increase CVD, at least when diuretic antihypertensive therapy has been used. Perhaps blood pressure control can trump lack of hyperglycemic control. Of course, because SHEP involved an elderly population, it is fair to say that its findings might not apply to younger patients.

CVd preVeNTioN iN persoNs wiTh diaBeTesWhat do we know about CVD prevention in diabetes? The UKPDS study tested approaches to CVD prevention in type 2 diabetes.22 Although it was completed more than a decade ago, with treatment goals less aggressive than those that are common today, its principle findings remain unchallenged. Disappointingly, superior glycemic control, to a HgA1c level of ~7, did not significantly reduce cardiovascular events, although microvascular outcomes were improved (Table 3). In contrast, better blood pressure reduction, albeit to levels above those sought today (systolic 154 mm Hg), did reduce CVD events. The Hypertension Optimal Treatment (HOT) study, which included 1,500 subjects with diabetes, was designed to determine whether the setting of lower target levels for blood pressure would improve CVD outcomes.23 Overall, a lower BP did not improve CVD outcomes for the participants in the HOT study. However, the minor-ity of the participants with diabetes did realize a dramatic stepwise reduction in CVD at target blood pressures rang-ing from 90 to 85 to 80 mm Hg (Figure 6). Lipid-lowering therapy is another approach proven to provide cardioprotec-tion in patients with diabetes. The Heart Protection Study Collaborative Group carried out a 5-year randomized trial of statin therapy in 20,500 participants at high risk of CVD, with a wide range of LDL-C levels; 6,000 of the subjects were diabetic.24 A statin- induced drop of 38 mg/dl (1.0 mmol/l) in LDL-C was associated with a 27% fall in CVD morbidity and mortality both in those with diabetes and those without. These studies indicate that, for patients with diabetes, as for all others, it is control of conventional risk factors rather than glycemic control, that best prevents CVD.

20

27

18

30

22

0

10

20

30

40

B-L DM

Placebo

ActiveF-U DM NO DM

*

** 18

figure 5 | Cardiovascular-related death (%) in the SHEP (Systolic Hypertension in Elderly Program) follow-up study.21 B-L DM, diabetes at baseline; F-U DM, diabetes at follow-up; NO DM, normoglycemic throughout.

Table 3 | uKpds: relative risks for endpoints by intensive and conventional glycemic treatment22

Aggregate Endpoint

Patients with clinical endpointsAbsolute risk events per 1,000

patient years

Log-rank (P) RR for intensive policy (CI)aIntensive

(n = 2,729)Conventional

(n = 1,138) Intensive Conventional

any diabetes related endpoint 963 438 409 460 0029 0.88 (0.79–0.99)

Diabetes related deaths 285 129 104 115 034 0.90 (0.73–1.11)

all cause mortality 489 213 175 189 044 0.94 (0.80–1.10)

Myocardial infarction 387 186 147 174 0062 0.84 (0.71–1.00)

Stroke 148 55 56 50 052 0.65 (0.81–1.51)aRelative risk (RR) and confidence interval (CI) = 95%.

at Uppsala U



Dow

nloaded from




CoNClusioNs aNd CliNiCal impliCaTioNsNOD is common among hypertensive patients, whether they are being treated for hypertension or not, and it is more likely to occur with the use of diuretics or β-blockers than with drugs that block the renin–angiotensin system and with calcium channel blockers. This probably reflects the fact that diuretic-induced hypokalemia causes hyperglycemia, and the other drug classes do not have major effects on glucose homeostasis. NOD is associated with increased CVD risk in hypertensive persons, but there is no evidence, in short-term clinical trials, that excess NOD in diuretic-treated subjects has adverse CVD consequences. It has also been found, in a long-term observa-tional study, that antihypertensive treatment with a diuretic actually eliminates the increase in CVD that would be expected in association with NOD. In fact, there is substantial evidence to indicate that blood pressure control (and lipid lowering) are the best means of preventing CVD in subjects with diabetes, whatever the circumstances in which the diabetes occurs. In summary, there is currently insufficient evidence to conclude that the potential occurrence of NOD should influence antihy-pertensive treatment choices. Clinical trials have consistently failed to show that newer agents provide cardioprotection superior to diuretics in patients with or without diabetes.

That notwithstanding, it is also true that NOD is an unwel-come outcome of antihypertensive treatment. It frightens patients and probably generates microvascular complications. Although the magnitude of these consequences has not been quantified, they are likely to justify hypoglycemic therapy and thereby bring on the side effects (e.g., hypoglycemia) related to these agents. In short, NOD is a burden for patients as well as for the health care system.

While apprehension about the potential occurrence of NOD should not determine initial antihypertensive therapy,

its appearance will inevitably have therapeutic consequences. β-Blockers are frequently associated with NOD. When indi-cated for cardiac care, they must be maintained, and appro-priate glycemic control efforts must be added. However, when not otherwise essential, blood pressure control can probably be achieved without a β-blocker. Thiazide-like diuretics are different. NOD is fairly common over the long course of diuretic use, and the discontinuation of diuretics may reverse the hyperglycemia. However, a diuretic is often an important component of optimal blood pressure treatment. If this is the case, either a lower dose, or a potassium sparing diuretic, and/or potassium supplementation, may correct the hyper-glycemia. It is also sometimes possible to correct NOD by adding a potassium-conserving antihypertensive agent such as an ACEI, ARB, or an anti-aldosterone agent, along with the diuretic. This may mitigate hypokalemia, correct hypergly-cemia, and permit reduction of the diuretic dose, without com-promising blood pressure control. If all these stratagems fail (or are unsuitable), and a diuretic is essential for blood pressure control, then appropriate management of hyperglycemia may be initiated, keeping in mind that even the non-cardiovascular consequences of hyperglycemia are not to be accepted lightly.

Disclosure: The author declared no conflict of interest.

1. Geiss LS, Herman WH, Smith PJ. Mortality in non-insulin dependent diabetes. In: Diabetes in America, 2nd edn, National Institutes of Health: Bethesda, MD, 1995, pp 233–257.

2. Fang J, Alderman MH. Impact of increasing burden of diabetes on acute myocardial infarction in New York City 1990–2000. Diabetes 2006; 55:768–773.

3. Centers for Disease Control and Prevention (CDC). Prevalence of self-reported cardiovascular disease among persons aged ≥35 years with diabetes—United States 1977–2005. MMWR Morb Mortal Wkly Rep 2007; 56:1129–1132.

4. SHEP Cooperative Research Group Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension: final results of the Systolic Hypertension in the Elderly Program (SHEP). JAMA 1991; 265:3255–3264.

5. Hansson L, Hedner T, Lund-Johansen P, Kjeldsen SE, Lindholm LH, Syvertsen JO, Lanke J, de Faire U, Dahlöf B, Karlberg BE. Randomised trial of effects of calcium antagonists compared with diuretics and β-blockers on cardiovascular morbidity and mortality in hypertension: the Nordic Diltiazem (NORDIL) study. Lancet 2000; 356:359–365.

6. Dahlöf B, Devereux RB, Kjeldsen SE, Julius S, Beevers G, de Faire U, Fyhrquist F, Ibsen H, Kristiansson K, Lederballe-Pedersen O, Lindholm LH, Nieminen MS, Omvik P, Oparil S, Wedel H. Cardiovascular morbidity and mortality in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE): a randomized trial against atenolol. Lancet 2002; 359:995–1003.

7. Gillespie EL, White CM, Kardas M, Lindberg M, Coaleman CI. The impact of ACE inhibitors or angiotensin ii type 1 receptor blockers on the development of new-onset type 2 diabetes. Diabetes Care 2005; 28:2261–2266.

8. The ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group. Major outcomes in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calcium channel blocker vs diuretic. The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). JAMA 2002; 288:2981–2997.

9. Barzilay JI, Davis BR, Cutler JA, Pressel SL, Whelton PK, Basile J, Margolis KL, Ong ST, Sadler LS, Summerson J. Fasting glucose levels and incident diabetes mellitus in older nondiabetic adults randomized to receive 3 different classes of antihypertensive treatment. Arch Intern Med 2006; 166:2191–2201.

10. The DREAM Trial Investigators. Effect of Ramipril on the incidence of diabetes. N Engl J Med 2006; 355:1551–1562.

9.9 10 9.3

24.4*

*

18.6

11.9

≤90 ≤85

Non-diabetic Diabetic

≤80 ≤90 ≤85 ≤80

85. 83. 81.10

5

10

15

20

25

Eve

nts

per

1,00

0 pt

yea

rs

Goal diastolic diabetic (mm Hg)

figure 6 | Major cardiovascular event rate in HOT (Hypertension Optimal Treatment).23 *P = 0.005 51% risk reduction.

at Uppsala U



Dow

nloaded from




11. Alderman MH, Cohen H, Madhavan S. Diabetes and cardiovascular events in hypertensive patients. Hypertension 1999; 33:1130–1134.

12. Verdecchia P, Reboldi G, Angeli F, Borgioni C, Gattobigio R, Filippucci L, Norgiolini S, Bracco C, Porcellati C. Adverse prognostic significance of new diabetes in treated hypertensive subjects. Hypertension 2004; 43:963–969.

13. Eberly LE, Cohen JD, Prineas R, Yang L; Intervention Trial Research group. Impact of incident diabetes and incident nonfatal cardiovascular disease on 18-year mortality: the multiple risk factor intervention trial experience. Diabetes Care 2003; 26:848–854.

14. Almgren T, Wilhelmsen L, Samuelsson O, Himmelmann A, Rosengren A, Andersson OK. Diabetes in treated hypertension is common and carries a high cardiovascular risk: results from a 28-year follow-up. J Hypertens 2007; 25:1311–1317.

15. Aksnes TA, Kjeldsen SE, Rostrup M, Omvik P, Hua TA, Julius S. Impact of new-onset diabetes mellitus on cardiac outcomes in the Valsartan Antihypertensive Long-Term Use Evaluation (VALUE) trial population. Hypertension 2007; 50:467–473.

16. Alderman MH, Lamport B. Labelling of hypertensives: a review of the data. J Clin Epidemiol 1990; 43:195–200.

17. Zillich AJ, Garg J, Basu S, Bakris GL, Carter BL. Thiazide diuretics, potassium, and the development of diabetes: a quantitative review. Hypertension 2006; 48:219–224.

18. Rowe JW, Tobin JD, Rosa RM, Andres R. Effect of experimental potassium deficiency on glucose and insulin metabolism. Metabolism 1980; 29:498–502.

19. ALLHAT Collaborative Research Group. Diuretic versus α-blocker as first-step antihypertensive therapy: final results from The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). Hypertension 2003; 43:239–246.

20. Blood Pressure Lowering Treatment Trialists’ Collaboration. Effects of different blood-pressure-lowering regimens on major cardiovascular events: results of prospectively-designed overviews of randomised trials. Lancet 2003; 362:1527–1535.

21. Kostis JB, Wilson AC, Freudenberger RS, Cosgrove NM, Pressel SL, Davis BR; SHEP Collaborative Research Group. Long-term effect of diuretic-based therapy on fatal outcomes in subjects with isolated systolic hypertension with and without diabetes. Am J Cardiol 2005; 95:29–35.

22. UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 1998; 352:837–853.

23. Hansson L, Zanchetti A, Carruthers SG, Dahlöf B, Elmfeldt D, Julius S, Menard J, Rahn KH, Wedel H, Westerling S. Effects of intensive blood-pressure lowering and low-dose aspirin in patients with hypertension: principal results of the Hypertension Optimal Treatment (HOT) randomized trial. Lancet 1998; 351:1755–1762.

24. Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20536 high-risk individuals: a randomized placebo-controlled trial. Lancet 2002; 360:7–22.

at Uppsala U



Dow

nloaded from


new onset diabetes during antihypertensive therapy

Documents