Challenges and Risks in Attaining the Systolic Blood Pressure Goal of <130 mm Hg in All Diabetic Patients

In their paper, “Attempted Forced Titration of Blood Pressure to <130/85 mm Hg in Type 2 Diabetic Hypertensive Patients in Clinical Practice: The Diastolic Cost,” Osher and colleagues ¹ have resurrected an old issue in an important new context. The old issue is the J‐curve effect, in which a diastolic blood pressure (BP) of <85, <70, or <55 mm Hg in treated patients is purportedly associated with increased risk of coronary heart disease and/or overall cardiovascular (CV) disease. ² , ³ The novel context raised is the potential risk of a diastolic BP <70 mm Hg in diabetic hypertensives when attempting to achieve the recommended systolic BP goal of <130 mm Hg. ⁴ , ⁵ , ⁶ , ⁷ , ⁸ The concern about diastolic BP <70 appears to be based largely on reanalysis of SHEP data by Somes, et al., ³ who reported that a diastolic BP of 65–69 mm Hg was associated with a marginally significant increase in CV risk. In that report, diastolic BP of 60–64 mm Hg was not associated with heightened CV risk, whereas readings <60 were more consistently associated with excess risk.

The investigators confirm the difficulty in achieving the systolic BP target in diabetics, as only 33% reached the goal of <130 mm Hg. ¹ Simultaneously, they are to be commended for achieving a mean systolic BP of 132±1 mm Hg in the entire group of 257 diabetic hypertensives, which surpasses the 138±1 mm Hg we reported among 19,864 diabetic hypertensives in our regional primary care network. ⁹ Moreover, their patients have obtained substantially better systolic BP control than diabetic volunteers in several clinical trials (Table), ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ with the exception of the intensively treated group in the Steno study. ¹⁴

Table.

Summary of Selected Trials in Diabetic Hypertensive Patients

Study	N	BPi (mm Hg)	BPf (mm Hg)	Within‐GroupΔ BP* (mm Hg)	Between‐GroupΔ BP** (mm Hg)	Δ CVD (%)
HOT 12	1501	170/105	140/81	−30/−24	−4/−4	−51†
UKPDS 13	1148	160/94	144/87	−16/−7	−10/−5	−33††
Steno 14	160	146/85	132/77	−14/−8	−14/−8	−53‡
SHEP 11	583	170/77	143/68	−27/−9	−10/−2	−34
Syst‐Eur 16	292	175/84	153/77	−24/−7	−9/−4	−69
Syst‐China 17	98	172/86	150/79	−22/−7	−6/−5	−74
BP=blood pressure; BPi=baseline BP; BPf=on‐treatment BP, which in some of the trials (HOT, SHEP) represents a best estimate from data provided in the various publications cited; Δ=change; CVD=cardiovascular disease; HOT=Hypertension Optimal Treatment; UKPDS=United Kingdom Prospective Diabetes Study; SHEP=Systolic Hypertension in the Elderly Program; Syst‐Eur=Systolic Hypertension in Europe; Syst‐China=Systolic Hypertension in China; *vs. baseline in treated group; **=vs. placebo, untreated, or less intensively treated group; †patients assigned to diastolic BP <80 vs. <90 mm Hg group; ††vs. usual care (less intensively treated) group; ‡patients who had more intensive treatment and better control of lipids and glucose in addition to BP. Adapted from Arch Int Med. 2000;160:2447–2452. 10

Osher and colleagues ¹ emphasize the potential risks to the 21% of diabetics with on‐treatment diastolic BP <70 mm Hg. ¹ , ³ In this group, the mean treated BP was 128±2/60±1 mm Hg, while the initial BP was 166/86 mm Hg. Thus, “forced titration” reduced BP ≅38/28 mm Hg. It is of interest to compare this group to a subset of hypertensives in the Individual Data Analysis of Antihypertensive Intervention Trials (INDANA) dataset ¹⁸ with a greater‐than‐median fall in systolic BP and an achieved diastolic BP below the 10th percentile. These patients were compared with controls with a similar CV risk profile (571 pairs) in INDANA. In this comparison, treated patients attained pressures of ≅124/62 mm Hg, compared with values of ≅154/84 mm Hg in the controls, a difference of ≅30/22 mm Hg. In this analysis, effective treatment reduced total mortality 54%, CV mortality 66%, all CV events 41%, and stroke 65% (all p<0.02). Myocardial infarction, a particular concern raised by proponents of the J‐curve, was reduced by 14% in the treated group, which, however was not statistically significant (p=0.61).

Findings from INDANA, which includes SHEP, ¹⁸ appear to conflict with reanalysis of SHEP. ³ In the latter report, ³ patients with a treatment‐induced reduction of 5 mm Hg diastolic had significantly more CV events. The differences may be explained, in part, by patient selection. The SHEP reanalysis included all patients, whereas INDANA examined only patients with a greater‐than‐median decline in systolic BP and diastolic BP below the 10th percentile (approximately 62 mm Hg).

Concerns about adverse CV effects from excessive reduction in diastolic BP with aggressive treatment of systolic BP in diabetics are mainly directed at patients with isolated systolic hypertension. As a group, these diabetic patients experience substantially larger reductions in systolic than diastolic BP with treatment (Table). The INDANA analysis, which included nondiabetic patients, suggests that groups with larger‐than‐median reductions in systolic BP derive impressive CV benefits from treatment, even when diastolic BP falls below the 10th percentile. ¹⁸

Low diastolic BP may impair organ perfusion, especially in elderly patients with isolated systolic hypertension who have a high probability of vascular remodeling and heart disease, which places them at high risk for overt CV disease independently of treatment. ¹ , ¹⁹ , ²⁰ With regard to the heart, the presumed reduction in oxygen supply due to a low diastolic BP is likely offset by a decline in oxygen demand in patients with a greater than median reduction in systolic pressure.

Diabetics with the lowest diastolic BP (<70 mm Hg, 21% of total) in the paper by Osher et al. ¹ had a substantially greater‐than‐mean decline in systolic BP (38 vs. 27 mm Hg). They likely derived CV benefit, based on extrapolations from the experience of patients with roughly comparable BP readings in the INDANA dataset, as discussed above. ¹⁸

At this time, there are no definitive answers about the risk‐benefit ratio of lowering diastolic BP to <70 or <60 mm Hg in diabetic patients, as may occur with forced titration of antihypertensive medications to achieve a systolic target. The ongoing Action to Control Cardiovascular Risk in Diabetes (ACCORD) study is comparing outcomes in diabetic hypertensives with a systolic BP goal of <140 vs. <120 mm Hg, based on clinical epidemiologic evidence that the lower readings are linked to better outcomes. ²¹ , ²² The results of ACCORD should better illuminate the risks and benefits of treating to more aggressive targets. In the interim, I recommend treating hypertensive diabetic patients to an initial goal of <140/80 mm Hg, based on trial results summarized in the Table. The BP goal of <140/80 mm Hg for diabetic patients coincides with recommendations of the British Hypertension Society. ²³ If diastolic BP remains ≥60 mm Hg, a systolic goal of <130 mm Hg is a reasonable target consistent with several guidelines.