We are currently updating our website, and will have our new version online soon. Please check back later this fall.

Mailing List

Subscribe to the KT Canada mailing list




Once you have signed up, you will receive a confirmation email with your username and password. To activate your account, follow the instructions in the email.


Centre for Evidence-
Based Medicine

Hypertension - Therapy with diuretics and beta blockers decreases the risk of stroke

Clinical Bottom Line

Treatment of isolated systolic hypertension with diuretics and/or beta blockers decreases the risk of stroke in the elderly.


SHEP Co-operative Research Group. Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension. JAMA 1991;265:3255-64.

Clinical Question

In a patient with hypertension, will therapy with diuretics decrease the risk of stroke and death?

Search Terms

"stroke" and "hypertension" in Best Evidence

The Study

  • Double-blinded concealed randomised controlled trial with intention-to-treat.
  • patients = 60 years with isolated systolic hypertension
  • Control group (N = 2371; 2371 analysed): placebos
  • Experimental group (N = 2365; 2365 analysed):
    4 consecutive steps
    • chlorthalidone12.5 mg/d;
    • increase to chlorthalidone 25 mg/d;
    • addition of atenolol 25mg/d or reserpine 0.05 mg/d;
    • increase to atenolol 50mg/d or reserpine 0.1 mg/d.
  • Goal of therapy was SBP < 160 mm Hg and a reduction in SBP of at least 20 mmHg. Any patient with sustained SBP > 220 mmHg or DBP > 90 mmHg was given active treatment

The Evidence

Outcome Time to Outcome CER EER RRR ARR NNT
Stroke 5 years 0.067 0.044 34% 0.023 43
95% Confidence Intervals 15% to 54% 0.107 to 0.036 28 to 100
Deaths 5 years 0.090 0.102 -13% -0.012 -83
95% Confidence Intervals -32% to 5% -0.029 to 0.005 NNT = 209 to INF
NNH = 35 to INF


  1. decreased risk of nonfatal MI or coronary heart disease death with treatment (RR 0.73 [0.57 to 0.94])
  2. only enrolled 1% of the patients that were screened - included the very well elderly and therefore each clinician must assess each patient's individual baseline risk and individualise the NNT using this estimate
  3. at 5 years, 90% of the study group and 44% of controls were on active treatment and therefore the study likely underestimates the benefits of therapy since an intention to treat analysis was done
  4. was there some unblinding because of the need for some patients(more often the ones receiving diuretics) to receive potassium supplements

Appraised By

Straus, 1998

Expiry Date