Rationale Against the Drug Treatment of Marginal Diastolic Systemic Hypertension Edward D. Freis, MD i.z t-r&r' ih I c'i -4 F'J I T he Joint National Committee on Detection, Evalu- ation and Treatment of High Blood Pressure' does not recommend drug treatment for uncem- plicated hy~rtension when the diastolic blood pressure (BP) is in the range of 90 to 94 mm Hg (marginal hypertension). They advise instead nonpharmaceutical therapy with periodic examinations to detect possible progression of hypertension. Most physicians, however, still use 90 mm Hg as the level at which they will begin treatment with drugs if necessary and at least 2 authori- ties*,3 emphasize the importance of decreasing BP using drug treatment when needed in all hypertensive pa- tients, including those with diastolic levels as low as 90 mm Hg. This report will review some of the background data concerning the effectiveness of treatment of pa- tients with diastolic BP in the range of 90 to 94 mm Hg. The question is of considerable importance because pa- tients with diastolic BP in this marginal range comprise approximately 40% of the hypertensive population4*" Level of dfssfok blood pressure and coronary heart disease mortality: The leading cause of death in marginal hy~rtension is coronary heart disease, which is 3 times greater than stroke, the second leading cause. However, cardiovascular risk is lower in marginal hy- pertensive patients compared to patients with higher levels of diastolic BP.h The curve relating diastolic BP and coronary heart disease mortality increases steeply at levels above 100 mm Hgh Anderson7 using 1978 data from the Fra- mingham study, pointed out that the curve flattened out at diastolic BP levers between 70 and 90 mm Hg; that is, there was no increase in coronary heart disease mor- tality between 70 and 90 mm Hg. No such inflection or "dogleg" in the curve was seen with respect to systolic BP where coronary heart disease mortality increased continuously from the lowest levels. This inflection in the diastolic curve was not recognized before because of the previous practice of drawing a smooth linear regres- sion curve, which will never disclose a dogleg. This au- thor has been informed by Dr. A. D'Agastino of the Framingham study that their most recent, still unpub- lished data also show a dogleg or J-shaped diastolic curve. Furthermore, 4 other epidemiologic studiesR also show in the unsmoothed curves relating diastolic BP From the Veterans AdnIinistr~tion Medical Center and the Depart- ment of Medicine, CIcorgetown Ilniversity, Washington. DC. This study was supported by the Medical Research Service, Veterans Ad- ministration, Washington, DC. Manuscript received Jnnrrary 8, 1990. Revised manuscript received and accepted March 26, 1990. Address for reprints: Edward D. Frcis, MD, I lypertcnsion Clinic. Veterans Adtllinistr~tion Medical Center, 50 Irving Street, N.W., Washington, DC 20422. and coronary disease mortality an inflection in the range of 95 to 105 mm Hg diastolic BP (Figure I), The relatively flat portion of the dogleg in all of the studies included the range of marginal hy~rtension, suggesting that the risk of coronary heart disease is no greater in marginal hypertensive patients than in the normotensive population. Evidence for a dogleg or even J-shaped curve also has been reported for treated hypertensive patients. The inflection occurs at approximately 90 mm Hg with an increasing incidence of coronary heart disease deaths at progressively lower levels of diastolic BP.9,`o Cruick- shank'" reported that this phenomenon occurred only in patients with evidence of preexisting ischemic heart dis- ease. He suggests that in patients with narrowed coro- nary arteries, myocardiai perfusion, which occurs main- ly in diastole, may become critical at a diastolic BP of approximately 85 mm Hg.`a Other groups, however, have noted a J-shaped curver'.i3 in the absence of clini- cally evident coronary heart disease. One study in the elderly found a J-shaped curvei while another did not.15 Because the epidemiologic studies described before found J-shaped or dogleg curves, it is not surprising that the same phenomenon is seen when the diastolic BP is decreased to <90 mm Hg by treatment. Although final proof is still lacking, the recent evidence suggests that diastolic BP possibly should not be decreased to <90 mm fig. Effect of drug treatment on heart disease! mortality in marginal hypertension: Most multiclinic trials (Ta- ble I) indicate that antihypertensive drug treatment has not been effective in preventing coronary heart disease morbidity-mortality.`~ r* Opposed to 7 negative trials, only 2 found that treat- ment was effective in preventing coronary heart disease, the relatively small European Working Party on Hypcr- tension in the Elderly TrialIF and the lfypertension De- tection and Follow-up Program.?' However, the latter was not a controlled trial in the accepted sense because the general care of the control group was quite different from that of the treatment group. The 2 most definitive trials from the point of view of adequate numbers and appropriate design were the Australian'~ and Medical Research CounciIr9 (MRC) trials. Their overall results indicated no significant pro- tection against major coronary heart. disease events from the use of antihypertensive drugs. These negative results are supported by the 5 remaining trials listed in Table [.I7.fR.Zf' 22 Effect of drug treatment on stroke prevention: It is generally agreed that antihypertensive therapy de- creases stroke morbidity and mortality. 368 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 66 r TABLE I Effect of Treatment on Major Coronary Heart Disease Events: Summary of Trials Trial No of Entry DW Subjects (mm 4) Incidence CHD Events Treatnlent Conlrol % Dlfference Showing no or negative benefit VA 380 USPHS 389 Oslo 785 Australian 3.427 MRFIT 7,012 MPPCDM 1,222 MRC 17,354 Showmg benefit EWPHE 840 HDFP 7,825 9010 114 90to115 9oto109 95 to 109 9ot 95t 90 to 109 90to119 7 16 56 90 to 104 86 107 20 11 7 20 33 115 19 222 13 13 6 -17 13 -54 33 0" 124 5 9 -111 234 12 Includes eather mortahty atom? or morbldlly plus mortatlty when data for both are gwn ' Interlllorl to treat Ctill = coronary heart dtsease. DBP = dwtotlc blood pressure, EWPtiE = European WorktrIg Party II@ Blood Pressure ,,, the Elderly. HOF,' = Wypertwwn "e,e,.,,,,n ,x,I in 850. These results further call into question the cost-effec- tiveness of treatment of marginal hypertensive patients even for the prevention of stroke. Adverse effects and cost of drug treatment: Al- though most patients tolerate antihypertensive drugs quite well, moderate to severe adverse effects may occur in a few. In the Veterans Administration trial,*" for ex- ample, of 180 actively treated patients, 2 were with- drawn because of presumed toxic reactions, and side ef- fects such as lethargy, weakness and nasal stuffiness FIGURE 1. Coronary he.arl disease death rates/l,000 patient years for 12 years re- corded in 4 epidemiotogk shaties (Albany, Chicago Gas, Chkago Western Etectric and Tecumseh) related to diastolic blood pressure (DBP). Graphic representation of tabular data as presented et al? The 4 curves all showed a dogleg 4i6clion with a relatively ftat portion be- tween <70 and >95 mm Hg DBP. Mortal- ity rates then increased steeply beginning at levets of DBP ,100 mm Hg. DBP, therefore, correlated directly with rates of coronary heart disease only at kvds 2 100 mm Hg. Mortality rates were no higher in the marginal hypertensive range than in the normal range of DBP. were significantly more frequent in the treated patient as compared to the placebo control subjects. The MRC trial*' reported that in men, adverse effects with ben- droflumethiazide or propranolol were significantly in- creased as compared to placebo. The cost of lifelong treatment for such a large num- ber of patients is also a major consideration. This cost is magnified by the current trend to prescribe new, quite expensive drugs such as calcium antagonists and con- verting enzyme inhibitors in place of the less costly, old- er agents such as generic thiazide diuretics, propranolol and reserpine. Drug treatment for the 20 million mar- ginal hypertensive subjects could total, therefore, several billion dollars per year. The course of untreated marginal hypertension: In the Australian trial,26 1,943 placebo-treated patients had diastolic BP for the 3 initial visits averaging be- tween 95 and 109 mm 1 fg and remaining at 295 mm fig at the third visit. The investigators also followed, without active treatment, 325 patients whose diastolic I m--l 90-4 loo.4 llO+ <707wl 60-4 9&4 1004 1 0 ALl?ANY 0 CHICAGOGAS 0 CHICAGOW E 0 TECUMSEH DIASTOLIC BP mmHg er -- THE AMERICAN JOURNAL OF CARDIOLOGY AUGUST I. 1990 369 BP averaged <95 mm Hg during the 3 initial visits. After 3 years, 48% of the total untreated patients had reverted to normal BP without treatment. Many physicians regard 3 visits as being sufficient to characterize the average BP. The Australian trial indi- cated that the BP of some patients will require 14 months to decrease to normal. None of the patients with initial diastolic BP <95 mm Hg at baseline developed elevation >109 mm Hg over 3 years of follow-up, whereas 94 of the patients exhibiting baseline diastolic BP levels between 105 and 109 mm Hg increased to >109 mm Hg during the trial. Thus, marginal hypertensive subjects rarely exhibit se- vere acceleration of BP over 3 years, whereas moderate- ly severe hypertensive subjects often do. A lesser in- crease of 95 to 109 mm Hg occurred in some of the marginal patients (Table 111). However, these more moderate increases can be safely managed by C-month visits when they would be detected and treated before the BP reached a high-risk level. The placebo group who remained in the range of 90 to 94 mm Hg throughout the trial, that is, the untreated marginal hypertensive subjects, exhibited no significant- ly greater rate of trial end points than the placebo pa- tients who maintained normotensive levels of diastolic BP of 85 to 89 mm Hg (Table IV). These data further suggest the low risk associated with untreated marginal hypertension. By contrast, in the placebo patients main- taining higher levels of 95 to 99 mm Hg, the event rate per year nearly doubled as compared to the marginal hypertensive subjects and then increased more steeply with a diastolic BP >I00 mm Hg. At similar levels of diastolic BP, the treated patients in general had higher rates of trial end points than the placebo patients. Marginal hypertension combined with systolic hy- pertension: Systolic hypertension is defined as a systolic BP of 160 mm 1lg or higher. Although often associated, there are important differences between systolic and diastolic hypertension.27 The latter is characterized by constriction of arterioles while systolic hypertension re- sults from loss of compliance of the aorta causing a steep systolic increase during cardiac ejection. Diastolic BP may remain normal if the arterioles are not con- stricted. Systolic hypertension in the presence of normal or marginal diastolic hypertension often occurs in the aged. Under the stress of the first 1 or 2 medical exami- nations, these individuals may have an increased cardiac output, resulting in systolic hypertension. This response is amplified by impaired barorcceptor moderation of the BP elevation with aging. During subsequent visits, as elderly patients become accustomed to the clinic envi- ronment, the increased cardiac output often moderates and the systolic BP reverts to normal. Systolic hyperten- sion, however, may persist in many other patients. Epidemiologic studies indicate that systolic hyper- tension poses as great or greater a risk as diastolic hy- pertension.*s However, there are presently no data from controlled clinical trials for determining the effective- ness of drug treatment in isolated systolic hypertension. It is, therefore, uncertain whether treatment will be 370 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 66 TABLE II MRC Trial: Stroke Rates/l.000 Patient Years in Relation to Average Entry Diastolic Blood Pressure Average Fnlry DRP(mm l4p.j Strokes/ I.000 Pt Yrs <95 95 to99 100t0 104 105 In II-N __-. Active treatment 03 18 14 I4 Placebo 1.4 2.3 31 44 Difference ____~ .._._ ~~~~. TABLE Ill Australian Trial ~- Placebo Group: Avrragc Diastolic Blood Pressure After 3 Years Compared to Avernfle Diastolic Blood Pressure at Entry I - TABLE IV Data from Australian Trial-Trial End Points, Rates by Average Diastolic Blood Pressure Throughout the Study for Active and Placebo Subjects TEP Rates/ I .XXl Yrq Average DRP (mm Ilg) Throughout the Trial Active rlar+ <85 123 II 8 85 to 89 134 188 90 to 94 29 7 I6 : 95 lo99 75 8 78 7 ?I00 84 5 60 4 Imp - (,1.7~,011< hlnorl prnsrurr IEI- 7 ,,,a, Cllll p01nt', beneficial or at what level to begin. Some physicians initiate drug treatment of persistent systolic hyperten- sion on the reasonable but unproven assumption that such treatment will prevent stroke and heart fniltlre, which are both prevalent in aged hypertensive patients. Conclusions: Few patients with uncomplicated mar- ginal hypertension require drug treatment. Ilowever, other hygienic measures such as weight reduction if needed, salt and alcohol restriction, regular exercise. cessation of cigarette smoking and especially dietary re- striction of saturated fats and cholesterol are indicated. Also, antihypertensive drug treatment is advisable in marginal hypertensive subjects with diabetes mellitus or with renal impairment from other causes or congestive heart failure or with moderate to severe left ventricular hypertrophy or other evidence of organic changes secon- dary to hypertension. Aside from these exceptions. most of which are infrequent among marginal hypertensive subjects, there is little evidence that these patient? will achieve enough benefit to justify the costs and adverse effects of antihypertensive drug treatment. REFERENCES 1. Jom~ National Commirtee on Detection, Evaluation and Trealmenl of High Bhul Pressure. The 1988 Report. Arch lr~rerrr Afeed /988:/48:1023-1038. 2. Gifford RW. Essen!ial hypertension: cost-effective evaluation and Irealmenl. .A,,, J ,&fed /986.8l(su/~p/ C-C-23~ C-37. 3. hloser hl. Is drug trealment indicated for mild hyperlension with diaslolic blowl pressure 90 mm IIg 10 100 mm llg? An affirmative view. J Fonr fro0 1988.26 449 459. 4. IJnited Stales Deparrmenl of II&h and lluman Services. Public lleallh Swice lleahh, United States with prevenlion profile. W'arhington DC Gowm- nw~ Priufrng O//ice. 1981 35(/M///S prrblicorion no (P//S) RI-1232135 39. S. llyperrension Delectiorr and Follow-up Program Cooperative Group The ff)perccnsiorr Dekckn and Follow-up Program. A progress reprf. C7rc Kcr I977.4fJ(.rlrp/d I) 1-106-1-109 6. The Pooling Project Research Group. Relalionship of blood pressure, serum cholcshxd, smoking habit. &rive weight and ECG abnormaliries to incidence of major coronary events: final report of the poling project. J Chronic Dis 1478.3/:20/ ~306. 7. Anderson TW. Re-examination of some of Ihe Framingham blood pressure data. Lnncw 1978.2 1139.//4/. 8. hlcGee D, Gordon T. The results of rhe Framingham Srudy applied to four other II S -based epidemiological studies of cardiovascular disease. Section 3 I: 7 he I-ramingham Study: an epidemiological investigation of cardiovxculnr dis- use In: Kannel WR. Gordon T. eds. H'oshingron IX', US 0cpar1nwrrr of Ilrcrlrh, Edwarion and We//ore, I976 N/H Pub/k arion 76. I.083. 9. Stewnr~ JhlcDG Relation of reduction in pressure in tirs~ rnyocardinl infarc- tiun in patients receiving trealment for severe hypertension. Lancrr /979;/ 861. ,865 10. (`wickshank Jhl. Thorp JM. Zacharias FJ Benefits and potential harm of Iouering high blood pressure. Lancer /987./;58/-584. 1 I. Wailer PC, Isles CC;. Lever AF, Murray GD, hfclnnis GT. Does therapeulic rcduc~inn in blood pressure caux death from coronary artery disease? J /fwrru~ I!,pcw~~tsion /988,2.7 10. 12. t`lercher AE. Bceners DC. Bulpiu CJ. Burler A. Coles EC, tlunr D, Munro- I;lurc AD, News MR. O'Riordan PW, P&e JC. Rategoplan B, Rylance PD. Tuallin G. Webster J. Dollery CT. The relationship between a low lreated blood pres~orc and IIID mortality; a report from DIISS llyprtension Care Computing (;roup (Dt ICCP). J Ilurnnn Iiypertensiott 1988:2. I I IS. 13. Samuelson 0. Wilhehnson I. Anderson OK, Peunert K, Bergland G. Cardio- vaxxdar morlnlily in blwd pressure and strum choleslcrol levels in Ireared hyper- tension. Resulrs from the Primary Prevention Trial in Goteborg. 5,4AfA I987.258 1768-1776. 14. Coope J. Warrender TS. Lowering blood pressure. Lancer /987;/~/380. 15. Amery A. Birkenhager W. Brixko P, Bulpilt C. De Schaepdryver A, Dollery C, Fogard R, Foretle I', Forte J. tlandy R. Henry JF. Joossens JV. Lconetti G. Lund-Johansen P. O'Malley K. Petrie J. Strasser T. Tuomilehto J. Williams R. hlortaliry and morbidiLy results from European Working Parly on High Bled Pressure in the Elderly Trial Lower /985./:/349 1354. 16. Reprl by the Managemen Commirlee The Auslralian rherapeutic trial in mild hypertension lancer 1980;1./26/ 1267. 17. Velerans Administration Cooperative Sludy Group on Antihypcrtensive Agenls. Effecls of treatment on morbidily in hypertension. II Rewl~s of palients with diastolic blood pressure averaging 90 through I I4 mm Ilg J.4nfA /970;2/3.//43-l/S/. 18. Smith WM for Ihe U.S. Public flealth Service Ilospifals Coopralive Srudy Group. Treatmen of mild hypertension. Results of a ten-year inlervenlion trial Circ RCP /977,40(rup/~l f) l-98 /-/OS. 19. Medical Research Council Working Party On Mild I lypertcnsion. MRC trial of treatment of mild hyperwnsion, principal resulls f/r Mt,d J I985.291 97-/(J4. 20. Iielgeland A. Trealmenl of mild hyperlension: a tive-year coalrolled drug trial. The Oslo sludy. Ant J &fed /980.69.725- 732. 21. Miettinen TA. Ilurrunen JK. Naukkarinen V, Strandberg T. hlittilla S. Kumlin T. Sarna S. Multifxtorial primary prevenlion of cardiovascular diseases in middle-aged men. Risk facIor changes, incidence ;and mortalily JA,tf:l /985;254 2097 2102. 22. Multiple Risk f'aclor and Intervention Trial Research Group. Multiple Risk Faclor Intervention Trial Risk faclor changes and morralicy results J.4M 4 /982;248 1465 I477 23. llyperlension Detection and Follow-up Program Cooperative Group. five- year lindings of rhc Ilyprlension Dcbxtion and l~~,llow-up-program. I Reduc- tion in morlalily of persons with high blood pressure. including mild hypertension. J.4M.4 1979.242 2.562 2571. 24. Veterans Administration Cooperative Study Group on Anlihypertcnsive Agents. III. Influence of age. diastolic blood pressure and prior cardiovascular discase; further analysis of side effects Crrcrrlorion 1972.45 991 1004. 25. Medical Research Council Working Rrly on Mild to M&rate Ilypcrtcn- sion. Adverse reactions IO txndrolluazide and propranolol for Ihe rreatmenr of mild hypertension. ffirrcer 1981.2 539 543. 26. Managcmenl CornmilleeofIhe Auslralian Therapculic'l`rial in Mild llyper- lension. Untrcarcd mild hylxwcnsion. lower 1982.I 185 191. 27.Gifford RW. Isolated systolic hypertension in the elderly. Somccon~roversial issues. J/IMA 1982:247 781 785 28. Kannel WR Role of blood pressure in cardiovascular morbidity and mortali- my. Prog Cardirwasr DIS 1974.18.X 23. THE AMERICAN JOURNAL OF CARDIOLOGY AUGUST 1, 1990 371