Intensive blood pressure treatment

In 1987, Cruickshank et al. [1] maintained that a J-shaped curve was present when mortality was related to diastolic blood pressure in hypertensive patients on drug treatment. In the highest pressure levels, from approximately 120–80 mmHg, mortality fell with a fall in blood pressure, but below approximately 80 mmHg it increased again, producing a ‘hockey-stick’ or J-curve [2]. This paradoxical increase in risk at the lowest diastolic blood pressure levels was assumed to be due to the presence of diseased, sclerotic and partially stenotic arteries in these hypertensive patients. Too drastic a reduction in perfusion pressure by antihypertensive drugs caused a reduction in blood flow to vital areas, possibly resulting in myocardial infarction or a cerebral stroke. In particular, elderly hypertensives would be at risk for this unwanted effect of blood pressure reduction. In the following years, the J-shape curve was strongly debated, but appeared be completely refuted in 1991 by the results from the large SHEP Study (Systolic Hypertension in the Elderly Program) [3]. In this group of elderly subjects blood pressure was reduced from approximately 170/77 to 144/68 mmHg (a reduction of −26/−9 mmHg) without any J-shape in the pressure-mortality curve. A fall in mortality was observed throughout [4]. In 1997, the results from another trial in systolic hypertension in the elderly, the SYST-EUR trial [5], also appeared to contradict the J-shape curve concept. However, the discussion about the J-shape problem still continues [6], and a recent analysis of the SHEP [7] and results from a more recent observational study [8] have shown that when diastolic pressure is reduced < 65 mmHg, the risk of cardiovascular complication increases [6,9]. It should also be emphasized that some elderly subjects with particularly low pressure are ill from various diseases, and therefore have a high mortality rate [10], and this may contribute to the J-shape phenomenon. In this issue of the journal, the HOT-investigators [11] draw attention to another important hypertensive group where the J-shape curve is seen again, and within a reasonable range of blood pressure values, namely in smokers. While no J-shape phenomenon was seen in any subgroup in the non-smoking cohort treated to similar systolic and diastolic levels as the smokers, the smoking cohort had higher total mortality, cardiovascular mortality, sudden death and more strokes among those who belonged to the most intensive treatment group (diastolic pressure < 80–85 mmHg). Before addressing the results in the smoking cohort, which comprised only 16% of all subjects, the results from the large non-smoking cohort of 15 807 patients are discussed first. The non-smokers The new subanalysis showed that when the smokers were excluded, the best treatment results were achieved in the most intensively treated subjects. This trend was seen for most risk categories. The relative risk for major cardiovascular events became significant for diabetics, very high-risk subjects, patients with ischemic heart disease, in women and also in older subjects (> 65 years). All of these subgroups were fairly large. The apparent lack of any J-phenomenon in these many subgroups could be due to several factors. The HOT-population was not very old, with the mean age at the inclusion being approximately 60 years. Only approximately 10% had previously known cardiovascular disease, and approximately 48% had had previous antihypertensive treatment. The baseline blood pressures were only approximately 170/105 mmHg. Thus, it is reasonable to believe that a large part of this non-smoking cohort without previous ischemic heart disease did not have high pressure compartment vessels that were markedly sclerotic or narrowed, and thus not too vulnerable to a reduction in perfusion pressure. Furthermore, the achieved blood pressures after ≥ 6 months in the different treatment groups, even the most aggressive ones, were not excessively low: approximately 145–138/83–81 mmHg, and thus far above 65 mmHg diastolic where perfusion problems in vital areas might be a threat [6]. The particularly good results observed in the diabetic cohort fit well with observations from other large trials [9,12]. The smokers The results in the smoking cohort are puzzling in many respects, and some of them are not easy to explain. First, it should be stressed that in a previous HOT publication [13], it was clearly demonstrated that the highest total mortality was seen in smokers, the mortality being approximately twice as high as in non-smokers. Even during the short observation period of 3.8 years, the killing effect of smoking was clearly shown. The new subanalysis showed that the incidences of death and cardiovascular events per 1000 patient-years were much higher in the aggressive treatment groups. When the two treatment goal groups (intensive versus less intensive) were compared, the relative risk (RR) was increased in the most intensively treated group for total death, cardiovascular death, sudden death, major cardiovascular diseases, and stroke (RR of 2.03, 2.67, 2.93, 1.74 and 2.44, respectively). The baseline characteristics in the different treatment groups were superimposable. Another most puzzling observation was that the RR for myocardial infarction went in the opposite direction (RR 0.74). The investigators state that the type of antihypertensive drug did not seem to give different results when they compared monotreatment with felodipin versus felodipin plus an angiotensin-converting enzyme inhibitor or versus felodipin plus a diuretic. The effects of smoking on the high-pressure compartment vessels The authors cannot rule out that the observations in the smokers might represent a chance finding. This is possible, but seems unlikely, because the group of smokers was fairly large (approximately 3000 subjects) and, thus, approximately 1000 in the less effective treatment group and 2000 in the most intensively treated groups. Because the mean age was approximately 60 years, and usually few people start to smoke after the age of 20 years, their vascular system has most likely been submitted to the damaging effect of smoking for more than 40 years. Their high-pressure compartment vessels probably were grossly abnormal at baseline, although this was not clinically discovered. In general, atherosclerotic disease in the high-pressure compartment in smokers develops approximately 10 years earlier than in non-smokers, and the damage starts early [14]. In a large ongoing study in the USA, the so-called PDAY study, Zieske et al. [14,15] performed meticulous studies of the arteries in young individuals (aged 15–34 years) who had died in accidents. Even among teenagers, smokers had significantly more severe wall changes than non-smokers! It is well known that smoking accelerates atheromas in the carotid, intra-cerebral and coronary arteries. Obviously genetic factors are important but, in a unique study from Finland in homozygote twins with disconcordant smoking habits, the smoking twin had significantly more carotid artery lesions than the non-smoking twin [16]. In smokers, severely stenotic areas in the carotid arteries before the age of 50 years have been reported [17]. Smoking also amplifies the damaging effect of other risk factors, such as high cholesterol and high blood pressure [18]. Thus, in hypertensive subjects, smoking is particularly harmful [19]. In a consensus statement from the National Stroke Association about prevention of the first stroke, smoking is declared as being one of the most important risk factors [20]. It is well established that smoking induces pathological alterations in the thrombogenic mechanism, and that fibrinogen increases, platelet adhesiveness is increased and fibrinolysis is impaired [20]. Interestingly, some of the same type of abnormalities are seen in hypertensive patients [21]. Thus, the combination of hypertension and smoking could be particularly harmful for the high-pressure compartment vessels, making them more sensitive to reductions in perfusion pressure than in subjects with just moderate hypertension. Why the J-curve only versus stroke and not versus myocardial infarction? It is well known that smoking also damages the coronary circulation, and it is most intriguing why only stroke and not myocardial infarction was more frequent when blood pressure was lowered most effectively in the HOT smoking cohort. The following attempt to a possible explanation is admittedly speculative. It is well known that, in hypertensives, the autoregulation curve in the cerebral circulation is shifted to the right (i.e. cerebral blood flow falls when blood pressure is reduced below a certain level), which is higher in hypertensives than in normals [22]. It is possible that smoking could act on the cerebral vessels in similar way, and that the autoregulation curve could remain shifted to the right in spite of reduction in blood pressure. The results in the HOT study are particularly puzzling because a dihydropyridine (felodipine) was used as the first drug. Other studies on different calcium antagonists (diltiazem and nitrendipine) have shown that they are particularly effective in reducing the incidence of stroke [9,23]. In earlier studies (MRC and IPPPSH), mainly based on diuretics and/or beta-blockers, the worst treatment results were observed in the smoking cohort [24,25]. No harmful effects of smoking in the pressure target group < 90 mmHg? Still another puzzle in this new subanalysis is that, in the cohort with the highest blood pressure goal (≤ 90 mmHg), smokers and non-smokers both had similar incidences for myocardial infarction, stroke, cardiovascular disease and total mortality. Total mortality per 1000 patient-years tended to be slightly lower in the smokers than in the non-smokers! It would be expected that a substantial fraction of deaths in the smokers should have been due to non-cardiovascular causes, such as lung cancer and lung insufficiency, but this did not appear to happen. The authors state that the results in this treatment group were remarkably good because the rate of cardiovascular complications in the smokers was so low. It would indeed be most sensational if antihypertensive drugs given to smokers protected against non-cardiovascular diseases, but this would obviously have to be confirmed. Stopping smoking: easy to recommend, difficult to achieve The authors recommend that the smoking hypertensives should be persuaded to stop or referred to non-smoking programs [11]. Consequently, they might also be submitted to intensive blood pressure-lowering therapy. Even if it is unknown whether the J-shaped curve would then disappear, it is obvious from previous and recent data that stopping smoking improves health and increases lifespan, irrespective of blood pressure. Several recent studies provide support for this. Thus, in a 35-year follow-up study from Finland, smoking increased premature death in middle-aged men, and giving up smoking earlier in life prevented smoking attributable deaths and prolonged life [26]. In a 16-year follow-up study from Norway in 22 300 middle-aged men, current smoking increased mortality irrespectively of body mass index, while ex-smokers had a total mortality similar to never-smokers [27]. The earliest beneficial effect from stopping smoking is probably related to rapid normalization of the thrombogenic mechanisms, such as normalization of fibrinogen and platelet adhesiveness [19]. As previously mentioned, hypertension confers a prothrombotic state [21], and the combination of hypertension and smoking is particularly bad for several reasons [28]. Conclusions This recent subanalysis from the HOT presents several surprising and puzzling observations. However, in my opinion, the most important message concerning smoking from this large study is what the investigators already have presented in the report from 2001 [13]: that smokers have a twice as high total mortality compared to non-smokers. Every effort should be made to persuade smokers to stop, and help those who are so addicted that they need a firm stop-smoking program to have a chance of success. However, in spite of such efforts, the majority will not succeed and, in those who continue to smoke, it would appear reasonable to have a somewhat higher target level than for non-smokers with respect to antihypertensive treatment regime. For a long time, hypertension has been known as a ‘silent killer'. We should emphasize that smoking is also a ‘silent killer'. The medical profession should fight vigorously against both of them!