Share this article on:

What really matters is not what the authors meant to say, it is how their work is read

Zanchetti, Alberto

doi: 10.1097/HJH.0b013e3283635290
Editor's Corner

Istituto Auxologico Italiano and Centro Interuniversitario di Fisiologia Clinica e Ipertensione, Università di Milano, Milan, Italy

Correspondence to Professor Alberto Zanchetti, Centro di Fisiologia Clinica e Ipertensione, Università di Milano, Via F. Sforza, 35, 20122 Milano, Milan, Italy. Tel: +39 02 50320484; e-mail:

The title of this editorial is borrowed from a New York Times article by the famous economist, Paul Krugman, commenting on the correlation between national debt and economic growth. The point raised by Paul Krugman is one of general application in science. Indeed, he points out that ‘a negative correlation between debt and economic performance need not mean that high debt causes low growth. It could just as easily be the other way around with poor economic performance leading to high debt’. He also remarks that, although the economists who originally described the correlation claimed they never asserted a causative relation between debt and depression, this was the way how the relation was read by policy makers in most countries in the world, and particularly in Europe.

Discussing whether the widespread austerity policies fathered by this correlation are sound or are worsening economic depression is out of the scope of your editor's competence and of the scientific interest of this journal. However, a point of broader relevance is whether correlative, cross-sectional studies, so frequent in medical research, should still have a place and suitable space in a reputable medical journal, or whether longitudinal prospective studies and, especially, randomized interventional studies deserve attention. The answer is that all scientific approaches have their own intrinsic values and limitations, and although the level of the evidence they provide is different (being greatest in randomized intervention trials), what really matters is how their results are read. Scientific results need to be correctly read but this need becomes increasingly difficult because of the ever mounting mass of scientific publications, and the authority of an article is often wrongly assumed from the authority of the journal where the article is published. Because of these problems, more than 1 year ago the Journal of Hypertension decided to publish brief summaries of the Reviewers’ opinion on the strengths and limitations of each published article at the end of the article, and for a number of years editorial commentaries have been invited, and still are invited, to discuss some of the most relevant issues raised by published articles.

Also, the current issue of the Journal of Hypertension publishes the results of a number of correlative cross-sectional, longitudinal, and interventional studies, as well as invited commentaries on the pros and cons of these various approaches.

Among cross-sectional studies, a systematic review and meta-analysis by Beauchet et al. (pp. 1502–1516) finds a significant association of high blood pressure levels with regional brain volume, particularly in key areas involved in cognition, such as the fronto-temporal lobes and the hippocampus. In the evaluation of the cardiovascular risk profile, Salah Mansour et al. (pp. 1584–1592) provide correlative evidence that aortic pulse wave velocity is a strong marker of cardiovascular disease and adds significantly to the explicative Framingham model. In the same area of cardiovascular risk assessment, Lind (pp. 1570–1574) reports that both endothelium-dependent and endothelium-independent vasodilatations are related to the Framingham score, thus casting doubt on the assumption that it is endothelial dysfunction rather than impaired vasodilatation per se that is a marker of vascular damage. Morillas et al. (pp. 1611–1617) show circulating levels of two biomarkers of collagen metabolism are proportionally higher the greater the number of organs involved in hypertension-related damage. In a study by Tatasciore et al. (pp. 1653–1661), early depressed left ventricular function was found to be associated with higher awake and 24-h SBP variability. Another correlative cross-sectional study by Sjöberg et al. (pp. 1547–1553) found that several persistent organic pollutants were related to abnormal left ventricular geometry, but all except one (hexachlorobenzene) lost significance following adjustment for established risk factors, thus indicating, if ever necessary, the careful and cautious attitude investigators must have in cross-sectional analyses. Finally, an epidemiological survey by Yip et al. (pp. 1539–1546) has investigated prevalence, awareness, and treatment of hypertension in two ethnically similar populations living in different geographic regions: Indians living in Singapore or in rural India. The much higher awareness and treatment among Indians in Singapore can likely be attributed to their better socio-economic setting.

There is a number of interesting longitudinal studies in this issue of the journal. Sundström et al. (pp. 1603–1610) report data from 34 009 consecutive patients with type-2 diabetes, followed during up to 11 years, and have correlated death incidence with blood pressure achieved during follow-up, finding a U-shaped relation with a nadir at 135–139/74–76 mmHg. In his editorial comment, P. Sleight (pp. 1527–1528) compares the results of this prospective observational study with those that were previously obtained by analyzing data provided by randomized interventional studies (in which original randomization is lost, however), and remarks the new data may be more representative of the patients’ population being derived from primary care rather than hospitals. He concludes, however, that the important issue about the optimal blood pressure target for antihypertensive therapy can only be solved correctly by a trial in which patients are randomized to different targets.

Another important article, adding data from a cohort of over 1.2 million individuals to what had been previously learnt from randomized trials, is published by Rao et al. (pp. 1669–1675): in this large nationwide cohort of United States Veterans no evidence was found to support any concern of increased lung cancer among users of angiotensin receptor blockers compared with nonusers, the authors’ findings being consistent with a protective effect, instead. In his editorial comment, K.K. Teo (pp. 1532–1534), who recently authored a meta-analysis of almost all trials employing an angiotensin receptor antagonist in one arm searching for cancer incidence [1], remarks the close similarity of the results of the two approaches, and concludes that the two studies can reassure individuals receiving these agents about their safety.

Prospective studies of antihypertensive therapy in patients with signs of organ damage are likely to shed light on the natural history or the ‘continuum’ of hypertensive cardiovascular disease. A study published by Rodilla et al. (pp. 1683–1691) clearly shows that regression of microalbuminuria and left ventricular hypertrophy is strongly related to blood pressure reduction, and microalbuminuria regression significantly increases the probability of left ventricular hypertrophy regression. Wachtell and Okin (pp. 1535–1536), in their commentary, remind us of some related observations in the Losartan Intervention For Endpoint Reduction in Hypertensives study, in which reductions of in-treatment albuminuria and left ventricular hypertrophy were associated with decreased risk of cardiovascular events independently of each other, thus suggesting regression of each organ damage sign independently contributes to improve prognosis [2]. It is also likely that microalbuminuria is an earlier sign of vascular involvement, and its treatment-related regression occurs more promptly, thus heralding the favourable response to treatment of more slowly developing phenomena such as left ventricular hypertrophy regression.

Noguchi et al. (pp. 1593–1602) have investigated the predictive power of baseline and on-treatment clinic and home blood pressure in a cohort of hypertensive patients, half of whom had impaired glucose metabolism or diabetes. Home blood pressure was found capable of more accurately predicting cardiovascular events during the following years, but the small number of incident events requires these interesting observations to be confirmed in a larger study. In a small but stimulating study Graff et al. (pp. 1629–1636) have investigated sophisticated ways of measuring heart rate variability as predictors of short-term and long-term neurological outcome in patients having recently suffered a stroke. The conventional (linear) and complex (nonlinear) measures of heart rate variability used by Graff et al. are commented upon in an in-depth editorial by Avolio (pp. 1529–1531), who explains how chaos theory and fractals can now be used for a better assessment of autonomic function.

The other longitudinal studies published in the current issue of the Journal of Hypertension have investigated dietary aspects of hypertension. Tielemans et al. (pp. 1564–1569) have explored in the large Prevention of Renal and Vascular Endstage Disease cohort if protein intake, as assessed by urinary urea excretion, is associated with development of hypertension, reporting negative data. Maneton et al. (pp. 1554–1563) report discouraging observations in a prospective sample of French patients to whom dietary recommendations had been made: although patients with dyslipidemia or overweight were more prone to adopt more protective diets, hypertensive patients were found to maintain unhealthy patterns tending to perpetuate their disorder.

Results from prospective interventional studies are also reported in this journal issue. In one of these studies (Schwieler et al., pp. 1676–1682) an angiotensin-converting enzyme inhibitor and an angiotensin receptor blocker were compared for their ability to reduce stress-induced platelet activation, with negative results for both types of compound. Also inflammatory and endothelial function biomarkers were not influenced by either drug. In a randomized study in Ontario, the success of Canadian guidelines was compared with that of Simplified Treatment Intervention to Control Hypertension (STITCH-2)-care, a treatment algorithm using single pill combinations (including the combination calcium channel blocker/statin) in achieving targets for both hypertension and dyslipidemia. Although the STITCH-2 approach obtained a significantly greater reduction of systolic blood pressure, LDL-cholesterol reduction was only marginally greater than in patients treated according to guidelines (Dresser et al., pp. 1702–1713). Briasoulis and Bakris (pp. 1537–1538) remark that a better blood pressure control is anyway an important achievement, and is in line with the very high rate of blood pressure control obtained in Avoiding Cardiovascular Events in Combination Therapy in Patients Living with Systolic Hypertension study, the only single pill combination outcome trial [3].

There are other interesting clinical studies in this issue. Nilsson et al. (pp. 1517–1526) update their stimulating concept of early vascular ageing and illustrate how the concept can be used in clinical practice. Persson et al. (pp. 1646–1652) investigate how measurement of urinary renin and angiotensinogen may have an added diagnostic or prognostic role in type-2 diabetes. Verloop et al. (pp. 1662–1668), on the basis of a large experience with resistant hypertensive patients referred to their clinic for renal denervation, report that a high percentage of patients were excluded because of secondary hypertension, white-coat hypertension, or blood pressure below the currently advised thresholds. Agarwal and Weir (pp. 1692–1701) publish an individual level meta-analysis of trials associating either hydrochlorothiazide or amlodipine with the angiotensin receptor blocker olmesartan, using both clinic and ambulatory blood pressure measurements. Although both types of combination were found to reduce ambulatory blood pressure to the same extent, the olmesartan–amlodipine combination appeared to be significantly more effective on clinic blood pressure, thus suggesting stronger mitigation of the white-coat effect. Doubova et al. (pp. 1714–1723) report the use of electronic health record to evaluate quality of care for hypertensive patients in Mexico.

Finally, a series of experimental animal studies also appear in this issue. Shin et al. (pp. 1575–1583) have silenced ATP2B1, a candidate gene among those identified in recent genome-wide studies, in mice, and found a rise in blood pressure and an increased wall: lumen ratio in the mesenteric arteries. Cova et al. (pp. 1618–1628) describe enhanced neurogenesis in the brain of stroke prone spontaneously hypertensive rats after a few weeks of a high salt diet. Although this increased neurogenesis can be interpreted as an attempt to contrast early brain degeneration, with disease progression only newborn astrocytes can survive. Vavrinec et al. (pp. 1637–1645), studying Fawn-Hooded rat strains prone or resistant to heart failure, report the heart failure-resistant strain maintains renal artery myogenic vasoconstriction protecting the kidney.

Back to Top | Article Outline


Conflicts of interest

There are no conflicts of interest concerning this paper.

Back to Top | Article Outline


1. The ARB Trialists CollaborationEffects of telmisartan, irbesartan, valsartan, candesartan, and losartan on cancers in 15 trials enrolling 138 769 individuals. J Hypertens 2011; 29:623–635.
2. Olsen MH, Wachtell K, Ibsen H, Lindholm LH, Dahlof B, Devereux RB, et al. Reduction in albuminuria and in electrocardiographic left ventricular hypertrophy independently improve prognosis in hypertension. J Hypertens 2006; 21:775–781.
3. Jamerson K, Weber MA, Bakris GL, Dahlof B, Pitt B, Shi V, et al. Benazepril plus amlodipine or hydrochlorothiazide for hypertension in high-risk patients. N Engl J Med 2008; 359:2417–2428.
© 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins