Beauty in Simplicity: Abnormal Neutrophil to Lymphocyte Ratio in Resistant Hypertension

Resistant hypertension has a well‐known association with increased cardiovascular risk.1, 2 While many of the underlying mechanisms for resistant hypertensions are well‐defined (eg, primary hyperaldosteronism and renal artery stenosis), the majority of resistant hypertension remains poorly understood. Thus, the attempt to debunk an “essential” cause of resistant hypertension is even more timely. Neutrophil to lymphocyte ratio (NLR) is an emerging risk marker in multiple diseases with a wealth of associations reported in the recent years. White blood cell count and absolute neutrophil count have been shown to predict new‐onset hypertension in a large Japanese cohort over a 40‐year follow‐up.3 Similar findings have been shown from our local population of the Atherosclerosis Risk in Communities (ARIC) cohort, especially for African American patients.4 The paper published in this current issue of the Journal by Belen and coworkers5 is striking, most of all, for what appears to be a potential weakness of the paper––its simplicity. It examines a cohort of 150 patients: 50 with normotension (NT), 50 with controlled hypertension (CHT), and 50 with resistant hypertension (RHT). The authors have applied vigorous exclusion criteria and performed meticulous work in ruling out potential secondary causes of RHT or major comorbidities potentially skewing NLR, such as diabetes mellitus, coronary artery disease, chronic kidney disease, peripheral vascular disease, and major heart or valvular disease, which would have confounded their findings. Therefore, only essential (ie, otherwise not well‐explained) cause(s) may have driven the difficult‐to‐control BP in the RHT group. In this regard, the authors presented three very well‐matched groups with regard to age, sex, and body mass index (BMI). Not surprisingly, and as a testimony of excellent matching along those parameters (a clear advantage over statistical procedures to “adjust” for these parameters), the patients appeared similar on many biochemical parameters, such as renal function and lipid studies. Nevertheless, an important distinction emerged between the three groups (NT, CHT, and RHT), where differences in NRL and neutrophil granulocyte counts were maintained after adjusting for confounding variables during logistic regression analysis. Further, it confirmed a graded relationship between the degree of difficulties to obtain control of hypertension and the elevation of NLR. Ratios, however, are known to potentially exaggerate differences. In this regard it is worthy to emphasize that neutrophil granulocyte count alone showed a significant difference between the NT, CHT and RHT cohorts (P<.001, during multivariate logistic regression). While the findings of the paper are thought‐provoking, it underlines the need for further studies and virtually begs for a more detailed and pathophysiologically oriented explanation. Even the all‐encompassing term of endothelial dysfunction may just be, at least in part, an epiphenomenon of a larger, more systemic process of the inflammatory state, yet poorly understood. What Belen and coworkers have achieved in their work is to point out that there may be a cellular mechanism to essential hypertension, not necessarily a humoral etiology, and that perhaps we may need to further investigate the cellular mechanism. We may now be at the dawn of a paradigm shift in our understanding of essential hypertension and this attention to NLR may be the beginning of such.

So what does the NLR or absolute neutrophil number actually mean at face value? At first look, it could imply increased production. Alternatively, it may also imply demargination of the neutrophil white blood bold cells, ie, a decreased tendency to adhere to vascular endothelium, akin to what is observed with increased white blood cell counts and absolute neutrophil counts after exogenous glucocorticoid hormone administration. Of note, the class of antihypertensive agents are also known to have an influence on microinflammation, eg, C‐reactive protein in treated hypertensive patients.6, 7 The current cohort size likely would not afford for further adjustment of an antihypertensive class agent effect, potentially modulating the ratio of NLR. Nonetheless, diuretic use, as expected, was more common among patients with RHT. We found it instructive, however, that when C‐reactive protein was incorporated into multivariate analysis, CRP lost its significance. Red cell distribution width (RDW) may be another, similar marker of inflammation and may potentially improve under effective antihypertensive therapy. The extreme degree of this latter phenomenon is observed in malignant hypertension with massive RBC fragmentation creating the phenotype thrombotic microangiopathy; however, this may be the result, not the cause, of malignant hypertension. In this current cohort there was no association between RDW and hypertension status, perhaps as a result of the good achieved BP control. It is uncertain, however, whether NLR would have an association with RDW in similar cohorts.

An additional merit of the paper rests in the ambulatory blood pressure monitoring (ABPM) performed in the cohort. ABPM is a safe,8 effective procedure with expanding wordwide use during the past decades. Overnight BP and heart rate on ABPM is positively associated with white blood cell count.9 It helps to avoid misclassification of hypertension in patients2 and it further reaffirmed the generally good achieved BP control in the RHT group (132.3±11.4/83.6±21.1 mm Hg) in this study. As one would expect, the BP values in the office were approximately 5 mm Hg higher than those obtained during ABPM recording in this study.

The clinical utility of the findings of Belen and colleagues,5 however, appears to be of limited value for the practicing physician for the time being: the values of absolute neutrophil counts reported certainly would have fallen within the normal range in our laboratory. Nonetheless, while the NLR may be difficult to interpret for one single individual, it may have a relatively large contribution on the level of entire populations, to the burden of attributable cardiovascular risk. What, then, may link these two phenomenon––RHT and an elevated NLR––together?

Obstructive sleep apnea (OSA)––or, broadly speaking, sleep‐disordered breathing––is an obvious candidate for RHT10, 11 and should be in the differential workup of resistant hypertension.12 OSA is still a vastly underappreciated entity and perhaps a not sufficiently emphasized entity during most medical schools' curriculum. The authors excluded patients with BMI >30 kg/m² (among many other exclusion criteria), making OSA less likely (but not excluded) among these enrollees. However, BMI may not capture all aspects of obesity and the waist to hip ratio (or neck size) would have been additional helpful parameters. Additionally, historical information, such as the Berlin questionnaire or the modified/shortened version of it can be helpful in estimating the predicted burden of OSA.13, 14

It is also possible that the observed neutrophil‐lymphocyte shift may be reflective of increased glucocorticoid secretion (herewith, we would assume still within the physiologic range). Subclinical cases of excessive cortisol excretion may overlap with other endocrine causes and may be difficult to rule out without 24‐hour urine collections.15 Perhaps additional, future studies may examine this phenomenon in these patients, similar to past efforts.16, 17

Along with the merits of this papers, some of the limitations should also be emphasized. Largest among the paper's limitations, is the cross‐sectional nature of the study. At present time, these findings are only associative in nature and causality cannot be implied from the current data. In addition, the recruitment procedure for the study was not well described. It is uncertain to the reader how many patients were approached, screened, and/or rejected from enrollment. Without knowing these details, the readers have to assume that patients with CHT and RHT were recruited from a highly selected referral population and the associations observed in this paper may not be immediately applicable to community‐dwelling individuals or those from different sociocultural or ethnic backgrounds. Similar limitations apply to the control group of normotensive patients. While the authors clearly excluded obese patients (BMI >30 kg/m²), waste circumference was not reported in the paper. Urine proteinuria within the “normal” range was also not reported. Among the biochemical parameters, uric acid was not measured. Uric acid is associated with visceral obesity18 and known to be elevated in patients with OSA.13, 19, 20, 21 Accordingly, measurement of uric acid may provide a helpful hint regarding underappreciated visceral obesity or OSA burden in the examined cohort. Finally, it is also unclear whether any particular “cutoff” value exists for NLR to discriminate RHT from NT vs CHT. In this regard, receiver operating characteristic curves may be helpful in future studies on NLRs.

In summary, while the concept of the study appears seemingly simple, the excellent matching between the three groups with respect to major demographic, anthropometric, and biochemical parameters lends credence to the study and the author's findings. It reminds us that meaningful clinical research can be performed without sophisticated or novel tests, using commercially available technology during routine care, with meticulous matching of subcohorts and asking the right questions. While additional, confirmatory data are clearly needed––including from other populations and larger cohorts––and the performance characteristics of NLR better understood, the study by Belen and colleagues5 is a welcome addition to the existing field of resistant hypertension and an important step forward.