Causal relationship between hemoglobin and hypertension: A Mendelian randomization study

Hypertension (HTN) is an important risk factor for heart disease and stroke1,2. Globally, 26.1% of adult women and 26.6% of adult men in the year 2000 suffered from HTN, and the prevalence of HTN among women was projected to be higher than that of men by 20253. Not only that, the Global Burden of Disease study has shown that non-optimal blood pressure is the biggest contributor to global disability-adjusted life-years, leading to 212 million lost healthy life years each year4. Several modifiable risk factors for HTN include low potassium intake, high sodium intake, alcohol consumption, obesity, lack of physical activity and unhealthy diet5. Indeed, intervening on risk factors for HTN plays an important role in preventing hypertension, stroke, and heart disease. 
In the past decade, the effect of hemoglobin (HGB) on HTN has received attention. A large cohort study6 in the Netherlands composed of 101,377 subjects showed that HGB was positively correlated with systolic and diastolic blood pressure in both healthy men and women. HGB is an important component of red blood cells and transports oxygen and carbon dioxide. The association between HGB and blood pressure has also been investigated by others7-10. Results of the study by Ren et al. which involved 4,186 women and 4,851 men from China showed that HGB was an important risk factor for increased blood pressure for both men and women7. Another study reported that the treatment of anemia in hemodialysis patients with erythropoietin, the hormone that plays a key role in the production of red blood cells, is associated with increased blood pressure11. In an animal study, an increased dose of cross-linked HGB during cardiopulmonary bypass surgery resulted in an increase in mean arterial pressure12. In the past two years, we recruited subjects of new-onset and first-diagnosed HTN cases to explore risk factors of HTN in perimenopause in Shantou, south of China. For the new-onset HTN design, we found that higher hemoglobin (HGB) concentration was strongly associated with HTN in perimenopause. In the second study on first-diagnosed HTN, we found that total bilirubin (TBIL) was negatively associated with HTN in this period. Since TBIL is an index of the breakdown of red blood cells, we believe these results in Shantou indicating that HTN is associated with higher HGB and lower TBIL in new-onset and first-diagnosed HTN data, respectively, were consistent. These results were in agreement with yet another study13, in which a positive relationship between HGB and risk of HTN was observed for non-anemic Japanese men and women. Several other studies have similarly discovered the association between HGB and HTN 8,9. Kim et al. surveyed 4,899 Korean residents, and although their cross-sectional analysis confirmed that HGB was associated with HTN, their longitudinal analysis showed a negative result 10. 
Most of the aforementioned studies are based on observational data, which can be influenced by unmeasured confounding. A method to overcome this problem is Mendelian randomization (MR), which is an analytical method based on Mendel’s law of random segregation using genetic variants as instrumental variables (IVs), which are independent of potential confounders. It is considered to be an alternative design for clinical randomized controlled trials14,15. MR leverages the fact that the direction of the causal effect of genetic IVs is clear, since randomization of genes is completed in the fertilized egg prior to being submitted to any harmful exposures. To date, it is unknown whether the relationship between HGB and HTN is causal. In this study, we aim to explore the causal relationship between these two variables in both men and women based on one-sample and two-sample MR analyses.