Abstract
Patients with masked hypertension (MH) and masked uncontrolled hypertension (MUCH) are easily overlooked, and both cause target organ damage. We propose a prediction model for MH and MUCH patients based on clinical features at a single outpatient visit. Data collection was planned before the index test and reference standard were after. Thus, we retrospectively collect analyzed 804 subjects who underwent ambulatory blood pressure monitoring (ABPM) at Renmin Hospital of Wuhan University. These patients were divided into normotension/controlled hypertension group (n = 121), MH/MUCH (n = 347), and sustained hypertension (SH)/sustained uncontrolled hypertension group (SUCH) (n = 302) for baseline characteristic analysis. Models were constructed by logistic regression, a nomogram was visualized, and internal validation by bootstrapping. All groups were performed according to the definition proposed by the Chinese Hypertension Association. Compared with normotension/controlled hypertension, patients with MH/MUCH had higher office blood pressure (BP) and were more likely to have poor liver and kidney function, metabolic disorder and myocardial damage. By analysis, [office systolic blood pressure (OSBP)] (P = .004) and [office diastolic blood pressure (ODBP)] (P = .007) were independent predictors of MH and MUCH. By logistic regression backward stepping method, office BP, body mass index (BMI), total cholesterol (Tch), high-density lipoprotein cholesterol (HDL-C), and left ventricular mass index are contained in this model [area under curve (AUC) = 0.755] and its mean absolute error is 0.015. Therefore, the prediction model established by the clinical characteristics or relevant data obtained from a single outpatient clinic can accurately predict MH and MUCH.
Keywords: masked hypertension, masked uncontrolled hypertension, nomogram, prediction model
1. Introduction
Cardiovascular diseases are the leading cause of death and account for an estimated 30% of deaths worldwide.[1] Hypertension is an overarching risk factor for cardiovascular morbidity and mortality, and it is the most critical contributor to the burden of disease and the leading cause of mortality worldwide.[2,3] Studies have shown that hypertension is responsible for almost 13% of all deaths.[1,4,5] Based on the current international guidelines, ambulatory blood pressure monitoring (ABPM) is widely accepted as the gold standard for the diagnosis of hypertension.[6–8]
Masked hypertension (MH) and masked uncontrolled hypertension are used to define people with normal office blood pressure (BP) but hypertension out-of-office, which ABPM determines.[9] MH refers to patients in the untreated or initial stage of treatment, and masked uncontrolled hypertension (MUCH) refers to patients with prior hypertension treatments.[10,11] A large number of studies have shown that the risk of target organ damage and cardiovascular and cerebrovascular disease in patients with MH and MUCH is similar to that in patients with sustained hypertension (SH) or sustained uncontrolled hypertension (SUCH), and is significantly higher than that in patients with normotension (NH) and CH.[12–15]
However, when performing ABPM, the number of measurements is higher, and the measurements are also taken during sleep, so it can produce discomfort and affect the patient’s sleep, acceptance, and compliance. Moreover, few primary care providers screen for MH, and little is known about primary care providers’ awareness, knowledge, and attitudes toward MH.[16] Then, MH and MUCH patients are always ignored because of normal office BP. In conclusion, the above reasons often limit the popularity of ABPM.
The purpose of this study is to identify MH/MUCH patients through the prediction model established by the clinical characteristics of a single outpatient and the data of examination results.
2. Methods
2.1. Population
Ethics Committee of Renmin Hospital of Wuhan University approved the study, and the study involves patient consent. Data collection was planned before the index test and reference standard were after. The patients who underwent ABPM in our hospital from September 1st, 2020 to December 31st, 2020 were selected retrospectively. The inclusion criteria included age over 18 years’ old and complete and accurate ambulatory BP data (more than 20 BP measurements during the day, more than 7 BP measurements at night, and more than 27 BP measurements throughout the day).[6,17,18] The exclusion criteria include secondary hypertension, life expectancy less than or equal to 12 months, serious diseases (such as acute myocardial infarction, acute renal failure requiring renal replacement therapy, and end-stage malignant tumors), mental diseases, inability to cooperate with ABPM, and other diseases leading to inaccurate BP measurement.[17,19,20]
2.2. Definition
The diagnostic criteria of hypertension proposed in this study are implemented according to the diagnostic criteria of Europe and China, that is, the diagnostic threshold of hypertension is 140/90 mm Hg.[20–22] (Refer to Supplementary File 1a and Supplementary File 1b, Supplemental Digital Content 1, http://links.lww.com/MD/I59 for specific diagnostic criteria and hypertension subtypes.)
2.3. Measurement of BP in the office
The clinical BP was recorded by trained nursing staff in 3 separate visits on 1 day. During each measurement, use the verified upper arm electronic sphygmomanometer with a standard cuff to read the sitting BP twice, with an interval of 2 min. Rest quietly for 5 min before measurement, and record the average value of the two BP readings as the measured value of this BP. The mean value of three BP measurements was taken as the study’s consulting room BP value.[6,19,23]
2.4. Measurement of ambulatory BP
Participants completed 24-h ABPM using an ABPM instrument with an appropriately sized cuff in the arm where their BP measurements were high. Ambulatory BP monitors were read at least once per hour and at least 27 times throughout the day (at least 20 times during the day and 7 times during the night). ABPM was uniformly divided into a whole day, daytime (8:00‐22:00), and night periods (22:00‐8:00).[6,24]
2.5. Measurement of observation index
All tests were obtained as fasting in the early morning before the patient underwent ABPM. The collected parameters included the following: liver function (alanine aminotransferase (ALT), aspartate aminotransferase (AST)), lipid profiles (total cholesterol (Tch), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), small dense low-density lipoprotein cholesterol (sdLDL-C), arteriosclerosis index (AI, TG/HDL-C-1)), renal function (uric acid (UA), urea, creatinine (Cr), estimated glomerular filtration rate (eGFR)), fasting plasma glucose (Glu), electrocardiographic indicators (electric axis, QRS duration, PR interval, QT interval, QTc, SV1, RV5, and RS)), Echocardiographic parameters (diameter of ascending aorta (AAOD), left atrial diameter (LAD), left ventricular end diastolic diameter (LVDD), right atrial diameter (RAD), right ventricular end diastolic diameter (RVDD), end diastolic interventricular septal thickness (IVSD), left ventricular posterior wall thickness at end diastole (LVPWD), left ventricular ejection fraction (LVEF), left ventricular mass index (LVMI)).
2.6. Statistical analysis
Because this study had partially missing data, imputation of missing values took the method of multiple imputations five times, and the numerical value of these imputations was taken as the final data of the study while missing categorical variables took the method of deletion and were excluded from this study.[25] Statistical analysis was performed by R 4.1.3 and SPSS 27.0 in this study. Quantitative variables are expressed as mean ± SD (x ± s), and categorical variables are expressed as percentages (%). Continuous parametric data were compared using an unpaired Student t test. Non-parametric data were compared using the Mann–Whitney U test. Categorical variables were analyzed using the Chi-square test or Fisher exact test. Spearman rank correlation coefficients were calculated between candidate variables. Statistical significance was inferred at a two-sided P value < .05.
Univariate logistic regression analysis was used to find out the clinical characteristics and relevant data that have a potential relationship with MH. The clinical features strongly correlated with MH were found by multivariate logistic regression analysis. The model is simplified and the error caused by the overfitting of the model is prevented by the backward step method of logistic regression.
The risk prediction model is visualized by transforming the regression coefficient into the Nomo coefficient. According to the degree of influence of each variable on the model, the corresponding score is given, and the score of each patient is calculated after adding all the scores, to determine the probability of disease of the patient.[26]
Area under curve (AUC) value, sensitivity, specificity, and a cutoff value of the model were calculated by receiver operating characteristic (ROC) analysis. A new sample is established by repeated return sampling, and the model is tested with the new sample.[27]
3. Result
3.1. Clinical characteristics of hypertension subtypes
Of all 804 patients, 121 were NH and CH, 343 were MH and MUCH, and 302 were SH and SUCH. No adverse events occurred when index tests or reference standards were performed. Among these 804 patients were 458 (57%) males with an age interval distribution of 61 ± 0.5. In the subsequent subgroup analysis, we found that in terms of baseline characteristics, compared with patients with NH and CH, MH and MUCH patients were younger, had higher office BP, and presented with poorer metabolic profiles, including higher body mass index (BMI) and serum TG levels. Meanwhile, MH showed that the structure of the left ventricle was more abnormal. And the difference between MH/MUCH and SH/SUCH is the same as the difference between NH/CH and MH/MUCH. However, SH/SUCH patients presented with medication nonadherence, and more of them did not take any antihypertensive drugs (Table 1).
Table 1.
Baseline clinical characteristics of patients.
| NH/CH (group 1) |
MH/MUCH (group 2) |
SH/SUCH (group 3) |
P value (group 1–2) |
P value (group 2–3) |
|
|---|---|---|---|---|---|
| Age, yr | 64.24 ± 1.15 | 61.14 ± 0.74 | 59.33 ± 0.87 | .008 | .139 |
| Male, % | 47 (39%) | 210 (61%) | 187 (62%) | <.001 | .7 |
| OSBP, mm Hg | 123.45 ± 0.85 | 128.61 ± 0.45 | 151.06 ± 0.65 | <.001 | NA |
| ODBP, mm Hg | 69.1 ± 0.7 | 74.67 ± 0.45 | 86.43 ± 0.68 | <.001 | NA |
| OHR, bpm | 78.13 ± 1.13 | 79.73 ± 0.82 | 83.72 ± 0.9 | .464 | .001 |
| Hypertension duration, yr | 8.31 ± 1.01 | 7.79 ± 0.57 | 8.33 ± 0.61 | .957 | .415 |
| No. of antihypertensive drugs | .116 | .007 | |||
| 0 | 45 (37%) | 111 (32%) | 68 (23%) | ||
| 1 | 24 (20%) | 68 (20%) | 60 (20%) | ||
| 2 | 31 (26%) | 83 (24%) | 78 (26%) | ||
| 3 | 17 (14%) | 54 (16%) | 68 (23%) | ||
| 4 | 4 (3.3%) | 26 (7.5%) | 23 (7.6%) | ||
| 5 | 0 (0%) | 5 (1.4%) | 5 (1.7%) | ||
| Previous history | |||||
| AF, % | 11 (9.1) | 28 (8.1) | 12 (4) | .7 | .03 |
| CHD, % | 60 (50) | 143 (41) | 98 (32) | .11 | .021 |
| Stroke, % | 27 (22) | 84 (24) | 83 (27) | .7 | .3 |
| CKD, % | 4 (3.3) | 21 (6.1) | 26 (8.6) | .2 | .2 |
| TDM, % | 22 (18) | 91 (26) | 74 (25) | .075 | .6 |
| Hyperlipidemia, % | 28 (23) | 112 (32) | 96 (32) | .059 | .9 |
| BMI | 22.77 ± 0.36 | 25.05 ± 0.25 | 24.97 ± 0.25 | <.001 | .921 |
| Laboratory tests | |||||
| ALT | 19.22 ± 0.96 | 23.48 ± 0.88 | 24.12 ± 0.99 | .014 | .809 |
| AST | 21.51 ± 0.65 | 22.9 ± 0.56 | 22.93 ± 0.55 | .237 | .952 |
| ALT/AST | 0.88 ± 0.03 | 0.99 ± 0.02 | 1.01 ± 0.02 | .007 | .618 |
| Urea | 5.49 ± 0.18 | 5.79 ± 0.11 | 6.16 ± 0.16 | .057 | .444 |
| Cr | 64.5 ± 1.46 | 72.98 ± 1.58 | 87.49 ± 5.66 | .003 | .143 |
| Urea/Cr | 0.09 ± 0 | 0.08 ± 0 | 0.08 ± 0 | .285 | .339 |
| UA | 337.74 ± 10.16 | 372.81 ± 5.59 | 395.61 ± 5.78 | .001 | .001 |
| eGFR | 89.69 ± 1.62 | 89.88 ± 1.07 | 88.61 ± 1.52 | .726 | .557 |
| Glu | 5.07 ± 0.1 | 5.55 ± 0.1 | 5.76 ± 0.11 | .005 | .104 |
| K | 4.02 ± 0.03 | 4 ± 0.02 | 5.48 ± 1.43 | .416 | .161 |
| Na | 142.43 ± 0.38 | 143.06 ± 0.19 | 142.67 ± 0.18 | .358 | .168 |
| Cl | 107.32 ± 0.46 | 107.51 ± 0.17 | 107.15 ± 0.17 | .514 | .065 |
| Ca | 2.24 ± 0.01 | 2.23 ± 0.01 | 2.25 ± 0.01 | .951 | .063 |
| Tch | 4.16 ± 0.09 | 4.2 ± 0.06 | 4.45 ± 0.06 | .596 | .003 |
| TG | 1.38 ± 0.08 | 1.8 ± 0.07 | 1.89 ± 0.07 | .001 | .051 |
| HDL-C | 1.18 ± 0.03 | 1.02 ± 0.01 | 1.07 ± 0.02 | <.001 | .067 |
| LDL-C | 2.32 ± 0.08 | 2.37 ± 0.04 | 2.51 ± 0.05 | .41 | .044 |
| sdLDL-C | 0.67 ± 0.03 | 0.75 ± 0.02 | 0.83 ± 0.02 | .018 | .011 |
| AI | 2.71 ± 0.1 | 3.25 ± 0.08 | 3.33 ± 0.07 | <.001 | .211 |
| Echocardiography | |||||
| AAOD | 32.08 ± 0.4 | 33.18 ± 0.2 | 33.45 ± 0.22 | .003 | .39 |
| LVEF, % | 58.45 ± 0.49 | 58.55 ± 0.23 | 58.91 ± 0.19 | .551 | .734 |
| LVMI | 83.65 ± 1.95 | 90.14 ± 1.2 | 94.13 ± 1.44 | .01 | .021 |
| LVH, % | 21 (17) | 71 (20) | 73 (24) | .5 | .3 |
| Electrocardiogram | |||||
| Electric axis | 33.99 ± 4.09 | 28.8 ± 1.99 | 31.4 ± 2.1 | .121 | .283 |
| QRS duration | 97.4 ± 1.43 | 101.39 ± 0.93 | 102.21 ± 0.91 | .001 | .319 |
| PR interval | 158.97 ± 2.55 | 162.06 ± 1.43 | 162.93 ± 1.59 | .19 | .943 |
| QT interval | 391.53 ± 3.37 | 387.31 ± 1.97 | 383.3 ± 2.1 | .261 | .138 |
| QTC | 427.33 ± 2.76 | 426.67 ± 1.5 | 429.47 ± 1.52 | .928 | .083 |
| SV1 | 0.69 ± 0.06 | 0.71 ± 0.03 | 0.81 ± 0.03 | .329 | .001 |
| RV5 | 1.35 ± 0.05 | 1.39 ± 0.03 | 1.51 ± 0.04 | .455 | .017 |
| RS | 2.05 ± 0.08 | 2.1 ± 0.04 | 2.32 ± 0.05 | .368 | .001 |
AAOD = diameter of ascending aorta, AF = atrial fibrillation, AI = atherosclerosis index, ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, CH = controlled hypertension, CHD = coronary heart disease, CKD = chronic kidney disease, Cr = creatinine, eGFR = estimated glomerular filtration rate, Glu = glucose, HDL-C = high-density lipoprotein cholesterol, LDL-C = low-density lipoprotein cholesterol, LVEF = left ventricular ejection fraction, LVH = left ventricular hypertrophy, LVMI = left ventricular mass index, MH = masked hypertension, MUCH = masked uncontrolled hypertension, NH = normotension, ODBP = office diastolic blood pressure, OHR = office heart rate, OSBP = office systolic blood pressure, sdLDL-C = small dense low-density lipoprotein cholesterol, SH = sustained hypertension, SUCH = sustained uncontrolled hypertension, Tch = total cholesterol, TG = triglycerides, TDM = diabetes mellitus, UA = uric acid.
3.2. Results of univariate and multivariate logistic regression
In the univariate logistic regression analysis, gender, age, office BP, BMI, ALT, ALT/AST, Cr, UA, Glu, HDL-C, AI, AAOD, LVMI, and QRS duration were associated with the presence of MH and MUCH. In the multivariate logistic regression analysis, the office BP was significantly associated with the presence of MH and MUCH (Table 2).
Table 2.
Logistic regression results.
| Dependent: type | OR | 95% CI (univariable) |
P value (univariable) |
OR | 95% CI (multivariable) |
P value (multivariable) |
|---|---|---|---|---|---|---|
| Female | 0.41 | 0.27‐0.63 | <.001 | 0.68 | 0.38‐1.21 | .191 |
| Age | 0.98 | 0.97‐1.00 | .032 | 1 | 0.97‐1.02 | .718 |
| OSBP | 1.06 | 1.04‐1.09 | <.001 | 1.04 | 1.01‐1.07 | .004 |
| ODBP | 1.08 | 1.06‐1.11 | <.001 | 1.05 | 1.01‐1.09 | .007 |
| BMI | 1.14 | 1.08‐1.20 | <.001 | 1.06 | 1.00‐1.14 | .073 |
| ALT | 1.02 | 1.01‐1.04 | .01 | 1.02 | 0.95‐1.12 | .592 |
| ALT/AST | 2.64 | 1.41‐5.13 | .003 | 0.55 | 0.06‐4.62 | .594 |
| Cr | 1.02 | 1.01‐1.03 | .002 | 1.01 | 0.99‐1.03 | .385 |
| UA | 1 | 1.00‐1.01 | .002 | 1 | 1.00‐1.00 | .529 |
| Glucose | 1.26 | 1.08‐1.52 | .008 | 1.14 | 0.93‐1.42 | .224 |
| TG | 1.47 | 1.17‐1.92 | .002 | 1.41 | 0.86‐2.47 | .203 |
| HDL-C | 0.14 | 0.06‐0.29 | <.001 | 0.03 | 0.00‐1.05 | .06 |
| AI | 1.59 | 1.28‐1.98 | <.001 | 0.59 | 0.20‐1.67 | .321 |
| AAOD | 1.08 | 1.02‐1.14 | .01 | 1.02 | 0.95‐1.10 | .537 |
| LVMI | 1.01 | 1.00‐1.02 | .016 | 1.01 | 1.00‐1.03 | .084 |
| QRS duration | 1.02 | 1.00‐1.03 | .027 | 1.01 | 0.99‐1.02 | .325 |
AAOD = diameter of ascending aorta, AI = atherosclerosis index, ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, Cr = creatinine, CI = confidence interval, HDL-C = high-density lipoprotein cholesterol, LVMI = left ventricular mass index, ODBP = office diastolic blood pressure, OSBP = office systolic blood pressure, OR = odds ratio, TG = triglycerides, UA = uric acid.
3.3. Predictive model construction and internal validation
We developed a prediction model by using the predictors from the best fitting model obtained by logistic regression backward stepping method, which contains office BP, BMI, Tch, HDL-C, and LVMI. To visualize the system, we convert the regression results of each factor in the model into a Nomogram coefficient (nomogram coefficient = regression coefficient B* (maximum value of factor - minimum value of factor)), set the maximum score of the maximum factor of the coefficient as 10 points, and reduce the other factors in equal proportions (Supplementary File 2, Supplemental Digital Content 2, http://links.lww.com/MD/I60). Through the above principles and steps, we developed two scoring systems to predict MH and MUCH (Fig. 1). The maximum score of each item is set to 10 points, and the maximum total score is set to 35 points. The total score corresponds to the corresponding prediction probability.
Figure 1.
Nomogram of the scoring system predicting for MH/MUCH. The highest score of each index is 10 points. The best truncation value of the model is 0.787, and the corresponding score is about 24 points. MH = masked hypertension, MUCH = masked uncontrolled hypertension.
The efficiency of the model was tested by the ROC curve. We found that the prediction efficiency of the model was positive (AUC = 0.755, 95% CI: 0.704‐0.806), and the sensitivity and specificity were within the acceptable range. Through ROC analysis, the cutoff value of our model is 0.787, and the score corresponding to the modified value on the nomogram is about 24 (Fig. 2a). The negative and positive cases obtained according to the gold standard ABPM and the cutoff value set by the diagnostic model can be seen in the Supplementary File 3, Supplemental Digital Content 3, http://links.lww.com/MD/I61. In addition, we internally validated the two sets of systems, and we found that both sets of models had good predictivity by the method of repeated sampling of the original sample 10000 times. As can be seen from the figure: from the figure, we can see that the apparent curve represented by the model constructed by the actual sample and the bias-corrected curve represented by the model of the new sample constructed by repeated sampling closely fit the ideal curve, which indicates that our model has good extrapolation (Fig. 2b).
Figure 2.
The results of testing prediction model accuracy. (a) ROC curve of the prediction model. The AUC of the ROC curve was 0.755, 95% CI: 0.704‐0.806, the TPR was 0.627, the (TNR) was 0.775, the Youden index was 0.402, and the optimal cutoff value was 0.787. (b) Internal validation using 10000 bootstraps resamples showed that only a small amount of the mean absolute error (0.015). AUC = area under curve, ROC = receiver operating characteristic, TPR = true positive rate, TNR = true negative rate.
4. Discussion
The main findings of this study are as follows: compared to NH and CH, MH and MUCH are associated with higher office BP and more disordered lipid metabolism; office BP is an independent risk factor for MH and MUCH; office BP, BMI, Tch, HDL-C, and LVMI are significantly association with MH and MUCH. The risk prediction model constructed by these clinical characteristics and indicators obtained from a single outpatient clinic can accurately predict MH and MUCH.
According to the definition of the guidelines, MH and MUCH are characterized by normotension in the office but hypertension out of the office. The only difference between them is whether they have a history of taking hypertension drugs. However, as stated in the introduction above: previous studies have shown that MH and MUCH are highly correlated with more severe target organ damage and worse prognosis. Although recent guidelines designate ABPM as the gold standard technique for diagnosing hypertension and distinguishing the types of hypertensions, it also has some limitations (such as affecting night sleep and high price). Therefore, it is unrealistic to conduct an ABPM examination for all populations. It is important to identify those at high risk of MH through the clinical characteristics related to a single outpatient visit.
Our study found that high office BP was highly correlated with MH and was an independent risk factor for MH. The results of the Spanish ambulatory BP registry and two other cohorts are also consistent with our findings.[28,29]
Our results also found that patients with MH and much showed worse metabolic status. These patients often have higher BMI, lower HDL-C, and higher serum Tch. Some previous studies have found MH to be more likely to be present in hyperlipidemia.[30–32] This is also consistent with our findings. In addition, relevant studies have also confirmed that obesity or high BMI is highly correlated with MH and MUCH.[33,34]
In addition, our study also found that higher LVMI was associated with MH as well as much. Related studies on target organ damage of MH and MUCH found that MH and MUCH patients had higher LVMI.[35,36] This is consistent with our findings. However, we evaluated the prevalence of left ventricular hypertrophy (LVH) among each group according to the LVH standard (LVMI ≥ 115 g/m2 in males and LVMI ≥ 95 g/m2 in females), but we found no significant statistical difference. This may be related to the selection of the population, the size of the sample, and the length of the course of hypertension.
In addition to the above indicators included in the model, our research results are worth discussing. Our study found no significant difference in the use of antihypertensive drugs between MH/ much patients and NH/ CH patients. This also indicates that much is not related to drug noncompliance. Relevant studies have also confirmed our view.[37] In echocardiography, we found that the AAOD of MH/ MUCH patients was larger. Wang et al[38] used a machine learning approach to perform automatic aorta measurement in thoracic CT images at nine key positions of 801 patients to predict MH. Better results were also finally obtained (AUC = 0.78). Therefore, this indicator may also be potentially related to MH and MUCH.
Whereas, in theory, a considerable proportion of studies found a strong association between CKD and MH.[39,40] CKD and hypertension share similar risk factors and mutually contribute to each other. Still, our study did not find an association due to insufficient sample size and different study populations in our analysis. Our study only found that Cr was associated with MH and MUCH, and Cr in our study population was within the normal range. Furthermore, there is no relevant study evidence for a causal link between a history of chronic kidney disease (CKD) and MH.
In conclusion, our study provides a good evaluation model for predicting patients with hypertension outside the office. As far as we know, this is the first risk prediction model established by the relevant data obtained from a single outpatient service in China. By changing the scoring model, we can be sure to determine the patients who really should undergo ABPM. This can increase the patient’s medical experience and save medical costs.
5. Conclusion
Compared with NH/CH, patients with MH/MUCH had higher office BP and were more likely to have poor liver and kidney function, metabolic disorder and myocardial damage. The proposed model accurately predicts MH and MUCH based on the clinical features.
6. Limitation
This study is a retrospective study, and there will inevitably be a lack of relevant data. Although various statistical methods process these missing values, it is still difficult to avoid the relevant errors. Secondly, because the concept of occult hypertension is not popularized in China, most patients with normal BP in the clinic do not carry out ABPM in time, leading to a certain selection bias in the study. And this study only included the patient data for 4 months. Although the total sample number is more than 800, the sample number of this study is still small for the high prevalence of hypertension, which may also lead to some “no significant difference” results in this study. Therefore, future research in this area may include a larger sample size.
Author contributions
Conceptualization: Bin Kong.
Data curation: Hong Meng.
Formal analysis: Liang Guo.
Methodology: Bin Kong.
Project administration: He Huang.
Supervision: He Huang.
Software: Liang Guo.
Writing – original draft: Hong Meng.
Writing – review & editing: He Huang, Wei Shuai.
Supplementary Material
Abbreviations:
- AAOD =
- diameter of ascending aorta
- ABPM =
- ambulatory blood pressure monitoring
- AI =
- atherosclerosis index
- ALT =
- alanine aminotransferase
- AST =
- aspartate aminotransferase
- AUC =
- area under curve
- BMI =
- body mass index
- BP =
- blood pressure
- CH =
- controlled hypertension
- Cr =
- creatinine
- HDL-C =
- high-density lipoprotein cholesterol
- LVMI =
- left ventricular mass index
- MH =
- masked hypertension
- MUCH =
- masked uncontrolled hypertension
- NH =
- normotension
- SH =
- sustained hypertension
- SUCH =
- sustained uncontrolled hypertension
- Tch =
- total cholesterol
- TG =
- triglycerides
- UA =
- uric acid
Supplemental Digital Content is available for this article.
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
HM and LG contributed equally to this work.
This work was supported by grants from the Fundamental Research Funds for the Central Universities (No. 2042021kf0119) and the National Natural Science Foundation of China (No. 82070330).
The authors have no conflicts of interest to disclose.
How to cite this article: Meng H, Guo L, Kong B, Shuai W, Huang H. Nomogram based on clinical features at a single outpatient visit to predict masked hypertension and masked uncontrolled hypertension: A study of diagnostic accuracy. Medicine 2022;101:49(e32144).
Contributor Information
Hong Meng, Email: 2020283020133@whu.edu.cn.
Liang Guo, Email: 1192993221@qq.com.
Bin Kong, Email: kongbin@whu.edu.cn.
Wei Shuai, Email: sw09120@163.com.
References
- [1].GBD 2013 Mortality and Causes of Death Collaborators. Global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990-2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2015;385:117–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].World Health Organization. 2003 World Health Organization (WHO)/International Society of Hypertension (ISH) statement on management of hypertension. J Hypertens. 2003;21:1983–92. [DOI] [PubMed] [Google Scholar]
- [3].Lim SS, Vos T, Flaxman AD, et al. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380:2224–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Poulter NR, Prabhakaran D, Caulfield M. Hypertension. Lancet. 2015;386:801–12. [DOI] [PubMed] [Google Scholar]
- [5].World Health Organization. Regional Office for South-East Asia. Global Health Risks: Mortality and Burden of Disease Attributable to Selected Major Risks. WHO Regional Office for South-East Asia; 2010. Available at: https://apps.who.int/iris/handle/10665/205781 [access date January 27, 2022]. [Google Scholar]
- [6].Joint Committee for Guideline Revision. 2018 Chinese Guidelines for prevention and treatment of hypertension-A report of the Revision Committee of Chinese Guidelines for prevention and treatment of hypertension. J Geriatr Cardiol. 2019;16:182–241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Reboussin DM, Allen NB, Griswold ME, et al. Systematic review for the 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults. J Am Coll Cardiol. 2018;71:2176–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Brouwers S, Sudano I, Kokubo Y, et al. Arterial hypertension. Lancet. 2021;398:249–61. [DOI] [PubMed] [Google Scholar]
- [9].Shimbo D, Artinian NT, Basile JN, et al. Self-measured blood pressure monitoring at home: a joint policy statement from the American Heart Association and American Medical Association. Circulation. 2020;142:e42–63. [DOI] [PubMed] [Google Scholar]
- [10].O’Brien E, Parati G, Stergiou G, et al. European Society of Hypertension position paper on ambulatory blood pressure monitoring. J Hypertens. 2013;31:1731–68. [DOI] [PubMed] [Google Scholar]
- [11].Williams B, Mancia G, Spiering W, et al. 2018 Practice Guidelines for the management of arterial hypertension of the European Society of Hypertension and the European Society of Cardiology: ESH/ESC task force for the management of arterial hypertension. J Hypertens. 2018;36:2284–309. [DOI] [PubMed] [Google Scholar]
- [12].Hansen TW, Kikuya M, Thijs L, et al. Prognostic superiority of daytime ambulatory over conventional blood pressure in four populations: a meta-analysis of 7030 individuals. J Hypertens. 2007;25:1554–64. [DOI] [PubMed] [Google Scholar]
- [13].Asayama K, Thijs L, Li Y, et al. Setting thresholds to varying blood pressure monitoring intervals differentially affects risk estimates associated with white-coat and masked hypertension in the population. Hypertension. 2014;64:935–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Zhang DY, Guo QH, An DW, et al. A comparative meta-analysis of prospective observational studies on masked hypertension and masked uncontrolled hypertension defined by ambulatory and home blood pressure. J Hypertens. 2019;37:1775–85. [DOI] [PubMed] [Google Scholar]
- [15].Li Y, Staessen JA, Lu L, et al. Is isolated nocturnal hypertension a novel clinical entity?: findings from a Chinese Population Study. Hypertension. 2007;50:333–9. [DOI] [PubMed] [Google Scholar]
- [16].Boyd R, Carter E, Moise N, et al. Awareness, knowledge, and attitudes toward screening and treatment of masked hypertension in primary care. Am J Hypertens. 2021;34:1322–7. [DOI] [PubMed] [Google Scholar]
- [17].Cai P, Zhong W, Wang Y, et al. Effects of white-coat, masked and sustained hypertension on coronary artery stenosis and cardiac arrhythmia. Hypertens Res. 2020;43:121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Hung MH, Shih LC, Wang YC, et al. Prediction of masked hypertension and masked uncontrolled hypertension using machine learning. Front Cardiovasc Med. 2021;8:778306. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Lee CJ, Ha JH, Kim JY, et al. Office blood pressure threshold of 130/80 mmHg better predicts uncontrolled out‐of‐office blood pressure in apparent treatment‐resistant hypertension. J Clin Hypertension. 2021;23:595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Stergiou GS, Palatini P, Parati G, et al. 2021 European Society of Hypertension practice guidelines for office and out-of-office blood pressure measurement. J Hypertens. 2021;39:1293–302. [DOI] [PubMed] [Google Scholar]
- [21].Unger T, Borghi C, Charchar F, et al. 2020 International Society of Hypertension global hypertension practice guidelines. J Hypertens. 2020;38:982–1004. [DOI] [PubMed] [Google Scholar]
- [22].Committee of ambulatory blood pressure monitoring guidelines of China hypertension Alliance. 2020 Chinese Hypertension League Guidelines on Ambulatory Blood Pressure Monitoring. Chin J Front Med Sci (Electronic Version). 2021;13:34–51. [Google Scholar]
- [23].Zhang DY, Cheng YB, Guo QH, et al. Treatment of masked hypertension with a Chinese herbal formula: a randomized, placebo-controlled trial. Circulation. 2020;142:1821–30. [DOI] [PubMed] [Google Scholar]
- [24].Parati G, Stergiou G, O’Brien E, et al. European Society of Hypertension practice guidelines for ambulatory blood pressure monitoring. J Hypertens. 2014;32:1359–66. [DOI] [PubMed] [Google Scholar]
- [25].Kim HJ, Shin JH, Lee Y, et al. Clinical features and predictors of masked uncontrolled hypertension from the Korean Ambulatory Blood Pressure Monitoring Registry. Korean J Intern Med. 2021;36:1102–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Hartaigh BO, Gransar H, Callister T, et al. Development and validation of a simple-to-use nomogram for predicting 5-, 10-, and 15-year survival in asymptomatic adults undergoing coronary artery calcium scoring. JACC Cardiovasc Imaging. 2018;11:450–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Lee Y, Shin JH, Park HC, et al. A prediction model for renal artery stenosis using carotid ultrasonography measurements in patients undergoing coronary angiography. BMC Nephrol. 2014;15:60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Banegas JR, Ruilope LM, de la Sierra A, et al. High prevalence of masked uncontrolled hypertension in people with treated hypertension. Eur Heart J. 2014;35:3304–12. [DOI] [PubMed] [Google Scholar]
- [29].Gkaliagkousi E, Protogerou AD, Argyris AA, et al. Contribution of single office aortic systolic blood pressure measurements to the detection of masked hypertension: data from two separate cohorts. Hypertens Res. 2021;44:215–24. [DOI] [PubMed] [Google Scholar]
- [30].Björklund K, Lind L, Zethelius B, et al. Isolated ambulatory hypertension predicts cardiovascular morbidity in elderly men. Circulation. 2003;107:1297–302. [DOI] [PubMed] [Google Scholar]
- [31].Kenny IE, Saeed S, Gerdts E, et al. Masked hypertension in obesity: potential predictors and arterial damage. Blood Press Monit. 2017;22:12–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].Furuki T, Kobayashi K, Toyoda M, et al. The influence of long-term administration of SGLT2 inhibitors on blood pressure at the office and at home in patients with type 2 diabetes mellitus and chronic kidney disease. J Clin Hypertens (Greenwich). 2020;22:2306–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [33].Narita K, Hoshide S, Ae R, et al. Simple predictive score for nocturnal hypertension and masked nocturnal hypertension using home blood pressure monitoring in clinical practice. J Hypertens. 2022;40:1513–21. [DOI] [PubMed] [Google Scholar]
- [34].Tepe D, Yilmaz S. Is office blood pressure measurement reliable in obese adolescents? Diabetes Metab Syndr Obes. 2021;14:3809–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [35].Jing L, Nevius CD, Friday CM, et al. Ambulatory systolic blood pressure and obesity are independently associated with left ventricular hypertrophic remodeling in children. J Cardiovasc Magn Reson. 2017;19:86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [36].van der Westhuizen B, Schutte AE, Gafane-Matemane LF, et al. Left ventricular mass independently associates with 24-hour sodium excretion in young masked hypertensive adults: the African-PREDICT study. Int J Cardiol. 2019;276:218–23. [DOI] [PubMed] [Google Scholar]
- [37].Siddiqui M, Judd EK, Dudenbostel T, et al. Masked uncontrolled hypertension is not attributable to medication nonadherence. Hypertension. 2019;74:652–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [38].Wang Y, Bai L, Yang J, et al. Artificial intelligence measuring the aortic diameter assist in identifying adverse blood pressure status including masked hypertension. Postgrad Med. 2022;134:111–21. [DOI] [PubMed] [Google Scholar]
- [39].Schnabel RB, Witt H, Walker J, et al. Machine learning-based identification of risk-factor signatures for undiagnosed atrial fibrillation in primary prevention and post-stroke in clinical practice. Eur Heart J Qual Care Clin Outcomes. 2022:qcac013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [40].Oh YK, Chin HJ, Ahn SY, et al. Discrepancies in clinic and ambulatory blood pressure in Korean chronic kidney disease patients. J Korean Med Sci. 2017;32:772–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.