Clinical Characteristics and Factors Associated with Uncontrolled Hypertension among Adult Patients in a Tertiary Care Centre: a Cross-Sectional Study

Abstract

Background

Achieving sustained blood-pressure control remains difficult in routine care, particularly in referral settings where patients often have longer di sease duration and multiple comorbidities. Cross-sectional clinical data can identify correlates of poor control, but such associations should not be interpreted as causal effects.

Objective

To describe the clinical characteristics of adults with treated hypertension and to examine factors associated with uncontrolled blood pressure in a tertiary-care population.

Methods

This cross-sectional observational study included 100 consecutively recruited adults with previously diagnosed hypertension at Government Medical College, Bhadradri Kothagudem, Telangana, India, from February to November 2025. Demographic, clinical, lifestyle, treatment, adherence, follow-up and office blood-pressure data were recorded. Controlled hypertension was defined as office blood pressure <140/90 mm Hg and uncontrolled hypertension as systolic blood pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg. Bivariate associations were summarized with unadjusted odds ratios (ORs) and 95% confidence intervals (CIs), followed by multivariable logistic regression.

Results

The mean age was 56.8 ± 11.4 years and 58.0% of participants were men. Uncontrolled hypertension was observed in 62.0% of the cohort. In bivariate analysis, body mass index (BMI) ≥25 kg/m², hypertension duration > five years, diabetes mellitus, high salt intake, physical inactivity, poor medication adherence, and missed follow-up visits were associated with uncontrolled blood pressure. After adjustment, poor medication adherence [adjusted OR (aOR) 6.84, 95% CI 2.02-23.15], BMI ≥25 kg/m² (aOR 3.49, 95% CI 1.11-10.99), physical inactivity (aOR 3.21, 95% CI 1.12-9.18), high salt intake (aOR 2.94, 95% CI 1.05-8.27) and hypertension duration > five years (aOR 2.88, 95% CI 1.02-8.15) remained statistically significant. The wide CIs indicate limited precision.

Conclusion

Several behavioural and clinical characteristics were associated with uncontrolled office blood pressure in this tertiary-care sample. Because this was a cross-sectional single-centre study based on a single-visit office assessment, its findings indicate associations rather than causal relationships and should not be extrapolated to the general hypertensive population.

INTRODUCTION

Hypertension is a major modifiable determinant of cardiovascular di sease, stroke, chronic kidney di sease and premature mortality. Despite the availability of effective antihypertensive medicines and well-established treatment strategies, population-level blood pressure (BP) control remains suboptimal. A pooled analysis of 1,201 population-representative studies showed that the global number of adults aged 30-79 years with hypertension doubled between 1990 and 2019, while substantial proportions remained undiagnosed, untreated, or inadequately controlled (1). Contemporary guidelines therefore place emphasis not only on pharmacotherapy but also on accurate BP measurement, cardiovascular-risk assessment, lifestyle modification, treatment adherence and continuing follow-up (2, 3).

The challenge is particularly relevant in India, where hypertension affects urban, rural and semi-urban communities and where important losses occur across the care continuum. Earlier systematic evidence documented low awareness, treatment and control, with pronounced deficiencies in rural populations (4). More recent national analyses have shown that diagnosis, treatment initiation and sustained control remain incomplete across many regions (5, 6). These observations suggest that office blood-pressure status among treated patients reflects a combination of clinical burden, health-seeking behaviour, adherence, access to follow-up, lifestyle practices and treatment complexity rather than drug prescription alone.

Tertiary-care populations may differ substantially from community samples because referral hospitals often receive patients with longer-standing di sease, diabetes, obesity, renal impairment, dyslipidaemia, cardiovascular di sease, or more intensive treatment. Indian epidemiological studies have reported associations between hypertension and age, excess body weight, diabetes, tobacco exposure, alcohol use and sedentary behaviour (7, 8). Among treated patients, nonadherence is also consistently associated with poor BP control (9). Excess adiposity, high sodium intake and inadequate physical activity are biologically plausible correlates of elevated blood pressure through renal, neurohormonal, vascular and metabolic pathways (10-13).

Local hospital-based data can help characterize the clinical profile of patients who remain uncontrolled during routine care, but the cross-sectional nature of such data requires cautious interpretation. Accordingly, the present study aimed to describe the demographic and clinical characteristics of adults with hypertension attending Government Medical College, Bhadradri Kothagudem, Telangana, India; estimate the proportion with uncontrolled office BP in this tertiary-care cohort; and examine patient and treatment characteristics associated with uncontrolled hypertension. The study was designed to identify associations, and not to establish temporal or causal relationships.

MATERIALS AND METHODS

Study design, setting, and period

A cross-sectional observational study was conducted at Government Medical College, Bhadradri Kothagudem, Telangana, India. Data collection was completed between February and November 2025. The institution is a tertiary-care teaching hospital that provides outpatient, inpatient, emergency, laboratory and specialist services to Bhadradri Kothagudem district and adjoining areas. Adults with hypertension attending the medicine outpatient department or admitted under general medicine services were considered for recruitment.

Participants and sampling

Adults aged ≥18 years with previously diagnosed hypertension or current antihypertensive treatment were eligible. Consecutive non-probability sampling was used until 100 eligible participants with complete study data had been enrolled. Patients with incomplete clinical records, hypertensive emergency, pregnancy-related hypertension, secondary hypertension under evaluation, or inability to provide a reliable clinical and treatment history were excluded. The pre-enrolment screening log did not preserve a reliable count of all patients initially assessed or excluded before enrolment, and the outpatient-versus-inpatient recruitment proportion was not retained; these omissions are acknowledged as limitations.

Sample-size rationale

No formal a priori power calculation was documented for this study; the sample of 100 participants was pragmatic and based on feasible consecutive recruitment during the prespecified study period. Sixty-two participants had uncontrolled hypertension. The final logistic-regression model contained seven covariate coefficients (plus an intercept), corresponding to approximately 8.9 uncontrolled-hypertension events per covariate. The adjusted estimates were therefore treated as exploratory and their precision was judged primarily from the 95% CIs.

Data collection and operational definitions

A structured case-record form captured age, sex, residence, body mass index (BMI), duration of hypertension, comorbidities, antihypertensive treatment pattern, number of antihypertensive medicines, lifestyle variables, medication-taking behaviour, follow-up attendance and office blood pressure. The BMI was categorized using conventional adult cut-offs: normal weight 18.5-24.9 kg/m², overweight 25.0-29.9 kg/m² and obesity ≥30 kg/m²; BMI ≥25 kg/m² was used in regression to represent excess body weight. Diabetes mellitus, dyslipidaemia, chronic kidney di sease (CKD) and cardiovascular di sease (CVD) were recorded from documented history, treatment records and physician diagnosis where available.

Medication adherence and lifestyle exposures

Medication adherence was assessed by self-report from the clinical interview and prescription history, including reported missed doses, irregular intake and interruption of prescribed therapy. A validated adherence questionnaire and a fixed recall-period score were not used; participants reporting irregular or interrupted medication use were classified as having poor adherence. High salt intake and physical inactivity were also recorded as self-reported binary variables on the study proforma. The original form did not include a validated dietary-sodium instrument, a quantitative sodium threshold, a validated physical-activity questionnaire, or an objective activity measure. These exposure definitions therefore carry a risk of recall, social-desirability and classification bias.

Antihypertensive treatment

Treatment intensity was available as monotherapy versus use of two or more antihypertensive medicines. Specific drug-class distributions and fixed-dose-combination use were not systematically captured in the dataset and could not be reported reliably. The number of prescribed agents was interpreted as a treatment characteristic rather than a causal exposure because treatment intensification can reflect underlying hypertension severity (confounding by indication).

Blood-pressure measurement and outcome definition

Office BP was measured with a calibrated sphygmomanometer after the participant had rested in the seated position. Two readings were obtained during the same assessment and their average was used for analysis. Controlled hypertension was defined as office BP <140/90 mm Hg and uncontrolled hypertension as systolic BP ≥140 mm Hg and/or diastolic BP ≥90 mm Hg in a participant with known hypertension or current antihypertensive treatment (2, 3). Home and ambulatory BP monitoring were not performed. The study record did not preserve the exact seated-rest duration, interval between readings, arm used, cuff size, sphygmomanometer model, calibration schedule, or the timing of measurement relative to the most recent antihypertensive dose. These protocol details therefore could not be analysed and are treated as measurement limitations rather than reconstructed retrospectively.

Statistical analysis

Continuous variables were summarized as mean ± standard deviation (SD) and categorical variables as frequency and percentage. Controlled and uncontrolled groups were compared using the independent-samples t test for continuous variables and the chi-square test or Fisher exact test for categorical variables, as appropriate. For Table 3, unadjusted odds ratios (ORs) with 95% confidence intervals (CIs) were calculated from the corresponding 2×2 distributions. The seven variables with statistically significant bivariate associations (p<0.05) – BMI ≥25 kg/m², hypertension duration > five years, diabetes mellitus, high salt intake, physical inactivity, poor medication adherence and missed follow-up visits – were entered simultaneously into the multivariable logistic-regression model; no additional variables were forced into the model. The model therefore estimated seven covariate coefficients plus an intercept. Statistical significance was set at a two-sided p<0.05. Ten bivariate comparisons were exploratory and no multiplicity adjustment was applied; effect sizes and 95% CIs were therefore emphasized. The archived analysis output did not contain a verifiable software-version record or formal variance-inflation-factor, calibration, or discrimination statistics. Accordingly, these diagnostics are not reported and possible overfitting and residual multicollinearity are acknowledged explicitly when interpreting the adjusted estimates.

Ethical considerations

The present study was approved by the Institutional Ethics Committee of Government Medical College, Bhadradri Kothagudem, Telangana, India (IEC/GMCBK/2025/02). Written informed consent was obtained from all participants before enrolment. Confidentiality was maintained, and anonymized data were used for analysis.

RESULTS

A total of 100 adults with hypertension were enrolled and all 100 had complete data for the final analysis. The mean age of the study population was 56.8 ± 11.4 years. Participants aged 41-60 years constituted the largest group (49.0%), followed by those aged >60 years (39.0%). Men accounted for 58.0% of the cohort and 63.0% were from rural areas. Sixty-three participants had BMI ≥25 kg/m² and 57.0% had hypertension for more than five years. Diabetes mellitus was the most frequent documented comorbidity (40.0%), followed by dyslipidaemia (34.0%) (Table 1).

Uncontrolled office hypertension was present in 62 participants (62.0%), whereas 38 had controlled BP. This proportion describes the present tertiary-care cohort and should not be interpreted as a population prevalence estimate for Telangana or India. The uncontrolled group had higher mean systolic and diastolic BP values and a longer mean duration of hypertension. Use of two or more antihypertensive agents was also more frequent in the uncontrolled group; this treatment pattern is compatible with greater underlying di sease severity and should not be interpreted as evidence that multidrug treatment caused poor control (Table 2).

In bivariate analysis, excess body weight, hypertension duration > five years, diabetes mellitus, high salt intake, physical inactivity, poor medication adherence and missed follow-up visits were associated with uncontrolled hypertension. Poor medication adherence showed the largest unadjusted association (OR 9.67, 95% CI 3.06-30.54). Age >60 years, current smoking and alcohol use had ORs above 1.0 but their 95% CIs included the null value. Because multiple exploratory comparisons were performed, these findings should be interpreted together with effect sizes and CIs rather than p-values alone (Table 3).

In the multivariable model, poor medication adherence, BMI ≥25 kg/m², physical inactivity, high salt intake and hypertension duration longer than five years remained statistically significant. Diabetes mellitus and missed follow-up visits, although significant in bivariate analysis, did not remain statistically significant after adjustment. The adjusted estimates were imprecise, as reflected by the broad CIs, and should not be interpreted as exact measures of causal effect (Table 4).

TABLE 1.

Baseline demographic and clinical characteristics of the study population (n=100)

TABLE 2.

Blood-pressure profile and treatment intensity by control status

TABLE 3.

Bivariate associations with uncontrolled hypertension

TABLE 4.

Multivariable logistic-regression analysis of variables associated with uncontrolled hypertension

DISCUSSION

In this tertiary-care cohort, 62.0% of participants met the study definition of uncontrolled office hypertension. This proportion should be interpreted as a feature of a selected hospital population rather than as an estimate of community prevalence. Patients referred to or admitted in a tertiary centre can have longer di sease duration, greater comorbidity and more complex treatment needs than individuals managed solely in primary care. The finding is nevertheless consistent with broader evidence that BP control remains incomplete despite increasing treatment availability (1, 5, 6, 14).

Medication-taking behaviour showed the strongest adjusted association. Participants classified as having poor adherence had higher odds of uncontrolled hypertension after adjustment (aOR 6.84, 95% CI 2.02-23.15). This direction is consistent with systematic evidence linking antihypertensive nonadherence with poor control (9). The magnitude, however, should be interpreted cautiously because adherence was assessed by non-validated self-report and the CI was wide. Self-report can be affected by recall and social-desirability bias, often leading patients to under-report missed doses; such misclassification can distort the observed association in either direction, although under-recognition of nonadherence is a particular concern (10).

Body mass index ≥25 kg/m², physical inactivity, high salt intake and hypertension duration longer than five years also remained associated with uncontrolled hypertension in the adjusted model. These findings are biologically plausible and align with evidence relating adiposity, sodium exposure and low physical activity to elevated BP (11-13). Nevertheless, the study did not quantify sodium intake or objectively measure physical activity, so the corresponding estimates reflect self-reported exposure categories rather than precise dose-response relationships. Diabetes mellitus and missed follow-up visits were significant in bivariate analysis but no longer reached statistical significance after adjustment; they should therefore not be described as independently associated factors in this dataset.

Treatment intensity deserves separate interpretation. Use of two or more antihypertensive agents was more frequent in the uncontrolled group, but this association is susceptible to confounding by indication: patients with more severe or difficult-to-control hypertension are more likely to receive treatment escalation. The cross-sectional design cannot determine whether multidrug therapy preceded poor control or was prescribed in response to it. Drug-class and fixed-dose-combination information was not available, preventing assessment of regimen composition. Similarly, socioeconomic status, medication affordability and access, laboratory measures of renal function and concomitant medicines capable of increasing blood pressure were not systematically captured, leaving potential residual confounding.

The present study also has important measurement constraints. Blood pressure was classified from an office assessment at a single encounter, without home or ambulatory monitoring. White-coat effect or white-coat uncontrolled hypertension therefore cannot be excluded and readings could also have been influenced by anxiety, pain, acute illness, recent activity, or medication timing. In addition, the exact cuff-selection procedure, arm, rest duration, interval between readings, device details and calibration schedule were not retained in the study record. These limitations reduce confidence in the precision of the office classification and reinforce the need for cautious interpretation.

The adjusted model identified several associations that are clinically actionable as targets for assessment, but the present study does not demonstrate that modifying them will reduce cardiovascular events or improve long-term outcomes. Prospective studies with repeated standardized BP measurements, home or ambulatory monitoring, validated adherence and lifestyle instruments, detailed treatment characterization and a larger sample would be better suited to assess temporality and intervention effects. Within routine tertiary-care practice, the current findings support careful review of adherence, weight, dietary salt, physical activity and continuity of follow-up as components of comprehensive hypertension management, while recognizing that the benefits of specific interventions require longitudinal evaluation.

Limitations of the study

The study has several limitations. Its cross-sectional design precludes temporal or causal inference. The sample was small, drawn by consecutive non-probability sampling from a single tertiary centre and may over-represent patients with more complex or difficult-to-control di sease. The retained screening record did not contain aggregate screening/exclusion counts or outpatient/inpatient recruitment proportions. Blood pressure status was based on a single-visit office assessment without home or ambulatory confirmation, so white-coat effects and short-term clinical influences cannot be excluded. Several exposures – including adherence, salt intake, physical activity, smoking, alcohol use and follow-up behaviour – were self-reported and measured without validated scales or fixed quantitative thresholds, increasing the potential for recall, social-desirability and misclassification bias. Socioeconomic status, medication access, detailed renal function, BP-raising concomitant medicines, specific antihypertensive drug classes and fixed-dose combinations were not systematically assessed. The seven-covariate regression model produced wide CIs and may be vulnerable to overfitting; formal multicollinearity, calibration and discrimination diagnostics were not retained with the archived analysis output. Multiple bivariate tests were performed without multiplicity correction. These issues, together with residual confounding, limit the precision and generalizability of the findings.

CONCLUSION

Uncontrolled office hypertension was frequent in this selected tertiary-care cohort. Poor medication adherence, BMI ≥25 kg/m², physical inactivity, high salt intake and hypertension duration longer than five years remained statistically associated with uncontrolled blood pressure after adjustment, whereas diabetes mellitus and missed follow-up visits did not. These results should be interpreted as cross-sectional associations rather than causal risk factors. The findings support systematic assessment of medication-taking behaviour and modifiable lifestyle characteristics during hypertension care, but they do not establish that targeted interventions will reduce cardiovascular outcomes. Larger prospective studies using standardized repeated blood-pressure measurements and validated exposure measures are needed to confirm the observed relationships.

Conflicts of interest

none declared.

Financial support

none declared.