Published Sep 30, 2026 | 7:00 AM ⚊ Updated Sep 30, 2026 | 7:00 AM
The study pooled data from 13 studies covering 14,311 patients and put the prevalence of resistant hypertension at 19 percent.
Synopsis: A new meta-analysis estimates that 19 percent of Indian hypertension patients may have resistant hypertension, but the figure varies sharply with how doctors define and measure resistance. Medication-adherence checks and ambulatory BP monitoring significantly alter prevalence estimates, highlighting the need to distinguish true resistance from apparent or pseudo-resistance.
Nearly one in five people with hypertension in India may have a form of high blood pressure that resists standard treatment, according to a new meta-analysis from researchers at the University of Hyderabad.
The study pooled data from 13 studies covering 14,311 patients and put the prevalence of resistant hypertension at 19 percent. But there’s a caveat: the numbers differ depending on how each study defined and measured resistance.
The study appears in Clinical Epidemiology and Global Health. Researchers from AIIMS Deoghar, AIIMS Jodhpur, Nitte University in Mangaluru, Mediciti Institute of Medical Sciences and the University of Nevada, Las Vegas, contributed to the analysis.
Also Read: Love your tea/coffee scalding hot? The habit might be tripling your risk of oesophageal cancer
Doctors classify a patient as having resistant hypertension (RH) when three drug classes, including a diuretic—a medication that helps your body get rid of extra salt and water through urine—fail to bring blood pressure into range. The label carries real consequences: it triggers referrals to specialists and changes how a doctor plans long-term care.
“RH is generally defined as uncontrolled blood pressure despite the use of three or more antihypertensive drugs, including a diuretic, at optimal dosages,” the authors wrote. “Some definitions also include patients who require four or more medications to achieve target blood pressure levels.”
Several factors can push patients into this category. The researchers named obesity, inactivity, smoking and salt intake as contributors. Diabetes, cholesterol imbalance, kidney disease and sleep apnoea also raise the risk.
One factor stands apart from the rest: missed doses. A patient who skips medicine can look resistant on paper without truly resisting treatment at all.
This is where the study lands its main finding. The prevalence estimate changes depending on how researchers classify resistance.
Doctors do not all use the same rulebook to decide who counts as resistant. Some demand strict proof: three specific drug classes at full doses, along with checks for missed pills. Others use a looser standard: just three drugs and a high reading. This gap in the rulebook explains much of the swing in this study.
Studies using strict definitions found a pooled prevalence of 14 percent. Simplified definitions pushed that figure to 20 percent. When researchers checked medication adherence—whether a patient actually takes pills as prescribed—and folded that check into their classification, the number fell to 11 percent. Skip that check, and it rises to 33 percent.
Ambulatory blood pressure monitoring, a small device a patient wears for a day that records blood pressure every 15 to 30 minutes rather than relying on one clinic visit, produced the largest gap. Studies using this method found a 45 percent prevalence. Studies that skipped it, or left it unclear, reported 16 percent or lower.
A single clinic reading can mislead in both directions. Some patients spike under the stress of a doctor’s visit, a pattern doctors call white-coat hypertension. Others show normal readings at the clinic but stay high the rest of the day, a pattern called masked hypertension. Wearable monitors can catch both patterns; a single clinic reading may catch neither.
Doctors call the gap between real and apparent resistance “pseudo-resistance.” A patient can fall into this category because of anxiety at the clinic, inconsistent dosing or a single faulty reading that never gets rechecked outside a hospital setting.
“This highlights the importance of distinguishing true resistant hypertension from apparent or pseudo-resistant hypertension,” the authors wrote, “particularly by assessing medication adherence, ambulatory blood pressure and other causes of persistently elevated readings.”
The analysis broke prevalence down by region. The South reported 20 percent, the West 22 percent, the North 17 percent and the East 14 percent.
Geography also skewed the evidence pool. Seven of the 13 studies came from the South. The North contributed just one. The East and West each supplied two.
A handful of studies from one part of the country cannot speak reliably for the rest of it. When most of the evidence comes from a single region, researchers cannot tell whether a number reflects a real pattern in that region or simply where scientists happened to run their studies.
“This imbalance limits how representative the findings are for the country as a whole,” the researchers said.
Also Read: Super El Niño: Karnataka may lose 1,870 lives, Andhra Pradesh and Kerala 1,680 each
The study tested five factors against resistant hypertension: diabetes, obesity, cholesterol, male sex and smoking.
Researchers measure a factor’s link to a condition through an odds ratio, a number that shows how much more likely one group is to have that condition compared with another. A ratio of 1 means no difference. A ratio above 1 means higher odds.
Obesity showed the strongest numerical link, with a pooled odds ratio of 3.81, meaning people with obesity showed nearly four times the odds of resistant hypertension compared with people without it. Diabetes followed at 2.95, while cholesterol imbalance stood at 1.83. Male sex sat at 1.44. Smoking barely moved the needle at 0.98, showing almost no difference between smokers and non-smokers.
These ratios alone cannot confirm a real link. Researchers also calculate a confidence interval, a range around each ratio that shows how much uncertainty surrounds the estimate. A narrow range indicates a more precise estimate; a wide range means the true value could fall across a much broader span.
When that range stretches down past 1, the point where a factor makes no difference at all, the finding cannot rule out coincidence. Every factor in this study crossed that line, which is why the authors called these links exploratory rather than confirmed.
“These findings should be considered exploratory signals rather than established risk factors,” the authors wrote, “given the use of crude study-level estimates, wide CIs, and substantial heterogeneity across several factors.”
Earlier research offers some support for these links. People with diabetes tend to carry more coexisting conditions overall, and past work ties sustained blood pressure exposure to heart events in this group. Obesity has separately emerged as a barrier to blood pressure control in other populations. This meta-analysis cannot confirm either link in India on its own.
Also Read: Goitre down, thyroid autoimmunity up: What four decades of iodised salt show in Indian children
A pooled result like 19 percent comes from averaging across 13 different studies. But researchers also want to know what might happen in a hospital or region none of those studies covered. For that, they calculate a prediction interval, a range that estimates how a result might land in a setting the researchers never studied.
Here, that interval ran from 2 percent to 70 percent, a range wide enough to cover almost any prevalence a future study might report. In other words, the prevalence in a new population could fall near either end of that range, and this analysis alone cannot narrow it further.
“The prediction interval ranged from 2 percent to 70 percent,” the authors noted, “indicating that the prevalence could vary substantially in different Indian populations and settings.”
To check whether one unusual study skewed the whole result, researchers ran the analysis again and again, each time removing one study from the pool. This test produced estimates between 16 percent and 22 percent, which told the researchers that no single study drove the result. The pooled figure held up regardless of which study was left out.
The studies still disagreed with each other significantly overall. Researchers measure this disagreement through a heterogeneity score, a number that shows how much studies differ beyond what chance alone would explain. A figure above 50 percent is generally taken as substantial heterogeneity; at 98.2 percent, this analysis sits far into that territory. In practical terms, the 13 studies produced results that differed sharply rather than painting one consistent picture of resistant hypertension.
Also Read: India has no new anaemia figures. The ICMR survey meant to provide them is still being analysed
Resistant hypertension carries a higher risk of stroke, heart attack and kidney damage compared with hypertension that responds to treatment. A shaky prevalence figure complicates decisions that follow from it, from how hospitals allocate specialist care to how health campaigns target patients.
The authors pointed to a fix: measure resistance the same way everywhere.
“India needs studies using standardised definitions of resistant hypertension, systematic assessment of medication adherence and appropriate use of ABPM,” the authors write.
