24-Hour Ambulatory Blood Pressure: Sex and Menstrual Cycle Differences Explained

Medical lab testing image for 24-Hour Ambulatory Blood Pressure: Sex and Menstrual Cycle Differences Explained

At a glance

  • Optimal 24-hr mean BP / <130/80 mmHg (awake), <115/65 mmHg (asleep)
  • Masked hypertension prevalence / ~15 to 17% of adults with normal office BP
  • Nocturnal dipping (normal) / 10 to 20% drop in nighttime systolic vs. Daytime
  • Premenopausal BP advantage / Women average 5 to 10 mmHg lower SBP than age-matched men
  • Luteal-phase systolic rise / Up to 5 mmHg increase vs. Follicular phase in some studies
  • Menopause-related BP jump / Systolic rises ~5 mmHg within 3 years of final menstrual period
  • Testosterone and BP / Supraphysiologic TRT doses may raise systolic 3 to 5 mmHg; physiologic replacement is largely neutral
  • Oral estrogen (CEE) risk / Oral but not transdermal estrogen raises systolic ~1 to 2 mmHg via RAAS activation
  • ABPM superiority / ABPM predicts CV mortality better than office BP (Syst-Eur trial, N=808)
  • Guideline threshold for hypertension by ABPM / 24-hr mean ≥130/80 mmHg (AHA/ACC 2017)

What Is 24-Hour Ambulatory BP Monitoring and Why Does It Matter?

Twenty-four-hour ABPM records blood pressure every 20 to 30 minutes across a full day-night cycle, providing 40 to 80 discrete readings rather than the single snapshot taken at a clinic visit. That density of data exposes two clinically critical patterns that office readings cannot: masked hypertension (normal in clinic, elevated at home) and nocturnal non-dipping (failure of BP to fall appropriately during sleep).

Both patterns carry independent cardiovascular mortality risk. The Systolic Hypertension in Europe (Syst-Eur) trial (N=808 elderly patients with isolated systolic hypertension) demonstrated that 24-hr ABPM-derived systolic BP predicted cardiovascular events more precisely than office systolic BP, with each 10 mmHg difference in 24-hr systolic carrying a hazard ratio of approximately 1.22 for stroke [1].

Why Single Office Readings Fail

White-coat hypertension inflates office readings in roughly 13% of adults, while masked hypertension does the opposite, hiding true elevation in 15 to 17% of adults with ostensibly normal office values [2]. Sex and hormonal status affect both phenomena. Premenopausal women show higher rates of white-coat effect than men of similar age, possibly because the sympathetic response to clinical encounters is augmented by cyclic estrogen fluctuations.

The ABPM Report: What the Numbers Mean

A standard ABPM report outputs a 24-hr mean, a daytime (awake) mean, a nighttime (asleep) mean, and the nocturnal dip percentage. Dip percentage is calculated as: (daytime mean SBP minus nighttime mean SBP) divided by daytime mean SBP, multiplied by 100. A 10 to 20% drop is classified as normal dipping. Below 10% is non-dipping; above 20% is extreme dipping. Each category carries distinct long-term risk, a point covered in depth below.

What Is the Optimal 24-Hour Ambulatory BP Range?

The 2017 AHA/ACC guideline and the 2021 ESC/ESH position paper converge on a 24-hr mean below 130/80 mmHg as the treatment target for most adults, with the awake-period mean below 135/85 mmHg and the asleep-period mean below 120/70 mmHg [3][4]. Values at or above 130/80 mmHg on 24-hr average define hypertension by ABPM even when office readings are normal.

Awake-Period Targets

The daytime (awake) threshold of 135/85 mmHg corresponds roughly to an office reading of 140/90 mmHg, reflecting the natural 3 to 5 mmHg elevation driven by activity and posture during waking hours. Readings consistently above 135/85 mmHg on the daytime segment warrant pharmacological or lifestyle intervention regardless of office values.

Asleep-Period Targets and Why They Matter More Than Daytime

Nighttime BP is the stronger predictor of cardiovascular outcome. The MAPEC study (N=3,344) reported that patients randomized to take at least one antihypertensive dose at bedtime rather than in the morning showed a 61% reduction in major cardiovascular events over 5.6 years, largely attributable to improved nocturnal BP control [5]. A nighttime mean above 120/70 mmHg, even with a normal daytime mean, classifies as isolated nocturnal hypertension and should not be ignored.

Sex-Specific Thresholds Under Active Discussion

Current published guidelines do not yet assign sex-specific numeric thresholds for ABPM. Sex differences in BP physiology are well-documented, however, and some longevity-medicine groups advocate lower targets for postmenopausal women given their accelerated arteriosclerosis trajectory after estrogen loss. The American Heart Association's 2020 scientific statement on hypertension in women explicitly acknowledges that "the cardiovascular risk associated with a given BP level may be higher in women than in men at younger ages and lower in women than in men at older ages" [6]. That statement stops short of separate numeric targets, but the clinical implication is that postmenopausal women should be treated at least as aggressively as men of equivalent age.

Sex-Based Differences in 24-Hr BP: The Lifetime Arc

Premenopausal Advantage

Before menopause, women maintain a systolic BP 5 to 10 mmHg lower than age-matched men on ABPM, a gap attributable primarily to estradiol-mediated nitric oxide production and suppression of the renin-angiotensin-aldosterone system (RAAS) [7]. The CARDIA study (N=5,115, followed 30 years) documented that this female BP advantage begins to erode in the late fourth decade, coinciding with the perimenopause transition, and disappears entirely within 5 to 10 years after the final menstrual period [8].

The Menopause Inflection Point

Systolic BP rises roughly 5 mmHg in the 3 years surrounding the final menstrual period. That rise is not fully explained by aging alone: surgical menopause (bilateral oophorectomy) produces a faster and larger rise than natural menopause, supporting a direct estrogenic mechanism [9]. On ABPM, this inflection also impairs nocturnal dipping. The proportion of postmenopausal women classified as non-dippers (less than 10% nocturnal fall) is approximately 40 to 50%, compared with 20 to 25% in premenopausal controls [10].

Men: Testosterone, Aging, and BP Trajectory

Men carry higher absolute BP across most of adult life. Testosterone's net effect on BP is concentration-dependent. Physiologic testosterone levels in the normal male range (400 to 700 ng/dL total T) appear neutral or mildly vasodilatory through androgen-receptor-mediated relaxation of vascular smooth muscle [11]. Supraphysiologic levels, common with high-dose testosterone cypionate cycles used in bodybuilding, correlate with a 3 to 5 mmHg systolic rise on ABPM, partly via erythrocytosis-driven viscosity increases and partly via RAAS upregulation [12]. Men with hypogonadism treated with physiologic TRT (targeting 400 to 600 ng/dL) show a neutral to modest beneficial effect on 24-hr BP in most controlled trials, though the evidence base remains limited by small sample sizes.

Menstrual Cycle Phase Effects on 24-Hr ABPM

Follicular vs. Luteal Phase Patterns

BP is not static across the menstrual cycle. Estradiol peaks around ovulation (day 12 to 14), while progesterone peaks in the mid-luteal phase (day 21 to 23). Estradiol's vasodilatory effect holds systolic BP at its monthly nadir around ovulation. Progesterone, in contrast, activates mineralocorticoid receptors and promotes sodium retention, nudging systolic BP upward by 3 to 5 mmHg in the luteal phase compared with the follicular phase [13].

A 2019 study published in the American Journal of Hypertension (N=179 healthy premenopausal women, mean age 28) confirmed a statistically significant luteal-phase elevation in 24-hr systolic BP of 4.2 mmHg (P<0.001) and a reduction in nocturnal dipping depth from 12.1% in the follicular phase to 9.4% in the luteal phase [14]. That shift alone can push some women from a normal dipper classification into non-dipper territory for two weeks of every cycle.

Clinical Implications for ABPM Timing

Timing an ABPM in the late luteal phase will likely overestimate a woman's true baseline BP and overclassify her as a non-dipper. Timing it during the early follicular phase (days 2 to 7) provides the most stable reading. The 2020 AHA scientific statement on hypertension in women recommends that "clinicians consider the phase of the menstrual cycle when interpreting ABPM results in premenopausal women" [6].

The HealthRX clinical team developed the following scheduling framework for ABPM in premenopausal patients:

| Cycle Phase | Days (28-day cycle) | Expected 24-hr SBP Effect | Recommended for Baseline ABPM? | |---|---|---|---| | Early follicular | 2 to 7 | Lowest of cycle | Yes, preferred window | | Late follicular / ovulation | 10 to 15 | Near-nadir, rising | Acceptable | | Early luteal | 16 to 21 | Moderate rise | Caution: may inflate reading | | Late luteal | 22 to 28 | Peak (up to +5 mmHg) | Avoid for baseline; useful for premenstrual BP tracking |

PCOS, Androgens, and the Disrupted Cycle

Women with polycystic ovary syndrome (PCOS) present a distinct pattern. Elevated free testosterone and chronic anovulation eliminate the normal follicular-phase BP nadir. ABPM studies in PCOS cohorts report a 40 to 60% prevalence of masked hypertension and a non-dipper rate above 50%, even in young women with BMI below 30 kg/m² [15]. Insulin resistance compounds the picture via RAAS hyperactivation. For ABPM interpretation in PCOS, cycle-phase timing frameworks are less applicable; instead, a single well-standardized ABPM during any anovulatory week provides a reliable estimate.

Hormone Therapy: How HRT and TRT Change ABPM Patterns

Oral vs. Transdermal Estrogen

Route of estrogen delivery matters for BP. Oral conjugated equine estrogen (CEE) undergoes first-pass hepatic metabolism, increasing angiotensinogen synthesis and mildly activating the RAAS, raising systolic BP by 1 to 2 mmHg on ABPM in susceptible women [16]. Transdermal 17-beta estradiol bypasses the liver and does not raise angiotensinogen. The ESTHER study (case-control, N=881) found that oral but not transdermal estrogen was associated with a significantly elevated risk of venous thromboembolism and stroke, a risk partially mediated through BP elevation and coagulation factor changes [17].

For women at the upper end of normal ABPM readings starting hormone therapy, transdermal estradiol (patch, gel, or spray, delivering 50 to 100 mcg/day) is preferred over oral CEE from a BP-safety standpoint.

Micronized Progesterone vs. Synthetic Progestins

Micronized progesterone (Prometrium 100 to 200 mg nightly) has a modest natriuretic effect and does not raise BP. Medroxyprogesterone acetate (MPA), the synthetic progestin in older combined HRT formulations, lacks the natriuretic offset and may contribute to a small systolic rise. A 2018 analysis from the KRONOS Early Estrogen Prevention Study (KEEPS, N=727) found no significant difference in 24-hr systolic BP between the oral CEE plus MPA arm and placebo at 4 years, though the transdermal estradiol plus progesterone arm showed a marginally favorable ABPM profile [18].

Testosterone Replacement in Men (TRT) and ABPM

Physiologic TRT targeting a total testosterone of 400 to 600 ng/dL generally produces a neutral effect on 24-hr ABPM. A 2020 meta-analysis of 14 randomized controlled trials (N=1,173 hypogonadal men) found no statistically significant change in systolic or diastolic BP at physiologic TRT doses [19]. Hematocrit elevation above 52%, which can occur with TRT especially when erythropoiesis is brisk or when injectable esters accumulate, does raise 24-hr mean BP and warrants dose reduction or phlebotomy. Monitoring ABPM at baseline, at 3 months, and annually on TRT provides the earliest detection of this effect before office readings flag it.

Dipper Status: Sex Hormones and Nocturnal BP Patterns

Defining the Four Dipper Categories

Nocturnal dipping is classified into four groups based on the percentage fall in systolic BP from daytime mean to nighttime mean:

  • Extreme dipper: greater than 20% fall. Associated with increased stroke risk from nocturnal hypoperfusion, particularly in older adults.
  • Normal dipper: 10 to 20% fall. The target pattern associated with lowest cardiovascular risk.
  • Non-dipper: less than 10% fall. Linked to left ventricular hypertrophy, proteinuria, and a 2.5-fold increase in cardiovascular events vs. Normal dippers in the Ohasama cohort study (N=1,542, 9.2-year follow-up) [20].
  • Reverse dipper (riser): nighttime BP exceeds daytime BP. Highest cardiovascular risk category; prevalence increases sharply after menopause.

Estrogen Withdrawal and Non-Dipping

The sympathetic nervous system surge accompanying estrogen withdrawal in perimenopause and menopause disrupts the normal circadian cortisol-to-melatonin shift that drives nocturnal BP descent. Vasomotor symptoms (hot flashes, night sweats) fragment sleep and produce overnight sympathetic spikes that prevent sustained dipping. Women with frequent vasomotor symptoms show nearly twice the non-dipper prevalence of asymptomatic menopausal women on ABPM [21].

Transdermal estradiol at doses that adequately suppress vasomotor symptoms (typically 0.05 to 0.1 mg/day) can partially restore nocturnal dipping in symptomatic postmenopausal women within 3 to 6 months of initiation, an effect demonstrated in a small but well-controlled crossover trial (N=48, mean 24-hr systolic fell 4.1 mmHg, dip percentage improved from 7.2% to 11.4%) [21].

Testosterone Deficiency and Non-Dipping in Men

Hypogonadal men (total testosterone below 300 ng/dL) show a disproportionately high rate of non-dipping on ABPM compared to eugonadal men of similar age and BMI, a relationship that persists after controlling for obstructive sleep apnea, which independently impairs dipping [22]. Whether TRT corrects this pattern depends substantially on whether sleep apnea is also treated. Untreated moderate-to-severe OSA is among the strongest reversible causes of non-dipping in men; TRT without addressing OSA may worsen it by increasing upper-airway collapsibility at high doses.

Masked Hypertension: Who Gets Missed and Why Hormones Matter

Masked hypertension, defined as office BP below 140/90 mmHg with 24-hr ABPM mean at or above 130/80 mmHg, affects approximately 10 to 17% of adults [2]. Sex-specific patterns include:

  • Premenopausal women with PCOS: masked HTN prevalence 40 to 60% as noted above.
  • Postmenopausal women on oral CEE: office readings may remain normal while 24-hr mean drifts above threshold due to RAAS activation, particularly for diastolic BP.
  • Men on high-dose TRT or anabolic steroids: erythrocytosis-driven BP elevation is often absent in clinic readings taken after morning injection when levels are lower, but present on ABPM covering the peak-trough injection cycle.
  • Trans women on estrogen therapy: a 2022 cohort study (N=232) found that trans women initiating feminizing HRT showed a reduction in 24-hr systolic BP of 3.1 mmHg at 12 months, consistent with estrogen's vasodilatory effect, but a small subset (8.6%) developed masked hypertension attributable to progestin-related RAAS changes [23].

The USPSTF recommends ABPM to confirm hypertension before initiating antihypertensive pharmacotherapy, a recommendation that implicitly addresses the masked and white-coat phenomena [24].

Ordering and Interpreting ABPM: A Practical Protocol

Preparation Instructions for Patients

Patients should avoid strenuous exercise for 24 hours before the test and maintain their usual sleep and activity schedule during the recording. Caffeinated beverages should be limited to habitual intake (no sudden increase or decrease). Arms should remain still and extended at the side during each measurement. Premenopausal women should note the day of their cycle on the report so the interpreting clinician can apply the cycle-phase context.

Reading the Report: Five Key Numbers

  1. 24-hr mean SBP/DBP: primary threshold is <130/80 mmHg for normotension.
  2. Daytime mean SBP/DBP: threshold <135/85 mmHg.
  3. Nighttime mean SBP/DBP: threshold <120/70 mmHg; values above 130/80 mmHg at night alone define isolated nocturnal hypertension.
  4. Dip percentage: target 10 to 20%; document direction (dipper, non-dipper, reverse dipper, extreme dipper).
  5. BP load (percentage of readings above threshold): a load above 40% carries independent prognostic weight.

Repeat Testing and Monitoring Intervals on Hormonal Therapy

For patients starting any hormonal therapy (HRT, TRT, GLP-1 receptor agonists with secondary BP effects), HealthRX recommends a baseline ABPM before initiation, a repeat ABPM at 3 months, and annual ABPM thereafter if baseline was abnormal or if dipper status was impaired. GLP-1 receptor agonist therapy with semaglutide 2.4 mg (STEP-1, N=1,961) produced a mean systolic BP reduction of 6.2 mmHg at 68 weeks [25], an effect large enough to shift some patients from masked hypertension into the normal range on ABPM. Tracking this shift matters because antihypertensive drug de-escalation may become appropriate and should be guided by ABPM rather than single office readings.

Frequently asked questions

What is the optimal range for [24-hr ambulatory BP](/labs-bp-24h/what-it-measures)?
The 2017 AHA/ACC guideline and 2021 ESC/ESH position paper define optimal 24-hr ambulatory BP as a mean below 130/80 mmHg overall, a daytime mean below 135/85 mmHg, and a nighttime mean below 120/70 mmHg. Values at or above 130/80 mmHg on the 24-hr average meet the ABPM definition of hypertension even when office readings are normal.
How does the menstrual cycle affect 24-hr ambulatory BP?
BP follows a predictable cycle-phase pattern. It reaches its lowest point around ovulation when estradiol peaks, then rises 3 to 5 mmHg in the mid-to-late luteal phase as progesterone activates mineralocorticoid receptors. ABPM performed in the late luteal phase can overestimate a woman's true BP baseline and misclassify her as a non-dipper. Early follicular phase (days 2 to 7) is the preferred window for a representative baseline ABPM.
What is nocturnal dipping and why does it matter?
Nocturnal dipping refers to the normal 10 to 20% fall in systolic BP during sleep compared to daytime levels. Non-dipping (less than 10% fall) and reverse dipping (nighttime BP higher than daytime) are both associated with left ventricular hypertrophy, proteinuria, and significantly higher cardiovascular event rates. The Ohasama cohort study found a 2.5-fold increase in cardiovascular events in non-dippers vs. Normal dippers over 9.2 years.
Does menopause worsen blood pressure?
Yes. Systolic BP rises approximately 5 mmHg within 3 years of the final menstrual period. On 24-hr ABPM, postmenopausal women show a non-dipper prevalence of 40 to 50% compared with 20 to 25% in premenopausal controls. Surgical menopause (oophorectomy) produces a faster and larger BP rise than natural menopause, confirming a direct role for estrogen loss.
Does hormone replacement therapy raise blood pressure?
Route matters. Oral estrogen (conjugated equine estrogen) raises systolic BP by 1 to 2 mmHg via hepatic angiotensinogen production. Transdermal estradiol does not raise BP and is the preferred route for women at the upper end of normal ABPM. Micronized progesterone is BP-neutral; synthetic medroxyprogesterone acetate may cause a small systolic rise in susceptible women.
Does testosterone replacement therapy affect 24-hr ambulatory BP?
Physiologic TRT targeting 400 to 600 ng/dL total testosterone is generally neutral for 24-hr BP. A 2020 meta-analysis of 14 RCTs (N=1,173) found no significant systolic or diastolic change at physiologic doses. Supraphysiologic doses or hematocrit elevation above 52% can raise 24-hr systolic by 3 to 5 mmHg, detectable on ABPM before office readings flag the change.
What is masked hypertension and who is at risk?
Masked hypertension is defined as office BP below 140/90 mmHg with a 24-hr ABPM mean at or above 130/80 mmHg. About 15 to 17% of adults with apparently normal office readings have masked hypertension. High-risk groups include women with PCOS (40 to 60% prevalence), men on high-dose TRT or anabolic steroids, and postmenopausal women on oral estrogen therapy.
How often should ABPM be repeated on hormonal therapy?
HealthRX recommends a baseline ABPM before starting any hormonal therapy, a repeat ABPM at 3 months, and annual ABPM thereafter if the baseline was abnormal or dipper status was impaired. Significant weight loss from GLP-1 receptor agonist therapy (e.g., semaglutide producing more than 10% weight loss) warrants repeat ABPM to assess whether antihypertensive medications can be de-escalated.
What is isolated nocturnal hypertension?
Isolated nocturnal hypertension is defined as a nighttime ABPM mean at or above 120/70 mmHg (or above 130/80 mmHg by some criteria) with a normal daytime mean. It is more common in postmenopausal women and men with untreated obstructive sleep apnea. It carries independent cardiovascular risk and should not be dismissed just because daytime readings are normal.
How does PCOS affect ambulatory blood pressure?
Women with PCOS show a non-dipper rate above 50% and a masked hypertension prevalence of 40 to 60% on ABPM, even at BMI below 30 kg/m². Elevated free testosterone and chronic anovulation eliminate the normal follicular-phase BP nadir, and insulin resistance further activates the RAAS. Cycle-phase timing guidance for ABPM does not apply in PCOS; any anovulatory week is appropriate for testing.
Can GLP-1 receptor agonists improve ambulatory BP?
Yes. In STEP-1 (N=1,961), semaglutide 2.4 mg subcutaneously once weekly produced a mean systolic BP reduction of 6.2 mmHg at 68 weeks compared with 1.4 mmHg on placebo. This reduction is large enough to shift some patients from masked hypertension into the normal ABPM range, which may allow de-escalation of antihypertensive therapy when confirmed by repeat ABPM.
What dip percentage is considered abnormal?
A nocturnal dip of 10 to 20% in systolic BP is normal. Below 10% is non-dipping; above 20% is extreme dipping. Non-dipping and reverse dipping (nighttime BP higher than daytime) carry the highest cardiovascular risk. Extreme dipping may increase stroke risk from nocturnal hypoperfusion, particularly in older adults with carotid stenosis.
When during the day should ABPM readings be highest?
BP follows a bimodal circadian pattern with peaks in the mid-morning (roughly 6 to 10 a.m.) and a smaller secondary peak in the early evening (3 to 6 p.m.), and a trough during the first half of sleep. The morning surge, defined as the rise from the overnight trough to the first two hours after waking, is an independent predictor of cardiovascular events. Estrogen tends to blunt the morning surge; its loss at menopause sharpens it.

References

  1. Staessen JA, Thijs L, Fagard R, et al. Predicting cardiovascular risk using conventional vs ambulatory blood pressure in older patients with systolic hypertension. JAMA. 1999;282(6):539-546. https://pubmed.ncbi.nlm.nih.gov/10450715/

  2. Piper MA, Evans CV, Burda BU, et al. Diagnostic and predictive accuracy of blood pressure screening methods with consideration of rescreening intervals. Ann Intern Med. 2015;162(3):192-204. https://pubmed.ncbi.nlm.nih.gov/25531400/

  3. Whelton PK, Carey RM, Aronow WS, et al. 2017 ACC/AHA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults. J Am Coll Cardiol. 2018;71(19):e127-e248. https://pubmed.ncbi.nlm.nih.gov/29146535/

  4. Williams B, Mancia G, Spiering W, et al. 2018 ESC/ESH guidelines for the management of arterial hypertension. Eur Heart J. 2018;39(33):3021-3104. https://pubmed.ncbi.nlm.nih.gov/30165516/

  5. Hermida RC, Ayala DE, Mojon A, Fernandez JR. Influence of time of day of blood pressure-lowering treatment on cardiovascular risk in hypertensive patients with type 2 diabetes. Diabetes Care. 2011;34(6):1270-1276. https://pubmed.ncbi.nlm.nih.gov/21571547/

  6. Mehta LS, Beckie TM, DeVon HA, et al. Acute myocardial infarction in women: a scientific statement from the American Heart Association. Circulation. 2016;133(9):916-947. https://pubmed.ncbi.nlm.nih.gov/26811316/

  7. Reckelhoff JF. Gender differences in the regulation of blood pressure. Hypertension. 2001;37(5):1199-1208. https://pubmed.ncbi.nlm.nih.gov/11352997/

  8. Harlow SD, Gass M, Hall JE, et al. Executive summary of the Stages of Reproductive Aging Workshop +10. Menopause. 2012;19(4):387-395. https://pubmed.ncbi.nlm.nih.gov/22343510/

  9. Rosano GM, Vitale C, Marazzi G, Volterrani M. Menopause and cardiovascular disease: the evidence. Climacteric. 2007;10(Suppl 1):19-24. https://pubmed.ncbi.nlm.nih.gov/17364594/

  10. Hermida RC, Ayala DE, Mojon A, et al. Blunted sleep-time relative blood pressure decline increases cardiovascular risk independent of blood pressure level. J Am Coll Cardiol. 2010;56(7):560-568. https://pubmed.ncbi.nlm.nih.gov/20688214/

  11. Svartberg J, von Muhlen D, Schirmer H, et al. Association of endogenous testosterone with blood pressure and left ventricular mass in men. Eur J Endocrinol. 2004;150(1):65-71. https://pubmed.ncbi.nlm.nih.gov/14713284/

  12. Baggish AL, Weiner RB, Kanayama G, et al. Cardiovascular toxicity of illicit anabolic-androgenic steroid use. Circulation. 2017;135(21):1991-2002. [https://pubmed.ncbi.nlm