Seasonal Variability of Blood Pressure in California


Joel Handler, MD, Kaiser Permanente, 411 Lakeview Avenue, Anaheim, CA 92807

A 69-year-old woman visiting Anaheim in southern California from Stockton in north central California in November requested a refill for lisinopril/hydrochlorothiazide (HCTZ) 20/25 mg daily on the basis of home blood pressures (BPs) that had been rising during the past month. She stated for the past 5 years that she had been taking lisinopril/HCTZ 20/25 mg from October through March each year, but did not require any antihypertensive medication between April and September. She first noticed her seasonal BP variability becoming extremely dizzy during warmer months with low blood pressures on antihypertensive medication.

Home BPs taken with a LifeSource UA-767 Plus monitor (A&D Medical, Milpitas, CA) twice a day 3 d/wk averaged 132/76 mm Hg during the warmer months off medication, but averaged 146/78 mm Hg during the months of October and November.

Physical examination revealed BPs of 152/80 mm Hg and 154/78 mm Hg with a normal cardiac examination, intact peripheral pulses, and no bruits. Her body mass index was 23 kg/m2 and she exercised regularly by walking 45 min 5 d/wk throughout the year. Electrocardiographic results were normal, creatinine was 1.2 mg/dL (normal 0.6 mg/dL to 1.1 mg/dL), estimated glomerular filtration rate 54 mL/min, sodium 142 meq/L (normal 135 meq/L to 145 meq/L), and potassium 4.2 meq/L (normal 3.5 meq/L to 5.0 meq/L).

The patient was given a prescription for lisinopril/HCTZ 20/25 mg×½ tablet daily with instructions to follow home BPs according to her usual method. If home BPs remained >135/85 mm Hg in 2 weeks, she was instructed to increase to a full tablet daily. She called back in 3 weeks, messaging that her home BPs were consistently running 118 mm Hg to 130/70–76 mm Hg on lisinopril/HCTZ 20/25 mg daily and that she was returning to northern California to resume her routine health care follow-up.


Observed Large-Population Seasonal BP Variability in Southern California

A periodic winter seasonal dip in hypertension control has been noted in the southern California Kaiser Permanente (KP) region since population data was collected in 2005. More than 660,000 adult patients with a diagnosis of hypertension have been identified of a total membership of approximately 3.4 million, representing an adult prevalence, aged 18 years and older, of 26%. Given the adult benchmark National Health and Nutrition Examination Survey prevalence of 29%,1 the southern California KP hypertension prevalence identification rate is 90%. Ninety percent of the potential adult population with hypertension has been identified. Despite overall improvement in hypertension control and prevalence identification since 2005, the difference between control, defined as last BP <140/90 mm Hg, in the nadir month of February compared with the zenith month of September, is consistently about 2% (Figure 1).

Figure 1.

  Hypertension control over time in the Southern California Permanente Medical Group, using two databases, demonstrating periodic winter seasonal dips in hypertension control with the overall rise in control. Permanente Online Interactive Network of Tools (POINT) presently includes approximately 660,000 adults 18 years and older with diagnosed hypertension. Clinical Strategic Goal (CSG) is a dataset designed as a surrogate for the Healthcare Effectiveness Data and Information Set (HEDIS) subpopulation and includes approximately 330,000 adults between the ages of 18 and 85 years with hypertension.

Various alternative explanations for the repetitive end-of-year performance dip have been entertained and found to be nonexistent or nonexplanatory: BP elevations associated with winter viral illness or use of decongestants, temporary reassignment of nursing support staff from BP duty to influenza vaccine administration, holiday-related alcohol consumption and dietary indiscretion, as well as the possibility of lack of ability to pay for medication on reaching the Medicare part D donut hole. End-of-year telephonic outreach programs to approximately 100,000 patients with uncontrolled hypertension have failed to blunt this performance dip. Although southern California, with a seasonal variation of only 15°F to 17°F in 3 major KP membership areas of Los Angeles, Anaheim, and San Diego (Table I) would appear to be an unlikely venue for seasonal BP variability, our conclusion is that is the best explanation.

Table I.   Monthly Mean Temperatures in 3 Southern Californian Population Centers
 Los AngelesAnaheimSan Diego
  1. Monthly temperature variability showing relatively steep dip between September and October, with warm rise occurring between March and April, in the context of reduced overall variability throughout the year.


Seasonal BP Variability in the Medical Research Council Trial

Large-scale seasonal variability in BP was noted in the Medical Research Council Mild Hypertension Treatment trial, which recruited more than 17,000 men and women with diastolic pressures of 90 mm Hg to 109 mm Hg.2 This trial, conducted in 190 centers throughout England, Scotland, and Wales, followed patients aged 35 to 64 during 5 years and found consistent seasonal variability for both men and women randomized to placebo, bendrofluazide, and propanolol. All regression coefficients were significant at P<.01 and P<.001 for mean deviations in both systolic and diastolic BPs, respectively, associated with seasonal variations of 20°C (36°F).2 Looking at age groups 35 to 44 years, 45 to 54 years, and 55 to 64 years, the largest BP variability was associated with systolic BP in elderly patients. Men aged 55 to 64 years had a seasonal systolic pressure difference of 6.1 mm Hg to 7.3 mm Hg, and women in this age range had a seasonal systolic pressure difference of 4.4 mm Hg to 6.9 mm Hg in the 3 randomized drug treatment groups.2

An analysis of the southern California KP uncontrolled hypertensive population was undertaken comparing a control nadir population in February with a control zenith population in September. In a standardized analysis template, characteristics including various age ranges, sex, patients seen by their primary care physician within 12 months, those without a primary care physician visit within 12 months, and various intensities of drug treatment are compared (Table II). Similar to the Medical Research Council analysis, older age was most closely associated with loss of hypertension control in the cooler season. A greater percentage of hypertension control was achieved for patients aged 18 to 50 years despite cooler weather, and there was very little seasonal variability associated with the 51- to 65-year age range. The overall increase in hypertension uncontrol was associated with the 66- to 80-year age group, with a tendency for seasonal uncontrol also associated with patients older than 80 (Table II).

Table II.   Analysis of Southern California Kaiser Permanente Uncontrolled Hypertension Population Comparing Zenith and Nadir Seasonal Hypertension Control Months
  1. Abbreviations: HTN, hypertension; PCP, primary care physician; Rx, number of antihypertensive drugs prescribed. Permanente Online Interactive Network of Tools (POINT) database analysis showing that the increase in Nadir month seasonal hypertension control occurs in the age 66- to 80-year and age >80-y strata.

Population HTN ControlZenithNadir
Months comparedSeptember 2010February 2011
Hypertension control rate all HTN patients aged ≥18 y81.1%79.1%
Age 18–50 y23.42% (21,785)22.03% (23,832)
Age 51–65 y38.65% (35,955)38.30% (41,431)
Age 66–80 y27.87% (25,929)29.05% (31,429)
Age >80 y10.07% (9369)10.62% (11,483)
Male45.53% (42,364)46.12% (49,895)
Female54.46% (50,669)53.87% (58,275)
Seen PCP in past 12 months81.99% (76,278)83.32% (90,129)
Not seen PCP in past 12 months18.01% (16,760)16.68% (18,046)
0 Rx10.81% (10,061)10.64% (11,513)
1 or 2 Rx55.09% (51,254)55.62% (60,163)
3 or 4 Rx28.95% (26,937)28.83% (31,187)
>4 Rx5.14% (4,786)4.91% (5,312)

Confirmation of Seasonal BP Variability in Temperate and Subtropical Climates

Several studies have confirmed the seasonal variability of BP in temperate and subtropical zones. In the Three-City Study in France, an 8-mm Hg variance in systolic pressure was seen in 8801 patients aged 65 and older with a mean systolic pressure of 150.1 mm Hg in winter and 142.1 mm Hg in summer between the lowest (<7.9°C≤17.8°F) and highest (≥21.2°C=70°F) quintiles of outdoor temperature.3 There was a greater difference in BP for patients 80 years and older.3 A similar range of seasonal systolic BP variation between 5 mm Hg and 7 mm Hg has been seen in the temperate climates of Japan, Great Britain, Minnesota, Austria, Italy, and Pennsylvania.4–6

In Japan, hypertensive patients who live mostly indoors with air conditioning, and with a summer/winter external temperature difference of only 12°F, still had winter season BPs that averaged 7.5/4.1 mm Hg higher than in the summer.4 Interestingly, that difference was comparable to a seasonal BP disparity of 8/5 mm Hg observed in 801 men in Finland experiencing external temperatures of −30°C (22°F) in winter and 30°C (86°F) in summer.7

Seasonal BP disparity has been described in the subtropical zone countries of India, Israel, and Iraq.8–10 For women aged 18 to 40 living in an urban slum of Delhi, India, with summer temperatures of 45°C (113°F) in summer and −4°C (25°F) in winter, BPs were 11.07/6.79 mm Hg higher.8 In a carefully performed study of 182 Israeli hypertensive patients aged 65 to 91 years residing for the most part in comfortable year-round air-conditioned ambient temperatures of 23°C to 24°C (73°F to 75°), and external temperatures of 28.6°C (83°F) in summer and 14°C (57°F) in winter, there was a sizable difference in BP. Winter BPs were 165±11.6 mm Hg /90±13.7 mm Hg in winter compared with 134±47.3 mm Hg/74±8.5 mm Hg in summer.9

In Mosul, Iraq, where daytime summer temperatures are 40°C (104°F) and daylight lasts 14 to 15 hours, compared with winter temperatures of 10°C (50°F) and daylight duration of 9 to 9.5 hours, the mean summer BP for hypertensive patients was 130.6/85mm Hg compared with 147.6/94.1 mm Hg for the same patients during winter.10 That is a seasonal BP difference of 17/8.9 mm Hg. Therefore, the seasonal BP disparity in subtropical areas is at least as large as in temperate zones, even with air conditioning.

Seasonal BP Variability in Different Patient Groups

Thus far, described studies have involved patients with and without hypertension at various age ranges and, for the most part, the largest seasonal variability affected older hypertensive patients. Seasonal BP variation also occurs in normal pregnancy,11 in children,4 and in patients with chronic kidney disease.12 In a Japanese pregnancy study, the expected physiologic decrease in second-trimester BP occurred throughout the year, but third-trimester BPs were 12/3 mm Hg higher in January than July.11 Home BPs during pregnancy were 2.5/2.5 mm Hg higher for every 10°C (18°F) decrease in external temperature.11 A winter increase in preeclampsia has also been noted.11

In a study describing 109 patients with chronic kidney disease in Beijing, China, the stage of kidney disease was not influential and all stages had similar seasonal BP variability.12 A systolic BP difference of 12 mm Hg was inversely correlated with an average monthly BP differential between 0°C (32°F) and 27°C (81°F)12 (Figure 2). Seasonal BP variability also occurs in hemodialysis and renal transplant patients.10,13

Figure 2.

 Variation of systolic blood pressure (SBP) and monthly temperature in chronic kidney patients. Reprinted with permission.12

Seasonal Variability Across Different Measurement Techniques

Most seasonal BP studies have used traditional office visit information, but studies using home BPs11,14 and 24-hour ambulatory BP monitoring (ABPM) have also been confirmatory.15 In the well-matched cross-sectional Pressione Arteriose Monitorate E Loro Associoni (PAMELA) study, 24-hour ABPMs in 2051 randomly selected individuals from the city of Monza, northeast of Milan, Italy, showed typical variability between the coldest and warmest months.15 Compared with clinic and home BPs in the PAMELA population, 24-hour ABPM seasonal variability was present, but diminished, as would be expected, due to the effect of many more BPs obtained that would reduce measurable deviation from the norm.15 Seasonal ABPM variability was greatest for undiagnosed hypertensive patients, intermediate for treated hypertensive patients, and least for normotensive patients.15

Possible Etiologies of Seasonal BP Variability

Much of the etiologic discussion of winter season increase in BP has centered around the cold pressor response and increased sympathetic nervous system activity. Cold pressor responses are age-related: younger patients exhibit more increase in peripheral vascular resistance, and older patients exhibit more increase in cardiac output.4 Older patients and thinner individuals probably have reduced capacity to respond to external cold.2,4 Support for increased sympathetic nervous system activity is the association of elevated plasma and urinary norepinephrine and epinephrine levels during the winter season.8,9,16 However, in the Medical Research Council trial there was as much of a winter increase in BP in the group of patients randomized to propanolol up to 120 mg twice a day as there was in the placebo group.2 Additionally, the 24-hour ABPM PAMELA study did not show an increase in heart rate during the cold weather season.15

One study that showed a winter season increase in catecholamines associated with increased BP also showed reduced urine sodium and urine volume during the winter, although summer sweating may explain the difference.16 Authors of a chronic kidney disease study showing no difference in seasonal BP variability related to the stage of kidney disease suggested that these observations constituted some evidence that volume status was not a key physiologic mechanism to explain this phenomenon.12

Youn and colleagues13 demonstrated that pulse wave velocity, a widely accepted surrogate measure of arterial stiffness, was correlated with winter-summer differences in systolic BP. Fibrinogen levels may increase to 23% in winter, and skin cooling can increase blood viscosity by increasing red blood cell and platelet counts.13 Increased physical activity during the warmer months may be associated with reduced BP, and pedometer counts are higher in the summer.17 Reduced sunlight and vitamin D exposure during the winter season may activate the renin-angiotensin-aldosterone system (RAAS) and cause increased BP,18 but there are also counterregulatory adaptations that mitigate these effects.19 Winter season aldosterone levels are also elevated.19 The vasodilator effect of higher temperatures probably also plays a role.20

Association With the Seasonal Incidence of Adverse Cardiovascular Events

Cardiovascular mortality, nonfatal myocardial infarction, nonfatal stroke, aortic dissection, and rupture of the abdominal aorta have been shown to fluctuate seasonally with winter season increases,3,21 although extreme weather data do show increased stroke related to extremely hot weather.22 In addition to the seasonal fluctuation in RAAS and sympathetic nervous system activation, elevated cooler season BP may play a role particularly in higher-risk cardiovascular risk populations. Given BP variability related to factors such as diurnal rhythm and body weight, seasonal fluctuations particularly in older hypertensive individuals may be another important adverse cardiovascular risk factor.7,20

Case Summary

The case presentation describes the unusual occurrence of a patient residing in California who was normotensive in the 6 warmer months of the year, but required 2-drug therapy to control her BP during the 6 cooler months. She fit the description of the older leaner patient phenotype who might be more susceptible to seasonal BP variability. In one study of older hypertensive patients, 38% had their antihypertensive medications increased during the wintertime and 18% had their medications decreased during the summer season.9 The large population Southern California KP Hypertension Registry has consistently shown a 2% winter season decrease in monthly hypertension control rates from October through March. This small but significant population-wide effect can be more marked for individual patients.