Utility of home blood pressure monitoring to evaluate postprandial blood pressure in treated hypertensive patients

Abstract

Postprandial hypotension, defined as a fall in systolic blood pressure (SBP) of 20 mmHg or greater within 2  hours after a meal, is a risk factor for stroke, coronary events and mortality. The clinical suspicion is typically raised by episodes of postprandial syncope or falls, whereas asymptomatic postprandial hypotension is mostly neglected. The magnitude of the postprandial fall in SBP, as detected by 24-hour recording in apparently healthy middle-aged to elderly subjects, was proportional to the severity of the silent cerebrovascular damage. Postprandial hypotension can also be detected by self-measured blood pressure before and within 2 hours after meals using automatic devices. The review highlights the value of home blood pressure monitoring (HBPM) as a screening test for asymptomatic postprandial hypotension in hypertensive patients. Using a HBPM protocol that included duplicated blood pressure measurements before and after three consecutive lunches, we detected unsuspected postprandial hypotension in 27.4% of the 230 hypertensive patients screened. The prevalence of postprandial hypotension was 13.2% in controlled and 42.2% in uncontrolled hypertensive patients (p < 0.001), raising the dilemma of further lowering blood pressure in the setting of postprandial hypotension. The inclusion of preprandial and postprandial measurements in the protocol of HBPM is useful to identify hypertensive patients with postprandial hypotension and may guide adjustments in antihypertensive treatment according to postprandial blood pressure.

Clinical implications of postprandial hypotension

Postprandial hypotension, defined as a decrease in systolic blood pressure (SBP) of 20 mmHg or greater within 2  hours after a meal, is an under-recognized risk factor for syncope, falls, dizziness, weakness, angina, stroke and mortality in older nursing home residents [Aronow and Ahn, 1997; Fisher et al. 2005]. Even smaller declines in SBP after meals (10–20 mmHg) as detected by ambulatory blood pressure monitoring (ABPM), are associated with greater number of lacunar infarction in middle-aged to elderly persons from the community [Tabara et al. 2014].

The prevalence of postprandial hypotension is widely variable among studies, depending on demographic factors (age, blood pressure, Parkinson disease, dysautonomia, drug consumption, symptoms, community versus nursing home residency), meal composition, site of evaluation (clinic versus home), and time interval of blood pressure measurements [Oberman et al. 2000; Van Orshoven et al. 2010; Luciano et al. 2010].

Doctors should take into account the recent intake of food and be aware of postprandial hypotension during the evaluation and treatment of hypertensive patients. Postprandial hypotension may be a source of visit-to-visit blood pressure variability and masked hypertension when the clinical appointment takes place during the subsequent hours after a meal [Fagan et al. 1986]. Achieving the blood pressure goal raises a therapeutic dilemma when postprandial hypotension is detected in a patient with poorly controlled hypertension. Otherwise, the attempt to control hypertension could inadvertently exaggerate the postprandial fall in blood pressure, conditioning a J-shaped curve.

Evaluation of postprandial blood pressure

As opposed to orthostatic hypotension, which can be assessed during the routine clinical evaluation, screening for postprandial hypotension is challenging. The hemodynamic, autonomic, and neurohumoral effect of standardized meals has been classically evaluated in the context of controlled laboratory conditions [Oberman et al. 2000; Fagan et al. 1986].

During the last two decades, ABPM has been used for the screening and prognostic implications of postprandial hypotension outside the office. ABPM can detect postprandial hypotension and evaluate its frequency, severity, duration, and relationship to drug intake and symptoms [O’Brien et al. 2013]. Several studies have used ABPM to investigate the prevalence and clinical implicance of postprandial hypotension [Kohara et al. 1998, 1999; Puisieux et al. 2000; Salvioli et al. 1996; Grodzicki et al. 1998; Vloet et al. 2005; Zanasi et al. 2012]. The prevalence of postprandial hypotension, as detected by ABPM, is greater in symptomatic elderly patients than in controls [Puisieux et al. 2000]. Nevertheless, the magnitude of asymptomatic postmeal reductions in ambulatory SBP increased in parallel with the number of lacunar infarctions without a clear cutoff [Tabara et al. 2014]. Patients should be instructed to record the timing of meals and related symptoms in the diary. Patients with postprandial hypotension show high daytime SBP variability at the expense of periods of normal or low blood pressure following meals preceded by periods of high SBP [Kohara et al. 1998; Puisieux et al. 2000]. Care should be taken to not misinterpret the siesta dipping as lunch-induced hypotension.

Alternatively, home blood pressure monitoring (HBPM) including premeal and postmeal measurements may be used to evaluate the effect of food on blood pressure [White et al. 2008]. We assessed postprandial hypotension in 255 unselected hypertensive patients using self-measured blood pressure at home (OMRON 705) [Barochiner et al. 2014]. Patients were instructed to take duplicated readings in the morning (before breakfast and drug intake) and evening during four consecutive days, and duplicated readings before and after lunch during the last 3 days of monitoring. Hypertension control was defined according to the mean of morning and evening blood pressure of the last 3 days of monitoring (the first day was discarded), using 135/85 mmHg as the cutoff value [Pickering et al. 2008]. Postprandial hypotension was defined as a decrease in SBP of 20 mmHg or greater than the corresponding premeal value in at least one lunch. A total of 25 patients (9.8%) had less than 2 days of preprandial and postprandial readings, therefore 230 patients remained for further analyses (91.3% under antihypertensive treatment; median age 73.6 years). A total of 63 patients (27.4%) had at least one and 20 patients (8.7%) had two or more episodes of postprandial hypotension. Patients with postprandial hypotension were older (78 versus 71 years; p < 0.001). There was no significant difference in gender, prevalence of diabetes or previous cardiovascular disease between patients with and without postprandial hypotension, with the exception of previous stroke, which was more frequent among subjects with postprandial hypotension (14% versus 4%; p = 0.003).

Patients with postprandial hypotension also had higher office (142 versus 134 mmHg; p < 0.001), morning (145 versus 134 mmHg; p < 0.001), prelunch (139 versus 127 mmHg; p < 0.001) and evening SBP (140 versus 133 mmHg; p = 0.01) than patients without postprandial hypotension. In contrast, postlunch SBP was comparable in both groups (124 versus 125 mmHg, respectively). The pattern of higher morning and prelunch SBP observed in patients with postprandial hypotension was reproducible across the 3 days of monitoring and it was independent of age (Figure 1).

Figure 1.

Bar graph comparing morning, prelunch, postlunch and evening systolic blood pressure (SBP) between patients with and without postprandial hypotension (PH) stratified by day of monitoring and age strata. The first day was discarded. Values are mean ± standard deviation (SD).

Further analyses were done in the subgroup of 140 patients receiving antihypertensive treatment and had blood pressure measurements during the three lunches [Alfie et al. 2015]. To evaluate the influence of age on the prevalence of postprandial hypotension, patients were divided according to the mean age of the sample in <74 years (mean age 64 ± 8 years, n = 69) and 74 years and older (mean age 81 ± 4 years, n = 71). The prevalence of postprandial hypotension was 2 times greater in older compared with younger patients (17.3% versus 35.4%, χ² < 0.02) and 3.2 times greater in uncontrolled compared to controlled hypertensive patients (13.2% versus 42.2%, χ² < 0.001). In contrast, postprandial hypotension was unrelated to the number of antihypertensive drugs (Figure 2).

Figure 2.

Bar graph comparing the proportion of episodes of postprandial hypotension (PH) according to age, hypertension control at home, and number of antihypertensive medications. *p < 0.05; † p < 0.001 by χ². The significance is for all of the proportions of episodes.

In patients with controlled hypertension, SBP fell from 118 ± 8 mmHg before lunch to 115 ± 9 mmHg after lunch (p < 0.001), whereas it fell from 140 ± 14 to 130 ± 13 mmHg (p < 0.001) in patients with uncontrolled hypertension. The postprandial fall in SBP was significantly greater in the latter group (–9.5 ± 10.5 versus –3.2 ± 7.8 mmHg, respectively, p <0.001). The difference remained significant when the postprandial change was expressed as a percentage from prelunch SBP (–6.4 ± 7.8% versus −2.6 ± 6.5%, respectively, p = 0.002). Two-factor analysis of variance (ANOVA) showed no significant interaction between hypertension control and age. This is consistent with previous data in rigorously screened healthy subjects showing that, in the absence of blood pressure elevation, the hemodynamic, autonomic, and neurohumoral responses to meal ingestion remained unchanged in elderly adults [Oberman et al. 2000].

The lunch-induced change in SBP was significantly correlated with the level of morning (r = −0.33, p < 0.001), evening (r = −0.25, p < 0.01) and particularly prelunch SBP (r = −0.52, p < 0.001). The strength of the correlation between prelunch SBP and its fall after lunch is consistent with previously published data using either conventional or ambulatory blood pressure measurements [Kohara et al. 1998; Puisieux et al. 2000; Vaitkevicius et al. 1991].

The agreement between the change in SBP after lunches 1, 2 and 3 is shown in Figure 3. The mean difference in the change in SBP was 0.66 mmHg between lunches 1 and 2, and −2.04 mmHg between lunches 1 and 3. Differences in meal composition and time interval of blood pressure measurements after lunch could contribute to the relatively wide limits of agreement [Luciano et al. 2010; Teunissen-Beekman et al. 2013]. This limitation could be compensated by increasing the number of monitored days.

Figure 3.

Bland–Altman plots showing the agreement between postprandial changes in SBP after lunches 1 and 2 (A) and after lunches 1 and 3 (B). Differences are plotted against mean postprandial changes. Mean differences are shown (solid lines) plus confidence intervals for 95% limits of agreement (dotted lines).

We used lunch rather than breakfast or dinner because it is a more consistent meal for our community. The lack of blood pressure measurements after breakfast and dinner represents a limitation of the study [Kohara et al. 1998]. Given the inverse correlation between preprandial and postprandial SBP, and that the highest SBP was recorded before breakfast, it is possible that episodes of asymptomatic postbreakfast hypotension may have passed undetected by our protocol of HBPM. Finally, 20% of our patients did not obtained blood pressure measurements during each lunch, indicating that some patients may require a more careful instruction on the schedule of measurements.

Therapeutic challenge

The routine use of HBPM with postprandial measurements unmasked the presence of postprandial hypotension in 42% of patients with uncontrolled hypertension. One might speculate that the attempt to control hypertension could inadvertently exaggerate the postprandial fall in blood pressure. This raises the dilemma of intensifying treatment to further lowering morning and evening blood pressure or down titrate treatment to attenuate postprandial hypotension. It is expected that uncontrolled hypertension would predispose to further postprandial hypotension, whereas tight control of hypertension would prevent or attenuate it. Repeated assessment of postprandial and upright blood pressure may help to avoid hypotension while antihypertensive treatment is adjusted to achieve target blood pressure.

Although we did not evaluate the timing of drug intake, a change in the timing of antihypertensive drug dosing from morning to bedtime is recommended. We observed a lack of association between the number of antihypertensive drugs and postprandial hypotension. Accordingly, a study of octogenarians and nonagenarians living in residential care facilities showed that antihypertensive therapy was not associated with an increased risk for orthostatic hypotension, postprandial hypotension, or falls [Fisher et al. 2003]. However, a prospective study in 4964 community-living status older than 70 years showed a significant association between intensity of antihypertensive medications and recurrent fall injury [Tinetti et al. 2014]. Patients with symptomatic or severe postprandial hypotension would benefit from a less stringent blood pressure target. Patients with postprandial hypotension should be advised to avoid high-carbohydrate meals [Teunissen-Beekman et al. 2013]. Pharmacological treatment with acarbose (100 mg), an alpha-glucosidase inhibitor that decreases glucose absorption in the small intestine, improves postprandial hypotension [Shibao et al. 2007]. Subclinical volume depletion due to diuretic agents or strict dietary sodium restriction should be considered as a contributing factor.

In conclusion, we suggest that elderly hypertensive patients should be routinely screened for asymptomatic postprandial hypotension. The addition of self-measurements before and after meals makes HBPM a suitable screening test for postprandial hypotension and offers a number of potential advantages over ABPM: first, it is better tolerated, easier to perform and cheaper; second, it evaluates blood pressure after meals over different days; third, patients must remain awake for self-measuring blood pressure, discarding the confounding effect of siesta; and most importantly, HBPM is more suitable for guiding treatment adjustments according to postprandial blood pressure response.