Why Is First Blood Pressure Reading Higher Than Second

Abstract

We evaluated the consequences of excluding the first of three blood pressure (BP) readings in different settings: a random population sample (POS, n=1525), a general practice office (GPO, due north=942) and a specialized hypertension center (SHC, north=462). Differences between systolic and diastolic BP (SBP and DBP) estimates obtained including and excluding the first reading were compared and their correlation with convalescent BP monitoring (ABPM) was estimated. The samples were divided into quartiles according to the difference betwixt the 3rd and the starting time SBP (3-1ΔSBP). SBP decreased through sequential readings, 3-1ΔSBP was −v.5±9.7 mm Hg (P<0.001), −5.1±10.4 mm Hg (P<0.001) and −vi.1±9.3 mm Hg (P<0.001) for POS, GPO and SHC, respectively. Yet, individuals included in the top quartile of 3-1ΔSBP showed their highest values on the third reading. The mean SBP judge was significantly college excluding the first reading (P<0.001), but the differences among both approaches were minor (ane.five–1.six mm Hg). Moreover, the correlation between SBP values including and excluding the first reading and daytime ABPM were comparable (r=0.69 and 0.68, respectively). Like results were observed for DBP. In conclusion, our study does non support the notion of discarding the first BP measurement and suggests that it should be measured repeatedly, regardless the starting time value.

Introduction

Accurate office claret pressure (BP) measurement remains the mainstay of the diagnosis and treatment of hypertension, and the BP judge in the first visit is the cornerstone of the initial management of hypertension as these values will exist used to decide if subsequent visits are necessary. The recognition of a reduction in the white-coat effect through successive measurements of BP^ane has led to the suggestion that discarding the commencement function reading of a set of three may improve the possibility of knowing the actual BP. If BP always decreases in the subsequent measurements, additional BP readings are not necessary when the first measurement is <140/ninety mm Hg. Nevertheless, the supposition that the starting time BP reading is systematically higher in all subjects was not widely studied. Two guidelines aimed at improving hypertension diagnosis and treatment in the clinical practice were recently published, the '2013 European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC) guidelines for the management of arterial hypertension'² and the '2013 Canadian Hypertension Education Program (CHEP) Recommendations for Claret Pressure Measurement, Diagnosis, Assessment of Risk, Prevention, and Treatment of Hypertension'.³ They agree on a value of ⩾140/90 mm Hg as hypertension threshold, but there is disagreement regarding whether the commencement reading should be discarded or not. The ESH/ESC guidelines recommend taking 'at least ii BP measurements, in the sitting position, spaced 1–two min apart, and additional measurements if the first two are quite dissimilar'. On the other manus, the Canadian guidelines suggested that 'At the initial visit for the assessment of hypertension, if systolic BP is ⩾140 and/or diastolic BP is ⩾xc mm Hg, more than 2 additional readings should be taken during the same visit' and 'The offset reading should be discarded and the latter ii readings averaged'.

Disagreement about whether to discard the first BP measurement or not was also observed in population studies; for example, ii recently published studies aimed at evaluating the variability of BP in the general population—both based on National Health and Nutrition Examination Survey (NHANES) data—estimated the BP using different criteria: the kickoff used the average of the second and tertiary measurement⁴ whereas the other used the average of the three readings.⁵

The aim of our written report was to evaluate the effect of discarding the commencement BP measurement. For this purpose the BP estimates including and excluding the first reading were compared in three unlike settings: (a) a random population sample, (b) a screening for high BP in general do offices (GPOs) and (c) a specialized evaluation in a hypertension center. In the last setting nosotros as well evaluated whether discarding the beginning BP reading improves the relationship betwixt office BP and ambulatory BP monitoring (ABPM) every bit was previously suggested.

Materials and methods

The study was performed using data from individuals who had had at least three consecutive BP readings on i occasion from 3 independent samples, (i) a random population sample, (two) consecutive patients in a GPO and (3) consecutive patients in a specialized hypertension center (SHC).

The methodology used in the random population sample to obtain measurements has already been published.^6,vii In brief, three BP measurements spaced effectually ii min from i some other were performed at home past trained nurses, after a minimum resting period of 5 min, using a mercury sphygmomanometer. Phase I and Five Korotkoff sounds were used to identify systolic BP (SBP) and diastolic BP (DBP) respectively.

In the GPOs, doctors using the OMRON HEM 705 CP devices (OMRON HEALTHCARE Co., Kyoto, Japan) took iii BP measurements separated by a minute in a unmarried visit from every patient examined regardless of the purpose of the visit.

In the SHC specially trained nurses performed iii BP measurements merely before the realization of ABPM with the aforementioned device and methodology used in the GPO setting. The ABPM registries were performed with the SpaceLabs 90207 monitor (Spacelabs Healthcare, Issaquah, WA, USA) programmed to read BP at intervals of 20 min during the 24-hour interval and 30 min during the night. Day and night was differentiated past taking into account the patient'due south diary.

To analyze the BP behavior through successive readings and evaluate if the first reading was systematically highest, mean and quartiles of the difference between the 3rd and first SBP (3-1ΔSBP) and DBP (3-1ΔDBP) readings were estimated. In order to decide the effects of antihypertensive drugs, individuals with and without treatment were analyzed separately. Age and trunk mass index were compared amid quartiles of three-1ΔSBP using analysis of variance with Tukey postal service hoc assay, and sex and current smoking using χ ²-test.

To estimate the effect of discarding the first BP reading, mean BP was estimated in each subject using two different approaches: (1) discarding the first measurement and (two) averaging the three readings.

Differences betwixt SBP and DBP values obtained using both approaches were calculated and compared using one-sample t-examination. Also, for each sample, Bland–Altman plots were constructed with the deviation of the ii approaches on the vertical centrality and the average of the ii approaches on the horizontal axis. Horizontal reference lines on the scatterplot showed the difference betwixt the measurements ±2s.d. In society to investigate the concordance of high BP diagnosis (SBP⩾140 mm Hg and/or DBP⩾90 mm Hg) among both BP estimates, we calculated the kappa coefficient (κ) after dichotomizing the office measurements equally 'high' or 'low' BP. The prevalence of high BP both including and excluding the first BP reading were compared using McNemar's test.

Finally, in the SHC sample, the relationship betwixt the office BP values obtained and the hateful daytime ABPM were evaluated using Pearson correlation coefficient (r); additionally, 95% confidence intervals (95%CI) for r values were calculated based on the Fisher r-to-z transformation.

Statistical analyses were performed using SPSS 18.0 software (SPSS, Chicago, IL, USA); 2-tailed P-values <0.05 were considered statistically significant.

Results

The study included 2926 individuals (1780 women, 46±17 years quondam and 1149 men, 47±17 years onetime, P between sex=0.566). The characteristics of the iii samples are described in Tabular array 1.

Tabular array 1 Characteristics of the three samples, random population, general do part and specialized hypertension office

Total size table

As Tables two and 3 prove, hateful BP decreased across the sequential readings; the mean subtract was similar in all settings; Δ3-1SBP was −5.v±9.7 mm Hg (P<0.001), −5.ane±ten.iv mm Hg (P<0.001) and −6.1±ix.3 mm Hg (P<0.001) in the random population sample, GPO and SHC, respectively; Δ3-1DBP was −iii.0±9.0 mm Hg (P<0.001), −ane.71±7.eight mm Hg (P<0.001), and −ii.one±v.7 mm Hg (P<0.001) in the random population sample, GPO and SHC, respectively. However, when the samples were divided into quartiles of the Δ3-1SBP, the individuals included in the top quartile had higher SBP values in the third reading compared with the first one (Effigy 1a); similar behavior was observed with DBP when it was divided into quartiles of Δ3-1DBP (Figure 1b). Individuals in the peak quartile of Δ3-1SBP were younger (46±18 vs 49±16 years erstwhile, P=0.007) and thinner than individuals in the get-go quartile of (body mass index 26.7±5.4 vs 28.ane±6.vi, P<0.001). The percentage of electric current smokers was higher in the tiptop quartile of Δ3-1DBP than in the remaining quartiles (26.4 vs 22.six, P=0.034). The pct of women was similar among quartiles of Δ3-1SBP (61.2, 62.v, 61.vi and 58.0, P=0.284).

Table 2 Outset, second and third SBP readings and SBP estimates using mean of three BP readings and discarding the first measurement

Total size table

Tabular array 3 First, 2d and 3rd diastolic blood pressure readings and diastolic blood pressure estimates using mean of 3 BP readings and discarding the first measurement

Total size table

Figure 1

Mean differences between the third and the first SBP (a) and DBP (b) readings according to quartiles of these differences in the three samples analyzed: random population sample (POS), general practice office (GPO) and specialized hypertension office (SHO).

Full size image

Tables ii and 3 also compare the departure between the hateful of all readings and the hateful excluding the first reading; SBP and DBP were significantly lower in all settings when the get-go reading was discarded. Although the mean differences for SBP including and excluding the outset BP reading seem only modest, 1.v (3.0), 1.six (2.8) and 1.6 (2.8) mm Hg for GPO, specialized hypertension and office random population sample, respectively (Table 2), the range was wide, and included positive and negative values (from −13.iii to 27.3 mm Hg) (Figure 2, Bland–Altman plots). Therefore, excluding the starting time reading did not yield lower values systematically. Like results were obtained for DBP (Table 3); the mean differences between BP estimates were 0.five (2.3), 0.6 (i.7) and 0.8 (2.vi) mm Hg for GPO, specialized hypertension and office random population sample, respectively.

Effigy 2

Bland–Altman plots comparing the SBP estimate using the mean of three readings and discarding the kickoff reading in the three samples analyzed: random population sample (POS) (a), general do role (GPO) (b) and specialized hypertension role (SHO) (c).

Full size image

The proportion of individuals with 'high' BP (⩾140/90 mm Hg) was lower when the first reading was excluded in all settings (42.three% vs 44.half-dozen% in the random population sample, P<0.001, 42.3% vs 44.3% in the GPO, P=0.005 and 45.v% vs 48.7% in the SHC, P=0.001). All the same, the concordance between approaches was loftier (κ=0.90, 0.91 and 0.91 for the random population sample, the GPO and the SHC, respectively); indeed, 2794 of 2929 subjects were classified concordantly; amidst the 135 subjects classified discordantly, 75.6% were considered as 'high' BP only when the mean of all three readings was used and 24.4% when the kickoff BP reading was excluded.

In the SHC sample, 446 ABPMs were considered valid. The correlations between daytime systolic ABPM and both approaches to estimate office SBP were modest but like, r=0.69 (95% CI 0.64–0.73, P<0.001) and r=0.68 (95% CI 0.61–0.73, P<0.001), for the hateful of 3 readings and excluding the first measurement, respectively (Figure 3). The correlations between daytime diastolic ABPM and office DBP estimates were also similar, r=0.69 (95% CI 0.64–0.73, P<0.001) and r=0.68 (95% CI 0.63–0.73, P<0.001), respectively.

Figure iii

Pearson's correlation betwixt systolic daytime ambulatory claret pressure level monitoring (ABPM) and the function systolic blood pressure (OBP) estimate obtained using (a) the mean of three readings and (b) discarding the first reading.

Total size paradigm

Discussion

Accurate office BP measurement, despite its shortcomings, remains the mainstay of diagnosis and handling of hypertension. Nonetheless, the simple question of whether the kickoff office BP measurement should exist discarded remains to be answered and there is no agreement amongst the varying hypertension guidelines. Thus, the Canadian³ and Overnice^eight guidelines recommend discarding the outset BP measurement, whereas the ESH/ESC one does non.² Our data evidence that, in all settings analyzed, the effect of excluding the first BP reading was pregnant but modest, ~1.5 and <1 mm Hg for SBP and DBP, respectively. Moreover, when BP was analyzed as a dichotomic variable in order to classify subjects as 'high' or 'low' BP, the vast majority were classified concordantly.

It has been suggested that excluding the first BP measurement could improve the correlation between office BP and out of the part BP. However, our data do not support this concept; using the second and third readings but (that is, discarding the start) did non improve the correlation with daytime ABPM when compared with the employ of hateful of all three readings; indeed, r values were well-nigh identical whether the first BP measurement was included or non. This argument is concordant with a previously published study about untreated hypertensive patients⁹ but expands the decision to non-hypertensive subjects and to subjects using antihypertensive drugs.

The subtract of BP in successive measurements has been shown repeatedly and with unlike methods,¹⁰ and it is widely internalized in medical thinking that successive measurements will give lower BP values. Thus, Parati et al. ⁴ suggested that discarding the commencement reading may improve the accuracy of the diagnosis of hypertension. Furthermore, in order to eliminate the increase of BP associated with the effect of alarm, ESH guidelines for home BP measurement recommend discarding the measurements performed on the first day.¹¹ Yet, the scientific evidence to support these assumptions is minimal and valuable information nearly BP variability could be lost.

Perhaps the virtually interesting finding in our written report was the fact that ~25% of the individuals did non decrease or increase their BP trough successive measurements and, consequently, in these individuals the first reading was not the highest. Remarkably, this behavior was found in all 3 settings (population sample, general practice and specialized hypertension office) and information technology was as well independent of whether BP was measured past medical doctors or nurses, in the function or at domicile. Despite this miracle being previously published,¹² it is non widely recognized and it has several implications for clinical practice.

First, the consequences of excluding the initial reading are unpredictable for an individual patient. Equally Bland–Altman plots show, the differences among BP estimate excluding or including the beginning measurement were broad, and yield both positive and negative values. Indeed, although the individuals who increased their BP in the tertiary reading were younger, thinner and more electric current smokers than those who decreased BP trough successive measurements (offset quartile of 3-1ΔSBP), at that place are no practical ways to identify them. Interestingly, the reclassification of individuals in categories of BP using the mean of the first and 2d, first through tertiary, and second and third readings was recently published.¹³

Another related consequence is whether or not a doctor should take more readings only when the showtime is ⩾140/90 mm Hg as the NICE guidelines and Canadian recommendations suggest; our findings do not support this arroyo. As the cardiovascular take a chance has a continuous relationship with BP starting with values as low every bit 115/75 mm Hg,^xiv taking but 1 BP measurement could significantly underestimate the subject'south 'true' risk of cardiovascular event.

Finally, the difference betwixt the first and third reading could give some important information about the BP variability. There is now some evidence that several such representations of BP variability, if augmented, increment cardiovascular take a chance independent of the average of BP readings conventionally acquired.^xv Visit-to-visit SBP variation was linked to increased cardiovascular and stroke bloodshed risk.¹⁶ Although the hazard associated with the short-term variability has been studied less, the BP variation in one visit results from changes in middle charge per unit, stroke volume and peripheral resistance in response to external and internal stimuli and their importance in terms of the take a chance of developing hypertension or cardiovascular disease remain to exist defined.

Some limitations of our study accept to exist pointed out. First, this was a post hoc analysis. Second, this study was performed in different settings with unlike methodologies to measure BP. However, our findings were concordant through all the samples, suggesting that the BP patterns observed were strongly consistent. Finally, our study was not able to define whether 1 arroyo would be more than appropriate than another; just long-term prospective studies with hard finish points could provide a definitive answer.

In conclusion, our study does not support discarding the outset BP measurement every bit the Canadian and NICE guidelines take suggested. Remarkably, although the mean BP decreased in successive BP readings, the beliefs of individuals was unpredictable and an appreciable proportion of subjects had the highest BP value in the last reading. We propose that 3 BP readings should exist taken in all individuals independently of the first reading value.

References

1. Parati 1000, Mancia G . Assessing the white coat event: which blood pressure measurement should be considered? J Hypertens 2006; 24: 29–31.

   CAS  Article  Google Scholar

2. Mancia G, Fagard R, Narkiewicz Thou, Redón J, Zanchetti A, Böhm K et al. 2013 ESH/ESC Guidelines for the management of arterial hypertension: the chore force for the management of arterial hypertension of the European Society of Hypertension (ESH) and of the European Guild of Cardiology (ESC). J Hypertens 2013; 31: 1281–1357.

   CAS  Article  Google Scholar

3. Hackam DG, Quinn RR, Ravani P, Rabi DM, Dasgupta K, Daskalopoulou SS et al. The 2013 Canadian Hypertension Educational activity Program recommendations for blood force per unit area measurement, diagnosis, assessment of run a risk, prevention and treatment of hypertension. Can J Cardiol 2013; 29: 528–542.

   Article  Google Scholar

4. Muntner P, Shimbo D, Tonelli Thousand, Reynolds One thousand, Arnett DK, Oparil Southward . The relationship betwixt visit-to-visit variability in systolic blood pressure and all-cause mortality in the general population: findings from NHANES Three, 1988–1994. Hypertension 2011; 57: 160–166.

   CAS  Article  Google Scholar

5. Muntner P, Levitan EB, Reynolds K, Mann DM, Tonelli M, Oparil Due south et al. Within-visit variability of blood pressure and all-cause and cardiovascular bloodshed among US adults. J Clin Hypertens (Greenwich) 2012; 14: 165–171.

   Article  Google Scholar

6. Salazar MR, Carbajal HA, Aizpurúa K, Riondet B, Rodrigo HF, Rechifort V et al. Decrease of blood force per unit area by customs-based strategies. Medicina (B Aires) 2005; 65: 507–512.

   Google Scholar

7. Salazar MR, Carbajal HA, Espeche WG, Aizpurúa M, Leiva Sisnieguez CE, March CE et al. Identifying cardiovascular disease risk and effect: use of the plasma triglyceride/loftier-density lipoprotein cholesterol concentration ratio versus metabolic syndrome criteria. J Intern Med 2013; 273: 595–601.

   CAS  Article  Google Scholar

8. National Clinical Guideline Heart (UK). Hypertension: The Clinical Management of Primary Hypertension in Adults: Update of Clinical Guidelines 18 and 34 [Internet]. Royal Higher of Physicians (Uk): London, 2011.

9. Graves JW, Grossardt BR . Discarding the first of three nurse-auscultatory or oscillometric blood pressure measurements does not better the association of office blood pressure with ABPM. Blood Press Monit 2010; fifteen: 146–151.

   Article  Google Scholar

10. Mengden T, Hernandez Medina RM, Beltran B, Alvarez E, Kraft Grand, Vetter H . Reliability of reporting self-measured blood force per unit area values by hypertensive patients. Am J Hypertens 1998; 11: 1413–1417.

    CAS  Article  Google Scholar

11. Parati Thou, Stergiou GS, Asmar R, Bilo Yard, de Leeuw P, Imai Y et al. ESH Working Grouping on Blood Pressure level Monitoring European Society of Hypertension practice guidelines for abode blood pressure monitoring. J Hum Hypertens 2010; 24: 779–785.

    CAS  Article  Google Scholar

12. Graves JW, Grossardt BR, Gullerud RE, Bailey KR, Feldstein J . The trained observer better predicts daytime ABPM diastolic blood pressure in hypertensive patients than does automatic (Omron) device. Blood Press Monit 2006; 11: 53–58.

    Article  Google Scholar

13. Handler J, Zhao Y, Egan BM . Impact of the number of blood pressure measurements on claret force per unit area classification in US adults: NHANES 1999–2008. J Clin Hypertens (Greenwich) 2012; xiv: 751–759.

    Article  Google Scholar

14. Lewington Due south, Clarke R, Qizilbash N, Peto R, Collins R . Prospective Studies Collaboration. Age-specific relevance of usual blood pressure to vascular mortality: a meta-assay of private information for i million adults in 61 prospective studies. Lancet 2002; 360: 1903–1913.

    Article  Google Scholar

15. Floras JS . Blood pressure variability: a novel and important hazard factor. Can J Cardiol 2013; 29: 557–563.

    Article  Google Scholar

16. Kikuya M, Ohkubo T, Metoki H, Asayama Chiliad, Hara A, Obara T et al. 24-hour interval-by-day variability of claret pressure level and heart rate at dwelling as a novel predictor of prognosis: the Ohasama Study. Hypertension 2008; 52: 1045–1050.

    CAS  Article  Google Scholar

Download references

Acknowledgements

This written report could not have been conducted without the help of the nurses from the 'Infirmary Municipal, San Andrés de Giles' and the 'Infirmary Municipal, Rauch' and the nurse Miriam Susana Cor from the 'Hospital Universitario San Martín, La Plata'. We appreciate the assistance of Sonja Zander in the revision of the English language of this manuscript.

Writer information

Affiliations

1. Clinica Medica, Infirmary San Martín, La Plata, Buenos Aires, Argentine republic

   M R Salazar, W G Espeche, C E 50 Sisnieguez, B C L Sisnieguez, C A Dulbecco, C Due east March, R N Stavile, E H Ferrari, M Correa, P M Maciel & E Balbín

2. Facultad de Ciencias Médicas, Universidad Nacional de La Plata, La Plata, Argentina

   M R Salazar, W K Espeche, C E L Sisnieguez, B C L Sisnieguez, C A Dulbecco, C E March, R N Stavile, East H Ferrari, Yard Correa & H A Carbajal

3. Hospital Municipal, Rauch, Buenos Aires, Argentina

   M Aizpurúa

Corresponding writer

Correspondence to Grand R Salazar.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Rights and permissions

About this article

Cite this article

Salazar, Thou., Espeche, Westward., Aizpurúa, Thou. et al. Should the first blood pressure reading be discarded?. J Hum Hypertens 29, 373–378 (2015). https://doi.org/10.1038/jhh.2014.98

Download citation

• Received: 05 May 2014

• Revised: 24 August 2014

• Accepted: 18 September 2014

• Published: 23 Oct 2014

• Issue Date: June 2015

• DOI : https://doi.org/10.1038/jhh.2014.98

wagnersquiter.blogspot.com

Source: https://www.nature.com/articles/jhh201498