HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Respond to This Article Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Schell, K. Articles by Simpson, K. Search for Related Content PubMed PubMed Citation Articles by Schell, K. Articles by Simpson, K.

Clinical Comparison of Automatic, Noninvasive Measurements of Blood Pressure in the Forearm and Upper Arm

	Abstract

Top Abstract Review of the Literature Methods Results Discussion Conclusions and Clinical... APPENDIX References

 
• Background When the upper arm (area from shoulder to elbow) is inaccessible and/or a standard-sized blood pressure cuff does not fit, some healthcare workers use the forearm to measure blood pressure.

• Objective To compare automatic noninvasive measurements of blood pressure in the upper arm and forearm.

• Methods A descriptive, correlational comparison study was conducted in the emergency department of a 1071-bed teaching hospital. Subjects were 204 English-speaking patients 6 to 91 years old in medically stable condition who had entered the department on foot or by wheelchair and who had no exclusions to using their left upper extremity. A Welch Allyn Vital Signs 420 series monitor was used to measure blood pressure in the left upper arm and forearm with the subject seated and the upper arm or forearm at heart level.

• Results Pearson r correlation coefficients between measurements in the upper arm and forearm were 0.88 for systolic blood pressure and 0.76 for diastolic blood pressure (P < .001 for both). Mean systolic pressures, but not mean diastolic pressures, in the upper arm and forearm differed significantly (t = 2.07, P = .04). A Bland-Altman analysis indicated that the distances between the mean values and the limits of agreement for the 2 sites ranged from 15 mm Hg (mean arterial pressure) to 18.4 mm Hg (systolic pressure).

• Conclusions Despite strict attention to correct cuff size and placement of the upper arm or forearm at heart level, measurements of blood pressure obtained noninvasively in the arm and forearm of seated patients in stable condition are not interchangeable.

The first method, the gold standard, is direct intra-arterial measurement. However, because of impracticalities and potential risks (eg, infection and trauma), indirect, noninvasive methods are commonly used. One indirect, noninvasive technique is use of a sphygmomanometer and auscultation. The blood pressure cuff is inflated by hand to a level that obliterates the arterial pressure or pulse. The cuff is then gradually deflated, and the systolic pressure is noted when Korotkoff sounds appear; the disappearance of the sounds indicates the diastolic pressure. An alternative indirect noninvasive method is use of an automated device and oscillometric measurements. Oscillations, movement in the arterial walls due to cardiac contractions, begin at the level of systolic blood pressure and reach their greatest amplitude at the level of the mean arterial pressure (MAP). The diastolic pressure is a derived value. The values are transmitted to a microprocessor, in conjunction with the cuff size, to determine the blood pressure.¹

Use of blood pressure cuffs that are too large yields readings lower than actual values, whereas use of cuffs that are too small yields readings higher than actual values.

 

The size and placement of the cuff are essential for accurate measurements. These features are specific to each manufacturer and are based on recommendations from the American Heart Association (AHA). Appropriate cuff size is key because cuffs that are too large yield readings lower than the actual values, and cuffs that are too small yield readings higher than the actual values.^1–3 Historically, the blood pressure cuff is placed on the upper arm (defined as the area from the shoulder to the elbow). However, the forearm (defined as the area from the elbow to the wrist) is increasingly becoming an alternative site because of the inaccessibility of the upper arm on some patients and the nonstandard, large-sized cuffs needed for the increasing numbers of morbidly obese patients.^2,4 Manufacturers of machines used in acute care to measure blood pressure noninvasively either do not recommend that the forearm be used or do not provide detailed, valid, and/or reliable directions for obtaining blood pressures in the forearm.^3,5 Cuffs designed specifically for measurements of blood pressure in the forearm are not available in most hospital settings. In the Protocol for Practice: Noninvasive Blood Pressure Monitoring⁶ published by the American Association of Critical-Care Nurses, Dobbin indicates that the forearm and the ankle can be used for measuring blood pressure but does not detail the procedure for cuff size or technique and cites only studies that validated the accuracy of blood pressures measured in the ankle. Additionally, the authors¹ of the AHA recommendations for noninvasive measurement of blood pressure mention use of the forearm for measuring blood pressures in obese patients but indicate that this method has not been validated.

Although the forearm is used for automatic noninvasive measurements of blood pressure, the accuracy of this procedure has not been verified.

 

	Review of the Literature

Top Abstract Review of the Literature Methods Results Discussion Conclusions and Clinical... APPENDIX References

 
An extensive electronic search of the database Cumulative Index to Nursing and Allied Health for research on the use of the forearm for measuring blood pressure noninvasively yielded 5 articles.^4,7–10 In an early study, Tachovsky⁷ compared indirect measurements of blood pressure obtained via auscultation in the upper arm with measurements obtained in the forearm. A convenience sample of 100 healthy female nursing students were placed in a supine position, and 3 blood pressure readings were obtained from each site (right upper arm vs right forearm), with a 3-minute rest between measurements. In addition to the 1967 AHA recommendations for obtaining blood pressure, Tachovsky used the Ravin procedure to augment auscultated sounds in the forearm but not the upper arm. A standard cuff (13 cm) with Velcro closure and a pediatric stethoscope were used to obtain measurements from both sites.

The results revealed significant differences between measurements of systolic blood pressure obtained in the 2 sites; measurements obtained in the upper arm were significantly greater than those obtained in the forearm (mean difference 7.35 mm Hg, P < .05). Conversely, measurements of diastolic pressure obtained in the forearm were significantly greater than those obtained in the upper arm (mean difference 14.1 mm Hg, P < .05). The clinical significance of these differences was less clear, and Tachovsky cautioned against using measurements of blood pressure obtained in the forearm in situations in which an accurate measurement was imperative for making clinical decisions. Although this study⁷ was one of the first on this topic, the current standard for measuring blood pressure often involves the use of automatic noninvasive instruments.

In 1996, Latman et al⁸ published their evaluation of the performance of an automatic, noninvasive instrument, the B/P Clinic (CardioAnalysis Systems, Winchester, Tenn), specifically designed to use the forearm for measuring blood pressure. The auscultatory method in the upper arm and the B/P Clinic in the forearm were used to measure blood pressure in a convenience sample of 106 subjects. In addition, heart rate was measured by palpating the radial artery and by using the B/P Clinic during blood pressure measurement. Patients with forearm circumferences less than 20.3 cm or greater than 33.0 cm and/or with systolic blood pressures of 190 mm Hg or greater were excluded. All subjects were seated during the data collection, with the upper arm and forearm positioned at the level of the heart for each measurement. In addition, a subset of the sample (n=30) had 2 sequential measurements in the forearm obtained with the B/P Clinic to determine the reliability of the instrument.

Correlation coefficients for measurements of blood pressure and heart rate obtained with the B/P Clinic and with the auscultatory method and palpation were 0.82 for systolic pressure, 0.75 for diastolic pressure, and 0.95 for heart rate. Reliability coefficients for the sequential measurements with the automatic device were 0.85 for systolic pressure, 0.81 for diastolic pressure, and 0.96 for heart rate. Latman et al⁸ concluded that the B/P Clinic was both accurate and precise and that the forearm could be used to measure blood pressure and heart rate.

This study⁸ provided important information on the reliability and validity of the B/P Clinic as an instrument for measuring blood pressure in the forearm. Sequential measurements have been used as the standard for comparisons of blood pressures measured in the upper arm and the forearm; however, Latman et al chose to compare the automated measurements with measurements obtained by using the auscultatory method. In reality, blood pressures in the upper arm are routinely measured by using an automatic noninvasive instrument.

In a similar study, Latman and Latman⁹ evaluated the accuracy and precision of an instrument used to measure blood pressure in the wrist that was developed for home blood pressure monitoring. The standard auscultatory method in the upper arm and the automatic noninvasive device in the wrist were used to measure blood pressure in a convenience sample of 150 subjects. In addition, heart rate was measured by palpating the radial artery and by using the noninvasive automatic device during blood pressure measurement. Patients with wrist circumferences less than 13.5 cm or greater than 19.5 cm and/or blood pressures of 280 mm Hg or greater and/or heart rates less than 40/min or greater than 200/min were excluded. All subjects were seated during the data collection, with the arm and wrist positioned at the level of the heart for each measurement. In addition, a subset of the sample (n = 51) had 2 sequential measurements in the wrist obtained with the automatic device to determine the reliability of the device.

Correlation coefficients for measurements of blood pressure and heart rate obtained with the automatic device and with the auscultatory method and palpation were 0.75 for systolic pressure, 0.76 for diastolic pressure, and 0.89 for heart rate. Reliability coefficients for the sequential measurements obtained in the wrist with the automatic device were 0.83 for systolic pressure, 0.92 for diastolic pressure, and 0.95 for heart rate. However, with the wrist instrument, hypotensive blood pressures tended to be overestimated and hypertensive pressures were significantly underestimated. The researchers concluded that the wrist instrument was easy to use and could have potential for home blood pressure monitoring. The major limitation of the earlier study⁸ also applies to this research.⁹

In a more recent study, Emerick¹⁰ evaluated the use of the wrist for determining blood pressure in a convenience sample of 85 nonhospitalized adults. Each subject was placed in the supine position, and upper arm and wrist circumferences were measured to determine proper cuff size. An oscillometric noninvasive device was used to obtain 3 sequential, paired measurements of blood pressure in the upper arm and the wrist of the right upper extremity, for a total of 255 paired readings. For each subject, measurements were obtained with the extremity at the subject’s side.

Mean differences between measurements obtained in the wrist and those obtained in the upper arm were 11.2 mm Hg (SD 7.56 mm Hg) for systolic blood pressure and 10.2 mm Hg (SD 5.60 mm Hg) for diastolic pressure. Emerick concluded that the mean differences exceeded an a priori standard (for accuracy) of a mean difference of less than 5 mm Hg set by the Association for the Advancement of Medical Instrumentation.¹¹ The author¹⁰ further concluded that if the wrist were used to measure blood pressure, adjustments would be required. Emerick recommended that 10 mm Hg be subtracted from the systolic and diastolic pressure measurements obtained in the wrist or that the wrist be elevated approximately 15 cm when a reading is obtained.

Emerick’s study¹⁰ was the first in which standards were incorporated to determine if observed differences between 2 measurements of blood pressure obtained at different locations on the upper extremity would be acceptable in practice. Unfortunately, Emerick did not validate the recommendations for adjusting the measurements obtained in the wrist or repositioning the wrist as a means of compensating for the observed differences.

In 1999, Singer et al⁴ reported the results of a study designed to determine if measurements of blood pressure in the forearm could be used to predict blood pressure in the upper arm. Measurements were obtained in a convenience sample of 151 subjects (mean age 35 years, SD 15.7 years) by a single triage nurse in a university-affiliated emergency department. Subjects were excluded if they were less than 5 years old, had arrived by ambulance, or had any condition that would preclude a blood pressure measurement (eg, amputation). An Omega 1400 automatic noninvasive blood pressure monitor (In Vivo Laboratories, Inc, Orlando, Fla) was used to obtain sequential measurements of blood pressure and heart rate in the left upper extremity (positioned at heart level) of the subjects while the subjects were seated. Cuff selection was based on recommendations from the AHA.

Mean systolic pressures were 129.8 mm Hg (SD 20.7) in the forearm and 126.2 mm Hg (SD 17.6) in the upper arm (P = .002). Mean diastolic pressures were 80.7 mm Hg (SD 14.5) in the forearm and 76.8 mm Hg (SD 13.4) in the upper arm (P < .001). Mean differences were 3.6 mm Hg for systolic pressures and 3.9 mm Hg for diastolic pressures. Correlations between the measurements obtained in the upper arm and the forearm were 0.75 for systolic pressures (P < .001) and 0.72 for diastolic pressures (P < .001). Heart rates recorded during the blood measurements in the upper arm and forearm did not differ significantly (P = .56). Measurements obtained in the 2 sites differed by 10 mm Hg or less in 58% of the patients for systolic pressure and in 70% for diastolic pressure and by 20 mm Hg or less in 86% of the patients for systolic pressure and in 94% for diastolic pressure. Differences in measurements obtained in the 2 sites were not related to demographic variables (age, sex, race, blood pressure, cardiac risk factors, and diagnosis). The researchers concluded that use of the forearm could be considered reasonable when the upper arm was not readily available.

The findings from this study⁴ contributed important information to the issue of using the forearm to measure blood pressure. It is the only study in which an automatic noninvasive monitor was used to sequentially measure blood pressure in the upper arm and the forearm and then the correlation between the 2 values was determined. It is unclear how the forearm was positioned in relation to the level of the heart and whether cuff selections for measurements in the forearm were based on forearm circumferences. Finally, Singer et al commented, rightfully so, that the Omega 1400 monitor was not designed specifically for measuring blood pressure in the forearm. Together, these aspects of the study⁴ may have introduced instrumentation and measurement errors.

In summary, the research on using the forearm to measure blood pressure is extremely limited. We have observed an increase in the use of the forearm as an alternative site for measuring blood pressure in a variety of clinical areas (the emergency department, medical-surgical units, labor and delivery units, and the operating room). Clearly the research to date does not provide the evidence needed to verify the accuracy of this practice. The purpose of our study was to compare automatic noninvasive measurements of blood pressures in the upper arm and forearm.

	Methods

Top Abstract Review of the Literature Methods Results Discussion Conclusions and Clinical... APPENDIX References

 
A descriptive correlational comparison study was conducted in the emergency departments of 2 acute care facilities (Wilmington Hospital and Christiana Hospital) of a 1071-bed teaching hospital system (Christiana Care Health System) located in Delaware. The study was approved by the institutional review board and was implemented in June of 2003.

Sample
A convenience sample of English-speaking patients 5 years or older whose medical condition was stable and who entered the triage or fast-track areas of the emergency department either on foot or in a wheelchair were recruited. Potential subjects were excluded if (1) they bypassed the triage or fast-track areas and/or their medical condition was unstable, (2) the noninvasive blood pressure cuffs available did not fit their upper arms or their forearms, (3) they could not expose the left upper extremity for measurement of blood pressure, and/or (4) they had conditions (eg, amputation, trauma to arm, previous mastectomy, arteriovenous fistula) that precluded measurement of blood pressure in the left upper extremity.

We determined that a sample of 189 patients would have 90% power to detect an effect size of 5 mm Hg (SD 15 mm Hg) at = .05. We considered a difference of 5 mm Hg or greater between upper arm and forearm blood pressures clinically significant.

Instruments
The Welch Allyn Spot Vital Signs 420 series noninvasive blood pressure monitor (Welch Allyn, Beaverton, Ore) was used for the study. The blood pressure accuracy of the monitor meets or exceeds SP10-1992 Association for the Advancement of Medical Instrumentation standards for noninvasive blood pressure accuracy (SE ±5 mm Hg, SD 8 mm Hg).¹¹ The accuracy of the measurements of blood pressure was validated by using the arm only.³ Four new monitors provided by Welch Allyn were used exclusively for study purposes. The monitors were calibrated by the hospital system’s clinical engineering department before the start of and at the conclusion of data collection, which occurred during a 2-week period.

A demographic/clinical data form was developed for this study. For each patient in the study, medical record number, age, sex, race, medical history, cardiac risk factors, medications, and emergency department medical diagnosis were collected retrospectively through electronic chart review. Retrospective chart review was more convenient because charts were not always accessible and complete during data collection periods. Clinical data collected included upper arm and forearm circumferences, upper arm and forearm cuff sizes, upper arm and forearm blood pressure measurements, upper arm and forearm heart rates, order of measurements, and time between measurements.

Procedure
Patients who met the inclusion criteria were approached after they had completed the emergency department registration process. They were given a brief, scripted description of the study to determine their interest in participating. If interested, triage patients were escorted from the waiting room to a private data collection station in the emergency department; fast-track patients were seen in the fast-track treatment areas. Data collectors provided patients a formal explanation of the study and obtained written informed consent. Permission from a legal guardian or parent was required for subjects between 5 and 18 years old. Assent was obtained from subjects ages 7 to 18 years.

Blood pressures were measured in accordance with the 1993 AHA standards.¹ Each subject was seated in a chair or assisted to a seated position (dangling) on a stretcher and was asked to expose the left upper extremity. Data collectors measured upper arm and forearm circumference and determined cuff size according to the manufacturer’s recommendations (Table 1). The extremity was placed at heart level. For measurements in the forearm, the extremity was positioned on a bedside table raised to the level of the phlebostatic axis; for measurements in the upper arm, the extremity was placed in the subject’s lap. Blood pressures were measured by using one of the Welch Allyn monitors. The order in which the measurements were obtained was alternated. For example, in the first subject, the blood pressure was measured first in the forearm and then, at least 1 minute but no more than 2 minutes later, in the upper arm. For the next subject, blood pressure was measured first in the upper arm and then in the forearm. Measurements were obtained in subsequent subjects in a similar manner. Subjects were asked to remain quiet during all measurements. During blood pressure measurement, pulse rate was also obtained from the monitor.

Before data collection, 8 data collectors were trained in the use of the blood pressure monitor by a Welch Allyn representative. Interrater reliability was established for measurement of upper arm and forearm circumferences, selection of cuff size, and positioning of the forearm at heart level (mean = .97). Two researchers abstracted demographic and clinical data from the electronic charts of the first 10 subjects, and this trial information was validated by a third researcher so that 100% agreement was obtained. Intercoder reliability was established at 97% agreement for the 4 data coders before data entry into Microsoft Word files.

Data Analysis
Data were entered into Microsoft Word and imported into SAS version 8.2 (SAS Inc, Cary NC). Descriptive statistics were used to characterize the sample and to summarize the demographic and clinical data. Measures of central tendency aggregated data and reflected the group of subjects. The Pearson correlation coefficient was used to determine the relationship between measurements of blood pressure, MAP, and heart rate in the upper arm and the forearm. The difference between mean blood pressures in the upper arm and the forearm was tested by using paired t tests. The influence of the variables of age, sex, race, cardiac risk factors, current medications, and emergency department diagnosis on the differences was tested by using analysis of covariance. Significance was set at P < .05.

In order to obtain a clearer picture of the measurements for individual subjects, a Bland-Altman agreement analysis¹² was used to determine the extent of agreement between measurements of blood pressure in the upper arm and the forearm. The purpose of the agreement analysis, recommended for comparisons of clinical measurements, is to answer the question, Are forearm blood pressures interchangeable with upper arm blood pressures? This analysis was performed by using MedCalc for Windows, version 7.4.2.0 [EC] (Med-Calc Software, Mariakerke, Belgium).

	Results

Top Abstract Review of the Literature Methods Results Discussion Conclusions and Clinical... APPENDIX References

 
The demographics of the sample (n = 204) are given in Table 2. Post hoc power was 80%. Approximately 52% of the subjects were male; subjects’ ages ranged from 6 to 91 years (mean 36.5 years). Seventeen children (subjects between ages 5 and 18 years) were included in the study. Forty-six percent of the subjects were black, 46% were white, 7% were Hispanic, and 1% were reported as unknown race. The remaining demographics (Table 3) were coded according to frequency; some subjects had more than 1 category identified under cardiac risk factors, current medications, and emergency department diagnosis. The most common cardiac risk factor was hypertension (n = 30). Medications related to pain relief (eg, acetaminophen, ibuprofen) were the drugs most often used by subjects. Of the 242 emergency department diagnoses, 34% were categorized as musculoskeletal (eg, muscle strain, back pain). Eyes-ears-nose-throat (eg, dental abscess, pharyngitis) and skin (eg, laceration, contusion) were the next most common diagnostic categories (17% and 14%, respectively).

The Bland-Altman procedure was used to determine agreement between the upper arm and the forearm for measurements of systolic blood pressure, diastolic blood pressure, and MAP. The bias (mean difference between upper arm and forearm measurements) for systolic pressures was –1.3 mm Hg (Figure 1). The computed upper and lower limits of agreement for systolic pressures were +17.1 mm Hg and –19.7 mm Hg, respectively. The limits of agreement for the diastolic pressures were +14.5 mm Hg and –16.5 mm Hg (Figure 2), and the bias was –1.0 mm Hg. For MAPs, the limits of agreement were +13.9 mm Hg and –16.0 mm Hg, and the bias was –1.1 mm Hg (Figure 3).

View larger version (25K): [in this window] [in a new window]   Figure 1 Bland-Altman analysis for systolic blood pressures

View larger version (26K): [in this window] [in a new window]   Figure 2 Bland-Altman analysis for diastolic blood pressures

View larger version (27K): [in this window] [in a new window]   Figure 3 Bland-Altman analysis of mean arterial blood pressures

Differences between measurements of systolic, diastolic, and mean arterial blood pressures in the upper arm differed widely from measurements in the forearm, indicating that measurements obtained in the 2 sites are not interchangeable.

 

Paired t tests revealed that as a group, smokers (n = 22) had significant differences between the upper arm and the forearm in measurements of systolic pressure (t = –2.42, P = .02) and MAP (t = –2.40, P = .03); measurements in the forearm were more than 3 mm Hg higher than those in the upper arm. Black subjects (n = 93) had significant differences between measurements in the upper arm and the forearm of MAP (t=–2.98, P = .004), systolic pressure (t = –3.28, P = .002), and diastolic pressure (t = –2.26, P = .03); measurements in the forearm were 1.7 to 3.4 mm Hg higher than those in the upper arm. Paired t tests indicated no significant differences in the blood pressures of children (<18 years old), elderly subjects (>65 years old), males, females, small and large cuff sizes, cuff order, emergency department diagnoses, current medications, or cardiac risk factors other than smoking.

	Discussion

Top Abstract Review of the Literature Methods Results Discussion Conclusions and Clinical... APPENDIX References

 
Use of the forearm for measuring blood pressure has been incorporated into practice by clinicians despite the lack of research evidence supporting its use. In this replication study, we compared automatic noninvasive measurements of blood pressure in the forearm and the upper arm. Our results indicated a strong correlation between measurements obtained at the 2 sites. Nevertheless, experts advise that correlation does not measure the agreement of 2 variables; rather it simply indicates the strength of the relationship.^12,13 Paired t tests, used to analyze group data, revealed no significant differences between the 2 sites in measurements of diastolic blood pressure or MAP and a statistically significant difference between measurements of systolic blood pressures (mean systolic blood pressure 122.5 and 123.9 mm Hg for the upper arm and the forearm, respectively; with a mean difference of 1.31 mm Hg between measurements obtained at the 2 sites). However, the Bland-Altman procedure, a more relevant analysis when measurements of individual subjects are examined, revealed different results (see Appendix).

The Bland-Altman procedure was used to calculate the amount of agreement between measurements of systolic pressure, diastolic pressure, and MAP in the upper arm and the forearm. The distances between the mean values and the limits of agreement ranged from 15 mm Hg (MAP) to 18.4 mm Hg (systolic pressure). This finding means, for example, that in approximately 95% of subjects, measurements of systolic pressure in the forearm will be within 18.4 mm Hg above and 18.4 mm Hg below the subjects’ mean measurements of systolic pressure in the upper arm and the forearm. Because differences of this magnitude could be clinically important, the analysis indicates that measurements of systolic pressure obtained by using the forearm are not interchangeable with measurements obtained by using the upper arm. In fact, the limits of agreement for measurements of systolic blood pressure, diastolic blood pressure, and MAP in the 2 sites all exceeded the difference of 5 mm Hg we specified as indicating a clinically significant difference. Thus, the degree of agreement for the measurements of all 3 blood pressures in the 2 sites is not clinically acceptable, and the forearm and upper arm measurements are not interchangeable.

A review of demographics for the 5 subjects who had dramatically significant differences between measurements in the upper arm and measurements in the forearm did not reveal any identifiable patterns. Possibly some unexplained or undisclosed vascular problems existed in this group.

In this study, we replicated the research of Singer et al⁴ with modifications to minimize some of the limitations of their study. In our study, we used pre and post hoc power analyses to determine adequacy of sample size, thus reducing the risk of type II error.¹⁴ Although it was unclear whether Singer et al⁴ selected cuff sizes for measurements obtained in the forearm on the basis of the circumference of the forearm or the circumference of the upper arm, we included separate measurements of the circumferences of the 2 sites and subsequently selected cuff sizes on the basis of those measurements; that is, cuff sizes for the forearm were based on forearm circumference, and cuff sizes for the upper arm were based on upper arm circumference. Additionally, Singer et al⁴ positioned the upper arm at heart level but did not indicate where the forearm was positioned with respect to heart level. In our study, we placed both the forearm and the upper arm at heart level, a step that necessitated a change in the position of the upper extremity, on the basis of the phlebostatic axis, for each measurement of blood in the forearm. Singer et al noted that a limitation of their study was that more subjects had blood pressure measured in the upper arm first. In our study, sequence of measurements in the forearm and upper arm was alternated for each subject. Finally, we used the Bland-Altman procedure to analyze the amount of agreement between measurements obtained at the 2 sites.

Our overall results differ from those of Singer et al,⁴ specifically when the results of the Bland-Altman analysis are included. Although like Singer et al, we found a statistically, but not clinically, significant difference between mean measurements of systolic pressure in the forearm and the upper arm, the Bland-Altman analysis indicated that in individuals, measurements obtained from the 2 sites are not interchangeable. Singer et al found that mean differences were unrelated to demographics, presence of cardiovascular risk factors, or clinical diagnoses. However, in our study, in smokers and blacks, mean measurements of systolic pressure in the forearm were higher than mean measurements in the upper arm.

Limitations and Recommendations
Our study has several limitations. A convenience sample allows minimal control for biases.¹⁵ We used demographic and other relevant information such as patients’ history of cardiovascular disease, medications, and emergency department diagnoses to help control these biases. Additionally, we obtained information on cardiac risk factors retrospectively through electronic chart review, and this information was not consistently recorded by emergency department practitioners during all history and physical examinations. Findings related to smokers should be considered cautiously because the number of smokers was small and was based on patients’ self-reports.

Generalizability of the findings to supine hospitalized patients may be limited because the subjects were seated when blood pressure was measured. Generalizability of findings to patients who require extra large blood pressure cuffs is also limited because of the small number of these subjects in the study. Last, noninvasive blood pressure monitors found in hospitals are not specifically designed for measuring blood pressure in the forearm.

Future Research
According to the Bland-Altman analysis, differences between the 2 sites in measurements of systolic pressure, diastolic pressure, and MAP ranged between approximately 14 and 20 mm Hg. Use of a predictive design in future investigations of the relationship between measurements of blood pressure in the forearm and measurements in the upper arm would provide a more concrete method of determining the accuracy of measurements obtained in the forearm.

In our study, both the upper arm and the forearm were placed at heart level for measurements of blood pressure. In practice, the forearm is often resting in the patient’s lap when the patient is seated. Future research is in progress to examine the effects of the position of the forearm on the accuracy of measurements of blood pressure and the effects of supine versus head-of-the-bed elevation on measurements of blood pressure obtained by using the forearm. In our study, the frequency of error codes displayed on the Welch Allyn monitor was increased when the extra large adult size cuff was used, suggesting that measurement of blood pressure in patients who are obese may be a problem. Influence of body mass index and race on differences between measurements obtained from the 2 sites (forearm and upper arm) should also be explored. Finally, future studies are indicated in children and in patients who are critically ill.

	Conclusions and Clinical Implications

Top Abstract Review of the Literature Methods Results Discussion Conclusions and Clinical... APPENDIX References

 
Our findings indicate that in seated, adult patients in stable clinical condition in the clinical setting, despite strict attention to correct cuff size and placement of the upper arm or forearm at heart level, noninvasive measurements of blood pressure obtained in the forearm are not interchangeable with measurements obtained in the upper arm. Realistically, situations in which the forearm is used to measure blood pressure will continue to arise in the clinical setting because of the inaccessibility of the upper arm and/or lack of proper-sized cuffs for the upper arm. Our findings underscore the importance of indicating which site was used to measure blood pressure, forearm or upper arm, when documenting blood pressure, and of using the same site for serial measurements to determine trends in blood pressure values. Additionally, proper selection of cuff size is imperative for accurate measurements.¹⁶ Clinicians who manage patients must realize that measurements of blood pressure obtained in the forearm may differ from measurements obtained in the upper arm by up to 20 mm Hg.

	APPENDIX

Top Abstract Review of the Literature Methods Results Discussion Conclusions and Clinical... APPENDIX References

 
The results reported in this article differ from those reported at the poster session of the May 2004 National Teaching Institute of the American Association of Critical-Care Nurses and subsequently published as an abstract in the American Journal of Critical Care. At that time, statistical analysis by paired t tests indicated that although measurements of blood pressure obtained by using the forearm differed statistically from measurements obtained by using the uper arm, this difference was not clinically significant for subjects as a group.

Upon editorial review of the study manuscript, it was suggested that the Bland-Altman procedure be included in the statistical analyses. This procedure revealed a wider variation in differences between measurements of systolic, diastolic, and mean arterial blood pressures obtained in the upper arm and measurements obtained in the forearm for individual subjects. Revisions to the discussion, conclusions, and clinical implications of the study were necessary and are reported in this article. We apologize for any confusion that these changes may have caused.

	ACKNOWLEDGMENTS

 
This research was funded in part by a grant from the Southeastern Pennsylvania Chapter of the American Association of Critical-Care Nurses, Philadelphia, Pa. Blood pressure monitors and cuffs were provided by Welch Allyn, Beaverton, Ore. We gratefully acknowledge the statistical assistance of Julie Waterhouse, RN, PhD, associate professor, Department of Nursing, University of Delaware, Newark, Del. We also thank the emergency department administration and staff at Christiana Care Health System who fully supported this study. Finally, we are especially grateful to the patients who participated in the study.

To purchase reprints, contact The InnoVision Group, 101 Columbia, Aliso Viejo, CA 92656. Phone, (800) 809-2273 or (949) 362-2050 (ext 532); fax, (949) 362-2049; e-mail, reprints{at}aacn.org.

Commentary by Mary Jo Grap (see shaded boxes).

	REFERENCES

Top Abstract Review of the Literature Methods Results Discussion Conclusions and Clinical... APPENDIX References

 

1. Perloff D, Grim C, Flack J, et al. Human Blood Pressure Determination by Sphygmomanometry. Dallas, Tex: American Heart Association; 1993.
2. Hill MN, Grim CM. How to take a precise blood pressure. Am J Nurs. February 1991;91:38–42.[Medline]
3. Spot Vital Signs 420 Series Operator’s Manual. Beaverton, Ore: Welch Allyn; 2001.
4. Singer AJ, Kahn SR, Thode HC Jr, Hollander JE. Comparison of forearm and upper arm blood pressures. Prehosp Emerg Care. 1999;3:123–126.[Medline]
5. Dinamap Compact Monitor Operation Manual. Tampa, Fla: Critikon; 1998.
6. Dobbin KR. Noninvasive Blood Pressure Monitoring. Aliso Viejo, Calif: American Association of Critical-Care Nurses; 1996.
7. Tachovsky BJ. Indirect auscultatory blood pressure measurement at two sites in the arm. Res Nurs Health. 1985;8:125–129.[Medline]
8. Latman NS, Coker N, Teague C. Evaluation of an instrument for noninvasive blood pressure monitoring in the forearm. Biomed Instrum Technol. 1996;30:160–163.[Medline]
9. Latman NS, Latman A. Evaluation of instruments for noninvasive blood pressure monitoring of the wrist. Biomed Instrum Technol. 1997;31:63–68.[Medline]
10. Emerick DR. An evaluation of non-invasive blood pressure (NIBP) monitoring on the wrist: comparison with upper arm NIBP measurement. Anaesth Intensive Care. 2002;30:43–47.[Medline]
11. White WB, Berson AS, Robbins C, et al. National standard for measurement of resting and ambulatory blood pressures with automated sphygmomanometers. Hypertension. 1993;21:504–509.[Abstract/Free Full Text]
12. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. February 8, 1986;1:307–310.[Medline]
13. Flemon WW, Littner MR. Measuring agreement between diagnostic devices. Chest. 2003;124:1535–1542.[Abstract/Free Full Text]
14. Lederman RP. Power analysis: I. MCN Am J Matern Child Nurs. 1993;18:313.[Medline]
15. Burns N, Grove SK. The Practice of Nursing Research: Conduct, Critique, and Utilization. 4th ed. Philadelphia, Pa: WB Saunders; 2001.
16. Graves JW. Prevalence of blood pressure cuff sizes in a referral practice of 430 consecutive adult hypertensives. Blood Press Monit. 2001;6:17–20.[Medline]

This article has been cited by other articles:

				K. L. Domiano, S. M. Hinck, D. L. Savinske, and K. L. Hope Comparison of Upper Arm and Forearm Blood Pressure Clin Nurs Res, November 1, 2008; 17(4): 241 - 250. [Abstract] [PDF]

				C. A. Rauen, M. Chulay, E. Bridges, K. M. Vollman, and R. Arbour Seven Evidence-Based Practice Habits: Putting Some Sacred Cows Out to Pasture Crit. Care Nurse, April 1, 2008; 28(2): 98 - 123. [Full Text] [PDF]

				K. Schell, D. Lyons, E. Bradley, L. Bucher, M. Seckel, S. Wakai, E. Carson, J. Waterhouse, M. Chichester, D. Bartell, et al. Clinical Comparison of Automatic, Noninvasive Measurements of Blood Pressure in the Forearm and Upper Arm With the Patient Supine or With the Head of the Bed Raised 45{degrees}: A Follow-Up Study Am. J. Crit. Care., March 1, 2006; 15(2): 196 - 205. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Respond to This Article Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Schell, K. Articles by Simpson, K. Search for Related Content PubMed PubMed Citation Articles by Schell, K. Articles by Simpson, K.