Comparison of Central and Peripheral Arterial Blood Pressure Gradients in Critically Ill Patients: A Systematic Review and Meta-Analysis

Abstract

OBJECTIVES:

Measurement of blood pressure taken from different anatomical sites, are often perceived as interchangeable, despite them representing different parts of the systemic circulation. We aimed to perform a systematic review and meta-analysis on blood pressure differences between central and peripheral arterial cannulation in critically ill patients.

DATA SOURCES:

We searched MEDLINE, Cochrane Central Register of Controlled Trials, and Embase from inception to December 26, 2023, using Medical Subject Headings (MeSH) terms and keywords.

STUDY SELECTION:

Observation study of adult patients in ICUs and operating rooms who underwent simultaneous central (femoral, axillary, or subclavian artery) and peripheral (radial, brachial, or dorsalis pedis artery) arterial catheter placement in ICUs and operating rooms.

DATA EXTRACTION:

We screened and extracted studies independently and in duplicate. We assessed risk of bias using the revised Quality Assessment for Studies of Diagnostic Accuracy tool.

DATA SYNTHESIS:

Twenty-four studies that enrolled 1598 patients in total were included. Central pressures (mean arterial pressure [MAP] and systolic blood pressure [SBP]) were found to be significantly higher than their peripheral counterparts, with mean gradients of 3.5 and 8.0 mm Hg, respectively. However, there was no statistically significant difference in central or peripheral diastolic blood pressure (DBP). Subgroup analysis further highlighted a higher MAP gradient during the on-cardiopulmonary bypass stage of cardiac surgery, reperfusion stage of liver transplant, and in nonsurgical critically ill patients. SBP or DBP gradient did not demonstrate any subgroup specific changes.

CONCLUSIONS:

SBP and MAP obtained by central arterial cannulation were higher than peripheral arterial cannulation; however, clinical implication of a difference of 8.0 mm Hg in SBP and 3.5 mm Hg in MAP remains unclear. Our current clinical practices preferring peripheral arterial lines need not change.

KEY POINTS

Question: What is the true difference between blood pressure obtained by arterial catheter in the central vs. peripheral arterial circulation in high-risk surgical and critically ill patients?

Findings: Our systematic review and meta-analysis found mean arterial pressure (MAP) and systolic blood pressure (SBP) to be higher in central arterial circulation by 3.5 and 8.0 mm Hg, respectively, while no statistical difference was observed in diastolic blood pressure.

Meaning: The comparison revealed a modest difference in SBP and MAP between central and peripheral arterial cannulation; however, the clinical relevance of these differences remains uncertain.

Maintaining organ perfusion is the cornerstone of resuscitation, both in the ICUs and operating rooms. Blood pressure monitoring is used as the key surrogate marker for tissue perfusion and is the trigger for many interventions, specifically fluid and vasopressors to achieve resuscitation targets. Hence, accurate blood pressure monitoring is essential for appropriate titration of these important therapies. Invasive arterial pressure monitoring is routinely used in high-risk surgical and critically ill patients, especially when patients are hemodynamically unstable, as it allows accurate and beat-to-beat assessment of arterial pressure. The radial artery is the preferred site because of its technical ease and lower risk of infection and complications compared with femoral artery, the other commonly used site (1–5). Measurement of blood pressure taken from different anatomical sites, are often perceived as interchangeable, despite them representing different parts of the systemic circulation, each with different vascular compliance.

Earlier studies have shown variable results regarding the concordance of blood pressure readings from different arterial sites in high-risk surgical and critically ill patients. Some studies found a good correlation between these measurements, while others observed genuine differences in systolic blood pressure (SBP) and mean arterial pressure (MAP) between central and peripheral sites (6–29). This disparity has the potential to influence clinical decision-making and patient outcomes as the accuracy of invasive blood pressure monitoring is crucial in determining the stability of the circulatory system, perfusion to major organs and hence guide the need for interventions. However, current understanding is fragmented due to inherent inconsistencies and variations in the studies conducted to date. Therefore, this meta-analysis aims to systematically review and synthesize the available evidence to determine the true gradient between central and peripheral arterial pressure. We hypothesized that the gradient between central and peripheral arterial pressure measurements in high-risk surgical and critically ill patients may not be significant enough to change our current practice preferring peripheral arterial cannulation.

MATERIALS AND METHODS

Protocol Registration

This study complied with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (30, 31) and the Meta-Analyses of Observational Studies in Epidemiology proposals (32). Our protocol was registered at PROSPERO (Center for Reviews and Dissemination : 42023394057) with no amendments.

Search Strategy

We searched MEDLINE, Cochrane Central Register of Controlled Trials, and Embase, from inception to December 26, 2023, using the search strategy that included the keywords: “radial or brachial or dorsalis pedis” or “femoral or axillary or subclavian” and “blood pressure or arterial pressure or pressure gradient or difference.” An information specialist developed and implemented the search strategy. We also conducted manual searches of reference lists of included studies and other systematic reviews. The detailed search strategy is described in Supplementary Table 1 (http://links.lww.com/CCX/B343). We only included articles published in English. We excluded conference proceedings as they lacked detailed analysis, and studies published in journals that were not peer-reviewed. Our search was updated on December 26, 2023.

Study Selection and Inclusion/Exclusion Criteria

Two authors (D.H., R.S.) screened the abstracts on the inclusion criteria. We then retrieved and reviewed the full texts. We performed those processes above on Covidence systematic review software, Veritas Health Innovation, Melbourne, VIC, Australia. Inclusion criteria were as follows:

• 1) Patient population: Adult (≥ 18 yr) critically ill patients who underwent central arterial catheter placement (femoral, axillary, or subclavian artery) and peripheral arterial catheter placement (radial, brachial, or dorsalis pedis artery), simultaneously.

• 2) Intervention/exposure: Central arterial pressure.

• 3) Comparison: Peripheral arterial pressure.

• 4) Outcomes: The pressure gradient defined as the difference calculated using formula (central arterial pressure–peripheral arterial pressure) for MAP, SBP, and diastolic blood pressure (DBP).

• 5) Study type: Randomized control trials, prospective and retrospective observational studies, and cross-sectional studies.

Exclusion criteria were as follows:

• 1) Studies involving healthy volunteers or cardiac arrest patients.

• 2) Studies measuring blood pressure noninvasively.

• 3) Studies with a total sample size of fewer than 20 patients to minimize heterogeneity and imprecision (33).

• 4) Studies published in a non-English language.

Data Extraction and Statistical Analysis

Two authors (D.H., R.S.) independently extracted the following data from the eligible studies: year of publication, number of participants, patient population and setting, and the mean and sds of actual numbers and pressure gradients of MAP, SBP, and DBP. The central arterial blood pressure was defined as the pressure measured using a catheter inserted in femoral, axillary, or subclavian artery, whereas the peripheral arterial blood pressure was defined as the arterial pressure using catheter placed in radial, brachial, or dorsalis pedis artery.

The primary outcome was the pressure gradients in MAP between central and peripheral arterial catheters. The secondary outcomes were the pressure gradients in SBP and DBP between central and peripheral arterial monitoring systems. If a study categorized patients into separate groups (e.g., based on gender, age), we merged these groups together before conducting the meta-analysis, to obtain the mean and sd of the measured values of blood pressure and the pressure gradients. If the study did not provide the gradients of arterial blood pressure between central and peripheral arterial lines, along with their sd, we calculated the pressure gradients using the measured values of blood pressure for both central and peripheral blood pressure, applying the formula (central arterial pressure–peripheral arterial pressure). To estimate the sd of pressure gradients, intraindividual correlation was used in the calculations (34). Intraindividual correlation was estimated with the newest study included into this meta-analysis, which provided the mean and sd for both the measured values of MAP in central and peripheral arterial blood pressures and the MAP gradients between central and peripheral arterial pressures (35). In case the study provided more than one-time blood pressure read, we used the last blood pressure recorded.

Subgroup analysis was also performed for MAP among distinct patient groups: those undergoing cardiac surgery, liver transplant surgery, and nonsurgical critically ill patients. For cardiac surgery patients, measurements were taken at pre-, on-, and off-cardiopulmonary bypass stages. For liver transplant patients, measurements were taken after induction of anesthesia, during reperfusion, and post-surgery. Subgroup analysis based on the vasopressor dose was not feasible given paucity of individual data and limited study number.

Three types of sensitivity analyses were conducted. The first analysis included all studies, even those with a sample size of fewer than 20, for the gradients of MAP, SBP, and DBP. The second analysis entailed the removal of one study at a time to assess its impact on our primary and secondary outcome (36). The third analysis focused on studies where the article provided the mean and sd for the gradients of MAP, SBP, and DBP.

Data were expressed as the pooled mean gradient using the random-effects method. We reported a 95% CI for all estimates. I² statistic with 95% CI was used to assess heterogeneity. The I² statistic ranges from 0% to 100% (I² < 25%: low heterogeneity, I² = 25–50%: moderate heterogeneity, and I² > 50%: substantial heterogeneity). Funnel plots were created for the primary outcome: the log odds ratios were plotted, and symmetry was tested using the Begg’s rank correlation. All analyses were performed using Comprehensive Meta-Analysis V.3 (Biostat, Englewood, NJ). A p value of equal to 0.05 was statistically significant.

Assessment of the Risk of Bias

The queries to assess the risk of bias were independently evaluated by two authors (D.H., R.S.) and verified by another author (S.D.). When there were disagreements, the discussion was held to reach a consensus. We assessed the study quality of each article using the quality of study with the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool (37).

RESULTS

Results of the Search

The described search strategy produced 2434 articles. Through citation searching, we added one more article. After removing the duplicate articles, 2253 articles remained. After screening the titles and abstracts, 2160 studies were found to be clearly irrelevant to our study. We retrieved the full texts of the remaining 93 studies and assessed the eligibility of these studies. Of these, one study was an editorial. One study was written in non-English language. Three studies were animal studies. One article was a correspondence. Thirteen articles were conference proceedings. Four studies investigated a different population group such as pediatric patients, healthy volunteers, or cardiac arrest patients. Seven articles were nonresearch letter article. Four studies were excluded because the total sample size was less than 20 patients (38–41). Thirty-five articles were excluded because outcomes of interest were not captured. Therefore, after assessing the eligibility of these studies, 24 studies that enrolled 1598 patients in total were included, as shown in Figure 1.

Figure 1.

Flow chart of the identification and selection of studies for inclusion.

Basic Features of Included Studies

The basic features of the included studies are summarized in Table 1; and Supplemental Table 2 (http://links.lww.com/CCX/B343). The earliest article was published in 1987 (29), and the latest article was published in 2022 (6–8). Twenty-three studies compared the femoral artery with the radial artery, whereas one study compared the femoral artery with the brachial artery. Two studies provided both the measured values of MAP in central and peripheral arterial blood pressures and the MAP gradients between central and peripheral arterial pressures. Nine studies dealt with those undergoing cardiac surgery, eight studies with liver transplant recipients, five studies with nonsurgical critically ill patients, and two studies with other patient populations. In studies where pressure gradients were not provided, the intraindividual correlation was inferred from the most recent study, which provided the mean and sd for both the measured values of MAP in central and peripheral arterial blood pressures. Baba et al (27) and Lee et al (15) provided both numbers. From the data presented by Lee et al (15), the intraindividual correlation was computed to be 90%. The predicted pressure gradients are also summarized in Supplemental Table 2 (http://links.lww.com/CCX/B343).

TABLE 1.

Summary of the Articles

References	Country	No. of Patients	Study Design	Comparison	Setting
Acosta et al (26)	Spain	47	Not specified	Femoral vs. radial artery	Liver transplantation
Ahmad et al (14)	Pakistan	60	Prospective, single-center, observational	Femoral vs. radial artery	Cardiac surgery, requiring high inotropic/vasopressor support on weaning from CPB
Arnal et al (24)	Spain	72	Prospective, single-center, observational	Femoral vs. radial artery	Liver transplant surgery
Baba et al (27)	Japan	75	Prospective, single-center, observational	Femoral vs. radial artery	CABG requiring CPB
Bhaskar et al (8)	India	80	Prospective, single-center, observational	Femoral vs. radial artery	Septic shock
Broch et al (19)	Germany	50	Prospective, single-center, observational	Femoral vs. radial artery	Elective CABG requiring CPB
Chauhan et al (25)	India	60	Prospective, single-center, observational	Femoral vs. radial artery	Cardiac surgery requiring CPB
Galluccio et al (21)	Australia	24	Prospective, single-center, observational	Femoral vs. radial artery	Sepsis 15 (62%), cardiogenic shock 5 (21%), other 4 (17%)
Gravlee et al (28)	United States	31	Prospective, single-center, observational	Femoral vs. brachial artery	Adults undergoing cardiac procedures requiring CPB
Jacquet-Lagrèze et al (11)	Canada	81	Prospective, multicenter, observational	Femoral vs. radial artery	Cardiac surgery 68 (84%), septic shock 5 (6%), other 8 (10%)
Kim et al (20)	Korea	31	Prospective, single-center, observational	Femoral vs. radial artery	Liver transplantation
Kim et al (18)	Korea	37	Prospective, single-center, observational	Femoral vs. radial artery	Septic shock
Kim et al (10)	Korea	80	Retrospective, single-center observational	Femoral vs. radial artery	Liver transplant
Kim et al (7)	United States	60	Retrospective, single-center observational	Femoral vs. radial artery	Liver transplant recipients
Lee et al (15)	Australia	25	Prospective, single-center, observational	Femoral vs. radial artery	Liver transplant recipients
Maddali et al (13)	Oman	27	Prospective, single-center, observational	Femoral vs. radial artery	CABG requiring CPB
Mignini et al (23)	Argentina	55	Prospective, single-center, observational	Femoral vs. radial artery	Critically ill patients
Mohr et al (29)	Israel	48	Not specified	Femoral vs. radial artery	CABG
Nakamura et al (12)	Japan	52	Prospective, single-center, observational	Femoral vs. radial artery	Minimally invasive cardiac surgery using CPB
Ruiz et al (17)	France	32	Prospective, single-center, observational	Femoral vs. radial artery	Aortic endografting under general anesthesia
Shin et al (22)	Korea	36	Prospective, single-center, observational	Femoral vs. radial artery	Liver transplantation
Sun et al (16)	China	401	Retrospective, single-center, observational	Femoral vs. radial artery	Cardiac surgery requiring CPB
Thomas et al (9)	India	102	Retrospective, single-center, observational	Femoral vs. radial artery	Liver transplant
Wisanusattra et al (6)	Thailand	32	Prospective, single-center, observational	Femoral vs. radial artery	Patients with refractory shock

CABG = coronary artery bypass grafting, CPB = cardiopulmonary bypass.

Primary Outcome

Out of the 24 studies conducted, all supplied information regarding gradients of MAP or data that facilitated the estimation of these gradients. The pooled mean MAP gradient was 3.5 mm Hg (95% CI, 2.1–4.8 mm Hg), as summarized in Table 2. No heterogeneity was observed (I² = 0%). The corresponding forest plot can be seen in Figure 2.

TABLE 2.

Main Results

Group	Systolic Pressure Gradient (mm Hg)	Mean Arterial Pressure Gradient (mm Hg)	Diastolic Pressure Gradient (mm Hg)
All patients	8.0 (3.9–12.0)	3.5 (2.1–4.8)	1.1 (–0.1 to 2.4)
Group	Stage	Mean Arterial Pressure Gradient (mm Hg)
Cardiac surgery	After induction	1.8 (0.8–2.8)
Cardiac surgery	On pump	5.1 (3.5–6.6)
Cardiac surgery	Post-pump	2.8 (0.3–5.3)
Liver transplantation	After induction	3.4 (0.4–6.4)
Liver transplantation	Reperfusion	5.0 (0.9–9.1)
Liver transplantation	Post-perfusion	3.0 (0.9–5.1)
Non-surgical critically-ill patients		4.6 (2.3–6.9)

Figure 2.

Forest plots for primary and secondary outcomes. DBP = diastolic blood pressure, MAP = mean arterial pressure, SBP = systolic blood pressure.

Secondary Outcomes

Fifteen articles provided information regarding gradients of SBP or data that facilitated the estimation of these gradients. The pooled mean SBP gradient was 8.0 mm Hg (95% CI, 3.9–12.0 mm Hg), as summarized in Table 2. This analysis showed low heterogeneity (I² = 8.5%). The forest plot for this is depicted in Figure 2.

Ten articles provided information regarding gradients of DBP or data that facilitated the estimation of these gradients. The pooled mean DBP gradient was 1.1 mm Hg (95% CI, –0.1 to 2.4 mm Hg), as summarized in Table 2. Heterogeneity was low (I² = 4.4%). The forest plot is also illustrated in Figure 2.

Subgroup Analysis

A subgroup analysis was conducted on the MAP gradient across various patient groups, including those undergoing cardiac surgery, liver transplant surgery, and nonsurgical critically ill patients.

In patients who underwent cardiac surgery, data were collated from nine studies for the pre-bypass stage, four for the on-bypass stage, and nine for the off-bypass stage. The mean MAP gradient with 95% CI was 1.8 mm Hg (0.8–2.8 mm Hg) at the pre-pump stage, 5.1 mm Hg (3.5–6.6 mm Hg) at the on-pump stage, and 2.8 mm Hg (0.3–5.3 mm Hg) at the off-pump stage, respectively.

In patients who underwent liver transplants, data were collated from eight studies for the post-induction stage, seven for the reperfusion stage, and six for the post-surgery stage. The analysis revealed a mean MAP gradient with a 95% CI was 3.4 mm Hg (0.4–6.4 mm Hg) post-induction, 5.0 mm Hg (0.9–9.1 mm Hg) during reperfusion, and 3.0 mm Hg (0.9–5.1 mm Hg) post-operation, respectively.

In nonsurgical critically ill patients, the mean MAP gradient with 95% CI was 4.6 mm Hg (2.3–6.9 mm Hg), based on the data pooled from five studies, as summarized in Table 2.

Sensitivity Analysis

Sensitivity analysis, which included all studies, irrespective of the sample size being fewer than 20, the pooled mean gradients were found to be as follows: the MAP gradient was 3.7 mm Hg (95% CI, 2.4–4.9 mm Hg) with no observed heterogeneity (I² = 0%), the SBP gradient was 8.7 mm Hg (95% CI, 5.1–12.4 mm Hg), showing substantial heterogeneity (I² = 60%), and the pooled mean DBP gradient was 1.4 mm Hg (95% CI, 0.3–2.5 mm Hg), with no heterogeneity (I² = 0%). The results are summarized in Supplemental Figure 1 (http://links.lww.com/CCX/B343).

The second sensitivity analysis involved sequentially removing one study at a time to evaluate its influence on the gradients of MAP, SBP, and DBP. The results are illustrated in Supplemental Figure 2 (http://links.lww.com/CCX/B343).

The third sensitivity analysis targeted studies where the publication provided the mean and sd for the gradients of MAP, SBP, and DBP. This particular analysis yielded a MAP gradient of 5.2 mm Hg (95% CI, 1.8–8.5 mm Hg) with an I² of 99.7% across eight studies. The SBP gradient was identified as 14.3 mm Hg (95% CI, 2.1–26.6 mm Hg) with an I² of 99.8% across six studies. For DBP, a gradient of 1.7 mm Hg (95% CI, 0.2–3.4 mm Hg) was observed with an I² of 98.8% across three studies. The results are illustrated in Supplemental Figure 3 (http://links.lww.com/CCX/B343).

Publication Biases

We detected no evidence of publication bias when assessing funnel plots visually as in Figure 3. We also statistically assessed publication bias using Begg rank correlation test (p = 0.06).

Figure 3.

Funnel plots for the primary outcome.

Risk of Bias of Included Studies

The quality of the included 24 studies was described using QUADAS-2 tool, as shown in Supplementary Figure 4 (http://links.lww.com/CCX/B343).

DISCUSSION

In this systematic review and meta-analysis, we demonstrated that central MAP was significantly higher than peripheral MAP with a mean gradient of 3.5 mm Hg. Furthermore, we found that central SBP was also significantly higher than peripheral SBP with a mean gradient of 8.0 mm Hg, whereas there was no statistically significant difference between central and peripheral DBP in high-risk surgical and critically ill patients. The subgroup analysis further revealed a higher MAP gradient over time with central catheters during the on-pump stage of cardiac surgery, reperfusion stage of liver transplant, and in nonsurgical critically ill patients. Our sensitivity analysis corroborated the overall findings. To the best of our knowledge, this is the first meta-analysis collating the existing evidence regarding the difference in arterial blood pressure based on the cannulation site. This new understanding of the gradient and the relationship between central and peripheral arterial pressures has important implications for the management of critically ill patients.

In healthy patients, it is known that SBP is higher in peripheral arteries than in central arteries (42–44), while MAP and DBP are relatively constant. However, in critical care settings such as during the on-bypass stage of cardiac surgery, the reperfusion stage of liver transplant, and nonsurgical critically ill conditions, our meta-analysis reported the reversal in the gradient in SBP, characterized by a higher central and lower peripheral pressure. Even though consensus regarding the underlying pathophysiology of this reversal remains elusive, previous studies have identified potential risk factors such as small arterial diameter caused by vasoactive medications, endogenous plasma norepinephrine levels, decreased vascular elasticity associated with aging, and female sex, and smaller stature, among others (45). Blood flow to peripheral regions diminishes in the setting of small arterial diameter due to multifactorial causes, leading to distinctly different blood circulation in peripheral arteries compared with the central artery. Despite these changes, the diastolic pressure can remain relatively unchanged between the peripheral and central arteries, as it is primarily determined by the resistance, rather than flow (46), which is less likely to be affected by the location of the arteries even in these critically ill conditions.

Another important aspect is the phenomenon of wave reflection back to the central compartment and its role in amplifying the disparity between central and peripheral arterial pressures, particularly under conditions of hypovolemia and vasoconstriction. As highlighted by Dorman et al (39), this phenomenon can result in significantly lower blood pressure in peripheral system impacting clinical management and hence outcomes during shock, reversing with resuscitation and recovery. Although relatively rare, failure to recognize these patients can lead to opposing treatment strategies.

Our meta-analysis highlighted that heterogeneity was significantly present in the sensitivity analysis for SBP, suggesting that SBP gradients may exhibit substantial variability across different patient populations and clinical settings. This variation underscores the need for more research to identify specific factors contributing to this variability in SBP. In contrast, for MAP and DBP, the heterogeneity was either absent or very small, indicating a notable consistency in these two values in variable clinical settings.

This study should be interpreted within the context of its limitations. First, the current study could not examine the potential influence of vasopressor use on the central and peripheral arterial pressure gradients. This is an important limitation given that vasopressors are commonly used in the management of critically ill patients to maintain adequate blood pressure and organ perfusion. Vasopressors can induce vasoconstriction, which might differently influence central and peripheral pressures, potentially introducing a confounding factor to our observed gradients (39). However, due to the limited number of studies that reported vasopressor usage and controlled for its effect, we could not perform a subgroup analysis for this population. Second, our findings may not be applicable to all critically ill patients as the included studies varied in their settings and patient populations. The effects of various diseases and treatments on the pressure gradient are not clear, and our findings may not be generalizable to all patient and clinical contexts. Third, the mean gradient between central and peripheral MAP was relatively small, and the clinical implication of this mean gradient remains unclear. However, when clinicians discuss targeted MAP values, it may be beneficial to specify the MAP target based on whether it pertains to central or peripheral measurements. This study should prompt further discussion aimed at enhancing the accuracy of invasive blood pressure monitoring. Last, arterial catheters provide not only pressure but also more information by waveforms and can imply perfusion by area under the curve and offer additional insight (47, 48). Finally, the intraindividual correlations used to predict pressure gradients in studies where this was not provided were inferred from Lee et al (15), which may not be representative of the wider population. This potentially introduces further errors into our calculations and may affect the validity of our findings.

CONCLUSIONS

This study finds significant SBP and MAP gradients between central and peripheral arterial pressures in critically ill patients. The clinical impact of these differences, with central pressures higher by 8.0 mm Hg for SBP and 3.5 mm Hg for MAP, is yet unclear, suggesting no immediate need to alter current preferences for peripheral cannulation. Further work is needed to better understand the factors influencing these gradients and to develop strategies to improve the accuracy of invasive blood pressure monitoring and reporting for documentation, patient safety, and research.