Introduction

Pediatric heart surgery is a vital therapeutic option for congenital heart disease, which is one of the most prevalent causes of death in children [1]. In pediatric cardiac surgery, arterial cannulation (AC) is required for close hemodynamic monitoring and frequent sampling for arterial blood gases (ABG). In addition, a central venous catheter (CVC) is needed for continuous monitoring of the central venous pressure (CVP) and replacement of fluid or blood products [2, 3].

The first choice for AC is the radial artery due to its easy access and fewer complications in children [4, 5]. Nevertheless, the use of radial artery cannulation has been linked to vasospasm. Therefore, a parallel trend has been observed in pediatric patients, where femoral artery catheterization has been associated with enhanced safety standards and accessibility [6,7,8]. Additionally, previous studies on pediatrics have predominantly focused on catheter line insertions through the internal jugular vein, which was associated with a high success rate and lower complications due to its wide diameter [9].

Traditionally, catheterization is performed using external landmarks and palpation techniques. However, using ultrasound (US) guidance during the femoral artery catheterization proved safer with strong supporting evidence [10]. Additionally, it is the standard of care for central venous catheterization [11,12,13]. Despite evidence-based recommendations, the adoption of US guidance remains limited, and its endorsement of palpation techniques, particularly in pediatric cases, needs more unequivocal supporting evidence [11, 14, 15]. Moreover, previous research in pediatrics has predominantly focused on catheter line insertions through the internal jugular vein, rendering comparative evidence for femoral artery catheterization techniques [11,12,13,14,15,16].

Recent investigations have aimed to fill this gap by comparing the success rates and complications associated with US-guided versus palpation-guided techniques in arterial and central venous catheterization in pediatric patients undergoing cardiac surgery. Hence, we conducted this systematic review and meta-analysis to synthesize the available evidence on the safety and efficacy of US-guided versus palpation-guided AC or CVC insertion in pediatric patients undergoing cardiac surgery.

Methodology

Protocol Registration

When reporting this systematic review and meta-analysis, we followed the preferred reporting items of systematic reviews and meta-analysis (PRISMA) statement guidelines [17]. We followed the Cochrane Handbook of Systematic Reviews of Interventions [18]. The protocol for this meta-analysis has been registered and published in PROSPERO with the following ID: CRD42024528227. The PRISMA checklist is demonstrated in Table S3.

Data Sources and Search Strategy

We have established a comprehensive search in various databases such as PubMed, EMBASE, Cochrane (CENTRAL), Scopus, and Web of Science Core Collection, that was systematically approached until the 21st of February 2024, using relevant search terms and keywords, as demonstrated in Table S1.

Eligibility Criteria

We included randomized controlled trials (RCTs) reported in English that fulfilled the following PICO criteria:

  1. 1.

    Population: children undergoing catheterization (cannulation) for cardiac surgery.

  2. 2.

    Intervention: US-guided technique in vascular access to the targeted vessel.

  3. 3.

    Comparison: palpation (landmark) guided technique.

  4. 4.

    Outcomes: our primary outcome was the success rate of cannulation. Secondary outcomes included duration of attempt, number of attempts and used cannula, any complications, artery or vein puncture defined as “An unintentional puncture or perforation of the wall of a blood vessel (artery or vein) during an attempt to cannulate (insert a needle or catheter into) the vessel.”, surgical cutdown, puncture failure, failure to pass the wire, and safety outcomes (mortality and occurrence of any adverse events).

We excluded the following types of articles: (I) studies lacking a comparison group, (II) those containing unreliable, non-extractable, duplicated, or overlapped data sets, (III) articles with unavailable full texts, (IV) conference posters/abstracts, case reports/series, review articles, and protocols of clinical trials with unpublished results.

Study Selection

This review was achieved using Covidence online software. The obtained studies were independently screened by (A.M., A.A., A.W.H., and M.S.R.) in two phases. The first phase was title/abstract screening for potential clinical studies on Covidence. In the second phase, we retrieved the full-text articles of the selected abstract for further eligibility screening using separate Google sheets. Any conflicts have been resolved by consensus and discussion.

Data Extraction

Data were extracted by at least two authors of (A.M., A.A., A.W.H., and M.S.R.), using separate Google sheets under three main domains: firstly, the summary characteristics (name of the first author, year of publication, study design, number of centers, country, total participants, venous or arterial canulation, name of the canulated vessel, interventional details, anesthesia used, main inclusion criteria, and primary outcome). The baseline information of the targeted population (sample size, age, weight, height, gender, heart rate, SBP, DBP, and vessel diameter). Eventually, outcome data, as previously illustrated, were included in the third section.

Risk of Bias and Certainty of Evidence

The quality assessment of studies was independently conducted using the Cochrane RoB2 tool [19] by (A.M., A.A., A.W.H., and M.S.R.). Moreover, they evaluated five domains, including deviation from the intended intervention, the risk of bias linked to the randomization process, outcomes measuring, missing outcome information, and choosing the reported outcomes and results. To evaluate the certainty of the evidence, the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) was utilized [20, 21] by (M.A., B.A.). Any conflicts have been resolved by consensus and discussion.

Statistical Analysis

For the statistical analysis, R version 4.3 was utilized using meta, metafor, and dmetar packages. Using the random-effects model, we pooled the results of dichotomous outcomes using the risk ratio (RR) and the continuous outcomes using the mean difference (MD), both with a 95% confidence interval (CI). We employed the Chi-square and I-square tests to evaluate heterogeneity; the Chi-square test determines whether heterogeneity exists, while the I-square test evaluates its degree. We considered an alpha level below 0.1 for the Chi-square test to denote significant heterogeneity.

We used both influence analysis and the brute force approach to identify the outlier for the sensitivity analysis. Additionally, our study's heterogeneity patterns were assessed using the Baujat plot. The Baujat plot's (y-axis) displays each effect size’s influence on the pooled result, while the x-axis displays each effect size’s total heterogeneity contribution. Studies or effect sizes that have high values on both the x and y axes could be regarded as influential cases; studies or effect sizes that have a high contribution to heterogeneity (x-axis) but little effect on the overall results could be regarded as outliers and could be eliminated to reduce the amount of heterogeneity between studies.

Results

Search Results and Study Selection

Figure 1 illustrates the process of screening literature. Initially, 295 relevant studies were found by searching databases. After duplicate entries were removed, titles and abstracts were screened to exclude 116 irrelevant articles, and the full texts of the remaining 30 articles were viewed. After the full-text screening, 13 RCTs were included [15, 22,23,24,25,26,27,28,29,30,31,32,33].

Fig. 1
figure 1

PRISMA flow chart of the screening process

Full size image

Characteristics of the Included Studies

A total of 13 RCTs involving 1060 patients fulfilled the inclusion criteria, of which 537 were allocated to the US group and 553 patients to the Palpation Group. These studies varied in the sample size from 40 to 201 and included children who were cannulated with either arterial or venous approach. Tables 1, 2 summarise the main features of the included literature and the baseline characteristics of the included population.

Table 1 Summary of the included RCTs
Full size table
Table 2 Baseline characters of the included population
Full size table

Risk of Bias and Certainty of Evidence

Among the 13 studies reviewed, eight were categorized as having low risk regarding the five domains [15, 22, 23, 25, 26, 28, 29], whereas the remaining studies were assessed to have some concern of selection bias [24, 27, 30,31,32,33]. The risk of bias for each of the included studies is shown in Fig. 2. Certainty of evidence is demonstrated in a GRADE evidence profile (Table 3).

Fig. 2
figure 2

Quality assessment of risk of bias in the included trials. The upper panel presents a schematic representation of risks (low = green, unclear = yellow, and high = red) for specific types of biases of each of the studies in the review. The lower panel presents risks (low = green, unclear = yellow, and high = red) for the subtypes of biases of the combination of studies included in this review

Full size image
Table 3 GRADE certainty of evidence assessment
Full size table

Primary Outcomes: Successful Cannulation

The US-guided technique significantly increased the successful cannulation in arterial cannulation [RR: 1.31 with 95% CI (1.10, 1.56), P < 0.0001]; however, there was no significant difference between both groups in venous cannulation [RR: 1.13 with 95% CI (0.98, 1.30), P = 0.10] (Fig. 3A). Moreover, the US-guided technique significantly increased the first-attempt success in arterial cannulation [RR: 1.88 with 95% CI (1.35, 2.63), P < 0.0001]; however, there was no significant difference between the two groups in venous cannulation [RR: 1.53 with 95% CI (0.86, 2.71), P = 0.15] (Fig. 3B).

Fig. 3
figure 3

Forest plot of the primary efficacy outcomes, RR risk ratio, CI confidence interval

Full size image

Pooled studies were heterogeneous in both outcomes (P < 0.001, I2 = 74%) and (P < 0.001, I2 = 78%), respectively. Sensitivity analysis was not applicable in both outcomes (Figs. S1–S4). We performed a meta-regression analysis against successful cannulation based on known baseline characteristics such as age (m) and weight (kg), with no obvious impact on the effect size (Table S2). A funnel plot was used in successful cannulation to detect possible publication bias. We found significant asymmetry by inspection, indicating significant publication bias (Egger’s P value = 0.029) (Fig. S5). Moreover, the trim and fill method was employed to address this, as shown in Fig. S6. Finally, the test of subgroup analysis was insignificant in successful cannulation either based on the vessel cannulated (arterial vs. venous) or the artery cannulated (femoral vs. radial) (P = 0.20), (P = 0.93) respectively (Fig. S7), and in first-attempt success based on the vessel cannulated (arterial vs. venous) (P = 0.54).

Secondary Outcomes

Efficacy Outcomes

The US-guided technique significantly decreased the number of attempts either in arterial cannulation [MD: − 0.73 with 95% CI (− 1.00, − 0.46), P < 0.0001] or in venous cannulation [MD: − 1.34 with 95% CI (− 2.55, − 0.12), P = 0.03] (Fig. 4A). However, there was no significant difference between both groups in arterial cannulation regarding the number of cannulas used [RR: − 0.31 with 95% CI (− 0.68, 0.05), P = 0.09] (Fig. 4B). Pooled studies were heterogeneous in both outcomes (P < 0.001, I2 = 71%) and (P < 0.001, I2 = 87%), respectively.

Fig. 4
figure 4

Forest plot of the secondary efficacy outcomes, RR risk ratio, CI confidence interval

Full size image

Regarding the number of attempts, heterogeneity was best resolved after omitting the study by Law et al. (2014) (I2 = 43%) (Aouad et al. (2010) and Law et al. (2014) were detected as possible outliers [random‐effect model]) (Figs. S8–S9). Results with these outliers removed are shown in Fig. S10. Moreover, in the number of cannulas used, sensitivity analysis revealed that the heterogeneity was best resolved after omitting the study by Siddik-Sayyid et al. (2016) (I2 = 0%) (no outliers detected [random effect model]) (Figs. S11–S12).

We performed a meta-regression analysis against the number of attempts based on known baseline characteristics such as age (m) and weight (kg), with a noticeable impact on the effect size with [P < 0.0001], [P < 0.0001] respectively (Table S2). A bubble plot of meta-regression is shown in Figs. S13–S14.

By inspection of funnel plot, we found significant asymmetry, indicating that there was significant publication bias (Egger’s P value = 0.005) (Fig. S15); the trim and fill method was employed to address this as shown in Fig. S16. Finally, the subgroup analysis test was insignificant in the number of attempts, either based on the vessel cannulated (arterial vs. venous) or the artery cannulated (femoral vs. radial) (P = 0.34), and (P = 0.80), respectively (Fig. S17).

Procedural Outcome: Time of Attempted Cannulation

The US-guided technique was associated with a significantly decreased the time of attempted cannulation either in arterial cannulation [MD: − 2.27 with 95% CI (− 3.38, − 1.16), P < 0.0001] or in venous cannulation [MD: − 4.13 with 95% CI (− 7.06, − 1.19), P < 0.0001] (Fig. 4C). Pooled studies were heterogeneous (P < 0.001, I2 = 82%). Sensitivity analysis was not applicable (Sadeghi et al. [29] and Verghese et al. [32] were detected as possible outliers [random effect model]) (Figs. S18–S19). Results with these outliers removed are shown in Fig. S20.

We performed a meta-regression analysis against the time of attempted cannulation based on known baseline characteristics such as age (m) and weight (kg), with no obvious impact on the effect size (Table S2). By inspection of funnel plot, we found significant asymmetry, indicating that there was significant publication bias (Egger’s P value = 0.001) (Fig. S21); moreover, trim and fill method was employed to address this as shown in Fig. S22. Finally, the subgroup analysis test was insignificant in the number of attempts based on the vessel cannulated (arterial vs. venous) or the artery cannulated (femoral vs. radial) (P = 0.25) and (P = 0.77), respectively (Fig. S23).

Safety Outcomes

Failure Causes

The US-guided technique significantly decreased the incidence of failure to pass a guide wire either in arterial cannulation [RR: − 0.20 with 95% CI (0.05, 0.86), P = 0.03] or in venous cannulation [RR: 0.24 with 95% CI (0.08, 0.71), P < 0.001] (Fig. S24). However, there was no significant difference between both groups in arterial cannulation regarding the incidence of failure to puncture the vessel with [RR: 0.20 with 95% CI (0.01, 3.91), P = 0.31] (Fig. S25). Pooled studies were heterogeneous in both outcomes with (P = 0.74, I2 = 0%), (P = 1.00, I2 = 0%) respectively.

Complications

The US-guided technique significantly decreased the incidence of any complications either in arterial cannulation [RR: 0.36 with 95% CI (0.18, 0.71), P < 0.001] or in venous cannulation [RR: 0.30 with 95% CI (0.11, 0.84), P = 0.02] (Fig. 5A). Furthermore, the US-guided technique significantly decreased in the incidence of hematoma formation in arterial cannulation [RR: 0.26 with 95% CI (0.12, 0.57), P < 0.001]; however, there was no significant difference between both groups in venous cannulation [RR: 0.82 with 95% CI (0.34, 1.94), P = 0.64] (Fig. 5B). Moreover, the US-guided technique significantly decreased the incidence of vessel puncture in arterial cannulation [RR: 0.22 with 95% CI (0.08, 0.63), P < 0.001]; however, it significantly increased the incidence of vessel puncture in venous cannulation with [RR: 1.94 with 95% CI (1.09, 3.47), P = 0.03] (Fig. 5C). Moreover, there was no significant difference between both groups in surgical cutdown in arterial cannulation [RR: 0.43 with 95% CI (0.07, 2.75), P = 0.37] (Fig. S26).

Fig. 5
figure 5

Forest plot of the safety outcomes, RR risk ratio, CI confidence interval

Full size image

Pooled studies were homogenous in any complications (P = 0.31, I2 = 16%), hematoma formation (P = 0.15, I2 = 36%), vessel puncture (P = 0.05, I2 = 47%), and in surgical cutdown (P = 0.84, I2 = 0%). Regarding any complication in venous cannulation (I2 = 63%), sensitivity analysis revealed that the heterogeneity was best resolved after omitting the study by Grebenik et al. (2004) (I2 = 0%) (no outliers detected [random effect model]) (Fig. S27).

By inspection of funnel plot of vessel puncture, we did not find significant asymmetry, indicating that there was no significant publication bias (Egger’s P value = 0.88) and (Egger’s P value = 0.80), respectively (Figs. S28–S29). Finally, the test of subgroup analysis was not significant across all the outcomes (P > 0.1) (Figs. S30–S31), except for vessel puncture, the test of subgroup analysis was significant based on the vessel cannulated (arterial vs. venous) (P < 0.001) with arterial cannulation being associated with a lower incidence of vessel puncture [RR: 0.22 with 95% CI (0.08, 0.63), P < 0.001] compared to venous cannulation [RR: 1.94 with 95% CI (1.09, 3.47), P = 0.03].

Discussion

Our meta-analysis on ultrasound-guided versus palpation-guided techniques for vascular access in pediatric cardiac surgery revealed that ultrasound-guided technique significantly increases the incidence of successful cannulation and first-attempt success compared to palpation-guided methods. Additionally, ultrasound guidance significantly reduced the number of attempts and procedural time required for arterial and venous cannulation. Furthermore, it was associated with decreased complications and procedure failure. Despite these advantages, ultrasound-guided venous cannulation was associated with a higher incidence of vessel puncture.

Our findings align with several studies that have reported similar results, indicating the superiority of ultrasound guidance in enhancing vascular access success rates [34,35,36,37,38,39]. Additionally, a meta-analysis on adult populations undergoing central venous catheterization has consistently shown improved success rates and reduced complication rates with ultrasound-guided techniques compared to palpation-guided methods [40]. Furthermore, an updated meta-analysis conducted by Gao et al. [41] and included both adults and pediatric populations demonstrated the benefit of the ultrasound-guided technique over the palpation technique. The extrapolation of these findings to pediatric populations, particularly in the context of cardiac surgery, underscores the generalized applicability of ultrasound guidance in optimizing procedural outcomes.

The suggested mechanisms underlying the superior success rates associated with ultrasound-guided techniques in the literature can be attributed to ultrasound providing real-time visualization of vascular structures, enabling clinicians to accurately identify vessel anatomy, size, and depth, thereby facilitating precise needle placement. This direct visualization minimizes the risk of inadvertent punctures, reducing procedural complications and optimizing cannulation success [42, 43].

The number and time of attempts during vascular access procedures are crucial metrics that reflect procedural efficiency. Our findings indicate that ultrasound-guided techniques significantly reduce both the number of attempts and the time required for vascular access compared to palpation-guided methods. This aligns with previous, where ultrasound guidance has consistently demonstrated advantages in procedural efficiency and success rates [44]. Like successful canulation, decreased time and number of attempts in US-guided procedures could be attributed to real-time visualization of vascular structures, allowing precise needle placement and minimizing the need for multiple insertion attempts.

Regarding the incidence of failure and complications, our findings indicate that ultrasound-guided techniques are associated with a significant decrease in the incidence of failure and complications such as hematoma formation and vessel puncture, particularly in arterial cannulation. This aligns with previous meta-analyses demonstrating the benefits of ultrasound guidance in reducing the incidence of complications [44, 45]. In contrast to our findings regarding hematoma, the previous meta-analysis conducted by Pacha et al. [40] on the adult population demonstrated no significant difference between both techniques regarding the incidence of hematoma. Nevertheless, our study's subgroup analysis revealed differences based on factors such as the type of vessel cannulated, and the specific artery targeted. For example, ultrasound-guided venous cannulation was associated with a higher incidence of vessel puncture than palpation-guided methods, suggesting the need for cautious consideration of procedural techniques and anatomical factors.

Strength and Limitations

Our study was the first meta-analysis conducted to compare US and palpation-guided techniques in vascular access in children undergoing cardiac surgery, encompassing a comprehensive literature search across multiple databases, ensuring the identification of relevant RCTs. The study further conducted subgroup analyses to explore outcome variations based on factors such as the type of vessel cannulated and the specific artery targeted, thus providing granularity to the analysis. Furthermore, sensitivity analyses and meta-regression were performed to assess the robustness of the findings and evaluate the impact of individual studies and baseline characteristics on the overall results, enhancing the reliability and validity of the meta-analysis. However, we faced some limitations. The included studies exhibited heterogeneity, which may have influenced the robustness of the findings despite attempts to account for heterogeneity through subgroup and sensitivity analyses. Additionally, publication bias cannot be ruled out, as studies with positive results may be more likely to be published, potentially leading to overestimating the effect sizes associated with ultrasound-guided techniques.

Clinical Implications

Our results underscore the importance of incorporating ultrasound guidance into clinical practice to improve procedural efficiency and patient outcomes. To maximize the benefits of this approach in pediatric cardiac surgery settings, healthcare providers should consider adopting standardized protocols for ultrasound-guided techniques and ensuring adequate training.

Conclusion

Ultrasound guidance improves successful cannulation rates and first-attempt success in arterial access while reducing the number of attempts and procedural time for both arterial and venous access. It was also associated with a lower incidence of complications and procedure failure, particularly in the arterial setting. However, it was associated with a higher incidence of venous puncture.