Bedside ultrasonography prior to abdominal paracentesis is associated with low complication and high success rate: Experience in a National Health Service District General Hospital in the United Kingdom from 2013 to 2019

Abstract

Aims/Background:

To analyse data from a bedside ultrasound-assisted ascites procedure service in a National Health Service District General Hospital and compare them to results of studies in the medical literature.

Methods:

A retrospective review of audit data collected (January 2013 to December 2019) of the practice of paracentesis in a National Health Service District General hospital. All adult patients referred to the ascites assessment service were included. Bedside ultrasound detected location and volume of ascites, if present. Abdominal wall diameters were determined in order to select the appropriate needle length for procedures. Results and scan images were recorded on a pro-forma. Patients who underwent a procedure were followed up for 7 days, with complications documented.

Results:

Seven hundred and two scans were performed on 282 patients – 127 (45%) male and 155 (55%) female. In 127 (18%) patients, an intervention was avoided. Five hundred forty-five (78%) patients underwent a procedure: 82 (15%) were diagnostic aspirations and 463 (85%) were therapeutic (large volume) paracentesis. Most scans were performed between 08:00–17:00. Average time from patient assessment to diagnostic aspiration was 4 hours 21 minutes. Complications included three failed procedures (0.6%) and one iatrogenic peritonitis (0.2%), but no bowel perforation, no major haemorrhage or death.

Conclusion:

It is possible to introduce a bedside ultrasound-assisted ascites procedure service to a National Health Service District General Hospital with a high success and low complication rate.

Keywords

Ascites abdominal wall ultrasound patient safety guideline advanced clinical practitioner

Introduction

Abdominal paracentesis is an invasive procedure performed frequently in secondary care for diagnostic and therapeutic indications.¹ Traditionally, combinations of clinical detection techniques and anatomical landmark identification have been used to identify a point for needle insertion.^2–4 Complications are rare and include bowel perforation, haemorrhage and death.^5–13 Ultrasound-assisted paracentesis was first recommended in 1986.¹⁴ Subsequently, comparative studies have demonstrated a better safety profile, reduced length of stay and lower hospitalisation costs, when using ultrasound.^15–19 Many articles support the use of ultrasound prior or during abdominal paracentesis for detection of ascites to reduce failed paracentesis attempts and identify the safest site for needle insertion.^{15–17,20–22} Furthermore, in several countries, guidelines supporting the use of ultrasound-assisted abdominal paracentesis have been published.^23–26

The authors present their 7-year experience of using bedside ultrasound prior to diagnostic and therapeutic abdominal paracentesis, on hospital wards and an Ambulatory Care Unit (ACU) and compare the results with previous studies.

Methods

This is a retrospective review of anonymised data collected (January 2013 to December 2019) for an internal audit of the practice of paracentesis in a National Health Service (NHS) District General hospital, serving a community of around 200,000 people. The data analysed were acquired as part of a local service evaluation, and therefore, no ethical approval was sought. Written consent was obtained from all patients undergoing therapeutic paracentesis. Written consent for publication for the images in Figures 1 –4 was also obtained.

Figure 1.

Patient with a distended abdomen and sufficient ascites to perform a procedure on ultrasound assessment.

Figure 2.

Patient with a distended abdomen and sufficient ascites to attempt a procedure on only one side of the abdomen.

Figure 3.

Ultrasound finding of a large cystic mass – consequently, a paracentesis procedure was not indicated.

Figure 4.

Abdominal wall measurement.

Included in the review were all data sheets (n = 702) collected by the doctors and Advanced Clinical Practitioners (ACPs), when they performed an assessment for an abdominal paracentesis procedure on patients’ referred to the trust’s ‘Ascites assessment and procedure service’. Accepted were referrals for non-pregnant adults (>18 years of age) with suspected ascites made by General Practitioners, doctors of all grades across inpatient wards, outpatient clinics and the Emergency Department. Procedures were performed on the Acute Assessment Unit (AAU), inpatient wards and on the ACU.

Ultrasounds and procedures were performed by medical doctors and ACPs. All had received training fulfilling the requirements of the Royal College of Radiology ‘Focused Ultrasound Training Standards’ guidelines (2012) for performing the ultrasound and the drain placements.²⁷ The number of accredited bedside ultrasound practitioners increased from one practitioner in 2013 to eight in 2019, of which four were ACP. In addition, specialist trainees in gastroenterology assisted the procedures as well, who were ‘Emergency Ultrasound level one’ accredited. The team of operators was working mainly from 08:00 to 17:00 on weekdays, but frequently at least one operator was available out of hours beyond this on Ambulatory Care until 21:00 and on weekends. However, there was no specific ‘on call’ rota. Referrals were made via a phone call to the Ambulatory Care coordinator, or contacting the service lead consultant (phone calls, email and direct person-to-person referral).

Ultrasound machines used were SonoAce PICO (Medison) and Antares Acuson X300 (Siemens), with 3.5 MHz curved linear probes. The portable ultrasound machines used were SonoAce PICO (Medison) and Antares Acuson X300 (Siemens), with 3.5 MHz curved linear probes. Both sides of the abdomen were scanned in each patient. Ascites volume, procedure performed, patient location, date and time were documented on a pro-forma. An ultrasound image of the optimal insertion site was stored, printed and attached to the data collection sheet.

Ascites was categorised as ‘large enough volume’, when a certain fluid filled distance (>10 mm) between abdominal wall and abdominal structures was detected, ‘trace only’, if this distance was small (<10 mm) or ‘none’. All operators made their own decisions regarding the safety of a procedure, but if in doubt, consulted with a senior clinician. Radiology consultants were not involved in the decision-making in daily practice.

Patients with sufficient ascites had the safest point of entry marked immediately prior to paracentesis. On three occasions, ultrasound was performed during paracentesis to guide the needle.

Therapeutic paracentesis was performed using a Bonanno^© catheter in all cases. Platelets below 50, and international normalised ratios (INRs) greater than 3.0 were corrected prior to procedures (local policy). In 2017, a formal pathway pro-forma was introduced in order to standardise the pre- and post-procedure safety checks, the documentation of the procedure itself, volume of ascites drained and immediate complications.

In order to select the appropriate needle length for diagnostic aspirations, formal abdominal wall measurements were obtained in the majority of patients. In the remaining patients, the abdominal wall diameter was estimated by looking at the centimetre scale on the scan monitor. Needles used for the aspirations were either 21G × 38 mm (green Monoject™ needle) or 18G × 45 mm (green Venflon™ cannula) according to the abdominal wall diameter measurements.

Patients undergoing a procedure were followed for 7 days to identify complications, by looking retrospectively at the medical records and digitalised laboratory results. Complications were defined as follows: significant haemorrhage (haemoperitoneum/haematoma formation necessitating a blood transfusion or an operation), solid organ injury, bowel perforation, peritonitis and death. Inability to obtain an aspirate, or insert a drain, was documented as procedure failure.

A further four parts of this study were done retrospectively:

Determining the time interval between the decision to perform a diagnostic aspirate and the procedure itself. The time of the decision to perform a procedure was taken from the documented time entry in the medical notes, by doctors of any grade. In order to establish, if during the study there was a change in this time interval, we compared, due to small numbers, the years 2013–2017 (n = 27) to the years 2018 and 2019 (n = 48).

To compare the percentage of procedures done with ultrasound assistance in the year 2013 and 2019, trust coding data were analysed to identify non-ultrasound-assisted procedures.

All ascitic fluid samples received in the laboratory for analysis in the years 2013 and 2019 were reviewed to further identify patients, not already flagged up by coding, who underwent non-ultrasound-assisted procedures.

The images of ultrasound scans which showed insufficient ascites to safely perform an intervention were collected and shown, prior to submitting this publication, to a senior interventional radiology consultant for review, to ascertain if potentially procedures could have been performed on those patients.

To compare complication rates with other studies, a literature search with PubMed.gov was performed, using the terms: paracentesis, ascites, aspirate, complication, haemorrhage, infection, perforation, peritonitis and procedure failure, summarised in Table 5. All studies which did not include bedside paracentesis were excluded.

From the review, we identified one study of similar size and patient mean age by De Gottardi et al.⁹ published in 2009 (Table 6). Not all items in the original table by De Gottardi were compared, since some complications are not dependent on the insertion technique, for example, ‘slipping of the catheter from the abdominal wall’. The data from both cohorts were analysed for odds ratios of procedural complications. The Fisher Exact test was also used to compare both cohorts for statistical significance.

Results

Two hundred eighty-two patients were referred on 702 occasions to the ascites assessment and procedure service. Bedside abdominal ultrasound was performed on all 702 occasions.

Table 1 summarises the patient demographics and underlying pathologies causing ascites or abdominal distension clinically mistaken for ascites. More female than male patients were referred (n = 155 and n = 127, respectively). The patients’ ages ranged from 25 to 94 years, with a median age of 65 years. Overall, the commonest cause of ultrasound-confirmed ascites was alcohol-related liver disease and underlying malignancy both in equal numbers (n = 117).

Table 1.

Patients’ demographics and underlying pathologies.

	Number of patients	Male	Female
	n (%)	n (%)	n (%)
Total scans	702 (100%)
Gender distribution of scans		338 (48%)	364 (52%)
Number of patients scanned	282 (100%)	127 (45%)	155 (55%)
Age range (years)	25–94	28–88	25–94
Age mean (years)	62	65	59
Age median (years)	65	67	60
Underlying pathology:
Non-cancer-related	165 (59%)
ALD	117 (71%)
Non-alcohol cirrhosis	32 (19%)
CCF	13 (8%)
Bowel gas distension	2 (1.2%)
Polyserositis	1 (0.8%)
Cancer-related	117 (41%)
Ovarian cancer	32 (27%)
Breast cancer	12 (10.2%)
Bowel cancer	12 (10.2%)
Pancreatic cancer	12 (10.2%)
Peritoneal cancer	7 (6.0%)
Cancer unknown primary	11 (9.5%)
Hepatocellular carcinoma	6 (5.5%)
Endometrial cancer	7 (6.0%)
Lymphoma/myeloma	4 (3.5%)
Cholangiocarcinoma	4 (3.5%)
Gastric/oesophageal	7 (6.0%)
Mesothelioma	2 (1.6%)
Bladder cancer	1 (0.8%)

ALD: alcohol-related liver disease; CCF: congestive cardiac failure.

Table 2 shows the location where the scans were performed, the findings during the ultrasound procedures and technical problems encountered. All patients were referred because sufficient ascites to warrant paracentesis was clinically suspected. Ascites was identified on ultrasound in 96% (n = 674) of patients. Nonetheless, a large enough volume of ascites, to warrant paracentesis, was only detected in 581 patients on ultrasound. This means that clinical detection techniques (e.g. anatomical landmark technique) for large volume ascites were overall only correct in 83% of patients.

Table 2.

Patient location, ascites volumes and technical problems.

	Number of patients	Male	Female
	n (%)	n (%)	n (%)
Total scans	702 (100%)
Location:
Hospital wards (including AAU)	389 (55%)
Ambulatory Care Unit	313 (45%)
Ascites detected	674 (96%)
Large volume ascites	581 (83%)	295 (51%)	286 (49%)
Trace ascites	93 (13%)	33 (35%)	60 (65%)
No ascites	28 (4%)	10 (36%)	18 (64%)
Procedures	545 (78%)	282 (52%)	263 (48%)
Ascitic drain insertions	463 (85%)
Successful first drain insertions	456 (98.5%)
Failed first drain insertions	7 (1.5%)
Fully unsuccessful drain insertions	2 (0.4%)
Aspirations (diagnostic)	82 (15%)
Successful first aspiration	79 (96.3%)
Failed first aspiration	3 (3.7%)
Fully unsuccessful diagnostic aspiration	1 (1.2%)
Total of failed procedures	3 (0.6%)
Intervention avoided	127 (18%)
None/trace ascites	124 (17%)
Abdominal wall >5 cm (with small/moderate ascites)	3 (0.4%)
Deferred/declined or other treatment	30 (4.3%)

AAU: Acute Assessment Unit.

In 545 (78%) patients, a procedure was performed: 82 (15%) had only a diagnostic aspiration and 463 (85%) underwent therapeutic paracentesis. In 20 patients, it was safe to proceed with paracentesis only on one side of the abdomen, because bowel loops extended to the abdominal wall on one side, posing the risk of a bowel perforation.

Examples of the discrepancies between clinical ‘detection’ of ascites and findings at ultrasound are demonstrated in Figures 1 –3. Regarding technical difficulties, in the therapeutic paracentesis cohort 98.7% (n = 458) of drain insertions were successful at the first attempt and only two (0.4%) did fail, despite a repeated attempt. Regarding diagnostic aspirations (n = 82), only three (3.7%) failed first pass and one (1.2%) at repeated attempts.

In one drain insertion attempt aspiration with a green needle did reveal ascitic fluid, but it was not possible to aspirate ascites with the drain needle and the procedure was abandoned. One further drain insertion and one attempted diagnostic aspiration failed, because, in both cases, a small bleed occurred that needed prolonged pressure applying and both patients then declined further procedure attempts. There were no major haemorrhages in either subgroup.

In 157 (23%) patients, a procedure was not performed. The reasons were insufficient ascites on ultrasound to perform a procedure (n = 124), patient refusal/deferred procedure (n = 30) and three patients had an abdominal wall diameters greater than 5 cm and were referred for computed tomography (CT) guided intervention.

An interventional consultant radiologist reviewed 124 scans, where our operators found insufficient ascites. It showed 31 patients had been referred for a diagnostic aspiration, 39 for an aspiration and (if possible) drain insertion and 54 for a drain insertion only. It concluded that in four patients (4.3%), a therapeutic paracentesis possibly could have been undertaken and regarding diagnostic aspirations, six (8.6%) out of 70 referred patients could possibly have had an aspiration attempted. The scans in question had been performed by multiple operators; no singular individual was identified as lacking in diagnostic accuracy. Formal abdominal wall measurements were obtained in 438 patients (80%) of the intervention group (Figure 4).

Table 3 summarises the timings and locations of 700 ultrasound scans, on two occasions the times were not recorded. On ACU, 97% (n = 304) and on the wards, 70% (n = 272) of the scans performed were done between 08:00 and 17:00. A total of 18% of scans (n = 124) were undertaken out of working hours (after 17:00; on weekends and bank holidays). The latest scan on ACU was at 21:10 and on the wards at 22:22; no scans were performed after midnight.

Table 3.

Timings and locations of ultrasound scans (n = 700)^a.

Table total (n = 700^a)	Ambulatory Unit (n = 312)			Out of ours total	Hospital wards (including AAU) (n = 388)			Out of hours total
Table total (n = 700^a)	Weekdays	Weekdays (out of hours)	WE/BH Friday >17:00	Out of ours total	Weekdays	Weekdays (out of hours)	WE/BH Friday >17:00	Out of hours total
	08:01–17:00	17:01–00:00			08:01–17:00	17:01–00:00
US scans total	n = 304	n = 1	n = 7	n = 8	n = 272	n = 55	n = 61	n = 116
US scans only	n = 53	n = 1	n = 5	n = 6	n = 69	n = 11	n = 17	n = 28
Procedures	n = 251	n = 0	n = 2	n = 2	n = 203	n = 44	n = 44	n = 88

AAU: Acute Assessment Unit; WE/BH: scans on weekends and bank holidays; US: ultrasound.

In two patients, time was not recorded: one on hospital wards and one in Ambulatory Care.

Table 4 displays the time interval from the decision to perform a diagnostic aspiration to intervention. In six procedures (7.4%), it was not possible to establish the time to aspiration. The time interval varied from 10 to 1410 minutes. The mean time to procedure was 4 hours 21 minutes.

Table 4.

Time to successful diagnostic ascites aspiration (n = 81)^a.

	Ambulatory Unit		Hospital wards		Total^a
	08:01–17:00	17:01–00:00	08:01–17:00	17:01–00:00
n=	6	0	51	18	75
Average time (minutes)	178	–	272	231	261
Minimum time (minutes)	20		10	10	10
Maximum time (minutes)	300		1410	660	1410

Unable to establish time for six procedures.

The time to aspiration in 2013–2017 was on average 3.5 hours and in 2018–2019 4.6 hours. This applied to both the whole cohort of patients (n = 75; successful aspirations, where timing was recorded) and the patients with decompensated liver cirrhosis (n = 58). Time to aspiration after 17:00 on weekdays and weekends/bank holidays (n = 19) increased on average to 5.5 hours and in six (32%) out of the nineteen patients time to aspiration was longer than 6 hours, the longest delay was 22 hours. No aspiration was performed beyond 24 hours from the time the decision was made.

Table 5 summarises the complications that occurred in our cohort, compared to other studies. We included data from nine articles.^{5,7,9,12,13,15,16–18} In these nine articles, data were found regarding ‘Procedure failure’ in three studies, regarding ‘Bowel perforation’ and ‘Peritonitis’ in five, respectively, and ‘Death’ in seven studies.^{5,7,9,12,13,15,16–18} Major or severe haemorrhage was defined in the literature as blood loss into the peritoneal cavity needing blood transfusions, abdominal wall haematoma or requiring hospitalisation of the patient and were documented in eight studies.^{5,7,9,12,13,15,17,18}

Table 5.

Paracentesis complication rate comparisons.

Authors	Year	Procedure technique	(n=)	Procedure failure % (n=)	Haemorrhage % (major) (n=)	Bowel perforation % (n=)	Peritonitis % (n=)	Death % (n=)
Droste JC et al. (This study)	2021	All US	482	0.6% (3)	0% (0)	0% (0)	0.2% (1)	0% (0)
Mallory and Schaefer⁵	1987	noUS	242	N/A	1.65% (4)	0.41% (1)	0.82% (2)	0.82% (2)
Tito et al.¹³	1990	noUS	38	N/A	5.2% (2)	0 (0%)	2.63% (1)	5.2% (2)
Webster et al.⁷	1995	noUS	179	0 (0%)	2.23% (4)	N/A	N/A	0 (0%)
Lin et al.¹¹	2015	noUS	602	N/A	2.9% (18)	N/A	N/A	0.3% (2)
De Gottardi et al.⁹	2009	88% noUS and 12% US (no comparison)	515	6.2% (32)	5 (1%)	0.4% (2)	0.2% (1)	0 (0%)
Nazeer et al.¹⁶	2005	Comparison	44 noUS and 56 US	39% (17) and 3.57% (2)	N/A	N/A	N/A	N/A
Patel et al.¹⁸	2011	Comparison	574 noUS and 723 US	N/A	0.87% (5) and 0% (0)	0 (0%) and 0 (0%)	2.44% (14) and 0.41% (3)	0 (0%) and 0 (0%)
Mercaldi and Lanes¹⁷	2013	Comparison	38,210 noUS and 31,649 US	N/A	1.25%(478) and 0.27% (87)	N/A	N/A	N/A
Sudulagunta et al.¹²	2015	Comparison	4089 noUS and 100 US	5.6% (69) and 0% (0)	2.8% (35) and 0% (0)	0% (0) and 0% (0)	2.6% (32) and 0%(0)	0 (0%) and 0 (0%)

N/A: not analysed; US: ultrasound; noUS: no ultrasound used.

The percentage ranges for the different complications were as follows: ‘Procedure failure’ 0.0%–6.2%; ‘Major haemorrhage’ 0.0%–5.2%; ‘Bowel perforation’ 0.0%–0.41%; ‘Peritonitis’ 0.0%–2.6%; ‘Death’ 0.0%–5.2%.^{5,7,9,12,13,15,16–18} Procedure failure attempts in this study were as described in the explanation for Table 3. There was one case of iatrogenic peritonitis (0.2% of all procedures), but no major haemorrhage, bowel perforation or death in our cohort.

Table 6 is a more in-depth comparison with the study by De Gottardi et al. It shows that ultrasound had been used in their study much less frequent (overall in 12%), compared to this study.⁹ In summary, there were less technical problems and complications in this study, which reached statistical significance regarding six compared criteria; regarding a further six criteria, the numbers were too small to identify a statistical significance.

Table 6.

Complication comparison to De Gottardi study (published 2009).

	DeGottardi et al.⁹	Droste et al.^a (this study)	OR	CI	p-value
Study period	2005–2007	2013–2019
Number of patients	171	193
Age mean (years)	59.6 years	63 years
Male	126 (73.7%)	89 (46.6%)
Female	45 (26.6%)	104 (53.4%)
Aetiology of ascites: liver disease	171 (100%)	113 (59%)
Total number of procedures	515 (100%)	545 (100%)
Diagnostic aspiration (total)Ultrasound used	45 (8.8%) and 20 (44.4%)	82 (15%) and 82 (100%)
Therapeutic aspiration (total)Ultrasound used	470 (91%) and 40 (15%)	463 (85%) and 463 (100%)
Technical problems	29 (5.6%)	10 (1.8%)	0.3132	0.1511–0.6495	0.002
No ascites at the first puncture attempt	32 (6.2%)	4 (0.7%)	0.1116	0.0392–0.3178	<0.001
Need for catheter repositioning	33 (6.4%)	3 (0.5%)	0.0721	0.0221–0.2354	<0.001
Need for a second puncture (all procedures)	50 (9.7%)	8 (1.5%)	0.1385	0.0650–0.3584	<0.001
Incomplete procedure	9 (1.7%)	3 (0.5%)	0.3112	0.0838–1.1560	0.081*
Need for a longer catheter	6 (1.2%)	0 (0.0%)	0.0718	0.0040–1.2787	0.073*
Slipping of the catheter from the abdominal wall	3 (0.6%)	N/A	–	–	–
Catheter residue broken into abdominal wall	1 (0.2%)	N/A	–	–	–
Small bowel perforation	2 (0.4%)	0 (0.0%)	0.1883	0.0090–3.9311	0.282*
Iatrogenic percutaneous infection of ascites	1 (0.2%)	1 (0.2%)	0.9449	0.0589–15.1463	0.968*
Ascites outflow at the puncture site after the procedure	26 (5.0%)	N/A	–	–	–
Bleeding after/during the procedure	17 (3.3%)	2 (0.4%)	0.1079	0.0248–0.4694	0.003
Minor	12 (2.3%)	2 (0.4%)	0.1544	0.0344–0.6932	0.015
Major	5 (1.0%)	0 (0.0%)	0.0851	0.0047–1.5425	0.100*
Death	2 (0.4%)	0 (0.0%)	0.1883	0.0090–3.9311	0.282*

N/A: not analysed, because not related to insertion technique.

Only patients included who had a procedure (=78% of patients scanned).

Numbers too small to identify statistical difference.

In 2013, 105 paracentesis procedures were coded, of which only 10 had been undertaken with ultrasound assistance. In 2019, 288 ultrasound-assisted procedures were undertaken, of which only 185 were coded. Further analysis of coding and laboratory data, compared with the medical records, revealed 26 procedures that were not in our ultrasound cohort. We conclude that 90% (n = 90 of 105) of procedures were done without ultrasound in 2013, falling to 8.2% (n = 26 of 314) in 2019.

Discussion

It is now 50 years since ultrasound was first used to confirm ascites in adult patients, and it can accurately detect volumes of ascites in quantities as little as 100 mL.²⁸ Despite physical examination techniques, such as shifting dullness to percussion and the presence of a fluid wave, having a reported accuracy in detecting ascites of between only 58% and 80%, the need for ultrasound prior to or during paracentesis is still debated between healthcare professionals.^{2–4,6,15,16,17,19,23,29} A pilot survey by the authors suggests that this is reflected in a marked variation in paracentesis practice between different NHS trusts: out of 33 NHS trusts, only 17 (51%) use ultrasound routinely prior to, or during diagnostic and therapeutic paracentesis.³⁰

Sufficient ascites for safe paracentesis in patients referred for an invasive procedure was identified with ultrasound in 83% (n = 581) of patients in this study (Table 2 and Figure 1). This is an equal percentage compared to a smaller study (n = 100) in 2005, which also found only 84% of patients enrolled having sufficient ascites to undergo an intervention.¹⁶ Illustrated in Figure 2 is the finding in a further 20 patients that ascites was accessible on only one side of the abdomen, due to prominent bowel loops on the opposite side. Ultrasound found insufficient ascites to safely perform an intervention in 124 (17%) patients, and in four of those patients, an alternative diagnosis was found, an example is demonstrated in Figure 3 (cystic mass). These four patients not only avoided the discomfort and potential complications of a failed procedure attempt, but ultrasound provided additional diagnostic information, allowing for prompt further investigations.

The British Society of Gastroenterology advocates diagnostic aspirates to be performed ideally within 6 hours and not more than 24 hours.²⁶ Patient outcomes are worse, when diagnostic aspirations are delayed, particularly beyond 24 hours.^31,32 In this study, the majority of ultrasound scans occurred within normal working hours (08:00–17:00), as summarised in Table 3. Scans took place ‘out of hours’ in 30% (n = 116) of the ward patients. That most of the inpatient scans were done ‘in hours’ on weekdays certainly does reflect the availability of ultrasound accredited practitioners within the trust. The overall mean time to aspiration was 4 hours 21 minutes, but out of hours, 1 hour longer on average, with no patient waiting over 24 hours for a procedure. The increase in ‘time to aspiration’, of about 1 hour, from the first four to the last 2 years of this study can be most likely be explained by the fact that the workload increased more than tenfold, from 22 ultrasound assessments in 2013 to 288 in 2019. This could not be entirely compensated by the increase from one to eight practitioners, over the time period documented.

Procedure failure is a clearly defined event, allowing direct comparison between studies. Failure to successfully perform paracentesis is a frequent problem encountered in up to 39% of attempts, when not using ultrasound (Table 5).^9,12,16 There are many reasons this occurs, including inadequate insertion technique, for example, not inserting the needle at a 90 degree angle to the skin, inadequate needle length, and finally due to the low accuracy of clinical examination in reliably identifying ascites.^12,16

As demonstrated in Tables 2, 5 and 6, in this study, we encountered three initial unsuccessful diagnostic aspiration attempts (3.7%), of which two were subsequently successful, leaving one procedure failed (1.2%). Of 463 therapeutic aspirations (ascitic drains), 98.7% (n = 457) were successful at first attempt, six needed a second attempt and two failed (0.4%), which compares very favourably to the failure rates of patient cohorts in studies using less frequent or no ultrasound (Table 5).^9,12,16

The review article of Ennis et al.²¹ shows visual examples of the usefulness of measuring abdominal wall diameters, also illustrated in Figure 4. Knowledge of the abdominal wall thickness allows the operator to modify their equipment accordingly and helps to reach deeper pockets of ascites by improving confidence to advance the needle further, when needed.²¹ Sakai et al.³³ note: ‘The evolving obesity epidemic has changed the physique of many patients with cirrhosis and ascites such that needles inserted into the abdominal wall may not reach fluid’. The authors believe that there is an important advantage of measuring abdominal wall diameters, and hypothesise that this has contributed to the low procedure failure rate observed by this study. Nonetheless, this would need to be statically proven in further research.

Table 5 shows data comparing paracentesis complication and failure rates between studies using in varying degrees ultrasound-assisted and physical examination techniques. The incidence of major haemorrhage related to abdominal paracentesis is reported as 0%–5.2%.^{5,7,9,12,13,15,17,18} In the largest comparison study of 69,859 patients undergoing paracentesis, the physical examination technique incurred over four times the number of haemorrhages, than that of the ultrasound-assisted technique (1.27% and 0.27%, respectively; Table 5).^12,16–18 However, it is not clear, whether they were classified as ‘minor’ of ‘major’. In our data collection, two minor haemorrhages (0.36%), but no major haemorrhage was documented.

Death and peritonitis are difficult to compare between studies, as these are not necessarily directly linked to the paracentesis technique used. There was one case of peritonitis in this study (0.18%), thought to be due to an iatrogenic percutaneous infection, but no bowel perforation or death. This is similar to other ultrasound-assisted paracentesis studies and supports the evidence for a good safety profile regarding ultrasound-assisted paracentesis (Table 5).^12,15–17

De Gottardi et al.⁹ conducted a study documenting complications related to diagnostic and therapeutic paracentesis, which we compared to this study in Table 6. The mean patient age, number of patients and total number of procedures are similar in both studies. Dissimilar between the studies is the male to female patient ratio, and the aetiology of ascites. The main difference regarding procedure technique is that ultrasound prior to paracentesis was used only in 12% by De Gottardi et al.⁹

In summary, Table 6 shows that the rate of technical problems/complications (with exception of iatrogenic infection) is three to ten times lower in our cohort and reached statistical significance regarding six criteria. We think this difference is likely due to the use of ultrasound prior to paracentesis in all our patients, compared to only 12% in the study of De Gottardi et al.⁹

Case reports of complications, despite using ultrasound, are rare, and should not lead to the conclusion that all ultrasound guided procedures need to take place in the radiology department.^34,35 To the contrary, our data, similar to other articles, support the use of bedside ultrasound-assisted paracentesis leading to a high success rate for diagnostic aspirations and drain insertions.^{12,14–17,19–23,35}

Regarding the analysis of procedure activity, unfortunately coding data were not reliably reflecting the activity in 2019. This is not necessarily surprising, since a lack of accurate procedure activity documentation by the coding process in NHS trusts has been demonstrated in other studies.^36,37 Therefore we attempted to obtain a more complete approximation of procedure activity by analysing laboratory data of ascitic samples in 2019. This combined approach, of looking at coding and laboratory data, revealed that one therapeutic and 25 diagnostic paracentesis procedures were performed without ultrasound assistance in our hospital in 2019. Even though this approach may not represent entirely the paracentesis activity, it still appears to reflect an overall very substantial increase of ultrasound-assisted, compared to non-ultrasound-assisted, procedures from 2013 to 2019 within the trust. This supports the notion that ultrasound-assisted paracentesis can satisfy the demand for most procedures, even in a medium-sized district general hospital.

Apart from the above inaccuracy of activity recording, this study has other limitations that need to be acknowledged. There is a potential sample bias arising from using a convenience sample. In addition, the sample is from a single site and includes a relatively limited cohort compared to some other publications.^12,17,18 Cost-effectiveness has not been addressed by this study, but other researchers have shown reduced hospitalisation costs for patients undergoing ultrasound-assisted paracentesis.^17,18 Some studies have linked this cost benefit to the reduction in complications incurred in by ultrasound-guided procedures, and this may warrant further investigation in an NHS-based study.^17,18

Albeit in the majority of patients (n = 438; 80%) abdominal wall measurements were formally taken, this was not performed prior to all procedures, mainly because on occasions, the operators felt ‘eyeballing’ the diameter was sufficient. Neither did we record the needle length used in each procedure, nor how often abdominal wall measurements led to an adjustment in needle length. This needs to be addressed in a future study.

Regarding the results of the scan reviewed by a radiologist, a small number of patients potentially could have had either a diagnostic or a therapeutic procedure performed, if a more experienced operator (e.g. radiologist) would have been involved. In the authors’ opinion, this potential disadvantage to a small number of patients is nonetheless far outweighed by the advantage of avoiding unnecessary and potentially unsafe procedures with ultrasound assistance in the vast majority of patients.

In hindsight such a review, as a quality control measure, should probably have been conducted earlier and in regular intervals during the data collection period. Furthermore, there was no direct comparison made between a cohort of ultrasound-assisted intervention versus physical examination techniques, but this has been investigated in another study by Nazeer et al.¹⁶

Paracentesis in the setting of oncology patients has yet to have guidance issued regarding the use of ultrasound. The authors welcome the recent recommendation by the British Society of Gastroenterology and the British Association for the Study of the Liver (BSG-BASL) to consider the use of ultrasound prior to paracentesis in patients with ascites due to liver cirrhosis, when available.³⁸

Nevertheless, we support the stronger recommendation of the American Society of Hospital Medicine to use ultrasound prior to all paracentesis, since we are convinced that this reduces the risk of serious complications, avoids attempting paracentesis with insufficient ascites and increases the overall procedure success rate.²³ This is indirectly supported by the extensive experience with other invasive procedures, for example, thoracentesis and central line insertion, where standardisation with ultrasound has been shown to contribute to improve patient safety.^39,40 The authors are convinced a national guideline promoting unequivocally the use of ultrasound prior to all diagnostic and therapeutic abdominal paracentesis is needed.

Conclusion

This retrospective review of data, spanning from 2013 to 2019, shows that it is possible to introduce an ultrasound-assisted abdominal paracentesis service to an NHS District General Hospital with a high successful procedure and low complication rate. Our findings support the evidence demonstrated in multiple studies that bedside ultrasound enables clinicians to assess ascites accurately, leading to a considerable number of patients avoiding unnecessary and potentially harmful procedures.

Supplemental Material

sj-docx-1-ult-10.1177_1742271X221095405 – Supplemental material for Bedside ultrasonography prior to abdominal paracentesis is associated with low complication and high success rate: Experience in a National Health Service District General Hospital in the United Kingdom from 2013 to 2019

Supplemental material, sj-docx-1-ult-10.1177_1742271X221095405 for Bedside ultrasonography prior to abdominal paracentesis is associated with low complication and high success rate: Experience in a National Health Service District General Hospital in the United Kingdom from 2013 to 2019 by Jan C Droste, Christy Riggott, Tracey Maxfield and Anna Zoltowski in Ultrasound

Footnotes

Acknowledgements

The authors thank the Advanced Clinical Practitioners of the Ambulatory Care Unit in Airedale NHSFT for their practical support for this study. The authors thank the library staff, namely, Paul Stevenson and Martyn Greenwood, for their excellent help with the literature search and obtaining medical journal articles. The authors thank Beth Old, Johan Lempen and Holly Whitaker for their help with the data entry to the excel tables. The authors also thank the audit department for their help and the medical records staff, namely, Carol and Angela for their sustained efforts of finding all the notes needed. Special thank you to Dr Nithin Narayan for help with the statistical data analysis.

Contributors

C.R., T.M. and A.Z. contributed to this article.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

Ethics approval

The data analysed were acquired as part of a local service evaluation and therefore no ethical approval was sought.

Permission from patient(s) or subject(s) obtained in writing for publishing their case report.

YES: written consent obtained for Figures 1 –.

If your answer is NO, then please explain reasons:

If your patient(s) or subject(s) is/are a minor(s), then written consent must be obtained from their parents or their designated responsible guardian and this should made clear under item 7(d).

Normally, we would not need to see the written permission but do want you to declare that you have obtained such permission.

Permission obtained in writing from patient or any person whose photo is included for publishing their photographs and images

YES: written consent obtained for Figures 1 –.

If your answer is NO, then please explain reasons:

If your patient or subject is a minor, then written consent must be obtained from their parents or their designated responsible guardian, and this should be made clear under item 7. Publication of anonymised and de-identified ultrasound images does not need permission except when they are part of a case report.

Normally, we would not need to see the written permission but do want you to declare that you have obtained such permission.

Confirm that you are aware that permission from a previous publisher for reproducing any previously published material will be required should your article be accepted for publication and that you will be responsible for obtaining that permission

Not applicable.

Guarantor

Dr Jan C Droste

ORCID iD

Jan C Droste

Supplemental material

Supplemental material for this article is available online

References

Habeeb

Herrera

. Management of ascites. Paracentesis as a guide. Postgrad Med 1997; 101(1): 191–192.

Cattau

Benjamin

Knuff

, et al. The accuracy of the physical examination in the diagnosis of suspected ascites. JAMA 1982; 247(8): 1164–1166.

Williams

Simel

. The rational clinical examination. Does this patient have ascites? How to divine fluid in the abdomen. JAMA 1992; 267(19): 2645–2648.

Chongtham

Singh

Kalantri

, et al. Accuracy of clinical manoeuvres in detection of minimal ascites. Indian J Med Sci 1998; 52(11): 514–520.

Mallory

Schaefer

. Complications of diagnostic paracentesis in patients with liver disease. JAMA 1978; 239(7): 628–630.

Runyon

. Abdominal paracentesis of Ascitic Fluid. A Safe Procedure. Arch Int Med 1986; 146(11): 2259–2261.

Webster

Brown

Lucey

, et al. Hemorrhagic complications of large volume abdominal paracentesis. Am J Gastroenterol 1996; 91(2): 366–368.

Pache

Bilodeau

. Severe haemorrhage following abdominal paracentesis for ascites in patients with liver disease. Aliment Pharmacol Ther 2005; 21(5): 525–529.

De Gottardi

Thevenot

Spahr

, et al. Risk of complications after abdominal paracentesis in cirrhotic patients: a prospective study. Clin Gastroenterol Hepatol 2009; 7(8): 906–909.

10.

Sharzehi

Jain

Naveed

, et al. Hemorrhagic complications of paracentesis: a systematic review of the literature. Gastroenterol Res Pract 2014; 2014: 985141.

11.

Lin

Wang

Zhu

, et al. Hemorrhagic complications following abdominal paracentesis in acute on chronic liver failure: a propensity score analysis. Medicine 2015; 94: e2225.

12.

Sudulagunta

Sodalagunta

Bangalore Raja

, et al. Clinical profile and complications of paracentesis in refractory ascites patients with cirrhosis. Gastroenterol Res 2015; 8(3–4): 228–233.

13.

Tito

Gines

Arroyo

, et al. Total paracentesis associated with intravenous albumin management of patients with cirrhosis and ascites. Gastroenterology 1990; 98(1): 146–151.

14.

McGahan

Anderson

Walter

, et al. Portable real-time sonographic and needle guidance systems for aspiration and drainage. Am J Roentgenol 1986; 147: 1241–1246.

15.

Bard

Lafortune

Breton

. Ascites: ultrasound guidance or blind paracentesis. CMAJ 1986; 135(3): 209–210.

16.

Nazeer

Dewbre

Miller

. Ultrasound-assisted abdominal paracentesis performed by emergency physicians vs the traditional technique: a prospective randomised study. Am J Emerg Med 2005; 23(3): 363–367.

17.

Mercaldi

Lanes

. Ultrasound guidance decreases complications and improves the cost of care among patients undergoing thoracentesis and paracentesis. Chest 2013; 143(2): 532–538.

18.

Patel

Ernst

Gunnarsson

. Evaluation of hospital complications and costs associated with using ultrasound guidance during abdominal paracentesis procedures. J Med Econ 2012; 15(1): 1–7.

19.

Tirado

Noble

, et al. Ultrasound-guided procedures in the emergency department-diagnostic and therapeutic asset. Emerg Med Clin North Am 2013; 31(1): 117–149.

20.

Sekiguchi

Suzuki

Daniels

. Making paracentesis safer: a proposal for the use of bedside abdominal and vascular ultrasonography to prevent a fatal complication. Chest 2013; 143(4): 1136–1139.

21.

Ennis

Schultz

Perera

, et al. Ultrasound for detection of ascites and for guidance of the paracentesis procedure: technique and review of the literature. Int J Clin Med 2014; 5: 1277–1293.

22.

Adams

Roggio

Wilkerson

. 368 Sonographic assessment of inadvertent vascular puncture during paracentesis using the traditional landmark approach. Ann Emerg Med 2015; 66: S132–S133.

23.

Cho

Jensen

Reierson

, et al. Recommendations on the use of ultrasound guidance for adult abdominal paracentesis: a position statement of the society of hospital medicine. J Soc Hospital Med 2019; 14: E7–E15.

24.

Österreichische Gesellschaft fuer Gastroenterologie und Hepatologie. Aszites und Spontan-Bakterielle Peritonitis, http://www.oeggh.at/files/aszites_und_spontan-bakterielle_peritonitis__sbp___2018.pdf (2018, accessed 20 January 2022).

25.

Societe National Francaise de Gastro-Enterologie. Conseil de pratique: Prise en charge de l’ascite de la cirrhose chez l’adulte, https://www.snfge.org.content/cp043-prise-en-charge-de-lascite-de-la-cirrhose-chez-ladulte-en-dehors-du-syndrome (2016, accessed 20 January 2022).

26.

Tito

Gines

Arroyo

, et al. Total paracentesis associated with intravenous albumin management of patients with cirrhosis and ascites. Gastroenterology 1990; 98(1): 146–151.

27.

Goldberg

Goodman

Clearfield

. Evaluation of ascites by ultrasound. Radiology 1970; 96(1): 15–22.

28.

Conduit

Wesley

Christie

, et al. PTU-002 Large volume paracentesis (LVP) can be safely performed by junior doctors without ultrasound guidance. Gut 2013; 62: A42.

29.

Droste

Riggott

. Variation in paracentesis practice in the United Kingdom could endanger patient safety. J Patient Safety 2021; 17(1): e42.

30.

British Society of Gastroenterology: BSG-BASL decompensated cirrhosis care bundle – first 24 hours, https://www.bsg.org.uk/wp-content/uploads/2019/12/BSG-BASL-Decompensated-Cirrhosis-Care-Bundle-First-24-Hours.pdf (2019, accessed 20 January 2022).

31.

Reggio

Angeloni

. Ascitic fluid analysis for diagnosis and monitoring of spontaneous bacterial peritonitis. World J Gastroenterol 2009; 15(31): 3845–33850.

32.

Kim

Tsukamoto

Mathur

, et al. Delayed paracentesis is associated with increased in-hospital mortality in patients with spontaneous bacterial peritonitis. Am J Gastroenterol 2014; 109(9): 1436–1442.

33.

Sakai

Sheer

Mendler

, et al. Choosing the location for non-image guided abdominal paracentesis. Liver Int 2005; 25(5): 984–986.

34.

Berry

. Hemorrhagic complications of paracentesis: aberrant anatomy versus aberrant technique – a fatal case of abdominal hemoperitoneum. Cureus 2020; 12(6): e8827.

35.

Barsuk

Cohen

Feinglass

, et al. Clinical outcomes after bedside and interventional radiology paracentesis procedures. Am J Med 2013; 126(4): 349–356.

36.

Cheema

Khwaja

. Implications of miscoding urological procedures in an era of financial austerity – ‘every penny counts’. JRSM Open 2015; 6(6): 2054270415593463.

37.

Hamilton

Arnold

. Accuracy of clinical coding of pleural empyema: a validation study. J Eval Clin Pract 2020; 26(1): 79–80.

38.

Aithal

Palaniyappan

China

, et al. Guidelines on the management of ascites in cirrhosis. Gut 2021; 70(1): 9–29.

39.

Havelock

Teoh

Laws

, et al. Pleural procedures and thoracic ultrasound: British Thoracic Society pleural disease guideline 2010. Thorax 2010; 65(Suppl. 2): ii61–ii76.

40.

Brass

Hellmich

Kolodziej

, et al. Ultrasound guidance versus anatomical landmarks for internal jugular vein catheterization. Cochrane Database Syst Rev 2015; 1(1): CD006962.