Blood transfusion, embolisation and nephrectomy after percutaneous nephrolithotomy (PCNL)

Authors


Correspondence: Stephen R. Keoghane, Department of Urology, Queen Alexandra Hospital, Portsmouth, Hants PO18 9HA, UK.

e-mail: Stephen.Keoghane@porthosp.nhs.uk

Abstract

What's known on the subject? and What does the study add?

  • Percutaneous treatment for renal stone disease is associated with a risk of significant morbidity.
  • Our large UK series provides contemporary data on the risk of vascular complications and admission to the Intensive Care Unit (ICU) after PCNL. When compared with recent international databases, these data support the current evidence that better outcomes can be achieved in centres performing large numbers of procedures. These data add to the debate for the centralisation of specialist stone surgery.

Objective

  • To audit the outcome of percutaneous nephrolithotomy (PCNL) at a UK stone centre over a 10-year period, and provide patients with understandable contemporary data on blood loss and vascular risk.

Patients and Methods

  • A single centre retrospective analysis of all PCNLs undertaken between April 2000 and December 2010.
  • The association between transfusion and patient age, operative duration and positive preoperative mid-stream urine (MSU) sample was subject to statistical analysis.

Results

  • Data on 568 patients was analysed. 21 were paediatric cases with a mean (range) age of 8 (2–16) years; 547 were adult cases with a mean (range) age of 55 (17–84) years.
  • 3.8% of adult patients (21/547) received a blood transfusion; mean age 60 years (55 years in those not transfused) with a mean operative duration of 119 min (103 min in those not transfused).
  • 23.8% of patients transfused had a confirmed preoperative urinary tract infection compared with 16.1% of those not transfused.
  • Seven patients underwent angiography, with five having selective arterial embolisation (0.9%).
  • There were no deaths in this series although one patient (0.2%) required an urgent nephrectomy due to cardiovascular instability from bleeding.

Conclusions

  • Large UK series that provides contemporary data for consent on vascular risk at PCNL.
  • The risk of transfusion is associated with increased patient age, operative duration and the presence of a positive preoperative MSU sample.
  • Data compares favourably with other large published series, and supports the argument for centralisation of percutaneous stone management.
Abbreviations
ASA

American Society of Anesthesiology

CROES

Clinical Research Office of the Endourological Society

ICU

intensive care unit

MSU

mid-stream urine

PCNL

percutaneous nephrolithotomy

Introduction

Percutaneous access to the kidney was first described in 1865 by Thomas Hillier who repeatedly drained a 4-year-old boy's kidney, which he thought at that time to be congenitally obstructed. Percutaneous nephrolithotomy (PCNL) for stone disease was first described in the 1970s by Fernstrom and Johannson [1]. Presently, the clinical practice guideline report for the management of staghorn calculi, by the AUA and European Urological Association, recommend that percutaneous treatment of staghorn calculi should be first-line treatment for most patients [2] (Fig. 1). Despite this, PCNL does carry the risk of significant morbidity, with contemporary series describing a complication rate of 20.5% [3], and transfusion rates varying enormously between <1% and 55% [4-10] (Table 2).

Figure 1.

Plain abdominal radiograph of the kidneys, ureters and bladder showing a right staghorn renal calculus.

Data on outcome are now considered an essential tool to aid patient choice before surgery, and a discussion of unit/surgeon outcome data should be an inherent part of the process of informed consent. This audit of outcome of PCNL over a 10-year period aims to provide patients with understandable data on blood loss and vascular risk.

Patients and Methods

Data on procedures between April 2000 and December 2010 were extracted and cross referenced from three separate hospital databases and checked against operating theatre diaries. This list of patients was then individually checked against the hospital transfusion, radiology and intensive care unit (ICU) databases.

The association between transfusion and patient age and operative duration was subject to statistical t-test; and association with positive preoperative mid-stream urine (MSU) sample subject to statistical chi-squared test (Table  1).

Table 1. Comparison of those patients transfused and not transfused with regards to age, operative duration and positive preoperative MSU sample
 TransfusedNot transfusedP
Number of patients21526 
Mean age, years60550.07
Mean operative duration, min1191030.16
Positive preoperative MSU, %23.816.10.53
Table 2. Comparison of transfusion and embolisation rates with other published series
SeriesNTransfusion rate, %Embolisation rate, %
Keoghane et al. 2012 (present study)5473.80.9
de la Rosette et al. 2011 [20]58035.7 
Soucy et al. 2009 [19]13380.7/1.2 
Stoller et al. 1994 [4]12723 
Jones et al. 1990 [6]100029/550.6
Lee et al. 1987 [5]58212 
Martin et al. 2000 [25]808 1
Kessaris et al. 1995 [26]2200 0.8

Blood transfusion was classified as a Clavien class II complication, radiological embolisation under local anaesthesia as IIIa and ICU admission or life-threatening complication as a class IV [11, 12].

Results

In all, 595 patients underwent PCNL between April 2000 and December 2010. Data were available on 568 patients during this period. In all, 21 were paediatric cases with a mean (range) age of 8 (2–16) years; 547 were adult cases with a mean (range) age of 55 (17–84) years. The procedures were performed by three surgeons, one specialising in paediatric stone surgery. Radiological access was obtained in a very small number of patients between June and December 2003. All other access to the kidney was obtained by a urological surgeon using both telescopic dilators and balloon dilatation.

The mean operative duration, calculated from when the patient entered the operating theatre to leaving, was 104 min in adults and 135 min in the paediatric cases.

In all, 3.8% of adult patients (21/547) received blood transfusion; mean age 60 years (55 years in those not transfused); the mean operative duration was 119 min (103 min in those not transfused).

In all, 23.8% of patients transfused had a confirmed preoperative UTI compared with 16.1% of those not transfused.

One bowel puncture was reported, and one case of a pneumothorax in a horseshoe kidney. Seven patients underwent angiography, with five having selective arterial embolisation (0.9%), and there were eight admissions to ICU (1.5%). There were no deaths in the present series, although one patient (0.2%) required an urgent nephrectomy due to cardiovascular instability from bleeding.

Discussion

The benchmark paper based on the early days of PCNL in the UK was published in 1990. A series of 1000 PCNLs performed between 1981 and 1988 reported transfusion rates of between 29 and 55%. Six patients required massive transfusions (8–14 units), but all were successfully managed with embolisation [7]. The BAUS section of Endourology recently published pooled data on 987 prospectively reported cases undertaken at 41 UK centres over 18 months, with a transfusion and embolisation rate of 2.5% and 0.4%, respectively [13]. Other published transfusion rates after PCNL have varied enormously and are quoted between <1% and 55% [4-10]. The present rate of 3.8% represents contemporary data from a UK regional stone centre.

Published data suggests that patient age, American Society of Anesthesiology (ASA) grade, stone burden and operative duration, are associated with an increased risk of vascular complication. There is contradicting data in regards to this risk and the number of punctures undertaken.

In a small series of 26 procedures in men aged 70–94 years, an increase in the frequency and Clavien grade of complications was noted in the elderly when compared with a younger control group, but the small sample size did not allow for meaningful statistical comparison [14]. Similarly the present analysis, although showing a trend towards an increase in transfusion risk with age is not statistically significant (P = 0.07).

It is postulated that the aggressive treatment of large stone burdens with multiple punctures is associated with greater blood loss, although very variable figures of between 3 and 46% have been reported [15-18]. Martin et al. [19] documented a 28% increase in the incidence of bleeding when the number of punctures rose above two. The transfusion rate also rose by 22% in these patients to 42%. However, a large contemporary Canadian series of 1338 patients treated between 1990 and 2005 found no statistical difference in transfusion rates in patients who were treated using single or multiple tracts [20]. In the present study the electronic patient records have not allowed a comparison between the two methods of track dilatation, nor an assessment of stone burden, as the older cases pre-dated digital radiological imaging. Only two cases with multiple tracts were included in the present study and therefore scientifically valid comments on the relationship between the number of punctures and transfusion were not possible.

The contemporary Clinical Research Office of the Endourological Society (CROES) database comprises a series of 5803 international PCNLs. The relatively higher transfusion rate of 5.7% probably reflects the dilution of outcomes by the pooled nature of the data. An increased risk of transfusion was associated with higher ASA scores, stone burden and, as with the present study, a trend towards operative duration (P = 0.16) [21]. Of great interest, and currently unexplained, is the difference in transfusion rate between telescopic dilatation at 4.9% and balloon dilatation at 7% [22]. A trend replicated by the recent BAUS audit: 0.8% vs 3.2% [13]; when, in contrast, a recent non-randomised comparison and literature review between balloon and serial dilatation showed no clear trend demonstrating superiority [23]. A rate of 0% in solitary kidneys is difficult to explain and again highlights some of the anomalies that have arisen from the CROES data [24]. A further subset analysis from CROES of 102 patients with ectopic, horseshoe or malrotated kidneys had a transfusion rate of 5% [25].

The present series highlights a previous unpublished trend between preoperative UTI and risk of haemorrhage requiring transfusion (Table 1). Possible explanations include the hyperaemic nature of inflamed urothelium, or distorted anatomy secondary to oedema.

Management of percutaneous renal haemorrhage has advanced in recent years with the introduction and increasing use of minimally invasive endovascular techniques to accurately diagnose and effectively treat vascular complications, such as, arteriovenous fistulae and false aneurysms. The present embolisation rate of 0.9% compares favourably with that of 0.6% described in the 1990 large UK series by Jones et al. [7]. Contemporary series have described rates of 0.4–1% [13, 26, 27].

A relatively recent development has been the introduction of the mini-PCNL and the use of the supine or lateral decubitus position. A series of 650 mini-PCNL cases was reported in 2011 using a 12-F sheath. The reported mean stone surface area was 4.1 cm2 and a transfusion rate of 1.4% was noted with only two cases of embolisation described [28]. Similarly a transfusion rate of 1.4% was reported in the first large series of 557 supine PCNLs. One open incision to stem bleeding (0.18%), one nephrectomy (0.18%) and one selective embolisation (0.18%) were described by Valdivia et al. [29]. De Sio et al. [30] have performed one of the few randomised control trials of the supine vs the prone position. They found no difference in complication rate between the two approaches. A literature review of PCNL studies between 1998 and 2008 (nine supine and 25 prone) reported transfusion rates of between 0 and 9.4% for the supine position. The results of that analysis of 34 papers are difficult to evaluate as the comparison with the prone position was based on surgery in the obese [31].

What does the present article add to the current literature? It describes a large contemporary UK series whose vascular complications compare favourably with others [32]; providing data to form part of the process of informed consent (Table  3). The risk of transfusion is associated with increased patient age, operative duration and the presence of a positive preoperative MSU. The retrospective nature of the data collection and its limitation on stone burden and puncture data is acknowledged. The comparatively higher rate of transfusion and embolisation observed in comparison with the BAUS database could be explained by the latter's possible reporting bias, and incomplete data set in almost a quarter of its patients. In the UK, there has been a move towards the centralisation of urological cancer services, but this has yet to happen in the field of stone surgery. Data now exist to support a better outcome in centres performing large numbers of percutaneous stone procedures, both for operative duration and transfusion rate [22]. Future interrogation of national databases will make comparisons between centres feasible. The authors would advocate other units explaining personal outcome data as part of the process of informed consent and indeed allowing the public to access individual outcome data. It is no longer acceptable to quote large national or international series when obtaining consent for complex surgery and the patient should know the capability of the operating surgeon when compared to his or her peer group.

Table 3. Take home message: summary of audited complication risk
Transfusion rate3.8%
Embolisation rate0.9%
Nephrectomy rate0.2%
ICU admission rate1.5%

Acknowledgements

Mr Stephen Chiverton performed some of the operations. Mr Neil Harris, Tim Crook and Christopher White were involved in various stages of data collection as urological trainees.

Conflict of Interest

None declared.

Ancillary