Discussion

Parts of this chapter are reproduced with permission from Wiseman et al.1 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) licence, which permits others to distribute, remix, adapt and build upon this work, even for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The text below includes additions and formatting changes to the original text.

When patients present with lower pole renal stones, if active treatment is pursued there are three possible technologies available within the NHS which might be used: ESWL, FURS with laser lithotripsy, and PCNL. The choice of treatment can be guided by stone size, likely stone composition, the anatomy of the drainage system of the affected kidney, clinician and patient preference, and availability of equipment and expertise.12

Guidance issued by the EAU12 at the time of study design, in 2014, which was widely followed in UK clinical practice, recommended ESWL or flexible ureteroscopy (URS) as first line options for LPSs ≤ 10 mm. For stones > 20 mm, PCNL was recommended as the first-line treatment option. For stones between these two sizes, in the lower pole, the recommendation depended on whether there were unfavourable factors for ESWL, previously detailed. If unfavourable factors were present, then PCNL or URS were the first-line recommended options, with ESWL being second line. If unfavourable factors were not present then ESWL or PCNL or URS were all possible first-line treatment options.12 The EAU Guideline also commented that there remained considerable uncertainty regarding the management of LPSs, with each treatment option having advantages and disadvantages.

A Cochrane review and meta-analysis in 2014 of RCTs compared ESWL with either FURS or PCNL for the treatment of renal stones13 included three studies looking at LPSs and concluded that included trials were small and of low methodological quality. Another review (Donaldson, 2015)31 included trials comparing PCNL with FURS and identified four further studies. Looking at the seven trials as a whole, the GRADE certainty of evidence ratings for the outcome of SFRs indicated they were of ‘moderate’ quality. The included trials reported few data on patient outcomes (such as QoL) or on resource use, and none on cost-effectiveness. The PUrE trial therefore set out to compare the three treatment options for lower pole renal stones, separated into two (RCT1 and RCT2) pragmatic, multicentre, patient-randomised, open-label superiority RCTs. RCT1 compared FURS with ESWL, as first treatments options in the pathway, recruiting patients with stones of maximum dimension ≤ 10 mm. RCT2 compared FURS with PCNL, as first treatments options in the pathway, recruiting patients with stones of maximum dimension > 10 and ≤ 25 mm. It aimed to provide robust evidence on health status, QoL, clinical outcomes, and resource use to both the NHS and society to close the gap in evidence. The guidance published by NICE in 201979 recommended that for stones anywhere in the kidney under 10 mm, ESWL should be offered, with FURS as an option only in specific situations. For stones between 10 and 20 mm FURS or ESWL could be considered, with PCNL offered only if URS or ESWL have failed. For stones over 20 mm, again not specifically LPSs, they recommended that PCNL should be offered, with FURS to be considered if PCNL is not an option.

Both RCT1 and RCT2 used a pragmatic outcome for stone clearance defined as ‘Complete clearance’ if the target stone had been treated such that no further action (including no need for observation of it) was required; ‘Acceptable clearance’ was chosen if observation was needed for a residual fragment but no further intervention was deemed necessary; and ‘Unacceptable clearance’ was recorded as the outcome for a residual stone that was deemed to require further intervention.

The remainder of the discussion chapter is split into a section on RCT1 then a section on RCT2, except for some general strengths and limitations which are considered for the whole PUrE study.

RCT1

Primary outcome

In the PURE study, RCT1 showed a small benefit in favour of FURS to 12 weeks post treatment compared to ESWL of 0.024 (−0.004, 0.053). When the model was adjusted for the differential waiting times, in effect removing the potential impact of waiting times on the PO, the difference increased slightly to 0.04 (0.01, 0.07) in favour of FURS. However, this small improvement in health status did not reach the pre-determined threshold of clinical significance. There was a notable decrease in the weekly participant reported health status in both groups in the initial postoperative period, which might be expected when intervening for a stone that in many patients is relatively asymptomatic. This decrease was greater in the FURS group than the ESWL group and was no longer apparent at 2 weeks after treatment.

Secondary outcomes

At 12 weeks post intervention, RCT1 showed that patients who underwent FURS were significantly more likely to have complete stone clearance compared to ESWL (72% compared to 36%). The FURS group also required significantly fewer additional treatments compared to ESWL (9% compared to 27%). There was no difference in the complications recorded between the two groups.

Self-reported health status from randomisation up to 12 weeks was the same in both groups and the precision of estimate ruled out any benefit in favour of either intervention.

The pain measures recorded postoperatively inversely reflect the pattern of the health status (EQ-5D score) plots, being higher than baseline in both groups, and greater in the FURS group than the ESWL group, with a significant reduction in pain by week 2, and with levels continuing to decrease over the subsequent weeks. The worst level of pain since the intervention was consistently higher in the FURS group at each week.

Cost-effectiveness

This study has shown that the adjusted mean costs per participant were £2223 and £3362 for the ESWL and FURS groups, respectively, resulting in an adjusted cost difference of £1138 and a non-significantly higher adjusted QALY difference of 0.017 for the 12-month follow-up period. The ICER between FURS and ESWL was £65,163 per QALY gained by FURS, which is above the usual cost-effectiveness threshold considered in the UK decision-making (i.e. below £30,000 per QALY gained).65 We found that at a threshold value of £20,000 per QALY, ESWL has a 99.9% chance of being cost-effective. The 5-year Markov model extrapolation showed that FURS generates more expected QALYs but is also more costly than ESWL with an incremental cost per QALY higher than the usual cost-effectiveness thresholds used for the NHS decision-making (i.e. £20,000 per QALY gained). The extrapolation analysis also showed the probability of ESWL being cost-effective reducing to 74% at £20,000 threshold value.

Comparison of findings with other studies

There are few studies reporting health status and QoL in kidney stone disease. The only directly comparable study is by Pearle et al.,30 which showed that patients undergoing ESWL had a faster time to ‘100% recovered’, as well as quicker return to work, driving and non-strenuous activity, in contrast to the patient-reported outcomes in the study. In their RCT of stones between 6 and 20 mm at any location in the kidney, Bosio et al. did not show any difference in patient satisfaction between ESWL and FURS, though it was not possible to break this down by location in the kidney or by stone size.80 Another study of 70 symptomatic adults with larger LPSs, of 10–20 mm, who underwent FURS or ESWL, showed the average time to return to normal activity and voiding symptoms were significantly higher in the FURS group, and overall satisfaction score was significantly higher in the FURS compared to the ESWL group.36 Our systematic review (see Appendix 10, Tables 82–84, Figures 39–46) shows that the evidence base on QoL remains inconclusive and of very low certainty due to risk of bias (RoB) and heterogeneity of measurement.

In contrast to PUrE, where there was a superior SFR in the FURS group compared to the ESWL group, the study by Pearle et al., in 2005, did not show any difference in SFRs.30 The meta-analysis by Donaldson et al. included three studies which are directly comparable to the study and showed a marginally superior SFR within 3 months for FURS compared to ESWL for LPSs under 10 mm.31 The RCT undertaken by Bosio and colleagues showed higher SFRs at 1 and 6 months, but not at 1 year for FURS compared to ESWL, but this was for stones between 6 and 20 mm at any location in the kidney. The SFR was higher for stones 6–10 mm at 1 month for FURS, but not subsequently, and the SFR for only LPSs from 6 to 20 mm was not higher at any time point.80 Our systematic review includes three studies reporting on SFR at 3 months, when restricting stone size to ≤ 10 mm, that is, the inclusion criteria for RCT1 (see Appendix 10, Figure 41) and the SFRs were not statistically different between ESWL and FURS, with moderate certainty of evidence – which was downgraded due to high RoB (see Appendix 10, Figure 40). The present study RCT1 found significantly higher SFRs at 3 months for FURS (72.1%) vs. ESWL (36.4%). The literature is inconsistent when reporting on SFRs, both in terms of imaging used and whether fragments up to a certain size are allowable or clinically significant. The SFRs in our study are likely higher as imaging was not mandated, but need for further treatment or observation of the stone was the determining factor of whether the patient was stone-free. This is a pragmatic and meaningful outcome to patients and clinicians, especially in terms of what happens in a health system when not all patients are followed up with imaging. Additionally, it is accurate and reproducible across studies. However, the use of this definition of stone free could explain the higher SFRs reported in the RCT1.

The present study also showed that the FURS group had a convincingly reduced number of required additional treatments 12 weeks after intervention compared to ESWL [RR: 0.32 (0.18, 0.58); < 0.001]. While some previous comparable studies also showed higher retreatment rates in the ESWL arm, these did not reach statistical significance in contrast to PUrE.34,81 This is likely due to the size of the present study in comparison. This point was also examined in our systematic review, where a total of seven studies reported on retreatment rates (see Appendix 10, Figure 43). Problems with outcome definition and heterogeneity make meta-analysis inappropriate.

The complication rate was low and similar in the two arms of the present study, with seven complications in each arm of the study at Clavien 3 or greater. This is in line with previously published studies.34,81 In our systematic review (see Appendix 10, Figure 44), it remains uncertain which intervention is more effective in terms of reducing the risk of complications, and the certainty of evidence remains very low. However, the majority of complications for both treatments in RCT1 were minor.

The results from RCT1 support guidance published in 2019 by NICE,79 which recommended that ESWL should be offered in the first instance for stones < 10 mm, anywhere in the kidney. This is because of the evidence that ESWL offered a better balance of benefits and costs, and that FURS should be considered only if there are contraindications for ESWL, or anatomical reasons or multiple stones, or if a previous course of ESWL has failed.

Comparing the costs of ESWL and FURS in RCT1 to previous studies is challenging, mainly due to the complexity of the procedures and differing equipment used even within a single healthcare system, let alone the differences across healthcare systems outside the UK. These challenges are shown in the meta-analysis by Geraghty.82 The authors attempted to tabulate measurements of URS compared with ESWL for ureteric stones and stones located anywhere in the kidney in studies from several countries and found that different costs of treatment were included in different studies. It should be noted, however, that URS is ureteroscopy with a rigid instrument, and this is usually undertaken for stones in the ureter, and differs from FURS which uses a flexible instrument, with FURS usually being costlier due to increased number of single-use items used.

On average, the FURS care pathway cost more than the ESWL care pathway in PUrE, and this cost difference was mainly driven by the unit cost of FURS. While the mean costs of additional and subsequent hospital admissions were higher for the ESWL group, these higher costs did not overturn the initial difference in cost in favour of ESWL. Koo et al.83 reported the cost in a UK setting of the SWL and flexible ureteroscopic laser lithotripsy procedure for LPSs under 20 mm. The total mean costs of the interventions were £2602 for FURS and £426 for ESWL, at 2010 costs, which is similar to the costs in our PUrE study for FURS, which were £3362, when adjusting for inflation between 2010 and 2020. However, the costs of ESWL in the present study were significantly greater at £2223, likely due to the fact that they incorporate additional per-protocol treatments as well as any additional costs in the year after randomisation.

In the TISU study, which looked at the treatment of ureteric stones of any size,84 the cost of ESWL was calculated at £1550, based on the costs of a six-month follow-up after randomisation. The costs of ESWL in PUrE are on average £672 higher, which can likely be accounted for by increased costs related to the longer duration of follow-up, together with an increased chance of requiring additional treatment due to the location of the stone in the lower pole, a position from where it is less likely to clear without intervention.

A study by Zhang et al. in 201885 undertaken in the Chinese healthcare system showed that the cost of ESWL was $535, and that for FURS was $4657 but for stones which were larger than in the present study, at between 10 and 20 mm. Again, the costs of treatments across healthcare systems are difficult to compare, and the greater size of the stones in this study also makes direct comparison of costs less valid.

In summary, the current study shows that FURS is more expensive than ESWL, and while the absolute cost of ESWL is higher than that reported in many studies, the relatively long follow-up and inclusion of further interventions in the costs, account for increase in ESWL costs in RCT1.

RCT2

Primary outcome

In the PUrE RCT2 study, patients with larger kidney stones (i.e. from 10 mm up to 25 mm) were randomised to treatment with FURS or PCNL. It was difficult to recruit to this study and the final sample size was well short of the planned size. Despite the reduced sample size, the results are of the same order (mean, SD, and effect sizes) as in RCT1 and the results are powered sufficiently for significance and detect marginal difference. The difference in PO significantly favoured PCNL, the health status being 0.07 (0.02, 0.11) higher over the 12-week post-intervention period. These estimated effect sizes should be considered alongside the original MCID, which was originally deemed to be around 0.07. This suggests that although significant, the difference in health status is marginal.

It should be noted that RCT2 participants did not have a particularly low health status score, even to begin with, meaning large improvements would be difficult to achieve.

Waiting times in RCT2 were longer than anticipated but very similar in both treatment groups with a mean interval of approximately 6 months until their treatment but with large SD indicating a wide range in the distribution of waiting time, adjusting for waiting time gave similar results for the PO.

Secondary outcomes

Both RCT1 and RCT2 used a pragmatic outcome for stone clearance defined as ‘Complete clearance’ if the target stone had been treated such that no further action (including no need for observation of it) was required; ‘Acceptable clearance’ was chosen if observation was needed for a residual fragment but no further intervention was deemed necessary; and ‘Unacceptable clearance’ was recorded as the outcome for a residual stone that was deemed to require further intervention. At 12 weeks post intervention, the FURS care pathway had fewer patients with complete stone clearance than PCNL (47.9% vs. 70.9%) with the overall effect indicating that for the larger stones in this arm of the study PCNL was superior to FURS [OR: 0.27 (0.16, 0.43); p < 0.001] for stone clearance. Treatment-related complications occurred in 22% (16/73) of participants randomised to FURS and in 20% (17/86) randomised to PCNL. These were mostly minor (17 Clavien-Dindo grade 1 or 2 accounting for 21% of the FURS arm and 15% for PCNL), 3% of patients in the FURS arm (2/73) and 5% in the PCNL arm (4/86) had a complication with Clavien–Dindo grade 3 or above.

The level of stone-related pain that participants experienced on the day (using a scale of 0–10, with lower scores representing less pain) at baseline was low with a median score of zero (IQI 0–3). Pain was experienced in the previous week by just less than half of the patients (44% overall), and this did not usually require analgesia – the median for days requiring pain medication in the previous was 0 days (IQI 0–7). At the time of intervention, patients still had a low level of pain with a mean of 1.6 for FURS and 1.8 for PCNL preoperatively. The majority of patients had had no pain in the 7 days before treatment (56.2% FURS/55.8% PCNL), and accordingly few patients had required pain relief in the prior 7 days (1.3 requiring treatment for FURS vs. 1.5 days requiring treatment for PCNL). Following treatment, the ‘worst pain’ since intervention was worse for FURS than PCNL. QoL appears primarily affected by pain: the weekly pain measures mirrored health status for both FURS and PCNL. As is intuitively logical, pain was worst in the early postoperative period for both procedures, with pain was very slightly higher in the first weeks for those allocated to FURS compared to those who underwent PCNL but which then decreases. This is reflected by an initial worsening of QoL as already discussed.

Simple profiles of health status and the EQ-5D-VAS score both indicate slight improvement in both arms from baseline to the intervention date, possibly due to the reassurance that a plan was in place the anticipated treatment would make the patient better. There was a fall during the first week post intervention, particularly for patients in the FURS treatment arm as expected after such surgery. Thereafter, levels improved to a steady state similar to their original baseline level but remained lower for FURS patients than those in PCNL arm, whose steady state was slightly greater than their baseline before intervention.

Additional treatments

For the PUrE study, additional intervention was defined as any procedure required that was not part of the treatment pathways. The RCT2 treatments compared PCNL (which was expected to completely remove stones up to 25 mm in a single procedure) and FURS (which was also expected to be a single procedure although, an additional FURS procedure was accepted as part of the original treatment provided it took place with 6 weeks of the initial FURS procedure). We found that there were fewer unplanned procedures following PCNL, such that the FURS group had a larger risk of undergoing additional treatments compared to PCNL [RR: 3.96 (2.07, 7.57); p < 0.001]. Again though, as with the 12-week results, FURS had a lower rate of stone clearance compared to PCNL [OR: 0.24 (0.16, 0.35); p < 0.001] and an increased rate of requiring additional treatment and/or had additional treatment [RR: 2.78 (1.69, 4.57); p < 0.001].

Patient-reported pain, health status, and quality of life

The secondary outcomes of health, measured by the EQ-5-5L score and the VAS scale, and QoL as the SF-12 scores (physical and mental domains) are also reported Table 15. Both EQ-5D measures show similar same patterns with initial detriment seen in FURS followed by improvement after 2–5 weeks statistically comparable albeit a bit lower to the PCNL group. The SF-12 measures do not show any difference by 12 weeks.

Cost-effectiveness

The QoL results using the mean EQ-5D scores were slightly higher for the FURS group than PCNL at baseline and pre intervention, and then reduced for both FURS and PCNL 1 week after treatment, followed by a steady recovery up to week 6. Overall, the week-by-week post-treatment scores were slightly higher for PCNL. At 3 months post treatment, the EQ-5D scores were higher for both groups compared with their scores at baseline and pre treatment before converging back to their baseline values at 12 months. Our analysis showed mean QALYs of 0.775 and 0.773 for PCNL and FURS, respectively, equating to an adjusted mean difference of −0.001 QALYs (95% CI: −0.004 to 0.042).

Since there was a negligible difference in QoL improvement between the two treatments, the cost-effectiveness is driven by the cost of the procedure. The treatment costs for FURS and PCNL were established as ‘micro-costings’ using the details of the patient journey within the trial, as well as using more generalised HRG costs.

Using detailed micro-costing for the surgical intervention and postoperative treatment, the 12-month mean cost per participant was £5298 for ITT with FURS and £4565 for ITT with PCNL, resulting in an adjusted mean difference that treatment by PCNL was £733 (95% CI: −£508 to £1973) cheaper than treatment with FURS, with an ICER between FURS and PCNL of £692,724. Therefore, ITT with FURS is, on average, more costly and does not produce additional QALYs compared with ITT with PCNL. Further analysis to address uncertainty (e.g. 1000 bootstrapped iterations of the regression analysis nested multiple imputation) showed PCNL is highly likely to be cost-effective at the usual cost-effectiveness threshold values used for decision-making in the UK NHS (e.g. probability of 0.87 at £20,000 per QALY gained) up to 12 months post randomisation. A sensitivity analysis was conducted using HRG costs for both the index and additional interventions (e.g. repeat FURS). For this, the ITT with PCNL mean total cost was £6703 and with FURS was £5769, such that PCNL was now £934 more expensive than FURS. Since the mean difference in QALYs remained unchanged as this analysis only varied the calculation of costs, this change in cost attribution meant that FURS was highly likely to be considered cost-effective (i.e. 90% at £20,000 cost per QALY threshold) if HRG costs are used rather than the detailed micro-costings summarised above.

Since the more expensive intervention between FURS and PCNL were reversed when based on HRG rather than micro-costings, the cost-effectiveness depends on the method used for assessing the cost.

While data on the time in anaesthetic, theatre and recovery room were meticulously collected and related staff, equipment and disposable costs incorporated in the micro-costing, the approach might underestimate the cost of the procedures by failing to include cost categories such as staff being on call when the patient is in the recovery room after surgery. While the micro-costing could have underestimated the cost of, for example, PCNL, it is unclear why the HRG cost associated to PCNL (LB75 ‒ Elective inpatient; £8285) is almost two and a half times the HRG cost allocated to FURS (LB65 ‒ Day case; £3350) when the time to complete the procedures and the length of hospital stay are not substantially different between the two procedures.

Since there was a negligible difference in QoL improvement between the two treatments, the cost-effectiveness is driven by the difference in cost between the trial groups, and this was different for the micro-costing approach (FURS £5298 and PCNL costing £4565) and HRG-based costing (where FURS costs £5769 and PCNL costs £6703).

The discrepancy between the micro-costings measured in the trial and the more general HRG costing is perhaps even more relevant at a hospital level as opposed to the wider system, where both represent a cost. For hospital providers, the micro-costings are an expense, whereas the HRG cost represents income for NHS trusts/hospitals. Considering these differences between the costs of performing PCNL for a hospital, and the relatively higher reimbursement via the HRG, PCNL is the intervention which should be chosen over FURS from a purely economic point of view.

Lessons: HRG for PCNL is ‘good value’ for trusts (unless the micro-costings have underestimated the cost, for example, with an interventional radiologist time, or more consumable use, for example, laser and/or basket for miniaturised PCNL track). Nevertheless, these data should drive a greater use of PCNL as both the more cost-effective and better-paid intervention.

Five-year extension

Since residual fragments are associated with a higher risk of recurrent symptoms and intervention67 economic assessment was extrapolated to anticipate additional events over a 5-year time horizon using a cohort Markov model to reflect possible care pathways after the initial intervention and the potential need for further intervention due to recurrence.

The RR for stone clearance was established using PUrE trial participants who had undergone CT imaging at 12 weeks to provide SFRs and patients with residual fragments of size ≤ 4 or > 4 mm after 12 weeks follow-up. The model also incorporated the care pathways observed in the trial including the observed waiting times.

The extrapolation exercise showed the cost-effectiveness of FURS was dependent on the method used to estimate costs with PCNL still being cost-effective when micro-costing was used. Moreover, the results using HRG-based costing show PCNL cost-effectiveness improving for longer time horizons. However, while at 5 years, PCNL generates more QALYs, it does this at an additional cost that is too high to be considered cost-effective. Finally, our analysis suggests the RR SFR should be above 0.6 (FURS generates 60% of the SFR of PCNL at 12 weeks post surgery) for FURS to be cost-effective when HRG-based costing was used and above 0.95 when micro-costing was used.

Comparison of findings with other studies

Our up-to-date meta-analysis described in Appendix 10, Tables 82, 83 and 85, Figures 39–40 and 47–52 showed that the operative time for PCNL was shorter in all five out of the six studies that reported procedure time (see Table 85 and Figure 52), ranging from 6 minutes longer to 38 minutes shorter than FURS. Bozzini et al. reported that PCNL showed a non-significant difference in operative time (72 vs. 56 minutes) but significantly longer fluoroscopy exposure time (176 vs. 32 seconds), and a longer hospital stay (3.7 vs. 1.3 days) than FURS.86 Zeng also found no significant difference in operating time between Super mini PCNL (SMP) and FURS.87

However, Jiang showed that the ‘all-seeing needle’ micro-PCNL patients had a shorter operative time than FURS,88 while Fayad et al. also found that operating time was significantly shorter for mini PCNL (72 minutes) than FURS (110 minutes).89 Zhang also found that the operating time was shorter in Ultramini PCNL (UMP, 69 minutes) than FURS (93 minutes) although ESWL had the shortest procedure time (46 minutes).85

Zhang showed an increasing length of stay (see Table 85 and Figure 51) from 1.1 days for ESWL to 3.2 days for FURS and 5.3 days for Ultra-mini PCNL for LPSs 1–2 cm.85 Likewise, Yavuz et al. showed that hospital stay increased with increasing invasiveness of the intervention from 1 day for FURS, 1.5 days for micro PCNL, 2 days for both ultra-mini and mini PCNL and 3 days for standard PCNL.90 In contrast, both Fayad et al. and Zeng et al. reported that the mini PCNLs performed in their respective studies had a comparable hospital stay to FURS.87,88 Jiang et al. also showed no significant difference in length of stay between micro PCNL and FURS, although the LOS was longer than usual for UK practice with more than 3 days in hospital for both procedures.88

It is likely that these data reflect depend on national systems including reimbursement and are therefore not a particularly useful marker of early procedure complexity and outcome. Given this, a validated PROM might act as a more useful comparator for future research.

As reported in PUrE RCTs, the complications reported in the literature for FURS and PCNL are usually Clavien–Dindo grades I and II, with no clinically relevant difference between the two treatments. Although eight RCTs reported complications (see Table 85 and Figure 50) and in our revised meta-analysis, there is a lack of standardised reporting which makes these difficult to compare against.

Bozzini et al. found an increasing complication rate from 6.7% for ESWL, 14.5 for FURS and 19.3% for PCNL.86 Whereas Zhang et al. noted 6.7% in SWL and 8.3% in FURS and 16.7% in UMP.85 These were mostly Clavien–Dindo Grade I and II, with FURS having a higher rate of postoperative fever than UMP. Jiang et al. found no significant difference between micro PCNL and FURS for LPSs < 20 mm for complications including haemoglobin drop.88 In keeping with this, blood transfusion is uncommon with treatment of stones of this size and was not required in either supermini-PCNL or FURS, which had a comparable complication profile overall.87

The auxiliary procedure rate in the Zeng et al. study between SMP and FURS was lower in the SMP group,87 whereas the auxiliary procedure rate was comparable for ESWL and FURS but lower for PCNL.86 As stones become larger, the differences are likely to further favour PCNL over FURS, although a smaller track intervention compared with standard PCNL seems reasonable. For example, in patients with stones larger than 20 mm in the lower pole who had mini-PCNL had a significantly lower likelihood of having a postoperative nephrostomy (2.8% vs. 21.1%), a shorter duration of nephrostomy when one was inserted, and a shorter hospital stay than standard PCNL.91

The retreatment rate in ESWL was significantly higher at 30% than FURS 5% or UMP 1.6% and Bozzini et al. also showed significantly higher retreatment rate in ESWL (61.3%) compared to FURS (4.3%) and PCNL (2.2%).85,86

In their study comparing SWL versus FURS versus PCNL for single LPS 1–2 cm, Bozzini found that overall SFR (a composite between a < 3 mm residual fragment and negative urine culture) improved from 61.8% for ESWL to 82.1% for FURS and 87.3% for PCNL.86 Their exclusion criteria differed from PUrE RCT2 by excluding patients with unfavourable lower pole anatomy (patients with a steep infundibular-pelvic angle < 30°, a longer calyx more than 10 mm, or a narrow infundibulum < 5 mm), which may contribute to their improved results. Furthermore, patients with uric acid stones had additional post-operative urinary alkalization with potassium citrate therapy.86

Similarly, Zhang et al. noted a significantly better SFR (defined as no residual stone or < 0.3 cm on computed tomography at 3 months postoperatively) from 73% for SWL to 92% for FURS and as high as 98% for UMP following the treatment of a single 10–20 mm radio-opaque lower caliceal stone.85 Yavuz et al. also (see Figure 47) reported improved SFRs of 90.1% for ultra-mini PCNL and 94.1% and 94% for mini and standard PNL compared with 76% for FURS and 77% for Micro-PCNL.90 Micro-PCNL was also shown to have similar outcomes to FURS by Jiang et al., who reported FURS SFR of 84.5% and 79.3% for micro PCNL.85

In their prospective, multicentre trial including 160 patients with 10–20 mm lower pole renal calculi Zeng et al. evaluated stone-free status defined as residual fragments ≤ 3 mm 1 day postoperatively [on plain kidney ureter bladder (KUB) X-ray and ultrasound] and at 3 months (by CTKUB). The mean stone diameters were similar to those in the PUrE RCT2 trial at 15.0 mm for the SMP patients and 14.3 mm for the FURS patients. SMP achieved a significantly better clearance at day 1 (91.2% vs. 71.2) and at 3 months review (93.8% vs. 82.5%) compared with FURS.87 In contrast, Fayad et al. found stone clearance (defined as ≤ 2 mm on CT at 12 weeks) for lower calyceal stones < 20 mm was not significantly different for mini PCNL (92.7%) and FURS (84.3%).89

The up-to-date meta-analysis showed that two studies also demonstrated that FURS was more expensive to perform than PCNL. Zhang found that FURS was more expensive than PCNL (US$ 4660 vs. 4090) which was related to extra consumable/disposable costs with FURS, whereas much of the equipment for PCNL is reusable.85 Similarly, Yavuz found that the cost of the procedure decreased with increasing invasiveness of treatment, such that FURS was more than twice as costly as ultra-mini, mini, and standard PCNL, with no significant differences between these different PCNL sizes. However, micro-PCNL cost was mid-way between the cost of FURS and the other three percutaneous interventions.90 The cost analysis included in Zhang et al. also found that the total cost of SWL was significantly lower than that of UMP, but that the overall cost of FURS was the highest.85

Our original systematic review and meta-analysis for the PUrE trial design showed that QoL was seldom reported in the studies analysed. This review concluded that information regarding QoL should be an area for improvement in knowledge of treatment outcome and was therefore made a PO in the PUrE RCT. In the PUrE study, the AUC from treatment to 12 weeks and found that QoL for PCNL was marginally superior.

Bosio reported a patient-centred assessment of their procedure and found no significant difference in patient perception of the procedures that 88.2% of patients undergoing ESWL and 87.1% of those undergoing FURS were satisfied with the procedure itself, that 80.9% versus 88.6% were satisfied with the results and 86.8% and 77% of ESWL and FURS patients would choose the same procedure again.80

In our up-to-date systematic review/meta-analysis prepared for this monograph and included as Appendix 10, four RCTs provided some data related to QoL or satisfaction outcomes. In more detail, Bosio et al. (2022) reported no clinically or statistically significant difference in patient satisfaction with the procedures80 that is, similar to the PUrE RCT2 findings. Yavuz et al. (2020) found that return to daily activities was significantly faster for FURS (3.9 days) than standard PCNL (13.5 days).90 Interestingly the recovery time depended on track size, with a significant increase from 4.5 days for Micro PCNL, 6.5 days for ultra-mini PCNL, 9.3 days for Mini PCNL and 13.5 days standard PCNL to return to normal daily activity. Reporting findings that were opposite to our RCT2 data, both Jin and Zeng found that FURS had the advantage over miniaturised PCNL with less postoperative pain.87,92

For larger renal stones from 10 to 20 mm, NICE 2019 recommends FURS or SWL as equivalent options with PCNL the next step if FURS or SWL have failed. In renal stones > 20 mm the NICE 2019 guidelines concur with both the EAU and AUA guidelines that PCNL has the strongest evidence to support it.79 PCNL is therefore the first-line intervention to offer for these larger stones, although FURS should be considered if PCNL is not appropriate.

Strengths of the PUrE study

• RCT1 was a large pragmatic multicentre randomised trial, adequately powered for a difference in health status, making the findings generalisable.
• Although RCT2 did not reach the target sample size, the results indicate sufficient power given the significant marginal differences in health status.
• Both RCTs had a PO which was focused on patient health status and other QoL outcomes, hitherto an under-researched area.
• These were pragmatic trials embedded within current urological practice across the UK; therefore, delivered findings that are relevant to the NHS.
• The randomised treatment allocation, the multicentre pragmatic trial design that facilitates the generalisability of the trial results to individuals routinely treated in the UK NHS, and the prespecified Health Economic Analysis Plan that followed established economic evaluation methods are strengths of the health economic evaluation.
• Provides granular detail on the costs of treatment throughout the patient pathway across both trials.
• Baseline characteristics show that the trial population was broadly similar to that in previously published studies.
• The definition used for SFR was pragmatic and related to additional hospital visits/treatment and was not always imaging-based. These patient-centred outcomes are clear, meaningful, and clinically relevant, based on whether additional treatment was needed (‘unacceptable’), observation was needed (‘acceptable’) or the stone had been sufficiently well treated that no further intervention or observation of any small fragments was deemed necessary (‘complete clearance’).

Limitations of the PUrE study

• Treatment received by participants could not be blinded; however, this reflects the routine use of these interventions in an NHS setting.
• There may have been heterogeneity of clinical practice across the multiple centres which would affect both RCTs.
• The heterogeneity in clinical practice will also mean that the costs of the individual procedures will vary between units, a factor that will need to be considered when applying the outcomes of the study to decision-making in an individual unit basis.
• The trials were completed during the COVID-19 pandemic; while the majority of participants were recruited and completed follow-up prior to the pandemic, recruitment was curtailed, and the study timelines were extended to accommodate the longer waiting times to treatment.
• The differing waiting times for the different treatments in RCT1 may have affected patient and clinician willingness to take part in the study.
• Response rates to health status and QoL questionnaires became lower as their stone episode became less recent and patients recovered from their treatment.
• Variability in imaging used for follow-up, but again, this reflects how decisions are made in the NHS.
• The cost–utility within trial analysis was for 12 months of follow-up, which might be too short to observe all relevant cost and consequences of the treatments to be recorded and therefore a modelling extrapolation was attempted.

Waiting times

The waiting times to intervention, particularly for FURS and PCNL, were longer than expected and did impact on completion of treatment and outcomes within the 12-month post-randomisation period for a small proportion of participants, compounded by the COVID-19 pandemic for the small proportion who were waiting for treatment or were recruited during the pandemic. We think that the long waiting times for FURS and PCNL made it harder for some sites to capture the intervention date in a timely fashion to trigger collection of the PO, with the PO completion rates being lower for those randomised to FURS in RCT1, and for both FURS and PCNL in RCT2 than for ESWL. The longer time from recruitment to intervention may also have had an impact on participant motivation to engage and complete their study questionnaires. Health status and QoL from the time of the decision to treat is what the patient perceives whereas the doctor’s perspective is more likely to be weighted by the apparent patient QoL from the time of treatment itself, it was important to consider both these starting points in our analyses.

Reporting equality, diversity and inclusion

The participant population in both RCT1 and RCT2 were similar in age and sex to the UK population with kidney stones requiring treatment. We optimised participation in the trial through the use of recordings of recruitment appointments in participating sites to feedback to the recruiters’ potential conscious/unconscious biases. A ‘top tip for recruitment’ sheet was generated from this and other such initiatives in similar trials to enable a broad range of participants to be recruited into the trial.

The research team included PPI partners who were actively involved throughout the study contributing to the study design, development of study materials and contributing to discussions of the study results, as well as preparation of the plain English summary. The study team represented a broad range of expertise in quantitative and qualitative methodologies. Less experienced members of the team, including a summer student wanting to gain experience in the clinical trial arena, were encouraged to lead discrete components of the study (under senior supervision) and lead the study SWATs.

Implications for health care

The choice of health status as the PO measure provides important data for patient counselling, assisting joint decision-making. It also allowed for the cost-effectiveness of the treatments for LPSs to be assessed.

Consent/patient expectations and experience

The relatively high baseline health status and low preoperative pain scores for most patients meant that there was relatively little room for improvement following treatment. Indeed, there was a dip in patients’ health status in the first week, seemingly driven by postoperative pain, and this is important for preoperative counselling. Furthermore, the health status and QoL scores returned towards baseline levels at 12 months post intervention, rather than showing a substantial improvement. The treatment of largely asymptomatic LPSs should therefore be considered as a preventative intervention for the future rather than in the expectation of an immediate QoL or healthcare benefit.

It is therefore important in preoperative counselling and patient consent to emphasise that, for patients in whom symptoms are slight, the goal of active stone treatment is to avoid renal/ureteric colic from < 10 mm stones treated by ESWL or FURS in RCT1, or to avoid a more extensive FURS or a PCNL if the stones were to enlarge in the future. For those patients with larger stones, such as those in RCT2, the goal of treatment is to prevent stone growth, with the requirement of a more extensive operation in the future, and/or to try and prevent some of the complications of renal stones, such as infections or loss of renal function. The benefits of these interventions may not be readily appreciated by the patient or the healthcare system, therefore.

The consistent drop in health status at week 1 in both trials, particularly for those in the FURS treatment arms, has important implications for practice during consent to treatment. Based on these data, it is important for clinicians to emphasise that patients are likely to feel worse before they feel better. While this is true of most surgical interventions, and therefore likely to be expected by most patients, the PUrE study provides detailed information regarding the anticipated post-treatment trajectory of QoL that will help make patients’ consent to be more fully informed.

The results from RCT1 shows that there is no significant difference in health status based on treatment allocation. However, in RCT2, the health status over 12 weeks post intervention was better for the PCNL group than the FURS group. Furthermore, when considering secondary outcomes, patients should be counselled that for smaller LPSs < 10 mm, FURS leads to a higher SFR than ESWL, and for LPSs > 10 mm, PCNL leads to higher SFRs than FURS.

Resource allocation including theatre scheduling and inpatient beds

RCT1 demonstrated that ESWL was the more cost-effective treatment for LPSs < 10 mm. This complements the recommendation from NICE Renal Stone Guidelines 2019 that ESWL should be offered as the first-line intervention for renal stones of this size.79

The cost-effectiveness difference between FURS and PCNL is less clear-cut, because of the discrepancy between a micro-costing analysis (which showed PCNL was more cost-effective) or HRG-based analysis (which showed that FURS was more cost-effective).

While data on the time in anaesthetic, theatre and recovery room were meticulously collected and related staff, equipment and disposable costs incorporated in the micro-costing, the approach might have underestimated the cost of the procedures by failing to include cost categories, such as staff being on call when the patient is in the recovery room after surgery or theatre preparation time. While the micro-costing could have underestimated the cost of, for example, PCNL, it is unclear why the HRG cost associated to PCNL (LB75 ‒ Elective inpatient; £8285) is almost two and a half times the HRG cost allocated to FURS (LB65 ‒ Day case; £3350) when the time to complete the procedures and the length of hospital stay are not substantially different between the two procedures.

More transparency in the way the HRG costing are estimated and the reasons behind procedures being associated with corresponding HRG codes and cost would be beneficial for decision-making.

Implications for research

Further research is needed to answer the following research questions:

What is the patient preference in terms of therapy and what level of patient tolerability and acceptance does each treatment modality enjoy?

What effect will suction devices, improvements in laser technology, and intraoperative pressure monitoring have on postoperative pain, QoL, SFRs, complications and costs of FURS?

What is the effect of miniaturisation of PCNL on postoperative pain, length of stay, complications, SFRs and costs?

What is the clinical and cost-effectiveness of full metabolic assessment compared with standard advice alone, in people who have undergone treatment with LPSs?

Conclusion

Taken together, these trials have shown that ESWL, FURS and PCNL (including smaller track varieties) are all safe and effective in the treatment of < 10–25 mm lower pole kidney stones but have various attributes that require the ‘pros and cons’ to be evaluated for individual patients.

Extracorporeal shockwave lithotripsy is less invasive, and less expensive than FURS or PCNL, but has a lower SFR, and consequently higher retreatment rate than these surgical interventions. Operation time is shorter for PCNL than FURS but is associated with a longer hospital stay and more complications.

The PUrE study shows that ESWL for the treatment of lower renal pole stones under 10 mm was more cost-effective than FURS because it was cheaper and there was no meaningful difference in patient health status when comparing the two. SFRs were higher, however, with FURS. From an overall NHS perspective, the costs savings of treating all patients with these stones with ESWL as their first treatment would be substantial. For larger stones, 10–25 mm, PCNL gave better health status over 12 weeks post intervention and SFRs than FURS and was more cost-effective on a micro-costing basis.