Creatinine

Spinal cord injury

One study (n=86) investigated the effectiveness of office ultrasonography of the bladder and kidneys to provide routine urological follow-up in the outpatient spinal cord injury clinic ²¹⁴.

106 scans were performed on 86 asymptomatic spinal cord injury patients. Of the patients, 68 had a blinded excretory urography for comparison, including 20 who underwent additional studies (computerised tomography scans of the abdomen and pelvis, and/or radiologist-performed ultrasound examination of the kidneys and bladder). Office ultrasound detected 5 of 6 kidney stones, 6 of 6 hydronephrotic kidneys, 5 of 7 renal masses (4 of 6 cysts and 1 of 1 renal tumour), 3 of 3 bladder stones and 3 of 3 bladder diverticula. Subtle changes of chronic renal infection noted on excretory urography in 4 patients were not detected on corresponding ultrasound scans but voiding cystourethrograms revealed no reflux, and comparison to prior studies confirmed that these renal units were stable ²¹⁴.

One study (n=54) compared ultrasound findings with those obtained from excretory urogram (IVP) and/or voiding cystourethrogram in spinal cord injury patients ²¹⁵. Kidneys: For 15/54 there were concerns regarding renal abnormalities based on the excretory urogram (IVP). Of these 15 patients ultrasound confirmed the radiographic findings in five (two with renal calculi, one with chronic pyelonephritis, one with peripelvic cyst and one with focal pyelonephrtitis), ruled out questionable radiographic findings in six and revealed abnormalities not present radiographically in four (one with renal cyst, one with hydronephrosis, one with cortical atrophy and one with renal calculi). Ureters: Of the 15 patients in whom the ureters were examined nine had different degrees of vescioureteric reflux on voiding cystourethrography, which was confirmed by ultrasound in five (56%) and not demonstrated in four. The remaining 6 patients had ureterctasis on an IVP, which was confirmed by ultrasound in two (33%) and not noted successfully in 4. In two patients with a known allergy to the contrast medium ultrasound demonstrated vesicoureteral reflux in one, and hydroureter and hydronephrosis in one. Bladder: The bladder was examined in 32 patients during ultrasound voiding cystourethrography but was imaged adequately in only 30. Ultrasound confirmed the positive radiographic findings in 23 (six with bladder calculi, three with trabeculated bladders and 12 with normal bladders), ruled out questionable radiographic findings in three and yielded additional information in four (one with bladder calculi, two with lithogenic bladder sediment and one with calcific crust on the Foley catheter balloon) ²¹⁵.

One study (n=162) reported on the results of a comparison between renal ultrasound (RUS) and renal nuclear scans (RNS) as part of upper tract surveillance in spinal cord injury patients ²¹⁶.

Only the results of the renal ultrasound scan are reported here. A RUS scan was judged to be positive if it demonstrated any degree of caliectasis or pyelocaliectasis; parenchymal disease; or the presence of complex cysts, calculi, solid masses, or other renal and/or peri-renal processes. Simple renal cysts were not considered an abnormality because they did not dictate any change in patient management. RUS abnormalities were found in 57/162 patients (35.2%). Of the 75 positive ultrasound studies, 39 were positive for hydronephrosis, 39 revealed parenchymal disease, 22 revealed renal stones, and 8 revealed solid renal mass (renal malignancy found in 2 of these 8 patients). Many ultrasounds had more than one pathologic finding ²¹⁶.

One study (n=109) reported on the diagnostic accuracy of ultrasound and radioisotope renography compared to intravenous urography to detect hydronephrosis in patients with spinal cord injury ²²⁶.

Of 235 kidneys studied, 43 kidneys in 23 patients showed hydronephrosis on the final findings. The estimated prevalence was 21% (23/109) in the study. The diagnostic accuracy of sonography and renal ultrasound are summarised in table 2.

Table 2

Diagnostic accuracy of ultrasound and radioisotope renography compared to intravenous urography.

One study (n=100) reported on the findings from routine radiological surveillance in patients with spinal cord injury ²¹⁷. In paraplegics, 26/47 patients had abnormalities (upper tract changes, calculi, bladder abnormalities, persistent post-voidal residual urine > 100 ml) detected on routine radiological screening. 24/26 abnormalities were detected 0 to 10 years after the injury compared with only 2/26 after 10 yrs of injury. For tetraplegics, 35/50 abnormalities were detected. All of these were detected within 10 yrs after the injury ²¹⁷.

One study (n=75) reported on patients with spinal paralysis who had undergone intravenous urography (IVU) and renal ultrasonography as part of routine assessment of the upper urinary tract ²²⁰.

The results are presented in table 111, table 112 and table 5 below.

Table 111

Normal IVU and abnormal ultrasound.

Table 112

Abnormal IVU and normal ultrasound.

Table 5

Abnormalities demonstrated by IVU and also indicated or shown by ultrasound.

One study compared Kidney, Ureter, Bladder (KUB) radiography with ultrasound in 100 consecutive patients with spinal cord injury ²²⁵. A total of 199 kidneys and 99 urinary bladders were examined. On average, less than 50% of the renal area and about 70–75% of the urinary bladders were visualised. Five patients had renal stones identified on KUB radiograph, and of these two were seen on ultrasound. There were no stones seen on ultrasound only. Ultrasound identified renal abnormalities in a further 14 patients. There were seven patients with renal scarring in eight kidneys. There were five patients with hydonephrosis in six kidneys; all cases were mild to moderate. There were two patients with a small kidney with thinned cortex. The KUB identified none of these patients. Ultrasound identified a number of other abnormalities. There was one patient with a duplex renal collecting system, one case of nephrectony, one case of adrenal myolipoma, one situs inversus, one case of abnormally high echogenicity of the liver and two cases of gallstones. In one of these an additional gallbladder polyp was seen. One of the cases of gallstones was also identified on the KUB; all other abnormalities were not seen on the radiographs. Abnormalities of the urinary bladder were seen in 20 cases. A total of 19 cases showed evidence of bladder wall hypertrophy, and one case of incomplete bladder emptying. There was one case of previous cystectomy and a neobladder. KUB did not identify any of the abnormalities. Therefore, apart from the renal stones and one patients with gallstones, KUB did not identify any of the other abnormalities seen on ultrasound ²²⁵.

One study (n=108) reported on patients who underwent ultrasound who had no urinary symptoms compared with patients who had urinary symptoms ²²⁷.

In the asymptomatic group no abnormalities were reported in 63 patients. The following findings were reported in 24 patients (Table 113).

Table 113

Ultrasound findings in asymptomatic patients.

There were 21 spinal cord injury patients who exhibited urinary symptoms (passing purulent urine, temperature, rigors, passing blood in urine, severe kidney/bladder pain, recurrent urine infections) when they underwent ultrasound examination of the urinary tract. Abnormalities such as hydronephrosis, pyonephrosis, bladder calculi, or bladder polyp were detected in 20 of 21 patients and, subsequently, all 20 patients required therapeutic intervention on the basis of ultrasound findings ²²⁷.

Spina bifida

One study (n=25) reported ultrasound on children with myelodysplasia with normal urodynamics at birth ²²⁴. The mean follow up was 9.1 yrs (range 1 to 18.6 yrs). No child had hydronephrosis or reflux ²²⁴.

One study (n=40) reported on ultrasound and serum creatinine in adults with spina bifida who were using clean intermittent catheterisation ²²¹. In patients with normal ultrasound and normal serum creatinine (1.5 mg/dl), there were no individuals (0/20) whose average catheterised volume corresponded to a bladder pressure of >40 cm H₂O on cystometry. However, in patients with hydronephrosis and/or elevated creatinine, 30% (6/20) had average catheterised volumes corresponding to a bladder pressure of >40 cm H₂O ²²¹.

Spinal cord injury

One study (N=59) reported on the results of an annual health maintenance evaluation to include cystoscopy on patients with spinal cord injury who were continuously catheterised for 10 more years, or were smokers and catheterised for 5 or more years ²²⁹ Ninety three bladder biopsies and 18 urine cytologies were obtained, none of which demonstrated malignant changes. No bladder cancers were diagnosed through screening. During the same six year period four spinal cord injury patients were diagnosed at the hospital with bladder cancer, all outside of the surveillance protocol ²²⁹.

One study ²³⁰ reviewed all 14 SCI patients diagnosed with squamous cell bladder cancer during a 16 year period and divided them into those who had 1) been given annual cytoscopy for at least 10 years and were asymptomatic when diagnosed with cytoscopy (n=5), and 2) those who had not been routinely screened and were symptomatic at presentation (n=14). Those receiving annual screening had a higher cancer-specific survival (100% compared to 50% in the symptomatic group). In addition, the stage of disease was less advanced at presentation in the screened group. In the non-screened group pathological stage was more advanced at diagnosis, with 7 patients at stage pT3a or pT3b, 1 with pT1N0M0, and 1 at stage pT2N0M0. By contrast in the screened group 3 patients were at pT1n0M0, 1 was at pT3aN0M0 and 1 had pT3bN0M0 disease. However these differences were not statistically significant.

Spinal cord injury

One study (n=160) reported that there were no significant differences between patients with spinal cord injury who had missed two or more consecutive annual examinations compared with patients who were compliant with their annual examinations on mean-adjusted Effective Renal Plasma Flow (ERPF) (left kidney 311 vs 308 mL/min, right 301 vs 276; ns) ²²⁸.

Base case results

The base-case analysis, using probabilistic data (Table 119) showed that over a ten year period the least costly monitoring strategy was the strategy that monitored low risk patients in the MS population (strategy 3a). This strategy however is only monitoring low risk patients; the high risk patient population was considerably more expensive, almost double the cost. The lowest cost strategy that considered a mixed population was strategy 2, in spinal cord injury patients. If we consider strategy 4 separate due to it being in a paediatric population, the most costly strategy is strategy 1 for general neurogenic lower urinary tract disorders. The probabilistic analysis enables us to fit confidence intervals around both the costs and the difference in costs. It shows when each of the strategies are compared to strategy 1 - as it is the most commonly followed guideline - the average differences are significant at the p=0.05 level in all strategies apart from 4b. For the low risk strategies strategy 3a remains the lowest cost and for the mixed and high risk strategies, strategy 2 is the least costly.

Table 119

Base-case results (Probabilistic).

As most of these strategies are for quite heterogeneous populations considering them together carries heavy limitations. Therefore considering the absolute costs is more informative. Threshold Analysis

In order to account for the lack of data on outcomes, a threshold analysis was carried out that calculated the incremental QALYs that each strategy would have to generate in order for them to be cost effective compared to “do nothing.” The incremental QALYs per patient that would be required to make a strategy cost effective are calculated at two thresholds, £20,000/QALY and £30,000/QALY. It is possible to see from the results in Table 120 that the more expensive strategies would have to generate more additional QALYs compared to ‘do nothing’ in order to account for the increased cost.

Table 120

Threshold Analysis on number of incremental QALYs needed for a strategy to be cost effective compared to ‘do nothing’.

Summary of results

The probabilistic base-case results show that among non-paediatric strategies, strategy 1 is the highest cost strategy at every horizon period, apart from year 1. When comparing strategies in the low risk population, strategy 3a emerges as the least costly, while in high risk populations, strategy 2 is the least costly.

Limitations and interpretation

The results obtained in this analysis give an indication of the cost of monitoring strategies over a given time period (10 years as base case). The absolute cost of monitoring strategies over 10 years ranges between £800 and £3000 per patient depending on risk, age and condition. The cost of paediatric monitoring is considerably high, particularly in the high risk population. Even in the low risk group, paediatric monitoring has a similar cost to the strategy for high risk patients with Multiple Sclerosis (MS). The GDG believed that the strategy in the paediatric population overstated the importance of imaging and monitoring.

Strategy 1 was the most costly strategy when considering a mixed population with mixed risk levels. This strategy was also the most costly at different time horizons, different frequencies of monitoring and different risk profiles. It was the opinion of the GDG that the regular urodynamics and physical examination in a specialist urological centre determined the high cost of this strategy. This was considered to be an over-use of specialist, invasive testing.

The least costly strategy when considering a mixed population with mixed risk levels was strategy2, while in a low risk population the least costly was strategy 3, which was the lowest cost strategy overall. The GDG noted that this shows that the risk profile is important when defining the monitoring strategy for a patient. A clear definition of high and low risk is crucial and it has been described elsewhere in the guideline (See Introduction).

No clinical outcome associated with any of the monitoring strategies was available, so it is not possible to conclude which is the most cost effective strategy. Another limitation of our analysis is that it does not consider the inevitable extra or unnecessary treatment associated with the monitoring strategies. As in any screening programme, the more often tests are done the more likely it is that false positives results will be picked up requiring an unnecessary treatment. This adds to the cost and impacts treatment effectiveness and patient quality of life. A further limitation is that the strategies are themselves based on guideline recommendations that are largely consensus driven.

In cases where the guidelines were unclear on the testing frequency, assumptions were made by the GDG. These assumptions were, however, tested extensively in probabilistic and deterministic sensitivity analyses. A further point to make is that all the populations for which the strategies are recommended are different. This limits the validity of comparisons between the strategies but not the validity of the absolute costs.

Generalisability to other populations/settings

The analysis was conducted from a UK perspective using: one international, one European, one US and one Scottish guideline. The strategies also made recommendations in different populations: general neurogenic incontinence patients, MS patients, Spinal cord injury patients and paediatric patients.