Radiographic and Laboratory Data (“the Megaprofile”) Can Accurately Guide Medical Management in the Absence of Stone Analysis

Introduction

Targeted therapy for the prevention of stone recurrence has been demonstrated to be effective. ^1,2 The medical evaluation of nephrolithiasis has traditionally included stone analysis, most commonly determined via infrared spectroscopy or X-ray diffraction. However, in this era of lithotripsy, stone analysis is not always available. In patients undergoing shockwave lithotripsy (SWL), postoperative straining of the urine to capture fragments may be unsuccessful. Ureteroscopic techniques such as stone dusting may deter stone analysis and lead to fragments that are passed but not collected. ³ Another clinical situation frequently encountered by the modern urologist is the incidental discovery of renal calculi in asymptomatic patients undergoing cross-sectional imaging for unrelated reasons. Many of these patients may be appropriate candidates for medical rather than surgical therapy but may not have stone analysis.

Even when stone analysis is available, there is evidence to suggest that these data may not always be accurate. ^{4 –7} Krambeck and colleagues reported significant variation in the compositions reported by five different commercial laboratories who were sent similar specimens. ⁶ These inaccuracies were particularly common when stones contained more than one component, as is the case in the majority of stones. ⁷

While experts have raised concerns about the accuracy of stone analysis, fortunately there are other tools to predict stone composition and aid in the application of appropriate treatment regimens for stone patients. For example, numerous authors have advocated the use of Hounsfield Unit Density (HUD) to predict either stone composition or response to surgical treatments, such as SWL. Thus, radiographic data, combined with pertinent historical and laboratory information, could effectively obviate the traditional need for stone analysis.

We hypothesize that appropriate therapeutic recommendations could be made in the absence of stone analysis when imaging, medical history, and laboratory data were available.

Methods

After obtaining institutional review board approval, a data set of stone formers from our multidisciplinary stone-prevention clinic was queried to identify 100 patients who had undergone chemical stone analysis (Dianon Systems, CT) and had a congruent 24-hour urine collection (within a mean of 3 months) that was deemed accurate using accepted 24-hour urinary creatinine excretion ranges. In an effort to test our hypothesis across a broad range of stone types, uric acid and struvite stone formers were deliberately “over-represented” (12% and 9%, respectively), beyond what might be seen in a typical population of stone formers in the United States. Cystine stone formers were excluded. For the purpose of constructing the study cohort, stone formers were categorized by our research team as calcium oxalate, calcium phosphate, uric acid, or struvite based on a single analysis (Table 1). This classification scheme was used only in the patient selection and data set construction process to ensure adequate representation of less-common stone components and was not used for determining therapeutic recommendations.

Once patients were identified, our research team composed a “Megaprofile” for each, consisting of age, gender, body mass index (BMI), select comorbidities, serum chemistries, and 24-hour urine studies. The 24-hour urine measurements included were urine volume, pH, calcium, oxalate, uric acid, citrate, sodium, magnesium, phosphorus, potassium, creatinine, and sulfate, as well as calculated superaturations of calcium oxalate, sodium urate, brushite, and uric acid. Serum chemistries, from as close as possible to the date of each patient's stone composition testing, included calcium, parathyroid hormone, uric acid, and glucose whenever available. Imaging information was also included for each patient. In particular, radiographic data, including stone size, number, and Hounsfield density (from computerized tomography scans [CT]), were recorded. When available, plain abdominal radiographs and/or CT scout images were reviewed for stone radio-opacity that would suggest a uric acid stone. Attenuation:size ratios were calculated to help predict stone composition, as previously described. ^{8 –12} The specific medical diagnoses that were included were gout, diabetes mellitus, recurrent urinary tract infections (UTIs), and conditions of rapid gastrointestinal transit, such as inflammatory bowel disease, significant bowel resection, and gastric bypass.

Each Megaprofile was then reviewed twice, in random order, by a panel of 3 urologists who had not participated in the cohort selection process. In the first “round,” the panel was blinded to the stone analyses when they reviewed the Megaprofile and made medical treatment recommendations on the basis of medical history, labs, and imaging data. Recommendations were generated as a group in a process similar to our stone clinic, and a consensus was reached in all cases. Nutritional recommendations were made to address specific 24-hour urinary risk factors and included the following: increasing fluid intake for low urine volumes (<2 L); limiting sodium intake if urinary calcium excretion was high, especially if urinary sodium excretion was also relatively high; and strategies to reduce urinary oxalate excretion if it was high. Limited meat, fish, and poultry intake was recommended in patients with hyperuricosuria (>600 mg/day), especially if urinary sulfate excretion was also relatively high. Medications that were recommended included potassium citrate, thiazides, allopurinol, and antibiotics. Potassium citrate was recommended to enhance urinary citrate excretion in patients with hypocitraturia (<320 mg/day) and to alkalinize the urine of those with hyperuricosuria or acidic urine (pH<6) plus suspected uric acid nephrolithiasis. Thiazides were prescribed for hypercalciuria (>250 mg/day) if, by reviewing other urinary parameters, diet was not suspected to be the primary contributor. Allopurinol was reserved for cases of very high urinary uric acid or when elevated serum uric acid was documented. An extended course of antibiotics, with the recommendation that they be culture specific, was advised for struvite stone formers (SFs) postoperatively.

After the first round of recommendations were made using only Megaprofile data, a repeat evaluation was performed, again in random patient order, this time with stone analyses data available to the panel. The first-round treatment recommendations were then compared with those made using stone composition data (considered the “gold standard”). Characteristics were recorded concerning patients who received treatment that was judged as incorrect after stone composition was known.

Results

The primary stone types of the 100 patients included in our investigation are listed (Table 1). Ninety-four percent of patients had mixed stone composition; five patients had 100% uric acid stones and one was 100% calcium oxalate. Fifty-four of the patients were men while 46 were women. Mean patient age was 51±14 years. Mean BMI was 30.2±7.1 kg/m². The mean maximal diameter of the largest stone measured (some patients had multiple stones) was 12±10 mm. Stone attenuation was measured in the 90% of patients who had a CT scan for review; the mean attenuation was 724±389 Hounsfield Units (HU). The mean attenuation:stone diameter ratio was 89±64 HU/mm. When assessing for comorbid conditions associated with stone formation, we found that 2% of patients had a history of gout, 15% had diabetes mellitus, 6% had gastrointestinal disease, and 19% had a history of recurrent UTIs. Thirty-seven percent of patients were obese (BMI>30).

Ninety-five percent of patients had at least one risk factor identified on their 24-hour urine analysis, though the sole factor identified was low urine volume (<2 L per day) in 17%. Most patients had multiple risk factors identified and these are listed (Table 2).

When stone composition was unknown, the panel recommended targeted nutrition therapy in 91% of cases, which remained unchanged once composition was revealed. Thus dietary recommendations were 100% accurate using Megaprofile data alone (Table 3). Medications were prescribed in 68% of cases when stone composition was unknown. In seven patients (7%), pharmacotherapy was changed after stone composition was known. In 21% (4/19) of patients with recurrent UTIs, knowledge of stone composition altered therapy. Three patients with calcium stones and recurrent UTIs were overtreated with antibiotic therapy, based on the incorrect supposition that they were struvite stone formers. The other metabolic risk factors in these patients (low urine volume, mild hyperoxaluria, and severe hypocitraturia, respectively) were treated appropriately, so the error was one of overtreatment with antibiotics that might have been unnecessary. In two patients with hypercalciuria, the panel suspected calcium nephrolithiasis and treated with thiazides, but the stone composition data revealed uric acid. While both patients had a “normal” urinary uric acid, they had acidic urine, which may have resulted in the uric acid crystals precipitating out of solution. It was thus reasoned that these patients required urinary alkalinization, which was not prescribed prior to knowledge of stone composition. Interestingly, the CT scan attenuation values of the stones in these patients were suggestive of uric acid nephrolithiasis (450 and 350 HU), suggesting a particularly valuable role for radiographic data when laboratory values seem contradictory. Finally, the last error was in a patient with hyperoxaluria (81 mg/day) and recurrent UTIs. The panel suspected a mixed calcium oxalate and struvite stone and prescribed antibiotics (correct) and also calcium and pyridoxine supplementation. The stone composition revealed a mixed struvite and calcium phosphate stone. Without an oxalate component to the patient's risk profile, the calcium and pyridoxine may be viewed as overtreatment.

Overall, therapy based on the Megaprofile without stone composition data was appropriate 93% of the time.

Discussion

The importance of medical management for stone disease has been established over several decades, and the use of biochemical data including stone composition data to support therapeutic recommendations is historical. The number of diagnostic tests that must be ordered and interpreted for metabolic evaluation has been described by some as time consuming, and costly. ¹³ Efforts to reduce costs and simplify treatment have been made and have resulted in “simplified” and “comprehensive” evaluation approaches, ^13,14 including the decreased use of tests such as the calcium “fast and load” test and bone densitometry.

The traditional metabolic evaluation of stone formers has relied on the determination of the patient's stone type. In recent years, the availability of stone composition has been reduced, and the accuracy of some reports has been questioned. ^{4 –6} Whereas traditional surgical techniques involved removing intact calculi, which could then be sent for analysis, the advent of lithotriptic techniques has resulted in more fragments than whole stones. The likelihood of missing components is inherently increased, potentially leading to misleading conclusions. Moreover, patients sometimes do not form the same stone types or they have variable components in repeated stone events, rendering knowledge of past stone composition of limited value in predicting new-stone formation. Finally, in some cases, obtaining stone material for chemical analysis may be confounding, such as in patients passing minute stone fragments and in those discovered incidentally to have renal stones and who do not require surgical intervention.

With improvements in CT imaging technology, there is new information for the physician to use in evaluating and counseling patients. In many cases, CT data, such as stone density/attenuation and attenuation:size ratio, accurately predict stone composition. ^{8 –12} This information, along with 24-hour urine studies as in our Megaprofile, can allow for appropriate treatment recommendations in the absence of stone analysis in most cases. Our study validates that this is possible 93% of the time in a typical cohort of patients with a history of urolithiasis. In this way, appropriate directed medical therapy may be feasible more quickly, just after 24-hour urine and lab data are obtained.

Finally, medical treatment regimens are similar across multiple stone types, so knowledge of composition often does not alter treatment. For example, SFs with low urine volume may be safely recommended to increase their fluid consumption, regardless of their stone type. Similarly, low urine pH may be treated in the typical calcium or uric acid SF with potassium citrate. Just as thiazides are prescribed for patients with hypercalciuria without the use of the calcium fast-and-load test to differentiate renal leak from absorptive hypercalciuria, so may other therapies be targeted to risk factors by using imaging and other data without knowing the exact chemical stone analysis.

As in any study, our investigation has several limitations. One of these is the retrospective nature, which was designed as such to ensure that a wide variety of stone types were included. Also, in order to assess the applicability of Megaprofile to more rare stone types (uric acid and struvite), these were deliberately over-represented. The included profiles may not therefore match what is seen in the general population. The patients included were mostly referred to our stone clinic and thus could have had more “metabolically active” stone disease than first-time SFs or those not otherwise referred for metabolic evaluation. This could have introduced some selection bias. Finally, a limited dataset (the Megaprofile) including a single 24-hour urine study was purposefully used to make treatment recommendations in this investigation. Having access to additional data, such as multiple 24-hour urinalyses over time, may result in superior therapeutic recommendations and would allow adjustments to treatment plans.

One outlying subpopulation of patients was those with a history of recurrent UTIs, for whom our recommendations were altered by stone analysis (4/19; 21%). Specifically, three patients with calcium stones and recurrent UTIs were prescribed an extended course of antibiotics because they were incorrectly assumed to have infection-related stones (though interestingly, one of these patient's next stone analysis did show struvite). While incorrectly treating stone disease with antibiotics may not be an egregious error, the contrary case, failing to prescribe them to a struvite stone former, as occurred in the fourth case, may be problematic. Although struvite stone formers may have additional stone risk factors that should be managed, the importance of sterilizing the urinary tract after stone removal, often with a multimonth course of antibiotics, is unique to these patients. ^15,16 Thus, in stone patients with UTIs, we recommend stone analysis. In the absence of stone composition data, obtaining urine culture data may be particularly helpful as the presence of urea-splitting organisms may suggest struvite stone composition. Megaprofile-based therapy was 96% accurate in patients without a history of recurrent UTI.

Conclusion

Medical, laboratory, and radiographic data used in the Megaprofile can be used to accurately prescribe medical and nutritional therapy for calcium, uric acid, and struvite stone formers in the absence of chemical stone analysis 93% of the time. Stone formers with recurrent UTIs benefit from stone analysis to identify and appropriately treat any infectious component. By doing this, the Megaprofile serves a powerful resource for managing the vast majority of patients without waiting for stone analysis. Future investigations might consider assessing outcomes and stone recurrence rates in patients medically managed without the use of stone composition data and see how their outcomes compare to traditional evaluation.