Crystalluria and Urinary Tract Abnormalities Associated with Indinavir

Methods

Patients

We collected information on all adults (n = 240) who received indinavir while participating in clinical studies of HIV infection sponsored by the National Institute of Allergy and Infectious Diseases and the Critical Care Medicine Department at the Warren Grant Magnuson Clinical Center of the National Institutes of Health (NIH) between October 1994 and December 1996. The mean duration of indinavir therapy was 30 weeks. None of the patients received more than the recommended 2400 mg of indinavir per day in three or four divided doses. To evaluate the frequency of asymptomatic crystalluria, we obtained urine specimens from a subgroup of these patients (n = 142) who did not have symptoms and were available to provide urine during the study period. We also obtained urine specimens from 40 patients who did not have symptoms and were not receiving indinavir.

Laboratory and Radiographic Studies

Microscopic urinalyses were done in a renal clinical research laboratory by using standard methods, including a modified Sternheimer-Malbin stain (Sedi-Stain, Becton Dickinson, Franklin Lakes, New Jersey). Urinary crystals were examined by using a microscope equipped with crossed polarizing filters and a λ plate (Leica, Deerfield, Illinois).

Urine specimens containing abundant crystals and renal stones were washed with water, dissolved in acetonitrile, and analyzed by high-performance liquid chromatography (HPLC) using a reverse phase C-18 column; elution peaks were compared with those produced by the free-base form of indinavir (Merck & Co., Rahway, New Jersey). Mass spectrometry was performed by using fast atom bombardment and an SX102 mass spectrometer (JEOL, Peabody, Massachusetts). Fragmentation patterns were obtained by doing B/E linked scan analysis.

Indinavir crystals were produced in vitro at room temperature from a solution of indinavir sulfate in phosphate-buffered saline at a pH of 8.0. Urine specimens that contained grossly apparent crystals were monitored microscopically as they were acidified by the addition of 1 molar of hydrochloric acid.

Computed tomography (CT) was done using 5-mm collimation on General Electric HiSpeed helical CT scanners (GE Medical Systems, Milwaukee, Wisconsin).

Statistical Analysis

The Fisher exact test was used to compare the proportion of patients who had crystalluria and were receiving indinavir with the proportion of patients who had crystalluria and were not receiving indinavir. This test was also used to compare the proportion of indinavir recipients who had urinary crystals with the proportion of indinavir recipients who did not have urinary crystals. The means for these patients were compared by using the Mann-Whitney test (for duration of indinavir therapy) and the Student t-test (for urinary pH).

Results

Symptomatic Patients

Among 240 patients treated with indinavir, we identified 19 patients with symptomatic urinary tract disease (Table 1). These patients had CD4+ lymphocyte counts ranging from 25 to 903 cells/mm³ (median, 472 cells/mm³), and none reported having nephrolithiasis before beginning indinavir therapy. Symptoms first developed after 1 to 47 weeks (mean ±SD, 18 ± 14 weeks) of indinavir use. Although some overlap was seen, three clinical syndromes could be distinguished: nephrolithiasis with frank renal colic, flank pain or back pain without evidence of renal stones, and dysuria or urgency.

Seven of the 19 symptomatic patients (patients 1 to 7) presented with renal colic; 6 of these patients passed a stone or gravel or had radiographic evidence of a stone (Figure 1). Seven others (patients 8 to 14) presented primarily with dull back or flank pain; 2 of these patients also had dysuria. Six had contrast-enhanced CT; 4 had bilateral filling defects in the renal parenchyma (Figure 2). Some defects were purely cortical, some were purely medullary, and others were full-thickness corticomedullary lesions. Three patients with abnormal scans had repeated CT after indinavir was withheld for 7 days (patient 8) and 25 days (patients 9 and 10); these scans showed partial alleviation or complete resolution of the lesions. The remaining 5 patients (patients 15 to 19) presented primarily with lower urinary tract symptoms of dysuria or urgency. One of these patients also had mild backache. These symptoms were frequently associated with the presence of opalescent urine and full-field crystals (Figure 3).

Thirteen of the 19 symptomatic patients were available to provide a urine specimen during the acute episode, and all had urinary crystals. Three other patients had urinary crystals at a subsequent clinic visit while they were still receiving indinavir. The remaining 3 patients did not have urinalysis while receiving indinavir. Serum creatinine levels were mildly and transiently elevated (range, 1.4 to 2.5 mg/dL) in 3 patients with stones and 2 patients with parenchymal lesions.

Of the 19 symptomatic patients, 18 improved with hydration alone or hydration and withdrawal of indinavir. One patient (patient 2) had cystoscopy for placement of a temporary ureteral stent, after which gravel was passed.

Sixteen of the symptomatic patients restarted indinavir therapy after the acute episode: Five had recurrent back or flank pain, and 4 had recurrent dysuria. Urine specimens were available from 6 of these 9 patients, and all had recurrent crystalluria. We also found recurrent crystalluria in 3 patients whose symptoms did not recur. Indinavir therapy was permanently discontinued in 4 patients because of recurrent urologic symptoms.

Eight symptomatic patients were receiving interleukin-2 intermittently. The acute symptoms were temporally remote from the administration of interleukin-2 in six patients. One patient (patient 5) developed symptoms within 24 hours of receiving interleukin-2, and another (patient 13) developed symptoms on the fifth day of interleukin-2 therapy.

Identification of Crystals

The urinary crystals that we observed were distinctive tapered or flat rectangular plates, typically with slightly irregular margins and a pattern of internal layering (Figure 4 A), and circular or bilaterally symmetrical fan-shaped and starburst forms (Figure 4 B). In some cases, the plate forms were seen alone; in others, both structures were present, sometimes in large aggregates (Figure 4 C). These crystals showed birefringence with straight extinction (that is, the colors of the crystals disappeared when the polarizers were oriented in the north-south and east-west positions) and were length negative (that is, the slow polarized light ray was perpendicular to the crystal axis) [9]. The size of the crystals, particularly the plate forms, varied widely.

Pelleted crystals from 10 patients (7 symptomatic patients and 3 asymptomatic patients) and stones from 2 patients were further characterized. In all specimens, HPLC showed elution peaks identical to those produced by authentic indinavir, and mass spectrometry indicated a molecular mass of 613 and a molecular formula and fragmentation pattern identical to those of authentic indinavir [10]. Indinavir crystals in urine at a pH of 7.0 became soluble in vitro upon acidification to a pH less than 4.5.

Rectangular, plate-like crystals morphologically identical to those found in patients' urine specimens could be grown in vitro in a buffered saline solution (Figure 4 D).

Asymptomatic Crystalluria

One hundred forty-two asymptomatic patients who were receiving indinavir submitted one or more urine specimens for examination. Twenty-nine of these patients (20%) had indinavir crystals in at least one specimen. For comparison, urine specimens were obtained from 40 patients who were not receiving indinavir, and none of these specimens contained similar crystals (P < 0.001). Nineteen of these 40 patients subsequently received indinavir, after which 3 developed crystalluria. Approximately 25% of these asymptomatic patients with indinavir crystals had crystalluria to a degree similar to that shown in Figure 3. Although crystals were evident in some specimens immediately after the specimens were voided, the crystals became visible in other specimens only as the urine cooled to room temperature. In some patients, crystals were evident only microscopically.

A range of possible risk factors for asymptomatic crystalluria was considered, but no factors were identified (Table 2). When multiple urine specimens were available, we analyzed the data from the first urine specimen containing crystals for patients with crystalluria and the last urine specimen for patients without crystalluria. This was done to demonstrate that the two groups were followed for similar periods of indinavir treatment. Although 4 (14%) of the patients with crystalluria and 14 (12%) of the patients without crystalluria were participating in clinical trials of interleukin-2, none of the asymptomatic patients was receiving interleukin-2 when urine specimens were collected.

Seventy of the 142 asymptomatic patients receiving indinavir submitted a urine specimen on more than one occasion (mean ±SD, 2.6 ± 0.9 specimens; range, 2 to 6 specimens). Of these patients, 53 did not have crystalluria in any specimens, 3 had crystalluria in every specimen (2 patients submitted two specimens and 1 patient submitted three specimens), and 14 had crystalluria in some specimens (10 patients submitted two specimens, 2 patients submitted three specimens, and 2 patients submitted four specimens). Five of the indinavir recipients who were asymptomatic when they submitted their first urine specimen eventually developed symptoms. One of these presymptomatic specimens contained crystals, and four did not.

Discussion

We identified distinctive urinary crystals in patients receiving indinavir. These crystals are birefringent and have plate and starburst structures unlike the structures known to be associated with other endogenous or exogenous substances [11, 12]. The identification of the crystals as indinavir was suggested by the strong association between the occurrence of these crystals and indinavir use and was confirmed by HPLC, mass spectrometry, and the growth of similar crystals in vitro.

We also report that indinavir crystalluria is associated with a spectrum of urologic findings. Symptomatic urinary tract disease developed in 19 (8%) of our patients who received indinavir. Seven patients (3%) had signs or symptoms of nephrolithiasis; this is similar to the frequency of nephrolithiasis previously reported among indinavir users [5, 6]. Although the renal colic was severe, all patients but one promptly responded to conservative therapeutic measures.

Seven patients had a novel syndrome that consisted primarily of back or flank pain with crystalluria but without nephrolithiasis. Six of the seven patients had CT, and four had renal parenchymal defects. We believe that this syndrome is due to crystals within the distal renal tubule, where urinary concentration promotes sludging. It is not known whether asymptomatic indinavir recipients or patients with symptoms other than flank or back pain develop similar parenchymal lesions, nor do we know whether some patients who develop this complication subsequently develop chronic deficits in renal function.

Five patients had another previously unreported syndrome that consisted primarily of dysuria and urgency with crystalluria. We believe that these symptoms are caused by dense crystalluria that irritates the mucosa of the bladder and urethra. This entity can easily be misdiagnosed as acute infectious urethritis and thus can lead to the unnecessary use of antibiotics.

Asymptomatic crystalluria was present in 20% of patients receiving indinavir. The appearance of crystalluria was not predictive of the subsequent development of symptoms and is therefore not an indication for withdrawal of indinavir. On the other hand, the presence of crystalluria is confirmatory that the symptoms are probably due to indinavir and suggests the need for increased fluid intake and possibly temporary drug withdrawal.

Indinavir in its native free-base form is poorly soluble in aqueous solutions [13]. The pharmaceutical preparation is admixed with H (₂) SO₄ to enhance solubility in the gastrointestinal tract. Cumulative urinary excretion of intact indinavir represents approximately 12% of the administered dose [14]. Limited solubility coupled with the relatively high proportion of renal excretion probably promotes crystal and stone formation. The manufacturer recommends a fluid intake of at least 1.5 L daily [6]. Although dehydration was not apparent in the symptomatic patients that we evaluated, it is possible that reduction in fluid intake contributes to crystal formation. Acid solutions are expected to promote the solubility of indinavir, and we found that the crystals dissolve upon in vitro urinary acidification to a pH less than 4.5, which cannot be achieved in vivo. We have no information on the efficacy of in vivo urinary acidification to a lesser degree in preventing crystal formation or in treating acute symptomatic crystalluria.

Renal stones are most commonly caused by the crystallization of endogenous substances in urine [15, 16]. Although certain drugs can be constituents of renal stones, drugs are not a major cause of nephrolithiasis [17]. Indinavir is exceptional in this regard, as has been shown by previous reports and by our experience [5, 6]. Crystalluria may also be caused by drugs, including some medications commonly taken by HIV-infected patients. Nevertheless, the structure of the crystals that we observed was unlike the structures associated with the use of vitamin C, acyclovir, or sulfa-containing drugs [18-20]. We cannot rule out the possibility that other, less frequently used medications or dietary components may produce a milieu that fosters the formation of indinavir crystals. Indinavir may act as both a substrate for and an inhibitor of the hepatic cytochrome P-450 enzyme system. This provides the potential for numerous drug–drug interactions, some of which might increase serum concentrations-and therefore urine concentrations-of the drug [21].

Tashima and associates [7] recently described a patient who had back pain and renal insufficiency while receiving indinavir. Renal biopsy showed interstitial nephritis and plate-like crystals within cortical and medullary collecting ducts. The presence or absence of urinary crystals was not reported. Renal function [unlike that in our patients] remained abnormal after indinavir therapy had been discontinued. This case report provides evidence to show that indinavir crystals form in the distal nephron, although the precise relation between 1) the crystals and interstitial inflammation seen in this patient and 2) the radiographic findings suggesting crystal sludging and reversible renal dysfunction in our patients is unknown.

Although our patients represent a select group referred to the NIH, we believe that our findings can be generalized to other clinical settings. We know of no factors that should have particularly predisposed our patients to indinavir crystalluria. Other than indinavir, interleukin-2 was the only investigational agent used, and we found no temporal relation between interleukin-2 administration and urinary tract symptoms and no increased risk for asymptomatic crystalluria in patients receiving interleukin-2.

Of the four currently approved protease inhibitors, indinavir has been the most widely used because of its efficacy, its favorable pharmacologic properties, its tolerability, and the relatively few drug–drug interactions associated with it. Our findings do not substantially alter this profile. Symptoms, when they develop, can usually be treated conservatively with hydration and drug withdrawal, and many patients restart indinavir therapy. Nonetheless, it is important that physicians who treat HIV-infected patients with indinavir be aware of the possible spectrum of urologic complications. When symptoms do occur, this information should allow formulation of a narrowed differential diagnosis and development of an appropriate therapeutic plan.