Detection of Indinavir Crystals in Urine Dependence on Method of Analysis Glen L. Hortin, MD, PhD; Christine King, MT, MS; Kirk D. Miller, MD; Jeffrey B. Kopp, MD

● Objectives.—To determine the frequency of crystalluria in patients treated with the human immunodeficiency virus protease inhibitor indinavir and to compare methods of detecting crystalluria. Methods.—A total of 308 freshly voided urine specimens from 168 patients treated with indinavir were evaluated by manual microscopy of sediment and microscopy with an automated workstation and by dipstick analysis. Results.—Crystals were detected in 22%, 31%, or 32% of specimens using, respectively, an automated workstation, manual microscopy, or both methods. Proteinuria or hemo-

globinuria occurred significantly more often in specimens with (28%) than without (18%) crystals. Frequency of crystalluria was unrelated to specific gravity, but it increased at higher pH. Crystals were detected in 21% of specimens with pH less than 6 and 42% of specimens with pH of 6 or higher. Conclusions.—Crystalluria occurs in more than 30% of urine specimens from patients treated with indinavir, but detection rates vary substantially with method of analysis. Manual microscopy detected crystalluria 41% more often than did an automated workstation. (Arch Pathol Lab Med. 2000;124:246–250)

I

vir-associated nephrolithiasis—50%—recently has been noted for hemophiliacs with HIV infection,28 suggesting that underlying liver disease may raise the risk for nephrolithiasis. Urinary crystals and renal stones in patients treated with indinavir consist primarily of precipitates of indinavir base.5,7,8,11 This has been established by infrared spectroscopy,7,8,14,21 chromatographic analysis,5,7,16 and mass spectrometry.5,7,16 About one fourth of indinavir stones are reported to consist of mixed deposits of indinavir with calcium oxalate or calcium phosphate.7 As a consequence of their composition, most indinavir stones are not detected by radiologic studies without use of contrast dyes, and structural analysis of stones by x-ray examination is not informative. Fourier-transform infrared spectroscopy can identify indinavir as a constituent of stones; previous reports have published the infrared absorption spectrum of indinavir.7,8 In routine clinical practice, urine crystals provide less material for analysis than do stones, and identification of the composition of crystals usually relies on crystal morphology and clinical history of indinavir use as in the present study. An issue that has not been addressed by the many studies of indinavir-associated nephrolithiasis is whether the frequency of detecting indinavir crystalluria depends on the method of urinalysis. Microscopic analysis of urine is not a well-standardized procedure. Many laboratories perform microscopic analysis of urine using automated workstations that assist with the selection of images of particles in urine passing through a flow cell.30–32 Other laboratories manually examine varying dilutions of urine sediment, volumes of sediment, and numbers of fields of view with or without the use of stains or polarization. Some laboratories perform microscopic analysis only on specimens that have abnormalities on dipstick analysis.33–36 Comparisons of automated, manual, and prescreening methods of

nhibitors of the human immunodeficiency virus (HIV) proteinase have become important elements in combination drug treatment regimens of HIV infection.1,2 One of the most widely used drugs in the class of HIV inhibitors is the sulfate salt of indinavir (Crixivan). Its pharmacokinetics have been investigated extensively.2–4 The drug has a relatively short serum half-life of 1.5 to 2 hours. The usual adult dose of indinavir is 800 mg every 8 hours. Treatment usually suppresses HIV infection, but occasional adverse effects have been noted, including nephrolithiasis, hepatitis, lipodystrophy, and hyperglycemia.1,2 Occurrence of urinary tract stones has been the most common major adverse effect. Although clearance of indinavir is considered to be primarily via metabolism by cytochrome P450 3A4, about 20% of an oral dose is cleared by the kidneys.1–4 Most of the indinavir in urine occurs as parent drug, although several metabolites also are present.3 Indinavir has a low solubility at neutral pH that can lead to precipitation within the urinary tract. There have been many reports of the occurrence of kidney stones, renal failure, interstitial nephritis, dysuria, and asymptomatic crystalluria attributed to precipitation of indinavir in the urinary tract.5–29 These reports have noted about a 10% to 20% frequency of crystalluria and an incidence of kidney stones of about 4%, which may be increased in a setting where dehydration is promoted by high ambient temperatures.12 Also, a much higher incidence of indinaAccepted for publication August 3, 1999. From the Clinical Pathology Department (Dr Hortin and Ms King) and Critical Care Medicine Department (Dr Miller), Warren Grant Magnusson Clinical Center, and Metabolic Diseases Branch, National Institute of Diabetes and Digestive and Kidney Diseases (Dr Kopp), National Institutes of Health, Bethesda, Md. Reprints: Glen L. Hortin, MD, PhD, Clinical Pathology Department, National Institutes of Health, Bldg 10, Room 2C-407, 10 Center Dr, Bethesda, MD 20892-1508. 246 Arch Pathol Lab Med—Vol 124, February 2000

Indinavir Crystals in Urine—Hortin et al

urinalysis have been published regarding detection of frequent microscopic elements in urine such as red blood cells, white blood cells, and bacteria.30–36 However, there are few data regarding the relative sensitivity of detecting crystals by different methods of urinalysis. The detection of crystals usually is difficult to study because of its low frequency and heterogeneity of crystal types. We prospectively compared the ability of 2 common methods for microscopic analysis of urine to detect crystalluria in patients treated with indinavir and sought to gain information about the clinical significance of this laboratory finding. MATERIALS AND METHODS Patients with HIV infection receiving treatment with indinavir were identified through the hospital information systems and by test orders specifically for microscopic examination for indinavir crystals. Dosage was 2400 mg of indinavir daily in 3 divided doses. A total of 308 specimens from 168 patients were examined. Urine specimens were well mixed and split between the automated workstation (Yellow IRIS Model 500, International Remote Imaging Systems, Chatsworth, Calif), which stains specimens with crystal violet, and the manual analysis of sediment with the Kova system (Hycor Biomedical, Garden Grove, Calif) without staining. The counting algorithm of the automated workstation was adjusted to provide optimal sensitivity for crystal detection. The number of high-power frames examined by the automated workstation was increased from 75 to 2500 to increase the sensitivity for crystal detection. Each specimen was analyzed at 1003 and 4003 magnification. A volume of 7 mL of unspun urine was analyzed by the automated workstation. For manual microscopy, a volume of 6 mL was centrifuged at 4003g for 5 minutes. Volume was decreased from the usual 12 mL volume to allow for adequate specimen volume for comparative analyses. A volume of 5 mL of supernatant was decanted and used for dipstick analysis. The sediment was resuspended in a volume of 1 mL urine and was placed in Kova Glasstik slides. Microscopic examination was with a Nikon Labophot-2 microscope (Nikon Corp, Tokyo, Japan). The entire well of the microscope slide was examined under low power (1003). The number of crystals per low-power field of manual inspection was recorded as rare when there were 0 to 1, few for more than 1 to 3, moderate for more than 3 to 4, and many for more than 4 crystals per low-power field. Using the workstation, specimens were classified as having rare crystals when there were 1 or fewer, few for 1 to 2, moderate for more than 2 to 4, and many for 4 or more crystals per edit screen. Indinavir crystals were in the size range of 27 to 103 mm or larger on the workstation screens and were displayed on the low-power field 1, 2, or 3 edit screens. Specimens were classified as negative by either method when no crystals were found. Indinavir crystals were identified by having a morphology consistent with the description by Kopp et al5 under brightfield and polarized microscopy. Urine dipstick analysis was performed using Chemstrip 10 UA test strips and the Chemstrip Urine Analyzer (Roche-Boehringer Mannheim Corp, Indianapolis, Ind). Results for pH, urine protein, and hemoglobin were recorded. Dipstick results for pH are recorded in only 5 steps for a pH of 5, 6, 6.5, 7, and 8. Specific gravity was measured by a harmonic oscillation method on the automated workstation.

RESULTS Crystals were detected in 100 (32%) of 308 urine specimens by either manual microscopy or the urine workstation (Table 1). Manual examination had higher sensitivity (P , .01 by x2 analysis), detecting 97 of the 100 positive specimens vs detecting 68 of the positives with the workstation. Typical appearance of crystals using manual microscopic examination and the urine workstation is shown in the Figure. Panels A and B of the Figure show radial Arch Pathol Lab Med—Vol 124, February 2000

Table 1. Comparison of Detection of Urine Indinavir Crystal Detection by Manual Microscopy or Automated Workstation Automated

Manual Negative Positive

Negative

Positive

208 (68%) 32 (10%)

3 (1%) 65 (21%)

clusters of crystals forming a starburst shape often by brightfield (A) or polarizing (B) microscopy of urine sediment. Using the urine workstation (Figure, C), indinavir crystals were usually observed as individual needles, and clusters of crystals were rarely observed. Passage of urine into the flow cell may disrupt clusters of crystals, the clusters may be excluded from passage into the flow cell of the urine workstation, or centrifugation for manual analysis may promote cluster formation. Comparison of the semiquantitative results from the 2 methods of analysis (Table 2) showed a high correlation between methods, but the manual analysis had a greater sensitivity and a bias toward higher semiquantitative scoring. The workstation found crystals in 3 specimens, all with rare crystals, that were not detected by manual analysis, while manual analysis detected crystals in 32 specimens identified as negative by the workstation, and the quantity of crystals in these specimens ranged from rare to many. Manual analysis had a higher semiquantitative scoring for 48 specimens vs a higher scoring by the urine workstation for 6 specimens. The higher scoring by manual analysis may reflect the urine concentration by centrifugation prior to manual examination. The average specific gravity of specimens with crystals was less than that of specimens without crystals (Table 3). This is surprising as more concentrated urine specimens should contain a higher concentration of indinavir. The only possible linkage of specific gravity to crystalluria noted was a higher specific gravity for specimens with large numbers of crystals. Urine pH appeared to be a significant factor in the occurrence of crystals (Table 4). Crystals occurred twice as frequently in specimens with a pH of 6 or higher than in specimens with a pH less than 6. Also, the semiquantitative scoring of crystalluria was higher in specimens with pH of 6 or higher, with only 17% of positive specimens classified as rare crystals vs 48% of positive specimens with pH less than 6 being classified as having rare crystals. There did not appear to be a major difference in the rates of crystalluria in specimens with pHs ranging from 6 to 8. Crystals appeared in 40%, 43%, 48%, and 33%, respectively, of specimens with pHs of 6, 6.5, 7, and 8. Occurrence of crystals was not randomly distributed among different subjects (Table 5). Among 78 subjects with more than 1 specimen tested, if there was a negative result on the first specimen, there was an 82% probability that the second specimen would be negative. If the first result was positive, there was a 52% probability that the second test would be negative. These figures are significantly different (P , .01) than the average probability of 73% for a negative result on the second test. The tendency of some subjects to have recurrent crystalluria could lead to a bias for a high rate of positive crystal findings in this study if a greater number of repeat analyses for these subIndinavir Crystals in Urine—Hortin et al 247

Table 2. Correlation of Quantitative Results for Manual Microscopy and the Automated Workstation Automated

Manual None Rare Few Moderate Many

None

Rare

Few

Moderate

Many

208 18 8 3 3

3 5 7 2 0

0 0 15 4 1

0 0 1 12 2

0 0 2 0 14

Table 3. Relationship of Occurrence of Crystals to Urine Specific Gravity* No. of Specimens

Specific Gravity (Mean 6 SD)

Negative

208

1.0177 6 0.0066

Positive All Rare Few Moderate Many

100 26 33 21 20

1.0156 1.0154 1.0144 1.0151 1.0183

Crystals

6 6 6 6 6

0.0060 0.0063 0.0053 0.0061 0.0063

* Crystals were detected by either manual analysis or with an automated workstation. Semiquantitative results are from the manual analysis, except for 3 additional specimens with rare crystals detected by the workstation.

Table 4. Relationship of Crystalluria to pH* pH Crystals

Negative

5

$6

111

97

Positive Rare 14 12 Few 12 21 Moderate 3 18 Many 0 20 All 29 71 (%) (21) (42) * Numbers of specimens in pH ranges with our without crystals. Semiquantitative results were from manual analysis except for 3 specimens with rare crystals found with the automated workstation alone.

Table 5. Consistency of Results for Repeated Specimens Examined for Indinavir Crystals* First Specimen Negative

Microscopic appearance of indinavir crystals by brightfield (A) or polarizing (B) microscopy of urine sediment or using an automated urine workstation (C). Photos were taken at 1003 magnification.

248 Arch Pathol Lab Med—Vol 124, February 2000

Positive

Second specimen Negative 45 12 Positive 10 11 * Results from analysis of the first specimen were compared with analysis of a second specimen on the same patient at a later date, usually 1 to 3 months later. For results above with 78 data subjects: If first result is negative, then 82% chance next will be negative. If first result is positive, then 52% chance next will be negative. Overall, 73% of second results were negative.

Indinavir Crystals in Urine—Hortin et al

Table 6. Relationship of Indinavir Crystalluria to Positive Dipstick Reactions for Protein or Hemoglobin or to Increased Number of White Blood Cells (.10 per Low-Power Field) Dipstick No. Negative

No. (%) Positive

No. (%) With Increased White Blood Cells

No crystals

170

38 (18)

4 (2)

Crystals All Rare Few Moderate Many

72 22 21 15 14

28 4 12 6 6

(28) (15) (36) (29) (30)

10 4 5 1 0

(10) (15) (15) (5) (0)

jects were obtained. However, it does not appear that greater numbers of specimens were analyzed for crystalluria-prone subjects. Reviewing only a single analysis for each of the 168 subjects found a rate of crystal detection of 33%, almost identical to the overall rate of detection of 32% when multiple samples are included for some subjects. Positive dipstick reactions for urine protein or hemoglobin were detected for 28% of specimens in which crystals were observed and for 18% of specimens without crystals (Table 6). Rates of dipstick positivity were about the same for samples with rare crystals (15%) as for specimens without crystals and were 36%, 29%, and 30% for specimens with few, moderate, or many crystals, respectively. An increased number of white cells (.10 per low-power field) was noted for 2% of specimens without crystals and for 10% of specimens with crystals (Table 6). There was substantial overlap between specimens positive by dipstick and white blood cell evaluation. Two of 4 specimens with increased white blood cells and no crystals had dipstick abnormalities, and 8 of 10 with increased white blood cells and crystals had positive dipstick findings. The frequency of increased white blood cells did not increase with higher numbers of crystals observed, but the sample size is relatively small. COMMENT It has been difficult to compare the sensitivity of urine crystal detection by different methods of analysis because crystals are a relatively infrequent finding in urine and there is a diverse range of types of crystals. In an unselected group of more than 13 000 urine specimens, the frequency of crystalluria was less than 2%.36 Another smaller study found crystalluria in 9% of fresh specimens from patients with urinary stones and 2% of a control population37 and emphasized the importance of timely analysis of specimens; rates of crystal detection increased severalfold when samples were stored 24 hours before analysis. Specimens for the present study were analyzed within 1 to 2 hours of collection, so that delays in analysis did not contribute to high rates of crystal detection. Patients treated with indinavir serve as a population with a high frequency of crystalluria of a uniform type, and specimens from these patients serve as an excellent test sample for comparing methods of crystal detection. Results here provide evidence that manual analysis of a urine sediment can detect about 40% more specimens with indinavir crystals than can an automated urine workstation. There are Arch Pathol Lab Med—Vol 124, February 2000

several reasons why the workstation could have lower sensitivity. First, the workstation possibly samples a smaller total volume of specimen than is reviewed by manual microscopy after centrifugation to concentrate particulate matter. Also, in the workstation, urine must pass through a narrow flow cell, and some larger crystals may be too large to enter the flow cell. In particular, clusters of crystals as observed by manual microscopy may not enter the flow cell, and this may account for detection only of individual needlelike crystals with the workstation. Another factor is software that determines the number of fields of view or the number of particles examined and selection of the particle sizes for examination. Instrument settings of the workstation were adjusted in the current study to try to achieve maximal crystal detection by examining a large number of fields. Our conclusion is that manual microscopy of urine sediment is the preferred method for detecting indinavir crystals, and this may give a general reflection about the comparative ability of methods to detect other types of crystals as well. Microscopic examination of urine is a relatively poorly standardized procedure, and our findings illustrate 1 example how methodologic differences in performing this examination can have substantial effects on findings. Also, morphology of crystals is a nonstandardized and imperfect method for classification of crystals, and indinavir crystals may be mixed with other types such as calcium oxalate crystals. The method of choice for definitive identification of indinavir as a component of crystals or urinary tract stones is infrared or Fourier-transform infrared analysis.7,8,14,21 The frequency of detecting crystals in urine from patients treated with indinavir was 32% when both methods of analysis were applied. This is higher than in most previous reports5–8 but represents a minimal frequency for the true occurrence of crystalluria because even application of the duplicate analysis may not be 100% sensitive. The sensitivity of the manual analysis was 97% or less and of the urine workstation was 68% or less. False-negative specimens by the duplicate analysis are likely to be few in number and to represent specimens with very few crystals, and that may have decreased clinical significance. Several factors relating to the development of crystalluria were examined. It was expected that crystalluria would be more frequent in highly concentrated urine reflected by increased specific gravity. There has been a previous suggestion that deposition of indinavir stones in the kidney may be more common in hot climates where dehydration may be more common.12 However, in our study the average specific gravity of specimens with crystals was less than that of specimens without crystals. This is unexpected, but it may be that patients with dysuria increase their fluid intake. The only linkage noted between urine concentration and crystals was that specimens with large numbers of crystals had a higher specific gravity than specimens with low numbers of crystals. Urine pH appeared to be a more important factor than specific gravity. Specimens with a pH less than 6 had half the frequency of crystalluria and had fewer crystals than did specimens with a pH of 6 or higher. This may relate to the increased solubility of indinavir at lower pH. Solubility of indinavir at pH 5 is 0.1 mg/mL, which is about 3-fold higher than the solubility at pH of 6 or higher.4 For subjects with multiple specimens, there was a tendency for some subjects to have positive results and others to have negative results at a greater frequency than exIndinavir Crystals in Urine—Hortin et al 249

pected by random probability. This agrees with a previous study that noted that among 70 asymptomatic patients treated with indinavir and having multiple urine specimens examined, 53 never had crystals, 14 had crystals intermittently, and 3 always had crystals.5 This may reflect subjects who have higher or lower than average excretion of unmetabolized drug. Usually, renal excretion of the parent drug is a minor pathway for clearance of indinavir,3 but this may become proportionately more important if primary routes of hepatic metabolism are impaired. Recent finding of a high rate of indinavir-related nephrolithiasis in the subset of patients with HIV infection and hemophilia, who also have a high risk of hepatitis C infection, provides evidence for higher risk of nephrolithiasis in patients with hepatic injury.28 Semiquantitative description of the number of indinavir crystals in a specimen primarily appeared to be useful for distinguishing specimens with rare crystals from those with larger numbers of crystals. Rare crystals in urine may have less potential for injury to the urinary tract as these specimens showed no increase in the frequency of dipstick abnormalities vs specimens without crystals. Little difference was noted in the rates of dipstick findings between specimens with few, moderate, or many crystals, however. There are limited data about the clinical significance of indinavir crystalluria. In a previous study, indinavir crystals were present in urine from all of 13 patients with symptomatic flank pain and evidence of intrarenal indinavir deposition, while it was observed in only 20% of all subjects on indinavir.5 When there are symptoms such as flank pain, the finding of indinavir crystalluria therefore contributes evidence that there is renal deposition of indinavir. However, it is not clear whether the finding of crystals in urine from asymptomatic patients is likely to presage patients who are at high risk of developing stones. The linkage of crystalluria to other types of kidney stones has been controversial.36–39 Crystalluria is noted in some subjects without kidney stones, and crystalluria is not always observed in subjects with stones. However, detection of crystalluria, particularly crystals other than calcium oxalate or calcium phosphate, probably should be considered an indicator of increased risk of nephrolithiasis. Higher rates of crystalluria or occurrence of larger crystals and clustering of crystals have been associated with nephrolithiasis.36 The finding of a more than 50% increase in the occurrence of positive dipstick reactions for blood and protein in specimens with crystals vs those without crystals suggests that, in some but not most instances, the crystals contribute to renal injury. Further study is needed to determine whether detection of crystals predicts risk of kidney injury. References 1. Panel on Clinical Practices for Treatment of HIV Infection. Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents. MMWR Morb Mortal Wkly Rep. 1998;47:40–82. 2. Deeks SG, Smith M, Holodniy M. HIV-1 protease inhibitors. JAMA. 1997; 277:145–153. 3. Balani SK, Ariso BH, Mathai L. Metabolites of L735,524, a potent HIV-1 protease inhibitor, in human urine. Drug Metab Dispos. 1995;23:266–270. 4. Lin JH, Chen I-W, Vastag KJ, et al. pH-dependent oral absorption of L735,524, a potent HIV protease inhibitor, in rats and dogs. Drug Metab Dispos. 1995;23:730–735.

250 Arch Pathol Lab Med—Vol 124, February 2000

5. Kopp JB, Miller KD, Mican JAM, et al. Crytalluria and urinary tract abnormalities associated with indinavir. Ann Intern Med. 1997;127:119–125. 6. Tashima KT, Horowitz JD, Rosen S. Indinavir nephropathy. N Engl J Med. 1997;336:1138–1140. 7. Daudon M, Estepa L, Viard JP, Joly D, Jungers P. Urinary stones in HIV-1positive patients treated with indinavir. Lancet. 1997;349:1294–1295. 8. Trainor LD, Steinberg JP, Austin GW, Solomon HM. Indinavir crystalluria: identification of patients at increased risk of developing nephrotoxicity. Arch Pathol Lab Med. 1998:122:256–259. 9. Gagnon RF, Tsoukas CM, Watters AK. Light microscopy of indinavir urinary crystals [letter]. Ann Intern Med. 1998;128:321. 10. Berns JS, Cohen RM, Silverman M, Turner J. Acute renal failure due to indinavir crystalluria and nephrolithiasis. Am J Kidney Dis. 1997;30:558–560. 11. Gentle DL, Stoller ML, Jarrett TW, Ward JF, Geib KS, Wood AF. Protease inhibitor-induced urolithiasis. Urology. 1997;50:508–511. 12. Bach MC, Godofsky EW. Indinavir nephrolithiasis in warm climates [letter]. J Acquir Immune Defic Syndr Hum Retrovirol. 1997;14:296–297. 13. John H, Muller NJ, Opravil M, Hauri D. Indinavir urinary stones as origin of upper urinary tract obstruction. Urol Int. 1997;59:257–259. 14. Sutherland SE, Reigle MD, Seftel AD, Resnick MI. Protease inhibitors and urolithiasis. J Urol. 1997;158:31–33. 15. Ascher DP, Lucy MD. Indinavir sulfate renal toxicity in a pediatric hemophiliac with HIV infection. Ann Pharmacother. 1997;31:1146–1149. 16. Bruce RG, Munch LC, Hoven AD, et al. Urolithiasis associated with the protease inhibitor indinavir. Urology. 1997;50:513–518. 17. Grunke M, Valerius T, Manger B, Kalden JR, Harrer T. Renal dysfunction in a human immunodeficiency virus-infected patient who was treated with indinavir. Clin Infect Dis. 1997;25:1270–1271. 18. Rich JD, Ramratnam B, Chiang M, Tashima KT. Management of indinavir associated nephrolithiasis. J Urol. 1997;158:2228. 19. Chen SC, Nankivell BJ, Dwyer DE. Indinavir-induced renal failure [letter]. AIDS. 1998;12:440–441. 20. Martinez F, Mommeja-Marin H, Estepa-Maurice L, et al. Indinavir crystal deposits associated with tubulointerstitial nephropathy. Nephrol Dial Transplant. 1998;13:750–753. 21. Lerner LB, Cendron M, Rous SN. Nephrolithiasis from indinavir, a new human immunodeficiency virus drug. J Urol. 1998;159:2074. 22. Blake SP, McNicholas MM, Raptopoulos V. Nonopaque crystal deposition causing ureteric obstruction in patients with HIV undergoing indinavir therapy. AJR Am J Roentgenol. 1998;171:717–720. 23. Vigano A, Rombola G, Barbiano di Belgioioso G, Sala N, Principi N. Subtle occurrence of indinavir-induced acute renal insufficiency [letter]. AIDS. 1998;12: 954–955. 24. Witte M, Tobon A, Gruenenfelder J, Goldfarb R, Coburn M. Anuria and acute renal failure resulting from indinavir sulfate induced nephrolithiasis. J Urol. 1998;159:498–499. 25. Clayman RV. Crystalluria and urinary tract abnormalities associated with indinavir. J Urol. 1998;160:633. 26. Marroni M, Baburri M, Mecozzi F, Baldelli F. Acute interstitial nephritis secondary to the administration of indinavir [letter]. Ann Pharmacother. 1998;32: 843–844. 27. Noble CB, Klein LT, Staiman VR, New N, Hensle TW, Berdon WE. Ureteral obstruction secondary to indinavir in the pediatric HIV population. Pediatr Radiol. 1998;28:627–629. 28. Brodie SB, Keller MJ, Ewenstein BM, Sax PE. Variation in incidence of indinavir-associated nephrolithiasis among HIV-positive patients. AIDS. 1998;12: 2433–2437. 29. Hanabusa H, Tagami H, Hataya H. Renal atrophy associated with longterm treatment with indinavir. N Engl J Med. 1999;340:392–393. 30. Deindoerfer FH, Gangwer JR, Laird CW, Ringold RR. ‘‘The Yellow IRISTM’’ urinalysis workstation—the first commercial application of ‘‘automated intelligent microscopy.’’ Clin Chem. 1985;31:1491–1499. 31. Elin RJ, Hosseini JM, Kestner J, Rawe M, Ruddel M, Nishi HH. Comparison of automated and manual methods for urinalysis. Am J Clin Pathol. 1986;86:731– 737. 32. Williams KH. Comparison of automated and manual urine microscopy. Am J Clin Pathol. 1991;96:421. 33. Loo SYT, Scottolini AG, Luanphinith S, Adam AL. Performance of a urinescreening protocol. Am J Clin Pathol. 1986;85:479–484. 34. Wenz B, Lampasso JA. Eliminating unnecessary urine microscopy. Am J Clin Pathol. 1989;92:78–81. 35. Morrison MC, Lum G. Dipstick testing of urine—can it replace urine microscopy? Am J Clin Pathol. 1986;85:590–594. 36. Widran S, Kyle J, Diendoerfer F, Koontz A. A retrospective nationwide comparison of effectiveness of reagent test strip screening for urine sediment abnormalities [abstract]. Clin Chem. 1996;42:S153. 37. Abdel-Halim RE. Crystalluria and its possible significance. Scand J Urol Nephrol. 1993;27:145–149. 38. Fogazzi GB. Crystalluria: a neglected aspect of urinary sediment analysis. Nephrol Dial Transplant. 1996;11:379–387. 39. Robert M, Boularan A-M, Delbos O, Monnier L, Grasset D. Evaluation of the risk of stone formation: study on crystalluria in patients with recurrent calcium oxalate urolithiasis. Eur Urol. 1996;29:456–461.

Indinavir Crystals in Urine—Hortin et al