Risk Factors for Urological Symptoms in a Cohort of Users of the HIV Protease Inhibitor Indinavir Sulfate: The ATHENA Cohort

Background  Nephrolithiasis is a well-known complication of indinavir treatment and may result in urological symptoms ranging from renal colic to renal insufficiency.

Objective  To obtain further knowledge regarding the incidence and risk factors of urological symptoms associated with indinavir sulfate use.

Methods  This study was performed in the ATHENA (AIDS Therapy Evaluation National AIDS Therapy Evaluation Centre) cohort of patients infected with human immunodeficiency virus (HIV) receiving antiretroviral therapy in the Netherlands. The incidence rate of urological symptoms was assessed in a subcohort of 1219 patients starting HIV protease inhibitor treatment after 1996. Urological symptoms were defined as an initial report of nephrolithiasis, renal colic, flank pain, hematuria, renal insufficiency, or nephropathy. Using multivariate Cox regression analysis, risk factors for urological symptoms during indinavir treatment were subsequently studied among the subset of 644 patients who started indinavir treatment after 1996.

Results  The incidence of urological symptoms was 8.3 per 100 treatment-years for indinavir vs 0.8 per 100 treatment-years for other HIV protease inhibitors. Risk factors for urological symptoms during indinavir treatment were low weight (relative risk [RR], 2.1; 95% confidence interval [CI], 1.1-3.9), low lean body mass (RR, 1.7; 95% CI, 1.0-2.9), undetectable HIV-1 RNA when starting indinavir treatment (RR, 3.2; 95% CI, 1.5-6.0), prior treatment change because of intolerance (RR, 2.4; 95% CI, 1.2-5.1), indinavir regimens of 1000 mg or more twice daily (RR, 3.1; 95% CI, 1.3-8.2), and warm environmental temperatures (RR, 3.9; 95% CI, 1.7-8.8). Risk estimates were highest among patients with a low lean body mass.

Conclusion  Increased alertness for urological symptoms is warranted for patients starting indinavir treatment, particularly among those with a low lean body mass, during indinavir regimens of 1000 mg or more twice daily, and in warm weather environments.

ADVERSE EFFECTS of HIV protease inhibitor–containing highly active antiretroviral treatment constitute a threat to the quality of life and treatment adherence of patients infected with human immunodeficiency virus (HIV).^1-3

Nephrolithiasis is a well-known complication of treatment with indinavir sulfate, 20% of which is cleared by the renal system.^4-6 Symptoms range from renal colic, flank pain, dysuria, and gross hematuria to the passing of a kidney stone. Other symptoms related to the renal effects of indinavir use are crystalluria and renal insufficiency, which may go undetected without specific monitoring. Symptoms may occur early or late during treatment and occasionally occur after treatment discontinuation.^7-12 The incidence of urological symptoms among patients taking 800 mg of indinavir 3 times daily was estimated at 4% in clinical trials and 8% to 16% in other studies.^4,9-13

Urological symptoms are generally assumed to be secondary to sludging of indinavir crystals in the urinary tract, probably because of the low solubility of indinavir in aqueous conditions.^14,15 Therefore, the major risk factor for urological symptoms seems to be insufficient fluid intake, which leads to concentrated urine and a diminished urinary flow, thus facilitating indinavir precipitation and crystal aggregation.^4,16 Nevertheless, abundant fluid intake is not always sufficient for the prevention of urological symptoms, which suggests that other factors play a role as well.¹²

Warm environmental temperatures, female sex, concurrent hepatitis C infection, and high indinavir plasma concentrations have also been suggested as risk factors for urological symptoms.^8,9,13,17-20 Indinavir plasma concentrations depend on the indinavir dosing regimen used, gastric pH, distribution volume, and hepatic and renal clearance.^5,6,21-23 Any factors affecting these aspects are potential determinants for the development of urological symptoms.

We performed a population-based cohort study to estimate the incidence of urological symptoms among patients taking HIV protease inhibitors as part of highly active antiretroviral treatment and to identify potential risk factors for urological symptoms during indinavir treatment. In particular, we studied the role of various indinavir dosing regimens, environmental temperatures, and lean body mass (LBM).

We conducted a study within the ATHENA (AIDS Therapy Evaluation National AIDS Therapy Evaluation Centre) cohort. This is a large national cohort of patients infected with HIV in the Netherlands who have been or are being treated with at least one of the antiretroviral drugs that were introduced after July 1996 (all HIV protease inhibitors, all nonnucleoside reverse transcriptase inhibitors, and the newer nucleoside analogue reverse transcriptase inhibitors such as lamivudine and stavudine) and who gave written informed consent. Patient entry began in May 1998 and continues to date. All 22 Dutch hospitals providing treatment to patients infected with HIV participate in the ATHENA project. The project has been approved by the ethics committees of all participating centers.

According to the national guidelines for HIV treatment in the Netherlands, patients are seen at approximately 3-month intervals for regular follow-up.²⁴ Data for the ATHENA cohort are collected from the medical records on standardized forms by trained research nurses and treating physicians. This is performed retrospectively for the period before cohort entry and prospectively thereafter and continues to date. The resulting database contains information on sex, age, route of HIV transmission, height, and weight. Start and stop dates and dose frequency of any antiretroviral medication or prophylactic treatment against opportunistic infections, reasons for stopping such treatments, date of onset and resolution of HIV-related diseases, CD4 cell counts, plasma HIV-1 RNA load, and abnormal laboratory values are recorded in a standardized manner. Database information on adverse events comprises all events that lead to a change in antiretroviral treatment and several specified adverse events, including nephrolithiasis. In addition, physicians are requested to report otherwise remarkable events and abnormal laboratory values. On-site data monitoring of at least 10% of completed study forms by central data monitors takes place at regular intervals. In addition, central data verification through automated database consistency checks is performed, and any resulting queries are resolved by local research nurses. Data that had been entered, monitored, and verified before March 1, 2000, were available for the present analysis.

On the basis of the treatment data in the ATHENA database, we selected 2 cohorts. The first cohort comprised patients commencing treatment with any of the HIV protease inhibitors and was used to estimate the incidence rate of urological symptoms during use of different HIV protease inhibitors. Patients were eligible for entry if they began taking the first HIV protease inhibitor no earlier than 1997, when the HIV protease inhibitors became available as prescription drugs through community pharmacies in the Netherlands. Patients entering the cohort were followed up until development of urological symptoms, death, loss to follow-up, or the date of last data collection, whichever came first. Cohort participants were allowed to change HIV protease inhibitors and to have treatment interruptions during the follow-up. The duration of use of each HIV protease inhibitor was calculated. Consequently, each person could contribute to multiple protease inhibitor exposure categories.

Because urological symptoms occur primarily in indinavir users, we formed a second cohort to further identify incidence rates and risk factors for urological symptoms during indinavir treatment. These patients first began indinavir treatment in 1997 or thereafter, independent of previous HIV protease inhibitor and previous urological symptoms. Follow-up lasted from the start of indinavir treatment until the development of urological symptoms, discontinuation of indinavir use, an interruption of indinavir use for more than 7 days, death, loss to follow-up, or the date of last data collection, whichever was earliest.

Patients for whom we had inadequate information regarding treatment or who had urological symptoms at cohort entry were excluded from both cohorts.

A range of clinical and subclinical symptoms has been attributed to the effects of indinavir treatment on the renal system. Given the method of data collection, the ATHENA database captures most overt clinical symptoms. There may be a relative underreporting of subclinical symptoms, such as abnormal laboratory values and renal insufficiency, depending on the intensity of patient monitoring. Assuming that any resulting misclassification is random within hospitals, the primary outcome measure of this study was the occurrence of urological symptoms (ranging from nephrolithiasis, renal colic, flank pain, and gross hematuria to a diagnosis of renal insufficiency, nephritis, or nephropathy), as diagnosed by the treating physician. The date of onset of first symptoms was defined as the index date.

Potential determinants for urological symptoms are sex, weight, HIV-1 RNA load (which may be considered as a marker for compliance; the lower limit of quantification was set at 1000 copies/mL for this analysis), high plasma concentrations, primary reason for prior treatment change (possible reasons included intolerance, treatment failure, patient request, pharmacological reason, other specified reason, or unknown), concomitant use of ritonavir and co-trimoxazole, and dosing regimen of indinavir. Body mass index (BMI) (defined as weight in kilograms divided by the square of height in meters) and LBM ([height in centimeters − 152] × [0.9 + 50] for men; [height in centimeters − 152] × [0.9 + 45.5] for women) were calculated from available data. Other factors investigated were the route of HIV transmission, Centers for Disease Control and Prevention disease classification,²⁵ CD4 cell counts, occurrence of any type of hepatitis, calendar time, and concomitant use of other antiretroviral medications. The mean monthly environmental temperatures from 1997 through 1999 were retrieved from the Internet site of the Royal Netherlands Institute of Meteorology.²⁶

Indinavir regimens were categorized as the standard regimen of 800 mg 3 times daily, 800 mg indinavir combined with 100 mg ritonavir twice daily, 400 mg indinavir combined with 400 mg ritonavir twice daily, 1000 mg or more of indinavir twice daily, or any other indinavir regimen (including combinations with other HIV protease inhibitors or nonnucleoside reverse transcriptase inhibitors). Continuous variables were dichotomized based on the lower quartile or the median.

Crude incidence rates of urological symptoms, with 95% confidence intervals (CIs) and relative risks (RRs), were calculated based on a Poisson distribution. The 1-year cumulative incidence was estimated using life table analysis. Potential risk factors for urological symptoms in the indinavir cohort were identified by means of univariate and multivariate Cox regression analysis, with the duration of indinavir use as follow-up time. We matched by treatment centers to control for potential bias from selective prescribing behavior, increased diagnostic attention for urological symptoms, or intensity of patient monitoring. Factors that may vary over time and potentially alter the risk of urological symptoms (weight, BMI, CD4 cell count, HIV-1 RNA load, indinavir regimen, type of nucleoside reverse transcriptase inhibitors, and environmental temperature) were included in the models as time-varying covariates. Effect modification was explored by means of stratified multivariate Cox regression analysis. All analyses were conducted using the statistical computer program SPSS 8.0 (SPSS Inc, Chicago, Ill).

The ATHENA source population comprised 2470 patients, among whom 2362 (96%) ever used an HIV protease inhibitor. Among these, 1239 patients (53%) began initial HIV protease inhibitor treatment in 1997 or thereafter. We excluded 18 patients because of incomplete data on HIV protease inhibitor treatment and 2 because of urological symptoms present at the start of HIV protease inhibitor treatment. The final HIV protease inhibitor cohort therefore comprised 1219 patients, among whom 445 used indinavir at any time.

For the indinavir cohort, we included all 445 patients from the HIV protease inhibitor cohort who used indinavir. In addition, we included 4 patients who had incomplete data on initial HIV protease inhibitor treatment but had complete data on indinavir treatment, and 195 patients who started taking HIV protease inhibitors before 1997 but who obtained the first prescription for indinavir in 1997 or thereafter. The final indinavir cohort therefore comprised 644 patients.

Within the HIV protease inhibitor cohort, a first occurrence of urological symptoms was described in 49 of 1219 patients (4%), among whom 45 (92%) developed symptoms during HIV protease inhibitor treatment. Other reasons for end of follow-up (right-censoring) were death (3%), loss to follow-up (0.4%), and last data collection (93%). The overall incidence rate of urological symptoms was 2.8 (95% CI, 2.0-3.7) per 100 person-years of HIV protease inhibitor treatment. Table 1 gives the specific HIV protease inhibitor incidence rates. Most symptoms (74%) occurred during treatment with indinavir. The risk of urological symptoms during use of indinavir-containing regimens was 8.7-fold higher (95% CI, 7.4-10.2) compared with regimens without indinavir.

The indinavir cohort was larger than the indinavir component of the HIV protease inhibitor cohort because inclusion criteria allowed patients to have used another HIV protease inhibitor before 1997. Conversely, the indinavir cohort had shorter follow-up because follow-up ended on discontinuation of indinavir treatment. The indinavir cohort was used for further analysis of the symptom incidence rates during indinavir treatment and the identification of risk factors for indinavir-associated urological symptoms. A first occurrence of urological symptoms during indinavir treatment was found in 58 of 644 patients (9%). Symptoms included nephrolithiasis (n = 38), renal colic or flank pain (n = 6), gross hematuria (n = 6), renal insufficiency (n = 6), nephritis (n = 1), and nephropathy (n = 1). Other reasons for end of follow-up (right-censoring) were discontinuation of indinavir use (33%), interruption of indinavir use for more than 7 days (6%), death (2%), loss to follow-up (1%), and last data collection (50%).

The overall incidence rate of urological symptoms was 8.3 (95% CI, 6.3-10.8) per 100 person-years of indinavir treatment. The highest incidence rates were observed with indinavir regimens of 1000 mg or more twice daily and during the first 6 months of indinavir treatment (Table 2). After the first 6 months, urological symptoms continued to occur, but at a steady lower rate, as illustrated in Figure 1.

Table 3 summarizes the baseline characteristics of the indinavir cohort. Overall, patients had a median age of 39 years, were predominantly male and homosexual, and had Centers for Disease Control and Prevention disease classification²⁵ A or B. The median weight, BMI, and LBM were 70 kg, 22 kg/m², and 73 kg, respectively. Most patients had more than 200 CD4 cells per microliter and more than 1000 HIV-1 RNA copies/mL. Most patients had had prior antiretroviral treatment and started indinavir treatment in 1997, and the most common first indinavir dosing regimen was 800 mg indinavir 3 times daily combined with lamivudine and zidovudine, usually without concomitant co-trimoxazole.

Baseline characteristics associated with urological symptoms were LBM below the lowest quartile (68 kg), undetectable HIV-1 RNA (≤1000 copies/mL), absence of co-trimoxazole use, and prior change of an antiretroviral treatment component because of intolerance. Prior change because of intolerance and the absence of co-trimoxazole use were closely related to low HIV-1 RNA (Pearson χ² test, P<.001 and P = .002, respectively), whereas low LBM was not (P = .65).

In addition to the baseline variables, we assessed whether time-varying covariates, such as weight, BMI, CD4 cell count, HIV-1 RNA load, indinavir dosing regimen, type of reverse transcriptase inhibitors, and environmental temperatures, altered the risk of urological symptoms. Of these, low weight (but not BMI) indinavir regimens of 1000 mg or more twice daily and warm environmental temperatures were associated with urological symptoms.

Table 4 gives the crude and adjusted RRs for all factors that were univariately associated with the development of urological symptoms at P = .10. Adjustment for age and sex alone (data not shown) and adjustment for univariately associated variables did not change any of the associations. Weight was excluded from the multivariate analysis because of its close relation to LBM.

Restricting the analysis to clinically overt symptoms, such as nephrolithiasis, renal colic, flank pain, or gross hematuria, did not lead to different conclusions (data not shown). Unfortunately, the small numbers of patients did not allow further breakdown and analysis of urological symptoms.

Stratification for LBM showed that the effects of undetectable HIV-1 RNA load (≤1000 copies/mL), indinavir regimens of 1000 mg or more twice daily and 800 mg indinavir combined with 100 mg ritonavir twice daily, and environmental temperatures greater than 18°C were more pronounced in the low LBM stratum, as reflected by higher RRs.

Further analysis of the 58 patients in the indinavir cohort who experienced urological symptoms provides insight into symptom reversibility. Twenty patients (35%) discontinued indinavir treatment, among whom 16 (28%) had subsequent resolution of their symptoms within 86 days. However, urological symptoms also resolved in 37 patients (64%) without discontinuation or before discontinuation of indinavir treatment. Among 5 patients (9%) in whom resolution of urological symptoms was not recorded, 4 had discontinued indinavir use.

In this large population-based cohort of patients infected with HIV in the Netherlands, the risk of urological symptoms during indinavir treatment was 8-fold higher compared with use of other HIV protease inhibitors. This difference is an expected finding, although the incidence among indinavir users was lower than reported elsewhere.^4,9,12,13,20 Most urological symptoms occurred within the first 6 months of indinavir use, but the incidence remained elevated thereafter, albeit at a steady reduced rate.

Risk factors for urological symptoms were low LBM, undetectable HIV-1 RNA (≤1000 copies/mL) at the start of indinavir treatment, prior stopping of an antiretroviral drug because of any type of intolerance, and the absence of concomitant co-trimoxazole use. In addition, indinavir regimens of 1000 mg or more twice daily, low weight, and monthly environmental temperatures greater than 18°C were identified as risk factors. Neither sex nor concomitant use of co-trimoxazole, suggested elsewhere as risk factors,^8,18 was associated with urological symptoms in our study.

The finding that LBM rather than BMI was associated with urological symptoms can be explained by the fact that a low LBM reflects a low distribution volume, which may lead to higher indinavir plasma concentrations. Moreover, indinavir has low lipid solubility, resulting in low fatty tissue penetration and therefore little protective effect of fatty tissue.⁶

Patients with undetectable HIV-1 RNA at the start of indinavir treatment represent those who previously had successful treatment with antiretroviral therapy. These patients are more likely to have therapeutic or high plasma drug concentrations compared with patients without prior viral suppression. As a consequence, they may be more susceptible to plasma concentration–dependent adverse effects, such as indinavir-associated urological symptoms.¹³ We can only speculate on this association because we did not measure indinavir plasma concentrations. However, a large proportion of patients with viral suppression had prior intolerance, leading to a treatment change, which confirms the fact that these patients may be more prone to development of adverse effects.²⁷ Another possible cause of urological symptoms in this group of patients is nausea, leading to less fluid intake and increased urine concentration.

The finding that indinavir dosing regimens of 1000 mg or more twice daily, which achieve higher indinavir plasma peak concentrations and larger areas under the curve, were associated with urological symptoms is consistent with earlier reports^13,28,29 of the relation between urological symptoms and high plasma indinavir concentrations. It is not known which pharmacokinetic factor, high indinavir plasma peak concentration or increased area under the curve, is the more important determinant of the risk for urological symptoms. Unfortunately, the number of patients using the combination of 800 mg indinavir and 100 mg ritonavir twice daily, a regimen with an indinavir plasma peak concentration comparable to that of 800 mg of indinavir 3 times daily and with a large area under the curve,^29,30 was too small to differentiate between the effects of the 2 pharmacokinetic factors.

Warm environmental temperatures reportedly increase the risk of urological symptoms.^17,20 Reduced urine production as a consequence of increased perspiration is the most likely explanation for the association between environmental temperatures and urological symptoms. The current advice for patients to drink at least 1.5 L of fluids per day may not be sufficient in warm weather.³¹

Potential limitations of this study concern information bias and bias because of diagnostic suspicion for uro logical symptoms among indinavir users vs users of other HIV protease inhibitors. To avoid information bias, we included only patients who started treatment in 1997 or thereafter, thus eliminating effects of accustomation to the new HIV treatment and preventing major underreporting because of relative unawareness of the symptoms. Nevertheless, subclinical urological symptoms of indinavir treatment, such as mild renal insufficiency (which may go undetected if renal function is not regularly assessed), may have been underreported. Furthermore, some urological symptoms may have been misdiagnosed as urinary tract infection. Underreporting or misclassification may have led to an underestimation of the incidence rates. However, it is unlikely to have affected the identification of risk factors for urological symptoms because we matched by treatment centers, within which the degree of reporting and preference for indinavir regimens can be expected to be similar among patients. Diagnostic suspicion bias is also unlikely to have played a role in the identification of risk factors, as we restricted this analysis to indinavir users only. Potential effect-modifying factors in the observed associations, such as dietary changes and fluid intake, were not available for analysis.

Because indinavir plasma concentrations play an important role in the development of urological symptoms, use of inhibitors of cytochrome P450 and hepatic dysfunction may have aggravated the symptoms observed in this study as effect modifiers.^9,16,32-34 Since the use of interacting drugs is rare (an automated warning system for incompatible drug combinations is operated by the community pharmacists in the Netherlands³⁵) and the prevalence of clinically overt hepatitis is low in our cohort (2%), these factors are unlikely to have substantially affected our results.

In conclusion, an incidence rate of urological symptoms of up to 8.3 per 100 person-years of indinavir use was found in this large population-based cohort study. Low LBM, minimal HIV-1 RNA at the start of indinavir treatment, the use of indinavir dosing regimens higher than the standard dosing of 800 mg 3 times daily, and warm environmental temperatures were independent risk factors for the development of urological symptoms. Because high indinavir plasma concentrations appear to play a key role, it may be useful to include indinavir plasma concentration monitoring as a preventive strategy for urological symptoms. Other indinavir and ritonavir dosing regimens, with lower indinavir plasma peak concentrations or smaller areas under the curve, may be considered. Our study focused on the identification of risk factors for urological symptoms. Any of the proposed control measures should be tested for their efficacy to reverse and prevent urological symptoms. Finally, the results endorse the need for increasing fluid intake in environments or under other circumstances in which urine output may be decreased.

Accepted for publication November 19, 2001.

This study was funded in part by a grant from the Dutch Inspectorate for Health Care, The Hague (Dr Dieleman). The ATHENA project is funded by the Dutch Health Insurance Council, Amstelveen, the Netherlands.

Ms Dieleman has received reimbursement of travel expenses from Merck Sharp & Dohme BV and has received an honorarium from Merck & Co Inc for her participation in an expert workshop on indinavir-associated nephrotoxicity organized by Merck & Co Inc.

We thank the patients, physicians, nurses, and data collectors participating in the ATHENA project and the ATHENA project team for their contributions to this study. We are grateful for critical comments from the ATHENA Toxicity Working Group.

Clinical and Epidemiological Working Group (*Site Coordinating Physician)

W. Bronsveld, Medical Centre-Alkmaar; H. Weigel,* K. Brinkman, P. Frissen, Onze Lieve Vrouwe Gasthuis; J. ten Veen,* M. Hillebrand, S. Schieveld, Onze Lieve Vrouwe Gasthuis–Location Prinsengracht; J. Mulder,* E. van Gorp, P. Meenhorst, Slotervaart Hospital; A. van Eeden, Jan v. Goyen Kliniek; S. Danner,* F. Claessen,* R. Perenboom, Academic Hospital Vrije Universiteit; J. K. Eeftinck Schattenkerk, E. Gisolf, M. Godfried, J. van der Meer, J. Nellen, D. Notermans, T. van der Poll, M. van Praag, J. Prins, P. Reiss, M. Reijers, T. Ruys, M. van der Valk, A. Verbon, F. Wit, Academic Medical Center–Amsterdam; C. Richter,* R. van Leusen, Hospial Rijnstate–Arnhem; R. Vriesendorp, Westeinde Hospital–Den Haag; R. Kauffmann,* E. Kogger, Hospital Leyengurg–Den Haag; B. Bravenboer, Catharina Hospital–Eindhoven; C. ten Napel,* K. Pogany, Medisch Spectrum Twente–Enschede; H. Sprenger,* G. Law, University Hospital–Groningen; R. W. ten Kate, Kennemer Gasthuis–Haarlem; M. Leemhuis, Medical Centre–Leeuwarden; F. Kroon, Leiden University Medical Centre; G. Schrey,* S. van der Geest, A. van der Ven, University Hospital–Maastricht; P. Koopmans,* M. Keuter, D. Telgt, University Hospital–Nijmegen; M. van der Ende,* I. Gyssens, S. de Marie, Erasmus University Medical Centre–Rotterdam; J. Juttmann,* C. van der Heul, St Elisabeth Hospital–Tilburg; M. Schneider,* J. Borleffs, I. Hoepelman, C. Jaspers, University Medical Centre–Utrecht; W. Blok, Hospital Walcheren–Vissingen.

Sub-Working Groups (*Chair):

J. Tijssen,* G. Bonsel, M. Dijkgraaf, S. Heisterkamp, Academic Medical Centre–Amsterdam (Working Group on Cost-effectiveness); J. Lange,* M. Jambroes, G. J. Weverling, Academic Medical Center–Amsterdam (Clinical Working Group); M. Mulder, Dutch HIV Patient Association; J. Dieleman, I. Gyssens, Erasmus University Medical Centre–Rotterdam; K. Brinkman, Onze Lieve Vrouwe Gasthuis; P. Koopmans, H. ter Hoffstede, University Hospital–Nijmegen; P. Reiss,* G. J. Weverling, M. Jambroes, Academic Medical Centre–Amsterdam (Toxicity Working Group).

Medical Psychology Working Group (*Chair)

M. Sprangers,* P. Nieuwkerk, Academic Medical Centre–Amsterdam.

Pharmacology Working Group (*Co-Chair)

D. Burger,* R. Aarnoutse, P. Hugen, University Hospital–Nijmegen; R. Hoetelmans,* R. van Heeswijk, A. Veldkamp, Slotervaart Hospital–Amsterdam.

Virological Working Group (*Co-Chair)

P. Rietra, K. Roozendaal, Onze Lieve Vrouwe Gasthuis; W. Pauw, A. van Zanten, Slotervaart Hospital; B. von Blomberg, P. Savelkoul, Academic Hospital Vrije Universiteit; F. de Wolf,* J. Goudsmit, L. van der Hoek, S. Jurriaans, Academic Medical Centre–Amsterdam; L. Nohlmans, Hospital Rijnstate–Arnhem; C. Jansen, Westeinde Hospital–Den Haag; P. Franck, A. Lampe, Hospital Leyenburg–Den Haag; E. Boel, A. Janz, Catharina Hospital–Eindhoven; R. Hendriks, Regional Laboratory Twente–Enschede; J. Schirm, Regional Laboratory–Groningen; H. Storm, Medical Centre–Leeuwarden; D. Veenendaal, LVF–Leeuwarden; A. Kroes,* Leiden University Medical Centre; C. Bruggeman, V. Goossens, University Hospital–Maastricht; J. Galama, University Hospital–Nijmegen; A. Osterhaus,* H. Niesters, Erasmus University Medical Centre–Rotterdam; A. Buiting, St Elisabeth Hospital–Tilburg; C. Boucher,* N. Back, R. Schuurman, University Medical Centre–Utrecht.

Steering Committee (*Chair)

J. Ruitenberg,* C. Boucher, D. Burger, S. Danner, R. Hoetelmans, R. W. ten Kate, R. Kauffmann, F. Kroes, J. Lange, A. Osterhaus, J. Tijssen, F. de Wolf.

Coordinating Centre

J. Lange, J. Tijssen, F. de Wolf (Project Leaders); M. Jambroes, E. van der Ven (Project Coordinators); S. Brouwer, M. Overveld, R. van Boxtel (Clinical Research Associates).

Data Collection Assistants

R. Runia, N. Wijdenes, Medical Centre–Alkmaar, N. Troost, R. Regez, Onze Lieve Vrouwe Gasthuis; M. Beerepoot, Onze Lieve Vrouwe Gasthuis–Location Prinsengracht; E. Oudmaijer, Slotervaart Hospital; J. Troon, Jan v. Goyen Kliniek; A. van Diggelen, Academic Hospital Vrije Universiteit; J. Ruijs, L. Veenenberg, Academic Medical Centre–Amsterdam; N. Langebeek, Hospital Rijnstate–Arnhem; M. Groot, S. Wildebeest, Westeinde Hospital–Den Haag; A. de Haas, Hospital Leyenburg–Den Haag; W. van Schaik, N. Slegers, Catharina Hospital–Eindhoven; H. Heins, T. Lansink, Medisch Spectrum Twente–Enschede; A. Bakker, S. Moolenburgh, University Hospital–Groningen; E. Kloosterhuis, M. Schoemaker, Kennemer Gasthuis–Haarlem; J. de Groot, A. Ketser, Medical Centre–Leeuwarden; W. Dorama, Leiden University Medical Centre; C. Leenders, University Hospital–Maastricht; M. Meeuwissen, B. Zomer, University Hospital–Nijmegen; T. Royaards, Erasmus University Medical Centre–Rotterdam; R. Santegoets, B. van der Ven, St Elisabeth Hospital–Tilburg; F. Bär, University Medical Centre–Utrecht; S. Baas, C. Ruissen, Hospital Walcheren–Vlissingen.