Abstract
BACKGROUND AND OBJECTIVES:
Trichomoniasis is a prevalent cause of vaginitis among adolescents that increases the risk of acquiring other sexually transmitted diseases and of negative pregnancy outcomes. Treatment of trichomoniasis is therefore essential for improving sexual and reproductive health outcomes. A timely, sensitive diagnostic test for T vaginalis may increase the accuracy of clinician’s treatment decisions, resulting in more infected women receiving treatment and fewer uninfected women receiving treatment.
METHODS:
This study was a retrospective observational assessment of electronic medical records before and after point-of-care (POC) implementation of the rapid antigen test. Records were collected from women aged 14 to 20 years who received a T vaginalis test in the emergency department during either study period. The main outcome measures were rates of accurate treatment, inaccurate treatment, and missed treatment of trichomoniasis in each study period.
RESULTS:
Overall rates of accurate treatment increased from 78.7% pre-POC to 87.7% post-POC (P = .02). Specifically, rates of not treating uninfected women increased from 61.4% pre-POC to 70.4% post-POC (P = .06), and rates of treating infected women were the same pre-POC (17.3%) and post-POC (17.3%; P = .99). Rates of inaccurate treatment decreased from 23.1% pre-POC to 13.1% post-POC (P = .02). Changes in missed treatment rates (14.0% pre-POC; 8.8% post-POC; P = .73) were not statistically significant.
CONCLUSIONS:
POC testing can improve clinical care by decreasing the use of antibiotics in uninfected women. The results of this study support the use of a T vaginalis rapid antigen POC test for adolescents presenting to the emergency department.
What’s Known on This Subject:
Trichomonas vaginalis is a common cause of vaginitis among adolescents associated with several negative health outcomes. Studies have shown that T vaginalis rapid antigen testing outperforms T vaginalis culture and wet preparation of a vaginal swab as a diagnostic tool.
What This Study Adds:
The impact of the T vaginalis rapid antigen test on clinical care has not been determined. This study showed that after implementation of rapid testing, improvement in clinical care occurred through a decrease in antibiotic treatment of uninfected women.
Vaginal trichomoniasis is a sexually transmitted disease (STD) caused by the protozoan Trichomonas vaginalis. Treating women with antibiotics is important to cure infection, reduce risks associated with infection, and lessen the possibility of developing antibiotic-resistant T vaginalis.1–8 Treating women for trichomoniasis when they are not infected (inaccurate treatment) exposes them to potential side effects of medication and the emotional distress associated with an STD diagnosis while also generating unnecessary costs.5,6 Failure to treat women for trichomoniasis when they are infected (missed treatment) places them at higher risk for symptomatic trichomoniasis, pregnancy complications, and acquisition of HIV.1,2,9 Inaccurate treatment and missed treatment of STDs are especially concerning in a population in whom follow-up for treatment is difficult, such as in adolescent patients presenting to an emergency department (ED).
Studies have demonstrated that T vaginalis rapid antigen testing outperforms T vaginalis culture and wet preparation of a vaginal swab (ie, wet mount) as a diagnostic tool.10–13 However, little is known about the impact of T vaginalis rapid antigen testing on clinical care, specifically in the ED setting. To determine the effects of T vaginalis rapid antigen testing on clinical care, we examined treatment outcomes both before and after implementation of the T vaginalis rapid antigen test in our institution’s ED. Comparisons before and after point-of-care (POC) implementation were possible because the institution’s clinical laboratory replaced the use of T vaginalis culture in clinical care with T vaginalis rapid antigen testing in February 2010. The primary aim of this study was to examine treatment outcomes, including accurate treatment, inaccurate treatment, and missed treatment, among all women tested for T vaginalis in the ED before and after implementation of the T vaginalis rapid antigen test. We hypothesized that the rate of accurate treatment would be higher and rates of inaccurate treatment and missed treatment would be lower after implementation of T vaginalis rapid antigen testing.
Methods
The study design was a retrospective observational review of electronic medical records (EMRs) approved by the hospital’s institutional review board.
Study Setting
The study was conducted in the ED of an urban pediatric academic center managing ∼90 000 visits per year. Testing for STDs is the standard of care for any woman presenting with lower abdominal pain or genital complaints who is sexually active. Testing was ordered in the ED by using an order set within the EMR that included a wet mount with either T vaginalis culture (pre-POC) or T vaginalis rapid antigen testing (post-POC). In our institution, 3 diagnostic tests for T vaginalis have been used in the clinical laboratory: the wet mount, the Inpouch TV Culture System (BioMed Diagnostics, Inc, San Diego, CA), and a T vaginalis rapid antigen test (OSOM Trichomonas Rapid Test; Sekisui Diagnostics, LLC, Lexington, MA). Wet mount was used during both time periods because it also served as a diagnostic test for other causes of vaginitis and vaginosis. Specificity of T vaginalis cultures approaches 100%, and sensitivity ranges between 75% and 95%.10,14 However, culture results are not available for 1 to 5 days. Wet mount is a POC test, with sensitivity ranging from 51% to 65% and specificity up to 100%.11,12,14 The T vaginalis rapid antigen POC test also tests vaginal swabs and has the highest sensitivity (82%–95%) of the 3 diagnostic tests as well as high specificity (97% to 100%).13,14 Samples for testing were collected in the ED, and all samples were delivered to the hospital’s central laboratory for processing and results.
Participants
Patient records were eligible for review if the ED encounter occurred between November 11, 2009, and February 3, 2010 (pre-POC) or between November 11, 2012, and February 3, 2013 (post-POC). Records from each time period were reviewed if the patient was female, aged 14 to 20 years, and had ≥1 order placed for either T vaginalis culture and/or wet mount (pre-POC) or T vaginalis rapid antigen test and/or wet mount (post-POC). Records from encounters that included a diagnosis of sexual abuse or suspected sexual abuse were excluded from analyses.
Data Collection
The start date of the pre-POC time period was chosen to coincide with the introduction of an EMR system and order set for STD testing, which occurred in November 2009; the end date coincides with the discontinuation of culture testing and the introduction of T vaginalis rapid antigen testing. The post-POC time period began after quality improvement research related to implementation of POC testing concluded. The months of the post-POC time period are the same as the pre-POC time period.
Data extracted from the EMRs included age, date of visit, length of ED stay, self-reported race, insurance status, T vaginalis tests ordered and date of testing, T vaginalis test results and date of results, orders for and documentation of dispensing metronidazole and/or tinidazole, and prescriptions for metronidazole and/or tinidazole.
Main Outcome Measures and Predictor Variables
The primary outcome measures were the rate of accurate, inaccurate, and missed treatment of T vaginalis among women in each study period within 30 days of testing. Women were considered to have received accurate treatment in 2 situations: (1) treatment given for a positive T vaginalis test result or (2) no treatment given for a negative T vaginalis test result. Women were considered to have received inaccurate treatment when treatment was given for negative T vaginalis test results. Women were considered to have missed treatment when no treatment was given for a positive T vaginalis test result. A woman was considered positive for T vaginalis if results of ≥1 of the ordered tests were positive, and she was considered negative for T vaginalis if all of the ordered test results were negative. Treatment was considered to have occurred if metronidazole or tinidazole was either given to the patient in the ED or prescribed within 30 days of the T vaginalis test result. Predictor variables included age, date of visit, length of ED stay, self-reported race, and insurance status.
Analysis
Data were analyzed by using SAS version 9.3 (SAS Institute, Inc, Cary, NC). Descriptive statistics generated frequencies and percentages for categorical variables as well as means and SDs for continuous variables. Differences between predictor variables in the pre-POC and post-POC time periods were determined by using χ2 tests for categorical variables and Wilcoxon rank-sum tests for continuous variables. The χ2 tests and 2-sided Fisher’s exact tests were used to test for significant differences in accuracy of treatment between the pre-POC and post-POC time periods. Unadjusted logistic regression models were used to evaluate the associations of the predictors of age, race, and insurance status with accuracy of treatment in the pre-POC and post-POC time periods. A post hoc analysis was conducted to further describe the number of infected women who were treated during the ED encounter. Differences in the proportions of infected women treated during their ED encounter in the pre-POC and post-POC periods were tested for significance by using a 2-sided Fisher’s exact test.
Results
Records from 249 encounters in the pre-POC time period and 179 encounters in the post-POC time period were used in the analyses. Characteristics of the women and description of the clinical encounter in each time period are shown in Table 1; there were no statistically significant differences in age, race, insurance status, or length of ED visit. Prevalence of T vaginalis was 20.1% (n = 50) in the pre-POC time period and 19.0% (n = 34) in the post-POC time period (P = .78). Of the 249 encounters in the pre-POC time period, 95.2% (n = 237) had results from both T vaginalis culture and wet mount, 2.4% (n = 6) from T vaginalis culture alone, and 2.4% (n = 6) from wet mount alone. Of the 179 encounters in the post-POC time period, 86.0% (n = 154) had results from both the T vaginalis rapid antigen test and wet mount, 14.0% (n = 25) from the T vaginalis rapid antigen test alone, and none from the wet mount alone.
TABLE 1.
Characteristics of the Study Population and Clinical Encounters
| Characteristic | Pre-POC, (n = 249) | Post-POC, (n = 179) | P |
|---|---|---|---|
| Race,a % | |||
| Black | 80.3 | 80.4 | .64 |
| Other | 4.4 | 2.8 | |
| White | 15.3 | 16.8 | |
| Insurance status,a % | |||
| Private | 15.7 | 16.8 | .95 |
| Public | 73.9 | 72.6 | |
| Uninsured | 10.4 | 10.6 | |
| Length of stay, mean ± SD, minb | 288.0 ± 112 | 308.5 ± 126.5 | .09 |
| Age, mean ± SD, yb | 17.2 ± 1.7 | 17.3 ± 1.6 | .56 |
Evaluated by using χ2 tests.
Evaluated by using the Wilcoxon rank-sum test.
Table 2 shows the rates of accurate, inaccurate, and missed treatment in the pre-POC and post-POC time periods. Rates of accurate treatment were higher post-POC (87.7%) compared with pre-POC (78.7%; P = .02). Among those who received accurate treatment, rates of not treating uninfected women were higher post-POC (70.4%) compared with pre-POC (61.4%; P = 0.06), and rates of treating infected women were the same post-POC (17.3%) and pre-POC (17.3%; P = 0.99). Rates of inaccurate treatment were lower post-POC (13.1%) compared with pre-POC (23.1%, P = .02). Rates of missed treatment were not significantly different between pre-POC (14.0%) and post-POC (8.8%; P = .73). In unadjusted logistic regression model analyses, the variables of age, race, and insurance status were not significantly associated with accurate treatment in the pre-POC or post-POC time period.
TABLE 2.
All Women Who Received Accurate Treatment, Inaccurate Treatment, or Missed Treatment
| Treatment Group | Pre-POC | Post-POC | P |
|---|---|---|---|
| Total no. of women tested | 249 | 179 | |
| Accurate treatment: infected and treated or uninfected and not treated | |||
| Total | 196 (78.7) | 157 (87.7) | .02a |
| Uninfected and not treated | 153 (61.4) | 126 (70.4) | .06a |
| Infected and treated | 43 (17.3) | 31 (17.3) | .99a |
| Inaccurate treatment: uninfected and treated | |||
| Total uninfected, n | 199 | 145 | |
| Total uninfected and treated | 46 (23.1) | 19 (13.1) | .02a |
| Missed treatment: infected and not treated | |||
| Total infected, n | 50 | 34 | |
| Total infected and not treated | 7 (14.0) | 3 (8.8) | .73b |
Data are presented as n (%) unless otherwise indicated.
Evaluated by using χ2 tests.
Evaluated by using Fisher’s exact tests.
During the pre-POC period, 30 of the 43 infected women who received treatment (69.8%) were treated during the ED encounter, and the remaining 13 infected women were treated after they had left the ED. During the post-POC period, 30 of the 31 infected women who received treatment (96.8%) were treated during the ED encounter, and 1 (3.2%) was treated after she had left the ED. The proportion of infected women who received treatment during the ED encounter was higher in the post-POC period than in the pre-POC period (P = .005).
To understand whether patients with inaccurate treatment may have received treatment of a diagnosis other than trichomoniasis, we first evaluated records of patients with inaccurate treatment to determine whether metronidazole or tinidazole may have been used to treat other infections. Because it was not possible to ascertain the clinical reasoning for metronidazole in each case, we conducted a second analysis in which patients with inaccurate treatment who received a treatment other than 2 g of metronidazole were considered not treated because a single 2-g dose is the standard trichomoniasis treatment in the ED. None of the encounters documented tinidazole as treatment. After removing these patients from the inaccurate treatment group, the results were similar: rates of accurate treatment were higher post-POC (94.4%) compared with pre-POC (83.1%; P < .001), and rates of inaccurate treatment were lower post-POC (4.83%) compared with pre-POC (17.6%; P < .001).
Because no single test or criterion standard was used during testing, and sensitivity of the wet mount, T vaginalis culture, and T vaginalis rapid antigen tests are known to differ, we performed a post hoc analysis to determine the impact of using multiple tests with differing sensitivities in the study population. The post hoc sensitivity of the T vaginalis culture and T vaginalis rapid antigen test were calculated to be 100%. Wet mount sensitivity was calculated to be 64.1% (95% confidence interval, 53.5–74.8). All women were diagnosed by using a wet mount as well as either T vaginalis culture or T vaginalis rapid antigen test except for 6 women who were tested with only the wet mount approach. Using the calculated test sensitivities from this population to determine the number of projected false-negative test results among women tested only with a wet mount, we projected that 2.2 (95% confidence interval, 1.5–2.8) of these 6 tests would produce false-negative results. To determine the impact of 2 false-negative test results on accurate, inaccurate, and missed treatment, 2 scenarios were considered: If 2 patients in the accurate treatment group had false-negative results, they would be assigned to the missed treatment group in the first scenario. If 2 patients in the inaccurate treatment group had false-negative results, these patients would be assigned to the accurate treatment group in the second scenario. When these 2 scenarios were considered, the direction of comparisons of the treatment rates did not change.
Discussion
This study evaluated the treatment outcomes of women tested for T vaginalis both before and after a reliable, valid POC test, the T vaginalis rapid antigen test, replaced T vaginalis culture in a pediatric ED. Use of a reliable, valid POC test improved treatment outcomes. Rates of accurate treatment (ie, no treatment given to women with a negative T vaginalis test result) were higher and rates of inaccurate treatment (ie, treatment given to women with a negative T vaginalis test result) were lower after T vaginalis rapid antigen testing was implemented. This outcome notably resulted in a decrease in the overuse of antibiotics. To our knowledge, this report is the first to examine the clinical impact of POC T vaginalis testing on trichomoniasis treatment outcomes.
Our findings of improved treatment outcomes associated with POC testing for T vaginalis are supported by studies of testing and treatment of other STDs. When test results for gonorrhea and Chlamydia are not available at the time of the visit, high rates of inaccurate treatment have been reported in an ED setting.5,15,16 After the implementation of T vaginalis rapid antigen testing, we observed a decrease in the treatment of uninfected women. This outcome is consistent with previous findings demonstrating that empirical treatment of T vaginalis and other STDs can result in significant rates of inaccurate treatment, especially in asymptomatic women, because clinicians must rely on their clinical suspicion as to whether to empirically treat the patient.15,16 We believe that the decrease in the treatment of uninfected women was largely due to clinicians having access to a POC test with higher sensitivity and specificity than wet mount. Although culture testing has a high sensitivity, the delayed turnaround time of the culture test results precludes the culture result informing the treatment decision made by clinicians when the patient is being evaluated in the ED.
Avoiding the use of antibiotics in women without infection promotes the individual and public health benefits of antibiotic stewardship. Increasing appropriate use of antibiotics shields patients from exposure to possible side effects of antibiotics and lowers medical costs. It can also help prevent T vaginalis antibiotic resistance.5,6 Low-level in vitro T vaginalis metronidazole resistance has already been reported in several sites around the country, and clinical treatment failure has been demonstrated in T vaginalis strains with even lower in vitro minimum lethal concentrations.17,18 The Centers for Disease Control and Prevention has estimated that 4% of T vaginalis cases have some level of metronidazole resistance.17 Because the nitroimidazoles are the only approved treatment of T vaginalis, clinical resistance to metronidazole presents formidable challenges for clinicians and patients. The T vaginalis rapid antigen test improves accurate treatment rates, and implementation of this test therefore has the potential to increase antibiotic stewardship.
Our study showed that after the implementation of the T vaginalis rapid antigen test, a higher proportion of infected women received treatment during their ED visit. This finding suggests that using a reliable, valid POC test for T vaginalis can affect clinical care in numerous ways. POC testing allows women to receive treatment more quickly and efficiently, and it provides medical staff and patients the opportunity for timely education and counseling regarding results. We have shown previously that women with positive STD results from a POC test accurately recalled their positive test result, and this knowledge was associated with abstinence, partner discussion, and partner treatment.19 In addition, the use of the POC test eliminates the need for laboratory resources to train personnel to prepare, incubate, and interpret T vaginalis cultures and lowers the need for staffing resources to contact women to notify them of positive results, provide education, and arrange for treatment.
The patient population served by the study pediatric ED is predominantly African American. T vaginalis rates are known to be higher in this population, which may limit generalizability.20 Individual provider practices may also affect results. However, there can be multiple trainees and handoffs for 1 patient, making assessment by provider impractical in this retrospective observational review. In addition, the introduction of EMR may have effects on the pre-POC results that cannot be measured. Future studies should consider conducting a cost analysis of the T vaginalis rapid antigen test to evaluate its cost-effectiveness in other clinical settings.
Conclusions
This study revealed significant improvement in decreasing the use of antibiotics in uninfected adolescent female subjects in an ED setting after implementation of T vaginalis rapid antigen testing. Because ED populations have historically low follow-up rates and carry high STD risk, methods to improve the accuracy of treatment, such as POC testing, are essential. In addition, decreasing the rate of inaccurate treatment leads to better antibiotic stewardship. Consequently, the success of T vaginalis rapid antigen testing supports its use in the ED setting for adolescent patients along with the further development of POC tests for treatment of other STDs.
Glossary
- ED
emergency department
- EMR
electronic medical record
- POC
point-of-care
- STD
sexually transmitted disease
Footnotes
Ms Postenrieder participated in data collection and interpretation, and drafted the initial manuscript; Dr Reed, Ms Hesse, and Dr Widdice conceptualized and designed the study, participated in data collection and interpretation, and reviewed and revised the manuscript; Drs Kahn, Gaydos, and Rompalo conceptualized and designed the study, and reviewed and revised the manuscript; and Dr Ding conducted data analysis and reviewed and revised the manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by grants from the National Institute of Biomedical Imaging and Bioengineering (2U54EB007958) and the National Institute of Allergy and Infectious Diseases (U01-068613; Gaydos) of the National Institutes of Health. Student contributions were supported by the Summer Undergraduate Research Fellowship at Cincinnati Children’s Hospital Medical Center. Funded by the National Institutes of Health (NIH).
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
References
- 1.McClelland RS, Sangaré L, Hassan WM, et al. Infection with Trichomonas vaginalis increases the risk of HIV-1 acquisition. J Infect Dis. 2007;195(5):698–702 [DOI] [PubMed] [Google Scholar]
- 2.Cotch MF, Pastorek JG II, Nugent RP, et al. ; The Vaginal Infections and Prematurity Study Group . Trichomonas vaginalis associated with low birth weight and preterm delivery. Sex Transm Dis. 1997;24(6):353–360 [DOI] [PubMed] [Google Scholar]
- 3.Centers for Disease Control and Prevention Incidence, prevalence, and cost of sexually transmitted infections in the United States. Available at: www.cdc.gov/std/stats/sti-estimates-fact-sheet-feb-2013.pdf. Accessed November 16, 2014
- 4.Centers for Disease Control and Prevention Sexually transmitted disease surveillance 2012. Available at: www.cdc.gov/std/stats12/Surv2012.pdf. Accessed November 16, 2014
- 5.Levitt MA, Johnson S, Engelstad L, Montana R, Stewart S. Clinical management of chlamydia and gonorrhea infection in a county teaching emergency department—concerns in overtreatment, undertreatment, and follow-up treatment success. J Emerg Med. 2003;25(1):7–11 [DOI] [PubMed] [Google Scholar]
- 6.Dunne RL, Dunn LA, Upcroft P, O’Donoghue PJ, Upcroft JA. Drug resistance in the sexually transmitted protozoan Trichomonas vaginalis. Cell Res. 2003;13(4):239–249 [DOI] [PubMed] [Google Scholar]
- 7.Wang CC, McClelland RS, Reilly M, et al. The effect of treatment of vaginal infections on shedding of human immunodeficiency virus type 1. J Infect Dis. 2001;183(7):1017–1022 [DOI] [PubMed] [Google Scholar]
- 8.Moodley P, Wilkinson D, Connolly C, Moodley J, Sturm AW. Trichomonas vaginalis is associated with pelvic inflammatory disease in women infected with human immunodeficiency virus. Clin Infect Dis. 2002;34(4):519–522 [DOI] [PubMed] [Google Scholar]
- 9.Workowski KA, Berman S; Centers for Disease Control and Prevention (CDC) . Sexually transmitted diseases treatment guidelines, 2010. [published correction appears in MMWR Recomm Rep. 2011;60(1):18] MMWR Recomm Rep. 2010;59(RR-12):1–110 [PubMed] [Google Scholar]
- 10.Huppert JS, Biro F, Lan D, Mortensen JE, Reed J, Slap GB. Urinary symptoms in adolescent females: STI or UTI? J Adolesc Health. 2007;40(5):418–424 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Nathan B, Appiah J, Saunders P, et al. Microscopy outperformed in a comparison of five methods for detecting Trichomonas vaginalis in symptomatic women. Int J STD AIDS. 2015;26(4):251–256 [DOI] [PubMed] [Google Scholar]
- 12.Huppert JS, Batteiger BE, Braslins P, et al. Use of an immunochromatographic assay for rapid detection of Trichomonas vaginalis in vaginal specimens. J Clin Microbiol. 2005;43(2):684–687 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Campbell L, Woods V, Lloyd T, Elsayed S, Church DL. Evaluation of the OSOM Trichomonas rapid test versus wet preparation examination for detection of Trichomonas vaginalis vaginitis in specimens from women with a low prevalence of infection. J Clin Microbiol. 2008;46(10):3467–3469 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Association of Public Health Laboratories Advances in laboratory detection of Trichomonas vaginalis. Public Health Lab Issues in Brief. Available at: www.aphl.org/AboutAPHL/publications/Documents/ID_2013August_Advances-in-Laboratory-Detection-of-Trichomonas-vaginalis.pdf. Accessed June 16, 2014
- 15.Pattishall AE, Rahman SY, Jain S, Simon HK. Empiric treatment of sexually transmitted infections in a pediatric emergency department: are we making the right decisions? Am J Emerg Med. 2012;30(8):1588–1590 [DOI] [PubMed] [Google Scholar]
- 16.Madhivanan P, Krupp K, Hardin J, Karat C, Klausner JD, Reingold AL. Simple and inexpensive point-of-care tests improve diagnosis of vaginal infections in resource constrained settings. Trop Med Int Health. 2009;14(6):703–708 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kirkcaldy RD, Augostini P, Asbel LE, et al. Trichomonas vaginalis antimicrobial drug resistance in 6 US cities, STD Surveillance Network, 2009-2010. Emerg Infect Dis. 2012;18(6):939–943 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lumsden WH, Harrison C, Robertson DH. Treatment failure in Trichomonas vaginalis infections in females. II. In-vitro estimation of the sensitivity of the organism to metronidazole. J Antimicrob Chemother. 1988;21(5):555–564 [DOI] [PubMed] [Google Scholar]
- 19.Reed JL, Simendinger L, Griffeth S, Kim HG, Huppert JS. Point-of-care testing for sexually transmitted infections increases awareness and short-term abstinence in adolescent women. J Adolesc Health. 2010;46(3):270–277 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sutton M, Sternberg M, Koumans EH, McQuillan G, Berman S, Markowitz L. The prevalence of Trichomonas vaginalis infection among reproductive-age women in the United States, 2001-2004. Clin Infect Dis. 2007;45(10):1319–1326 [DOI] [PubMed] [Google Scholar]