FemLab Clinical Trial Results

 

The performance of the FemLab test kit to detect Bacterial, yeast, Chlamydia and Trichomonas vaginitis in vaginal fluid specimens collected from symptomatic and asymptomatic patients, was determined in a multi-center study conducted at three geographically separate clinical sites. A total of 300 women were enrolled and tested.  Study site personnel performed the FemLab test according to a specific protocol, and also performed traditional tests for comparison to the FemLab results.  The study personnel performed control tests on each patient, and reported diagnoses of the most important vaginitis disease entities:  Bacterial Vaginitis, Gardnerella, Yeast Vaginitis, Trichomonas vaginitis, and Chlamydia Vaginitis.  In addition, clinical evaluations of the severity of symptoms in the presenting patients were recorded. 

 

The traditional tests used as control methods included commercially available pH and Gardnerella tests, wet mount microscopic evaluation of samples for Trichomonas parasites, and culture systems for detection of Chlamydia and yeast.  A pelvic exam was performed on each patient to determine overall level of vaginal health. 

 

The clinical trial results are discussed below, first with a general comparison between results, and then specific comparisons of the ability of each method to diagnose infectious organisms.  The results for each diagnosis as found by the FemLab Test Kit are compared to the results found by the control methods, and are reported in a 2 by 2 chart from which the statistical model may be calculated. 

The terms positive (+) and negative (-) are used to refer to the presence or absence of the condition of interest, the presence of an infectious diagnosis. Thus there are true positives and true negatives, where both methods of diagnosis agree. There are also two cells on each chart where the diagnoses disagree.  These proportions are described with the following terms:

Sensitivity is the proportion of true positives that are correctly identified by the test, compared to the control method.

Specificity is the proportion of true negatives that are correctly identified by the test, compared to the control method. 

Sensitivity and specificity are one approach to quantifying the diagnostic ability of a diagnostic test. In clinical practice, however, the test result is all that is known, so we want to know how good the test is at predicting the abnormal condition.  The comparison of sensitivity and specificity provide this means.

 

Performance Characteristics- Symptomatic vs. Asymptomatic  Patients

 

As part of the clinical trial, pelvic exams were performed by gynecologists on each of the 300 patients enrolled, and the results were reported on a scale of 1 to 4.  The pelvic examination reported the general condition of vaginal cavity, the color, amount and any odor of discharge, and other visible observations.  The results were reported as the following;

 

I+: Slight discharge with no color and odor

II+: Slight discharge with redness and no odor

III+: White discharge with redness, slight odor and itching

IV+: White discharge with pus, redness, strong odor, burning and itching

 

Thirty eight women were found to be asymptomatic (I+).  Of these women, however, 23 were also diagnosed by the hospital as having some form of vaginitis.  Fifteen were found to be completely free from any form of vaginitis.  FemLab results on these asymptomatic women agreed:  21 women were found to have some form of vaginitis, and 17 were healthy.  The remainder of the trial population (262 cases) had clinical symptoms of vaginal disease, with pelvic exam reports of II+, III+ or IV+.  Upon testing, only 10 of these cases were diagnosed as healthy with no form of infectious vaginitis by the hospital control methods, and by the FemLab test, and these cases were in the II+ pelvic exam category.  Thus, 96.2% of symptomatic patients had at least one, and sometimes more, diagnosis of Bacterial Vaginitis, Yeast Vaginitis, trichomoniasis, or chlamydia infections.  The conclusion that can be drawn from this analysis of the pelvic exam data is that pelvic exams are generally very accurate at predicting the presence of vaginitis, but that some forms of vaginitis are asymptomatic, and may not be diagnosed by this method alone.  This suggests that pelvic exams are not sufficient in themselves in the final determination of vaginitis.

 

Performance Characteristics – Healthy vs. Diseased Patients

 

The ability of the FemLab test kit to detect any of the various forms of vaginitis in vaginal fluid specimens collected in the multi-center study was compared to the results of the traditional tests for diagnosing any form of vaginitis.

 

Out of the 300 patients tested for any of the various forms of vaginitis, the FemLab and hospital control diagnoses agreed in 284 cases, resulting in an overall agreement of 94.7%. In 18 cases (6% of the total), both FemLab and reference methods concluded with a "Healthy" vaginal diagnosis.  However, FemLab found 7 cases with some form of vaginitis that the control methods missed, while the control methods found 9 cases with some form of vaginitis that were diagnosed "Healthy" by FemLab.  These results are shown in Table III below.  The significant level of overall agreement – 94.7% overall agreement - between the two methods suggests that FemLab is very capable of diagnosing vaginitis.

 

Table III - Healthy Diagnosis

 Comparison between FemLab and Control Methods

 

		Control	Diagnosis	Method
		Healthy	%	Diseased	%	Total	%
Femlab Diagnosis	Healthy	18	6%	9	3%	27	9%
	Diseased	7	2%	266	89%	273	91%
	Total	25	8%	275	92%	300	100%

 

 

Performance Characteristics – Bacterial Vaginitis and Gardnerella

 

The ability of the FemLab test kit to detect Bacterial Vaginitis and Gardnerella in vaginal fluid specimens collected in the multi-center study was compared to the results of the control test methods for diagnosing Bacterial Vaginitis.

 

Bacterial Vaginitis is the most common vaginitis and has been associated with pelvic inflammatory disease, cervicitis, postoperative infection, abnormal cytology (cellular structure), and increased acquisition of human immunodeficiency virus infection and other sexu­ally transmitted diseases. Bacterial Vaginitis can also cause obstetric complications such as preterm labor and low-birth-weight infants.

 

Common symptoms for Bacterial Vaginitis include: a milky, thin discharge at times; or a heavy, gray discharge and a "fishy" odor which may become more noticeable during inter­course. However, any individual may experience symptoms differently.

 

The performance of the FemLab test for Bacterial Vaginitis was compared to the control group reference methods for diagnosing Bacterial Vaginitis. 

Out of the 300 patients tested for Bacterial Vaginitis, the FemLab and control method diagnoses agreed in 247 cases, resulting in an overall agreement of 82.3%. In 93 cases, both reference methods produced positive diagnoses for Bacterial Vaginitis (31% of the total).  Thus, FemLab and the control methods agreed on positive diagnoses for 93 of 104 cases that the control method found positive, yielding a sensitivity = 89.4%. Also, FemLab and control methods agreed on negative findings 154 times out of 196 cases found negative by the control methods, yielding a FemLab specificity of 78.6% for Bacterial Vaginitis.  These results are shown in Table IV below.

 

Table IV  - Bacterial Vaginitis(Bv) Diagnosis

Comparison between FemLab and Control Methods

 

			Control	Methods	Diagnosis
		Bv +	(%)	Bv -	(%)	Total
FemLab	Bv +	93	31%	42	14%	135	45%
Diagnosis	Bv -	11	4%	154	51%	165	55%
	Total	104	35%	196	65%	300	100%

 

The FemLab test produced the following results for diagnosis of Bacterial Vaginitis: sensitivity = 89.4%; specificity = 78.6%; positive predictive value = 68.9%; negative predictive value = 93.3%; and overall agreement = 82.3%.

 

Performance Characteristics – Chlamydia Vaginitis

Another common form of vaginitis is caused by the bacterium Chlamydia Trachomatis. These infections are often asymptomatic. This makes diagnosis difficult, thus prolonging detection and treatment, as well as raising the probability of further spreading of the disease. The following are the most common symptoms of Chlamydia: increased vaginal discharge; light bleeding, especially after intercourse; pain in the lower abdomen or pelvis; burning during urination; pus in the urine; and redness and swelling of the ure­thra and labia.  Since the symptoms of Chlamydia may resemble other vaginitis conditions, it is critical that this be diagnosed correctly. 

 

Of the 300 cases enrolled in the FemLab clinical trial, 23 were diagnosed with Chlamydia infections by the FemLab test, and 2 also were diagnosed by the hospital control procedures.   The FemLab test correctly identified 23/25 of the cases identified positive by the control methods (92.0% sensitivity), and also had a specificity of 99.3%. (Sensitivity is the proportion of true positives that are correctly identified by the test.  Specificity is the proportion of true negatives that are correctly identified by the test). The results summarized in Table V below shows a comparison of the FemLab results to control methods for the diagnosis of Chlamydia.

 

Table V - Chlamydia Vaginitis (Cv) Diagnosis

Comparison between FemLab and Control Methods

 

		Control	Diagnosis	Method
		Cy+	%	Cy-	%	Total	%
Femlab Diagnosis	Cy+	23	8%	2	1%	25	8%
	Cy-	2	1%	273	91%	275	92%
	Total	25	8%	275	92%	300	100%

 

The FemLab test produced the following results for diagnosis of Chlamydia Vaginitis: sensitivity = 92.0%; specificity = 99.3%; positive predictive value = 92.0%; negative predictive value = 99.3%; and overall agreement = 98.7%.

 

Performance Characteristics – Yeast Vaginitis

Yeast Vaginitis is very common.  Even in asymptomatic, reproductive age women without recent yeast infection, there can be a 25-30% incidence of vaginal yeast colonization.  Culture is an accurate method to diagnose yeast infections, but this only applies to symptomatic patients because there is a background of false positive diagnoses by culture methods in women without yeast problems. The cause of Yeast Vaginitis has at least two components. One is the presence of a yeast species growing in the vagina and the other is some change in the vaginal biochemical or immune environment that allows the yeast organisms to overgrow and produce symptoms.  The most common yeast organism is candida albicans but other species of yeast also produce symptoms such as C. glabrata, C. tropicalis, C. uilliermondii and C. parapsilosis and others.

The performance of the FemLab test for Yeast Vaginitis was compared to the Culture reference method for diagnosing Yeast Vaginitis.  All of the 300 patients in the clinical trial population were evaluated for yeast infections. The FemLab test kit diagnosis agreed with the reference control method in 260 cases (86.7% overall agreement). In the 128 cases where the hospital control methods produced a positive diagnosis for Yeast Vaginitis, the FemLab test also showed positive results in 117 cases,  91.4% sensitivity. Also, in the 172 cases where the hospital reference methods produced negative test results for Yeast Vaginitis, the FemLab results agreed in 143 cases (83.1% specificity). Overall agreement between FemLab and the results of the reference tests, in diagnosing both positive and negative results, was 86.7%.

 

Table VI - Yeast Vaginitis (Yv) Diagnosis

Comparison between FemLab and Control Methods

 

			Control	Methods	Diagnosis
		Yv +	(%)	Yv -	(%)	Total
FemLab	Yv +	117	39%	29	10%	146	49%
Diagnosis	Yv -	11	4%	143	48%	154	51%
	Total	128	43%	172	57%	300	100%

 

Comparison To Control Methods: The FemLab test produced the following results for diagnosis of Yeast Vaginitis: sensitivity = 91.4%; specificity = 83.1%; positive predictive value =80.1%; negative predictive value = 92.9%; and overall agreement = 86.7%.

 

Performance Characteristics – Trichomonas Infections

 

The traditional method of diagnosing trichomonal infections is a wet mount microscopic examination of vaginal secretions in saline.  Careful examination may show moving organisms - trophozoites are about the size of a white blood cell with 3 flagella that cause the movement.  Compared to DNA testing, which can pick up about 87% of positive Trichomonal cases, wet mount microscopy usually identifies only about half of cases. 

Symptoms can be similar to a yeast infection vaginitis. pH measurement of vaginal discharge for both is usually greater than 4.5. 

 

Trichomonal Vaginitis is often asymptomatic, and even when symptoms are present, they correlate poorly with a clinical diagnosis of vaginitis. The relative lack of specificity of symptoms precludes a differential diagnosis based on symptoms. In the clinical study, Bacterial Vaginitis, Yeast infections, and Trichomoniasis were diagnosed on the basis of reference tests and the FemLab test results. The overlap of the various diagnoses from both test methods shown on Table VII below shows the results of different test methods in patients stratified by clinical diagnosis.

 

The ability of the FemLab test kit to diagnose Trichomonal infections was compared to the traditional method of wet mount microscopy used by the control laboratory. 

 

In the trial population of 300 patients analyzed for Trichomonal Vaginitis, the FemLab diagnosis agreed with the reference control method in 264 cases (88.0% overall agreement). In the 85 cases where the hospital control methods produced positive tests for Trichomonal Vaginitis, the FemLab test also showed positive in 72 cases (84.7% sensitivity). (Sensitivity is the proportion of true positives that are correctly identified by the test.  Specificity is the proportion of true negatives that are correctly identified by the test). In contrast, in the 215 cases where the hospital reference methods produced negative test results for Trichomonal Vaginitis, the FemLab results were positive in 23 cases (89.3% specificity).

 

Table VII - Trichomonas Vaginitis (Tv) Diagnosis

Comparison between FemLab and Control Methods

 

		Control	Diagnosis	Method
		Ty+	%	Ty-	%	Total	%
Femlab Diagnosis	Ty+	72	24%	23	8%	95	32%
	Ty-	13	4%	192	64%	205	68%
	Total	85	28%	215	72%	300	100%

 

Comparison To Control Methods: The FemLab test produced the following results for diagnosis of Trichomonal Vaginitis: sensitivity = 84.7%; specificity = 89.3%; positive predictive value 75.8= %; negative predictive value = 93.7%; and overall agreement = 88.0%.

 

Performance Characteristics – Other Vaginal Diseases

 

Sexually transmitted diseases such as syphilis, gonorrhea, HPV or Herpes are not detected directly by the FemLab test kit.  In some cases of such diseases, such as an aggressive Herpes infection, the FemLab test may report pH, blood, or protein as positive due to the symptoms of the Herpes.  In such cases, pelvic exams will clearly reveal the cause of the symptoms, and additional tests are widely available to confirm diagnoses.  In cases of syphilis or gonorrhea also, symptoms will often be very specific for the disease state.  Thus, even though FemLab may reveal no positive results when sexually transmitted diseases are causal factors of symptomology, an accurate diagnosis of the disease state can result from additional test procedures specific for such diseases. 

 

Performance Characteristics – Multiple Diagnoses

 

Significant numbers of patients in the clinical trial were diagnosed with multiple forms of vaginitis.   Eleven cases were diagnosed with 3 forms of vaginitis by FemLab – Bacterial, Chlamydia, and Trichomonal Vaginitis.  An additional two cases also had yeast infections as well as the prior three.  These diagnoses were confirmed by the control methods of testing.  The two cases with four diagnoses found by FemLab also had four diagnoses by control methods.  Of the remaining eleven FemLab cases with three or more diagnoses, eight were picked up by the control methods as showing two or three diagnoses.  In 11 of the 13 cases, the pelvic exam reported a IV+ result, showing severe symptoms.

 

Performance Characteristics – Summary

 

In Table VIII below, the statistical results of the comparisons between FemLab diagnoses and control method diagnoses for the four major infectious organisms responsible for vaginitis are shown. 

 

Table VIII

Statistical Measures of the Sensitivity, Specificity and Overall Agreement of the FemLab Test kit Compared to Control Methods

 

Vaginitis	Sensitivity	Specificity	Positive Predictive Value	Negative Predictive Value	Overall Agreement	Prevalence in trial population
Bacterial Vaginitis	89.4%	78.6%	68.9%	93.3%	82.3%.	31%
Chlamydia Vaginitis	92.0%	92.0%	92.0%	99.3%	98.7%.	8%
Trichomonal Vaginitis	84.7%	89.3%	75.8%	93.7%	88.0%	24%
Yeast Vaginitis	91.4%	83.1%	80.1%	92.9%	86.7%.	39%

 

 

The Sensitivity and Specificity for each infectious organism are the key factors for evaluation of the efficacy of FemLab.  The Sensitivity of the FemLab test kit – the agreement with the control methods for the positive diagnosis of the four various disease states - ranges between 84.7% and 92.0%.  This means that FemLab can be considered very effective in the identification of the infectious organisms; a clinician can have a high degree of confidence in the results.  Indeed, given the known uncertainties in the control methods of diagnosis, this level of agreement is excellent.  In addition, the Specificity – the ability to obtain negative diagnosis agreement – is also very good, ranging between 78.6% and 92.0%.  From the perspective of the FemLab test kit, many of these cases can be considered misdiagnosed by the control methods, giving a clinician additional confidence in the FemLab accuracy. 

 

In addition, many cases of FemLab "misdiagnosis" compared to the control method for one particular disease state are in fact correctly diagnosed for another disease.  For instance, many cases diagnosed with both Bacterial Vaginitis and Trichomoniasis by FemLab, only were diagnosed with Bacterial Vaginitis by the control methods.  It is very possible that in fact these women did have Trichomonal infections as well, since the control method of diagnosis of Trichomoniasis – microscopic evaluation – is well known to be sensitive to technician error. 

 

These results can be interpreted as providing a high level of confidence to a clinician for use of the FemLab test kit as a screening method.  Only a very small proportion of patients are misdiagnosed compared to the control method, and many of these are diagnosed correctly for another infection.  In summary, the clinical data show clearly that FemLab is very effective in diagnosing the major forms of vaginitis.