Urine Congo red test for the detection of preeclampsia in pregnant women presenting with suspected preeclampsia

Abstract

Objectives

To determine the predictive performance of the urine Congo red point-of-care test for the identification of preeclampsia in women presenting with suspected preeclampsia.

Methods

A prospective multi-center cohort study was conducted to include women with suspected preeclampsia (n = 244). The urine Congo red test was determined (score range 1–8). The diagnosis of preeclampsia was based on criteria proposed by The American College of Obstetricians and Gynecologists. The primary outcome was the predictive performance (sensitivity, specificity, negative and positive predictive values, as well as likelihood ratios) of the Congo red kit test for the diagnosis of preeclampsia.

Results

Fifty-four percent (131/244) of women with suspected preeclampsia subsequently developed preeclampsia. The sensitivity and specificity of the urine Congo red test were 49.6% and 94.7%, respectively, when using a cutoff for Congo red ≥4. The test had a significant positive correlation with the level of urine protein (Pearson correlation 0.61, p-value <.01). Intra- and inter-observer reliabilities were good (intra-class correlation coefficient and Cohen’s kappa coefficient of 0.88 and 0.75, respectively; p < .01).

Conclusion

The urine Congo red kit test has a high positive predictive performance for the identification of preeclampsia with high reproducibility. This test may be used as a bed side test to rule-in the diagnosis of preeclampsia in women presenting with suspected preeclampsia.

Keywords:

• Congo red kit test
• misfolded proteins
• point-of-care test
• preeclampsia
• urine congophilia

Introduction

Preeclampsia (PE) is a pregnancy-induced hypertension disorders that occurs in approximately 2%–8% of pregnancies [1], and it is associated with maternal and perinatal mortality and morbidity worldwide [2]. Each year, approximately 50,000 maternal deaths from PE occur globally, with approximately 9% of such deaths observed in Asia and Africa [1]. PE is characterized by the new onset of hypertension and significant end-organ dysfunction with or without proteinuria, typically presenting after 20 weeks of gestation (GA) [3]. The only definitive treatment for PE is delivery of the placenta [4]. In general, women with suspected PE can present with a wide variety of symptoms, such as headache, visual disturbances, abdominal pain or new-onset hypertension, and patients will be evaluated and examined for other organ systems and fetal involvement for the diagnosis of PE. These tests include complete blood count (for the assessment of hemoconcentration and thrombocytopenia), liver function, serum creatinine and urine analysis [5–7]. However, such tests are not predictive of PE or associated adverse complications [8]. In addition, patients are often hospitalized for additional investigations, causing substantial financial burden on the health care ­system [9–14].

An imbalance in angiogenic/anti-angiogenic factors, soluble fms-like tyrosine kinase-1 (sFlt-1) and placental growth factor (PlGF), is proven to play a role in predicting PE and adverse neonatal outcomes in women with suspected PE [15–18]. According to a multi-center study (PROGNOSIS), a low sFlt-1/PlGF ratio (<38) may be used to predict the absence of PE within 4 weeks with a negative predictive value and sensitivity of 99.3% and 80%, respectively [19]. In contrast, a sFlt-1/PlGF ratio of ≥85 had a sensitivity and specificity of 88% and 99.5%, respectively, for the identification of subsequent early PE (PE with delivery <34 weeks) [17,20]. However, use of the sFlt-1/PlGF ratio for the prediction of PE in the clinical setting is limited because of the lack of specific devices and costs. Hence, an alternative point-of-care test should be considered.

Congo red stain has been used to identify amyloid protein in the tissue [21,22], such as amyloid deposits in the brain tissue of patients with Alzheimer’s disease [23]. It has been reported that amyloid proteins are present in the urine of pregnant women diagnosed with PE [24–26]. A rapid test of urine Congo red staining has been developed, known as the Congo Red Dot paper test [24]. A US-based study showed that the urine Congo Red Dot paper test was superior (sensitivity and specificity of 80.2% and 89.2%, respectively) to maternal serum sFlt-1, PlGF and urine sFlt-1 measurements in both diagnosing and ruling-out PE [24]. However, this previous study also included women with a diagnosis of PE at the time of Congo Red Dot testing, and this may affect predictive performance. The objective of this study was to evaluate the performance of the urine Congo red test for the identification of PE in women presenting with suspected PE.

Materials and methods

This was a prospective, multi-center, cohort study of women with suspected PE. We enrolled 244 pregnant women from Ramathibodi Hospital (Mahidol University, Bangkok, Thailand), King Chulalongkorn Memorial Hospital (Chulalongkorn University, Bangkok, Thailand), HRH Maha Chakri Sirindhorn Medical Center (Srinakharinwirot University, Nakhornnayok Thailand), who presented with suspected PE from January 27th, 2022 to February 10th, 2023. The inclusion criteria were ≥18 years of age singleton pregnant women at the gestational week ≥20^+0 days with a suspicion of clinical diagnosis of PE based on one of the criteria proposed by The American College of Obstetricians and Gynecologists [3], including new-onset hypertension, proteinuria, headache, right upper quadrant of abdomen pain or tenderness, pulmonary edema, abnormal liver function test, increased serum creatinine, and thrombocytopenia. We excluded multiple pregnancy, patients with the diagnosis of PE, or presence of a fetal chromosomal abnormality. The study was approved by Human Research Ethics Committee of Mahidol University, Thailand (IRB No. COA. MURA2022/39 and COA. MURA2021/258), Chulalongkorn Memorial Hospital (IRB No. 612/64) and HRH Maha Chakri Sirindhorn Medical Center (SWUEC/E/M-002/2564).

Outcomes measured

The primary objective was to determine the predictive performance (sensitivity, specificity, positive or negative likelihood ratio and predictive values for each cutoff scores) of urine Congo red test for subsequent PE. Secondary objectives included determination of predictive performance in early (GA <34 weeks) and late onset (GA ≥34 weeks) PE. Furthermore, correlation coefficients between the urine Congo red test and the level of proteinuria, and maternal and fetal outcomes between PE and non-PE group were analyzed. Data on pregnancy outcomes were collected and the obstetric records of all women with preexisting or pregnancy-associated hypertension were examined to determine the diagnosis of PE, which was based on the definition proposed by The American College of Obstetricians and Gynecologists [3]. Specifically, PE was diagnosed by elevated blood pressure (systolic ≥140 mm Hg or diastolic ≥90 mm Hg on two occasions for at least 4 h), along with proteinuria (≥300 mg in a 24-h urine sample or a protein/creatinine ratio ≥0.3). In the absence of proteinuria, signs of severe features, such as severe elevated blood pressure (systolic ≥160 mm Hg or diastolic ≥110 mm Hg on two occasions for at least 4 h), thrombocytopenia (platelet level <100 × 10⁹/L), impaired liver function (two-fold increase in serum liver transaminase levels above the normal concentration), renal insufficiency (serum creatinine levels surpassing 1.1 mg/dL or a two-fold increase in serum creatinine concentration in the absence of concurrent renal disorders), pulmonary edema or cerebral/visual symptoms, may indicate severe preeclampsia. Diagnosis typically occurs after 20 weeks of GA [3].

Urine specimen collection and processing for Congo red test

Midstream urine specimens were collected from pregnant woman with suspected PE before administration of any medications. The Shuwen Biotech’s point-of-care test (Preeclampsia Detection Kit, Shuwen Biotech Co. Ltd, Zhejiang, China) was used to determine urine congophilia. The urine Congo red test was performed immediately or urine specimens were kept at 4 °C for tests performed within 24 h according to kit instructions [27]. A urine drop was mixed with Congo red dye using a plastic pipette included in the device package. Then, the mixture along with Congo red free solution (a negative control) were placed on the cellulose paper. The kit test result was obtained within 3 min, as defined by the pattern of Congo red dye on the cellulose paper (8 patterns according to the kit reference) (Figure 1). Originally, pattern number 5–8 were defined as a positive test according to the manufacturer. Other patterns were defined as a negative result. The results of the urine Congo red test were blinded to clinicians and were not used in clinical management.

Figure 1. Urine Congo red kit test.

Reproducibility

The reproducibility of the urine Congo red kit test was assessed by intra- and inter-observer reliability. For intra-observer reliability, the same researcher who classified the first result of the urine Congo red test was asked to classify 50 random pictures of the kit test again (P.T.). For inter-observer reliability, 50 pictures of the urine Congo red test were randomly selected and assigned to four different researchers (NY, NC, KR, PC) to classify the results using the reference picture from the kit test. The results of the inter-observer reliability and intra-observer reliability were assessed with the intraclass correlation coefficient and Cohen Kappa coefficient, respectively.

Statistical analysis

The sample size was calculated by the screening performance of the urine Congo red test according to a previous study, which showed 80.2%, 89.2%, and 27.7% for the sensitivity, specificity and incidence rate, respectively, of confirmed PE in women with suspected PE [24]. With an alpha of 5% and a power of 80%, at least 228 cases of suspected PE were required.

Descriptive data are summarized as medians (interquartile range) for numerical variables and count (percentage) for categorical variables. The comparisons of categorical and continuous data were assessed by the χ² test and T-test, respectively. Predictive performance (sensitivity, specificity, positive and predictive values, as well as likelihood ratio) of the urine Congo red test was calculated. A probability value (p-value) of less than 0.05 was considered statistically significant. Correlation coefficients were assessed with the Pearson correlation. Intra-class correlation coefficient and Cohen’s kappa coefficient were used to evaluated intra-observer and inter-observer reliability, respectively. The statistical software STATA V17 (California, USA), IBM SPSS version 18 (Armonk, N.Y., USA), and MedCalC version 20.218 (Mariakerke, Belgium) were used for statistical analysis in this study.

Results

Two hundreds and forty-four women with suspected PE were recruited (Figure 2). Subsequent PE was developed in 53.7% (131/244) of these women. Characteristics of participants stratified by diagnosis of preeclampsia are shown in Table 1. The PE group had higher systolic and diastolic blood pressure, an increased urine protein creatinine ratio and cesarean delivery rate, but a lower platelet count compared with the group without PE. In addition, patients with PE had significantly lower gestational age at delivery and birth weight, but higher rates of neonatal intensive care unit admission, neonatal respiratory distress syndrome and neonatal intraventricular hemorrhage compared with those without PE. The results of the urine Congo red test are shown in Table 2 and Figure 3. The score of the urine Congo red test was associated with a greater diagnostic odds ratio for subsequent PE (Table 3). When using a cutoff of urine Congo red ≥5 (traditional cutoff score), the sensitivity, specificity and positive predictive value were 45.0% (59/131), 96.5% (109/113) and 93.7% (59/63), respectively, for the identification of PE. This test has a positive likelihood ratio (LR) of 12.7 for PE. In addition, patients with a urine Congo red test ≥5 had a 22 times greater risk of developing PE compared with those with a negative test (<5). For a cutoff score ≥4, the sensitivity, specificity, positive predictive value and positive LR were 49.6% (65/131), 94.7% (107/113), 91.5% (65/71) and 9.3, respectively. We also examined the predictive performance of the urine Congo red test for early-onset (GA <34 weeks) and late-onset PE (GA ≥34 weeks). In general, the performance of the urine Congo red test was similar between early vs. late onset PE (supplementary Tables 1 and 2). The urine Congo red test score had a significant positive correlation with the urine protein creatinine ratio (Pearson correlation of 0.61; p-value <.01) (Figure 4). For reproducibility, the urine Congo red test had high intra- and inter-observer reliabilities (intraclass correlation coefficient and Cohen’s kappa coefficient were 0.879 and 0.752, respectively, p-value <.01 for both).

Figure 2. Flow chart of study participants.

Figure 3. Proportion of the diagnosis of preeclampsia based on using the urine Congo red test.

Figure 4. Pearson correlation [with 95% confidence interval (CI)] between Congo red and proteinuria (UPCR).

Table 1. Characteristics of the study population stratified by the diagnosis of preeclampsia.

Characteristics	Non-PE group (n = 113)	PE group (n = 131)	p-value
Age (years)^a	33.0 (26.0–36.0)	35.0 (29.0–38.0)	.05
Multiparity^b	31.9% (36)	48.9% (64)	.21
BMI (kg/m^2)^a	24.1 (21.4–29.6)	24.2 (21.1–28.5)	.45
Highest SBP (mmHg)^a	145.0 (140.0–150.0)	160.0 (150.0–172.0)	<.01
Highest DBP (mmHg)^a	90.0 (86.0–94.0)	98.0 (90.0–109.0)	<.01
Proteinuria (UPCR ≥0.3 mg/dL)^b	8.8% (10)	87.8% (115)	<.01
UPCR (mg/dL)^a	0.2 (0.1–0.2)	0.9 (0.4–2.3)	<.01
Serum platelet level (x10^3/µL)^a	258.5 (217.0–308.3)	243.0 (193.0–290.0)	.04
Creatinine (mg/dL)^a	0.5 (0.5–0.6)	0.6 (0.5–0.7)	.21
AST (U/L)^a	19.0 (16.0–23.0)	20.0 (17.0–29.3)	.06
ALT (U/L)^a	14.0 (10.0–19.0)	11.0 (8.0 − 23.0)	.13
FGR^b	9.7% (11)	18.3% (24)	.06
GA at delivery (weeks)^a	38.0 (37.0–39.0)	36.0 (33.0–39.0)	<.01
Birth weight (gram)^a	2985.0 (2615.0–3275.0)	2685.0 (1728.0–3080.0)	<.01
Apgar at 1 min (Apgar score)^a	9.0 (8.0–9.0)	8.0 (6.8–9.0)	.07
Apgar at 5 min (Apgar score)^a	10.0 (9.0–10.0)	9.0 (9.0–10.0)	.09
Cesarean Delivery^b	53.1% (60)	77.9% (102)	<.01
NICU admission^b	5.3% (6)	26.7% (35)	<.01
Stillbirth^b	0.0% (0)	1.5% (2)	.19
RDS^b	1.8% (2)	12.2% (16)	<.01
IVH^b	0.0% (0)	3.8% (5)	.03
Neonatal sepsis^b	3.5% (4)	6.9% (9)	.14

^a Data are presented as median (IQR).

^b Data are presented as % (absolute number), Abbreviations. GA, gestational age; AST, aspartate aminotransferase; ALT, alanine aminotransferase; FGR, fetal growth restriction; NICU, neonatal intensive care unit; RDS, respiratory distress syndrome; IVH, intraventricular hemorrhage; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; UPCR, urine protein creatinine ratio.

Table 2. Frequency of urine Congo red test results in preeclampsia and non-preeclampsia groups.

Urine Congo red test^a	Non-PE group (n = 113)	PE group (n = 131)
≥2+	108	130
≥3+	47	80
≥4+	6	65
≥5+	4	59
≥6+	2	53
≥7+	1	38
8+	0	5

^a Data are presented as absolute number.

Table 3. Performance of the urine Congo red test for the prediction of preeclampsia.

Urine Congo red Result	Odds Ratio [95%CI]	Sensitivity (%) [95%CI]	Specificity (%) [95%CI]	Positive LR [95%CI]	Negative LR [95%CI]	PPV (%) [95%CI]	NPV (%) [95%CI]
≥ 2+	6.0 [0.7 − 52.3]	99.2 [95.8 − 100]	4.4 [1.5 − 10.0]	1.0 [1.0 − 1.1]	0.2 [0.2 − 1.5]	54.6 [48.1 − 61.1]	83.3 [35.9 − 99.6]
≥ 3+	2.2 [1.3 − 3.7]	61.1 [52.2 − 69.5]	58.4 [48.8 − 67.6]	1.5 [1.1 − 1.9]	0.7 [0.5 − 0.9]	63.0 [54.0 − 71.4]	56.4 [46.9 − 65.6]
≥ 4+	17.6 [7.4 − 41.8]	49.6 [40.8 − 58.5]	94.7 [88.8 − 98.0]	9.3 [4.2 − 20.7]	0.5 [0.4 − 0.6]	91.5 [82.5 − 96.8]	61.8 [54.2 − 69.1]
≥ 5+	22.3 [8.1 − 61.5]	45.0 [36.3 − 54.0]	96.5 [91.2 − 99.0]	12.7 [4.8 − 33.9]	0.6 [0.5 − 0.7]	93.7 [84.5 − 98.2]	60.2 [52.7 − 67.4]
≥ 6+	37.7 [8.9 − 159.3]	40.5 [32.0 − 49.4]	98.2 [93.8 − 99.8]	22.9 [5.7 − 91.7]	0.6 [0.5 − 0.7]	96.4 [87.5 − 99.6]	58.7 [51.4 − 65.8]
≥ 7+	45.8 [6.2 − 339.7]	29.0 [21.4 − 37.6]	99.1 [95.2 − 100]	32.8 [4.6 − 235.0]	0.7 [0.6 − 0.8]	97.4 [86.5 − 99.9]	54.6 [47.6 − 61.6]
≥ 8+		3.8 [1.3 − 8.7]	100.0 [96.8 − 100.0]		1.0 [0.9 − 1.0]	100 [47.8 − 100]	47.3 [40.8 − 53.8]

Abbreviations. CI, confidence interval; LR, likelihood ratio; PPV, positive predictive value; NPV, negative predictive value.

Discussion

Principal findings of the study: (1) The frequency of a positive urine Congo red test was higher in women with PE than in those without PE; (2) Using a cutoff score of the urine Congo red test of 4 or above resulted in a sensitivity and specificity of 49.6% and 94.7%, respectively, for the identification of PE, and with a positive and negative LR of 9.3 and 0.5, respectively; (3) The urine Congo red test exhibited a similar performance for the identification of early and late PE; (4) The urine Congo red score was positively correlated with the urine protein creatinine ratio and; (5) The urine Congo red test had appropriate inter- and intra-observer reliabilities.

The current findings showed that the urine Congo red test had high specificity and a reasonable sensitivity, consistent with previous studies [24,27]. Several studies have determined the predictive performance of urine Congo red for the identification of PE. Rood et al. determined the predictive performance of urine Congo red for the evaluation of PE in 346 pregnant women presenting with suspected PE [24]. The urine Congo red test had a sensitivity and specificity of 80.2% and 89.2%, respectively, with a positive and negative LR of 7.43 and 0.22, respectively. Moreover, the urine Congo red test had superior performance for the detection of PE compared with the performance of angiogenic factors (sFlt-1, PlGF). Subsequently, Li et al. examined the predictive performance of the urine Congo red test for the identification of PE in 1532 pregnant women with suspected PE and asymptomatic at 20–41 weeks of GA [27]. In this previous study, the frequency of PE was 9.1% (140/1532) and the detection rate of the urine Congo red test of 74% (103/140) with a false positive rate of 3.0% (42/1392) [27]. However, in Li et al. and Rood et al.’s studies, the reference pictures of positive or negative urine Congo red test were used for the result determination. In contrast, the current urine Congo red test results were classified using a score from 1 to 8. The present results showed that the proper cutoff score for the urine Congo red test was “4 or above”.

Subsequently, Döbert et al. evaluated the screening performance of the urine Congo red test for the identification of PE in 2140 asymptomatic pregnant women at 35⁺⁰ to 36⁺⁶weeks of GA [28]. In this previous study, the frequency of PE was 2.1% (46/2,140), and the authors reported that the urine Congo red test score of “≥3” had a detection rate of 66.7% (8/12) and 23.5% (8/34) for PE within 2 weeks and after 2 weeks, respectively, of the urine test assessment, with a false positive rate of 17% (346/2094). The authors concluded that the screening performance of urine Congo red was very poor for the prediction of PE in asymptomatic pregnant women [28].

Recently, Wong et al. evaluated the urine Congo red test for the prediction of PE in 216 women presenting with suspected PE from 20–36⁺⁶weeks of GA. The frequency of PE was 36.1% (78/216). The urine Congo red test score of ≥5 (defined positive according to the manufacturer reference) had 10.6%, and 100% of the sensitivity and specificity, respectively, for the development of subsequent PE within 28 days [29]. The current results are consistent with the Wong et al. study, showing that the urine Congo red test had a very high specificity but fair sensitivity [29]. However, the current study included both preterm and term pregnant women presenting with suspected PE. In addition, the present study selected the cutoff score based on the predictive performance of each cutoff score.

In comparison to sFlt-1/PlGF results, the urine Congo red test had a high specificity but lower sensitivity. Previous work reported that the sFlt-1/PlGF had a sensitivity and specificity of 90% and >90%, respectively, for the diagnosis of PE in women presenting with suspected PE [30]. The urine Congo red may be used as a test to rule-in PE in a low resource setting in which the sFlt-1/PlGF test is not available.

The current study found that 5.3% (6/112) of patients had a false positive urine Congo red test and, importantly, all these women had chronic kidney disease (CKD) with the urine protein creatinine ratio ranging from 0.7 to 1.4. Fergus et al. reported that women with CKD may have urine congophilia. A similar amount of congophilia is present in pregnant women with CKD and in women with PE [31]. In addition, the current study found that the urine Congo red test score was positively correlated with the level of proteinuria, which may be an explanation for cases with a false positive urine Congo red test.

We propose that the urine Congo red test may be used as a point-of-care test for ruling in PE in women presenting with suspected PE, because the test: (1) has a high specificity; (2) is simple to perform (bedside test); (3) is inexpensive to set up; (4) is operator independent; (5) provides rapid results (within 3 min); and (6) is low maintenance (for the kit). Larger studies are required to confirm these findings.

The strength of this study includes the multi-center prospective cohort design, available obstetric and neonatal outcomes of participants and reproducibility of results. This study also determined the proper cutoff score of the urine Congo red test based on predictive performance. A limitation of this study is that most participants developed subsequent PE by the first visit of assessment, so the time interval from a positive urine Congo red test to the diagnosis of PE could not be determined.

Conclusions

The urine Congo red kit test is a bedside test with high specificity and reproducibility. This test may be used to rule-in the presence of PE in women with suspected PE.

Authors contributions

Conceptualization, sample collection, patient management, original draft preparation and finalization: P.T. and P.C. Substantial contributions to sample collection, patient management and finalization of the manuscript: K.R., N.Y. and N.C. All authors read and approved the final manuscript.

Acknowledgments

We thank all participants, our staff and faculty for their support.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author, P.T., upon reasonable request.

Additional information

Funding

This research was supported by the Faculty of Medicine Ramathibodi Hospital, Mahidol University (RF 65118 and A26/2564) and the Office of the Permanent Secretary, Ministry of Higher Education, Science, Research and Innovation [Grant No. RGNS 64-157].