Investigation of simple,objective,and effective indicators for predicting acute paraquat poisoning outcomes

Abstract

Initial symptoms of paraquat (PQ) toxicity are often not obvious, and the lack of advanced testing equipment and medical conditions in the primary hospital make it difficult to provide early diagnosis and timely treatment. To explore simple, objective, and effective indicators of prognosis for primary clinicians, we retrospectively analyzed acute PQ poisoning in 190 patients admitted to our hospital from 2008 to 2017. Based on their condition at the time of discharge, patients were categorized into either the survival group (n = 71) or the mortality group (n = 119). Age, PQ ingested amount, urinary PQ, urinary protein, white blood cell (WBC), and serum creatinine (Cr) were the key factors associated with the prognosis for PQ poisoning. We identified specific diagnostic thresholds for these key indicators of PQ poisoning: PQ ingested amount (36.50 mL), urinary PQ (semiquantitative result “++”), urinary protein (semiquantitative result “±”), WBC (16.50 × 10⁹/L), and serum Cr (102.10 µmol/L). Combining these five indicators to identify poisoning outcomes was considered objective, accurate, and convenient. When the combined score was <1, the predicted probability of patient death was 6%. When the combined score was ≥3, the predicted probability of patient death was 96%. These findings provide metrics to assist primary clinicians in predicting outcomes of acute PQ poisoning at earlier stages, a basis for administering treatment.

Keywords

Paraquat acute poisoning predictive indicator biological indicators

Introduction

Paraquat (PQ), or 1,1′-dimethyl-4,4′-bipyridinium cation, is a broad spectrum herbicide and is easy to use without any harm to the plant root. It is widely used in agricultural weed control and in the process of returning farmland to forest in more than 100 countries around the world (Kanchan et al., 2015). Using PQ at a price of 70–100 yuan (US$9.8–$14) per hectare of rice field can increase a farmer’s income by 2500 yuan (US$350) (Yang, 2015). Its superior performance and economic benefits have made PQ the second most popular herbicide in the world. There is currently no alternative herbicide with the same benefits of PQ, making it one of the most widely used by farmers around the globe (Jamshidi et al., 2017).

Despite its benefits as a herbicide, PQ is a fatal poison for humans and animals when used without proper care. The first case of PQ poisoning was reported in the 1960s, and subsequent cases have shown a globally increasing incidence since then. In fact, bibliometric analyses indicate a total of 4328 publications about PQ from 1962 to 2015, of which 1971 were related to PQ poisoning. There were 338 publications on PQ poisoning in the United States and 159 in China (Zyoud, 2018). The high suicide rates in rural parts of Asia and the Pacific coast, which constitute a major public health issue, are generally associated with PQ poisoning (Gawarammana et al., 2017). PQ poisoning is most often the result of either accidental misuse or suicide by oral consumption. Only 5–15% of total PQ consumed orally is absorbed by the gastrointestinal tract. After the absorbed PQ enters the circulatory system, 80–90% is excreted by the kidneys within 6 h, assuming normal renal function (Zhang et al., 2017). The PQ that is not excreted accumulates in the lungs, kidneys, liver, muscles, and other tissues. The pulmonary system is one of the most commonly affected organ systems, and a typical cause of death from PQ poisoning is pulmonary fibrosis with multi-organ failure (Liu et al., 2020; Rashidipour et al., 2020). The biological mechanisms of PQ poisoning are still not well understood, despite extensive investigation from experts in various fields over the past few decades. At the same time, there is no specified treatment for PQ poisoning. The mortality rate of PQ poisoning is as high as 60–80%, making it the leading cause of poisoning death in developing countries (Yu et al., 2014).

Most early symptoms of PQ poisoning are not obvious. Early assessment and effective intervention are key to significantly improving the prognosis of patients with acute PQ poisoning (Hou et al., 2017). Studies have demonstrated that patients are curable if their PQ toxicity level is between the highest among survivors and the lowest among non-survivors (Gil et al., 2014). Indicators that effectively predict the prognosis and outcome of acute PQ poisoning in its early stages could prove critical in effectively treating patients in early stages, preventing overtreatment of terminally ill patients, and reducing conflicts between doctors and patients (Cao et al., 2019; Senarathna et al., 2009). The hospitals that initially receive PQ poisoning patients are mostly local primary hospitals, lacking advanced testing equipment and diagnostic conditions, screen simple and efficient predictive indicators from routine indicators of blood and urine and rapid urine PQ qualitative testing. It is of great significance to guide these first-time hospitals how to treat patients with PQ poisoning efficiently.

Materials and methods

Ethics statement

This retrospective, descriptive analysis was submitted based on the guidelines of the Declaration of Helsinki and was approved by the Medical Ethics Committee of Guangzhou Twelfth People’s Hospital. Written informed consent was obtained from all individual participants included in the study. For patients with an altered sensorium, informed consent was obtained from their first-degree relatives. All primary data from PQ poisoning patients were collected in accordance with procedures outlined in the epidemiology guidelines. Data used in this study were anonymous and no identifiable personal data of the patients were available in the analysis.

Baseline date

Our hospital is an occupational disease prevention and treatment hospital. It is also the chemical poisoning rescue center for the region. In this study, we selected 340 patients suffering from acute PQ poisoning from July 2008 to December 2017. We recorded patient baseline information, including medical card number, gender, age, amount of ingested PQ, time of arrival at a local hospital after poisoning, time of arrival at our hospital after poisoning, time of arrival at our hospital from local hospital, time of first gastric lavage, whether blood purification (hemodialysis, blood perfusion or combination therapy) was performed, time of first blood purification (if performed), hormone therapy, vomiting, dizziness/fatigue/chest tightness, and abdominal pain. We also recorded laboratory test results performed at admission, including white blood cell (WBC) count, alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood glucose, creatinine kinase (CK), lactate dehydrogenase (LDH), urea, serum creatinine (Cr), urine protein, and urine PQ. Additionally, we recorded patient outcome at discharge. Gender, blood purification performed, vomiting, dizziness/fatigue/chest tightness, abdominal pain, urinary protein, and urinary PQ were graded variables. They are represented, depending on severity, using the following values: −, ±, +, ++, +++, and ++++. All other variables are reported as quantitative.

The inclusion criteria for enrollment in this study were as follows: the patient was poisoned by PQ orally, there was a clear diagnosis of PQ poisoning, the patient arrived at our hospital within 24 h of poisoning, complete baseline data and routine blood tests were collected, laboratory test results were obtained within 24 h of admission, and the patient actively cooperated with the treatment. Patients were excluded from the study if the patient was pregnant or lactating, consumed other poisons (e.g. rat poison, herbicides, alcohol, sleeping pills), possessed any preexisting conditions (e.g. heart disease, lung disease, liver disease, kidney disease, metabolic diseases, cancer), was discharged without hospitalization, or could not provide approximate oral PQ consumption and other necessary information.

For those patients who could not recall the amount of PQ ingested, we estimated the oral dose from experimental experience (Xu et al., 2011). The amount constituting “a small mouthful” was estimated to be 20 mL for men and 10–15 mL for women; “a moderate mouthful” was 40 mL for men and 30 mL for women; and “a large mouthful” was 60 mL for men and 40 mL for women.

According to the inclusion and exclusion criteria, 100 males and 90 females with acute PQ poisoning were included in this study. Patients’ age ranged from 31.36 ± 13.94 years (median, 26.50 years), and the approximate oral PQ dose ranged from 50.02 ± 50.36 mL (median, 30.00 mL). Based on the patient’s condition at the time of discharge, they were categorized into either the survival group (n = 71) or the mortality group (n = 119). The study flowchart is shown in Figure 1. The mortality group included patients who died during hospitalization, as well as critically ill patients that elected to abandon treatment.

Figure 1.

Flow diagram of the study patients of acute paraquat poisoning.

Methods

In this retrospective study, we collected information on general conditions, laboratory test results, and discharge outcomes of patients admitted to the hospital within 24 h of acute PQ poisoning. We also compared the differences between baseline data and laboratory test results of the survival group and the morality group. We fitted the equation for early prediction of death probability by acute PQ poisoning. We also searched for predictive indicators and optimal diagnostic values that were effective in judging the outcome of poisoning and explored ways to easily, quickly, and accurately assess acute PQ poisoning diagnosis.

Statistical analysis

The data were collected and analyzed using Microsoft Excel and SPSS22.0. Categorical variables are presented as ratios/percentages. Continuous variables are presented as mean ± standard deviation or median and interquartile range, based on their distribution determined by the Shapiro–Wilk test. Univariate comparisons were performed with use of the Student’s t-test, the χ ² test, and the Wilcoxon rank-sum test. The variables that were statistically significant (p < 0.10) were included in the multivariate analysis. Multiple logistic regression analyses were performed to parse out the main influencing factors affecting the prognosis of patients with acute PQ poisoning and to build the predictive equation for probability of death in these patients. Sensitivity and specificity tests in conjunction with a receiver operating characteristic (ROC) curve were computed to analyze the value of these main effect indicators in judging the prognosis of patients with acute PQ poisoning. The best diagnostic values for signaling the prognosis were calculated by maximizing Youden’s index. All p-values were two-sided and those less than 0.05 were considered to indicate statistical significance.

Results

Basic information on research participants

From July 2008 to December 2017, a total of 340 patients with acute PQ poisoning were included in the study. Of these, 190 were included in the final analysis. Based on the condition of the participants at the time of discharge, they were enrolled either into the survival group (n = 71) or the mortality group (n = 119). We observed statistically significant (p < 0.05) differences between the two groups with respect to amount of PQ ingested (21.92 ± 23.88 vs. 66.74 ± 54.51), vomiting (38:33 vs. 97:22), and whether blood purification was performed (46:25 vs. 97:22). Other indicators were not identified to be statistically different, as shown in Table 1.

Table 1.

Comparison of characteristics between the survival group and the mortality group at the time of admission.

Variable	Survival group (n = 71)	Mortality group (n = 119)	t/χ ²	p-Value^a
Age (years)	29.20 ± 12.42	32.65 ± 14.67	1.66	0.09
Gender (male:female, n)	36:35	64:55	0.17	0.68
Amount of PQ ingested	21.92 ± 23.88	66.74 ± 54.51	6.56	<0.01
Time from poisoning to arrival at local hospital (h)	1.95 ± 3.53	2.30 ± 3.11	0.72	0.47
Time from poisoning to arrival at Guangzhou Twelfth People’s Hospital (h)	10.01 ± 5.47	10.88 ± 6.08	1.00	0.32
Time to arrival at Guangzhou Twelfth People’s Hospital from local hospital (h)	8.06 ± 5.44	8.58 ± 5.90	0.61	0.54
Vomiting (Y:N, n)	38:33	97:22	16.94	<0.01
Dizziness/fatigue/chest tightness (Y:N, n)	12:59	32:87	2.49	0.11
Abdominal pain (Y:N, n)	11:60	22:97	0.28	0.60
First gastric lavage time (h)	2.36 ± 3.94	2.67 ± 4.43	0.48	0.63
Blood purification performed (Y:N, n)	46:25	97:22	6.68	0.01
First blood purification time (h)^b	0.67 ± 0.33	0.82 ± 1.27	0.77	0.44
Hormone therapy (Y:N, n)	56:15	103:16	1.92	0.17

PQ: paraquat.

^a The variables that are statistically significant (p < 0.10) were included in the multivariate analysis.

^b Blood purification time: compared the first blood purification time of the two groups of patients, excluding patients who did not receive blood purification (survival group = 46 and mortality group = 97).

Laboratory test results

Results from the Student’s t-test analysis regarding differences between the two groups implicated urinary PQ, urinary protein, WBC, and other biochemical indicators. Table 2 displays the statistical difference (p < 0.10) of these factors between the two groups. The values of these factors are higher in the mortality group than in the survival group.

Table 2.

Comparison of laboratory test results in the survival group and the mortality group at the time of admission.

Variable	Survival group (n = 71)	Mortality group (n = 119)	Z/t	p-Value^a
Urine PQ (total rank)^b	4340.50	13,804.50	6.83	<0.01
Urine protein (total rank)^b	4060.00	14,085.00	7.66	<0.01
WBC (10⁹/L)	12.11 ± 4.48	21.15 ± 10.65	7.41	<0.01
ALT (U/L)	19.35 ± 12.02	64.74 ± 92.58	4.11	<0.01
AST (U/L)	23.60 ± 12.22	98.66 ± 123.53	5.10	<0.01
Urea (mmol/L)	4.27 ± 1.80	8.03 ± 6.89	4.51	<0.01
Serum Cr (µmol/L)	82.71 ± 42.55	188.94 ± 112.72	7.62	<0.01
Blood glucose (mmol/L)	6.83 ± 2.09	8.36 ± 3.40	3.42	<0.01
CK (mmol/L)	190.42 ± 177.50	294.20 ± 202.97	3.57	<0.01
LDH (U/L)	179.42 ± 54.22	315.69 ± 174.62	6.39	<0.01

PQ: paraquat; WBC: white blood cell; ALT: alanine aminotransferase; AST: aspartate aminotransferase; Cr: creatinine; CK: creatinine kinase; LDH: lactic dehydrogenase.

^a The variables that are statistically significant (p < 0.10) were included in the multivariate analysis.

^b Graded data, described by the total rank; statistical result was Z value from the Wilcoxon rank-sum test.

Analysis of factors affecting the prognosis of PQ poisoning

Univariate analyses, including age, amount of PQ poisoning, vomiting, blood purification, urine PQ, urinary protein, WBC, ALT, AST, urea, serum Cr, blood glucose, CK, and LDH, are shown in Tables 1 and 2. Multiple linear regression analysis was performed to test the multicollinearity between independent variables prior to the execution of multiple logistic regression analysis. All variance inflation factors (VIFs) were less than 10 (tolerances were higher than 0.1), meaning that no significant multicollinearity between independent variables was identified.

To exclude the effects of age, vomiting, and blood purification on the logistic regression results for both groups, we used the conditional exclusion method to screen out the main influential factors. We used the method of mandatory inclusion for age, vomiting, and blood purification and then used the main influencing factors selected in the previous logistic regression analysis step. Multiple logistic regression analysis demonstrated that amount of PQ poisoning, urine PQ, WBC, urinary protein, serum Cr, and age are the main risk factors affecting the prognosis of hospitalized patients with acute PQ poisoning within 24 h. These indicators could be used to predict the probability of patient death from PQ poisoning, as shown in Table 3. The equation, with an R ² of 74.5%, for predicting the probability of death among patients with acute PQ poisoning is shown as equation (1). Its accuracy rate of prediction of death is as high as 88.9%.

p = \frac{e^{(- 7.079 + 0.024 X_{1} + 0.675 X_{2} + 0.107 X_{3} + 0.542 X_{4} + 0.017 X_{5} + 0.059 X_{6})}}{1 + e^{(- 7.079 + 0.024 X_{1} + 0.675 X_{2} + 0.107 X_{3} + 0.542 X_{4} + 0.017 X_{5} + 0.059 X_{6})}},

where X ₁ is the PQ ingested amount; X ₂, urine PQ; X ₃, leucocyte count; X ₄, urine protein; X ₅, serum Cr; and X ₆, age.

Table 3.

Multifactorial logistic regression models affecting prognostic factors of acute PQ poisoning (n = 190).

Variable	B	SE	p-Value^a	Exp (B)	95% CI
Amount of PQ ingested	0.024	0.009	0.009	1.024	1.006–1.043
Urine PQ	0.675	0.241	0.005	1.964	1.226–3.148
WBC	0.107	0.052	0.039	1.113	1.005–1.233
Urine protein	0.542	0.202	0.007	1.719	1.157–2.555
Serum Cr	0.017	0.005	<0.001	1.017	1.008–1.027
Age	0.059	0.020	0.003	1.061	1.020–1.103
Vomiting	−0.566	0.586	0.334	0.568	0.180–1.791
Blood purification	1.262	0.857	0.141	3.533	0.659–18.948
Constant	−7.079	1.360	<0.001	—	—

PQ: paraquat; B: regression coefficiect; SE: standard error; Exp (B): odds ratio; CI: confidence interval; WBC: white blood cell; Cr: creatinine.

^a p < 0.05 were considered to indicate statistical significance.

Diagnostic values of main indicators affecting the prognosis of patients with acute PQ poisoning

To further analyze the significance of the main indicators assessed by multiple logistic regression for predicting the prognosis of acute PQ poisoning patients, we generated an ROC. PQ ingested amount, urinary PQ, urinary protein, WBC, and serum Cr all possess high sensitivity and specificity in determining the prognosis of acute PQ poisoning. These indicators can independently predict the outcome, as shown in Figure 2 and Table 4. Among the five major predictors, urine protein was identified as the most sensitive indicator at 0.90. High urinary protein may indicate poor prognosis in a patient (test result “±”). The indicator with the highest specificity was WBC count at 0.92. Patients with WBC counts less than 16.5 × 10⁹/L are likely to improve. WBC was identified as having the highest prognostic value, with an area under the ROC of 0.85 ± 0.03. The WBC of 16.5 × 10⁹/L is the best diagnostic threshold value for determining prognosis among PQ poisoning patients. Patients with WBC more than 16.5 × 10⁹/L indicated high mortality, while those with lower WBC indicated a better prognosis.

Figure 2.

Receiver operating characteristic (ROC) curve analysis of age, paraquat ingested amount, urine paraquat, urine protein, serum creatinine, white blood cell of patients with acute paraquat poisoning (n =190).

Table 4.

ROC analysis results of main influencing factors and best critical values to determine prognosis of patients with acute PQ poisoning (n = 190).

Variable	AUC ± SE	p-Value^a	95% CI	The best critical point
Variable	AUC ± SE	p-Value^a	95% CI	Diagnostic value	Sensitivity	Specificity	Youden’s index	PPV	NPV
Age	0.56 ± 0.04	0.17	0.48–0.64	—	—	—	—	—	—
PQ ingested amount	0.82 ± 0.03	<0.01	0.76–0.89	36.50	0.66	0.89	0.55	0.91	0.61
Urine PQ	0.79 ± 0.03	<0.01	0.72–0.85	1.50	0.59	0.85	0.44	0.87	0.55
WBC	0.85 ± 0.03	<0.01	0.80–0.91	16.50	0.70	0.92	0.62	0.94	0.65
Urine protein	0.82 ± 0.03	<0.01	0.76–0.88	−0.50	0.90	0.66	0.56	0.82	0.80
Serum Cr	0.84 ± 0.03	<0.01	0.78–0.90	102.10	0.75	0.86	0.61	0.90	0.67

ROC: receiver operating characteristic; PQ: paraquat; AUC: the area under the ROC; SE: standard error; CI: confidence interval; PPV: positive predictive value; NPV: negative predictive value; WBC: white blood cell; Cr: creatinine.

^a p < 0.05 were considered to indicate statistical significance.

To develop a simple, rapid, and accurate method for assessing the prognosis of acute PQ poisoning, we combined the diagnostic values of the five main influencing factors. For each factor, if the test value was greater than the diagnostic value, a value of 1 was assigned; if the test value was less than the diagnostic value, a value of 0 was assigned. We then analyzed the aggregate score for all five factors in each patient to predict the outcome. From Table 5, we show that when the joint score is less than 1, all indicators have a test value less than the diagnostic value. The predicted probability of patient survival in this case is as high as 94%. The predicted probability of death increases with each increase in the joint score. When the joint score is greater than or equal to 1, the predicted probability of death can be accurately estimated at 76%. When the score is greater than or equal to 3, the predicted probability of death is as high as 96%. When the test values of all five indicators are greater than the diagnostic values, creating a joint score of 5, the predicted probability of death is 100%.

Table 5.

Joint scores of five major influencing factors^a combined to assess prognosis in patients with acute PQ poisoning.

Joint score	Sensitivity	Specificity	PPV	NPV
≥1	0.98	0.48	0.76	0.94
≥2	0.92	0.76	0.87	0.84
≥3	0.81	0.94	0.96	0.74
≥4	0.61	0.99	0.99	0.60
5	0.27	1.00	1.00	0.45

PQ: paraquat; PPV: positive predictive value; NPV: negative predictive value.

^a Five major influencing factors: PQ ingested amount, urinary PQ, urinary protein, white blood cell, and serum creatinine.

Discussion

At present, there is no effective treatment for PQ poisoning resulting in a high mortality rate. In clinical practice, doctors often classify the severity of acute PQ poisoning into four types: mild, moderate, severe, and acute outbreaks. These are based on the dose of poison, symptoms at admission, and extent of damage to tissues and organs. However, there are no uniform criteria to classify patients into these groups. In one study, a few patients did not exhibit obvious clinical signs in the initial stages of PQ poisoning until organ damage was identified almost a week later (Jiang et al., 2012). At the same time, the quantity ingested in PQ poisoning cases is often self-reported by the patients or their family members, and lack of a uniform reference causes clinician confusion about estimating the amount of poisoning. Therefore, different physicians and institutions often grade the same patient differently. This makes it difficult for a physician to effectively diagnose, evaluate prognosis, and administer appropriate treatment.

In recent years, some scholars have attempted to assess the severity and prognosis of PQ poisoning through a variety of methods, including “Acute Physiology and Chronic Health Evaluation” (Huang et al., 2017), “Sequential organ failure assesment score” (Wang et al., 2018), metabolomics (Hu et al., 2017), or serum/plasma PQ concentrations (Shi et al., 2017), or these methods combined with some biochemical indicators (Hou et al., 2017; Kavousi et al., 2017; Liu et al., 2016; Liu et al., 2017b; Oghabian et al., 2019). Although these methods have a high diagnostic and prognostic index, they possess a cumbersome number of evaluation items, are limitated by equipment requirements and cost, the level of clinicians’ diagnosis and treatment, and poor applicability to primary medical institutions. And most of the first hospitals for PQ poisoning patients are often local primary hospitals. Therefore, exploring simple, objective, and effective prognostic indicators from routine blood and urine markers is of great significance in guiding primary clinical treatment and utilization of medical resources.

We selected patients with acute oral PQ poisoning from July 2008 to December 2017 to analyze the relationship between hospital outcomes, baseline data, and conventional biological indicators at the time of admission. We found that age, ingested amount of PQ, urinary PQ, urinary protein, WBC, and serum Cr were good predictors of mortality in acute PQ poisoning patients (R ², 74.5%). Moreover, high sensitivity and specificity of ingested amount of PQ, urinary PQ, urinary protein, WBC, and serum Cr means that each can predict prognosis independently. The combination of these five indicators was even more accurate and intuitive. When all five indicators were below their respective diagnostic values, the predicted probability of patient survival was as high as 94%. Conversely, when more than three indicators were above the diagnostic value, the predicted probability of patient death was as high as 96%. This method is very simple, rapid, and efficient for the primary clinician to distinguish the prognosis of PQ poisoning patients, and it is also conducive to the rational distribution of medical resources.

The lethal dose for PQ poisoning is relatively low: 8–10 mL of 20% PQ solution can cause fatality; 10–20 mL can lead to irreversible fibrosis, respiratory failure, and death in a few weeks; and >20 mL can cause multi-organ failure and death within 1–4 days (Choi et al., 2013; Yu et al., 2014). The survival rate of patients with acute PQ poisoning drastically decreases with increasing dose (Gil et al., 2014). The amount of PQ ingested is closely tied to the prognosis of patients and is an important factor in predicting outcome. However, some studies showed no association between the amount of PQ ingested and organ damage (Gil et al., 2008; Liu et al., 2012). This could be due to inaccurate and incomplete assessment of dosage. Our study identified that the amount of PQ ingested was greater in the mortality group than the survival group (p < 0.01). The high sensitivity (0.66) and specificity (0.89) of this factor showed that it can independently determine prognosis. When the amount of PQ ingested was greater than 36.5 mL (20% aqueous solution), 82% of patients had a poor prognosis. The amout of PQ ingested was an estimated indicator and the main limiting factor of the method. However, we used empirical methods to estimate the amount of PQ poisoning in patients in terms of the size of a “mouthful,” which is a relatively accurate index. Additionally, only those patients within 24 h of poisoning were included, because they could better recall regarding the amount ingested.

When PQ enters the tissues and organs, a series of oxidative stress reactions occur, resulting in a large number of reactive oxygen species. These attack tissue, leading to organ damage and/or failure. During this process, WBC, cytokines, and other inflammatory factors are produced, leading to acute lung injury and multi-organ dysfunction. Higher WBC count is associated with negative patient outcomes (Sun et al., 2016). We also found that WBC count at the time of admission was higher in the mortality group than that in the survival group in our study (p < 0.01). When the WBC count was higher than 16.5 × 10⁹/L, 90% of the patients had a poor prognosis. WBC count is a key indicator for early assessment of prognosis in acute PQ poisoning.

In the case of normal renal function, 80–90% of PQ in the blood is excreted unchanged in the urine within 6 h (Zhang et al., 2017). However, PQ can directly attack the renal tubules, leading to renal damage or acute failure. This reduces PQ excretion, which causes an accumulation of PQ in the tissues and organs and increases body damage (Molck and Friis, 1998; Sun et al., 2016). Urea nitrogen and serum Cr levels increase during this process. The rate of acute kidney injury caused by PQ poisoning can be as high as 64% (Kim et al., 2011). Kidney damage occurs secondary to other organ damage. Thus, the outcome of PQ poisoning depends largely on the severity of renal damage. Early renal dysfunction indicates severe disease and poor prognosis. Renal function is a key parameter in clinical practice to determine prognosis of PQ poisoning (Lee et al., 2002; Weng et al., 2017; Zhao et al., 2019). Some studies have pointed to urinary PQ as the most valuable indicator in estimating severity and prognosis at the time of admission (Liu et al., 2017b; Tan et al., 2013). These findings are similar to the results of our study, where urine PQ, urine protein, urea, and serum Cr were higher in the mortality group than in the survival group (p < 0.01). When other factors affecting prognosis did not change, the risk of death was doubled for every unit change of urinary PQ, increased by 72% for every unit change in urinary protein, and increased by 1.7% for every 1 µmol/L increase in serum Cr. Thus, these factors can be used as independent biomarkers for evaluating the prognosis of PQ poisoning.

Our study also identified that ALT, AST, blood glucose, CK, and LDH levels were significantly higher in the mortality group than in the survival group (p < 0.05). The degree of liver and heart function impairment in the mortality group was also more severe than in the survival group, but these factors were not included in the predictive prognosis equation. Early damage to heart and liver function might gradually improve with drug treatments and renal excretion of PQ. This again reinforces the significance of early assessing of renal function status in predicting the prognosis of acute PQ poisoning. The earlier that renal function becomes impaired, the worse the prognosis. Therefore, it is necessary to protect renal function as early as possible to reduce the concentration of PQ and inflammatory factors in the blood. Hormones, antioxidant drugs, and hemoperfusion combined with hemodialysis can be administered to replace the kidney detoxification function and reduce renal damage. Blood purification, in particular, can compensate for kidney dysfunction, effectively reducing the concentration of toxic molecules in the blood and promoting improvement.

In conclusion, this study could effectively predict the probability of death due to acute PQ poisoning. Age, amount of PQ ingested, urinary PQ, urinary protein, WBC, and serum Cr are the key factors affecting prognosis of patients. This study also identified the best diagnostic value thresholds for the amount of PQ ingested, urinary PQ, urinary protein, WBC, and serum Cr in assessing prognosis. The sensitivity and specificity of these five indicators were high and they independently predicted the outcomes of acute PQ poisoning. The combination of these five indicators were considered to be more objective, accurate, and predictive than any previously described metrics. This study is of great practical significance for primary clinicians in establishing a prognosis for patients with early acute PQ poisoning and a clinical treatment plan.

Footnotes

Author contribution

Yiwei Su and Weiwei Liu have contributed equally to this work and should both list as the first author.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Major Program of Industry-University-Research Collaborative Innovation of Guangzhou (201704020177), the Key Medical Disciplines and Specialities Program of Guangzhou, and Key Laboratories in Guangzhou.

ORCID iD

Yimin Liu

References

Cao

Song

Bai

, et al. (2019) Neutrophil-lymphocyte ratio as an early predictor for patients with acute paraquat poisoning: a retrospective analysis. Medicine 98(37): e17199.

Choi

Jou

, et al. (2013) The dose of cyclophosphamide for treating paraquat-induced rat lung injury. Korean Journal of Internal Medicine 28: 420–427.

Gawarammana

Buckley

Mohamed

, et al. (2017) High-dose immunosuppression to prevent death after paraquat self-poisoning—a randomised controlled trial. Clinical Toxicology (Philadelphia, Pa) 56: 1–7.

Gil

Hong

Jang

, et al. (2014) Diagnostic and therapeutic approach for acute paraquat intoxication. Journal of Korean Medical Science 29: 1441–1449.

Gil

Kang

Yang

, et al. (2008) Association between plasma paraquat level and outcome of paraquat poisoning in 375 paraquat poisoning patients. Clinical Toxicology (Philadelphia, Pa) 46: 515–518.

Hou

Tian

, et al. (2017) Factors influencing prognosis of patients with acute paraquat poisoning within 24 hours of admission. Journal of Clinical Research 34: 1051–1053.

Hong

Tang

, et al. (2017) Early metabolome profiling and prognostic value in paraquat-poisoned patients: based on ultraperformance liquid chromatography coupled to quadrupole time-of-flight mass spectrometry. Chemical Research in Toxicology 30: 2151–2158.

Huang

Xuan

, et al. (2017) The value of APACHE II in predicting mortality after paraquat poisoning in Chinese and Korean population: a systematic review and meta-analysis. Medicine (Baltimore) 96: e6838.

Jamshidi

Fathi

Davoodzadeh

(2017) Investigation paraquat poisoning in southwest of Iran from sign to mortality and morbidity. Archiwum Medycyny Sadowej i Kryminologii 67: 35–45.

10.

Jiang

Liu

, et al. (2012) Determination and clinical analysis of urinary paraquat concentration in 18 cases of paraquat poisoning. China Journal of Emergency Resuscitation and Disaster Medicine 7: 480–481.

11.

Kanchan

Bakkannavar

Acharya

(2015) Paraquat poisoning: analysis of an uncommon cause of fatal poisoning from Manipal, South India. International Journal of Toxicology 22: 30–34.

12.

Kavousi-Gharbi

Jalli

Rasekhi-Kazerouni

, et al. (2017) Discernment scheme for paraquat poisoning: a five-year experience in Shiraz, Iran. World Journal of Experimental Medicine 7: 31–39.

13.

Kim

Gil

Yang

, et al. (2011) Serum uric acid level as a marker for mortality and acute kidney injury in patients with acute paraquat intoxication. Nephrology Dialysis Transplantation 26: 1846–1852.

14.

Lee

Hwang

Yang

, et al. (2002) Predictors of survival after acute paraquat poisoning. Toxicology and Industrial Health 18: 201–206.

15.

Liu

Jin

, et al. (2012) Influence factors for prognosis in patients with acute oral poisoning with paraquat. Chinese Journal of Industrial Hygiene and Occupational Diseases 30: 686–687.

16.

Liu

Zhan

Gao

, et al. (2020) Cyclosporine attenuates paraquat-induced mitophagy and pulmonary fibrosis. Immunopharmacology and Immunotoxicolgy 42(2): 138–146.

17.

Liu

Wang

Zhou

, et al. (2016) Hyperamylasemia as an early predictor of mortality in patients with acute paraquat poisoning. Medical Science Monitor 22: 1342–1348.

18.

Liu

, et al. (2017a) Predictive values of urine paraquat concentration, dose of poison, arterial blood lactate and APACHE II score in the prognosis of patients with acute paraquat poisoning. Experimental and Therapertic Medicine 14: 79–86.

19.

Liu

Wang

(2017b) Hypokalemia is a biochemical signal of poor prognosis for acute paraquat poisoning within 4 hours. Internal and Emergency Medicine 12: 837–843.

20.

Molck

Friis

(1998) Transport of paraquat by isolated renal proximal tubular segments from rabbits. Pharmacology and Toxicology 83:208–213.

21.

Oghabian

Williams

Mohajeri

, et al. (2019) Clinical features, treatment, prognosis, and mortality in paraquat poisonings: a hospital-based study in Iran. Journal of Research in Pharmacy Practice 8(3):129–136.

22.

Rashidipour

Karami-Mohajer

Mandegary

, et al. (2020) Where ferroptosis inhibitors and paraquat detoxification mechanisms intersect, exploring possible treatment strategies. Toxiclogy 433-434: 152407.

23.

Senarathna

Eddleston

Wilks

, et al. (2009) Prediction of outcome after paraquat poisoning by measurement of the plasma paraquat concentration. QJM-an International Journal of Medicine 102:251–259.

24.

Shi

Bai

Zou

, et al. (2017) The value of plasma paraquat concentration in predicting therapeutic effects of haemoperfusion in patients with acute paraquat poisoning. PLoS One 7: e40911.

25.

Sun

Shin

Yoon

, et al. (2016) Predicting the probability of survival in acute paraquat poisoning. Kidney Research and Clinical Practice 35: 102–106.

26.

Tan

Letchuman Ramanathan

Choy

, et al. (2013) Paraquat poisoning: experience in hospital Taiping (year 2008–October 2011). Medicine Journal of Malaysia 68: 384–388.

27.

Wang

Zhang

Feng

, et al. (2018) Sequential organ failure assessment in predicting mortality after paraquat poisoning: a meta-analysis. PLoS One 13(11): e0207725.

28.

Weng

Chen

, et al. (2017) Predictors of acute kidney injury after paraquat intoxication. Oncotarget 8: 51345–51354.

29.

Liu

Wang

, et al. (2011) Clinical analysis of 292 patients of paraquat poisoning. Medical Journal of the Chinese People’s Armed Police Forces 22: 189–191.

30.

Yang

(2015) Paraquat price cycle return in 2015. Pesticide Data 22: 37–38.

31.

Kan

Jian

, et al. (2014) A case report of acute severe paraquat poisoning and long-term follow-up. Experimental and Therapeutic Medicine 8: 233–236.

32.

Zhang

Liu

, et al. (2017) Evaluation of the prognosis of patients with acute paraquat poisoning by semi-quantification test of urine paraquat. Practical Pharmacy and Clinical Remedies 20: 683–686.

33.

Zhao

Song

Gao

, et al. (2019) Monocytes as an early predictor for patients with acute paraquat poisoning: a retrospective analysis. Biomed Research International 2019: 6360459.

34.

Zyoud

(2018) Investigating global trends in paraquat intoxication research from 1962 to 2015 using bibliometric analysis. American Journal of Industrial Medicine 61: 462–470.