The Clinical Characteristics of Intestinal Flora Imbalance in Stable Chronic Obstructive Pulmonary Disease (COPD) and Construction of Early Warning Model

Bcackground There is still a lack of predictive models for early identication of intestinal ora imbalance in stable Chronic Obstructive Pulmonary Disease (COPD) patients. We assessed the risk factors related to intestinal ora imbalance in COPD patients, and established a simple predictive model and scoring rules suitable for clinical medical staff in the respiratory department. Methods From January 1, 2018 to December 31, 2019, COPD patients (195 cases), who attended the Outpatient Department of Respiratory and Critical Care of Yixing Hospital of Jiangsu University, were collected for a cross-sectional study. The patients were divided into the experimental groups (41 cases) and the control group (154 cases) based on the results of stool examination. By single-factor analysis and logistic regression analysis, the baseline data of two groups were performed to obtain a new prediction model, and then simplied it. Results The ve predictive factors including body mass index(BMI), serum albumin(ALB), charlson comorbidity index(CCI), gastrointestinal symptom score(GSRs), and Global Initiative for Chronic Obstructive Lung Disease(GOLD) classication entered the model. The area under the ROC curve of the model for predicting intestinal ora imbalance in patients with stable COPD is 0.953 [95%CI (0.924, 0.982)], further simplifying the scoring rules, and the area under the ROC curve is 0.767 [95%CI (0.676, 0.858)]. study, intestinal ora COPD, then the Furthermore, all and predicted PaO2:arterial BNP:B-type Hb:hemoglobin;WBC:white blood characteristic

being absorbed into blood; And TMAO was associated with long-term all-cause mortality in patients with COPD[6].On the other hand, the increased metabolic demand of patients with COPD could lead to ischemia and hypoxia of intestinal mucosal, damage of the intestinal tract integrity, and then cause the imbalance of intestinal ora [7].Therefore, even stable COPD patients should be alert to the occurrence of intestinal ora imbalance, and to the identi cation of vulnerable patients. In the study, an early warning model and speci c scoring rules was established for intestinal ora disorders in patients with stable chronic obstructive pulmonary disease, so as to be more intuitive and convenient for clinicians to identify early patients, and then carry out intervention to improve their prognosis. Furthermore, the model was also used as a risk strati cation basis for intestinal ora disorders in stable COPD patients. respiratory physicians helped to con rm that it was stable: there was no acute exacerbation in the past 8 weeks; there is no clear history of infection; sign the relevant informed consent form of this study; (2) the diagnostic criteria of intestinal ora disorders in the recommendations for diagnosis and treatment of intestinal ora disorders [9,10]: intestinal ora disorders can be diagnosed by meeting any of the following laboratory tests. Fecal laboratory examination: the number of enterococci and bacilli was observed and counted under oil microscope, and it was found that the proportion of cocci was more than 40%; Bi dobacterium/Enterobacter < 1; abnormal type of G − bacilli was 50%. Abnormal type of Grampositive bacilli < 68%; (3) complete clinical data [including demographic characteristics, fecal ora smear or culture results, partial pressure of oxygen, lung function] and so on. Exclusion criteria: there were primary intestinal diseases causing intestinal ora imbalance; diseases such as pulmonary interstitial brosis and active tuberculosis; antibiotics, systemic intravenous hormones, probiotics and so on were used 8 weeks before admission; the data were incomplete or dropped out. A total of 213 patients with stable COPD were enrolled.18 patients were deleted because of primary intestinal diseases, pulmonary interstitial brosis, use of antibiotics or probiotics in the past 8 weeks, incomplete data, midway withdrawal and so on.Finally,195 patients were divided into experimental group (41 cases) and control group (154 cases) according to whether they were diagnosed as intestinal ora imbalance or not, as shown in Fig. 1.

Chalson Comorbidity Index Questionnaire
Charlson Comorbidity Index (CCI) is a scale based on the risk of complications and death. In this study, the CCI score was calculated according to the basic condition, and the speci c score scale was as follows: myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic lung disease, connective tissue disease, ulcer, mild liver disease, diabetes mellitus, hemiplegia, moderate/severe kidney disease, diabetes with organ damage, tumor, leukemia, lymphoma, etc. The patients with moderate/severe liver disease were given 3 points, metastatic tumor and acquired immunode ciency syndrome were given 6 points. A total of 19 items were assigned 1, 2, 3 and 6 points respectively, with a total score of 0-36 points.
Chronic Obstructive Pulmonary Disease Assessment Test Score (Cat) CAT score is a comprehensive questionnaire on the health damage of COPD patients, including 8 items,which can be used to observe the effects of COPD on patients by evaluating cough,expectoration,chest tightness,sleep,energy,mood and activity ability. Patients score each item according to their own conditions (0-5), CAT scores range from 0-40).0-10:"slight impact";11-20:"moderate impact";21-30:"serious impact";31-40: "very serious impact".

Modi ed British Medical Research Council Questionnaire(Mmrc)
It is a questionnaire score scale used to evaluate the severity of dyspnea in patients with COPD. The details are as follows: it is divided into 0-4 grades, and the higher the grade, the more severe the dyspnea. Level 0: wheezing occurs only when exercising hard; level 1: di culty breathing when walking on at ground or climbing a small slope; level 2: walking on at ground is slower than others of the same age and needs to stop and have a rest; level 3: need to stop and rest after walking on at ground for about 100 meters or a few minutes; level 4: unable to leave home due to severe breathing di culties, or di culty breathing when getting dressed or undressing.

Stool And Blood Specimen Collection
All subjects collected 10ml of fasting elbow venous blood and placed it in the 10ml anticoagulant tube. After collection, the blood vessels were turned upside down several times to fully mix. After being placed at room temperature for 1 hour, the plasma was separated by centrifugation at a rotational speed of 3500 rpm/min for 10 min. The supernatant was absorbed and packed in the Eppendorf tube and stored in the refrigerator at -80℃. At the same time, the fecal samples are collected in the special and clean urinal, and only the middle part of the feces is sampled to remove the surface part. In the process of sampling, avoid contact with the inner surface of the bottom of the container, ensure no urine and other pollution in the whole process, use a special aseptic spoon to move the feces around 2ml to the Eppendorf tube and immediately cover them tightly, mark them, and quickly store them in the refrigerator at -80℃. The blood and fecal samples were sent for examination within 4 hours, and the feces were examined and cultured under microscope.

Statistical Analyses
According to the results of stool microscopy and culture, it was divided into two groups with stable COPD with intestinal ora imbalance and without intestinal ora imbalance. The general clinical data characteristics were analyzed. The data was entered using Excel 2016 and SPSS 22.0 statistical software was used for analysis. The measurement data conformed to the normal distribution, expressed as x̄ ± s, the comparison between the two groups was expressed by the t-test, the non-normal distribution was expressed by M (P25, P75), and the comparison of the differences between the groups was expressed by the Mann-Whitney U rank sum test. Count data is expressed in frequency (composition ratio). The comparison between the two groups is performed by χ² test, and statistically signi cant variables are included. ROC curve is drawn to obtain the best cut-off value and Youden index of the above variables, and the continuous variables in the variables according to the cut-off value transformed into a binary variable, a multi-factor logistic regression analysis was performed to establish a preliminary prediction model. The t of the obtained model is tested by Hosmer & Lemeshow. In the result, the corresponding βvalue is given as an integer to assign a score, and a simpli ed early warning model scoring rule is established. The ROC curve veri cation of the scoring model, the use of GraphPad Prism software to draw the ROC curve, and the ROC curve to verify the effectiveness of the early warning scoring model for intestinal dysbiosis in COPD. P<0.05 is considered as statistically signi cant.

Comparative analysis of base data of intestinal ora imbalance in patients with COPD
Compared with the control group, the case group had lower body mass index, worse pulmonary function, lower partial pressure of arterial oxygen and lower albumin concentration(all P < 0.05). The median CAT scores of the case group and the control group were 14 (9,19) and 9 (7,14). The median mMRC scores were 2 (1,3) and 1 (0,2). The CCI scores of 3(2,4)and (0.75,2) GSRs scores were 2(1,4) and 1(0,1).
Therefore, the case group had higher CAT, mMRC, CCI and GSRs scores than the control group (all P < 0. 05), The other factors were not signi cantly different between the two groups (P > 0.05),as shown in Table 1.
The ROC curve of the model and its variables for predicting the occurrence of intestinal ora imbalance in patients  Fig. 2,3 .

Simplify The Resulting Model
After the statistically signi cant regression coe cients of BMI, ALB, CCI score, GSRs score, and GOLD classi cation in the model were taken as integers (that is, the values were 2, 1, 3, 3, 3), a simpli ed scoring model(PRE_2)was established,Where BMI ≤ 23.25kg/m 2 is assigned 2 points, > 23.25kg/m 2 is assigned 0 points; ALB ≤ 32.5g/L is assigned 1 point, > 32.5g/L is assigned 0 points; CCI score ≤ 2 points are assigned 0 points,>2 points are assigned 3 points; GSRs score ≤ 2 points are assigned 0 points, > 2 points are assigned 3 points; GOLD classi cation: GOLD I-II are assigned 0 points, GOLD III-IV are assigned 3 points. The score of this simpli ed model is 0-12 points, as shown in Table 5.The AUC value of the simpli ed model is 0.767 [95%CI (0.676, 0.858)]. This model(PRE_2)has a moderate predictive power for the risk of intestinal ora imbalance in COPD patients, as shown in Fig. 4. Use SPSS software to determine that the best critical value of the above model is score > 7 ( nally rounded according to clinical analysis), which is used as the basis for risk strati cation of intestinal ora imbalance in patients with stable COPD, and patients can be divided For the occurrence of low-risk group (≤ 7 points) and highrisk group (> 7 points).The ROC curve is shown in Fig. 4.  GOLD I-II  0   GOLD III-IV  3 Total score 0-12

Discussion
In this study, it was found that BMI, ALB, CCI score, GSRs score and GOLD classi cation were the risk factors of intestinal ora imbalance in patients with COPD and had good predictive value for the prognosis of COPD patients. Among them, CCI score > 2 and GSRs score > 2 increased the risk of intestinal ora disorders in stable COPD patients by 25.633 times and 17.113 times, respectively, whereas BMI > 23.25kg/m 2 and ALB > 32.5g/L might be protective factors, which decreased the risk of intestinal ora disorders in stable COPD patients by 0.090 times and 0.278 times, respectively. BMI is often regarded as indicator of the body's nutritional status and reports also demonstrated the association of low BMI with exacerbation in patients with COPD [11]. A 3-year South Korean study of follow-up BMI to predict the prognosis of COPD found that the decrease in BMI was independently related to the deterioration of COPD and the increase of mortality (HR = 3.167) [12]. COPD patients is often accompanied with increased energy expenditure, when malnutrition occurs, the body's defense ability decreases, and then results in intestinal ora imbalance followed deterioration of COPD. In addition, serum albumin is another indicator of nutritional status, because it can well predict in-hospital mortality (ALB level is 30.5g/L) and readmission rate within 30 days (ALB level is 30.1g/L) in patients with COPD [13]. When hypoalbuminemia occurs, it not only leads to malnutrition and low immunity in patients with COPD, but also leads to decrease of colloidal osmotic pressure, edema of intestinal wall and structural disorder of intestinal mucosa, thus affecting the ecological stability of intestinal ora and resulting in the occurrence of ora imbalance [14,15]. Furthermore, albumin / globulin ratio has more predictive value than the single index of serum albumin, and it can be used as a reference index to evaluate the condition of elderly patients with COPD [16]. Although albumin / globulin ratio could determine whether COPD patients were complicated with infection and predicted prognosis, but it was not found in this study. An Italian study showed that CCI score is an independent risk factor for acute exacerbation and death in patients with COPD; the death rate of patients with COPD increased by 4.8% when CCI > 1 [17]. Both COPD and its complications can aggravate the weakness or disability progress of COPD patients [18], and further be more likely to lead to intestinal ora imbalance. The interaction between them accelerate the disease process and the number of acute exacerbation of COPD. In addition, some studies showed that the score of gastrointestinal symptoms in patients with COPD was higher than that in healthy individuals (2.12 ± 1.96), it showed that there was a negative correlation between gastrointestinal symptoms and mental health impairment (r=-0.49,p < 0.001) [19].COPD patients were more apt to psychological disorders, such as anxiety and depression, which could easily lead to the disappearance of intestinal dominant ora, the loss of intestinal wall barrier function and the release of intestinal endotoxin, which jointly promoted the occurrence of intestinal ora imbalance and leaded to the gastrointestinal symptoms [20]. Compared with non-COPD patients and COPD GOLD I patients, GOLD II-IV patients have signi cantly higher symptoms(CAT score) and lower lung function variables (including FEV 1 /FVC, FVC); Especially, COPD patients with GOLD III-IV are considered to have good clinical signi cance for judging the prognosis of COPD [21].Furthermore, Patients with moderate to severe pulmonary insu ciency tend to have more clinical symptoms and a higher risk of acute exacerbation. According to the Global Initiative for Chronic Obstructive Lung Disease (GOLD), this type of population is usually divided into GOLD D group. Compared with A group patients group A patients has a more higer risk of death (25 times) [22], which indicates that worse lung function can easily lead to acute exacerbation of COPD, and then aggravate intestinal ora imbalance. Moreover, age is also regarded as a risk factor for intestinal ora disorders in patients with stable COPD. The older the patients are, the more serious the decline of organ function and gastrointestinal function are; In addition, the decline of gastrointestinal function is prone to enterotoxin accumulation, and then leads to ora imbalance and accelerates COPD [23]. However, age is not found to be one of the risk factors attributed to the majority of elderly patients in this study (p>0.05).
The predictive value of the model for intestinal ora imbalance in patients with stable COPD is better than a single variable. The area under the ROC curve is 0.953, the sensitivity is 90.2%, and the speci city is 87.7%, which indicates that the predictive model has good clinical predictions. So it provides a basis for early screening and identi cation of prone patients. However, the calculation process of the prediction model is relatively cumbersome, and it is not suitable for rapid clinical judgment and early identi cation of prone patients. Therefore, we further simpli ed the model, and assigned the corresponding β-values of the 5 predictors to integer scores to obtain the stable period. Predictive scoring rules for intestinal ora imbalance in COPD patients is:BMI ≤ 23.25kg/m 2 (2 points),ALB ≤ 32.5g/L(1 point), CCI score > 2(3 points), GSRs score > 2(3 points), GOLD classi cation III-IV(3 points). Fast screening of high-risk patients who are prone to intestinal ora imbalance by the scores has good predictive performance, with an AUC of up to 0.767. The simpli ed scoring model allows us to intervene in this group of people early, so as to extend the stable period of COPD and improve them prognosis.

Limitation Of The Study
However, it is a single-center retrospective study and the number of observation samples is small; Besides, the factors, such as the use of antibiotics and the total dose of glucocorticoids, which has a certain impact on the authenticity of the conclusions of the study, are not included in the analysis. Therefore, it needs to be veri ed by further expanding the sample size, prolonging the follow-up time and conducting a multicenter prospective study.In this study, only the ROC receiver operating curve was used to evaluate the value of the simpli ed scoring model in predicting the occurrence of intestinal dysbiosis in patients with stable COPD.We will prospectively include stable COPD patients as veri cation in the next experiment.Cohort, to further verify the clinical test e cacy of the model, and to further optimize it for better clinical application.

Conclusion
Finally, the simpli ed scoring model(PRE_2)established by the ve indicators of BMI, ALB, CCI score, GSRs score, and GOLD classi cation showed moderate predictive value for intestinal dysbiosis in patients with stable COPD. All the indicators are easy to detect and obtain quickly. The scoring model can identify high-risk groups that are prone to intestinal ora imbalance, and implement intervention measures, such as oral probiotics, as soon as possible. By adjusting the steady state of the intestinal ora, the goal of reducing the number of acute attacks of COPD patients and improving the prognosis is achieved. Figure 1 Patient fow diagram ROC curve of simpli ed scoring rules(PRE_2) for risk prediction of intestinal ora imbalance in patients with stable COPD