Effect of pulmonary infection on convalescent patients with intracerebral hemorrhage
The disease course of intracerebral hemorrhage is divided into hyperacute phase, acute phase and recovery period based on the clinical progression of the disease. Studies have shown that 40% of ICH patients develop pulmonary infection during the recovery period; this greatly impacts rehabilitation and prognosis of the patients [14-16]. There are many causes for the occurrence of pulmonary infection during the recovery period of intracerebral hemorrhage, including age, disturbance of consciousness, tracheotomy, and nasal feeding [17-20]. In this case, the above factors can cause a decrease in immunity and an increase in the invasion and reproduction ability of bacteria, thereby resulting in pulmonary infection. Consistently, numerous studies have revealed a high incidence of pulmonary infections in the recovery period of intracerebral hemorrhage. Xu et al. found that the incidence of postoperative pulmonary infection in 264 patients with hypertensive cerebral hemorrhage was 19.70% [20]. Barlas et al. showed that out of 58,586 ICH patients, 9.6% developed pulmonary infection [21]. In this study, we observed that the incidence of concurrent pulmonary infection in convalescent patients with intracerebral hemorrhage was 33.77%, which was higher than that in the aforementioned studies. This difference may be attributed to several factors including advanced age, long bed rest, and a high percentage of coma patients with disturbance of consciousness in the present study. A high incidence of concurrent pulmonary infection in convalescent patients with intracerebral hemorrhage mainly increases the incidence of other adverse outcomes. It has been demonstrated that pulmonary infection did not affect the death of hospitalized patients in the acute phase of intracerebral hemorrhage, but increased the risk of death in hospitalized patients in the recovery period of intracerebral hemorrhage [22]. Another study showed that as one of the important factors, pulmonary infection leads to prolonged hospitalization of ICH patients, which makes patients more likely to be exposed to pathogens during hospitalization, thus aggravating pulmonary infection or causing other infections [23]. Pulmonary infection may also lead to sepsis, respiratory failure and epilepsy in convalescent patients with intracerebral hemorrhage [21, 24, 25]. Therefore, pulmonary infection has a huge impact on the prognosis of convalescent patients with intracerebral hemorrhage, and it is necessary for clinical staff to identify ICH patients at high risk of pulmonary infection as early as possible.
Analysis of risk factors for concurrent pulmonary infection in convalescent patients with intracerebral hemorrhage
Advanced age is almost the most frequent risk factor for nosocomial infections [26]. It has been shown that increasing age is highly correlated with the incidence of pulmonary infection in ICH patients [27]. Sui et al. found that for every one-year increase in age, the risk of stroke-related pulmonary infection increases by 1.113 times [28]. In this study, we observed that the risk of developing pulmonary infection was 1.683 times higher in the patients over 60 years old than in those under 60 years old. This observation may be explained by the fact that compared with the younger patients, the elderly ones have a significantly weaker immune system and a greatly reduced ability to eliminate pathogens for an effective control the infections. In addition, the decline in the ciliary motility of the bronchial mucosa, cough response, and lung tissue elasticity in the elderly patients weakens the expectoration function, thereby increasing the chance of pulmonary infection after illness.
A meta-analysis by Westendorp et al. in 2012 found that prophylactic antibiotic use reduced infection rates (hazard ratio = 0.58) [29]. On the contrary, another meta-analysis by Yuan et al. revealed a 7.10-fold increase in the risk of developing pulmonary infection in patients undergoing prophylactic antibiotic use compared to those who did not receive prophylactic antibiotics [30]. Similar to the research of Yuan et al, the present study found that the risk of pulmonary infection in patients undergoing prophylactic antibiotic use was 5.507 times as high as that in patients who did not receive prophylactic antibiotics. The opposite results between the above two meta-analyses may be attributed to the difference in the type of literature involved in the analyses (randomized controlled trial vs case-control and cohort study). Additionally, the study population and the disease severity differed in the two studies. In the present study, convalescent patients with intracerebral hemorrhage were selected as the study subjects. Thus, the overall condition of patients was relatively stable. In the analysis conducted by Westendorp et al., while prophylactic antibiotic use displayed a positive significance for patients with severe stroke, its significance for the mild patients may be exaggerated.
Convalescent patients with intracerebral hemorrhage with lower GCS scores have more severe disturbance of consciousness, and are more likely to die from ischemia and hypoxia in brain tissue due to cerebral edema and intracranial pressure. In turn, these complications may exacerbate damage to the central nervous system and impair the patient's cough reflex. In this study, we observed that the incidence of pulmonary infection in patients with moderate or higher disturbance of consciousness was 8.879 times as high as that in patients with mild or less consciousness, showing that the lower the score of disturbance of consciousness, the higher the incidence of lung infection. This observation is consistent with the results of the previous studies [28, 31]. In terms of tracheotomy, increased exposure to the external environment stimulates the production of more secretions from the airways, which compromises local defenses and provides conditions for bacterial colonization, thus increasing the incidence of pulmonary infection [32]. In this study, we found that the incidence of pulmonary infection in patients undergoing tracheotomy was 8.931 times as high as that in patients who did not receive tracheotomy. This finding was similar to the results of the previous researches [19, 33].
Numerous studies have identified dysphagia and bed rest time as the most important factors causing pulmonary infection in ICH patients [30, 34]. It has been demonstrated that the presence of normal chewing and swallowing not only prevents the accumulation of bacteria in the throat, but also ensures proper secretion of secretions. Therefore, patients with reduced or absent cough and swallowing reflexes are more prone to pulmonary infection. A study conducted by Walter et al clearly showed that dysphagia promotes an increased probability of aspiration and increases the risk of pulmonary infection by 10 times [35]. Brogan et al also found that about half of ICH patients had dysphagia, while 48% of them developed pulmonary infection [36]. It has not yet been determined whether the length of bed rest acts as an independent risk factor for pulmonary infection. Ward et al showed that prolonged bed rest led to the accumulation of secretions in the lower part of the trachea, which makes the patients be more prone to difficulty in expectoration, thus resulting in an increase in the incidence of pulmonary infection [37]. Mao et al found that prolonged bed rest had the highest prognostic value for pulmonary infection, with an AUC value of 0.908 [38]. In this study, we observed that the incidence of pulmonary infection in patients with dysphagia was 2.709 times as high as that in patients without dysphagia, while the incidence of pulmonary infection in patients who had been bedridden for more than 2 months was 3.402 times as high as that in patients who had been bedridden for no more than 2 months. This observation was similar to the findings of Mao et al [38].
The present study identified nasal feeding as another important risk factor for the pulmonary infection. A meta-analysis showed a 9.87-fold increase in the risk of pulmonary infection in ICH patients with nasal feeding [30]. Here, we found that the incidence of pulmonary infection was increased by 2.528 times in patients with nasal feeding compared to those without nasal feeding. This finding could be explained by the fact that nasal feeding leads to oral flora dysbiosis, making oral care difficult. Besides, nasal feeding may impair gag reflex and cough function of the patients. Studies have shown that the placement of a nasogastric tube allows for a direct connection between the respiratory tract and the external environment, which can serve as an important entry point for pathogenic microorganisms. Serum biomarkers such as PCT have been widely used to predict bacterial infection and severity [39]. A growing number of studies have clearly indicated that elevated serum PCT levels are highly correlated with pulmonary infections following intracerebral hemorrhage [40]. A study by LU et al suggested a high predictive efficacy of PCT for pulmonary infections after intracerebral hemorrhage based on the ROC curve area [41]. Similarly, the present study showed that the probability of pulmonary infection in patients with abnormal PCT was 2.406 times as high as that in patients with normal PCT.
The nomogram model has predictive ability for pulmonary infection in convalescence patients with intracerebral hemorrhage
The present study not only described the independent risk factors for pulmonary infection in convalescence patients with intracerebral hemorrhage, but also created a clinical predictive nomogram for early identification of high-risk patients through data analysis. In this study, both univariate logistic regression and Lasso regression were used to perform screening, and the screened risk factors were included in multivariate logistic regression analysis to build a prediction model. This study identified age, prophylactic antibiotic use, disturbance of consciousness, tracheotomy, dysphagia, length of bed rest, nasal feeding, and PCT as independent risk factors for convalescent patients with intracerebral hemorrhage. The analysis showed that the AUC of the prediction model for concurrent pulmonary infection in the recovery period of intracerebral hemorrhage was 0.901 with 95% CI of 0.878 to 0.924. The repeated sampling by Bootstrap for 1000 times yielded an AUC of 0.900 with 95% CI of 0.877 to 0.923. In this case, the AUC value decreased by only 0.001, indicating that the model has an excellent discrimination. The Hosmer-Lemeshow test revealed a high degree of agreement between the predicted and actual probabilities, which is almost close to 1 (P=0.982), showing that the model has a good goodness of fit. The prediction model constructed in this study based on the DCA decision curve can help clinical staff to make better clinical decisions. Given that a relatively complicated calculation is required for equation-based prediction model construction, the present study used the nomogram to visualize the prediction model. With this approach, clinical staff can perform dynamic risk assessment based on the identified risk factors for each individual patient, screen patients at high risk of pulmonary infection, and accurately predict the risk of pulmonary infection in convalescent patients with intracerebral hemorrhage. Therefore, this study can help clinical staff to target the risk factors and improve the precise prevention of concurrent pulmonary infection in convalescent patients with intracerebral hemorrhage.
In sum, convalescent patients with intracerebral hemorrhage are at a higher risk for concurrent pulmonary infection. Although the occurrence of pulmonary infection in convalescent patients with intracerebral hemorrhage cannot be eradicated at present, it is necessary to make appropriate predictive interventions based on risk factors. In this study, ICH patients in the recovery period were selected as the research subjects because a high prevalence of pulmonary infection occurs in the recovery period, and this period is a critical stage for prognosis of ICH patients. Identification of independent risk factors for pulmonary infection and construction of a clinical prediction model will enable clinical staff to perform early assessment and screening. This study still has several limitations. First, this was a retrospective analysis of single-center studies. A multicenter and prospective trial is needed to validate the accuracy of the model. Second, not all risk factors were included in this study. Other potential predictors such as infection need to be investigated in future studies. Lastly, external validation is needed to further evaluate the performance of the nomogram model.