LTH is a medical emergency with high mortality[6]. Although the mortality rate associated with hemoptysis has decreased with the development of technology and advances, the LTH mortality rate remains high (9–38%). In our study, in-hospital mortality was 7.5% (3/40). LTH symptoms are sometimes mild but can become severe over a short period and even quickly lead to death. LTH has a short therapeutic window and is not premonitory. Assessing whether it is LTH can be challenging. As a result, LTH management requires an experienced multidisciplinary team of physicians (intensivists), radiologists, pulmonologists, interventional radiologists, and thoracic surgeons in the intensive care unit (ICU) with specialized rescue procedures and pre-agreed plans. Skillful coordination between these doctors can improve patient management efficiency, simplify procedures, and save more time for patient treatment[11].
The average total volume of conducting airways is 150 mL in adults [12]. The amount of hemoptysis that can be transported by the airways is small. Therefore, even a small amount of blood can quickly overwhelm the lungs’ ability to oxygenate and ventilate. Thus, the severity of hemoptysis should be based not only on the amount of blood spat out but also on the clinical context. The patient’s ability to clear blood from the airway and the patient’s underlying physiological reserve are very important in assessing the severity of hemoptysis. The time window for LTH treatment is too small to conduct enough investigations to assess the patient’s airway and underlying physiological reserve. In the emergency setting, the patient’s airway and underlying physiological reserve can only be assessed through history and pulmonary radiological examination. For hemoptysis evaluation, a chest X-ray is not recommended and computed tomography (CT) is considered a better choice. In a retrospective study, CT scan identified the location and cause of bleeding in 70% and 77% of cases[13], respectively. If time permits, computed tomography angiography (CTA) is recommended because it can detect the cause and can lay down the vascular roadmap for therapeutic intervention[14].
For LTH, it is important to check the patient’s airway early and isolate the bleeding immediately to locate and control the bleeding[15]. Selective bronchial intubation and balloon tamponade with a bronchial blocker [1] are good options to control the airway and isolate bleeding in order to localize and control bleeding. Treatment under tracheoscopy is suitable for patients who require endotracheal intubation. However, when the amount of bleeding under the microscope is large, the field of vision is relatively limited; moreover, treatment under the tracheoscope often has certain limitations. BAE’s immediate clinical success varied from 70–99% [16]in various studies. BAE has good clinical success. It is considered as the first-line treatment in managing LTH. However, hemoptysis recurrence is common. The present study showed a recurrence rate of 18.9% (84 of 444 patients) after BAE within 3 years. For patients who underwent repeat BAE, the recurrence causes based on angiographic findings were categorized into the following: previously embolized vessel recanalization, missed culprit vessels, or new collateral circulation recruitment. Surgical therapy was selected as a last resort after BAE failure or bronchoscopy[17]. Andréjak et al. reported that in 111 lung resections with severe hemoptysis, the mortality rates were 35% when performed emergently compared with 0% in patients undergoing resection electively after hospital discharge [18].
According to our univariable analyses, LTH was associated with CRP, WBC, Hb, albumin, RBC, HCT, hemoptysis amount, and lung destruction. However, multivariable analysis demonstrated that only HCT, hemoptysis amount, and lung destruction were independent risk factors for LTH. Although the description of the hemoptysis amount varies among studies, hemoptysis amount remains an indicator of hemoptysis severity. Hemoptysis volume was shown to be an independent risk factor for LTH in our study. The cut-off value of amount was 225 ml. Hemoptysis amount > 225 ml was predicted to be more likely to develop LTH, whereas hemoptysis < 225 ml would be less likely to develop LTH. HCT was normal in 40–50% of men and 37–48% of women. Patients with hemoptysis often had decreased HCT due to the body's own compensatory mechanism or treatment with fluids during the resuscitative phase of shock. Blood loss = [(original measured HCT—post blood loss HCT)/original measured HCT] × weight × 7% × 1000. This study can truly reflect the hemoptysis volume and disease severity by HCT. It also showed that lung destruction was an independent LTH risk factor. Lung destruction reflects the decline of lung reserve capacity. Lung destruction would impair the lung’s ability to oxygenate and ventilate due to the increase of physiologic dead space. Lung destruction is more likely to develop LTH. In our opinion, pulmonary aneurysm[19] is an independent risk factor for LTH regardless of whether it presents with rupture or not. In this study, the pulmonary aneurysm factor was also included, but it was not statistically significant because the number of cases was too small.
According to the literature, different etiologies were associated with LTH, which will complicate the model establishment. Therefore, in order to construct a relatively simple and accurate prediction model, a subgroup analysis was performed with respect to tuberculous hemoptysis. In this study, a nomogram was developed for predicting LTH. This nomogram incorporated three variables, including HCT, hemoptysis amount, and lung destruction. These three variables were obtained from observing CT images, symptoms, and blood tests, which were all simple and easy to measure clinically. Because of the small data set, we were unable to conduct external validation. However, the nomogram has an AUC value of 0.814 on internal validation. The nomogram’s sensitivity and specificity were 90.1% and 62.5%, respectively. The nomogram showed good discriminatory ability and calibration, and the DCA evaluation showed its clinical usefulness.
After nomogram development, clinical application was conducted by our institution. Patients with hemoptysis were evaluated and triaged by the nomogram. Patients with a high LTH likelihood will be transferred to the ICU and those with a low likelihood will be transferred to a general ward[20]. Patients with a high LTH likelihood were treated according to the hemoptysis procedure and cared for 24 h. The materials and personnel needed for rescue were prepared in advance. Although further research is needed to confirm whether these policies are effective for LTH, it is meaningful that it is a practical attempt to help patients to gain enough time for treatment so appropriate treatment can be chosen. This preliminary model is expected to be a convenient decision-making tool to prevent LTH and make appropriate decisions to buy time for the patient's treatment.
This study has some limitations. First, to evaluate the model, an internal validation method was used which may have overestimated the model’s diagnostic accuracy. Therefore, prospective external validation of the model is needed. Second, data were collected from a single center; multi-agency, large-sample research is warranted. Prospective multi-center large sample studies might be needed in the future to further confirm the findings in our current study. Third, the lung damage index is difficult to quantify. If we can, we might be able to get more accurate models to predict LTH.
In conclusion, our study suggests that HCT, hemoptysis amount, and lung destruction are independent risk factors for LTH in patients with tuberculous hemoptysis. In addition, the predictive nomogram combining imaging features and clinical factors is expected to serve as a convenient decision-making tool to prevent LTH and make appropriate decisions to gain time for appropriate patient treatment.