Patient and Procedure-related Characteristics Correlated with Severity of Hemoptysis Following Percutaneous Transthoracic Needle Biopsy: Results of a 10-year Retrospective Review

Abstract

Purpose

A risk of percutaneous transthoracic needle biopsy (PTNB) is hemoptysis which can range from mild to life-threatening. The reported occurrence of hemoptysis is 1.7–14.5% and the demographic, patient, and procedure characteristics have not been extensively described. The purpose of this study was to assess the associations of demographic, patient, and procedure characteristics with the severity of hemoptysis.

Materials and Methods:

A single-institution, single group, retrospective, electronic medical record (EMR) review was performed on all hemoptysis events occurring between 2008 and 2018. Demographic, clinical, and procedure variables were extracted from EMRs. Outcome of hemoptysis events was graded using Common Terminology Criteria for Adverse Events (CT-CAE). Mild-moderate and severe hemoptysis were defined as CT-CAE classifications of 1–2 and 3–5, respectively. Associations were generated using logistic regressions and Likelihood Ratio Chi-Square tests.

Results:

In ten years, 14,665 PTNB resulted in 231 hemoptysis events occurring in 229 patients; 12.7% (n=29) of those were severe. The strongest and statistically significant variables associated with an increased likelihood of a severe event, if an event occurred, were cigarette pack years (OR=1.02, 95% C.I.=1.01–1.04, P=.020); history of chronic obstructive pulmonary disease (COPD) (OR=3.68, 95% C.I. = 1.53–8.82, P=.003); core biopsy technique (OR=8.13, 95% CI=1.07, 61.40, P=.042), and larger diameter needle (20g vs. 18g: OR= 2.60 (1.09, 6.17), P=.031).

Conclusions:

PTNB-associated hemoptysis was an uncommon event that was rarely life-threatening. The extent of the patient’s smoking history, the diagnosis of COPD, and core biopsy technique were associated with an increased likelihood of severe hemoptysis.

Introduction

Percutaneous transthoracic needle biopsy (PTNB) provides tissue needed for a pathologic diagnosis of suspicious lung lesions prior to the initiation of treatment. In cancer patients biopsy specimens are also used to identify molecular markers utilized in targeted therapies, either as part of established clinical treatment regimens or for novel cancer therapies under investigation.¹

The advantage of PTNB procedures is that they are minimally invasive with high rates of diagnostic accuracy and minimal risk of a complication.^2–6 Reported risks include pneumothorax, bleeding, air embolism, and tumor seeding.⁷ Bleeding following a PTNB has been described as 1) intraparenchymal or pulmonary hemorrhage which are radiologic diagnoses,^{8, 9} 2) hemoptysis, and 3) hemothorax.^{2, 4, 7, 10} Hemoptysis is the expectoration of blood that has originated from either the lungs or the bronchial tubes.¹¹ The reported frequency of hemoptysis is 0.3–15.5%.^{2, 5, 6, 8, 12–17} If bleeding following PTNB is detected early and controlled, patient outcomes are optimized. However, undetected or uncontrolled bleeding can be life-threatening and result in death.^{18, 19} The reported frequency of significant bleeds is 0%-1.6%^{14–16, 18, 20–22} with a death occurring at a rate of 1.1–2.0 in 10,000.^{18, 20}

Researchers have explored characteristics associated with bleeding following PTNB. Many studies have focused on the radiographic findings associated with hemoptysis including lesion depth, size, composition, and pulmonary features traversed. ^{2, 5, 6, 8, 10, 12–16, 22, 23} Research exploring host (patient and procedure) characteristics have consisted of retrospectively designed studies of patients who developed any type of pulmonary bleed.^{5, 6, 9, 10, 13–17, 23, 24} Female sex has had a positive effect on bleeding (OR 1.64–2.52 with male referent and OR 0.48–0.87 with female referent).^{6, 23, 24} The effect between females and bleeding is reported to be stronger with severe bleeding (OR 3.69).⁹ Correlations with age have varied among studies. Some researchers have found positive associations with older age and bleeding (OR1.52–2.92),^{6, 9} while others have found negative associations between older patients and hemoptysis (OR 0.61–0.72).^{14, 15} Positive associations have been reported with bleeding and the clinical diagnosis of COPD (OR1.44–1.64)^{9, 15, 23} and a smoking history (OR 2.30).¹⁵ Associations between the use of antithrombic agents prior to the procedure and bleeding have been inconsistent. Atwell et al.¹⁸ identified no cases of major hemorrhage in patients who had been on aspirin while other researchers found associations with hemoptysis and antithrombic use (OR 1.24–2.88).^{9, 14, 15} Finally, researchers have found that core biopsy technique compared to find needle aspiration (FNA) (OR 2.00 with FNA referent, OR 0.38–0.92 with core referent)^{3, 9, 13} and smaller gauge needles (larger diameter) (OR 1.05–2.30)^{9, 15, 23} were associated with hemoptysis.

Differences in study samples, evaluated characteristics, and definitions of bleeding and each characteristic makes comparison between studies challenging. Furthermore, most studies were limited by the sample size of rarely occurring, severe events (n=1 to 21).^{4, 6, 7, 9, 14–16, 18, 20} Understanding the differences between those who have self-limiting hemoptysis and those who develop severe hemoptysis can potentially help direct clinical interventions.

The purpose of this study was to describe associations of demographic, patient, and procedure characteristics with the severity of hemoptysis.

Methods

This study was an Institutional Review Board approved, retrospective review of all patients who underwent a percutaneous lung biopsy at a tertiary-care, National Cancer Institute-designated comprehensive cancer center. Patients were consented for the PTNB procedure but not for inclusion in this retrospective study.

Sample selection

PTNB procedure reports (N=14,665) PTNB procedure reports between October 1, 2008, and September 30, 2018 (Figure I). Patients who experienced hemoptysis were identified through a Radiology Information System (RIS) query of the procedure reports for these cases. All physician-dictated, lung biopsy procedure reports that contained the word “hemoptysis” were extracted. There were 262 procedure reports that used the word hemoptysis. The study team (first and second authors) reviewed each report and cases were removed from the cohort if the report indicated that hemoptysis did not occur or hemoptysis was the indication for the biopsy, resulting in 31 exclusions. Interventionalists reported hemoptysis that began after the needle was inserted or removed, usually while the patient was in the procedure room, however hemoptysis that occurred in the recovery area prior to the completion of follow-up chest radiographs would also be reported as part of the procedure report. Over the 10-year period, there were 231 hemoptysis events occurring in 229 patients (two patients had multiple biopsies and experienced hemoptysis both times, the second event was excluded from analysis of characteristics).

PTNB procedures were performed with CT, fluoroscopy, or CT-fluoroscopic guidance and breath holding was not consistently required of the patient. The utilized sedation was moderate sedation (narcotic and benzodiazepine), monitored anesthesia care with propofol, or general anesthesia. The interventionalist used either a core biopsy or FNA technique. At times, FNA was performed through a coaxial needle. A critical checklist was developed to guide clinician management of hemoptysis.¹⁹ Placing the patient in the lateral decubitus position, with effected side down, while enlisting support of airway experts (anesthesia providers, thoracic surgeons, and interventional pulmonologists) were identified as essential critical actions for patient rescue from severe hemoptysis.¹⁹

Tools and Variable Definitions

The Common Terminology Criteria for Adverse Events version 5.0 (CT-CAE) scale was utilized to describe severity of outcomes in the sample. CT-CAE is a classification system for adverse event reporting that uses a grading scale from 1 to 5.²⁵ Grade 1 is utilized for mild symptoms with no intervention indicated. Grade 2 is defined as moderate symptoms without the need for intervention. Grade 3 indicates those where a transfusion, procedure, surgery, or hospitalization are indicated. Grade 4 represent life-threatening complications requiring urgent interventions and grade 5 is death.²⁵ In this study, hemoptysis CT-CAE classifications of 1 or 2 were defined as “Mild-Moderate” hemoptysis; CT-CAE score of 3 or greater as “Severe”.²⁰

Electronic Medical Record Review

Subsequently, two investigators (first and second authors) collected all of the demographic, clinical, and procedural variables used in this study via a manual review of the electronic medical record (EMR) of all identified hemoptysis cases. The variables collected were based on previously reported variables of interest as well as newly hypothesized contributors (see Table A). Initially, the second author reviewed the EMR and recorded the variables on paper. The first author then transcribed those data into Research Electronic Data Capture (REDCap), “a secure, web-based software platform designed to support data capture for research studies, providing 1) an intuitive interface for validated data capture; 2) audit trails for tracking data manipulation and export procedures; 3) automated export procedures for seamless data downloads to common statistical packages; and 4) procedures for data integration and interoperability with external sources”.^{26, 27} The first author reviewed the final dataset to confirm accuracy.

Variables utilized in the study with their medical record source and definition are presented in Table I, including references to prior research that has described associations between the variable and bleeding.

Table 1.

Variable medical record sources and definitions

Variable	Source	Definition
Anti-thrombic medication use	EMR	Home medication list obtained from the nearest date before the procedure was reviewed to identify anticoagulants and anti-aggregates (9, 14, 15, 18). Pre-procedure clinical documentation was reviewed for patient instructions and patient report of hold times for medications. Non-aspirin antithrombic hold times were evaluated against protocol established by the Society of Interventional Radiology to determine compliance with current protocol (30).
Body mass indexa	EMR	Included value was either included in EMR or determined by inputting weight and height at the time of the procedure into an open-source BMI calculator (32).
Cigarette pack-year historya	EMR	Continuous calculated variable: number of years smoked multiplied by the reported number of packs smoked per day obtained from patient’s report of their smoking history (33).
Cigarette smoking history	EMR	Positive if there was evidence in medical record of a pack year history greater than 5 years (2, 15).
Demographic variables (age, sex)	EMR	Age (6, 9, 14, 15) was reported at time of the procedure, sex (6, 9, 23, 24) was biologic sex.
Malignancya	EMR	Any history of a malignancy was recorded, regardless if remote, since this information is necessary for pathological analysis of tissue sample.
Medical history	EMR	Positive for evaluated conditions (hypertension, atrial fibrillation, coronary artery disease, or COPD (9, 15, 23)) if recorded in the pre-procedure documentation.
Procedure characteristics	PPR/EMR	Indication for biopsy, lobe biopsied with laterality, smallest needle gauge (largest diameter) used intra-procedurally (9, 15, 23), biopsy technique (core biopsy and/or FNA) (3, 9, 13), and total number of specimens obtained (5, 8, 13, 18) were recorded. If lobe was not reported, most recent position emission tomography (PET) scan was utilized to identify target lesion.
Laboratory values	EMR	Hemoglobin, hematocrit, platelets, and INR were drawn and evaluated within 30 days of the procedure. Laboratory reports from outside institutions within time frame were accepted (14, 15, 18, 20).
Pneumothorax development and treatment	PPR/EMR	Immediate post-PTNB development of pneumothorax (within 2 hours) and required treatment (e.g. decompression with chest tube or hospital admission) (6, 12).
Sedationa	EMR	Type of sedation used during procedure including moderate sedation, monitored anesthesia care (e.g. propofol), or general anesthesia.
Specimen result	PATH	Pathologic and cytology reports were reviewed to determine pathological result of the PTNB (2, 5, 6).
Vital signs	EMR	Heart rate, systolic blood pressure, and oxygen saturation collected during the pre-procedure period were recorded.

FNA, fine needle aspirate; PATH, pathology reports; PPR, Physician procedure report; EMR, electronic medical record

^aNo know previously described associations with severity of hemoptysis.

Statistical analysis

Analysis was conducted using IBM SPSS Statistics for Windows, version 25.0.²⁸ Nominal and ordinal variables were summarized using frequency distributions; continuous variables using median (IQR). Associations of each variable with the hemoptysis severity (mild-moderate vs. severe) were conducted using binary logistic regressions. Given the very small sample of patients with severe hemoptysis and the large number of tests conducted the focus of this work is on the size of the effects (odds ratios) and 95% confidence intervals generated around those effects. Interpretations of statistical significance used a p < .05 criteria with no adjusted made for the multiple tests.

Results

The overall incidence of hemoptysis was 1.6% (15.8 events per 1000). The incidence of severe hemoptysis events (CT-CAE 3–5) was 0.2% (2.0 per 1000). The incidence of life-threatening and fatal hemoptysis events (CT-CAE 4–5) was 0.1% (7.5 per 10,000). Death occurred in 0.03% of procedures (2.7 in 10,000).

Of the 229 patients included in this study, 12.7% (n=29) were classified into the severe hemoptysis category; 87.3% (n=200) had ‘mild-moderate’ hemoptysis. Summaries of the demographic and medical history of the patients are shown in Table 2. Median age was 71.0 (IQR 60–78), 71.6% (n=164) were female, had a median BMI of 27.2 (IQR = 23–32), and approximately 71% (n=162) had a history of a solid tumor malignancy. Within this set of characteristics, the strongest and statistically significant associations with the severity of hemoptysis were observed for a history of COPD (OR=3.68, 95% C.I. = 1.53–8.82, P=.003) and within the subsample of patients who previously smoked, cigarette pack years (n=134, OR=1.02, 95% C.I.=1.01–1.04, P=.020).

Table 2.

Patient characteristics and medical history with hemoptysis severity

	Entire Sample	Mild-Moderate	Severe	OR (95% C.I.)	p-value a
	n (%) / N	n (%) / N	n (%) / N
Sex
Male	65 (28.4) / 229	59 (29.5) / 200	6 (20.7) / 29	Referent
Female	164 (71.6) / 229	141 (70.5) / 200	23 (79.3) / 29	1.60 (0.62 – 4.14)	.33
	Median (IQR) / N	Median (IQR) / N	Median (IQR) / N
BMI	27.2 (23, 32) / 229	27.0 (23, 32) / 200	27.4 (23, 34) / 29	1.04 (0.98 – 1.10)	.18
Age Time of Procedure	71.0 (60, 78) / 229	70.0 (59, 77) / 200	75.0 (69, 83) / 29	1.03 (0.99 – 1.06)	.14
	n (%) / N	n (%) / N	n (%) / N
Cigarette History
Never smoked	90 (39.3) / 229	80 (40.0) / 200	10 (34.5) / 29	Referent
Current / Former smoker	139 (60.7) / 229	120 (60.0) / 200	19 (65.5) / 29	1.27 (0.56 – 2.87)	.57
If smoked,	Median (IQR) / N	Median (IQR) / N	Median (IQR) / N
Years quit prior procedure	20.0 (7, 32) / 131	20.0 (6, 32) / 112	27.0 (8, 31) / 19	1.00 (0.97 – 1.04)	.82
Cigarette pack years b	30.0 (15, 45) / 134	30.0 (15, 44) / 116	38.8 (17, 93) / 18	1.02 (1.01 – 1.04)	.02
	N = 229	N = 200	N = 29
Medical History	n (%)	n (%)	n (%)
Hypertension	114 (49.8)	96 (48.0)	18 (62.1)	1.77 (0.79 – 3.95)	.16
Atrial fibrillation	23 (10.0)	21 (10.5)	2 (6.9)	0.63 (0.14 – 2.85)	.55
Coronary artery disease	33 (14.4)	28 (14.0)	5 (17.2)	1.28 (0.45 – 3.63)	.64
COPD	35 (15.3)	25 (12.5)	10 (34.5)	3.68 (1.53 – 8.82)	.003
Thoracic malignancy	69 (30.1)	58 (29.0)	11 (37.9)	1.50 (0.66 – 3.37)	.33
Thoracic surgery c	57 (24.9)	50 (25.0)	7 (24.1)	0.96 (0.38 – 2.37)	.92
Hematologic malignancy	32 (14.0)	27 (13.5)	5 (17.2)	1.34 (0.46 – 3.80)	.59
Solid tumor malignancy	162 (70.7)	142 (71.0)	20 (69.0)	0.91 (0.39 – 2.12)	.82

^aSpecific contrast, Likelihood Ratio Chi-Square

^bPack year history is calculated as the number of years smoked multiplied by number of packs per day

^cLobectomy, pneumonectomy, or wedge resection in past

Summaries of the lab values, vital signs, and antithrombic medication use are shown in Table 3. Pre-procedure laboratory values and vital signs were not shown to be associated with the severity of hemoptysis (OR 0.94–1.13, p > .140). Consistent with the similarities in the distributions of the values shown in Table 2 within each of the severity groups, no statistically significant associations of those variables with severity of hemoptysis were observed other than that for the number of days prior to the procedure that ASA was held within the subsample of patients who had been using ASA (n=75, p = .007). The subsample of those patients within the severe category was so small (n=12) that specific contrasts between the number of days could not be conducted. Nevertheless, the tendency observed was that the likelihood of a severe hemoptysis event increased as the days that ASA was held decreased prior to PTNB (see Table 3).

Table 3.

Laboratory values, vital signs, and antithrombic medication use with hemoptysis severity

	Entire Sample	Mild-Moderate	Severe	OR (95% C.I.)	p-value a
	N = 229	N = 200	N = 29
Laboratory values b	Median (IQR)	Median (IQR)	Median (IQR)
Hemoglobin g/dL	12.9 (11.1, 13.9)	13.0 (11.2, 13.9)	12.5 (10.9, 14.0)	0.94 (0.77 – 1.14)	.53
Hematocrit g/dL	39.3 (34.7, 41.7)	39.3 (34.8, 41.7)	38.6 (32.2, 42.4)	0.98 (0.92 – 1.06)	.63
Platelet (X103 microL)	225.0 (178, 283)	225.0 (179, 283)	212.0 (154, 289)	1.03 (0.99 – 1.06)	.14
INR	1.1 (0.9, 1.1)	1.0 (0.9, 1.2)	1.1 (0.9, 1.2)	1.00 (0.99 – 1.01)	.81
	n (%)	n (%)	n (%)
Platelet Category (p=.951)
< 50,000	6 (2.6)	5 (2.5)	1 (3.4)	-----	-----
50 – 100,000	7 (3.1)	6 (3.0)	1 (3.4)	-----	-----
> 100,000	216 (94.3)	189 (94.5)	27 (93.1)	-----	-----
		N=199
Baseline vital signs	Median (IQR)	Median (IQR)	Median (IQR)
Heart rate (bpm)	75.5 (66, 85)	75.0 (66, 86)	78.0 (65, 85)	1.00 (0.97 – 1.03)	.97
Systolic BP (mmHg)	148.0 (125, 169)	148.0 (125, 169)	151.0 (130, 171)	1.00 (0.98 – 1.02)	.89
Oxygen saturation (%)	97.0 (96, 97)	97.0 (96, 98)	97.0 (96, 99)	1.13 (0.91 – 1.40)	.26
	n (%)	n (%)	n (%)
Systolic BP > 190 mmHg	14 (6.1)	13 (6.5)	1 (3.4)	-----	-----
Aspirin use pre-procedure	81 (35.4)	67 (33.5)	14 (48.3)	1.85 (0.84 – 4.07)	.12
If ASA used, held (p = .007)	N=75	N=63	N=12
>= 5 days	52 (69.3)	48 (76.2)	4 (33.3)	-----	-----
2–4 days	10 (13.3)	5 (7.9)	5 (41.7)	-----	-----
<= 1 day	13 (17.3)	10 (15.9)	3 (25.0)	-----	-----
	N = 229	N = 200	N = 29
Non-ASA antithrombotic use	50 (21.8)	43 (21.5)	7 (24.1)	1.16 (0.46 – 2.91)	.75
If non-ASA used, held (p= .47)	N=49	N=43	N=6
Per department protocol	47 (95.9)	41 (95.3)	6 (100.0)	-----	-----
Not per department protocol	2 (4.2)	2 (4.7)	0 (0.0)	-----	-----

^aSpecific contrast, Likelihood Ratio Chi-Square

^bLaboratory values obtained within month proceeding PTNB

As shown in Table 4, slightly more than half were undergoing a biopsy to rule out metastatic disease or a new primary (52.4%, n=120); 67.1% (n=153 of 228) resulted in a malignant diagnosis, and 17.5% (n=40 of 228) demonstrated an active infection or inflammation. Most of the strongest associations with severity were observed in the set of procedural characteristics. Increased likelihood of severe hemoptysis was observed for core biopsy technique (OR=8.13, 95% CI=1.07, 61.40, P=0.042), pneumothorax post biopsy (OR=3.86, 95% CI=1.60, 9.28, P=.003), and chest tube placement (treated pneumothorax, OR=5.97, 95% CI=2.19, 16.26, P<.001). Needle size was statistically significantly associated with severity. Relative to the use of the 20g needle, use of the 18g needle increased the likelihood of a severe as opposed to a mild/moderate event by 2.60 (95% CI=1.09, 6.17, P=.031).

Table IV.

Procedural characteristics with hemoptysis severity

	Entire Sample	Mild-Moderate	Severe	OR (95% C.I.)	p-value a
Characteristics	N = 229	N = 200	N = 29
	n (%)	n (%)	n (%)
Biopsy indication (p = .14)
New diagnosis of cancer	75 (32.8)	65 (32.5)	10 (34.5)	Referent
Rule out new disease b	120 (52.4)	108 (54.0)	12 (41.4)	0.72 (0.29, 1.77)	.48
Molecular diagnostics	10 (4.4)	6 (3.0)	4 (13.8)	4.33 (1.03, 1.77)	.04
Clinical trial	25 (10.5)	21 (10.5)	3 (10.3)	0.93 (0.23, 3.70)	.92
Laterality (p = .54)
Right	122 (53.3)	105 (52.5)	17 (58.6)	Referent
Left	107 (46.7)	95 (47.5)	12 (41.4)	0.78 (0.35, 1.72)	.54
Lobe biopsied (p = .95)
Upper	107 (46.7)	93 (46.5)	14 (48.3)	Referent
Middle	14 (6.1)	12 (6.0)	2 (6.9)	1.11 (0.22, 5.48)	.90
Lower	96 (41.9)	85 (42.5)	11 (37.9)	0.86 (0.37, 2.00)	.73
Other c	12 (5.2)	10 (5.0)	2 (6.9)	1.33 (0.26, 6.71)	.73
Core biopsy technique used d	183 (79.9)	155 (77.5)	28 (96.6)	8.13 (1.07, 61.40)	.04
Largest needle size (p = .007)	N = 227	N = 198	N = 29
18g	112 (49.3)	91 (46.0)	21 (72.4)	2.60 (1.09, 6.17)	.03
20g	98 (43.2)	90 (45.5)	8 (27.6)	Referent
22g	17 (7.5)	17 (8.6)	0 (0.0)	-----	-----
	Median (IQR)	Median (IQR)	Median (IQR)
	N = 221	N = 192	N = 29
Number of Specimens e	2.0 (1, 3)	2.0 (1, 3)	2.0 (1, 3)	0.99 (0.82, 1.20)	.93
	n (%)	n (%)	n (%)
Position (p = .07)	N = 228	N = 200	N = 28
Supine	104 (45.6)	92 (42.9)	12 (42.9)	Referent
Prone	118 (51.8)	105(52.5)	13 (46.4)	0.95 (0.41, 2.19)	.95
Other f	6 (2.6)	3 (1.5)	3 (10.7)	-----	-----
Sedation (p = .14)	N = 229	N = 200	N = 29
Moderate sedation	208 (90.8)	184 (92.0)	24 (82.8)	Referent
Monitored anesthesia care	18 (7.9)	13 (6.5)	5(17.2)	2.95 (0.96, 9.00)	.06
General anesthesia	3 (1.3)	3 (1.5)	0 (0.0)	-----	-----
Pneumothorax g	34 (14.8)	24 (12.0)	10 (34.5)	3.86 (1.60, 9.28)	.003
Chest tube placement g	20 (8.7)	12 (6.0)	8 (27.6)	5.97 (2.19, 16.26)	< .001
	N = 225	N = 196	N = 29
Aborted procedure h	85 (37.8)	76 (38.8)	9 (31.0)	0.71 (0.30, 1.65)	.42
Specimen result (p = .38)	N = 228	N = 199	N = 29
Malignancy	153 (67.1)	133 (66.8)	20 (69.0)	1.11 (0.29, 4.11)	.88
Inadequate specimen	21 (9.2)	18 (9.0)	3 (10.3)	-----	-----
Normal tissue j	8 (3.5)	8 (4.0)	0 (0.0)	0.74 (0.23, 2.30	.60
Infection / inflammation	40 (17.5)	36 (18.1)	4 (13.8)	3.33 (0.57, 19.36)	.18
Non-malignant diagnosis	6 (2.6)	4 (2.0)	2 (6.9)	-----	-----

^aSpecific contrast, Likelihood Ratio Chi-Square

^bNew disease including recurrence, new primary, metastatic disease in those with history of malignancy

^cPerihilar, paramediastinal, lingular, and multiple lobes

^dFine needle aspirate technique used for other cases

^eTotal specimens obtained from target lesion (cores and FNA specimens)

^fPatient positioned in lateral decubitus or position not indicated

^gOccurred after biopsy completion

^hAborted procedure due to hemoptysis

ⁱSpecimen adequate for analysis

Discussion

This study is the largest-known published series of PTNB procedures investigated for incidence of hemoptysis. The reported overall incidence of hemoptysis (1.6%) at this institution that performs a high volume of PTNB procedures, mostly with a core biopsy technique, is at the lower end of reported incidence of hemoptysis (0.3–15.5%) in similar retrospective studies.^{2, 5, 6, 8, 12–17} Within the sample of only the people who experienced hemoptysis, females were 1.57 times more likely males to have hemoptysis be severe rather than mild/moderate hemoptysis. The incidence of severe hemoptysis events (CT-CAE 3–5) was 0.2% (2.0 per 1000) which is consistent with the reported incidence of severe hemoptysis (0%-1.6%).^{14–16, 18, 20–22} Few research studies included death as a complication, however the incidence of death following PTNB (0.03%) was found to be consistent with other studies (0.01%-0.08%).^{7, 18, 20} Bleeding-related mortality following PTNB is a rare event, 2.7 events in 10,000 cases, so it is possible that other retrospective reviews were not adequately powered to evaluate this outcome.

Statistically significant correlations with the severity of hemoptysis were identified among both those with a history of greater years of tobacco use (OR 1.02 per pack year) and those a history of COPD (OR 3.68). The positive association between COPD and bleeding was stronger than effects reported in other studies (OR 1.44–1.64).^{9, 15, 23} A history of smoking independent of amount of smoking was not found to be statistically significant correlated with severe hemoptysis (OR 1.27).

In this study, we demonstrated a relationship between increased hemoptysis severity and development of a pneumothorax after PTNB (OR 3.86). Although this finding will not inform preventative measures to minimize hemoptysis severity, prior studies have found that any amount of hemoptysis, regardless of severity, offered a protective effect from pneumothorax.^{6, 29} In cases of severe hemoptysis, this does not appear to be the case. Possible explanations for this finding may be that the path of blood did not impede the path of air to the pleura, the size of the needle defect was large enough to create both conditions, or smoking-related lung parenchymal changes may increase susceptibility to both severe hemoptysis and pneumothorax.

Core biopsy specimens (OR 8.13) correlated with greater hemoptysis severity with smaller needle gauges (larger diameter) showing a reduction in risk (OR 0.39). Other studies have reported that core biopsies (OR 2.00 with FNA referent, OR 0.38–0.92 with core referent)^{3, 9, 13} and smaller gauge (larger diameter) needles (OR 1.05–2.30)^{9, 15, 23} were associated with hemoptysis. It stands to reason that a larger diameter needle (smaller gauge) would increase the bleed vessel defect and therefore increase the time required to achieve hemostasis and therefore, severity of bleeding would be greater.

This study did not find statistically significant associations between hemoptysis severity and antithrombotic medications such as aspirin which is a consistent finding in most other studies.^{9, 16} However, longer aspirin hold times were correlated with decreased bleeding severity. Song et al. reported an effect of ASA on severe hemoptysis (OR=2.0, calculated).¹⁶ Song et al. also found a strong association between hemoptysis and patients who were on clopidogrel and aspirin (OR 10.09, 95% CI 2.33–43.62, adjusted) over aspirin (OR=1.38, 95% CI 0.78–2.45, adjusted) or clopidogrel alone (OR 2.26, 95% CI 0.56–9.08, adjusted).¹⁶ This finding supports the recent change in practice recommendations of the Society of Interventional Radiology supporting holding of ASA prior to PTNB when clinically possible.³⁰

Limitations

We acknowledge several limitations to this study. This was a retrospective chart review with the possibility of missing, inaccurate, incomplete, or difficult to extract data. This was particularly evident for earlier events in the study period when standardized documentation templates were not used for documentation. Multiple reviews of the dataset were undertaken to ensure all available information in the medical record was extracted in a reliable manner. Furthermore, the retrospective design of this study limited the ability to accurately report on some variables, such as pre-procedure antithrombic medication hold times and specific conditions, such as pulmonary hypertension. Another limitation included a historical threat to internal validity since clinical practice has evolved over the 10-year study period. For example, in 2012 and 2019 the Society of Interventional Radiology published updated thrombotic and bleeding risk management guidelines. The most current practice recommendations were not applied to this cohort of patients.^30–32

Finally, we acknowledge that that our sample size was insufficient for conducting multivariable associations that would have enabled us to explore important potentially interactive effects of some variables (such as sex). We also acknowledge that variables may have been omitted that are of interest to the reader and these omissions could inadvertently create bias. Our focus was on publishing the effect sizes we found with accompanying confidence intervals. Ideally, we would like to compare those effect sizes with findings from other studies, however those comparisons are limited by a lack of standardized definitions of hemoptysis and measurement of severity.³³ Future retrospective research designs should incorporate non-bleeding controls and future prospective studies should be specifically designed to answer targeted research questions designed evaluate effect of specific variables on hemoptysis.

Conclusion

In conclusion, this study represents the largest known sample of severe hemoptysis events (n=29). Where mild-moderate events are uncommon (1.6 events per 100) and self-limiting with most PTNB procedures using a core biopsy technique, rare severe events are a medical emergency and require rapid intervention to rescue patient from clinical deterioration. We described several novel positive associations with severe hemoptysis including a patient’s clinical pack-year smoking history, development of a pneumothorax after PTNB, use of a core needle biopsy technique, and smaller needle gauge (larger diameter). Understanding characteristics associated with increased severity can help clinicians identify those at high risk and perhaps modify procedure choices.

Research Highlights.

The overall occurrence of hemoptysis following percutaneous transthoracic needle biopsy (PTNB) was 1.6%. Severe hemoptysis events occurred in 0.2% of all biopsies performed over a 10-year period.

The strongest and statistically significant variables associated with an increased likelihood of a severe event, if an event occurred, were cigarette pack years (OR=1.02, 95% C.I.=1.01–1.04, P=.020); history of chronic obstructive pulmonary disease (COPD) (OR=3.68, 95% C.I. = 1.53–8.82, P=.003); core biopsy technique (OR=8.13, 95% CI=1.07, 61.40, P=.042), and larger diameter needle (20g vs. 18g: OR= 2.60 (1.09, 6.17), P=.031).