Risk factors for thoracic-osteoporotic thoracic vertebral compression fractures during the normalized prevention and control period of COVID-19

Abstract

BACKGROUND

Some scholars believe that there is a certain correlation between coronavirus disease 2019 (COVID-19) infection and the occurrence of osteoporotic vertebral compression fractures (OVCF), but no relevant domestic research reports have been published yet. This study retrospectively analyzed the clinical data of patients admitted for OVCF and treated with percutaneous vertebroplasty during the normalized prevention and control period of COVID-19 in Shanghai (after June 1, 2022), aiming to explore the individual differences of OVCF patients during this special period and analyze the risk factors for OVCF.

AIM

To explore risk factors for senile OVCF during COVID-19 normalization.

METHODS

Retrospective analysis of 230 OVCF patients (June 2022 to June 2023, percutaneous vertebroplasty, observation group) and 236 controls (June 2020 to June 2021, OVCF surgery). Observation group was split into nucleic acid-positive (n = 85) and negative (n = 145) subgroups. Fracture location/cause, visual analogue scale, activity scores were compared; risk factors identified.

RESULTS

Observation group had 42.2% thoracic fractures and 38.2% cough-induced fractures (both higher than controls, P < 0.05). No significant differences in post-op visual analogue scale (1 day, 1 month) or activity scores (P > 0.05). Nucleic acid results, fracture cause/location correlated (P < 0.05; cause-location r = 0.827). Logistic regression showed fracture cause, nucleic acid result, body mass index, bone mineral density as high-risk factors.

CONCLUSION

During COVID-19 normalization, cough is the top OVCF cause; thoracic fractures rise. The four factors above are independent risks. For elderly COVID-positive patients with recurrent cough, lung disease treatment, anti-osteoporosis therapy, and thoracic braces prevent thoracic OVCF.

Key Words: COVID-19; Normalized prevention and control; Osteoporotic vertebral compression fracture; Risk factor; Cough

Core Tip: During the normalized prevention and control period of coronavirus disease 2019 (COVID-19), compared with cases in the same period, cough is the main cause of osteoporotic vertebral compression fracture (OVCF), and most fractures are thoracic fractures. After percutaneous vertebroplasty treatment, the improvement of pain and function in patients is not significantly different from that in the same period. Regression analysis shows that cause of fracture, nucleic acid test result, body mass index, and bone mineral density are independent risk factors for thoracic OVCF. For elderly COVID-19-positive patients with recurrent cough, timely and active treatment of lung diseases, aggressive anti-osteoporosis therapy, and wearing a thoracic brace for protection are effective measures to prevent thoracic OVCF.

INTRODUCTION

In December 2019, coronavirus disease 2019 (COVID-19) broke out globally, including in China. The World Health Organization officially classified the COVID-19 epidemic as a public health emergency of international concern in January 2020, which poses a severe threat to human life and health. Through the implementation of sustained and effective epidemic prevention policies, significant achievements have been made in domestic epidemic prevention and control. With the adjustment of COVID-19 management from “class B infectious diseases managed with class A measures” to “class B infectious diseases managed with class B measures”, China has entered a new phase of epidemic prevention and control, shifting the focus from “preventing infection” to “protecting health and preventing severe cases”. Meanwhile, the number of people infected with the virus has increased on a large scale across the country[1]. The COVID-19 epidemic in Shanghai began in late February 2022. According to public information, 1 local asymptomatic infected case was reported in Shanghai on February 24, marking the start of this round of epidemic. After more than two months of continuous efforts, Shanghai fully resumed normal production and living order on June 1, 2022, indicating that this round of epidemic was effectively controlled. Since then, Shanghai has entered the normalized phase of epidemic prevention and control[2].

An epidemiological study in China confirmed that among COVID-19 patients, those aged 30-80 years account for approximately 86.6%, of which 66.5% are over 50 years old. Most of these elderly patients are accompanied by postmenopausal osteoporosis or senile osteoporosis, and some even have concurrent osteoporotic vertebral compression fractures (OVCF)[3]. Clinical data show that most patients still have recurrent cough symptoms during the recovery period of COVID-19 infection, which can last for more than 4 weeks. In severe cases, this may lead to thoracolumbar fractures[4]. Some scholars believe that there is a certain correlation between COVID-19 infection and the occurrence of OVCF, but no relevant domestic research reports have been published yet[5,6]. This study retrospectively analyzed the clinical data of patients admitted for OVCF and treated with percutaneous vertebroplasty (PVP) during the normalized prevention and control period of COVID-19 in Shanghai (after June 1, 2022), aiming to explore the individual differences of OVCF patients during this special period and analyze the risk factors for OVCF.

MATERIALS AND METHODS

General information

Patients who were admitted due to OVCF and underwent PVP from June 2022 to June 2023 were collected through the medical record query system as the observation group. Meanwhile, patients with the same diagnosis who were admitted and underwent surgery from June 2020 to June 2021 were collected as the control group. Inclusion criteria: (1) Confirmed diagnosis of OVCF with a single-segment fracture; (2) Meeting the diagnostic criteria for osteoporosis (T-score < -2.5) by dual-energy X-ray absorptiometry bone mineral density (BMD) testing, or a Hounsfield unit value < 90 by computed tomography; and (3) Complete medical records. Exclusion criteria: (1) Patients with multi-segment fractures; (2) Pathological fractures complicated with metastatic tumors; and (3) Severe trauma (high-energy injury), fractures compressing spinal nerves, old fractures, etc.

Observation indicators

General conditions of patients before surgery, including gender, age, body mass index (BMI), BMD, cause of onset, fracture location, and COVID-19 nucleic acid test result. Additionally, the visual analogue scale (VAS) score for pain and activity ability score were evaluated before surgery, at 1 day after surgery, and during the re-examination at 1 month after surgery. The Likert 4-point scale was used to assess activity ability, with higher scores indicating better activity ability: 1 point = able to get out of bed and move freely; 2 points = able to get out of bed and move with assistance; 3 points = unable to get out of bed and requiring a wheelchair; 4 points = bedridden and completely unable to get out of bed.

Data analysis

Descriptive analysis: Comparison of general information, fracture location, and cause of onset between the two groups.

Analysis of high-risk factors during the normalized COVID-19 prevention and control period: The observation group was divided into a positive subgroup (n = 85) and a negative subgroup (n = 145) based on COVID-19 nucleic acid test results. The causes of fractures between the two subgroups were compared, and binary logistic regression analysis was performed.

Comparison of surgical efficacy: Comparison between the two periods and between the COVID-19 nucleic acid-positive and negative subgroups.

Statistical analysis

SPSS 21.0 statistical software (IBM, United States) was used for data processing and analysis. The Kolmogorov-Smirnov test was used to verify the normal distribution of measurement data. Measurement data conforming to normal distribution and homogeneous variance were expressed as mean ± SD, and the independent samples t-test was used for comparison between the two groups; otherwise, the non-parametric test was used. Count data were expressed as the n (%), and the χ² test for fourfold tables was used for comparison. Logistic regression analysis was used for the analysis of related factors. The two-tailed test level α was set at 0.05.

Assignment of independent variables for regression analysis: Fracture segment: Thoracic segment = 1, thoracolumbar segment = 2, lumbar segment = 3; cause of fracture: Cough = 1, fall = 2, fall from height = 3, sprain = 4, lifting heavy objects = 5; nucleic acid test result: Positive = 1, negative = 2; BMI: Underweight (BMI < 18.5) = 1, normal weight (18.5 ≤ BMI < 24.0) = 2, overweight (24.0 ≤ BMI < 28.0) = 3, obesity (BMI ≥ 28.0) = 4; BMD: Osteoporosis ( ≤ -2.5) = 1, osteopenia (-2.5 to -1.0) = 2, normal bone mass ( > -1) = 3.

RESULTS

Comparison of general information between the two groups

The observation group included 230 patients, with 56 males and 174 females; the average age was 75.96 ± 6.00 years; the average BMI was 21.80 ± 2.31 kg/m²; the average lumbar BMD was -2.31 ± 0.83 g/cm². The control group included 236 patients, with 49 males and 187 females; the average age was 76.76 ± 5.44 years; the average BMI was 22.05 ± 2.41 kg/m², the average lumbar BMD was -2.40 ± 0.67 g/cm². There were no statistically significant differences in gender, age, BMI, or BMD between the two groups (P > 0.05), as shown in Table 1.

Table 1 Comparison of general data, mean ± SD.

Parameter	Observation group (n = 230)	Control group (n = 236)	χ2	P value
Gender	56 males, 174 females	49 males, 187 females	-1.539	0.125
Age (years)	75.96 ± 6.00	76.76 ± 5.44	-1.518	0.130
BMI (kg/m2)	21.80 ± 2.31	22.05 ± 2.41	-1.107	0.269
BMD (g/cm2)	-2.31 ± 0.83	-2.40 ± 0.67	1.280	0.201

BMI: Body mass index; BMD: Bone mineral density.

Comparison of fracture location and cause between the two groups

There were statistically significant differences in the cause and location of fractures between the observation group and the control group (P < 0.05). In the observation group, the number of fractures in the thoracolumbar segment, lumbar segment, and thoracic segment was 124 cases (53.9%), 9 cases (3.9%), and 97 cases (42.2%), respectively; in the control group, the number of fractures in the thoracolumbar segment, lumbar segment, and thoracic segment was 190 cases (80.5%), 26 cases (11.0%), and 20 cases (8.5%), respectively. The number of patients with thoracic fractures was significantly increased in the observation group (P < 0.001).

In the observation group, the number of fractures caused by falls at the same level, falls from height, sprains, lifting heavy objects, and coughs was 68 cases (29.6%), 26 cases (11.3%), 26 cases (11.3%), 22 cases (9.6%), and 88 cases (38.2%), respectively; in the control group, the number of fractures caused by the above factors was 111 cases (47.0%), 55 cases (23.3%), 31 cases (13.1%), 24 cases (10.2%), and 15 cases (6.4%), respectively. Cough was the main cause of fractures in the observation group, with a statistically significant difference (P < 0.001). Details are shown in Table 2 and Figure 1.

Figure 1 Comparison of fracture location and cause between the observation group and the control group. A: Fracture location; B: Cause of fracture. The horizontal axis represents the type of fracture location/cause, and the vertical axis represents the number of cases. ^aP < 0.05. T: Thoracic vertebrae; L: Lumbar vertebrae.

Table 2 Comparison of fracture location and cause between the two groups, n (%).

Item	Observation group (n = 230)	Control group (n = 236)	χ2	P value
Fracture location
Thoracolumbar segment (T11-L2)	124 (53.9)	190 (80.5)	37.485	< 0.01a
Lumbar segment (L3-L5)	9 (3.9)	26 (11.0)	8.462	0.0036a
Thoracic segment (T1-T10)	97 (42.2)	20 (8.5)	70.348	< 0.01a
Cause of fracture
Fall at the same level	68 (29.6)	111 (47.0)	15.024	0.0001a
Fall from height	26 (11.3)	55 (23.3)	11.681	0.0006a
Sprain	26 (11.3)	31 (13.1)	0.363	0.546
Lifting heavy objects	22 (9.6)	24 (10.2)	0.0478	0.826
Non-cough causes of fracture	142	221
Cough	88 (38.2)	15 (6.4)	68.864	< 0.01a

^aP < 0.05 vs the control group.

T: Thoracic vertebrae; L: Lumbar vertebrae.

Comparison of pain and activity ability before and after surgery between the two groups

Before surgery, the VAS score of the observation group was significantly higher than that of the control group, with a statistically significant difference (P < 0.05); there were no statistically significant differences in VAS scores at 7 days and 1 month after surgery, or in activity ability scores before and after surgery between the two groups (P > 0.05). Details are shown in Table 3 and Figure 2.

Figure 2 Comparison of visual analogue scale scores and activity ability scores between the observation group and the control group before surgery, 1 day after surgery, and 1 month after surgery. B: Activity ability score. The horizontal axis represents the time point (pre-operation, 1 day post-operation, 1 month post-operation), and the vertical axis represents the score value). 1d: 1 day; 1m: 1 month.

Table 3 Comparison of pain and activity ability at different time points after surgery between the two groups, mean ± SD.

Item	Observation group (n = 230)	Control group (n = 236)	t	P value
VAS score (points)
Pre-operation	7.8 ± 1.3	7.4 ± 1.5	2.097	0.037
1 day post-operation	4.4 ± 0.9a	4.2 ± 1.1a	1.457	0.146
1 month post-operation	2.3 ± 0.5a	2.2 ± 0.4a	1.681	0.094
Activity ability score (points)
Pre-operation	3.3 ± 0.8	3.4 ± 1.0	0.807	0.420
1 day post-operation	1.8 ± 0.5a	1.7 ± 0.6a	1.328	0.185
1 month post-operation	1.3 ± 0.4a	1.2 ± 0.3a	1.942	0.053

^aP < 0.05 vs pre-operation.

VAS: Visual analogue scale.

Comparison of the effects of fracture cause, fracture site, and positive nucleic acid test result

According to the COVID-19 nucleic acid test results, patients in the observation group were divided into the positive subgroup and the negative subgroup. There were significant differences in thoracolumbar fractures and thoracic fractures between the two subgroups with different nucleic acid test results (P < 0.01). All causes of fractures showed statistically significant differences between the two subgroups, as shown in Table 4 and Figure 3.

Figure 3 Comparison of different nucleic acid test. A: Comparison of causes of fractures between subgroups with different nucleic acid test results; B: Comparison of fracture locations between subgroups with different nucleic acid test results. The horizontal axis represents the type of fracture cause/Location, and the vertical axis represents the number of cases. ^aP < 0.05. T: Thoracic vertebrae; L: Lumbar vertebrae.

Table 4 Comparison of influencing factors of nucleic acid test results, n (%).

Item	Negative subgroup (n = 145)	Positive subgroup (n = 85)	χ2	P value
Fracture location
Thoracolumbar segment (T11-L2)	108 (74.5)	11 (12.9)	81.279	< 0.001a
Lumbar segment (L3-L5)	17 (11.7)	6 (7.1)	1.295	0.255
Thoracic segment (T1-T10)	20 (13.8)	68 (80.0)	99.437	< 0.001a
Cause of fracture
Fall at the same level	58 (40.90)	10 (28.24)	20.515	< 0.001a
Fall from height	30 (22.76)	0 (11.76)	20.224	< 0.001a
Sprain	21 (14.48)	1 (9.41)	10.968	0.0009a
Lifting heavy objects	14 (10.66)	1 (9.41)	6.149	0.0131a
Cough	2 (6.21)	73 (41.18)	174.127	< 0.001a

^aP < 0.05 vs the negative subgroup.

T: Thoracic vertebrae; L: Lumbar vertebrae.

Correlation analysis of fracture cause, fracture site (thoracic vertebra), and positive nucleic acid test result (rank correlation test)

Correlation analysis between cause of fracture and fracture location: The correlation coefficient r = 0.827 (P < 0.01), indicating that the correlation coefficient between cause of fracture and fracture location is statistically significant, and there is a positive correlation between cause of fracture and fracture location (Figure 4A).

Figure 4 Correlation analysis chart. A: Correlation analysis chart between cause of fracture and fracture location; B: Correlation analysis chart between nucleic acid test result and fracture location (thoracic vertebra); C: Correlation analysis chart between nucleic acid test result and cause of fracture. The horizontal axis represents fracture location/cause of fracture, and the vertical axis represents nucleic acid test result/cause of fracture/fracture location.

Correlation analysis between fracture location and nucleic acid test result: The correlation coefficient r = 0.332 (P < 0.01), indicating that the correlation coefficient between fracture location and nucleic acid test result is statistically significant, and there is a positive correlation between fracture location and nucleic acid test result (Figure 4B).

Correlation analysis between cause of fracture and nucleic acid test result: The correlation coefficient r = 0.287 (P < 0.01), indicating that the correlation coefficient between cause of fracture and nucleic acid test result is statistically significant, and there is a positive correlation between cause of fracture and nucleic acid test result (Figure 4C).

Logistic regression analysis of risk factors for thoracic OVCF

Comprehensive test of model coefficients: Significance < 0.01, indicating that the model is valid. Among the variables in the equation, cause of fracture [P < 0.01, odds ratio (OR) = 3.66], nucleic acid test result (P = 0.01, OR = 4.82), BMI (P < 0.01, OR = 4.69), and BMD (P < 0.03, OR = 3.52) were statistically significant, as shown in Table 5.

Table 5 Binary logistic regression analysis (enter method).

Variable	B	SE	P value	Exponentiated B
Gender	0.244	0.666	0.714	1.276
Age	0.410	0.383	0.284	1.507
Cause of fracture	1.298	0.197	0.000	3.660
Nucleic acid test result	1.573	0.613	0.010	4.821
BMI	1.546	0.406	0.000	4.691
BMD	1.259	0.424	0.003	3.521
Constant	-13.451	2.461	0.000	0.000

BMI: Body mass index; BMD: Bone mineral density.

Regression equation: logit (p) = (-13.451) + 1.298 × cause of fracture + 1.573 × nucleic acid test result + 1.546 × BMI + 1.259 × BMD. The probability area under the curve value of the regression equation was 0.961, and other area under the curve values are shown in Table 6 and Figure 5.

Figure 5 Receiver operating characteristic curve comparison of each influencing factor. The horizontal axis represents 100 - specificity, and the vertical axis represents sensitivity. Different curves represent different influencing factors: Probability, body mass index, bone mineral density, causes of fractures, and nucleic acid test. BMI: Body mass index; BMD: Bone mineral density.

Table 6 Predictive accuracy of each influencing factor.

Variable	AUC	SE	95%CI
Probability	0.961	0.0130	0.927-0.982
BMI	0.730	0.0304	0.667-0.786
BMD	0.684	0.0314	0.620-0.744
Cause of fracture	0.911	0.0237	0.867-0.945
Nucleic acid test result	0.738	0.0324	0.676-0.794

AUC: Area under the curve; CI: Confidence interval; BMI: Body mass index; BMD: Bone mineral density.

Typical case

A case, female, 87 years old, tested positive for COVID-19, was admitted to the hospital due to back pain after recurrent cough for 1 month. She was diagnosed with T6 osteoporotic fracture and underwent PVP treatment. The pain symptoms were improved after surgery, as shown in Figure 6.

Figure 6 Imaging results. A: Preoperative anteroposterior and lateral X-ray images of T6 fracture; B: Preoperative thoracic magnetic resonance imaging image; C: Anteroposterior and lateral X-ray images of the thoracic spine after T6 vertebroplasty; D: Thoracic computed tomography image after T6 vertebroplasty.

DISCUSSION

OVCF during the normalized prevention and control period of COVID-19 require attention

Currently, there are still few studies on vertebral fractures after severe acute respiratory syndrome coronavirus 2 infection. Some studies have reported that vertebral fractures are relatively common in COVID-19 patients, with 36% of patients showing vertebral deformities, but only 3% of patients have a previous diagnosis of osteoporosis. A study by Yu et al[7] suggested that COVID-19 patients have varying degrees of hypercoagulable state, leukocyte aggregation, and vascular inflammation, all of which may damage the blood flow of bone microvessels, thereby promoting the occurrence and development of osteonecrosis. At the same time, cytokines induced by COVID-19, such as interleukin-17, tumor necrosis factor-alpha, and C-X-C motif chemokine ligand 10, can reduce osteoblast proliferation and induce osteoclast formation, leading to a net decrease in BMD. In a study by Qiao et al[8], hamsters infected with COVID-19 experienced severe bone loss, which could persist until the recovery period. Additionally, the number of osteoclast precursor cells increased, and the type I interleukin-1β receptor was significantly upregulated in CD68-positive cells expressing tartrate-resistant acid phosphatase. This confirms that severe acute respiratory syndrome coronavirus infection activates the osteoclast cascade, thereby causing damage to the trabecular structure in long bones and axial bones. Therefore, there is a close relationship between COVID-19 infection and bone loss. A study by Patil et al[9] pointed out that the risk of vertebral fractures increases after COVID-19, and chest computed tomography shows that thoracolumbar fractures account for approximately two-thirds of all vertebral fractures. Another study suggested that recurrent cough can lead to osteoporotic thoracic fractures, and many OVCF patients were ignored and missed diagnosis during the epidemic, failing to receive scientific treatment[5,6]. On the other hand, this study shows that OVCF patients during the normalized prevention and control period of COVID-19 have higher VAS scores than those in the same period. Therefore, OVCF during the normalized prevention and control period of COVID-19 require attention.

Characteristics of senile osteoporotic thoracic vertebral compression fractures during the normalized prevention and control period of COVID-19

Previous studies have shown that OVCF are most common in the thoracolumbar segment, and the common causes include falls, drops, and lifting heavy objects[10,11]. However, the characteristics of OVCF during the normalized prevention and control period of COVID-19 are different. During this period, OVCF are most common in the middle thoracic spine, and cough is the main cause of fractures, both of which are significantly higher than those in the same period. World Health Organization led a collaboration of global medical scientists to complete the largest-scale study on long COVID to date. By following up self-reported data from 1.2 million confirmed COVID-19 patients in 22 countries and combining global COVID-19 infection data analysis, it was found that persistent respiratory problems are a group of long-term symptoms after COVID-19 infection[12]. Results from a large-scale domestic study showed that cough is the second most common symptom in COVID-19 patients during hospitalization, after fever, accounting for approximately 67.8%[13]. A study by Davis et al[14] found that the incidence of respiratory diseases in COVID-19 patients is twice that of the general population, and shortness of breath and cough are the most common respiratory symptoms, with 40% and 20% of long COVID patients experiencing these symptoms for at least 7 months, respectively. During this period, the COVID-19 virus spread, and patients developed cough after infection. Elderly patients had a long treatment cycle for COVID-19, and recurrent cough led to middle thoracic fractures[15]. Therefore, middle thoracic fracture location and cough as the cause of onset are the characteristics of senile osteoporotic thoracic vertebral compression fractures during the normalized prevention and control period of COVID-19 in Shanghai. On the other hand, elderly people have vertebral deformation and ligament relaxation due to osteoporosis, and further bone loss after COVID-19 infection leads to varying degrees of thoracic kyphosis. Coughing exacerbates this kyphosis, resulting in high stress at the middle thoracic spine, which in turn causes thoracic compression fractures[16,17].

No significant difference in the efficacy of vertebroplasty between the two periods

In recent years, with the acceleration of the aging process, the incidence of osteoporosis has been increasing year by year. OVCF is a common complication of osteoporosis, mainly manifested as sudden severe pain and progressive kyphosis, which seriously affects the quality of life of elderly patients and significantly increases the mortality rate[18]. PVP is currently the most commonly used method for treating OVCF, which can effectively relieve clinical symptoms and improve the quality of life. Statistical data show that there are up to 44.49 million OVCF patients in China, with more than 1.8 million new cases each year[19]. These patients are often complicated with various underlying diseases, organ dysfunction, and decreased body compensatory capacity[20]. During the COVID-19 epidemic, due to reduced mobility, decreased physical activity, and the combined effects of environment and stress, the skeletal muscle mass of the elderly population decreased, resulting in lumbodorsal muscle soreness. Moreover, recurrent cough after COVID-19 infection aggravated lumbodorsal pain[21].

The study results show that the VAS score for pain in OVCF patients during the normalized prevention and control period of COVID-19 in Shanghai is higher than that in the same period, which may be related to the cause and location of fractures in the two groups. Thoracic fractures caused by cough have higher VAS scores. However, there is no significant difference in VAS scores between the two periods after vertebroplasty. Therefore, vertebroplasty has a good therapeutic effect on thoracic OVCF caused by cough, and the therapeutic effect remains stable during the follow-up at 1 month after surgery.

Nucleic acid test result during this period has a significant correlation with fracture location and cause

Patients during this special period were grouped according to nucleic acid test results. The results showed that 80% of nucleic acid-positive patients had thoracic fractures, and 73% of them had cough as the cause of fracture. During the normalized prevention and control period of COVID-19, patients infected with the COVID-19 virus had positive nucleic acid test results, cough was the main clinical manifestation, and recurrent cough-induced thoracic stress fractures were the pathogenesis of fractures. Correlation analysis showed that there were positive correlations between nucleic acid test results, fracture location, and cause of onset. Among them, the correlation between cause of fracture and fracture location was the strongest (correlation coefficient r = 0.827, P < 0.01). The correlations between cause of fracture (r = 0.287) and fracture location (r = 0.332) with nucleic acid test results were weak. Therefore, for elderly patients with positive nucleic acid test results, it is necessary to attach great importance to the screening of OVCF, especially in the case of recurrent cough, and check whether there is percussion pain in the back.

Analysis of risk factors for thoracic vertebral OVCF

Logistic regression analysis revealed that fracture cause (P < 0.01, OR = 3.66), nucleic acid (P = 0.01, OR = 4.82), BMI (P < 0.01, OR = 4.69), and bone density (P < 0.03, OR = 3.52) were high-risk factors for thoracic OVCF. Among them, the odds ratio of nucleic acid was the highest. The probability area under the curve value of the regression equation was 0.961, indicating a good predictive effect. When coughing was the cause of fracture, nucleic acid was positive, BMI was less than 18.5, and bone density was less than or equal to -2.5, the incidence of thoracic fractures was as high as 96.1%. Additionally, it is worth noting that BMI and bone density are also high-risk factors for thoracic OVCF. These are patient’s own factors that cannot be controlled in the short term, while the cause of fracture and nucleic acid results can be prevented in the short term.

Risk factors for thoracic OVCF and prevention measures

To further identify the risk factors for thoracic OVCF during the normalized prevention and control period of COVID-19, binary logistic regression analysis showed that cause of fracture, nucleic acid test result, BMI, and BMD were risk factors for thoracic OVCF during this period. The OR from high to low were: Nucleic acid test result (OR = 4.82), BMI (OR = 4.69), cause of fracture (OR = 3.66), and BMD (OR = 3.52). The prediction accuracy of the regression equation was 0.961. Therefore, during the normalized prevention and control period of COVID-19, it is important to perform thoracic MRI examinations on elderly patients with cough, positive nucleic acid test results, and back pain to prevent missed diagnosis of thoracic OVCF.

In view of the adjustment of epidemic prevention policies, COVID-19 has been classified as a class B infectious disease managed with class B measures. After the full liberalization, medical institutions have received a large number of post-infection patients, among whom there are many patients with chronic cough as the main symptom. When these patients experience persistent and severe thoracolumbar pain during the course of the disease, it is necessary to be alert to the occurrence of OVCF. First, active treatment of COVID-19 and the use of antitussive drugs to relieve cough symptoms are essential. Thoracic brace fixation can well protect against thoracic stress fractures caused by cough. Anti-osteoporosis drugs such as teriparatide and denosumab can reduce the incidence of thoracic OVCF, and special attention should be paid to patients with low BMI. Therefore, during the normalized prevention and control period of COVID-19, reasonable and standardized diagnosis and treatment, as well as timely and effective intervention, are of great significance for protecting the rapid recovery of patients with thoracic OVCF.

Advantages and disadvantages of this study

This study retrospectively analyzed clinical cases, first proposed the relationship between COVID-19 infection and the causes and locations of thoracic OVCF, summarized the characteristics of thoracic OVCF during the normalized prevention and control period of COVID-19, and proposed effective prevention methods, which have high clinical guiding value. At the same time, there are some limitations. First, this study is a retrospective study, so bias may be introduced in data collection. Second, due to time constraints, no multi-center study was conducted, and the sample size was small, so the conclusions may not be comprehensive. A large-scale multi-center systematic analysis is still needed for the characteristics of OVCF patients during the COVID-19 epidemic.

CONCLUSION

During the normalized prevention and control period of COVID-19, compared with cases in the same period, cough is the main cause of OVCF, and most fractures are thoracic fractures. After PVP treatment, the improvement of pain and function in patients is not significantly different from that in the same period. Regression analysis shows that cause of fracture, nucleic acid test result, BMI, and BMD are independent risk factors for thoracic OVCF. For elderly COVID-19-positive patients with recurrent cough, timely and active treatment of lung diseases, aggressive anti-osteoporosis therapy, and wearing a thoracic brace for protection are effective measures to prevent thoracic OVCF.