Psoas CT Attenuation and Fatty Infiltration Rate Predict Incisional Hernia After Appendicectomy


 Background Incisional hernia (IH), especially giant incisional hernia negatively affects a patient’s quality of life. Previous literatures often used muscle area measured by CT to assess loss of muscle mass and to estimate its relationship with diseases. For now, a few literatures have discussed potential effect of psoas muscle area on IH, while very limited attention has been paid to the association between psoas fatty infiltration and incidence of IH. In this study, we aimed to investigate if psoas CT measurement parameters, including average CT attenuation, fatty infiltration rate and psoas muscle index, were associated with IH.Methods Adult patients (> 18 years old) who had a history of appendicectomy and underwent CT examination in Beijing Chaoyang hospital from January 2018 to December 2019 were enrolled. All patients were classified into IH group and non-IH group. Psoas CT attenuation, fatty infiltration rate (FIR) and psoas muscle index (PMI) were measured or calculated. Sarcopenia was defined according to PMI. Differences in indices between the two groups were compared. Logistic regression model was applied to assess the effects of psoas CT measurement parameters on the occurrence of IH.Results 120 consecutive patients were included in this study. Psoas CT attenuation (p = 0.031) and PMI (p = 0.042) in IH group were significantly lower than those in non-IH group, and FIR in IH group was significantly higher than that in non-IH group (p < 0.001). Psoas CT attenuation, FIR, PMI, age, gender and smoking were significant factors in univariate logistic regression analysis. After adjusting for confounding factors, multivariate logistic regression analysis demonstrated that psoas CT attenuation was an independent protective factor (p = 0.042), and FIR an independent risk factor (p = 0.018); while PMI (p = 0.118) and sarcopenia (p = 0.663) showed no significant effect on incidence of IH.Conclusion Decreased CT attenuation and increased FIR of psoas are risk factors for IH after appendectomy.


Background
Since incisional hernia (IH), especially giant incisional hernia, negatively affects a patient's quality of life, body image and even cardiopulmonary function, it has received increasingly clinical attention. Computed tomography (CT) is a major preoperative examination for patients with abdominal incisional hernia. Besides accurately and intuitively displaying speci c contents and sizes of hernias [1], helping to estimate potential risk of intestinal adhesion or intestinal ischemia [2,3], CT scan of the abdomen can provide information about lumbar and dorsal muscle, such as muscle area, CT density and extent of fatty in ltration. Previous literatures often used muscle area measured by CT to assess loss of muscle mass and to estimate its relationship with diseases [4]. However, area alone may be inadequate when evaluating the muscle. Because in addition to volume change, muscle loss is accompanied by fatty in ltration [5,6]. For now, a few literatures have discussed potential effect of psoas muscle area on IH, while very limited attention has been paid to the association between psoas fatty in ltration and incidence of IH. In this study, by measuring psoas parameters on CT images in IH and non-IH patients after appendectomy, we aimed to investigate if psoas CT measurement parameters, including average CT attenuation, fatty in ltration rate (FIR) and psoas muscle index (PMI), were associated with IH.

Methods
Adult patients (> 18 years old) who had a history of appendicectomy and underwent CT examination in Beijing Chaoyang hospital from January 2018 to December 2019 were enrolled. Exclusion criteria were (a) various factors that affect muscle mass, including malignant tumors, taking medicine such as glucocorticoids or thyroxine, endocrine diseases, chronic exercise limitations, organ failure and uncontrolled cardiopulmonary diseases; (b) lumbosacral transitional anatomy. The patients were classi ed into IH group (con rmed clinically or surgically) and non-IH group. Data on age, gender, height, weight, smoking, hypertension and diabetes were recorded. Body mass index (BMI, kg/m 2 ) was calculated by dividing the weight in kilograms by the square of the height in meters. The Research Ethics Review Committee of Beijing Chaoyang hospital had reviewed and approved this study.
CT scan and image measurements CT scan was performed on gemstone CT spectroscopy (Discovery CT 750 HD, GE Healthcare). Imaging protocols included: 120kVp, tube current modulation, pitch 0.984, slice thickness 5 mm, slice spacing 5 mm. Reconstructed image with 0.625 mm slice thickness was obtained for further analysis. On GE AW 4.6 workstation, bilateral psoas muscles were measured separately at the level of the third lumbar vertebra (L3), on which both the transverse processes were visible, as previously described [7]. By manually outlining the left and the right psoas, cross-sectional area (CSA, cm 2 ), CT attenuation, and FIR (%) based on a Houns eld unit (HU) threshold for adipose tissue (-190 HU to -30 HU) were automatically calculated. PMI (bilateral CSA adjusted for a patient's height squared, cm 2 /m 2 ), bilateral mean CT attenuation and FIR were calculated. According to previous literature about Asian [8], a male PMI < 5.923cm 2 /m 2 or a female PMI < 3.999 cm 2 /m 2 was de ned as sarcopenia.

Statistical analysis
Quantitative data were either represented as mean ± standard deviation (SD) or median with interquartile range (IQR). Qualitative data were reported as counts and percentages. Quantitative variables were compared between the two groups by using independent sample t-test or Wilcoxon signed rank test.
Qualitative variables were compared using chi-square test. Logistic regression model was created to assess potential in uence of psoas muscle measurement parameters on risk of IH. All the above statistical analyses were performed using SAS version 9.1 (SAS, Cary, NC, USA). The MedCalc software was used for receiver operating characteristic (ROC) curve analysis to compare diagnostic e ciency of CT attenuation and FIR for IH. Statistical signi cance was de ned at a p < 0.05 (2-sided).
Results 120 consecutive patients were included in this study.
Clinical data and CT measurements between IH and non-IH groups (Table 1)  Psoas CT attenuation (t = -2.18, p = 0.031) and PMI (S = 3242, p = 0.042) in IH group were signi cantly lower than those in non-IH group, and FIR in IH patients was statistically higher than that in non-IH patients (S = 4310, p < 0.001). IH patients were, on average, 5 years older than non-IH patients (t = 2.26, p = 0.026). In IH group, more than half was women (63.3%); by contrast, female accounted for only 45.0% in non-IH group (χ 2 = 4.06, p = 0.044). In IH group, the proportion of smokers was 1.4 times of that in non-IH group (χ 2 = 7.78, p = 0.005). However, BMI, diabetes and hypertension were comparable between the two groups (p > 0.05).

Discussion
Incisional hernia is a frequent and well described complication to abdominal surgery, which causes discomfort, pain, and reduction of quality of life. Nearly 350,000 hernia repairs are performed annually, costing approximately $3 billion dollars in the USA [9]. The main known risk factors for IH development include surgical site infection, obesity, smoking, older age, gender (mostly reported as female), hypertension, diabetes, and use of corticosteroids [10][11][12][13][14].
CT scan has become a routine examination for abdominal IH, especially for huge hernia. Besides demonstrating the location, size and content of hernia, as well as helping to determine the presence of intestinal obstruction and necrosis, CT examination of the abdomen can quantitatively measure body composition, in particular visceral fat area and abdominal wall subcutaneous fat content, abdominal wall muscle and psoas muscle mass. CT measurement of body composition is quite reliable. For example, cross-sectional CT measurements of muscle and subcutaneous fat tissue are consistent with autopsy results [15]. Furthermore, fat and fat-free tissue measurement at L3 level using CT strongly correlate with whole-body fat and fat-free mass [16]. Numerous studies have delineated that visceral fat area and subcutaneous fat area measured on CT image can predict re-herniation and surgical site infection [17,18]. Moreover, visceral obesity, rather than elevated BMI, is associated with IH after colorectal surgery [19,20]. To date, there is very limited knowledge on the relationship between psoas CT measurement parameters and susceptibility to IH.
In this study, we recruited patients who underwent appendicectomy due to appendicitis, and divided them into two groups based on whether they had developed IH or not. We explored the relationship between psoas CT measurements and IH. After adjusting for age, gender and smoking, psoas CT attenuation was a protective factor for IH, and FIR a risk factor for IH; while PMI and sarcopenia hardly affect the occurrence of IH.
Muscles are composed of muscle bers and intramuscular adipose tissue. Unlike erector spinae, psoas atrophy is mainly manifested by volume reduction and morphological changes, with unobvious fat deposition. Therefore, previous studies on IH had mainly focused on psoas muscle area and skeletal muscle index (SMI). However, in addition to volume reduction, muscle atrophy may pathologically exhibit fatty in ltration and muscle bers loss [5,6]. Some authors believe that CT attenuation may accurately re ect the number of muscle bers and the degree of fat deposition [21]. Muscle CT attenuation is related to various diseases: paraspinal muscle density was associated with facet joint osteoarthritis, spondylolisthesis and disc narrowing at the same level [22]; lower thigh muscle CT attenuation could increase the risk of hip fracture, and a 1 SD decrease in thigh muscle HU value conferred a nearly 40% increase in the risk of hip fracture [23]; for critically ill adult patients treated with mechanical ventilation, those with a lower skeletal muscle CT value at admission had a higher 6-month mortality rate, and a 10 HU increase in muscle density was associated with a 14% decrease in hospital length of stay [25]; cancer patients with cachexia and low muscle CT values had a poor prognosis [25].
Skeletal muscle density inversely correlated with length of hospitalization after complex abdominal wall hernia surgery [26]. In our study, patients with lower psoas attenuation were more likely to develop IH. CT attenuation of abdominal wall muscles may correlate with that of psoas [27]. Decreased CT attenuation of muscle is independently associated with muscle weakness [28]. Therefore, lower CT attenuation of psoas indirectly re ects weakness of the abdominal wall muscle, which is susceptible to hernia [29].
In our study, patients with higher FIR of psoas were prone to develop IH. Fatty in ltration of skeletal muscle has been identi ed as a possible cause of loss of muscle quality [28]. Fatty in ltration induces insulin resistance, which impairs normal capacity for protein synthesis, subsequently contributing to muscle atrophy [30,31]. Previous studies often applied magnetic resonance imaging (MRI) to quantitatively measure muscle fat content [32,33], and concluded that paraspinal fat in ltration, rather than muscle CSA, was associated with high-intensity pain/disability and structural abnormalities in the lumbar spine [34]. Moreover, in patients with L4-5 single-segment degenerative lumbar spinal stenosis, fatty in ltration in the multi dus muscles at L5-S1 may be correlated with the disc bulge at the stenosis segment and reduction of lumbar lordosis [35]. Since MRI examination is not routinely performed on hernia patients, MRI-measured muscle fat content in these patients is not widely-applied. Prior literatures about CT imaging mostly used HU value to indirectly represent muscle fat content. However, since muscle CT attenuation can be affected by previous surgery and deposition of high-density substances, such as calcium and bleeding, it may not effectively re ect fat content of muscle. WC et al reported that psoas CT attenuation in patients with osteoporosis fracture was unexpectedly higher than those without osteoporosis fracture [36], which may be related to intra-muscular hemorrhage or muscle repair after fracture. In this case, CT attenuation cannot accurately re ect muscle ber content as well as the degree of fat accumulation. Although relatively complex, measuring intramuscular fat area or muscle FIR by de ning CT threshold can more accurately re ect the degree of muscle fat in ltration. When intramuscular fat area was quantitatively measured by CT, rotator cuff fat in ltration conferred functional recovery after shoulder arthroplasty [37]. To our knowledge, this is the rst study to explore potential relationship between FIR and IH. We concluded that compared with CT attenuation, FIR was more closely related to IH.
Previous studies, to investigate the roles of abdominal muscles in malignancies, mostly measured muscle area of L3 or L4 cross-sectional image, including psoas, erector spinae, quadratus lumborum, transversus abdominus, rectus abdominus, as well as internal and external obliques. The corresponding SMI might predict prognosis of various malignancies [7,38]. Based on SMI (calculated by measuring L3 level cross-section muscles), sarcopenia was not a risk factor for IH [39]; but it prolonged postoperative hospital-stay [26].
Since the measurement method mentioned above needs to be performed on a speci c post-processing software by de ning the range of CT values to exclude intermuscular fat, which is very labor-intensive, its application in clinical is very limited. To measure psoas area at L3 level only is simple, and corresponding PMI correlates with the whole-body muscle mass [40]. A decrease in PMI indicates a decline in wholebody muscle mass (including abdominal wall muscles), which results in decreased functional capacity.
Thus, PMI may be a potential risk factor for IH. In our research, PMI at L3 level was a protective factor for IH in univariate regression analysis, but not statistically signi cant in multivariate regression analysis. Sarcopenia, as de ned by cut-off values in previous study (based on Asians), was not associated with IH neither in univariate nor in multivariate regression analysis.
This study has several limitations. First, wound infection after surgery can be prone to IH. However, some patients did not remember clearly whether they had postoperative infection or not, since more than ten years or even decades had passed from appendectomy to this admission. Taking recall-bias into consideration, wound infection was not included in statistical analysis. Second, cross-sectional retrospective study design limits our ability to ascertain causality, and may cause selection bias inevitably. Thus, our conclusion needs to be veri ed by a larger sample size.

Conclusion
In summary, our study indicates that decreased CT attenuation and increased FIR of psoas are risk factors for IH after appendectomy. For patients, who have a history of abdominal surgery, with low psoas CT attenuation or high psoas FIR, prophylaxis of hernia should be conducted to reduce the occurrence of IH.