臺灣中部地區骨科兵役複檢役男扁平足之放射線學分析

Radiographic Analysis of the Flatfoot of the Mid-Taiwan Military Recruits (Draftees) at an Orthopaedic Department

林育民(Yu-Min Lin)

指導教授 ： 周明智

中山醫學大學/醫學院/醫學研究所/碩士(2005年)

摘要

背景及目的： 在台灣，成人扁平足(adult flatfoot)並沒有系統性的大規模流行病學研究。自民國九十年六月一日起，全台兵役複檢由軍醫院轉移至七間大型醫院。因扁平足需接受複檢之役男應內政部之體位區分標準之規定皆需接受足部負重正側位X光攝影(weight-bearing true lateral view)，並規定足弓角 > 165度者為扁平足。藉此機會可分析中部地區複檢役男成人扁平足在骨科兵役複檢之盛行率並探討體位區分標準之足弓角定義之信度、效度等問題。

方法： 自92年1月1日起至93年12月31日止至台中榮民總醫院接受骨科兵役複檢之役男共2663人次，共2475人。其中關於足部問題而接受左右足X光正側位站立攝影共373人納入本研究，作者記錄其左右足足弓角之報告。作者並重新調閱其X光影像，遮蔽其原始報告，重新度量其足弓角、距骨蹠骨角及足跟角。使用統計軟體作單變數分析、相關分析、配對檢定、receiver operator characteristic (ROC) curve 、信度(intraclass correlation coefficient, Lin’s concordance correlation coefficient, Bland-Altman plot)、效度(ROC curve comparison)等分析，計算成人扁平足在骨科兵役複檢之觀察盛行率。

結果: 373位役男平均年齡23 +/- 1.9歲。作者可調閱363位役男X光影像重新度量其足弓角、距骨蹠骨角及足跟角。雙足足弓度同時>165度具有顯著一致性(P < 0.001)。以距骨蹠骨角< -4度為扁平足之黃金標準，雙足距骨蹠骨角同時< -4度具有顯著一致性(P < 0.001)，至少一足距骨蹠骨角< -4度335位，故成人扁平足在骨科兵役複檢觀察盛行率為13.53%。足弓角之度量：在同一影像不同度量者(interrater)彼此間一致性高(Kappa = 0.826, Lin’s concordance correlation coefficient = 0.9458, intraclass correlation coefficient = 0.9459)﹔在不同次攝影同一度量者其一致性良好(Lin’s concordance correlation coefficient = 0.8164 (報告)和0.8458 (作者), intraclass correlation coefficient = 0.8193(報告)和 0.8469(作者)。而距骨蹠骨角在不同次攝影同一度量者其一致性中等(Lin’s concordance correlation coefficient = 0.7224, intraclass correlation coefficient = 0.7262)。足跟角在不同次攝影同一度量者其一致性良好(Lin’s concordance correlation coefficient = 0.8993, intraclass correlation coefficient = 0.9023)。在不同次攝影同一度量者，在Bland-Altman plot 上，除了右側距骨蹠骨角平均差異稍大(1.1652度)，其他足跟角、足弓角、距骨蹠骨角平均差異皆 < 1度。以距骨蹠骨角 < -4度為扁平足之黃金標準所繪製的ROC曲線，足弓角面積 = 0.781，足跟角面積 = 0.837，兩者面積有統計上的差異(P < 0.001)。足弓角之最佳cutoff point 為 > 164度，敏感度47.9%，特異度92.6%。若足弓角之cutoff point 定為 > 165度，敏感度40.6%，特異度94.4%。以役男之年齡、受檢足部、足弓角、跟骨角這四個變數作出多元線性迴歸模型來預測距骨蹠骨角，此模型具統計上之顯著性(P < 0.001, 判定係數R2 = 0.4097)﹔標準化係數：足跟角0.73 (P< 0.001)、足弓角0.12 (P=0.432)、年齡 -0.06 (P = 0.0286)、受檢足部 0.04 (P = 0.1247)。

結論： 台灣中部地區役男扁平足的在骨科兵役複檢觀察盛行率較加拿大(1947年時23%)稍低。以疾病篩檢的角度言，足弓角較足跟角差﹔由多元迴歸分析可知足跟角較能反應出扁平足距骨及蹠骨間的變化。以兵役體檢角度言，足弓角較為公平，因特異度高。足弓角度量在不同度量者之間具有高度的一致性，在不同次攝影時其一致性也良好。

關鍵字

成人 ； 役男 ； 扁平足 ； 足弓角 ； 足跟角 ； 距骨蹠骨角 ； 敏感度 ； 特異度 ； 信度 ； 效度 ； 盛行率 ； Receiver operator characteristic (ROC) curve

並列摘要

Background and purpose: There is no systemic epidemiologic study of adult flatfoot in Taiwan. The obligation of physical check-up of draftees (military recruits) was transferred to the seven major hospitals from the military hospitals since June 1 2001. The Physical Classification Criteria of the Ministry of Interior demands the draftees with flatfeet to take the weight-bearing true lateral views of their feet, and they are classified as flatfeet when the calcaneal-5th metatarsal angles are greater than 165 degrees. With this unique opportunity, the author calculated the apparent prevalence of the adult flatfoot among the draftees in mid-Taiwan examined at an Orthopaedic department and analyzed the reliability and validity of the criteria of flatfoot classification used by the Ministry of Interior.

Methods: From January 1 2003 to December 31 2004, a total of 2475 draftees underwent physical checkup at the Orthopaedic Department of Veterans General Hospital Taichung. A total of 373 draftees took the true lateral radiographs of their bilateral feet was included in the study. The author charted down the results of the report of the radiologist. The author, blinded to the report, reviewed the radiographs and measured the calcaneal pitch, talus-1st metatarsal angle, and calcaneal-5th metatarsal angle. Statistical packages were used for univariate analyses, correlation analyses, paired t-test, chi square test, and receiver operator characteristic curve analyses. Reliabilities were examined with Lin’s concordance correlation coefficients and intraclass correlation coefficients. Validities were assessed with ROC curve analyses. The apparent prevalence of adult flatfoot at the Orthopaedic Department was calculated.

Results: The average age of the 373 draftees was 23 +/- 1.9 years. The radiographs of 363 draftees were available for calcaneal pitch, talus-1st metatarsal angle, and calcaneal-5th metatarsal angle measurements. The concordance of the calcaneal-5th metatarsal angle > 165 degrees of both feet is statistically significant (P < 0.001). With the talus-1st metatarsal angle < -4 degrees as the gold standard, the concordance of the talus-1st metatarsal angle < -4 degrees is also statistically significant (P < 0.001). There were 335 draftees at least with the talus-1st metatarsal angle < -4 degrees in one foot, so the apparent prevalence of adult flatfoot at the Orthopaedic Department was 13.53%. Concerning the measurements of the calcaneal-5th metatarsal angle, the interrater agreement was statistically significant (kappa = 0.826, Lin’s concordance correlation coefficient = 0.9458, intraclass correlation coefficient = 0.9459). The agreement of radiographs taken on two different occasions was good for the measurement of the calcaneal-5th metatarsal angle. Lin’s concordance correlation coefficients were 0.8164 and 0.8458 by the radiologist and the author, respectively; intraclass correlation coefficients were 0.8193 and 0.8469 by the radiologist and the author, respectively. The agreement of radiographs taken on two different occasions was fair for the measurement of the talus-1st metatarsal angle, and the Lin’s concordance correlation coefficient and intraclass correlation coefficient were 0.7224 and 0.7262, respectively. The agreement of radiographs taken on different occasions was good for the measurement of the calcaneal pitch, and the Lin’s concordance correlation coefficient and intraclass correlation coefficient were 0.8993 and 0.9023, respectively. With the same rater reviewing the radiographs taken on the two different occasions, the mean difference between the calcaneal pitch, calcaneal-5th metatarsal angle was less than 1 degree on the Bland-Altman plot, except the mild increased difference of the talus-1st metatarsal angle of the right foot. With the talus-1st metatarsal angle < -4 degrees as the gold standard, the area under the ROC curves was 0.781 and 0.837 for the calcaneal-5th metatarsal angle and calcaneal pitch, respectively. The area under these two curves differed statistically (P < 0.001). From the ROC curve of the calcaneal-5th metatatarsal angle, the best cutoff point was > 164 degrees with the sensitivity of 47.9% and specificity of 92.6%. If the cutoff point was selected for calcaneal-5th metatarsal angle > 165 degrees, the sensitivity and specificity would be 40.6% and 94.4%, respectively. With the age of the draftees, side of foot, calcaneal-5th metatarsal angle, and calcaneal pitch as multiple linear regression variables to estimate the talus-1st metatarsal angle, the multiple regression model is statistically significant (P < 0.001, R2 = 0.4097). The standardized coefficients were as follows: calcaneal pitch, 0.73 (P < 0.001); calcaneal-5th metatarsal angle, 0.12 (P = 0.432); age, -0.06 (P = 0.0286); side of foot, 0.04 (P = 0.1247).

Conclusion: The apparent prevalence of adult flatfoot in the Mid- Taiwan is lower than the prevalence of Canadian soldiers in 1947 (23%). From the viewpoint of the disease screening tools, the calcaneal-5th metatarsal angle was not better than the calcaneal pitch. From the results of the multiple linear regression, the calcaneal pitch might reveal the relationship of the talus between metatarsals in adult flatfoot. From the viewpoints of conscription, the calcaneal-5th metatarsal angle was better because of higher specificity. The concordance of the calcaneal-5th metatarsal angle between the observers and on the different radiographic occasion was good.

並列關鍵字

adult ； draftee ； military recruits ； calcaneal-fifth metatarsal angle ； calcaneal pitch ； talometatarsal angle ； sensitivity ； specificity ； reliability ； validity ； prevalence ； Receiver operator characteristic (ROC) curve

參考文獻

19 Incel NA, Genc H, Yorgancioglu ZR, Erdem HR. Relation between hallux valgus deformity and lumbar and lower extremity biomechanics. Kaohsiung J Med Sci 2002;18(7):329-33.

2 Mosca VS. Flexible flatfoot and skewfoot. Instr Course Lect 1996;45:347-54.

3 Kanatli U, Yetkin H, Cila E. Footprint and radiographic analysis of the feet. J Pediatr Orthop 2001;21(2):225-8.

4 Vanderwilde R, Staheli LT, Chew DE, Malagon V. Measurements on radiographs of the foot in normal infants and children. J Bone Joint Surg Am 1988;70(3):407-15.

5 Steel MW, 3rd, Johnson KA, DeWitz MA, Ilstrup DM. Radiographic measurements of the normal adult foot. Foot Ankle 1980;1(3):151-8.