Shann F, Mackenzie A
Goroka Base Hospital, Papua New Guinea.
Arch Pediatr Adolesc Med. 1996 Jan;150(1):74-8. doi: 10.1001/archpedi.1996.02170260078013.
To assess whether axillary and forehead temperatures accurately reflect the rectal temperature (the criterion standard)
Prospective study with calculation of paired axillary-rectal and forehead-rectal temperature differences and their SDs.
Referral hospital.
Convenience sample of 120 patients, with 20 patients in each of six age groups (ie, < 1 month, 1 to 5 months, 6 to 11 months, 12 to 23 months, 2 to 14 years, and adults)
In newborns, the rectal temperature was equal to the axillary temperature plus 0.2 degrees C for each week of age up to 5 weeks; forehead strip thermometers gave inaccurate readings in this age group. In patients older than 1 month, the mean difference (SD) between the rectal and axillary temperatures was 1.04 degrees C (0.45 degrees C); thus the axillary temperature was adjusted by adding 1 degree C, and no adjusted axillary temperature differed from the rectal temperature by more than 1 degree C. The mean difference (SD) between the forehead temperature that was measured by the best forehead liquid crystal strip thermometer (FeverScan) and the rectal temperature was 0.14 degrees C (0.60 degrees C); 10 forehead temperatures differed from the rectal temperature by more than 1 degree C.
Previous studies that have suggested that axillary and forehead temperatures do not provide a reliable guide to the rectal temperature have all used inappropriate methods of analysis (correlation coefficients or sensitivity and specificity); previous studies that have based their conclusions on the correct method of analysis (paired differences and their SDs) have all found that the axillary temperature gives a good indication of the rectal temperature. The axillary temperature can be measured safely at any age, and the axillary temperature plus 1 degree C is a good guide to the rectal temperature in patients older than 1 month. Forehead strip thermometers are easy to use, but they do not estimate the rectal temperature as accurately as the axillary temperature does.
评估腋窝温度和额头温度能否准确反映直肠温度（标准对照）。
前瞻性研究，计算配对的腋窝-直肠温度差和额头-直肠温度差及其标准差。
转诊医院。
便利样本，共120例患者，分为六个年龄组，每组20例（即小于1个月、1至5个月、6至11个月、12至23个月、2至14岁及成年人）。
在新生儿中，直至5周龄，直肠温度等于腋窝温度加上每周0.2摄氏度；该年龄组中额头条式体温计读数不准确。在1个月以上的患者中，直肠温度与腋窝温度的平均差值（标准差）为1.04摄氏度（0.45摄氏度）；因此，腋窝温度需加1摄氏度进行校正，校正后的腋窝温度与直肠温度相差不超过1摄氏度。最佳额头液晶条式体温计（FeverScan）测量的额头温度与直肠温度的平均差值（标准差）为0.14摄氏度（0.60摄氏度）；10例额头温度与直肠温度相差超过1摄氏度。
以往认为腋窝温度和额头温度不能可靠反映直肠温度的研究均采用了不恰当的分析方法（相关系数或敏感度及特异度）；以往基于正确分析方法（配对差值及其标准差）得出结论的研究均发现腋窝温度能较好地反映直肠温度。腋窝温度在任何年龄均可安全测量，1个月以上患者，腋窝温度加1摄氏度是直肠温度的良好指标。额头条式体温计使用方便，但对直肠温度的估计不如腋窝温度准确。
