Comments on "Effects of Patient and Tumor Characteristics on Central Lymph Node Metastasis in Papillary Thyroid Cancer: a Guide for Selective Node Dissection"

Recently, Altiner et alpublished an article entitled above and showed that gender (OR = 1.22), tumor subtype (OR = 0.44) and lymphovascular invasion (OR = 2.59) significantly increase the odds of central lymph node metastasis (CLNM), while intra-thyroidal localization of the tumor is not significantly associated with CLNM based on the multiple logistic regression.¹ However, these findings are doubtful due to the following reasons:

• The estimated crude odds ratios are not true for women (correct OR = 0.51), lymphovascular invasion (correct OR = 138.19) and extracapsular invasion (correct OR = 8.87).

• In Table 2, for not applicable ORs, 95% confidence intervals were reported!

• The negative values in the confidence interval of odds ratios are definitely wrong because the odds ratio is always positive.

• The adjusted OR for lymphovascular invasion is 2.59, but the 95% confidence intervals range from 4.47 to 7.9 which is impossible.

• In the multiple logistic regression, just one OR was reported for each of the quantitative variables with more than two categories. Considering a category as the baseline, four ORs must be reported for intra-thyroidal localization and two ORs for tumor subtype. So, the current ORs are not valid.

• The confidence interval of OR for gender ranges from 0.99 to 1.55 which contains one and cannot be significant; so, P value = 0.025 is not true and the correct P value is greater than 0.05, indicating gender is not statistically significant.

I hope this note shows some deleterious effects of inappropriate statistical analysis in medical research.

Cite this article as: Raeisi Shahraki H. Comments on "effects of patient and tumor characteristics on central lymph node metastasis in papillary thyroid cancer: a guide for selective node dissection". Arch Iran Med. 2024;27(8):465. doi: 10.34172/aim.27317