Chronic kidney disease (CKD) causes a continuous decline in kidney function and structural damage to the kidneys. The speech characteristics of CKD speakers will be different from those of non-CKD speakers because the typical characteristics of CKD, which are impairment of respiratory and laryngeal muscles, can affect respiration, the primary source of speech. In this paper, we identify the glottal characteristics of CKD speech and then investigate whether CKD can be automatically detected using the glottal features. Statistical analysis shows significant differences between groups in glottal source features, representing the breathy characteristic of CKD speech. Through the classification experiment, we compare the performance of solely using voice quality features (baseline) against additional glottal and spectral features. When glottal source features and voice quality features are used together, an F1-score of 0.88 with a 76% relative increase compared to the baseline is obtained.