Among the 52 recruited subjects, 35 (67%) were male and 17 (33%) females. Clinical characteristics and metabolic markers of the entire cohort at baseline and during the follow up are reported in Table 1. All subjects were receiving hydrocortisone at a mean dose of 14.6 ± 5.1 mg/m2 that did not change during the study period (p > 0.05). Forty-three out of the 52 21-OHD subjects were affected by the SW form and were also receiving fludrocortisone therapy (0.1 mg/day). Eight subjects at T0, 12 subjects at T1 and 11 subjects at T2 were also receiving cholecalciferol supplementation at the dose of 1000 UI per day.

In the Table 1 are reported the values of each variable at the three time points. A statistically significant difference was found for 17-OHP (p = 0.005), with median values ranging from 11.1 (3.0–25.1) ng/mL at T0 to 5.9 (2–20) ng/mL at T2, and for Δ4-androstenedione, with median values ranging from 0.9 (0.3–2.5) ng/mL at T0 to 0.5 (0.3–1.5) ng/mL at T2 (p = 0.014). ACTH levels showed a strong downward trend from T0 (median value 58.5 pg/mL) to T2 (median value 35 pg/mL), although a statistically significant difference was not observed. A statistically significant difference was also found for 25-OH vitamin D levels which showed an increase from T0 to T2 (p = 0.004), probably due to the cholecalciferol supplementation. PTH, Ca, P, Na and K serum concentrations were in the normal range for all patients and stable during the study time. Mean values of the Ca/Cr ratio were stable during the three time-points.

Of the 52 patients recruited for this study, 11 subjects (10 males) showed at least one ultrasound finding of nephrolithiasis, and only one subject showed a single finding of nephrocalcinosis.

At T0, 9 subjects had nephrolithiasis (17.3%), and 1 showed nephrocalcinosis (1.9%); during the follow up, 7 subjects showed nephrolithiasis (13.5%) at T1, and 6 subjects (11.5%) at T2.

Among the 21-OHD subjects showing nephrolithiasis, the most of them had the SW form of 21-OHD (8 at T0, 5 at T1, and 6 at T2), thus they were receiving both hydrocortisone and fludrocortisone.

Among the 11 21-OHD patients who presented with nephrolithiasis, 4 showed this condition in all the three follow-up visits, 2 subjects at T0 and T1, and 5 subjects had a single finding of nephrolithiasis (3 subjects at T0, and 2 subjects at T1). The child who showed nephrocalcinosis at T0 was found to have nephrolithiasis at T1 and T2.

The Table 2 shows the comparison of ACTH, 17-OHP, DHEAS and Δ4-androstenedione values among the three time points in 21-OHD patients with nephrolithiasis. No statistically significant difference among these metabolic markers was found, even if 17-OHP, DHEAS and Δ4-androstenedione levels showed a tendency towards a reduction from T0 to T2.

Although the mean dose of hydrocortisone remained stable in the entire cohort during the study period, the subjects with nephrolithiasis showed an increase in the mean dose even if not statistically significant.

Among the subjects with nephrolithiasis, none of them had a family history of kidney stones; 3 subjects, including the one with nephrocalcinosis, had mild hypercalciuria at T0, 2 subjects at T1 and 2 subjects at T2. During the follow-up period, serum Na, K, Ca, P, PTH levels were in the normal range for all subjects, as well as urinary excretion of Na and K. Urinary oxalate excretion was slightly increased in 2 subjects at T0, while urinary citrate excretion was reduced in 4 subjects at T0 and in 2 subjects at T1 and T2.

Table 3 shows comparisons between subjects with and without nephrolithiasis during the three time points.

Two principal components were fixed as observed by scree plot (Fig. 1A). The first and the second principal component (PC1 and PC2) explained 25.5 and 21.5% of overall variance, respectively, with a total variance explained equal to 47%. The importance of each variable to build principal components was investigated. Interestingly, as shown in Fig. 1B, DHEAS, Ca/Cr ratio, 17-OHP, Δ4-androstenedione variables explained, each one, more than 60% of variance, followed by ACTH (50% of variance explained), and the others that explained not much about variability over time. These results were confirmed observing loading plot in Fig. 2A: variables with longer vectors had a strong weight in the principal components, while markers with short vector did not have relevance in explaining variability and consequently differences over time. Moreover, from loading plot, correlations between variables were observed. In particular, two strong correlation patterns among 17-OHP, Δ4-androstenedione and ACTH, and among DHEAS and Ca/Cr ratio were found (Fig. 2A, blue and orange circles). Considering the score plot (Fig. 2B), a trend over time points (black, red and green circles) along negative direction of PC1 and PC2 was found. As the trend was observed along the negative directions of PC1 and PC2, moving from the first time point (black dot) to the end point (green dot), these results suggested that a decrease in the 17-OHP, Δ4-androstenedione and ACTH amounts occurred during the follow up.

A, B Scree plot: percentage of explained variance vs number of principal component of the model (in general the components number before the inflection point are retained); variance of each variable explained: the weight of each variable to the two principal components selected in the PCA model

A, B Loading and score plots of PCA (principal components 1 and 2) of variables. In the loading plot, blue and orange circles represent correlation patterns between variables. In the score plot, black, red and green circles (and respective dots A–C) represent groupings of the variables amount at starting time point, first time point, and the end point, respectively

The decrease in the variables like 17-OHP, Δ4-androstenedione and ACTH over time, matched with the decrease in the cases of nephrolithiasis during the follow up.