It has been brought to our attention that Figure 2b in our recently published paper [1] contains an error. Specifically, we inadvertently entered -6.4 rather than +6.4 for the Nadaskevich paper [2] reported in Figure 2b. Recalculation of the summary estimate for diffusing lung capacity of carbon monoxide (DLCO) in this figure revealed a summary mean difference of 3.74, with 95% confidence interval (0.09 to 7.40). This effect is now statistically significant, although the overall conclusion of our meta-analysis, that cyclophosphamide treatment did not result in clinically significant improvement of pulmonary function, is unchanged by this recalculation, as the improvement remains less than the predefined criterion of 10%. The corrected figure 2b is given here as Figure 1.

Figure 1
figure 1

Forest plot of the overall meta-analysis result in the randomized clinical trials. Comparison of the diffusing capacity for carbon monoxide (DLCO) at 12 months for patients with scleroderma lung disease treated with cyclophosphamide versus a control group. See Table 2 [1] for study details. RCT, randomized clinical trial; SE, standard error; CI, confidence interval; Chi2, chi-squared; df, degree of freedom; I2, I-squared; Z, Z value; Mean difference, weighted mean difference; Random, random-effects.

Full size image