Referee 2's Review:

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication, the manuscript has been revised in light of these comments and to address other editorial requirements.
**********

Review of the original submission:
In this study, the authors confirmed that hyper-hemolysis as measured by serum lactate dehydrogenase (LDH) levels in sickle cell anemia (HbSS) was associated with a vasculopathy (pulmonary hypertension, priapism, leg ulcers and an increase in systolic BP) in an independent group of patients. The observations were originally made in a cohort of NIH sickle patients and previously published. Patients recruited in the Cooperative Study of Sickle Cell Disease (CSSCD) recruited between 1978 and 1988 served as the validation set. The major limitation in this validation study is that it is retrospective, several of the parameters and assays used in the NIH cohort were not measured in the CSSCD group and surrogate measures had to be used, for example, BNP as a surrogate for pulmonary hypertension. As a result of various filters and exclusion measures, the original population of 4085 CSSCD patients distilled down to a population of 451. The authors selected patients with sickle cell anemia (HbSS) and filtered off those below 30 yrs and excluded those with ALT values greater than 80 IU/L. I find this paper extremely difficult to follow because of the numerous inclusion and exclusion criteria and filtered subsets.

Comments
1. Patients on hydroxyurea and blood transfusion therapy were included in the NIH group and then adjusted, in the analysis. How do the authors calculate the correction values? I would think it is more appropriate to exclude these patients and use the cohort as detailed in appendix table 2. In this way we can compare like for like since none of CSSCD patients were on HU therapy.
2. I note a minimum LDH value of 54 in the CSSCD cohort. This seems to be an outlier (see distribution in SCD-SC). Shouldn't this individual be excluded from the analysis?
3. Table 1 - patients are supposed to be homozygous for HbS, i.e. HbSS, so there should not be any HbA. The presence of HbA suggests that the genotype is HbSβ+ thalassaemia or a recent blood transfusion which confounds interpretation of the various lab measures.
Those patients with HbA should not be included in the analysis.
4. Fetal haemoglobin should be given as gm/L as for total haemoglobin, as well as percentage of the total haemoglobin rather than mol/mol.
5. Since this is a validation study, listing measures such as reticulocytes, arginine, arginine : ornithine ratio, plasma haemoglobin, etc. in Table 1, and haptoglobin etc, in Table 2, that are available only in the NIH but not the CSSCD cohort, is not valid.
6. Table 2 - the total number of patients in the whole study should be clearly stated. The 'n' column, i.e. number available for each category, should be explained in the footnote.
7. Table 2 should be Table 1, as it is cited earlier.
8. Finally, according to the tables, only leg unl;cers and pulmonary hypertension ( surrogate marker BNP) have been validated in the replication set

Review of the first revised manuscript:
All points addressed.