Background Macroalbuminuria predicts renal and cardiovascular events in patients with type 2 diabetes. We aimed to assess the albuminuria-lowering effects of salt restriction, paricalcitol therapy, or both, in this population. Methods In this randomised, double-blind, placebo-controlled, crossover trial, we recruited adult patients with type 2 diabetes from six diabetology outpatient clinics in northern Italy, with 24 h albuminuria of more than 300 mg despite 100 mg per day losartan therapy, blood pressure of less than 140/90 mm Hg, serum creatinine concentration of less than 2 mg/dL, stable renal function on stable renin-angiotensin system inhibitor therapy with a fixed dose of losartan, parathyroid hormone concentration of 20 pg/mL to <110 pg/mL, serum calcium concentration of less than 9·5 mg/dL, and serum phosphate concentration of less than 5 mg/dL, who had been more than 80% compliant with placebo treatment during a 1 month placebo run-in. We allocated patients 1:1 with computer-generated randomisation to an open-label 3 month high-sodium (>200 mEq [4·8 g] per day) or low-sodium (<100 mEq [2·4 g] per day) diet and, within each diet group, to a 1 month double-blind treatment period of oral paricalcitol (2 μg per day) or placebo, followed by 1 month of placebo washout and then a further 1 month double-blind treatment period of paricalcitol or placebo in which patients crossed over to the opposite treatment period. The primary outcome was 24 h albuminuria (median of three consecutive measurements). Analyses were modified intention-to-treat (including all randomly allocated patients who took at least one dose of study drug and had an efficacy measurement after the first treatment period). Patients and investigators were masked to paricalcitol and placebo assignment. Those assessing outcomes were masked to both study drug and diet assignment. This study is registered with ClinicalTrials.gov, number NCT01393808, and the European Union Clinical Trials Register, number 2011-001713-14. Findings Between Dec 13, 2011, and Feb 17, 2015, we randomly allocated 57 (50%) patients to a low-sodium diet (28 [49%] to paricalcitol then placebo and 29 [51%] to placebo then paricalcitol) and 58 (50%) to a high-sodium diet (29 [50%] to paricalcitol then placebo and 29 [50%] to placebo then paricalcitol). In the low-sodium group (30 mEq of daily sodium intake reduction, equivalent to approximately 1·7–1·8 g per day), 24 h albuminuria was reduced by 36·6% (95% CI 28·5–44·9) from 724 mg (441–1233) at baseline to 481 mg (289–837) at month 3 (p<0·0001), but no significant change occurred in the high-sodium group (from 730 mg [416–1227] to 801 mg [441–1365]; 2·9% [–16·8 to 16·4] increase; p=0·50). Changes between diet groups differed by 32·4% (17·2–48·8; p<0·0001) and correlated with changes in natriuresis (r=0·43; p<0·0001). On the high-sodium diet, paricalcitol reduced the salt-induced albuminuria increase by 17·8% (3·9–32·3) over the month of treatment compared with placebo (p=0·02), whereas on the low-sodium diet, paricalcitol did not have a significant effect versus placebo (increase of 4·1% [–9·3 to 21·6]; p=0·59). During placebo treatment, albuminuria decreased with the low-sodium diet (p=0·0002) and did not significantly change with the high-sodium diet, but changes were significantly different between diet groups (p=0·0004). Treatment was well tolerated and no patients withdrew from the study because of treatment-related effects. 67 adverse events occurred in 52 (45%) patients during paricalcitol treatment and 44 events occurred in 36 (31%) patients during placebo treatment. During paricalcitol therapy, 14 cases of hypercalciuria, six cases of hypercalcaemia, and five cases of hyperphosphataemia were reported in one patient each, all of which were possibly treatment related. One case of hypercalciuria was reported in one patient during the placebo treatment period. One stroke and one coronary event occurred during paricalcitol therapy. No patients died during the study. Interpretation In patients with macroalbuminuria and type 2 diabetes, moderate salt restriction enhances the antialbuminuric effect of losartan, an effect that could be nephroprotective and cardioprotective in the long term. The finding that paricalcitol prevents a sodium-induced increase in albuminuria provides support for trials to test the long-term risk-benefit profile of paricalcitol add-on therapy in patients with type 2 diabetes and macroalbuminuria refractory to dietary salt restriction, including patients refractory to even moderate salt restriction. Funding AbbVie.