Controlling BP in hypertensive kidney transplant patients is mandatory to prevent cardiovascular complications and graft failure . In most patients, several drugs are needed to reach a BP lower than 140/90 mmHg (or 130/80). In such particular patients with a single functioning kidney, BP control can be difficult to reach because of a restricting therapy including immunosuppressive and corticosteroid medications. Our purpose was to analyze the daily practice control of BP by two methods (OBP and HBP) and the factors which could influence such control, especially looking for the potential contribution of non pharmacological intervention such to restore a well-balanced sodium/potassium ratio intakes in treated by antihypertenvive drugs Kt patients, as it was described in non transplanted hypertensive population [19, 20].
Indeed, in Kt patients with a stable kidney function, taking into account several determinants of BP, even though controlled and uncontrolled patients seemed to consume both a similar amount of salt, the controlled patients had significantly higher intakes of potassium, this could be associated with their lower Na/K ratio compared to uncontrolled patients. On one hand, when controlled for age, BMI, antirejection drugs, smoking habit and Na + excretion, an inverse and significant relation was found between urinary potassium and HBP. On the other hand a positive but non significant relationship was observed between urinary sodium and HBP when controlled for K + excretion. As for many patients, out-of-clinic BP are recommended for Kt patients [21, 22]. In the present study, we favoured the use of two complementary BP measurement techniques i.e. office and home and defined controlled or uncontrolled BP when there was an agreement between both techniques. That classification of BP reinforced their truly status of normotensive or hypertensive treated patients taking into account the well known limitations of office BP and allowed the detection of masked and white coat hypertension .
The frequency of Kt patients remaining hypertensive despite a treatment requiring on average 2 or more drugs is too high, 58.6% based on OBP only but it decreases to 49% when based on both OBP and HBP (after excluding in the calculation of the proportion other forms such white-coat and masked hypertension).
Interpreting BP level in Kt patients is complex since many factors are susceptible to modify BP regulation.
Age, and BP status of donor, presence of native kidneys in recipients or number and kind of antihypertensive drugs did not differ between uncontrolled and controlled patients.
Our controlled patients are mainly treated with tacrolimus (80%) (Table 2) compared with uncontrolled ones (50%) and this could partly explain the higher BP of the uncontrolled group as cyclosporine has been considered as a more hypertensive drug . However in our study, on average, the number of antihypertensive drugs did not differ between the two BP control groups, moreover a comparison of BP (office and home) and of the number of antihypertensive drugs between patients treated by either cyclosporine or tacrolimus did not show any statistical difference. On average, our patients were transplanted for 5 (controlled) and 6 years (uncontrolled), it is known that incidence of hypertension increases with increasing time after transplantation, this is partly due to additional risk factors but also to the deleterious effects on the cardiovascular system caused by the sum of pre and posttransplant risk factors, so at this stage, it is difficult and even hazardous to attribute the uncontrolled BP to the only higher proportion of patients treated with cyclosporine rather than tacrolimus. Even if patients treated with tacrolimus were transplanted more recently than those treated with cyclosporin (P = 0.03), there were no significant differences according to age, GFR and time duration spent on hemodialysis.
Steroid treatment has been said to contribute to 15% in hypertension . The frequencies of patients still treated with steroid were similar in cyclosporine and tacrolimus treated patients. Adherence to medications is also a key point in reaching BP target, all the patients questioned on their regularity in taking medications were particularly convinced by the importance of immunosuppressive and antihypertensive treatments regarding the graft protection.
Urinary and dietary sodium and potassium
Urinary Na + excretion is considered as representative of the dietary intakes in stable conditions. On average, in our study, controlled patients did not excrete less Na + than uncontrolled ones. The mean salt consumption was however between 9 and 10 g per day, with 87% who had salt consumption higher than 5 g, so as in many countries, mean values exceeded largely the physiological needs  and this in spite of the fact that there are hypertensive treated by antihypertensive drugs.
Calculated sodium intakes were lower than urinary excretion. The difference can be explained by the lesser precision of the 24-h dietary recall intakes declared by patients recording their diet intake while they were collecting urine. Indeed, the real intakes were most certainly underestimated because the exact amount of food eaten by the patients was very difficult to know and moreover we could not assess the salt added in cooking and at table. However the patients were asked if they systematically added salt when cooking and/or at table, so 8 (53%) controlled patients admitted to add salt against 22 (65%) uncontrolled patients. Next the added salt, bread, high salted cheese and cooked meats and processed food were major diet sources of salt. Interestingly, the intakes and electrolytes excretions were reproducible on two occasions, 7 days apart. As sodium intake was less accurately quantified by dietary recall, a 24 h urine collection measuring sodium excretion would be better, after explaining the patient how to proceed.
On the contrary, the agreement between diet and urinary potassium seemed clearly better. This was documented in the study of Tasevska et al. who found a high correlation (r = 0.89) between dietary potassium and its urinary excretion in free living individuals .
Although all patients had potassium intakes lower that the 4700 mg recommended , patients with controlled BP had on average higher potassium intakes with a diet richer in fruits and vegetables and contrary to what is generally said i.e. the high sodium consumers consume less potassium, in our population, there was no inverse relation between sodium and potassium intakes. The low consumption of potassium for several patients was not necessarily an individual's preference for salted food but was probably an habit they have kept when they were treated by hemodialysis, a situation in which strict limitations of potassium intakes are mandatory.
Randomized large-scale control trials on the effect of a modified sodium and potassium diet on blood pressure in late kidney transplant patients are lacking, however dietary sodium restriction (80–100 mmol/day or 1.8 to 2.3 g/day) is recommended to contribute to the cardiovascular burden management and also to limit the cyclosporin-induced hypertension caused by sodium retention in early post-transplantation .
Although the legendary salt controversy is still a hot topic [30–34], numerous studies favored a beneficial effect on BP, even sometimes modest, with reduction of salt intake and increasing potassium intakes. A drop of 3.5/2.5 mmHG with a 1.7 g increase of potassium intake and a drop of 5.2/3.7 mmHg with a decrease of 4.5 g of NaCl could be reached, however effects were larger in hypertensive than normotensive individuals . Potassium supplementation was associated with a small but significant reduction in mean systolic and diastolic blood pressure ,-3.11 mm Hg and −1.97 mm Hg respectively. Effects of treatment appeared to be enhanced in studies in which participants were concurrently exposed to a high intake of sodium. The beneficial effect of K on BP could be explained by a vasodilation action due to hyperpolarization of endothelial and vascular smooth muscle cell of the arterial wall . Moreover, high potassium intake can decrease the salt sensitivity and also decrease the need for antihypertensive drugs .
Some observational studies did not find a correlation between salt intakes, estimated by 24-h urinary Na+, and prevalence of hypertension or BP in Kt recipients with stable allograft function [38, 39]. In these studies, measured office BP or hypertension defined by the prescription of antihypertensive medication, and the number of drugs were considered as a surrogate marker for severity of hypertension and no dietary analysis of sodium and potassium intakes was performed. In agreement with Prasad , we did not find any statistical difference of BP between patients treated with cyclosporine or tacrolimus and no difference in patients treated with or without a thiazide diuretic. However some methodological differences could explain our conclusions, indeed we defined the BP control on the basis of two measurement techniques and 24-h Na + and K + measurements were associated to dietary intakes assessments. We observed a positive but non significant relation between urinary sodium and systolic home BP when confounding factors such as age, BMI, antirejection drugs, smoking habit and potassium excretion were controlled. We also observed a significantly negative correlation between systolic home BP and potassium excretion when age, BMI, antirejection drugs, smoking habit and sodium excretion were controlled. Therefore our results plead in favor of a possible efficient non pharmacological intervention by restricting sodium intakes and increasing potassium ones in Kt recipients. This was documented for sodium in the study of Keven et al. who measured a decrease in BP with daily intakes limited to the recommended levels between 80 and 100 mmol . In a very recent paper , van den Berg et al. found that sodium intake was positively and significantly associated with systolic and diastolic office BP in renal transplant recipients compared to healthy controls. Authors concluded that decreasing sodium intake to recommended amounts could decrease systolic BP by 4–5 mmHg. Conversely it was noted in CKD patients that high salt intake can increase proteinuria and ESRD risk . In our study, proteinuria was not correlated with urinary Na excretion, but we have not tested the specific effect of a modification of such sodium intake on the proteinuria. This should be done to reinforce the argument for lowering sodium intake especially in hypertensive Kt patients who are proteinuric and exposed to higher cardiovascular and kidney disease progression risks.
Our uncontrolled patients had a higher Na+/K + ratio than controlled ones. The risk of subsequent cardiovascular disease related to a high Na+/K + ratio has been identified in adults with prehypertension, the effect of the ratio was even stronger than that of sodium or potassium alone . It has to be confirmed in our Kt population with a prospective study.
Our study limitations concern the small size of the sample notably due to our selection criteria to define controlled and uncontrolled patients based on two measurement techniques in late Kt recipients. Cases of masked and white coat hypertension are not discussed in the present paper and have to be studied apart since interestingly, in the context of our study, we did not find a significant relation between sodium, potassium excretion and BP in those particular forms of hypertension (data not shown). Either confounding factors or other determinants of BP level in these patients have to be identified. Our study shows that measurements of Na + and K + intakes with a diet recall fulfilled by patients in a free way leads to limitations when data are compared with urinary excretion. We think that the main limitations came from patients who did not mentioned their complete intakes while they were collecting 24 h urines. Another problem came from the estimation of real amounts of food intakes which were not always accurately described by several patients, so we had to use standardized portions defined by “Poids et mesures- quantification standardisée des denrées alimentaires (Conseil Supérieur d'Hygiène, janvier 2005). The discrepancy between intakes and urinary Na + is higher than with K, it can be explained by the fact that added salt was impossible to measure and that the amount of sodium in some processed foods was not mentionned, this was coupled with possible omissions from the patients in their recall diet document. The better correlation between intakes and urinary excretion of K + has been described in literature. Cooking methods are more conservative for potassium and added potassium sources are much more limited than those for sodium. It is possible that patients have been more reliable when they noticed their food consumption containing potassium rather than the one containing sodium. Obviously, those patients have kept habits from the time they were treated by dialysis when potassium intakes were strictly limited that is probably why they do not consume fruits and vegetables enough. We recognize that these considerations bear witness of a limited use of diet recall to calculate amounts of sodium and potassium. This has been mentionned in literature, so patients tend to underestimate their Na + intake by 30 to 50% while estimated K + intake better correlated with 24-h urinary excretion . These observations are in favour of the use of 24 h urinary excretion provided that renal function and medications have to be taking in account.
Our study is crossectional so it gives a snapshot of a situation influenced by the complex and atypical risk factors profile of Kt patients mixing traditional, predisposing and associated to transplantation risk factors. Moreover, the factors which contribute to elevated BP vary between the early and late posttransplant periods, conditioning a continuous increase of the cardiovascular risk throughout time. Our patients have a mean transplantation duration of 7 years (± 6.6), so in that late posttransplant period, next to the allograft function supervision, the management of the cardiovascular risk factors, among them hypertension, can beneficiate of the acquired experience with notably non pharmacological interventions in general population.