A best practice position statement on pregnancy in chronic kidney disease: the Italian Study Group on Kidney and Pregnancy

Pregnancy is increasingly undertaken in patients with chronic kidney disease (CKD) and, conversely, CKD is increasingly diagnosed in pregnancy: up to 3 % of pregnancies are estimated to be complicated by CKD. The heterogeneity of CKD (accounting for stage, hypertension and proteinuria) and the rarity of several kidney diseases make risk assessment difficult and therapeutic strategies are often based upon scattered experiences and small series. In this setting, the aim of this position statement of the Kidney and Pregnancy Study Group of the Italian Society of Nephrology is to review the literature, and discuss the experience in the clinical management of CKD in pregnancy. CKD is associated with an increased risk for adverse pregnancy-related outcomes since its early stage, also in the absence of hypertension and proteinuria, thus supporting the need for a multidisciplinary follow-up in all CKD patients. CKD stage, hypertension and proteinuria are interrelated, but they are also independent risk factors for adverse pregnancy-related outcomes. Among the different kidney diseases, patients with glomerulonephritis and immunologic diseases are at higher risk of developing or increasing proteinuria and hypertension, a picture often difficult to differentiate from preeclampsia. The risk is higher in active immunologic diseases, and in those cases that are detected or flare up during pregnancy. Referral to tertiary care centres for multidisciplinary follow-up and tailored approaches are warranted. The risk of maternal death is, almost exclusively, reported in systemic lupus erythematosus and vasculitis, which share with diabetic nephropathy an increased risk for perinatal death of the babies. Conversely, patients with kidney malformation, autosomal-dominant polycystic kidney disease, stone disease, and previous upper urinary tract infections are at higher risk for urinary tract infections, in turn associated with prematurity. No risk for malformations other than those related to familiar urinary tract malformations is reported in CKD patients, with the possible exception of diabetic nephropathy. Risks of worsening of the renal function are differently reported, but are higher in advanced CKD. Strict follow-up is needed, also to identify the best balance between maternal and foetal risks. The need for further multicentre studies is underlined.

ences and small series. In this setting, the aim of this position statement of the Kidney and Pregnancy Study Group of the Italian Society of Nephrology is to review the literature, and discuss the experience in the clinical management of CKD in pregnancy. CKD is associated with an increased risk for adverse pregnancy-related outcomes since its early stage, also in the absence of hypertension and proteinuria, thus supporting the need for a multidisciplinary follow-up in all CKD patients. CKD stage, hypertension and proteinuria are interrelated, but they are also independent risk factors for adverse pregnancy-related outcomes. Among the different kidney diseases, patients with glomerulonephritis and immunologic diseases are at higher risk of developing or increasing proteinuria and hypertension, a picture often difficult to differentiate from preeclampsia. The risk is higher in active immunologic diseases, and in those cases that are detected or flare up during pregnancy. Referral to tertiary care centres for multidisciplinary follow-up and tailored approaches are warranted. The risk of maternal death is, almost exclusively, reported in systemic lupus erythematosus and vasculitis, which share with diabetic nephropathy an increased risk for perinatal death of the babies. Conversely, patients with kidney malformation, autosomal-dominant polycystic kidney disease, stone disease, and previous upper urinary tract infections are at higher risk for urinary tract infections, in turn associated with prematurity. No risk for malformations other than those related to familiar urinary tract malformations is reported in CKD patients, with the possible exception of diabetic nephropathy. Risks of worsening of the renal function are differently reported, but are higher in advanced CKD. Strict follow-up is needed, also to identify the best balance between maternal and foetal risks. The need for further multicentre studies is underlined.

Introduction
In 1975, an editorial in the Lancet, entitled Pregnancy and Renal Disease, started with the following sentence: Children of women with renal disease used to be born dangerously or not at all-not at all if their doctors had their way [1]. The paper reviewed the previous grim evidence, with minimal chances of a positive outcome for children whose mothers started pregnancy with blood urea nitrogen (BUN) levels higher than 60 mg/dl and with hypertension. In spite of several negative reports, however, the authors gave space to new more positive data and concluded discussing the advice to be given to a woman with chronic kidney disease (CKD) who wants to undertake a pregnancy. The conclusions are still valid in our era of patient empowerment: the woman should be told that there is a considerable risk to her infant and a small risk to herself, but dogmatic prohibitions do not seem justified today. Instead, obstetrician and physician must batter down the hatches and prepare to ride out the storm together with those determined to set sail [1]. Many of the issues raised in 1975, notably the role of hypertension and proteinuria, the severity of CKD and the presence of specific diseases such as systemic lupus erythematosus (SLE) or autosomal dominant polycystic kidney disease (ADPKD), are still a matter of discussion. However, the context has changed throughout the decades, for several reasons including advances in maternalfoetal care, with a progressive expansion of the foetal viability zone to 22-24 weeks [2][3][4]. The progressive empowerment of patients with chronic diseases has shifted the attitude from a paternalistic protection of the mother to a shared choice. Meanwhile, the almost unexpected results of pregnancy on dialysis have highlighted the potentials of pregnancy even in the latest stages of CKD [5][6][7][8][9][10][11][12][13].
In this context, the Kidney and Pregnancy Study Group of the Italian Society of Nephrology undertook the present best practice review, aimed at combining the available evidence with an in-depth discussion on shared experiences and open questions.

Evidence-based medicine and pregnancy in CKD: methodological insights
The evidence on pregnancy in CKD shares several methodological problems with pregnancy in dialysis and, more generally, with pregnancy in rare and/or heterogeneous diseases [13,[19][20][21].
CKD is an etiquette gathering many different diseases that may differently affect pregnancy. Immunologic nephropathies are at risk of flares during and after pregnancy. Pyelonephritis and renal malformations share an increased risk of urinary tract infections. ADPKD presents complex ethical problems as regards prenatal counselling. Diabetic nephropathy is the only kidney disease associated with a higher risk for non-renal malformations [19][20][21][22][23][24][25][26][27][28][29][30]. The lack of shared definitions partially impairs the pooling of the data for systematic reviews, an issue that has to be considered during a critical analysis of the results [19-21, 28, 30].
Kidney disease, degree of kidney impairment, hypertension and proteinuria are well known factors in the pathogenesis of pregnancy-related adverse outcomes. Their interactions are complex, and so far not completely elucidated [31][32][33].
A major issue is kidney function assessment in pregnancy, a situation in which no validated formula exists and the physiologic hyperfiltration may alter CKD staging [34][35][36][37][38].
It is difficult to perform randomised controlled trials (RCTs) in pregnancy, and often these are ethically unfeasible [39][40][41]. Hence, our understanding on pregnancy outcomes in CKD patients mainly relies on observational studies, whose quality is not necessarily inferior as compared to RCTs [42][43][44][45][46][47][48]. As a consequence of the ''physiological'' lack of RCTs and of the paucity of large observational studies, we will deal with low levels of recommendation. We will include several not graded suggestions, whose contribution should however not be underestimated [13].
The research strategy on which the present position statement is based represents an updating, at June 2014, of a previous systematic review on CKD and pregnancy, to which we refer the reader for details on the search strategy and on the paper selection modality [20]. The heterogeneity of the literature and the high number of kidney diseases led us to focus on general features and on the main forms of CKD, highlighting the needs for future research. The Italian Study Group on Kidney and Pregnancy has already covered the issue of pregnancy on dialysis [13]. A further position statement will treat pregnancy in kidneytransplanted patients.
Which CKD patient should be considered at ''increased risk'' for adverse pregnancy-related outcomes?
1. All CKD patients should be considered at increased risk for adverse pregnancy-related outcomes, and followed-up accordingly (strong suggestion; evidence form relatively large reports from few settings). 2. The risks of adverse pregnancy-related outcomes increase in incidence and severity along with the worsening of CKD. Follow-up should be intensified in patients with severe CKD (strong suggestion; evidence from several reports in different settings, CKD stages and/or in different diseases). 3. Proteinuria, hypertension, type 1 diabetes and immunologic diseases are associated with increased risk. Follow-up should be intensified (strong suggestion; evidence from several reports in different settings).
The effect of CKD, even with normal renal function, is confirmed by the significant increase in adverse maternalfoetal outcomes, including preeclampsia (PE), pregnancy induced hypertension and pre-term delivery, in three large cohorts of patients who undertook pregnancy after kidney donation [57, 58, 67]. The reasons why women with single kidney (or kidney scars, glomerulonephritis in remission, etc.) with normal renal function, no proteinuria or hypertension should share a higher risk for adverse pregnancyrelated outcomes are not elucidated [32,38,57,58,67].
In such a context, the working group suggests that all patients with CKD should be considered at increased risk for pregnancy-related adverse events and should be followed, whenever possible, by a multidisciplinary team including nephrology and obstetrics staff. Since the risks increase along with the worsening of the renal function, follow-up should be intensified in severe CKD (see further recommendations). For the management of dialysis patients, please refer to the specific position statement [13].
These conclusions are mostly based upon scattered evidence and underline the need for further studies on large cohorts of CKD patients. Such studies are highly encouraged by the study group.
What are the main adverse pregnancy-related outcomes in CKD patients?
1. The risks of maternal death are low, and hardly quantified; they are mainly described in active or flared immunologic diseases, above all in lupus erythematosus (strong suggestion; scattered evidence). 2. The risks of worsening of the kidney function, of developing hypertension and of developing or increasing proteinuria increase along with CKD stage (moderate suggestion; scattered evidence). 3. The foetal risks are mainly related to prematurity (strong suggestion; scattered evidence). 4. Perinatal death may be increased in specific diseases (SLE, immunologic systemic diseases, and diabetic nephropathy) (strong suggestion; scattered evidence). 5. Malformations are not increased in CKD patients, with the possible exception of diabetic nephropathy (strong suggestion, large body of evidence from different sources on non diabetic CKD; one paper only reporting on increase in malformations in diabetic nephropathy). 6. Risks for adverse materno-foetal outcomes may be increased in multiple pregnancies (moderate suggestion, one large series only).
Pregnancy-related outcomes in CKD may be related to the mother and to the foetus-child. The maternal risks are death, and progression of CKD, including also increase in proteinuria, impaired control or onset of hypertension, eventually persisting after delivery. The risks for the foetus-newborn are perinatal death, malformations, pretermbirth, intellectual deficiencies, diseases in adulthood [18][19][20]. Maternal death is reported as a very low or non quantifiable risk, at least in Western Countries, where maternal mortality is already low at baseline [18][19][20]66]. According to most large series and systematic reviews, maternal deaths are almost exclusively reported in patients with systemic immunologic diseases, first of all SLE, also on account of the fact that SLE represents the most common systemic immunologic disease in pregnancy [18-20, 23, 68-70].
The risk for worsening of kidney function may increase along CKD stages; its entity has been differently calculated, from about 20 to over 80 %; however, a recent systematic review did not find a higher risk of kidney function reduction in CKD patients with versus without pregnancy in stages 1-3 [16-21, 25, 49-53, 60, 62, 71].
An increase in proteinuria is common; it is more frequent in patients already proteinuric at start of pregnancy and may be higher in diabetic nephropathy. However, the definition of a ''relevant increase'' in proteinuria is quite elusive, and the reported prevalence ranges from 20 to almost 100 % of the cases [25-29, 72-74].
Children from mothers affected by diabetes and systemic-immunologic diseases have a higher risk of perinatal death, also when born at-term or in the late preterm phase. This risk seems more related to the specific disease than to the degree of renal functional impairment [22,23,28,69,70]. There is no evidence that children from non diabetic CKD mothers have a higher risk of malformations, beside urinary tract malformations, including vescico-ureteral reflux, and related conditions, that are known to have a genetic background, albeit so far partially understood.
One recent study, on a large series, conversely suggests that the presence of nephropathy increases the risk for malformations in children of diabetic mothers. Several different malformations have been reported, with an odds ratio of 2-3 as compared with mothers without nephropathy [91].
According to one case series and a few case reports, multiple pregnancies may have higher risks of adverse pregnancy related outcomes [92][93][94][95]. This should be kept in mind in particular in the case of assisted reproduction techniques.
Very few studies have been specifically addressed at the health status of children from CKD mothers; they analysed different aspects of health, from kidney function to psychological well-being. Within the limits of the scattered evidence, it appears that most of the children from CKD mothers are able to attain normal developmental goals, at any level of kidney function impairment [11,12,[96][97][98].
According to the Study Group, there is need for further studies, focusing on the clinical and psychological longterm outcome of children from CKD mothers.
Indications and goals for follow-up 1. The main goals of follow-up are early identification and treatment of complications, including hypertension, anaemia, coagulation disorders, and timely planning of delivery (strong recommendation, indirect evidence). 2. Follow-up should be intensified in CKD patients with respect to normal pregnancies (strong recommendation, indirect evidence). 3. Follow-up should be increased along with the increase in CKD stages, and according to the presence of other risk factors, such as hypertension, proteinuria and systemic disease (strong recommendation, indirect evidence). 4. Follow-up should include at least one nephrological visit with blood and urinary tests every 4-6 weeks in non proteinuric, non hypertensive stage 1 CKD pregnancies and should be increased to weekly in intensely proteinuric, hypertensive patients or in CKD stages 4-5 (strong recommendation, indirect evidence).

No validated formula for glomerular filtration rate
(GFR) calculation in CKD pregnancy is available, hence GFR should be assessed by 24-h urine collection (strong recommendation, based on GFR in normal pregnancy and PE). 6. 24-h protein excretion is preferable for quantification of proteinuria (strong recommendation, indirect evidence). 7. Twice monthly to weekly urinalysis and urinary cultures may be indicated in the prevention of urinary tract infection (UTI) recurrences and for early detection of proteinuria (strong recommendation, evidence limited to one RCT).
Pregnancy is not a disease nor a disease complication; however, it may be complicated by the presence of maternal diseases. Hence, there is no specific ''treatment'' for pregnancy in CKD, and conversely several therapies that are used in CKD outside of pregnancy are contraindicated in pregnancy (Tables 1, 2 The assessment of kidney function in pregnancy is challenging, due to hyperfiltration, and increase in plasma volume and in volume distribution. To date, no formula is validated for GFR calculation in CKD pregnancy, and it is preferable to assess creatinine clearance on 24-h urine collection, at least in patients trained to follow the indications. This also allows a precise assessment of proteinuria [33-38]. The Study Group suggests a minimum requirement of one nephrology visit with blood and urinary tests every 4-6 weeks in non proteinuric, non hypertensive stage 1 CKD pregnancies, increased up to weekly in patients with various combinations of proteinuria, hypertension or CKD stages 4-5. Patients at risk of recurrent UTIs, of immunologic flares or with glomerulonephritis in remission should control urinalysis for proteinuria and/or urinary cultures every 1-2 weeks, to promptly identify UTIs and/or proteinuria (  4. Kidney biopsy is not recommended in pregnancy, either for the definition of the kidney disease or for the differential diagnosis between PE and CKD, because of risk of severe complications (strong recommendation, systematic review of scattered evidence).
The broad definition of PE includes both mildly affected subjects, whose situation is compatible with term delivery of ''appropriate for gestational age'' babies, as well as patients affected by stormy, devastating and even deadly disease, with ''small for gestational age'' (SGA) babies and risk of long-term complications [99][100][101][102][103][104][105][106]. Several attempts have been made to sub-classify PE, according to severity, gestational age, initiating events and pathophysiological origin; furthermore, PE may be included in a continuum of diseases, from pregnancy-induced hypertension to HELLP syndrome (named for its three features: haemolysis, elevated liver enzymes, and low platelet levels) [107][108][109][110][111][112].
The classic definition of PE states that it is a reversible condition characterized by the association of hypertension and proteinuria (above 300 mg/day), occurring after the 20th week of pregnancy in previously normotensive, non proteinuric pregnant women [101]. Proteinuria and hypertension should disappear within 1-3 months after delivery. Some authors, however, recently suggested following PE patients for at least 6 months after delivery, to ''change public (and clinician) thinking that pregnancy lasts, instead of 9 months, at least 15 months''. This idea is a great way of getting us all to consider the importance of post-partum care particularly for women with PE. In this way we can detect ongoing hypertension or renal issues [102]. The relationship with kidney diseases has been underlined by several authors, and is crucial for the differential diagnosis of PE [113][114][115][116].
A paradigm shift occurred when the American College of Obstetricians and Gynaecologists (ACOG) guidelines stated that PE may be diagnosed in the absence of proteinuria, basing diagnosis on new-onset hypertension accompanied by at least one of the following: serum creatinine increase, low platelet count or high liver enzymes, pulmonary edema or central nervous symptoms, once more Usually considered as relatively safe, when absolutely needed [341][342][343][344][345][346][347][348][349] Azathioprine This is the most widely used immunosuppressive drug. It is teratogenic in animal models, but not in humans, possibly because the foetal liver is not able to activate the drug. K-DIGO and European Best Practice Guidelines suggest switching from Mycophenolate to Azathioprine before pregnancy [341][342][343] D Cyclosporine A This Calcineurin inhibitor has not been associated with increased teratogenicity; however, small for gestational age babies and preterm delivery have been reported, possibly due to the maternal disease and not specifically to the drug; levels may vary in pregnancy and the hypertensive, hyperglycaemic and nephrotoxic effects should be mentioned [344] C

Tacrolimus
The drug has similar effects and side effects as Cyclosporine A; since it is a relatively new drug, experience is more limited than with the previous drug [345] C Steroids Together with azathioprine these are the most often employed and best known drugs. The most frequently used shortacting corticosteroids include prednisone, methylprednisolone and prednisolone, while betamethasone and dexamethasone are among the long-acting drugs. No major malformations have been reported, and the issue of labio-palatoschisis is debated. A higher risk of premature rupture of membranes has been reported. Other relevant side effects include infectious risk, and the increased risk of gestational diabetes [346] C

Hydroxychloroquine
This synthetic anti-malaric agent crosses the placenta but was not found to be associated with foetal toxicity [217][218][219] B To be avoided [341][342][343][344][345][346][347][348][349] Cyclophosphamide This alkylating agent is contraindicated in pregnancy; a few reports suggest that pregnancy termination is common in the case of inadvertent use or need for life saving therapy. A few positive reports, mainly in women with SLE are also available [68] D

Mycophenolate
Severe foetal malformations are reported, mainly involving cardiovascular and cranial malformations. Discontinuation for at least 6 months, to stabilize kidney function, is usually indicted after kidney transplantation [347,348] D

Rituximab
There are no data on whether rituximab can cause foetal harm. Rituximab was detected postnatally in the serum of infants exposed in utero: B-cell lymphocytopenia generally lasting less than 6 months can occur in infants. The manufacturer recommends contraception for up to 12 months following therapy [369,370] C m-Tor inhibitors Very few studies have considered their use in pregnancy. They are teratogenic in animals and discontinuation in humans is a matter of debate; KDIGO guidelines suggest discontinuation in anticipation of pregnancy [347,349] C FDA site of the Food and Drug Administration [340]; FDA rating: A, controlled human studies show no risk; B, no evidence of risk in studies; C, risk cannot be ruled out; D, positive evidence of risk; X, contraindicated in pregnancy underlining the central importance of kidney impairment in this context [117]. Independently from the definition, there are at least three conditions in which the differential diagnosis between PE and CKD is almost impossible: when no data before 20 gestational weeks is available; when CKD flares in pregnancy; when kidney disease develops in pregnancy [116,118]. According to the widespread pathophysiological view, a placentation defect is at the basis of PE; thus, utero-placental flow analysis may help distinguish PE (impaired flows, and stunted growth) from CKD (normal blood flows and growth) [118][119][120][121][122]. Several clinical and biochemical biomarkers of PE have been suggested and could support the differential diagnosis between PE and CKD [123][124][125][126][127][128][129][130][131]. The ratio between sFlt-1, a receptor for both vascular endothelial growth factor (VEGF), and PIGF is considered as one of the most promising predictors of PE that may support the differential diagnosis between PE and CKD which, at least in the early stages, is characterized by a normal placentation [116,118,[132][133][134][135][136][137][138][139].
Kidney biopsy is not recommended in pregnancy, according to a systematic review that highlights risks of severe bleeding complications. However, in early pregnancy, the pros and cons of diagnosis versus empiric therapy should be individually assessed, in particular in cases with progressive kidney function impairment. In late pregnancy, the advantages of an invasive manoeuver versus pre-term delivery should be considered [140]. A kidney biopsy should be performed in patients with persisting proteinuria [3-6 months after delivery (with the timing depending also on other renal parameters) without an established diagnosis of kidney disease. These patients are unfortunately often lost to follow-up, thus limiting the potential advantages of pregnancy as a first occasion for the diagnosis of potentially reversible CKD [32].
The Study Group therefore supports the need for systematic assessment of kidney function following PE, as well as in future pregnancies after PE or HELLP syndrome. While acknowledging the interest in serum biomarkers for the differential diagnosis between PE and CKD, the Study Group underlines that it is based on the results of few studies, mostly from the same groups, and envisages the organization of large multicentre studies on this issue.
Specific issues: hypertension as a complication of CKD and pregnancy 1. Arterial hypertension is a common complication of pregnancy in CKD, and should be early identified and treated (strong recommendation, indirect evidence from large observational studies). 2. ''Overtreatment'' should be avoided, due to the risk of reducing utero-placental flows, and inducing intrauterine growth restriction (strong recommendation, evidence from large observational studies). 3. Treatment should follow a stepwise approach, starting from the drugs with less contraindications in pregnancy (strong recommendation, based upon systematic reviews and guidelines). 4. When used for the antiproteinuric and nephronprotective effects, angiotensin-converting enzyme inhibitors (ACEi) or angiotensin receptor blockers (ARBs) should be discontinued at the first assessment of a positive pregnancy test (strong recommendation for discontinuation, contrasting evidence for the timing).
Target blood pressure values are periodically reviewed in the hypertensive disorders of pregnancy, a setting in which a shared recommendation is to avoid overcorrection of hypertension, for the risk of affecting foetal growth [141][142][143][144][145][146][147][148].
Conversely, blood pressure targets in hypertensive CKD pregnancies are less established. It is presumably wise to avoid overcorrection, but it should also be remembered that hyperfiltration and proteinuria, that are widely described in hypertensive CKD patients, may have a detrimental effect on the kidney function [21].
Considering the recent Control of Hypertension In Pregnancy Study (CHIPS) trial, that revealed a substantial equivalence between ''thigh and standard'' blood pressure control in pregnancy, our group suggests implementing a strict blood pressure control ('ideal' target \130/ 80 mmHg, acceptable \140/90 mmHg), under careful clinical surveillance, at least in patients with good compliance that control blood pressure at home [149][150][151][152].
The main list of anti-hypertensive drugs is reported in Table 1. None is considered as fully safe. The issue of Low doses during pregnancy needed for the treatment of diverse medical conditions have not been shown to cause foetal harm; may be protective against pre-eclampsia, favoring placentation (see text); discontinuation before delivery is recommended [161-163, 223, 224] NC LMWH Low molecular-weight heparin (LMWH) does not cross the placenta and is safe for the foetus, although bleeding at the utero-placental junction cannot be rued out. Individualized doses of LMWH are well tolerated and safe for prophylaxis and treatment of thromboembolic complications during pregnancy, and post-partum [355]. Twice-daily heparin should be discontinued prior to induction of labour or planned cesarean delivery and can be resumed after delivery [355,356] C

ESAs
In vitro studies suggest that recombinant erythropoietin does not cross the human placenta; higher doses may be needed in dialysis patients [357,358] C Allopurinol Adverse events were observed in animal studies. Allopurinol crosses the placenta [359]. An increased risk of malformations events has not been observed in humans (limited data) [360] C Vitamin D The role for vitamin D supplementation in pregnancy is controversial, but there is no evidence of a reduction in adverse pregnancy outcomes (e.g., preeclampsia, stillbirth, neonatal death) [361] or improvement in bone mineral content in children in studies in which supplementation was given independently from blood levels [362]. Cholecalciferol (vitamin D3) crosses the placenta but the transfer to the foetus from the mother is low. Maternal supplementation has not been shown to affect pregnancy outcomes [363]. Ergocalciferol Adverse events were observed in animal studies. The ergocalciferol (vitamin D 2 ) metabolite, 25(OH) D, crosses the placenta; maternal serum concentrations correlate with foetal concentrations at birth. Calcitriol Teratogenic effects have been observed in animal studies. Adverse effects on foetal development were not observed in women (N = 9) with pseudovitamin D-dependent rickets. Paricalcitol Adverse events were observed in some animal reproduction studies Vitamin D deficiency in a pregnant woman may lead to deficiency in the neonate [364,365]. Serum 25(OH) D concentrations should be measured in pregnant women at increased risk of deficiency [366]. Current guidelines recommend an intake of 1000-2000 units/day until more safety data is available [366,367]. Vitamin D and calcium levels should be monitored and kept in the lower normal range The Study Group strongly supports early discontinuation of ACEi and ARBs at the first positive pregnancy test (4th-6th gestational week) in proteinuric CKD patients, to maximize the benefits of control of proteinuria, at least in compliant patients [153][154][155][156][157][158].
Specific issues: proteinuria and low-protein diets

3.
A moderate protein restriction may be of help in counterbalancing the hyperfiltration of pregnancy and may be safely employed in proteinuric patients and in patients with advanced CKD (medium recommendation, evidence from one group only).
The use of low-dose acetylsalicylate for the prevention of PE is suggested by a large body of evidence, in the cases at risk for PE, a category encompassing CKD patients (Table 4) [159,160]. The timing is controversial. Most guidelines suggest starting therapy after the first trimester, because of the risk of haemorrhage in the case of early foetal loss. From a pathophysiological point of view, however, the effects on placentation should be maximal when aspirin is started early; therefore other authors suggest that an earlier start may be effective in preventing severe and early PE [161][162][163].
In the absence of clear evidence upon the best timing, the Study Group suggests to identify in each setting a shared policy between Obstetricians and Nephrologists,  taking also into account that our patients are more closely followed than the general population, thus potentially minimizing the risks of unexpected haemorrhage.
No study clearly identifies any benefit of albumin infusion in CKD and in PE, even if this empirical therapy is still diffused, with the aim of increasing albumin levels and reducing the risk of placental hypoperfusion. Therefore, the Study Group does not support the systematic use of albumin infusion in pregnant patients with nephrotic syndrome or PE, considering the potential risk of adverse reactions, and the risk of paradoxically increasing proteinuria by enhancing hyperfiltration [164][165][166][167].
In the absence of specific anti-proteinuric therapy in pregnancy, two non-randomised Italian reports suggest that a moderate protein restriction (0.6-0.8 g/kg/day), with a vegan-vegetarian diet supplemented with aminoacids and ketoacids, is feasible and safe in pregnant patients with CKD. Interestingly, the prevalence of small for gestational age babies was lower in on-diet patients as compared to patients on an unrestricted diet [168,169]. The data obtained in a relatively small group of patients are in line with reports on the advantages of vegetal proteins in reducing kidney disease progression in non-pregnant humans, with a systematic review on vegan-vegetarian diets in pregnancy, and with the position of the American Dietary Association, provided that B12, iron and vitamin D are controlled and supplied upon need [170][171][172][173][174][175][176][177][178].
The Study Group suggests considering the diet option, at least in cases in which hyperfiltration (absolute or relative to the remnant nephrons) is supposed to play an important role, and strongly suggests organising multicentre studies on this issue.
Specific diseases: glomerulonephritis (primary) 1. The appearance of proteinuria during the third trimester of pregnancy calls for a differential diagnosis between glomerular disease and PE (strong recommendation, scattered evidence). 2. During pregnancy, glomerulonephritis may appear for the first time or be a relapse or worsening of a previously identified disease: the differential diagnosis is important for management (strong recommendation, evidence from large studies).

Proteinuria and nephrotic syndrome in pregnancy may
suggest membranous glomerulonephropathy (MGN), minimal change nephropathy (MCN), or focal segmental glomerulosclerosis (FSGS) (strong recommendation, evidence from observational studies). 4. In patients with IgAN, the most common glomerulonephritis in the Mediterranean countries, pregnancy is not associated with a faster disease progression. Risk of disease progression after delivery is higher with GFR \70 ml/min/1.73 m 2 , poorly controlled hypertension, and proteinuria (strong recommendation, evidence from observational studies). 5. Immunosuppressive therapy for primary glomerulonephritis needs to be re-evaluated in patients who wish to start pregnancy.
Many young women discover haematuria and proteinuria on the occasion of their first urinalysis during pregnancy and these urinary changes are often considered as pregnancy-related missing a diagnosis of underlying glomerular disease. Microscopic haematuria is very common during pregnancy, detected in about 20 % of women, and disappears in 75 % after delivery [179]. Postpartum follow-up is recommended in women with persistent haematuria and presumed underlying glomerulonephritis, after excluding the presence of infection [180].
The development of proteinuria during pregnancy is often associated with PE. However, as previously discussed, isolated proteinuria or proteinuria with haematuria may reflect the presence of a renal disease.  The indication for renal biopsy is limited by the increase in bleeding risk; however, renal biopsy may be considered, in particular in early pregnancy, in the presence of rapidly progressive kidney function impairment or severe nephrotic syndrome [140].

IgA nephropathy
IgAN is the most prevalent primary glomerulonephritis worldwide, most often diagnosed in childbearing age. Proteinuria and/or hypertension in pregnancy are reported in up to 55 % of cases, prematurity in 30 % [181]. Several reports, including an Italian multicentre study on more than 200 pregnancies, suggest that pregnancy does not increase the progression of IgAN, at least in the earlier CKD stages. Favourable prognostic factors are normal blood pressure and GFR equal to or above 70 ml/ min before conception [60, [181][182][183][184][185][186]. Proteinuria at conception was associated with renal function impairment after delivery, stressing the role for its reduction before pregnancy, thus supporting discontinuation of ACEi and ARBs at the first positive pregnancy test (see previous paragraph). In the case of rapid progression, the differential diagnosis may consider both severe PE and a superimposed kidney disease (such as haemolytic uremic syndrome) [187].

Other forms of primary glomerulonephritis
The incidence of primary nephrotic syndrome is approximately 3-5 cases per 100,000 inhabitants per year in children and adults. Albeit rare, all the main causes of primary nephrotic syndrome may present in pregnancy.  Some authors presently classify MCN and FSGS as 'podocytopathies' due to the characteristic podocyte damage or dysfunction [188]. Interestingly, podocyte impairment has been described in PE [126]. This may at least partly account for the reported, albeit debated, development of FSGS after PE [188][189][190]. FSGS associated with PE is characterized by an increase in glomerular size, as is observed in other forms of secondary FSGS such as obesity, in which hyperfiltration plays a central role [191].

Main biochemical tests prescribed
While steroids are the first choice therapy in primary FSGS, prolonged treatment at high doses may have a detrimental effect on foetal growth [192]. Since the placental enzyme 11b-hydroxysteroid dehydrogenase 2 (11b-HSD2) protects the foetus reducing the transplacental passage of prednisolone, this drug may be preferred to betamethasone or dexamethasone [193,194]. In the case of steroid resistance, calcineurin inhibitors may be added ( Table 2).
Idiopathic membranous nephropathy (MN) often presents with a nephrotic syndrome; since, in over 70 % of patients with primary disease, autoantibodies against phospholipase A2 receptor 1 (PLA2R) are detectable by commercial kits, diagnosis may be performed without a kidney biopsy, which may, nevertheless, be required after pregnancy for staging [195]. A second target, thrombospondin type-1 domain-containing 7A (THSD7A) has been identified in a cohort of PLA2R-negative patients with MN, but is not yet clinically available [196]. In a study published in 1987, encompassing 33 pregnancies in 24 patients with MN, an increased foetal loss was observed, together with a risk of worsening of renal function [197]. This grim prognosis was not confirmed in a more recent cohort of nine MN patients with 51 pregnancies [198]. While steroids and calcineurin inhibitors are the milestones of treatment in pregnancy, addition of aspirin or low molecular-weight heparin should be considered, according to the clinical conditions.
Few studies and isolated case reports regard pregnancy in patients with primary MPGN, a multifaceted disease presently undergoing reclassification [199]. In a relatively old study on 123 pregnancies in 86 patients with biopsyproven various glomerular diseases, a higher incidence of complications in MPGN patients versus MGN was observed [200]. However, the outcome was favourable in recent reports, including a case in which the nephrological decision to start steroid therapy instead of pregnancy termination, allowed to safely reach the end of gestation [201][202][203].
In conclusion, the Study Group calls for attention to the differential diagnosis of glomerulonephritis in pregnancy, and underlines the importance of pregnancy as the first occasion for diagnosis in apparently healthy women. The limits of the current evidence and the need for further studies targeted to the different nephropathies has to be further underlined.

Patients should be in remission for at least 6 months
to reduce the risks of SLE flares in pregnancy (strong recommendation, scattered evidence). 5. Cyclophosphamide, mycophenolate and leflunomide have potential teratogenicity, and patients should be switched early to non teratogenic drugs (strong recommendation, scattered evidence). 6. Pregnant SLE women should be followed by a multidisciplinary team, in a tertiary care hospital (strong recommendation, indirect evidence). 7. Changes in immunosuppressive treatment in pregnancy should be limited to SLE flares; choice of treatment should be individualised (strong recommendation, indirect evidence). 8. Prednisone is considered safe. Intravenous immunoglobulin therapy is an effective and safe treatment for SLE pregnant patients with recurrent spontaneous abortions, especially if associated with antiphospholipid antibodies (strong recommendation, evidence from large series). 9. Low molecular-weight heparin with or without lowdose aspirin should be given to patients with SLE and antiphospholipid antibodies at high titres or triple positivity (strong recommendation, evidence from large series and RCTs). 10. An increase in glucocorticoid doses at time of delivery to prevent postpartum flares is a matter of debate (limited evidence). 11. Follow-up should be intensified postpartum since SLE flares may be more frequent in the first 6-12 months. Follow-up after PE should consider long-term cardiovascular complications (strong recommendation, evidence from large series).
12. Breastfeeding is allowed if prednisone does not exceed 20 mg/day. Azathioprine and cyclosporine are found in breast milk and lactation is not recommended (strong recommendation, evidence from large series).
SLE is a chronic autoimmune disease, primarily affecting young females with no influence on fertility. Pregnancy in women with SLE, and in those with lupus nephritis in particular, remains a high-risk situation for maternal and foetal complications [204][205][206][207]. The paucity of prospective studies on the subject makes evidence-based recommendations difficult; however, a number of retrospective cohort studies suggest that women with stable and prolonged remission of SLE, normal renal function, normal blood pressure and negative antiphospholipid antibodies have a high probability to have positive foetal and maternal outcomes [69, [204][205][206][207][208][209][210][211][212].
Since pregnancy-related risks are exceedingly high in SLE flares, effective contraceptive counselling is of paramount importance. Two large RCTs demonstrated the safety of oral contraceptives in SLE patients with inactive or mild active disease [213,214]. However, patients with severe active disease or with antiphospholipid antibodies should not be treated with these drugs.
An extensive clinical, biochemical and immunological evaluation, in particular regarding the extractable nuclear antigen (ENA) profile and presence of antiphospholipid antibodies is necessary before pregnancy, together with a careful re-evaluation of the ongoing treatments [207,[209][210][211].
Cyclophosphamide should be withdrawn at least 3 months before pregnancy and mycophenolate shifted to azathioprine, as the European Medicine Agency (EMA) recommended in its 23/10/2015 warning. The shift from mycophenolate to azathioprine should be done at least 6 weeks before pregnancy in order to minimize the risk of SLE flares due to this change of therapy [215,216]. Patients should start the pregnancy with the lowest possible dosage of the allowed drugs. Small birth weight and prematurity are associated with high-dose glucocorticoids, azathioprine and calcineurin inhibitors. Hydroxychloroquine reduces the rate of prematurity, intrauterine growth restriction and the risk of recurrent anti-SSA/Ro-antibodies associated cardiac manifestations in SLE pregnancy, and for these reasons should not be discontinued before pregnancy [217][218][219]. High-dose immunoglobulins are reported as safe and effective in SLE flares. Intravenous immunoglobulin therapy is an effective and safe treatment for SLE pregnant patients with recurrent spontaneous abortions, especially if associated with antiphospholipid antibodies [220][221][222].
Whether preeclampsia can be prevented by low dose aspirin is still a matter of debate, but current evidence supports its use in high risk populations [223,224]. Low molecular-weight heparin with or without low-dose aspirin should be added to patients with SLE and antiphospholipid antibodies at high titres or triple positivity [225,226].
Foetal loss includes spontaneous abortion, stillbirth and neonatal death, and, based on a systematic review of 37  [231].

Previous acute pyelonephriƟs and diseases at risk of UTI including malformaƟons and ADPKD
With the progressive reduction of foetal losses, the percentage of pregnancies that ends in premature birth has progressively increased, probably as a result of the immunological and vascular abnormalities associated with SLE. In lupus nephritis, approximately one-third of children are delivered preterm [205][206][207][208][209][210].
Neonatal lupus and congenital heart block are very rare complications, reported in less than 2 % of pregnancies, and are associated with positive SSA or SSB antibodies [232].
Several predictors of foetal outcome have been identified: high-activity of SLE in the first and second trimesters led to a threefold increase in pregnancy loss [233]. Active lupus nephritis at conception, and hypertension emerged as predictive factors for adverse foetal outcome in our as well as other authors' experience [206,227,234].
As already mentioned, the presence of GFR less than 40 ml/min/1.73 m 2 and of proteinuria greater than 1 g/d before conception are strong predictors of poor foetal outcomes whatever the underlying renal disease [52-56]. A relatively old review of 10 studies of 554 women with SLE found that foetal loss ranged from 39 to 59 % in patients with antiphospholipid antibodies versus 18 % in those with negative antiphospholipid antibodies [235].
Overall, as underlined above, history of nephritis, active nephritis or active SLE at conception are the strongest predictors of adverse foetal and maternal outcome [236,237]. For these reasons, a prerequisite for the success of pregnancy in patients with lupus nephritis is the planning of the pregnancy during periods of disease quiescence.
The study group therefore recommends to pay particular attention to pre-pregnancy counselling and to a strict multidisciplinary follow-up, in specialised Centres, in pregnancy in patients with SLE nephropathy.
Specific diseases: vasculitides 1. Pregnancies in woman affected by vasculitides are rare. The presence of active disease or of disease diagnosed in pregnancy is associated with a higher risk of adverse pregnancy-related outcomes (strong recommendation, scattered evidence). 2. A strict, multidisciplinary follow-up should be planned in tertiary care Centres to identify the best, personalised therapeutic strategies and the best balance between pregnancy termination, risks of early delivery and maternal risks (strong recommendation, indirect evidence).
Pregnancies are very rare in women affected by systemic vasculitides. The main reason is that, except for Takayasu disease, the peak age is typically over 40 years; the multiorgan involvement usually discourages the few patients in childbearing age from planning a pregnancy. Furthermore, the most frequently used immunosuppressive drugs, above all cyclophosphamide, contraindicate pregnancy and may induce premature menopause and sterility. However, with earlier diagnosis and new treatments, pregnancies in these patients will probably increase.
Because of the rarity of pregnancy, data on clinical management and maternal and foetal outcomes are based on few series, from experienced referral centres, and on case reports. Within these limits, some important points for risk assessment may be identified [238][239][240][241][242][243][244][245].
With regards to disease activity at conception, three situations are possible: pregnancies electively planned in remission and conducted under a strict multidisciplinary surveillance; unplanned pregnancies in patients on treatment; de novo or relapsing systemic vasculitis during pregnancy.
Pregnancy in patients affected by vasculitis in remission, undergoing planned multidisciplinary follow-up are reported to have good maternal and foetal outcomes, provided treatment is strictly monitored, to avoid relapses. Most published cases fall into this category (usually milder and often localized disease). However, some authors outline specific risks, mainly in the last weeks of pregnancy, including cardiac failure, asthma or decompensating subglottic stenosis.
The materno-foetal risks are higher in patients who get pregnant in a phase of active disease, who relapse or who develop de novo vasculitis during pregnancy. In this setting, ominous maternal and foetal complications have been reported. Among the life-threatening complications, pulmonary haemorrhage, rapidly progressive glomerulonephritis, polyneuritis and myocarditis have been described.
In the setting of kidney involvement and hypertension, the differential diagnosis between vasculitis and PE may be difficult; whatever the cause, hypertension and renal impairment may persist after delivery.
Therapeutic strategies need to be individualised. Patients should be aggressively treated in the case of flares or of active disease. Pregnancy termination or induction of delivery should be considered in these cases. Low-dose acetylsalicylic acid may be indicated for PE prevention. High-dose intravenous (iv) steroids, plasma-exchange, high-dose immunoglobulin, and azathioprine may allow continuation of pregnancy in an attempt to find a balance between maternal and foetal risks [238][239][240][241][242][243][244][245]. For further therapeutic issues, please refer to Tables 1, 2, 3 and 4. Acknowledging the scarce experience on this issue, the Study Group strongly recommends that patients with systemic vasculitis should be referred to tertiary care Centres to undergo a strict, personalised, multidisciplinary followup.
Specific diseases: diabetic nephropathy 1. Diabetic nephropathy is associated with a higher risk of perinatal death as compared with other primary causes of CKD (evidence from several studies on diabetes, no specific comparison with other causes of CKD). 2. There may be a higher incidence of malformations in diabetic nephropathy as compared to diabetes per se, in which an in increase in malformation has been reported (important recommendation, one large study only reporting higher risk of malformations in diabetic nephropathy). 3. Proteinuria may steeply rise in pregnant women with diabetic nephropathy, making the differential diagnosis with PE very difficult (strong recommendation, scattered data). 4. Prenatal counselling is fundamental to reduce the risks linked to poor glycaemic control and to optimize multidisciplinary follow-up (malformations, prematurity, large for gestational age babies).
The evidence regarding diabetic nephropathy in pregnancy is limited. In a recent systematic review, out of thirty-four papers retrieved, only two reported on more than 100 patients, while most studies described less than 50 cases [28]. When diabetic nephropathy is broadly defined as presence of any sign of renal disease including microalbuminuria, its prevalence ranges from 5 to over 25 % in type 1 diabetic pregnant women. During pregnancy, diabetes per se has been associated to a variety of complications, including malformations, foetal growth impairment, stillbirth, early mortality and preterm delivery while the incidence of PE has been reported as significantly increased, even if this is always within the limits of the non-univocal definition of this term [28-30, [246][247][248][249][250].
Maternal outcomes are widely scattered and probably reflect several factors: the lack of univocal definitions of diabetic nephropathy, CDK and PE, increasing maternal age over time, different study aims, patients selection and setting of several studies, small sample size and singlecentre and care settings.
While the high risk of prematurity, small for gestational age babies, or intrauterine growth restriction is shared also by other kidney disorders and increases along with the progression of kidney disease, stillbirth or foetal death are shared only by SLE, sharing with diabetes the important microvascular damage [62,242]. The risk is present also in diabetic patients without overt nephropathy, but appears to be increased in the setting of diabetic nephropathy [28][29][30]. Interestingly, even if in the last period, the risks of stillbirth and foetal death are reduced with respect to the '80s (from over 10 to about 5 %), the reduction is much lesser than that observed in the case, for example, of mothers on dialysis, which has decreased by 25 % each decade since the '80s, albeit starting from a much higher level [28-30].
Only one recent paper specifically reported on malformations in diabetic nephropathy, described as increased compared to diabetic patients without nephropathy [91]. While these data, obtained in a large cohort, are waiting for further confirmation, a note of caution should probably be made at counselling.
Acknowledging the need for further research in this field, the Study Group suggests a strict multidisciplinary followup, paying attention to the concomitant factors, such as hypertension and diabetes, that may further increase pregnancy related complications, and suggests strict follow-up also after delivery, in particular in patients who developed severe proteinuria and hypertension (whether or not labelled as PE) during pregnancy [251][252][253][254][255][256][257][258].
Specific diseases: autosomal-dominant polycystic kidney disease 1. ADPKD patients may have a higher risk of pyelonephritis in pregnancy and should be strictly controlled for positive urinary cultures (moderate suggestion, scattered evidence). 2. ADPKD patients may be at higher risk for PE, also when pregnancy starts with normal blood pressure and without proteinuria (moderate suggestion, scattered evidence). 3. The Study Group does not support a policy of systematic caesarean delivery in ADPKD patients (moderate suggestion, scattered evidence). 4. While acknowledging the interest in pre-implantation genetic selection, as a potential tool in autosomal recessive PKD, there is no experience on ADPKD, and the problems linked to assisted fertilization procedures have to be carefully weighed up (moderate suggestion, scant evidence).
The evidence on ADPKD in pregnancy is limited, and the largest recently published series encompasses 54 patients J Nephrol (2016) 29: 277-303 291 only [259]. In keeping with older observations, the current evidence suggests that hypertensive disorders of pregnancy, including PE, are more frequent in ADPKD patients. Preterm delivery and UTIs were also found to be increased, but the relationship between them is not completely clarified [259][260][261][262].
The old tenet that women with ADPKD should deliver by caesarean section because of the risk of intracystic bleeding, due to the increased abdominal pressure at parturition, is not supported by current evidence, also considering the risks of a surgical intervention and of urinary catheter. However, some cases have also recently been reported, and the opinion of the Study Group is that the risk of intracystic bleeding should be considered; therefore, we suggest ultrasound monitoring in particular of the largest cysts, also in non-symptomatic patients, at least in the proximity of and immediately after delivery [262][263][264].
The Group does not presently support the choice of implantation selection procedures, given their non-negligible rate of complications; such an approach has been recently discussed as for autosomal-recessive polycystic kidney disease, where, in spite of the higher severity of the kidney disease, it is likewise controversial [265,266].
Specific diseases: chronic pyelonephritis, kidney scars and diseases at risk for upper urinary tract infections in pregnancy (including kidney stones) 1. Asymptomatic bacteriuria, acute cystitis and acute pyelonephritis have been associated with an increased risk of abortion, preterm delivery, low birth weight and perinatal mortality. Gestational hypertension and PE are also increased (relative risk [RR] up to 5) (strong suggestion, from large observational studies). 2. Since untreated, asymptomatic bacteriuria persists in 60-85 % of cases and progresses to acute cystitis or pyelonephritis in up to 40 % of patients, all cases should be treated (strong suggestion, from large observational studies and RCTs). 3. Antibiotic therapy allows sterilization of urinary culture in 90 % of patients and reduces the incidence of acute pyelonephritis by 70-80 %, decreasing the incidence of low birth-weight and of preterm birth (strong suggestion, from large observational studies and RCTs). 4. The optimal treatment for asymptomatic bacteriuria, acute cystitis and acute pyelonephritis is not fully established. Empiric treatment should be tailored to the specific situation and take into account the local resistance patterns (strong suggestion, from large observational studies and RCTs).
Chronic pyelonephritis is now recognized as an umbrella term encompassing many chronic inflammatory conditions even in the absence of a clear history of renal infections [267]. Whatever the cause, kidney scars, identified by imaging techniques, define this condition [268][269][270][271]. Risk of bacteriuria is significantly higher in women with a history of UTI in childhood and adolescence, whether or not with kidney scars [271]. The incidence of asymptomatic bacteriuria is estimated to be 2-7 % of all pregnancies (more common in diabetic women: 8-14 %); acute pyelonephritis occurs in 1-2 % of low-risk pregnancies [272][273][274][275][276]. Pregnancy predisposes to UTI for several reasons: the physiological hydroureter of pregnancy, consequent to progesterone-induced hypotonia and decreased peristaltic activity and to the compression of the gravid uterus; pregnancy-induced glycosuria and aminoaciduria may facilitate bacterial colonization, immune-system changes, including reduced response to nitric oxide (NO) and Tolllike receptor 4 (TLR4) [277,278].
Asymptomatic bacteriuria has been associated with an increased risk of abortion, preterm birth, low birth-weight and perinatal mortality, thus leading to the current practice of repeated urinalysis and urinary cultures in all pregnancies [280][281][282][283][284].
The relationship between asymptomatic bacteriuria, low birth-weight and preterm delivery is however still elusive. Pro-inflammatory cytokines may initiate labour, while microorganisms may produce arachidonic acid, phospholipase A2 and prostaglandins that enhance cervical softening, which stimulates uterine contractions and preterm labour [285]. If untreated, bacteriuria persists in 60-85 % of cases, progressing to cystitis or pyelonephritis in up to 40 % of patients, and increasing the risk of PE [273,279].
The therapeutic protocols are controversial: single-dose therapy and 3-day courses were used with success in some studies. The pros are better compliance, lesser exposure for the foetus, fewer side effects, lower costs; the con is the higher recurrence rate. On this basis, the Study Group supports standard 6-7 days antibiotic courses until further data are available [300][301][302]. Follow-up requires at least monthly urine culture and, in case of persistence, suppressive therapy should be considered. For the most commonly used drugs, refer to Table 3.
Acute cystitis is frequently asymptomatic; conversely, acute pyelonephritis usually presents with flank abdominal or pelvic pain, and high fever; severe complications, including septic shock, respiratory distress and acute renal failure are reported in 2-25 % of cases [303][304][305][306][307][308][309][310][311][312][313][314][315][316][317][318]. Since Escherichia Coli is involved in about 80 % of the cases and other Gram negative germs in most of the remaining, empirical treatment should be started early on the basis of the local resistance pattern and of the eventual maternal intolerances (Table 3) [319,320]. Preeclampsia and preterm delivery are reported as increased after UTI in pregnancy, even in the absence of active infection at delivery [321].

Kidney stones
1. Kidney stones are rare in pregnancy, but they are associated with several negative pregnancy-associated outcomes, including PE and preterm delivery (strong suggestion, evidence from observational studies). 2. Within the limits of the scarce urethral view, ultrasound should be the first imaging modality, while magnetic resonance imaging (MRI) and computed tomography (CT) scans (II and III trimester) should be limited to selected cases (strong suggestion, scattered evidence). 3. While most of the stones are passed spontaneously, ureteroscopy is relatively safe in the case of need; small series suggest considering also the option of holmium laser (strong suggestion, large series for uteroscopy, small for holmium laser).
Nephrolithiasis is rare in pregnancy (1:1500-3000), despite several predisposing factors, including urinary stasis, hypercalciuria and increased urinary pH, increased excretion of uric acid, sodium and oxalates, out of proportion with citrate and magnesium excretion [322]. Calcium and phosphate are the most frequent components of kidney stones in pregnancy (up to 75 %) [322][323][324].
In most cases, the dilation of the urinary tract favours stone emission and only symptomatic therapy, based on hydration and analgesia is needed. Ureteral stents, percutaneous nephrostomy or ureteroscopy may be employed in selected cases, as well as holmium laser lithotripsy, according to small patient series [325][326][327][328][329][330][331].
Renal stents have been associated with increase in proteinuria to be considered in the differential diagnosis with PE [332]. The association with nephrocalcinosis should be kept in mind for allowing diagnosis of underlying metabolic disorders [333].
In view of the above, considering also the possible association between renal colic, premature rupture of membranes, pregnancy losses and PE, the Study Group suggests a strict multidisciplinary follow-up and personalised therapy for all patients with recurrent UTI or kidney stones [334][335][336][337].

Reflux nephropathy
1. Patients with reflux nephropathy are at higher risk for urinary tract infections even after reflux correction. 2. Reflux nephropathy shows a complex hereditary pattern and children from affected mothers should be screened before and after birth for UT and/or urinary malformations.
The frequent correction of reflux nephropathy has decreased its importance in pregnancy, and most of the studies related to this condition are relatively old. While the surveillance for infectious complications should follow the strict recommendations already mentioned in the case of recurrent UTI, the disease shows a complex hereditary pattern, and children from affected mothers should be screened before and after birth for UTI and/or urinary malformations [338][339][340].

Final comment
CKD is increasingly encountered in pregnancy and patients with chronic diseases, including CKD, are increasingly looking for pregnancies. Due to the changes in therapies, and in attitudes towards patients' self-determination as well as in the definitions of CKD and PE, the evidence on which counselling and follow-up are based is often scattered and fragmentary. This lack of ''certitudes'' should be kept in mind at counselling, and strongly supports shared choices in the context of multidisciplinary care. Clinical nephrologists need to be continuously updated in this emerging field, and the Study Group supports the establishment of educational programs and of multicentre research studies.