Causes of The Nephrotic Syndrome In Sweden 2014-2019 Depending On Clinical Presentation And Demographics

Background: Many different pathological processes can affect the integrity of the glomerular capillary wall and cause massive leakage of protein resulting in the Nephrotic Syndrome (NS). The prognosis and response to therapy differs depending on diagnosis, but renal biopsy cannot always be performed promptly. These is insucient knowledge to which extent clinical parameters can predict the diagnosis. Methods: Age, gender, haematuria, proteinuria, plasma creatinine, plasma albumin and nal diagnosis were retrieved for all adult patients with NS as indication for biopsy or massive albuminuria in conjunction with a low plasma albumin from the biopsy module of the Swedish Renal Registry (SRR) between 2014 and 2019. A basic calculator was developed to demonstrate the importance of clinical presentations in relation to the likelihood of having a specic diagnosis. Results: 913 unique patients were included in the study. Overall membranous nephropathy (17%) was the most common diagnoses, but when studying those <50 years old or women minimal change nephropathy (21 and 17 %) was the most frequent diagnosis. When examining those between 50 and 70 years-old, those with chronic kidney disease (CKD) 4 and those with negative dipstick tests for hematuria diabetic nephropathy (23, 30 and 21 %) was the most common underlying disease. Among those with high grade hematuria (grade 3-4 on dipsticks) Membranoproliferative glomerulonephritis was most common (14%), closely followed by IgA nephropathy (13%). Focal segmental glomerulosclerosis (9.7%) was less common than in many comparable studies. Conclusions: Clinical parameters have a profound impact on the likelihood of different diagnoses in adult patients with NS. Differences in clinical practice, inclusion criteria in studies and probably genetic background are important to account for when comparing data from different parts of the world.


Background
Massive leakage of proteins in the kidneys associated with oedema is referred to as nephrotic syndrome (NS) (1). NS is relatively common in renal practice with an incidence of 3-4 new cases per 100 000 each year (2)(3)(4) and is in many reports the most frequent indication for renal biopsy (5). In a recent study (6) on patients' perspectives of suffering from NS we found that it is perceived as a highly complex condition to grasp. Many patients describe the illness experience as being a stranger in an unfamiliar world of symptoms and medical treatments. This perception is at least partly shared by the physicians as the heterogeneity of the condition is a problem. Understanding how different clinical presentations are associated with different diagnoses and pathological processes is important.
The distribution of causes of NS varies between countries and over time which hampers comparisons. Membranous nephropathy (MN) was the most common cause of NS in India (7), Spain(8) and Denmark (9). Focal segmental glomerulosclerosis (FSGS) was the most common cause in USA (10) while IgA nephropathy (IgAN) was the most cause common in Czech Republic (11) and mesangial proliferative glomerulonephritis (MPGN) in Lebanon (12).
There are studies based on national registers (4,9,11,13), on regional registries (8, 14) and on data from single centres (7,10). Most of these studies did not focus speci cally at examining diagnoses behind NS and many are scarce on clinical data.
In some studies, secondary causes of NS such as diabetic nephropathy (DN) and systemic lupus erythematosus (SLE) were excluded (15,16). There are few attempts to analyse how demographic factors and clinical presentation affect the likelihood of different diagnoses among adult patients with NS. The aim of this study was to describe the causes of NS in the Swedish population and to analyse how the spectrum is dependent on factors readily available before return of biopsy results. To this end we have retrieved data on biopsy indication, albuminuria, haematuria, CKD stage, age and gender from the renal biopsy module of the Swedish Renal Registry.

Data source
Data was retrieved from the biopsy module of the Swedish Renal Registry (SRR-biopsy). The organisation of the SRR is described in previous publications (17) and the biopsy module was launched in 2015 for prospective registration of biopsies from native renal kidneys (18). It is also possible to enter older biopsies retrospectively, and it has been done in a systematic fashion at some hospitals for 2013 and 2014. For biopsy indication one of the following ve alternatives has to be chosen: nephrotic syndrome, acute nephritic syndrome, other acute kidney injury, CKD 1-2 and CKD 3-5. The result of the biopsy is entered both with a SNOMED (Systematized Nomenclature of Medicine) code from the pathologist and with an ERA-EDTA (European Renal Association − European Dialysis and Transplant Association) code for primary renal disease chosen by the nephrologist when combining clinical features, laboratory results along with biopsy results. When there are multiple ndings in the biopsy the one best explaining the biopsy indication should be entered as main diagnosis.

Study population
We included data from biopsies performed between January 1, 2014 and December 31, 2019. Three searches were performed in the SRR-biopsy (i) biopsy indication nephrotic syndrome, (ii) urine albumin/creatinine > 300 mg / mmol and (iii) urine albumin > 3.5 g/24 hours. Data retrieved from the registry at time of biopsy included patient age, gender, weight, renal diagnosis, blood pressure, dipstick haematuria, plasma creatinine, eGFR (CKD-EPI), plasma CRP and plasma albumin, urine albumin/creatinine, urine albumin (24 h collection), serum haemoglobin. Patients' biopsy indication other than NS were excluded from the study if plasma albumin was > 30 g/L.
If a patient had multiple biopsies clinical data was collected from the time of the rst biopsy, but the diagnosis was taken from the latest biopsy if performed on the same indication. The reason for performing a repeat biopsy is most often that the initial biopsy yielded insu cient material and unable to provide a clear diagnosis. We grouped the ERA-EDTA codes version 2018 for primary renal diseases into 12 groups (19) ( Table 1).

Study population
Our primary search criteria generated a total of 1734 biopsies entries. After removal of duplicates, repeat biopsies, incomplete data sets and patients not ful lling inclusion criteria, 913 unique patients were included in the analyses (Fig. 1). There were 735 patients with biopsy indication NS and 178 patients who had laboratory parameters compatible with NS but the main biopsy indication was not NS. In this "other indication group" the most common indication for biopsy was CKD 3-5 (48.3%) followed by acute nephritic syndrome (30.9%), CKD 1-2 (11.2%) and acute kidney injury (9.6%).

NS vs. other indication
There were only small and insigni cant differences in age and gender distribution ( Table 2) between those with biopsy indication NS and those with other indication. There were, however, a substantial difference in plasma creatinine concentration, where the other indication group had more than twice the level compared to the NS group, with reciprocal results regarding eGFR. The NS group also had lower systolic blood pressure (median 135 vs. 140 mmHg), lower proportion of patients with hematuria of grade 3-4 (26.2 vs. 43.5%) and higher blood haemoglobin (median 126 vs. 113 g/L).
Next, we compared those within the NS group that had laboratory features compatible with NS (n = 487) with those not ful lling the proteinuria (urine albumin/creatinine > 300 mg / mmol or dU-albumin > 3.5 g) and/or albumin criteria (plasma albumin ≤ 30 g/L) set up in this study at the time of biopsy (n = 248  When comparing relative proportions among those with a biopsy at CKD 1-2 with those with CKD stage 5 the largest decrease was seen for MN (26.0 to 2.0%) and MCN (27.1to 4.1%). Going in the other direction the largest differences were seen for vasculitis (0.3 to 16.3 %) and other vascular diseases (1.7 to 10.2%) and DN exhibited an increase from 6.6% in CKD 1-2 to 29.8% in CKD 4, but fell back to 23.5% in CKD 5. A relatively stable proportion in different CKD stages was seen for IgAN, being highest in CKD 4 with 8.5% and lowest in CKD1-2 with 5.8% (Table 4).

Nephrosis diagnosis calculator
When combining demographical and basal clinical chemistry data a calculator can be constructed (suppl 1). When age, gender, level of hematuria and CKD stage are entered the calculator provides the likelihood of different diagnoses and how the percentage change compared to no knowledge of clinical data.
For example, for a 30-year-old female with 3 + haematuria and CKD 2, SLE will be the most common diagnosis (50% chance) followed by IgAN (13%) and MCN (9%). If the same patient was male, the most probable diagnosis becomes IgAN (29%) followed by MCN and FSGS (14% each). Four other examples are shown in Fig. 2. To evaluate the calculator all patients with complete data were entered into the calculator. The biopsy veri ed diagnosis was found among the top three diagnoses in 62% of the cases, ranging from 22% for the collective diagnosis group "Other renal diagnoses" to 85% for vasculitis.

Discussion
In this study we present data on the causes of the nephrotic syndrome in Sweden and how this is affected by the clinical presentation, basic clinical features and demographic factors. These data facilitate the identi cation of the most probable diagnosis in cases when biopsy must be delayed or cannot be performed due to contraindications. Clinical practice, for instance differing indication for biopsy, will affect the reported patterns as well. In a cohort of Swedish patients that have performed a renal biopsy with the indication NS based on current clinical practice the most frequent diagnoses are MN (19%), MCN (18.5%) followed by DN (16.5%). However, the proportion varies considerably when subsets based on certain clinical parameters are studied.

Main ndings and comparison with previous studies
De nitions of disease and inclusion criteria are always crucial in studies concerning disease incidence. In our study MN and MCN were the most common diagnoses in patients with NS as indication for biopsy. In patients with heavy albuminuria but other indication for biopsy DN and IgAN were the most frequent diagnoses. This is in line with other studies (4,(7)(8)(9)(20)(21)(22)(23).
In contrast, a study from Lebanon (12) showed that MPGN was the most common diagnosis in NS, accounting for 27% of the cases.
There are differences in disease pattern on a regional and national level, due to socioeconomic factors and ethnicity (10). Compared with some other studies (7, 10, 12, 15, 16) we found a high proportion of patients with DN and low share of FSGS. It is well known that FSGS is correlated both to obesity (24) and to African ancestry (10,16). The SRR database contains no information on ethnicity, but in the general Swedish population approximately 18% have non-European ancestry, with a large population coming from the middle east, which might explain some of these differences.
Clinical practice, for instance differing indication for biopsy, will affect the reported patterns as well. DN was a common cause of NS (16.5 %) and the most common diagnosis in patients with heavy albuminuria with other indication for biopsy (21.9 %). We believe that this high share is explained by differences in clinical practice, and inclusion criteria for studies. Bandi et al. (7)

Diagnosis calculator
To demonstrate the usefulness of clinical data to predict the diagnosis for individual patients, we developed a basic calculator. By calculating the probability of the different diagnoses based on four clinical parameters (age, gender, haematuria and CKD), we can show the probabilities based on the distribution within these subgroups. As shown in Fig. 2 these probabilities exhibit substantial differences. Nonetheless, it is important to keep in mind that the calculator calculate probabilities and should not be used to set a diagnosis. Adding data from other cohorts would make it possible to enter more parameters such as serology, proteinuria or previous disease history (e.g. diabetes mellitus) and re ne the results. It may be further developed by using machine learning to calculate different weighting of the entered parameters and could potentially be used to nd the diagnosis in cases when it is not possible to perform a biopsy.

Limitations
This study has several limitations. The cohort stems from a single country with a primarily Caucasian population, and we do not have data on ethnicity. Indication for biopsy is entered at the discretion of individual clinicians and has not been validated. For instance, 4% of the patients with the indication NS had only moderate proteinuria (U-albumin/creatinine < 100 g/mol), even if time may have elapsed from the day the indication was set until the biopsy was actually performed and their cases with rapid response to therapy, we cannot exclude mistakes in the registration procedures. The study also has important merits. It contains a large number of cases collected over a relatively short time period, and thus re ects the current situation in our country. We provide data both on patients with a de ned set of laboratory values and of patients with a clinical diagnosis of NS and compare the differences between these two inclusion criteria. We thereby provide the entire spectrum of the NS.  Figure 1 Flow chart of study population

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. AJONephroticSyndromeDiagnosistool.xlsx