Introduction

Sri Lanka is a South Asia island country with an estimated population of 22 million in 2023 [7]. This investigation is based on the spectrum of findings from a decade (2011–2020) of native renal biopsies performed at Kandy Teaching Hospital, Sri Lanka. Kandy is the capital of Central Province, 100 km inland from the national capital Colombo, and is surrounded by extensive agricultural areas, used especially for rice cultivation. Kandy Teaching Hospital has strong links with the Medical Faculty at the adjacent University of Peradeniya Hospital, which are both public hospitals providing medical services without charge.

Prevalence and risk factors for impaired renal function in the district of Anuradhapura, Sri Lanka was assessed in a cross-sectional clustered population-representative survey in 2003–2008 of adults ≥ 19 years (n = 6,153) [1]. Participants with ≥ 1 + proteinuria (≥ 2/3 occasions) had blood taken for serum creatinine measurement from which estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet in Renal Disease (MDRD) simplified equation [15]. Proteinuric-chronic kidney disease’, as defined by proteinuria + 1–3 with reduced eGFR (< 60 ml/min per 1.73m2), was present in 5.1% in Medawachchiya n = 2600 (North Central Province), 9.5% in Yatinuwara n = 709 (Central Province), and 2.3% in Hambantota n = 2844 (Southern Province).

Prevalence rates of CKDu, as measured by low estimated glomerular filtration rate (eGFR < 60 ml/min per 1.73m2) calculated using the ‘CKD EPI equation’ [5, 15] from serum creatinine incorporating age and sex, in the absence of heavy proteinuria, diabetes or hypertension, has been estimated at 11% for men and 4% for women in a cross-sectional population prevalence study (n= 4,803) within affected areas in North Central Province [26].

The prevalence of CKDu in Sri Lanka has led to research into possible causes of CKD in the region and methods of prevention and control. These have involved expensive health service interventions, including early detection through population screening for renal abnormalities and diagnostic assessment as necessary (including renal biopsy), and implementation of various treatment modalities, including kidney replacement therapy [8, 12, 23, 30, 31]. Tubulointerstitial nephritis (TIN) has been reported as the predominant histopathological feature in renal biopsies from CKDu cases (n= 125, 74% male); and in 84% of cases in renal failure stage 3–4, from Anuradhapura, North Central Province, Sri Lanka 2008–2012 [27]. In a study of cases (n= 59) presenting with an acute illness with renal symptoms, and elevated serum creatinine (excluding those with known causes and/or small kidneys) from Kandy (Central Province, Sri Lanka), histopathology revealed a mixture of both acute and chronic interstitial nephropathy which, according to the authors, would imply that chronicity may be preceded by recurrent acute episodes [2].

There are numerous causes of TIN including, but not limited to, infections, autoimmune disease, environmental toxins and envenomation, among others [16]. Some renal biopsies may receive a syndromic diagnosis of CKDu when no known cause is evident without a recorded histopathological diagnosis. While TIN is associated with CKDu in Sri Lanka, TIN has not been reported as a predominant histopathological finding in most international renal biopsy studies (Table 1) [6, 13, 17, 21, 25, 28]. In Sri Lanka, the proportions of TIN reported in previous renal biopsy case series were: 4.3% in Colombo South Kalubowila [20]; 25.5% (if all TIN is included) in Nugegoda west of Colombo [24]; and 22.2% (if all TIN is included) in Kandy [4] (Table 1). Previous studies in Sri Lanka have not analysed differences in TIN by sex, despite epidemiological prevalence studies demonstrating that men are more likely to be affected by CKDu [26].

Table 1 Proportion of TIN in native renal biopsies, Sri Lanka and internationally
Full size table

The present study is the largest investigation (n = 3,095) of native renal biopsies (2011–2020) assembled in Sri Lanka, with cases drawn in major part from a large inland agricultural region with known areas of endemic CKDu. The objectives of the present study were to: (1) describe the range and relative frequency of histopathological diagnoses, and (2) provide stratified analyses of the different patterns of histopathology for adults ≥ 15 years (by sex), and for children < 15 years.

Methods

Study design

The present study is a retrospective study of renal biopsies performed at Kandy Teaching Hospital during 2011–2020, which includes a small number (n = 52) of biopsies in children < 15 years that were performed at University of Peradeniya Hospital.

Renal biopsy

Kandy Teaching Hospital provides nephrology services for patients who present directly to the hospital with possible symptoms of renal disease, as well as routinely servicing those referred from outpatient and outreach services, and from surrounding rural hospitals and clinics. The referral pathways for renal biopsies differed over the ten-year-study period because of sporadic population screening programs, and the establishment of outreach renal clinics in areas with significant occurrence of clinical chronic kidney disease. The spatial distribution of renal biopsy cases performed at Kandy Teaching Hospital (Fig. 1) shows the intake from surrounding agricultural areas. District of residence was known for 77% (n = 2,396) of biopsies, and of those biopsies 89% (n = 2,133) were from districts in, or directly surrounding, Central Province (Kandy n = 757, Matale n = 377, Nuwara Eliya n = 251, Badulla n = 214, Kegalle n = 177, Kurunegala n = 140, Ampara n = 115, and Polonnaruwa n = 100).

Fig. 1
figure 1

Number of renal biopsies by district (grey lines) examined at Kandy Teaching Hospital, 2011–2020. (Legend for Fig. 1) Province outlines shown in blue lines. Number and proportion of biopsies represented by each coloured segment is shown in the key

Full size image

The decision to progress to renal biopsy was at the discretion of the treating nephrologist based on clinical findings, urine analysis, serum creatinine and other blood tests, and results of renal ultrasound. Common reasons for biopsies not proceeding include small kidneys identified on ultrasound measurement (< 9 cm), lack of perceived clinical benefit where renal biopsy would not lead to a change in treatment, and patient refusal—including for socio-economic reasons, such as cost of transport and time off work to attend an appointment for the procedure.

The reasons for proceeding with renal biopsy were derived from data extracted from the Pathology Report Form (one page) and include either explicit statements of clinical indications (n = 2,047) or were derived from the clinical information contained elsewhere on the form (n = 1,048). The indications for renal biopsy were: nephrotic syndrome (n = 1,094, 35%), unexplained chronic renal failure (n = 573, 19%), unexplained proteinuria and/or haematuria (n = 638, 21%), to plan future therapy for those previously diagnosed and treated (n = 329, 11%), unexplained acute renal failure (n = 204, 7%), diabetes mellitus (without haematuria) (n = 174, 6%), unknown (n = 51, 2%), and renal diseases in pregnancy and postpartum (n = 32, 1%).

Preparation of biopsy

Two cores were taken from all renal biopsies: one in formal saline for routine tissue processing, and the other was a fresh sample transported in Michael’s solution or with ice packing within two hours of sample collection for immunofluorescence. Formalin fixed paraffin embedded (FFPE) tissue was used for haematoxylin and eosin (H & E) stains for general histopathological evaluation and Jones' methenamine silver stains for assessment of the glomerular basement membrane. Generally, eight to twelve H & E sections were examined and if required more tissue sections were obtained. Immunofluorescence staining was performed to detect immunoglobulins and complement components, using IgG, IgA, IgM, and C3. Although electron microscopy was not routinely performed on all samples, it had been previously performed on a batch of cases diagnosed with chronic kidney disease of unknown aetiology (CKDu) at the University of Erlangen–Nuremberg (Germany) using FFPE tissue. This study did not identify any unique glomerular or tubular interstitial pathology for CKDu. The comprehensive approach described ensures a thorough evaluation of renal biopsies under resource constrained settings, facilitating accurate diagnosis and guiding appropriate management.

Data acquisition

Reports by pathologists who reviewed renal biopsy material for each case were abstracted onto a structured form and entered into an electronic database. Information derived from the report included specific histopathological features and a transcription of written comments and summary diagnoses by the pathologists. Demographic and residential area data were recorded. No personal identifiable data were extracted. Consequently, it was not possible to ascertain if an individual was biopsied more than once over the ten-year study period, or to follow-up individual cases. In Sri Lanka, it is usual practice for patients to take medical records home with them, so no other information was available from the hospital other than the renal biopsy report form. Completeness of data was 94% for sex and 96% for age.

Data preparation

The pathologists’ written description and comments, specific histopathological features, and summary diagnoses were utilised to categorise the renal biopsies as a single predominant diagnosis. The array of diagnoses and their abbreviations reported from 52 international studies of renal biopsies (1980–2019), summarised by Muthukuda et al. [20], were taken as a guide for names, abbreviations and nosology of renal histopathological conditions. Information from published articles on particular diagnostic categories, and from the International Classification of Disease version 10 (ICD-10), were used where appropriate. Fogo et al. [9] Fundamentals of Renal Pathology was consulted extensively for criteria for histopathological diagnosis and for classification of diagnoses.

Certain diagnostic categories in the original pathology reports were individually reviewed in instances where glomerulonephritis (unqualified) was the sole or predominant diagnosis recorded (n = 150), tubular injury (unqualified) was the sole or predominant diagnosis (n = 50), multiple histopathological diagnoses were provided without selection of a predominant condition (n = 132), CKDu was the diagnosis with or without other conditions (n = 98), and histopathological features were reported without any diagnosis (n = 89). In these 519 cases, comments and diagnoses by pathologists on the report form, and categorical histopathological information, were used to arrive at a probable predominant histopathological diagnosis where possible.

The histopathological findings in all 98 cases of CKDu (syndromic diagnosis) in the present study were individually reviewed using commonly accepted criteria for tubulointerstitial nephropathy. The characteristics for TIN, if present, were indicated by the pathologist on the Histopathology Report Form. These indicators compromise: (a) lymphocytic interstitial invasion and/or tubilitis (active or active); (b) tubular atrophy and/or fibrosis (interstitial, peri-glomerular, perivascular) (chronic); and (c) both together. The histopathology review of these 98 CKDu cases indicated: 92 cases of tubulointerstitial nephropathy (84 cases of tubulointerstitial nephropathy only, and 8 cases as tubulointerstitial nephropathy with an additional diagnosis).

From 5,014 renal biopsies, 1,572 post-kidney transplant biopsies were excluded. Other exclusions included 188 biopsies with no renal tissue in the biopsy core, 95 biopsies with no histopathological abnormalities (normal), and 64 biopsies with insufficient abnormalities to indicate a specific histopathological diagnosis (Fig. 2). Thus 3,095 renal biopsies were included in this study.

Fig. 2
figure 2

Inclusion flow chart for renal biopsy record analysis, Kandy Teaching Hospital, 2011–2020

Full size image

Data analysis

For tabulation, histopathological diagnoses were categorised according to whether they primarily affected the glomerulus, the interstitium (or adjacent structures), or were renal manifestations of systemic disease. Each category was then arranged in sub-groups according to commonly used nomenclature, and then into reportable histopathological diagnoses following Fogo et al. [9]. Counts and proportions (%) of totals were provided for each category, subgroup, and diagnosis.

Tabulations were prepared for all specific reportable histopathological diagnoses arranged into related groups as set out in Tables 2, 3, 4 and 5. The number of cases and proportion (%) of biopsies are given for specific reportable histopathological diagnoses, subgroups, and main categories. Data were stratified by (1) adults ≥ 15 years by sex, (2) children < 15 years. Chi-squared tests were conducted to compare proportions in groups within this series, and to test the statistical significance of differences in proportions of TIN compared to other similar renal biopsy studies.

Table 2 Histopathology of all native renal biopsies, Kandy, Sri Lanka, 2011–2020
Full size table

Results

All cases (n = 3,095). Following documented exclusions (Fig. 2), 47.2% of the diagnosed histopathological entities were primary glomerular lesions (n = 1,460); tubulointerstitial and related histopathology (n = 808) comprised 26.1%, and systemic disease with renal involvement (n = 824) comprised 26.6% including secondary glomerular pathology (Table 2). The largest proportion of cases by individual category was accounted for by tubulointerstitial nephritis/nephropathy (TIN) (n = 760), which comprised 24.6% of cases. This category includes cases of active (acute), chronic, and acute/active-on-chronic TIN. Other important recorded conditions included SLE 13.9%, FSGS 9.6%, IgAN 7.8%, MCD 7.3%, DN 6.6%, and membranous nephropathy (MN) 5.7%.

Women aged ≥ 15 years (n = 1,353). Of the biopsy records for women ≥ 15 years included in analyses, 46.0% (n = 622) of all the diagnosed histopathological entities were (primary) glomerular lesions, 18.1% (n = 245) interstitial and related histopathology, and 35.8% (n = 484) systemic disease with renal involvement (Table 3). The highest proportion of cases comprised SLE (24.0%), followed by TIN (16.9%). Other important recorded conditions included FSGS 10.4%, MCD 7.5%, MN 7.2%, DN 6.7%, and IgAN 6.4%.

Table 3 Histopathology of renal biopsies, women aged ≥ 15 years, Kandy, Sri Lanka, 2011–2020
Full size table

Men aged ≥ 15 years. (n = 1,313). Of the biopsy records for men ≥ 15 years included in analyses, 46.1% (n = 605) of the diagnosed histopathological entities were (primary) glomerular lesions, followed by interstitial and related histopathology (n = 470, 35.8%), and systemic disease with renal involvement (n = 238, 18.1%) (Table 4). The largest proportion of cases was TIN (34.2%). Other important conditions recorded were IgAN 9.5%, FSGS 8.6%, DN 7.8%, and MCD 6.4%. The proportion of TIN was significantly higher (p < 0.001) in men than in women.

Table 4 Histopathology of renal biopsies, men aged ≥ 15 years, Kandy, Sri Lanka, 2011–2020
Full size table

Children < 15 years both sexes. (n = 189). Of the biopsy records for children < 15 years included in analyses, 74.1% (n = 140) of the diagnosed histopathological entities were (primary) glomerular lesions, 18.5% (n = 35) were systemic disease with renal involvement, and 7.4% (n = 14) were interstitial and related histopathology (Table 5). The most frequent individual conditions were: FSGS 14.8%, MCD 12.2%, mesangioproliferative glomerulonephritis (MesPGN) 11.6%, and SLE 10.1%. Other important conditions were IgAN 9.0%, proliferative glomerulonephritis (PGN) 7.9%, glomerulonephritis (GN n.o.s.) 7.9%, and TIN 6.7%.

Table 5 Histopathology of renal biopsies, children < 15 years, Kandy, Sri Lanka, 2011–2020
Full size table

Comparison of TIN to other renal biopsy studies

For adults ≥ 15 years of both sexes (n = 2,666) in the present study, 677 (25.4%) were recorded as TIN, which is similar in magnitude but statistically significantly higher (p = 0.008) than 22.2% (including all TIN) in the previous Kandy study (2010–2019) including all probable TIN = 596 cases for ages ≥ 12 years (n= 2680) (because of the large numbers in both studies) [4]. The proportion of TIN in the present study is not statistically significantly different (p = 0.44) from 25.5% (including all probable TIN = 148 cases) in the Nugegoda study (west of Colombo) of ages ≥ 12 years (n= 547)  [24]. The proportion of TIN in the present study is statistically significantly higher (p < 0.001) than the 2.9% from 4 cases of TIN in the Kalubowila study (n= 140) from Colombo south of ages ≥ 16 years [20]. Proportions of TIN reported in similarly sized renal biopsy case series from Korea (TIN = 0.04%) [32], India (5.3%) [18], and Pakistan (11.6%) [19] were all significantly lower than in the present study (p < 0.001).

Discussion

This retrospective analysis of 3,095 native renal biopsies performed at Kandy Teaching Hospital between 2011 and 2020 aimed to characterise the predominant histopathological all age findings, and to examine differences in diagnoses in adults ≥ 15 years by sex, and children < 15 years. Although the majority of documented histopathological entities were primary glomerular lesions (47%) of all biopsies, there were important differences between adults ≥ 15 years of each sex, and with children < 15 years compared to adults. In a quarter of all biopsies the main histopathological entity was TIN (25%), however, in adult males TIN was 34% (n = 448) compared to adult females 17% (n = 228). In children < 15 years TIN was 7% (n = 13).

In similar studies from other parts of Sri Lanka, TIN has been reported to vary from 4% in Colombo South Teaching Hospital (Kalubowila) (n= 140, 56% female) [20], to 22% in Kandy (n= 2680) [4], and 26% in Nugegoda (west of Colombo) (n= 771, 67% male) if all related TIN categories are included [24]. The present study in Kandy (n = 3,095) found the proportion of TIN to be statistically significantly higher (p < 0.001) than in Kalubowila (Columbo south), but not significantly different from that in the Nugegoda study (p > 0.05). Although the magnitude of the proportions of TIN in the present Kandy study 2011–2020 (n = 3095) at 25% is not substantially different from the 22% probable TIN in the previous Kandy study 2010–2019 (n = 2680), because of the large numbers the difference was statistically significant at p < 0.04.

Stratification of TIN by sex was not included in previous Sri Lanka biopsy studies, which also did not include children < 12 years [4, 24]. The breakdown by sex in the present study highlights a significantly higher proportion of TIN in men compared to women. The sex difference in proportions of TIN is likely influenced by more frequent occupational exposure to putative causes through farming activities in men than women, and the higher occurrence of lupus nephropathy in women than men.

In international renal biopsy studies, the proportion of tubulointerstitial nephritis has varied from 0.4% of total cases in Korea [32] to 18% of cases in South Africa [21] (Table 1). Protocols for referral for renal biopsy may differ between centres and countries, which may impact direct comparisons, but the evidence suggests that the proportion of TIN found in the present study is higher than renal biopsy studies from other countries. The findings of a male predominance in TIN have been reported in studies of CKDu in Sri Lanka [14] and in Nicaragua [10, 29]. The higher proportion of TIN found in Sri Lanka remains unexplained compared to international findings and further research is required into the causes.

Published studies from south India suggest occurrence of endemic interstitial nephropathy in areas of Tamil Nadu and Andhra Pradesh, although the report of ‘Tondaimandalam nephropathy’ is a clinical study of CKDu cases with no histopathology from renal biopsy, and ‘Uddanam Endemic Nephropathy’ is based on a pathology series of 6 cases (with raised serum creatinine) of chronic tubulointerstitial nephritis on renal biopsy [11, 22].

Strengths of the present study include the large number of renal biopsies (n = 3,095) over a ten-year period, which is the largest study of renal biopsy records in Sri Lanka. The size is comparable to large similar studies from other countries over multi-year periods (Table 1), including in Korea (n= 2053) [32], India (n= 3275) [18], and Pakistan (n= 1793)  [19]. The present study shows a statistically significant (p < 0.001) higher proportion of TIN compared to the similarly sized Korean, Indian and Pakistan studies. Other strengths include: coverage of extensive agricultural areas in central Sri Lanka where endemic CKDu has been observed; the examination of renal biopsies at Kandy Teaching Hospital allied to the Medical Faculty of the University of Peradeniya; and the high completion rate of 95% for age and sex in the data.

Inevitable limitations of studies of renal biopsies are that they capture only a part of those with renal disease since their inclusion depends upon the vagaries of: clinical presentation or selection for screening, access to nephrologists and hospital care, clinical decisions concerning the need for biopsy, and patient consent. Renal biopsy studies can also be used for detailed study of particular clinical syndromic presentations of kidney disease, such as, acute or chronic kidney failure, chronic kidney failure of unknown cause (CKDu) [27], or nephrotic syndrome [3].

In summary, the proportion of TIN was considerably higher than those reported in most published renal biopsy studies to date. The proportion of TIN in men (34%) was double that found in women (17%), a sex difference which has not been reported in other similar studies. These findings warrant further and intensive research to understand potential causes.