Introduction
Glomerular disease is an important contributor to chronic kidney disease (CKD) and kidney failure (KF) in sub-Saharan Africa.(1) Significant variation in the frequency of primary glomerular disease diagnoses has been reported in South Africa. Whilst membranoproliferative glomerulonephritis (MPGN) has been reported to be the most frequently encountered lesion at Groote Schuur Hospital, Cape Town,(2) primary focal segmental glomerulosclerosis (FSGS) has been reported as the dominant glomerulopathy at the Universitas Hospital in the Free State (3) and the Chris Hani Baragwanath Academic hospital, Johannesburg;(4) unpublished data from Charlotte Maxeke Johannesburg Academic Hospital similarly suggests a significant contribution from FSGS.(5) This variation may reflect differences in the demographics of the communities served by individual institutions and the contribution of genetic risk to the development of glomerular disease; for example, a lower proportion of people of Black African ethnicity resident in Cape Town has been advanced as an explanation for the lower prevalence of FSGS reported in that community.(2) In comparison, lupus nephritis and HIV-associated nephropathy have been reported as the most common secondary causes of glomerular disease across studies.(2,4)
Helen Joseph Hospital serves an ethnically diverse community resident in the western suburbs of Johannesburg. The pattern of glomerular disease in this population has not previously been prescribed, and it is unclear as to whether the ethnic diversity of the community resident in the hospital's drainage area translates to a unique pattern of glomerular disease. To answer this question, we retrospectively reviewed all native kidney biopsies undertaken at this institution over a 5-year period.
Methods
Convenience sampling was used to identify patients for inclusion through reference to the biopsy database held by the Division of Nephrology at Helen Joseph Hospital. All patients undergoing native kidney biopsy at this institution between 1 January 2014 and 31 December 2018, with an age over 14 years and a definitive pathological diagnosis were considered for inclusion. Indications for biopsy, patient demographics (age, sex, and ethnicity), HIV infection status, estimated glomerular filtration rate (eGFR) calculated using the CKD-EPI equation, urine protein : creatine ratio (UPCR), and histopathological diagnosis were extracted from the records. Biopsies undertaken in patients living with HIV were re-interpreted in accordance with the 2018 Kidney Disease: Improving Global Outcomes (KDIGO) controversies conference.(6) Histological diagnoses were grouped into primary glomerular, secondary glomerular, tubulointerstitial nephropathy, and miscellaneous nephropathy categories following the schema used by Okpechi et al.(2) After exclusions for inadequate data, the study database was imported into Statistica version 14.0 (TIBCO Software Inc., Palo Alto, California). The overall incidences of glomerular disease were described over the course of the study period and further categorised by patient age, ethnicity, and HIV infection status. Comparison between categories was undertaken using the Pearson Chi square and Kruskall-Wallis ANOVA tests. Where appropriate logistic or linear regression modelling were used to further characterise the significance of observed associations.
Permission to undertake this study was obtained from the Human Research Ethics Committee of the University of the Witwatersrand.
Results
Two hundred and five patients were included in this study. The demographics and biochemical data at the time of biopsy of this cohort are shown in Table 1.
Demographics and baseline characteristics of the included cohort (n = 205)
| Sex | |
| Ethnicity | |
| HIV infection status | HIV positive: 68,33.2% |
| Hepatitis infection status | |
| Syphilis serology status | Positive: 6,2.9% |
| Age (years) | 38.2 ± 13.4* |
| CKD-EPI eGFR (mL/min/1.73 m2) | 55.2 ± 44.3* |
| Urine protein : creatinine ratio (g/mmol) | 0.41 [0.23–0.78]** |
| CD4 count in HIV positive patients (×106/mm3) | 273.7 ± 234.5* |
Values are *mean ± SD and **median [interquartile range]
Nephrotic range proteinuria was the most common indication for biopsy (79 biopsies, 38.5%) followed by unexplained renal dysfunction (74, 36.1%), abnormal urinalysis comprising subnephrotic proteinuria and/or non-nephritic leukocyturia or haematuria (34, 16.6%). Nephritic syndrome was an uncommon indication for biopsy in this series (18 biopsies, 8.8%).
Secondary glomerular disease comprised the majority of histological injury in this study (131 biopsies, 63.9%). Primary glomerular disease contributed 55 cases (26.8%) and tubulointerstitial diagnoses 9 cases (4.4%). Miscellaneous nephropathies comprised 10 cases (4.9%). Secondary glomerulopathies were the most common diagnoses in biopsies undertaken for unexplained renal dysfunction (71.6%), abnormal urinalysis (79.4%), and the nephritic syndrome (66.7%). The incidence of primary and secondary glomerular disease was similar in those biopsies undertaken for investigation of nephrotic range proteinuria (48.1% and 49.4% respectively). Unexplained renal dysfunction was the most common indication for biopsy in patients with tubulointerstitial diagnoses (66.7%) and was the indication for biopsy in all cases of miscellaneous nephropathy. The individual histopathological diagnoses are set out in Table 2.
Incidences of histological diagnoses (n)
| Secondary glomerular disease (n = 131) | HIV-related kidney disease (n = 53) | HIVAN | 21 |
| Lupus nephritis in the setting of HIV | 11 | ||
| Membranous nephropathy in the setting of HIV* | 8 | ||
| Mesangial hypercellularity in the setting of HIV | 6 | ||
| FSGS in the setting of HIV | 3 | ||
| Exudative GN in the setting of HIV | 2 | ||
| Minimal change disease in the setting of HIV | 1 | ||
| Lupus-like GN in the setting of HIV | 1 | ||
| Lupus nephritis (n = 43) | Proliferative lupus nephritis | 31 | |
| Non-proliferative lupus nephritis | 12 | ||
| Other (n = 35) | Hypertension | 15 | |
| Diabetic nephropathy | 10 | ||
| Amyloidosis | 3 | ||
| Idiopathic crescentic glomerulonephritis | 2 | ||
| Post-infectious glomerulonephritis | 2 | ||
| Membranous nephropathy** | 1 | ||
| Minimal change nephropathy** | 1 | ||
| Pauci-immune glomerulonephritis | 1 | ||
| Primary glomerular disease(n = 55) | FSGS | 27 | |
| Membranous nephropathy | 9 | ||
| Minimal change nephropathy | 9 | ||
| Membranoproliferative glomerulonephritis | 9 | ||
| Mesangioproliferative glomerulonephritis | 1 | ||
| Tubulointerstitial (n = 9) | Interstitial nephritis | 8 | |
| Tubulopathy in the setting of classic HIVAN | 1 | ||
| Miscellaneous (n = 10) | Acute injury NOS | 9 | |
| Advanced chronic injury NOS | 1 | ||
*Includes patients with positive HBV, HCV, and syphilis serology (n = 1 each)
**Denotes patients serologically positive for Treponema pallidum
Lupus nephritis (LN) was the most common glomerular pathology encountered, contributing 54 cases overall (26.3% of biopsies) when including cases of lupus nephritis in the setting of HIV. LN was the most frequent diagnosis in patients undergoing biopsy for investigation of abnormal urinalysis (70.6%). Proliferative LN was more common than non-proliferative LN, comprising 17 cases of LN WHO class IV (8.3% of all biopsies) and 16 cases of LN WHO class III (7.8%). Non-proliferative LN classes comprised 2 cases of LN WHO Class I, 11 cases of LN II, and 8 cases of LN V.
FSGS was the second most common glomerular pathology encountered in this series (30 cases including those in HIV positive patients, 14.6% of all biopsies). Membranous glomerulopathy (18 cases, 8.9%), minimal change nephropathy (11, 5.4%), and MPGN (9, 4.4%) were less common podocytopathies diagnosed in this cohort.
Hypertensive nephropathy was not uncommon in this series, comprising 15 cases (7.3%). This category included 5 cases of malignant hypertension and 9 cases of benign hypertensive change. The majority of hypertensive nephropathy cases (60%) underwent biopsy for unexplained renal dysfunction; nephrotic range proteinuria was the second most common indication for biopsy in these cases (26.7%). Ten patients were diagnosed with diabetic nephropathy (4.9% of all biopsies). Unexplained renal dysfunction was the indication for biopsy in 8 of these cases, and nephrotic range proteinuria was the indication in the remaining two.
Significant differences in age were observed between histological diagnoses (p < 0.001 in Kruskall Wallis ANOVA testing, Table 3); significance was lost when LN and membranous nephropathy cases were excluded from analysis (p = 0.154). In univariate linear regression modelling, increasing age was associated with a decreasing odds of diagnosis with lupus nephritis (OR 0.936, 95% CI 0.908–0.966, p < 0.001), but not with increasing odds of diagnosis of membranous nephropathy (OR 1.02 95% CI 0.986–1.058, p = 0.23).
Demographics and baseline characteristics in selected histological categories
| Sex (F : M) | Ethnicity* | Age (years)** | eGFR (mL/min/1.73 m2)** | ||||
| B | M | W | A | ||||
| Lupus nephritis | 42:12 | 22.4 | 43.8 | 42.9 | 60 | 30 [25–37] | 96 [34–131] |
| Proliferative LN | 27:6 | 12.9 | 18.8 | 35.7 | 60 | 27 [24–33] | 92 [32–129] |
| Non-proliferative LN | 15:6 | 9.4 | 25 | 7.1 | 0 | 32 [28–41] | 107 [36–131] |
| FSGS | 16:14 | 15.3 | 18.8 | 7.1 | 0 | 32.5 [26–51] | 60.5 [25–104] |
| HIVAN | 12:9 | 12.4 | 0 | 0 | 0 | 34 [30–41] | 10 [6–18] |
| Membranous nephropathy | 9:8 | 9.4 | 0 | 0 | 20 | 45 [37–47] | 95 [55–123] |
| Minimal change nephropathy | 7:4 | 5.3 | 0 | 14.3 | 0 | 34 [24–54] | 128 [84–168] |
| MPGN | 3:6 | 4.1 | 6.3 | 7.1 | 0 | 31 [24–43] | 64 [63–81] |
| Hypertensive nephropathy | 5:10 | 7.7 | 12.5 | 0 | 0 | 51 [38–57] | 10 [7–47] |
| Diabetic nephropathy | 3:7 | 4.7 | 6.3 | 7.1 | 0 | 59 [51–61] | 13.5 [11–32] |
| Interstitial nephritis | 3:5 | 4.1 | 0 | 7.1 | 0 | 44 [36–45.5] | 21 [6.5–51] |
*Percentage of all patients of a particular ethnicity (B = Black African, M = mixed ethnicity, W = white, A = Asian)
**Values are median [interquartile range]
Certain histological diagnoses appeared to demonstrate sex clustering (Table 3). LN was more frequent in females (77.8% of LN cases occurred in females compared to 22.2% in males, p < 0.001), mainly due to cases of proliferative LN being more frequent in women (24.6% of biopsies in females demonstrated proliferative LN compared to 6.5% in males, p < 0.001, Figure 3); cases of non-proliferative LN were numerically more frequent in females (13.6% versus 6.5%, p = 0.07). As a result, women in this cohort were more likely to be diagnosed with LN than men (OR 4.16, 95% CI 2.03–8.52, p < 0.001). A trend towards a higher number of diagnoses of hypertensive nephropathy in males was detected (10.9% of biopsies in this group compared to 4.6% in females), although this did not reach statistical significance (p = 0.08).
Proliferative LN was disproportionately represented as a cause of glomerular disease amongst Asian (60% of all biopsies in this group) and white patients (37.5%) (Table 3). Asian ethnicity was however not independently associated with an increased odds of diagnosis with proliferative LN (p = 0.072). In contrast, non-proliferative LN contributed a comparatively greater proportion of cases in patients of mixed (25%) or Black African (9.4%) ethnicity than was the case in white (7.1%) or Asian (0%) patients (p = 0.20). Somewhat unexpectedly, FSGS was proportionally more frequent in patients of mixed ethnicity (18.8%) than was the case amongst Black Africans (15.3%), the contribution of FSGS to glomerular disease was more significant in these groups than was the case for either white (7.1%) or Asian (0%) patients (p = 0.62). Black African patients nevertheless contributed the majority of cases of FSGS in this series (26 cases, 86.7% of all cases of FSGS), whilst 10% of FSGS cases were recorded in those of mixed ethnicity. All cases of HIV-associated nephropathy (HIVAN) in this series were diagnosed amongst patients of Black African ethnicity; the association of HIVAN with black ethnicity was statistically significant (p = 0.02). Membranous nephropathy was a more frequent diagnosis in patients of Asian ethnicity (20%) than was the case in other groups (p = 0.27); minimal change nephropathy was a more common diagnosis in white patients (14.3%).
HIVAN was the most common diagnosis in patients living with HIV (21 cases, 30.9% of all biopsies in this group, Figure 2). This study was completed prior to adoption of the 2018 KDIGO Controversies Conference on kidney disease in the setting of HIV infection which removed HIV immune-complex kidney disease (HIVICKD) as a recognised diagnostic category.(6) Reinterpretation of biopsy diagnoses in this series in accordance with the 2018 KDIGO recommended diagnostic categories results in a slight preponderance of HIV-related podocytopathies (31 cases comprising 21 cases of Classic HIVAN, 3 cases of FSGS NOS in the setting of HIV, 6 cases of mesangial hypercellularity in the setting of HIV, and 1 case of minimal change disease in the setting of HIV); in comparison, immune complex-mediated glomerular disease contributed 22 cases comprising 11 cases of lupus-like nephritis in the setting of HIV, 8 cases of membranous nephropathy in the setting of HIV, 2 cases of endocapillary and exudative glomerulonephritis in the setting of HIV, and 1 case of lupus-like nephritis in the setting of HIV). Interstitial nephritis was not an infrequent diagnosis in patients living with HIV (7 cases, 10.2% of biopsies in this group). Median CD4 count at biopsy was lowest in those patients diagnosed with HIVAN (129 × 106/mm3) and better preserved in those with immune complex-mediated HIV kidney diseases (Figure 1).
Median eGFR was generally better preserved in patients with non-proliferative LN (107 mL/min/1.73 m2), membranous nephropathy (93 mL/min/1.73 m2), and minimal change nephropathy (128 mL/min/1.73 m2) (Table 3). Substantially reduced median eGFR was observed in patients presenting with HIVAN (10 mL/min/1.73 m2), hypertensive (10 mL/min/1.73 m2) and diabetic nephropathies (13.5 mL/min/1.73 m2), acute tubular necrosis/injury (16 mL/min/1.73 m2), crescentic glomerulonephritis (15 mL/min/1.73 m2), and interstitial nephritis (21 mL/min/1.73 m2).
Discussion
This study is the first description of biopsy-proven kidney disease at Helen Joseph Hospital and as such offers insight into patterns of glomerular disease in the community served by this institution.
Reported histological patterns of kidney disease are informed by biopsy practice. Nephrotic range proteinuria was the most frequent indication for biopsy in this series (38.5%), followed by unexplained kidney dysfunction (36.1%) and abnormal urinalysis (16.6%); similar biopsy practices have been reported at Groote Schuur (2) and Baragwanath (4) hospitals. Secondary glomerular disease was more common than primary, similar to previous South African series,(2,4) mainly due to significant rates of HIV-related kidney disease and lupus in this series.
The significant contribution of lupus nephritis in the present series likely reflects a low threshold to biopsy suspected cases, as evidenced by the association of this diagnosis with the biopsy indications of unexplained renal dysfunction and abnormal urinalysis, and the substantial contribution of these indications to the total number of biopsies performed (52.7% of all biopsies). Early biopsy ameliorates risk of kidney failure in lupus through early diagnosis.(7) As a result, biopsy in patients living with lupus is generally recommended for any acute deterioration in renal function or for active urine sediment or new onset proteinuria, regardless of quantified level.(7) Widespread adoption of these recommendations may contribute to the preponderance of LN reported in this and other series.(2,4,8,9)
LN class III and IV contributed the majority of cases of lupus nephritis in this series, followed by LN class V. Similar results were reported from Groote Schuur,(10) although the Baragwanath series reported a predominance of LN class V.(4) LN class II was not uncommon in this series, which may reflect in-centre referral practices for renal biopsy. The findings of the present study are broadly consistent with previous local reports which associate LN with female sex, younger age, and non-Black African ethnicity.(10) Male sex has been associated with more severe disease and adverse renal prognosis in LN,(11,12) which may find expression in the increased incidence of proliferative LN in males reported by some studies.(10) In this study, proliferative LN was more frequently diagnosed in female patients. It has been suggested that the apparent association of LN severity with male sex may reflect gender disparities in health-seeking behaviour and treatment compliance.(13) The preponderance of female patients in the present study may be evidence of gendered health-seeking behaviour in this community and may account for the increased number of cases of proliferative LN seen in this group.
FSGS was the most common primary glomerular pathology in this series. FSGS has been reported as the dominant primary glomerular lesion in a number of African studies in which investigation of nephrotic range proteinuria was the most frequent indication to biopsy.(3,4,8,14,15) This may reflect the genetic tendency for the development of FSGS in patients of Black African ethnicity induced by APOL1 variants.(16) The higher proportion of FSGS in patients of mixed ethnicity in this study is interesting and requires further investigation. Data regarding the penetrance of APOL1 variants in persons of mixed ethnicity in the local context is poor, but it is noteworthy that African American patients remain at increased risk of FSGS due to APOL1 variant inheritance from ancestral sub-Saharan populations.(16)
HIVAN was the third most common glomerular lesion in this series after LN and FSGS and the most common glomerulopathy diagnosed in patients living with HIV. High rates of HIVAN in the present study are reflective of HIV seroprevalence rates in the broader community: at the time of this study, HIV prevalence in Gauteng was 12.5%.(17) APOL1 risk variants are strongly associated with the development of HIVAN,(18) which contributes to the clustering of this diagnosis amongst patients of Black ethnicity.
Immune complex-mediated glomerulopathies contributed a significant proportion of cases of HIV-related kidney disease in this series (41.5%). CD4 count was generally higher in patients with immune complex-mediated diagnoses than in those with HIVAN, reflecting the role of immune effector replenishment in precipitating the latter disorders.(19) Tubulointerstitial diseases were not infrequent in patients living with HIV in this cohort. Recent local data suggests similarly substantial contribution of immune complex-mediated kidney disease and increasing tubulointerstitial diagnoses on a background of declining rates of HIVAN associated with antiretroviral rollout.(20)
Membranous nephropathy was not uncommon in this series. A non-significant trend towards older patient age was associated with this diagnosis. Age-related changes in degalactosylation of immunoglobulin are believed to contribute to the increased occurrence of primary membranous nephropathy in older adults.(21)
MPGN was relatively uncommon in this series but was marginally more prevalent in patients of white or mixed ethnicity. It has been suggested that the preponderance of MPGN in Western Cape series may reflect an underlying genetic predisposition (2) or unique toxin exposure in the form of methamphetamine in the biopsy population.(22) The frequency of APOL1 variants in persons of mixed ethnicity resident in the Western Cape has been reported to be low;(23) increased representation of this population group may account for the lower contribution of FSGS and higher prevalence of MPGN in the Groote Schuur biopsy series.
Hypertensive nephropathy was not infrequent in this series and was more common in patients of Black African and mixed ethnicity, consistent with other series.(2) Hypertensive nephropathy, like diabetic nephropathy, was non-significantly more frequent amongst older patients undergoing biopsy for renal dysfunction, reflecting age-related risks for kidney disease. The contribution of hypertension and diabetes to overall kidney disease is likely under-represented in this series by institutional practices which restrict biopsy in these conditions to cases deemed atypical in presentation.
This importance of hypertension and diabetes in the prevalence of significant kidney disease finds expression by the association in this series with low median eGFR in these cases at biopsy. The presentation of HIVAN with advanced renal dysfunction has also been noted in other biopsy series of patients living with HIV.(20)
Limitations of this study include the restriction to a single centre which has resulted in a small sample size. As a result, non-Black ethnicities are under-represented in this series which limits the interpretation of ethnicity on histological patterns. In addition, described histological patterns are determined by kidney disease screening and referral patterns as well as local biopsy practice. A lack of universal access to screening in the local context, and local practice which typically eschews biopsy for isolated haematuria, may have skewed the histological patterns described towards those presenting with a nephrotic phenotype. These factors likely explain the lack of IgA nephropathy detected in this series.
Conclusions
This first description of the histological patterns of kidney disease at Helen Joseph Hospital confirms the previously reported significant contribution of lupus nephritis and HIV-related kidney disease in South Africa. FSGS is the dominant primary glomerular lesion in this community, consistent with other series from Johannesburg; however, MPGN is uncommon in this setting which is discrepant with experience from Cape Town. The findings underscore the variability in glomerular disease in the local population and emphasize the need for renal biopsy in adult populations to facilitate diagnosis. Differences observed in the types of glomerular disease reported in Johannesburg and Cape Town further illustrate the need for a South African biopsy registry. Further evaluation of the geographic variation in glomerular disease in South Africa is required.
