Dr Finbar J. O’Callaghan, Department of Paediatric Neurology, Bristol Royal Hospital for Children, Level 6 UBHT Education Centre, Upper Maudlin Street, Bristol BS2 8AE, UK. e-mail: email@example.com
To report the frequency of renal symptoms and complications of patients with tuberous sclerosis complex (TSC), to describe the ultrasonographic appearance of the kidneys in a population-based sample, and to investigate the relationship between a history of renal haemorrhage and renal lesions identified by ultrasonography.
PATIENTS AND METHODS
As part of an epidemiological study, 179 patients with TSC were identified as living in the Wessex Region in the South of England. Patients were interviewed and examined in their homes, to elicit the presence of renal symptoms or a history of renal complications. Renal ultrasonography was used in consenting patients in their homes.
There was a history of renal complications in 16 (9%) patients; 149 consented to interview and examination, and 19 gave a history of renal symptoms in the previous year; 124 had renal ultrasonography; 86 (69%) had renal angiomyolipomas and 37 (30%) had renal cysts. Large lesions (>3 cm in diameter) were strongly associated with a history of symptomatic bleeding, although significant haemorrhage occurred in a 6-year-old child with small angiomyolipomas.
The formation of angiomyolipoma in TSC is common. Polycystic kidney disease, renal carcinoma and renal failure, although rare, occur in TSC. Most patients with angiomyolipomas have neither complications nor symptoms. There was no appreciable difference between the sexes in the risk of developing these lesions. Although less commonly seen in the very young, there is no identifiable relationship after adolescence between age and the risk of having a renal angiomyolipoma. Bleeding tends to occur from large lesions (>3 cm) but most such patients have remained asymptomatic to date.
Renal involvement in tuberous sclerosis complex (TSC) is common and potentially serious. One study suggests that it is second only to CNS complications as a cause of mortality in these patients . Angiomyolipomas (AMLs) and cysts are the two characteristic types of renal lesion in TSC. Polycystic kidney disease may occur and although uncommon is an important cause of morbidity and mortality. There is also evidence that renal carcinoma is more common in these patients than in the general population [2,3]. Despite the seriousness of these complications there have been no large population-based studies of the prevalence of renal lesions in TSC. A recent large-scale note review of 139 patients with TSC who had undergone renal ultrasonography (US)  suggests that 49% of patients with TSC will have AMLs and 32% will have cysts. Selection bias may have distorted these results, but similar figures have been obtained from other studies based on selected patients . Even higher prevalence rates for renal involvement (67–100%) have been reported from postmortem studies. One population-based study assessed the prevalence of renal lesions in TSC; Webb et al. found, using US, that in Bath (UK), 34% of the population of patients with TSC had renal AMLs and 14% had cysts. However, there were only 21 patients and the CIs about these proportions are wide. One study in particular reported a strong positive correlation between the prevalence of AMLs and age, with few appearing in young children , but others have found no association [5,7]. Some investigators have reported an increased prevalence of AMLs in females , but others have shown no sex bias . Two papers reported that AMLs develop at an earlier age in females [4,6]. There is no apparent correlation between the prevalence of cysts and either age or sex, but polycystic kidney disease occurs early in life in both sexes.
The lack of population-based studies also means that there are few reliable data on the frequency of symptoms and complications from kidney hamartomas. Cook et al. showed that 13 of 139 patients (9.4%) had a history of symptoms attributable to their renal disease, most commonly frank haematuria from bleeding AMLs, and three patients (2.2%) had renal carcinoma. Webb et al. identified the cumulative incidence of renal complications in their large population-based study (in the same area of Wessex as the present study) and found that eight of 131 patients (6%) had a history of renal complications related to TSC, two had polycystic kidney disease, and six had haemorrhages from renal AMLs.
Understandably there is much interest in the relationship between renal lesions and both symptoms and complications. There have been several studies of clinic-based patients; some have investigated patients with AMLs both with and with no underlying TSC , and van Baal et al. published a longitudinal study of 23 patients with AMLs and TSC. There is agreement that there is a relationship between AML size and its tendency to bleed. Lesions of >3.5–4 cm in diameter are more likely to bleed than smaller lesions . Van Baal et al. concluded that clinicians should intervene aggressively whenever they identify a patient with TSC who has a large AML. However, a potential problem is that previous studies reported on selected patients and the conclusions may not apply generally to patients with TSC. Moreover, evidence suggests that many individuals with large AMLs remain asymptomatic , but some with small AMLs develop symptoms .
The aim of this study was to establish the prevalence of renal AMLs and renal cysts, and to examine the relationship between the US appearance of the kidneys and symptoms in a population-based study of patients with TSC.
PATIENTS AND METHODS
As part of a large epidemiological study we identified 179 patients with TSC living in the Wessex region of the South of England on 1 August 1998. The population denominator was 3 679 172 and therefore the point prevalence was 4.9 per 100 000 population (95% CI 4.2–5.6). Thirty patients declined to be seen by the investigator; in these cases the relevant GP was interviewed and information collected about any history of renal complications. A renal complication was defined as any medical event or development attributable to a TSC lesion that required medical intervention; 149 individuals agreed to be interviewed and examined in their homes. The first author conducted all interviews, examinations and US. The interview included questions about any previous history of renal complications, renal surgery and renal investigations, and questions about any history of frank haematuria or flank pain in the preceding year. Renal US was offered to all patients; 124 agreed and were able to undergo this examination, using a portable machine (Kapasee, Toshiba, Japan) with a phased array scanner (3–5 mHz depending on size of subject). Longitudinal and transverse images were acquired. Scans were reported immediately and a copy of each stored on disk.
An AML was defined as an identifiable hyperechoic lesion (equal to that of the renal sinus fat) within the renal parenchyma. A renal cyst was defined as a completely anechoic, sharply circumscribed, smooth-walled lesion. Longitudinal and transverse diameters of lesions were measured using electronic callipers. The relationships between lesions seen by US and symptoms elicited by patient interview were explored. Proportions were compared using the chi-square or Fisher's exact tests where appropriate.
Sixteen (9%) of the 179 patients had a history of renal complications (Table 1). One patient had a renal carcinoma that had required a nephrectomy (Fig. 1). Nineteen (13%) of the 149 patients who were interviewed and examined gave a positive history of renal symptoms (Table 2). There were 67 males and 57 females in the sample of 124 that had US, and the median (range) age was 25 (0.25–86) years; 67 (54%) had learning difficulties. AMLs were found in 86 patients (69%; Fig. 2); 13 (15%) were unilateral and 73 (85%) bilateral. A slightly higher proportion of females (74%) than males (66%) had lesions, but the difference was not significant. When the sample was divided into quintiles according to age there was no significant relationship between age and prevalence of AMLs (Table 3), although the youngest group had a slightly lower prevalence of lesions. In the youngest group, seven males and six females had AMLs. Renal cysts were found in 37 patients (30%), six patients had cysts alone, and 31 had both cysts and AMLs. Two patients had polycystic kidneys (Fig. 3). The prevalence of renal cysts was not affected by age or sex.
Table 1. The 16 patients with renal complications attributable to TSC
Bleeding AML/embolization/renal abscess
Table 2. The 19 patients with a history of renal symptoms in the previous year
All patients had AMLs except this patient, for whom the underlying pathology was unknown.
There was a significant relationship between the size of the largest AML and a history of symptomatic bleeding (P < 0.001, Fisher's exact test, Table 4). The risk of bleeding appeared to increase markedly with lesions of ≥ 3 cm in diameter. No patients had a history of renal bleeding with no US evidence of an AML, and only one patient with a lesion of <3.0 cm had a history of haemorrhage. However, 16 patients with lesions of ≥ 3.0 cm had no history of haemorrhage.
Table 4. Relationship between the size of AMLs and history of renal haemorrhage
Largest AML, cm
History of haemorrhage
P < 0.001, Fisher's exact test.
The main strength of this study is that it was based on a population of patients with TSC rather than a selected sample. The sample that had US may not be representative of the population as a whole, but the sex and age distribution of this group is very similar to that in the whole population of the study. However, in the group examined by US, a slightly higher proportion had learning difficulties (54%) than in the population of the present study as a whole (43%). It was previously suggested that patients with learning difficulties are at greater risk of AMLs than patients of normal intellect , so the effect of the increased proportion of patients with learning difficulties would be to overestimate the prevalence of AMLs.
This study was also unusual in that US was used in the patients’ homes rather than in a hospital. During the study 42 patients also underwent US at the Royal United Hospital in Bath, for clinical reasons, and the concordance in results from research and hospital scans was substantial, with a κ statistic for interobserver agreement of 0.7 (data available on request). Theoretically, US should detect lesions of >0.2 cm in diameter, but two previous studies have suggested that AMLs need to be 0.4–0.5 cm in diameter to be detected reliably [13,14]. Therefore, very small AMLs will be missed by US, and the prevalence as determined by the present study will underestimate the true rate. The limitations of US probably explain the different prevalences reported by clinical and postmortem studies. Interestingly, the histology report on the kidney of the patient who had a nephrectomy for renal carcinoma mentions the presence of multiple AMLs that had not been identified on previous US or CT. Despite the limitations of US, the prevalence of AMLs in the present study was higher than in many previous studies, although Ewalt et al. found a similar prevalence in their group of 60 children. However, as the present is the only large population-based study of renal involvement in patients with TSC to date, comparisons with studies on highly selected groups may be inappropriate.
Most patients with AMLs had neither symptoms nor complications, but symptoms may have been under-reported because of the problems of eliciting accurate histories from patients with learning difficulties, and from their carers. The study confirms the finding of many others, that large AMLs (>3 cm) are more likely to be associated with a history of bleeding than small ones, but there was one case of a small AML in a 6-year-old child bleeding significantly, and therefore clinicians should be wary of ignoring the possibility of either small lesions bleeding, or young patients having particular problems. They should also be careful of ‘intervening aggressively’, as van Baal et al. recommend, when they detect large AMLs. Most patients in this study with large lesions, some of whom were 50–70 years old, had been asymptomatic. However, these large asymptomatic lesions (or, indeed, smaller lesions) might cause problems in the future, and a longitudinal study of a population-based sample of patients is needed to establish this.
The ‘two-hit’ hypothesis of hamartoma formation in TSC predicts that AMLs will be more common with increasing age; according to this hypothesis, the older the individual, the more time he/she has had to acquire a somatic mutation in the ‘good’ copy of their TSC gene. Cook et al. reported a positive correlation between the prevalence of renal AMLs and age, but although we found AMLs to be less prevalent in the pre-adolescent group, there was no continued positive relationship between age and AMLs in older groups. It is possible that ‘second hits’ in the kidney occur early in life, but the lesions may be too small to detect in the very young. However, by adolescence they can be seen by US more easily, and thereafter the prevalence does not change with age. If so, then an individual who has not developed an AML by late adolescence is unlikely to do so in the future.
Both Webb et al. and van Baal et al. found a higher prevalence of AMLs in females. Although a slightly higher proportion of women in the present study had AMLs than men, the difference was not significant. Unlike these two earlier studies, we found no sex difference in young patients who developed AMLs. There was no significant difference between the sexes in the proportion of people with lesions who had reported symptoms, but the increased proportion of females with complications was almost significant (P = 0.054, Fisher's exact test). Previous investigators may have found a higher prevalence of AMLs in female patients with TSC because of ascertainment bias; for a long time it has been known that sporadic AMLs (i.e. no underlying diagnosis of TSC) develop almost exclusively in females, and so clinicians may have looked harder for these lesions in female patients with TSC. Retrospective studies of the prevalence of renal complications in TSC that relied on previously collected clinic data would be likely to reflect this bias in their sex-specific prevalence rates.
Only one patient in the present study was known to be in advanced renal failure, a 62-year-old woman who had severe learning difficulty. Her renal failure had been documented since 1991 but there had been no attempt to identify the underlying cause. Her indices of renal function in August 1998 were a urea level of 40.6 mmol/L and a creatinine level of 558 µmol/L. She died shortly after the census date, before her scheduled research US. Unfortunately, she did not undergo postmortem examination and it is not clear whether she died of renal failure or another cause. Renal failure is rare in TSC but has been documented ; several patients with TSC have successfully undergone renal transplantation [15,16]. The prevalence of renal failure in patients with TSC of normal intellect has been estimated at 1%, but its prevalence in patients with learning difficulties is unknown. It occurs as a consequence of nephron replacement by AMLs, of treatment of AMLs, and from small dysplastic kidneys, but most commonly from polycystic disease . The risk of renal failure is greater if the patient has undergone nephrectomy, and seven patients in the present study had done so for renal haemorrhage. Sometimes nephrectomy is necessary to prevent loss of life but, in general, it is sensible to adopt a conservative approach in TSC. It is likely that both kidneys will be involved and so every attempt should be made to preserve functioning renal tissue. One patient in the present study had a renal haemorrhage which was her first presentation with TSC, and she had a nephrectomy before a diagnosis was made, and without considering a more conservative approach. She has now developed problems in her remaining kidney, requiring selective arterial embolization. Although her renal function was normal, she is at high risk of renal failure in the future. The preferred treatment for renal haemorrhage from AMLs in TSC is selective arterial embolization, because it preserves the most renal tissue. This treatment had been used in four of the present patients and in all four it stopped the renal bleeding; one patient had remained asymptomatic for 11 years. However, the procedure is associated with morbidity, and all patients had ischaemic pain after embolization.
Only one patient in the present study had been diagnosed with renal carcinoma. She was 54 years old at the time of diagnosis, which was made opportunistically when a hypoechoic lesion was noticed during research US; she had no symptoms. CT and renal biopsy confirmed the diagnosis; she underwent nephrectomy, and histology confirmed the presence of RCC that stained negatively for the melanocyte marker HMB-45. She had had no tumour recurrence in the 4 years after removal of the primary lesion.
The prevalence of renal cysts in patients with TSC was similar to that found in previous studies. Cysts rarely cause symptoms but occasionally they may cause renal failure, either in conjunction with AMLs or when they occur as part of the polycystic kidney phenotype. Patients with TSC who have polycystic kidney disease have a contiguous deletion of the APKD1 gene and TSC2 gene on chromosome 16 . As in previous studies, most patients who had cysts also had AMLs. Four patients had polycystic kidneys, and in one of these the contiguous deletion on chromosome 16 was identified.
In conclusion, AML formation in TSC is common; although less commonly seen in the very young, there is no identifiable relationship after adolescence between age and the risk of having an AML. Most lesions cause neither complications nor symptoms. Bleeding tends to occur from large lesions (>3 cm) but most patients with large lesions in the present study have remained asymptomatic to date.
The authors thank the Wellcome Trust, Bath Unit for Research in Paediatrics, The Tuberous Sclerosis Association and the Medical Research Council for their help. Dr F.J. O’Callaghan was supported by a Wellcome Trust Fellowship in Clinical Epidemiology for the duration of this study.