Clinical Features of Multiple Cutaneous and Uterine Leiomyomatosis:  An Underdiagnosed Tumor Syndrome FREE

In multiple cutaneous and uterine leiomyomatosis (MCUL) (also known as Reed syndrome, leiomyomatosis cutis et uteri, and multiple leiomyomatosis) (Online Mendelian Inheritance in Man [OMIM] 150800), affected women develop uterine leiomyomas and affected individuals of both sexes develop cutaneous leiomyomas. A disease variant involving aggressive renal cancer can occur, and MCUL is consequently also known as hereditary leiomyomatosis and renal cell carcinoma (HLRCC, OMIM 605839).^1- 2 Members or our research team have shown that patients with MCUL/HLRCC have germline mutations in a single copy of the fumarate hydratase (FH) gene.³ Fumarate hydratase is an enzyme of the Krebs cycle and plays an important role in intermediary metabolism. The mechanism by which FH mutations cause tumor predisposition is unknown. Patients with germline mutations of both copies of FH develop fumarate hydratase deficiency (FHD), (OMIM 606812), a rare, recessive, inborn error of metabolism characterized by neurologic dysfunction and survival of a few months or years but no tumors. However, members of our group have identified skin leiomyomas in parents of FH-deficient children.³FH is an unusual example of a gene in which germline mutations in one or both copies lead to distinct pathologic phenotypes.

The clinical features of the MCUL syndrome are not well characterized; many previous reports of multiple cutaneous leiomyomas have generally been restricted to a small number of isolated cases or pedigrees.^1,4- 10 As well as coexisting uterine leiomyosarcoma and renal carcinoma, single cases of MCUL have been reported in association with multiple endocrine neoplasia type I,⁵ rheumatoid arthritis,¹⁰ breast cancer, and prostate cancer.² For the present study, we identified a large number of patients with multiple skin leiomyomas, assessed them for family history of leiomyomas, and screened them for underlying FH mutation. Family members were also evaluated. We determined the clinicopathologic features of their skin and uterine leiomyomas and noted coexisting medical conditions. Since some patients with MCUL have in vitro FH activity nearly as low as that of some patients with FHD,¹¹ we assessed patients with MCUL for features of FHD. We also estimated the penetrance of FH mutations in predisposing to skin and uterine leiomyomas.

We identified study subjects by sending a circular letter to dermatologists in the United Kingdom requesting the participation of their patients with multiple skin leiomyomas. All subjects were examined for the presence of skin leiomyomas by a geneticist or dermatologist. Patients were additionally considered to have uterine leiomyomas if these had been previously diagnosed by pelvic ultrasound, hysteroscopy, or histopathologic examination after myomectomy or hysterectomy. We did not evaluate patients for the presence of subclinical fibroids.

Patients’ clinicopathologic features were determined from their medical records and from questionnaires concerning family history, specific features of skin and uterine leiomyomas, and other medical conditions. Seventy-five (69%) of the 108 affected individuals completed questionnaires providing detailed clinical information. Hematoxylin-eosin–stained sections of leiomyomas were reviewed by a gynecopathologist (S.M.) or a dermatopathologist (E.C.), as appropriate. Probands were screened for FH mutations and in some cases for reduced germline fumarase activity.^3,11 Data were analyzed using regression analysis of log-transformed outcome variables and nonparametric tests (Stata 7.0; StataCorp LP, College Station, Tex). The study was performed with full informed consent and ethical review board approval.

Forty-six white probands with skin leiomyomas were recruited to the study, 41 (89%) of whom were from the United Kingdom, and the remainder from France, Spain, and Turkey (Table 1). Twenty-three probands (50%) had a family history of both skin and uterine leiomyomas; 5 (11%) had a family history of uterine leiomyomas only; 5 (11%) had a family history of skin leiomyomas only; and 7 (13%) reported no known family history of leiomyomas. No family history information was available for the remaining 6 probands (15%) (Table 1).

Where sufficient data were available, all pedigrees were consistent with autosomal dominant inheritance or new mutations. Relatives were evaluated for skin and or uterine leiomyomas, and a case series of 46 probands, 62 affected relatives, and 42 apparently unaffected relatives was obtained. Of the 108 affected subjects, 31 were men with skin leiomyomas. Of the 77 affected women, 15 had only skin leiomyomas, 8 had only uterine leiomyomas, and 54 had both skin and uterine leiomyomas. Of the 42 apparently unaffected relatives, 17 were male and 25 were female. The mean age of all individuals with leiomyomas was 50.9 years (median, 48 years; range, 19-80 years).

A germline FH mutation was identified in 37 (80%) of 46 probands.¹¹ An assay for fumarase enzyme activity was carried out on 4 of the 9 probands without identified FH mutation and showed reduced FH activity in all 4 cases, suggesting underlying cryptic FH mutation. Thus, 41 (89%) of 46 probands had evidence of underlying FH mutation.

Of the 7 individuals who denied any family history of leiomyomas, 5 had evidence of an FH mutation. The other 2 patients were women with intact FH activity, only a single cluster of skin lesions, and no uterine leiomyomas. This, together with the negative family history of MCUL and intact FH activity in lymphoblastoid cell lines, suggests that these 2 (4%) of 46 may represent sporadic or mosaic cases of cutaneous leiomyomas.

We estimated penetrance in all 37 families with identified FH mutation. There were 93 FH mutation carriers (26 male and 67 female). All 26 men with FH mutations had skin leiomyomas at the time of the study. Of 67 women with identified FH mutations, 46 (69%) had both skin and uterine leiomyomas; 10 (15%) had only skin leiomyomas; 5 (7%) had only uterine leiomyomas; and 6 (9%) were apparently unaffected (ie, the mutation was nonpenetrant). It is possible that other women had undiagnosed fibroids, but we did not screen for subclinical lesions. There were 2 women from 2 different families who had fibroids but did not carry their families’ FH mutations and so were likely to represent phenocopies, coincidental sporadic cases of uterine fibroids.

The age-dependent penetrances of FH mutations in terms of skin leiomyomas, uterine leiomyomas, and hysterectomy are summarized in Table 2 and illustrated in Figure 1, Figure 2, and Figure 3. These may be used to estimate risk in patients with FH mutations. For skin lesions, penetrance estimates were based on data from nonprobands (since all probands by definition had skin lesions), but for fibroids and hysterectomy, all female gene carriers were included in the analysis. The risk of skin leiomyomas was higher in men than in women (100% in men and 55% in women by age 35 years). The mutation penetrance with respect to clinically apparent fibroids was high at a relatively early age (72% diagnosed with uterine fibroids by age 35 years), even making allowance for the common occurrence of fibroids in the general population. The severity of the uterine disease is illustrated by the high risk of hysterectomy (39% of all female FH mutation carriers by age 45 years) (Table 2).

The clinical features of skin leiomyomas are summarized in Table 3. Patients had a mean of 25 lesions each (median, 20; range, 1-150), and despite higher penetrance for skin disease among men, women overall had significantly more lesions (mean, 27.8; median, 24.5; range, 2-150 for women vs mean, 19.7; median, 10.0; range, 1-100 for men) (P<.007, Wilcoxon test). The distribution of skin lesions showed approximately equal numbers of patients with clustered leiomyomas only, scattered leiomyomas only, and a combination of clustered and scattered lesions. Clustered lesions were most common on the trunk, followed by the lower limb(s), upper limb(s), and head and neck, in that order. Scattered lesions were most often found on the upper limb(s), followed by the trunk, lower limb(s), and head and neck. A small proportion of patients had symmetrically distributed or unilaterally distributed lesions.

Although not measured formally, most lesions were between 2 and 20 mm in diameter and varied greatly in appearance from ill-defined, flesh-colored papules to well-circumscribed, reddish-purple nodules. Thus, while some lesions were barely visible, others were highly disfiguring (Figure 4). Skin leiomyomas were reported to develop at a mean age of 24.1 years (median, 25 years; range, 9-45 years), although the mean ages of symptom onset and diagnosis were 31.4 and 36.6 years, respectively. A minority of female subjects reported onset of skin leiomyomas after a hormonal change, for example puberty, pregnancy, or starting an oral contraceptive regimen.

An unusual and diagnostically helpful feature of leiomyomas was that they were characteristically painful: 73 subjects (89%) reported that 1 or more of their skin leiomyomas were painful. Painful lesions were not otherwise obviously different from nonpainful lesions. Most patients reported several stimuli that caused skin lesions to be painful: cold or trauma, light touch, and heat, in decreasing order of reported frequency. The nature of pain was most commonly described as sharp and shooting, which suggests neuropathic origin. Despite this, apart from some patients with near constant pain, 60 patients (73%) reported that their skin leiomyomas had no impact on their quality of life. Thirteen subjects (16%) reported that skin lesions were itchy.

Seventeen (71%) of 24 nonprobands were newly diagnosed by us during this study, suggesting that some cases of skin leiomyoma are never reported. All lesions required biopsy for diagnosis, and the most common prebiopsy differential diagnoses included dermatofibroma, sebaceous cyst, neurofibroma, and intradermal nevus (data not shown). Most patients did not receive any treatment for their skin leiomyomas. Surgical excision of lesions relieved pain. Surgical excision and skin grafting of severely affected areas were occasionally undertaken with moderate cosmetic success. Laser treatment was not generally reported to be of significant help in management of pain or cosmesis. Two patients reported good pain relief with nifedipine treatment.

The clinical features of fibroids are summarized in Table 4. The mean age at diagnosis of fibroids was 31 years (median, 30 years; range, 19-47 years), although most women had experienced gradually worsening gynecologic symptoms for several years, often since their middle teens or early 20s, before seeking medical advice. The most common symptom of uterine leiomyomas was menorrhagia, menstrual periods typically lasting 2 to 3 weeks. In 12 women (28%), gynecologists had ascribed reduced fertility—either difficulty achieving conception (7/43, 16%) or the occurrence of miscarriages (5/43, 12%)—to fibroids. These women reported that they would have requested screening for uterine fibroids if they had known that they were at risk. All of these women reported a severe negative impact of fibroids on their quality of life.

Ten (24%) of 41 individuals underwent myomectomy, as opposed to hysterectomy, in an attempt to preserve fertility, at a mean age of 28.1 years. Eight (80%) of these 10 subsequently underwent hysterectomy a mean of 9.1 years later. Sixty-five percent (n = 27) of all women with uterine leiomyomas underwent hysterectomy at a mean age of 35.3 years (median, 34 years; range, 27-47 years). Eighty-one percent (n = 33) of women with uterine leiomyomas reported either a moderate or a severe negative effect of uterine fibroids on quality of life. Women who had difficulty achieving conception and/or who experienced miscarriages and those who were housebound or unable to work during menstruation were most likely to report a severe negative effect on quality of life.

At the time of the study, 54 (87%) of 62 women with fibroids had skin leiomyomas. Of 69 women with skin leiomyomas, 54 (78%) had symptomatic uterine leiomyomas. Skin leiomyomas typically preceded diagnosis of uterine leiomyomas by a mean of 6.8 years (median, 7.5 years; range, −15 to 24) years (Table 4). Four women had no gynecologic symptoms and had been diagnosed with uterine fibroids incidentally after pelvic masses were noted during routine medical examination. Two other women with skin leiomyomas and long-standing gynecologic symptoms learned of the association through the present study, requested evaluation by their own physicians, and were diagnosed as having uterine leiomyomas. Thus the presence of skin leiomyomas was a clinically useful sign of predisposition to uterine fibroids.

Seven individuals from 3 Finnish families had previously been described² as having the HLRCC form of MCUL syndrome. None of our patients developed a leiomyosarcoma. We identified just 1 individual, the deceased daughter of proband 766, who presented with metastatic renal cell carcinoma at age 16 years and died at age 18 years. The histologic characteristics of this renal cancer were those of a collecting duct tumor and therefore different from the type II papillary renal cancer described in the Finnish families.¹¹ Although the woman with collecting duct renal carcinoma was not known to have leiomyomas, analysis of her renal tumor showed that she was an FH mutation carrier, and allele loss suggested that her carcinoma was related to FH mutation and not coincidental.¹¹ The Finnish families had been identified on the basis of type II papillary renal cancer in conjunction with skin or uterine leiomyomas, whereas the families in our study were identified on the basis of skin leiomyomas. Thus, renal cancer appears to have an uncommon association with skin leiomyoma. There were no obvious similarities between the underlying FH mutations in the British and Finnish families with renal cancer (N318K, 2–tase pair del codon 181, and R300X).¹¹

Although no tumors were obviously overrepresented in our families with MCUL, the example of renal cancer suggests that even a single case, particularly of an unusual condition, might represent a true association. In our largest family (304 members), FH mutation carriers with tumors included a woman with transitional cell carcinoma of the bladder at age 65 years, a man with an inoperable malignant brain tumor (no biopsy performed), a man with multiple lipomata, and a man with multiple lipomata and angiolipomata. In family 765.1, there were 2 female FH mutation carriers with benign breast tumors; precise histologic findings for these 2 patients were not available. Other families, including those with germline FH mutations identical to those of families 304 and 765.1 (mutations N64T and E312K, respectively), did not have any of these tumors. There were no other unusual or obviously overrepresented medical conditions in any of our patients. None of our patients was known to have a leiomyoma of any site other than the skin and uterus. There were no known cases of FHD or infant or childhood deaths. Several miscarriages in different families had been attributed to uterine fibroids. Genetic or metabolic assessment of the terminated fetuses had not been carried out. There was no indication of neurologic abnormality, developmental delay, or premature deaths even in individuals with constitutional in vitro FH activity comparable to that of FH-deficient patients.

Phenotypic variables (age of presentation, number of skin lesions, penetrance, and presence of both cutaneous and uterine disease in women) were examined for associations with FH genotype with respect to missense vs truncating changes; position of missense mutation in the gene; common changes (N64T, G354R) vs others; and FH in vitro function in the proband. We found that of 5 women with only uterine fibroids, 3 were from different families with the G354R mutation; in all, 3 of 9 women with the G354R mutation had only uterine fibroids compared with 2 of 44 other women (P = .03, Fisher exact test). No other significant genotype-phenotype correlations were found. There was no significant association between occurrence of fibroids and parity or age at first delivery (details not shown). The significantly higher numbers of skin leiomyomas in women than in men suggests a possible role for hormonal factors.

Seventeen (71%) of 24 cases of skin leiomyomas in probands’ relatives had not previously been reported (Table 3). In most cases, individuals had noticed the lesions but had not sought medical advice because the lesions were long-standing and only mildly symptomatic. A proportion of affected individuals had been unaware of their skin lesions. It is likely, therefore, that our probands had relatively severe skin disease, which caused them to seek medical attention, and that many mild, unreported cases of skin leiomyomas may exist in the population.

No patient who had presented with skin leiomyomas had been offered screening for uterine fibroids, probably because even if the association with uterine fibroids had been recognized, the strength of the association and the severity of uterine fibroids were not known. Similarly, most families were not further evaluated because of the paucity of information on the inheritance, penetrance, and molecular basis of MCUL and the absence until recently of a diagnostic test to identify affected individuals. With uterine fibroids, the situation was different. Here, most patients had presented for treatment with severely symptomatic uterine fibroids, but the syndrome had not been recognized because of little clinical awareness of the condition and because uterine leiomyomas, unlike skin leiomyomas, are commonly encountered lesions that are usually nonsyndromic.

Histologic examination showed that all cutaneous leiomyomas were pilar lesions occurring superficially in the dermis They were thought to originate from the pilo arrector muscles of the hair follicle (Figure 5) and were made up of bundles of well-differentiated smooth muscle cells in the reticular dermis. Most were diffuse lesions, whereas some were relatively well circumscribed or nodular (Figure 4 and Figure 5). In patients observed prospectively, uterine fibroids removed at hysterectomy were unusually numerous and/or large compared with sporadic cases (Figure 6). Retrospective analysis showed that details of numbers or size of fibroids were rarely recorded systematically in patients’ case notes. The microscopic appearance of uterine leiomyomas did not differ obviously from that of sporadic lesions (Figure 6).

Almost all patients who present with multiple skin leiomyomas have evidence of a germline FH mutation and hence can be diagnosed as having MCUL/HLRCC. The presence of an underlying FH mutation places an individual at a high risk of developing skin leiomyomas and, in women, early-onset fibroids that require hysterectomy. Leiomyosarcomas appear to be extremely rare in MCUL/HLRCC, having been absent from our own patients and from the case series of Toro et al¹² and Martinez-Mir et al.¹³ It is uncommon for patients with MCUL/HLRCC to have type II papillary or collecting duct cancer, but cases have been reported by Launonen et al,¹⁴ Toro et al,¹² and Alam et al.¹¹ Given that these renal cancers have a high mortality rate at a young age, difficulties arise in deciding which patients to screen for renal cancer, from what age, by what means, and how frequently. In the present study, a patient presented with metastatic disease at age 16 years, and so an aggressive strategy of screening all individuals with the FH mutation from their early teens might be considered. Less aggressive options would include screening only those FH mutation carriers whose families have either a history of renal cancer or FH mutations previously associated with renal cancer. Finally, the option of no screening at all might be considered. There is a difficult balance between identifying all possible cases and minimizing the expense, worry, and risk of investigation. Some type II papillary or collecting duct renal cell carcinomas cannot be identified by ultrasound and may therefore require computed tomography or magnetic resonance imaging for adequate screening. It will also be extremely difficult to determine whether earlier detection provides an improved outcome.

Uterine fibroids are a significant health issue. They are the most common tumors in women during the reproductive years.¹⁵ They are clinically apparent in up to 25% of women, but they have been identified in up to 77% of serial hysterectomy specimens.¹⁶ In most countries, fibroids are the most frequent indication for hysterectomy in premenopausal women.¹⁷ We have shown that germline FH mutations are strongly associated with a high risk of severely symptomatic uterine fibroids that require hysterectomy. It seems likely that screening by a relatively safe, inexpensive method such as ultrasound would allow diagnosis of subclinical fibroids and thus allow affected women to make informed family planning decisions.

Several features of skin leiomyomas are helpful in clinical diagnosis of these lesions. Onset of lesions occurs in the teens, 20s, or 30s, and most skin leiomyomas are intradermal papules or nodules 2 to 20 mm in diameter with a disseminated and/or a segmental distribution. Nearly 90% of patients report lesion-associated pain, characteristically precipitated by cold or trauma. A biopsy confirming skin leiomyomas should prompt referral of the proband to a clinical geneticist. If an underlying FH mutation is identified, family members should be contacted and screened for FH mutation and for associated tumors. A few patients who have only a single group of segmental skin leiomyomas without uterine leiomyomas or family history may not have a constitutional FH mutation.

The prevalence of MCUL is not known, although 1 study identified evidence of a heterozygous FH enzyme variant in 1 of 676 individuals in the United Kingdom.^18- 19 In the present study, we diagnosed a large number of new cases of skin leiomyomas in probands’ relatives, which suggests that a proportion of skin leiomyoma cases in the population at large are never identified. Although most cases of uterine leiomyomas in the present study presented clinically, these lesions are so common in general that the first examining physician did not usually consider the diagnosis of MCUL; thus, a proportion of cases of apparently sporadic uterine fibroids might actually be cases of MCUL. Increased recognition of the association of uterine leiomyomas with renal cancer may increase accurate diagnosis of MCUL cases on the basis of the histopathologically unusual renal cancers. In the present study, the G354R mutation was significantly associated with uterine fibroids without skin leiomyomas in women, although the explanation for this putative association is unclear. The significantly higher number of skin leiomyomas in women than in men suggests a role for hormonal factors. No other genetic or environmental factors causing variable expression of disease in MCUL are currently known.

Mutations in both copies of FH cause the inborn error of metabolism FHD. The absence of any neurologic symptoms or premature mortality in patients with MCUL suggests that 1 intact copy of FH is sufficient for normal neurologic development. The absence of cases of FHD in MCUL families is unsurprising given the relatively low population frequency of the FH mutation and the absence of consanguinity in our families.

In conclusion, MCUL/HLRCC is an underrecognized condition. Increased clinical awareness is important because of the risk of severe uterine fibroids and, in some cases, aggressive renal cancer. The identification of the FH mutations underlying this condition can now allow molecular diagnosis of affected individuals and informed screening and management.

Correspondence: Ian P. M. Tomlinson, MD, Molecular and Population Genetics Laboratory, Cancer Research UK, Lincoln’s Inn Fields, London WC2A 3PX, England (ian.tomlinson@cancer.org.uk).

Accepted for Publication: November 3, 2004.

Acknowledgment: We thank the patients and families who participated in this study and our colleagues in the Department of Histopathology and in the Equipment Park, Cancer Research UK. The following clinicians have very kindly identified patients who were suitable for this study: N. P. Burrows, MD; R. Charles-Holmes, MD; L. J. Cook, MD; B. M. Daly, MD; G. P. Ford, MD; S. E. Hadfield-Jones, MD; N. Hardwick, MD; A. S. Highet, MD; M. Keefe, MD; S. P. MacDonald-Hull, MD; E. D. A. Potts, MD; M. Crone, MD; S. Wilkinson, MD; F. Camacho-Martinez, MD; S. Jablonska, MD; R. Ratnavel, MD; A. MacDonald, MD; R. J. Mann, MD; K. Grice, MD; G. Guillet, MD; M. S. Lewis-Jones, MD; H. McGrath, MD; D. C. Seukeran, MD; P. J. Morrison, MD; S. Fleming, MD; S. J. Adams, MD; P. W. Bowers, MD; H. M. Nelson, MD; E. Healy, MD; P. S. Friedmann, MD; A. C. Pembroke, MD; K. Dalziel, MD; J. Anderson, MD; P. J. August, MD; M. G. Davies, MD; R. Felix, MD; C. S. Munro, MD; M. Murdoch, MD; J. Rendall, MD.

Financial Disclosure: None.