CASE REPORTS

To B or Not to B: Is Tylosis B Truly Benign?Two North American Genealogies

Roger H. Maillefer, M.D., and Martin P. Greydanus, M.D.Blodgett Memorial Medical Center, and Grand River Gastroenterology P.C.,St. Mary’s Hospital, Grand Rapids, Michigan

ABSTRACTTylosis is a rare, autosomal dominant syndrome presentingwith hyperkeratosis of the palms and soles of the feet. Twotypes have been identified. Late onset tylosis (type A) isreported to be associated with a high incidence of esopha-geal carcinoma, whereas early onset tylosis (type B) appearsto be a benign disorder. This distinction has significantimplications for surveillance and prognosis. We report twofamilies exhibiting early onset type B tylosis, spanning fiveand seven generations, respectively, and believe these to bethe first two extensive genealogies of tylosis type B reportedin North America. They serve to verify the features of typeB tylosis and its benign prognosis. The world literature isreviewed and clinical relevance is discussed. Recommenda-tions for follow-up of afflicted individuals are proposed.(Am J Gastroenterol 1999;94:829–834. © 1999 by Am.Coll. of Gastroenterology)

INTRODUCTION

Hyperkeratosis of the palms and soles of the feet, morecommonly known as tylosis, is a rare autosomal dominantsyndrome with full penetrance. Two types have been de-scribed: type A and type B. The disorder is of particularinterest and concern to the gastroenterologist because of thehigh incidence of esophageal squamous cell carcinoma withtylosis type A (1). Tylosis B, on the other hand, appears tobe a benign disorder. This little known distinction is signif-icant when considering the prognosis and appropriate sur-veillance of these two types. Only 10 type A tylotic gene-alogies are documented in the world literature. Onlyscattered reports of type B genealogies can be found, andnot all authors have recognized their reports as representingtype B. This, in part, arises from the fact that the dermato-logical literature does not focus on the associated risk ofesophageal cancer and does not recognize or address thetype A/B distinction. Additionally, tylosis has been used bysome as a generic term for all palmoplantar keratodermas, ofwhich 32 clinically distinct dermatological entities havebeen described (2). The genetics underlying these disordersis complex, and a cumbersome categorization by mode ofinheritance, age of onset and physical findings has beenproposed (2). This has resulted in considerable confusionwhen reviewing the literature on tylosis.

Most of these palmoplantar keratodermas are clearly dis-

tinct from the tylotic families originally described, whereasa few resemble them more closely. We find this categori-zation complicated and not clinically helpful in determiningtylosis associated either with or without esophageal malig-nancy. The original classification of type A and type Btylosis is easily understood and can be applied to nearly allof the palmoplantar keratodermas described. This usefuldistinction allows the rapid identification of individuals atrisk for esophageal carcinoma. Our two genealogies areamong the largest of their type and typify early onset type Btylosis. They serve to support and to clarify the distinctionbetween type A and type B, and confirm the apparentlybenign nature of type B tylosis. There is a notable paucity ofliterature regarding recommendations for follow-up of theseindividuals. These kindred provide sufficient data to allowus to propose appropriate surveillance of type B tylotics andthe need for specific approaches to the follow-up of thesetwo types. The world literature is reviewed to heightenawareness of the syndrome, the contrasting features of typeA and type B, and the prognostic significance of this dis-tinction.

HISTORICAL BACKGROUND

Esophageal squamous cell carcinoma is of unknown etiol-ogy, but strong evidence has pointed toward environmentalcauses including alcohol, tobacco, nitrosamines, candidal in-fection, vitamin A deficiency, and aflatoxin. Ingestion of toxicsubstances such as lye potentiates one’s cancer risk (3). Inci-dence and sex ratio vary widely among geographic regions.Despite extensive epidemiological studies, no inherited riskfactors have been found except for the association withtylosis as first described by Howel-Evans in 1957 (4).

Tylosis (keratosis palmaris et plantaris, Thost-Unna Syn-drome (5), Howel-Evans Syndrome (4), diffuse palmoplan-tar keratoderma) was first recognized by Krost (6) (1880)and further publicized by Cockayne (4) in 1933. It did notreceive widespread attention until 1954 when Clark (7) andHowel-Evans (4) (1957) identified two separate families inLiverpool, England, with tylosis. In these kindred, tyloticpatients exhibited a 50% incidence of developing esopha-geal malignancy by age 45 and a 95% incidence by age 65(1). Onset occurred as early as age 29 yr. These malignan-cies were predominantly in the distal esophagus (4, 8),whereas acquired squamous cell carcinomas are mostly

THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 94, No. 3, 1999© 1999 by Am. Coll. of Gastroenterology ISSN 0002-9270/99/$20.00Published by Elsevier Science Inc. PII S0002-9270(99)00009-X

mid-thoracic in location (3). These families were later iden-tified as having the late onset (type A) variant. Ten geneal-ogies (4, 9–15) of type A have been published worldwide.These range from two to five generations, afflicting betweentwo and 125 members. They exhibit a mean esophageal cancerincidence of 27%, varying from 12.8% in the largest familyto 50% in the smallest published clan of four. Overall incidenceis unreliable in that some family members had not yet reachedthe age of onset of malignancy. Sporadic cases of type Ahave also been identified (16–19). Type A tylosis has re-cently been linked to the keratin gene cluster on 17q23 (20).

Subsequent investigators then identified a benign variant,type B tylosis (21, 22). Published genealogies of type Btylosis are more scattered and involve from three to 35members over four to five generations (5, 6, 21–24). Thesekindred do not appear to have the associated cancer risk, andare easily identified from the malignant variant by differ-ences in the palmoplantar distribution and in the age ofonset. Genetic linkage of type B to the keratin gene clusteron 12q11-q13 has now been suggested (25). This reportdocuments the first two extensive American genealogies,spanning five and seven generations, exhibiting classic earlyonset type B tylosis (Appendices 1 and 2). No incidence ofesophageal carcinoma has been identified in these kindred,supporting and confirming the benign nature of type B tylosis.

CASE REPORT

The proband (Appendix 1), a 71-yr-old white man, wasreferred to the gastroenterology clinic for workup of spo-radic constipation and diarrhea. He denied dysphagia orabdominal pain. Initial laboratory studies showed a mildmicrocytic anemia (Hb 12.3, MCV 78). He did not smoke ordrink and was a vegetable farmer. His family history wassignificant only for a father who expired at the age of 77from colon cancer.

On examination, the proband appeared to be in good

health and was robust, consistent with his profession. Hy-perkeratosis involved the palmar and plantar surfaces of hishands and feet respectively (Fig. 1A, B). There was uniformthickness with a 0.5 cm rim of erythema (Fig. 1A). Theproximal extent of the callus was the wrist flexure and nofissures were seen. Our proband had mild gingival hyper-plasia (Fig. 2) but there was no evidence of oral leukoplakiaor other buccal lesions.

The initial workup consisted of a colonoscopy because ofhis family history of colon carcinoma, change in bowelhabits, and mild microcytic anemia. Concern about tylosisand its known association with esophageal carcinomaprompted an esophagogastroduodenoscopy (EGD) as well.Some distal esophageal inflammation was present, consis-tent with reflux esophagitis, and cytological brushings re-vealed inflammation with atypia. The patient was treatedwith omeprazole for reflux esophagitis, and follow-up EGDwas entirely benign. Because of chronic reflux with recur-

Figure 1. A andB: Diffuse palmar and plantar hyperkeratosis of Family J displaying uniform thickness, sharply demarcated edges, andthe proximal extent at the wrist flexure.

Figure 2. Gingival hyperplasia demonstrated in a member of Fam-ily J.

830 Maillefer and Greydanus AJG – Vol. 94, No. 3, 1999

rent esophagitis, some noncompliance with antireflux med-ications, and an ongoing concern about the potential devel-opment of malignancy, the patient underwent twosubsequent EGD examinations but showed no evidence ofdeveloping malignancy.

The proband’s American heritage began when his grand-mother emigrated from the Netherlands in the 1880s andsettled into farming in western Michigan. The descendantsall remain in this locale. Fifteen additional family members,spanning five generations of Family J (Appendix 1), wereidentified as having characteristic features of tylosis. Thefamily had coined this readily identified disorder as “doubleskin.” They had never sought treatment or evaluation for theaffliction and family physicians had not recognized thesyndrome or its potential significance. Family memberswere interviewed and multiple affected members were ex-amined and photographed. All affected members had uni-form involvement of hands and feet and reported an onset ofsigns and symptoms by 1 yr of age (Fig. 3A, B). Files andpumice stones were used to trim or abrade the calluses, andmost affected individuals used creams and emollients fre-quently. Some even reported using belt sanders to reduceexcessive callus. Sporadic family members had involvementof the knuckles, but all had normal nails. Some hyperkera-

toses peeled routinely every spring or fall, or if the affectedextremity had been placed in a cast, but in all cases wouldquickly regenerate. Gingival hyperplasia was only sporadicand the condition of teeth varied. No cases of esophagealmalignancy could be identified in either living or deceasedfamily members.

The second proband (Appendix 2), an 80-yr-old whitewoman, was also healthy and had a benign past medical andsocial history. She was discovered on a tip from a memberof Family J who recalled a chance meeting with this probandat some time in the past and recognized that she had asimilar disorder. The proband had no active medical com-plaints and was on no medications. She and numerousfamily members also exhibited changes of type B tylosis(Table 1), all with onset by age 1 yr (Fig. 4A, B). Usingsimilar methods, 42 additional affected family members,spanning seven generations of Family Z, were identified.Again, sporadic members had calluses of the knuckles orfriction-related calluses of the elbows (such as related toschool work). No nail changes were identified. Gingivalchanges were less consistent. Tracing Family Z to 1853when they emigrated from the Netherlands failed to identifyany premature deaths from malignancy or cases of esopha-geal carcinoma. Surprisingly, no link between these two

Figure 3. A andB: A 10-month old exhibiting ground glass appearance, yellow discoloration, and the rim of erythema.

Table 1. Comparison of Tylosis Types

Tylosis A Tylosis B

Onset 5–15 yr old (, 30 y/o) (4, 10–13) , 1 yr old (4, 10, 11)Inheritance Autosomal dominant (5) Autosomal dominant (4, 11)Penetrance Full (5) Full (11)Cancer risk Premature squamous cell esophageal cancer (9–15) No cancer risk (4–6, 21–24)Oral findings Buccal preleukoplakia and leukoplakia (9, 17, 26) Gingival hyperplasia (9, 12)Skin findings Hyperkeratosis of feet and/or hands (8–10, 13, 14) Hyperkeratosis of feet and hands (6) with sporadic

extension to knuckles and elbows (2)Hand findings Deep fissures (9, 19)

No erythema (21)Nonuniform thickness (11, 12)Weight-bearing areas (12)Blurred edges (12)

Fissures rare (4, 10, 11, 21)Ring of erythema (4, 6, 21)Uniform thickness (4, 11)Extends to wrist flexure (21)Sharply demarcated edges (4, 21)

Risk to nontylotic members None (1) None (2, 5, 6, 21, 22)

831AJG – March, 1999 Is Tylosis B Truly Benign?

families of Dutch origin could be found. Again, no medicalpersonnel had ever cautioned them about the disorder or itspossible association with malignancy.

All affected, interviewed members of both kindred de-scribed similar findings, could readily identify other affectedfamily members, and were familiar with nonliving membersand ancestors who also exhibited characteristic signs oftylosis. Although this ancestral incidence is dependent onthe memory of those questioned, it is assumed to be reliablebecause the physical findings of tylosis are so distinctive andconsistent that it is nearly inconceivable that affected indi-viduals would have gone unrecognized or been forgotten.Confirmation of involvement in deceased individuals wasobtained by cross-questioning multiple family memberswho affirmed that these persons exhibited traits consistentwith tylosis type B and an onset by age 1 yr. In neitherfamily were there any cases of esophageal malignancy, andthere was a relative paucity of malignancy of any type.

DISCUSSION

Familiarity with tylosis is important to identify the syn-drome and its associated risk of esophageal malignancy.Specific identifying features (Table 1) of early and late onsettylosis allow placement of an individual, whether sporadicor familial, into the appropriate risk group, with significantimplications for surveillance.

Type A typically presents between age 5 and 15 yr (4,10–13), with onset no later than the fourth decade. Physicalexamination reveals hyperkeratosis of the hands and/or feetand follows a weight-bearing distribution, lacking uniformthickness. There may be isolated patches confined only tothe fingertips (9), or only the feet may be involved (14). Theproximal palmar extent is not discrete. Painful, deep fissuresare frequently the presenting complaint (19). Buccal preleu-koplakia and leukoplakia are key findings with gingival

hyperplasia being absent (9, 17, 26). There appears to be apremature malignant degeneration of the distal esophagusinto squamous cell carcinoma occurring as early as age 25yr (17), with a 50% incidence by age 45 yr and as high as95% by age 65 yr (1). Cumulative review of the worldliterature suggests an average malignancy rate of 27%. Mi-croscopic examination reveals poikilocytes. Human papil-lomavirus, which exhibits similar poikilocytes, may be arisk factor for tylosis, but this has yet to be confirmed (27).Vitamin A deficiency, however, may play a role (4, 14, 22,28). Other malignancies have been reported sporadicallyand include the stomach (4), larynx, lung, tongue, tonsil,breast, and uterus (14). Prudent recommendations for theseindividuals with type A tylosis might include strict avoid-ance of environmental factors (2, 14). Because esophagealcancer may be harbored in 20% of visibly normal mucosa,annual endoscopy with multiple biopsies at various levelsseems reasonable (29). If precancerous or cancerous changesare identified, esophagectomy would be warranted. Someauthors have even recommended prophylactic esophagec-tomy by age 45 yr when qualified surgeons are available(13). Fortunately, nontylotic family members appear to be atno increased risk for the development of malignancy (8).

Type B tylosis presents by 1 yr of age (10) as a yellow,waxy, ground glass appearance and evolves with age to athick hypertrophic callus. All documented cases have uni-formly involved both the palms and the soles of the feet. Theedges are sharply demarcated and the palmar proximal ex-tent is the wrist crease. The hyperkeratosis is of uniformthickness and fissures are rare. Hyperhidrosis may or maynot be present but is more common on the feet and may beassociated with chronic dermatophyte infection. A thin rimof erythema is present (9, 21). Peeling may be related to thefrequency of friction, change in weather (21), or may occurwith high fevers. Type B may be associated with gingivaland palatal hyperplasia but preleukoplakia and leukoplakia

Figure 4. A andB: Palmoplantar keratoderma consistent with type B tylosis in Family Z.

832 Maillefer and Greydanus AJG – Vol. 94, No. 3, 1999

are notably absent (4, 12). These tylotics and their nontylotickin are at the same risk as the general population for ma-lignant change (2, 5, 6, 21, 22). (An additional single kin-dred with some features of type B tylosis has been describedbut is distinguished by clubbing, sclerodactyly, and an as-sociation with squamous cell carcinoma of the involvedskin. This subtype has been designatedsclerotylosis[23].)Our two large kindred support and establish type B tylosisas a benign disorder. No specific interventions in true typeB tylosis are necessary, and routine endoscopic evaluation isnot warranted.

In conclusion, we present two North American familieswith tylosis type B. Our findings support the contention that

there are two distinct types of tylosis, based on history andphysical examination. Our kindred confirm and solidify thecontention that type B tylosis is not associated with esoph-ageal malignancy. This previously little recognized distinc-tion has important prognostic relevance and helps to deter-mine appropriate surveillance of affected individuals. TypeA and type B tylosis are readily recognized and distin-guished by history and physical examination. If late onsettype A tylosis is identified, aggressive endoscopic surveil-lance seems warranted. The optimal interval between ex-aminations has not been determined. Patients with earlyonset type B tylosis, on the other hand, may be followedconservatively and without routine endoscopic intervention.

Appendix 1. Pedigree of Family J. (The numeric values represent the current age or age at death.)

Appendix 2. Pedigree of Family Z.

833AJG – March, 1999 Is Tylosis B Truly Benign?

Reprint requests and correspondence:Martin P. Greydanus,M.D., Grand River Gastroenterology P.C., 310 Lafayette, Suite400, Grand Rapids, MI 49503.

Received Mar. 23, 1998; accepted Sep. 18, 1998.

REFERENCES

1. Clarke CA, Howel-Evans W, McConnell RB, et al. Carcinomaof oesophagus in association with tylosis. Br Med J 1959;2:1100.

2. Stevens HP, Kelsell DP, Bryant SP, et al. Linkage of anAmerican pedigree with palmoplantar keratoderma and ma-lignancy (Palmoplantar ectodermal dysplasia type III) to17q24. Arch Dermatol 1996;132:640–51.

3. Cotran RS, Kumar V, Robbins SL. Robbins pathologic basisof disease, 4th ed. Philadelphia: WB Saunders, 1989;835–6.

4. Howell-Evans W, McConnell RB, Clarke CA. Carcinoma ofthe oesophagus with keratosis palmaris et plantaris (tylosis): Astudy of two families. QJ Med 1958;27:413–29.

5. Berth-Jones J, Hutchinson PE. A family of palmoplantar epi-dermolytic hyperkeratosis. Clin Exper Dermatol 1984;14:313–6.

6. Sukumar P, Premalatha S, Thambiah AS. Genetic tylosis inthree siblings. Indian Ped 1983;20:145–7.

7. Clarke CA, McConnell RB. Six cases of carcinoma of theoesophagus occurring in one family. Br Med J 1954;2:1137–8.

8. Clarke CA, Howel-Evans AW, McConnell RB. Carcinoma ofoesophagus associated with tylosis. Br Med J 1957;1:945.

9. Young WG, Newcomb GM, Daley TJ. Focal palmoplantar andgingival hyperkeratosis syndrome: Report of a family, withcytologic, ultrastructural, and histochemical findings. OralSurg 1982;53:473–82.

10. Marger RS, Marger D. Carcinoma of the esophagus and tylo-sis. Cancer 1993;72:17–9.

11. Shine I, Allison PR. Carcinoma of the oesophagus with tylosis.Lancet 1966;1:951–3.

12. Laskaris G, Vareltzidis A, Avgerinou G. Focal palmoplantarand oral mucosa hyperkeratosis syndrome: A report concern-ing five members of a family. Oral Surg 1980;50:250–3.

13. Goldberg HI. Case 3: Tylosis. AJR 1981;136:1026–7.14. Harper PS, Harper RMJ, Howel-Evans AW. Carcinoma of the

oesophagus with tylosis. QJ Med 1970;39:317–33.15. Ritter SB, Peterson G. Esophageal cancer, hyperkeratosis, and

oral leukoplakia: Follow-up family study. JAMA 1976;236:1844–5.

16. Schwindt WD, Bernhardt LC, Johnson SAM. Tylosis andintrathoracic neoplasms. Chest 1970;57:590–1.

17. Ritter SB, Peterson G. Esophageal cancer, hyperkeratosis, andoral leukoplakia: Occurrence in a 25-year-old woman. JAMA1976;235:1723.

18. Veneri RJ, Gordon CG. Cervical columnar epithelium andtylosis. Lancet 1989;105.

19. Fred LH, Gieser RG, Berry WR, et al. Keratosis palmaris etplantaris. Arch Intern Med 1964;113:866–71.

20. Risk JM, Field EA, Whittaker J, et al. Tylosis oesophagealcancer mapped. Nat Genet 1994;8:319–21 (letter).

21. Muir VML. Tylosis in the Orkney Islands. J Biosoc Sci 1978;10–6.

22. Anderson IF, Klintworth GK. Hypovitaminosis-A in a familywith tylosis and clinodactyly. Br Med J 1961;1:1293–7.

23. Yesudian P, Premalatha S, Thombiah AS. Genetic tylosis withmalignancy: A study of a south Indian pedigree. Br J Dermatol1980;102:597–600.

24. Camisa C, Williams H. Epidermolytic variant of hereditarypalmoplantar keratoderma. Br J Dermatol 1985;112:221–5.

25. Lind L, Lundstrom A, Hofer PA, et al. The gene for diffusepalmoplantar keratoderma of the type found in northern Swe-den is localized to chromosome 12q11-q13. Hum Mol Genet1994;3:1789–93.

26. Tyldesley WR, Hughes RO. Tylosis, leukoplakia and oesoph-ageal carcinoma. Br Med J 1973;4:427.

27. Ashworth MT, McDicken IW, Southern SA, et al. Humanpapillomavirus in squamous cell carcinoma of the oesophagusassociated with tylosis. Clin Pathol 1993;46:573–5.

28. O’Mahoney MY, Hellier M, Ellis JP, et al. Familial tylosis andcarcinoma of the oesophagus. JR Soc Med 1984;77:514–7.

29. Lightdale CJ, Winawer SJ. Screening diagnosis and staging ofesophageal cancer. Semin Oncol 1984;11:105.

834 Maillefer and Greydanus AJG – Vol. 94, No. 3, 1999