1330

Phenotype and genotype heterogeneity in autosomal dominant polycystic kidney disease

It is

clear that mutations of at least two genetic loci can lead to autosomal dominant polycystic kidney disease (ADPKD). We have compared the clinical features of ADPKD caused by mutations at the PKD1 locus (linked to the &agr;-globin complex on chromosome 16) with those of disease not linked to the locus (non-PKD1). now

We identified 18 families

(285 affected members)

with mutations at PKD1 and 5 families (49 affected individuals) in which involvement of this locus could be dismissed. Non-PKD1 patients lived longer than PKD1 patients (median survival 71·5 vs 56·0 years), had a lower risk of progressing to renal failure (odds ratio 0·35, 95% Cl 0·13-0·92), were less likely to have hypertension (odds ratio adjusted for age and

family of origin 0·29, 0·11-0·80), were diagnosed at an older age (median 69·1 vs 44·8 years), and had fewer renal cysts at the time of diagnosis. Although most of the PKD1 families were ascertained through clinics treating patients with renal impairment, no non-PKD1 family was identified through this source. Non-PKD1 ADPKD has a much milder phenotype than that linked to PKD1. Partly as a result of this difference in severity, the reported prevalence of this genotype is probably an underestimate. Introduction Autosomal dominant polycystic kidney disease (ADPKD) is the underlying cause of renal failure in 8-10% of patients with end-stage kidney disease in Europe and Australia.1,2 Mutations of at least two genetic loci can lead to this disorder. The locus more commonly involved (PKD1) has been mapped to the distal end of the short arm of chromosome 16 (16pl3), just proximal to the ot-globin locus.3 The availability of many DNA probes and some oligonucleotide sequences closely linked to the PKD11 locus4,senables identification of restriction fragment length polymorphisms (RFLPs) and AC repeat length polymorphisms, which can be used to track a PKD11 mutation from parent to child. The identification of a large kindred in which ADPKD and PKD1-linked RFLPs were inherited independently provided the first evidence of a second ADPKD locus.6 Other families in which there was no linkage of ADPKD to the PKD1locus7,8 confirmed the existence of another, non-PKD 1, locus. There are still too few data for the prevalence of non-PKDto be established reliably. Heterogeneity testing in 42 North American ADPKD families produced an estimate of 4-1%,9 whereas the estimate for 273 European families was 15%.10 The limited clinical data available suggest that the non-PKD1 phenotype is less severe than that of PKD1." A

study of ADPKD in Melbourne, set up to assess the sensitivity and specificity of ultrasound diagnosis in comparison with gene tracking through linkage studies, provided an opportunity to compare the clinical features of PKDand non-PKD 1. We also studied the effect of ascertainment source on the prevalence estimates of the PKD1 and non-PKDgenotypes.

Patients and methods Families were tested for evidence of linkage or non-linkage to DNA probes identifying polymorphic sites close to the PKDI locus on 16pl3; 26-6, 218EP6, VK5B, and 24-1 on the centromeric sidel2-15and 65CMMb and 3’HVR, which identify polymorphisms distal to PKD 1.16,1’ Later in the study 16AC2.5, a polymorphic AC repeating sequence, was used in 5 families. The 16AC2.5 locus is between 65CCMb and VK5b, but its position relative to PKD1 is unknown.18 Lod scores were calculated by pairwise analysis with the Mlink program from the Linkage program package (version 5.2).’ Bayes’ theorem was used to combine the Lod score obtained with the prior probability (C’t) that the family had PKD11 (a = the estimated proportion of families in which the mutation causing ADPKD is linked to the PKD 1 locus).20,21 A conservative N value of 0-7 was used, representing a point just beyond the 95% lower confidence limit of the European estimated" This conservative estimate was used because it would increase the stringency for identifying PKD1families. Families with probability of linkage greater than 0-95 were accepted as having PKD 1. Simple inspection of pedigree linkage results proved satisfactory for identifying the random assortment of unlinked loci in non-PKDfamilies. In PKD1-linked families, individuals born with a 1 in 2 risk of PKD 1 and previously undiagnosed were further tested with the linked markers. When closely linked DNA markers flanking the PKD1 locus were informative, the inheritance of a PKD1 allele could be inferred with an error risk, due to double recombination, of much less than 1%. When only non-flanking probes were informative, and in other complex settings, risk estimates were obtained with the Mlink and Unknown programs from Linkage.22 52 families with members who had ADPKD were identified through the renal clinic of the Royal Melbourne Hospital, nephrologists in private practice, a paediatric nephrologist, the genetics clinic of the Victorian Clinical Genetics Service, and other public hospital renal clinics. No linkage information was available in these families before they joined the study. After the family history had been recorded from each index patient, other known affected family members and those at 50% risk were approached and asked to cooperate in a research study. They were allowed to decide whether or not they wanted to know the results and receive counselling. The study was approved by the ethics committees of the Royal Melbourne and Royal Children’s Hospitals. Information was collected from three categories of ADPKD-affected individuals in each family. ADDRESSES. Murdoch Institute, Royal Children’s Hospital, Melbourne (D. Ravine, FRACP, S. M. Forrest, DPhil, L. J. Sheffield, FRACP, Prof D. M. Danks, MD); Departments of Nephrology (R. G. Walker, MD, Prof P. Kincaid-Smith, MD); and Radiology (R. N. Gibson, FRACR), Royal Melbourne Hospital; and Department of Cytogenetics and Molecular Genetics, Adelaide Children’s Hospital (R. I. Richards, PhD, K. Friend, BSc), Australia. Correspondence to Dr David Ravine, Institute of Medical Genetics, University Hospital of Wales, Heath Park, Cardiff CF4 4XW, UK.

1331

TABLE I-EVIDENCE SUPPORTING NON-LINKAGE IN FAMILIES WITHOUT APPARENT LINKAGE TO PKD1

TABLE II-NUMBER OF DOUBLE RECOMBINANTS IF LINKAGE HAD BEEN PRESENT

I

In

_

i

i

i

’Diagnosis based on ultrasound evidence.

Newly diagnosed-These were previously undiagnosed were diagnosed by ultrasound examination or DNA-linkage tests during the study. Their medical history was recorded, together with details of any existing renal symptoms. Physical examination included measurement of supine and erect systolic and diastolic (phase V) blood pressure. Renal and hepatic individuals who

ultrasound was done; 95% were scanned with 3 MHz or 5 MHz mechanical sector probes on a single ultrasound scanning unit (ATL Ultramark 8, Bothel, Washington, USA) and 5% were scanned in their local districts. All positive ultrasound scans and those for individuals at high risk of having PKD 1 were reviewed by

single radiologist (R. N. G.). Previously diagnosed-The medical history and current renal symptoms of previously diagnosed individuals were recorded, including age at diagnosis, the factors leading to diagnosis, and known complications of ADPKD, especially hypertension requiring treatment, macrohaematuria, renal infection, renal calculi, and stroke. Age at onset of renal failure was recorded in those who had started dialysis or undergone renal transplantation. No physical examination was done. When necessary, further medical details were obtained from the attending doctor. a

Deceased affected-Information was collected on affected individuals who had died, including dates of birth and death, mode of and age at diagnosis, and cause of death. The age at onset of end-stage renal failure was taken to be the age at which long-term replacement therapy for renal function became necessary or, in people who died of renal failure, the age at death. Before analysis, index cases were excluded from the patient population. The product-limit method of survival analysis23 was used to calculate survival probabilities. Differences between the PKD 1 and non-PKD 1 survival curves were assessed by the logrank test. Fisher’s exact test and two-sample t tests were done where appropriate. For the comparison of frequencies of hypertension in PKDland non-PKD1 patients with ADPKD, random effects logistic regression analysis was used to adjust the estimated odds ratio for age and family of origin.

I

I

pedigrees 4 and 5 markers known to be proximal were uninformative

small Lod scores, positive in 21 and negative in 2. The summed Lod score for these 23 families was 8 18, which suggests that most of these smaller families had PKD 1. Of the 18 PKD 1 probands, 10 were identified at the Royal Melbourne Hospital renal clinic, and 2 each by private adult nephrologists, paediatric nephrologists, genetics clinics, and other sources. By contrast, no non-PKD1 proband was referred to the study from the Royal Melbourne Hospital renal clinic (p = 0076); 4 were referred by adult nephrologists and 1 by a genetics clinic. All the PKD1 and non-PKD1 families were of European descent. Among the 18 PKDfamilies, ancestors with ADPKD were known to have migrated from the UK (8), Ireland (4), and Italy (1); 5 families did not know the country of origin of the ADPKD-affected ancestor. Among the 5 non-PKD1 families, 3 originated from the UK and 1 from Italy, and 1 was unaware of the country of origin of the affected ancestor. Within the 5 non-PKDlfamilies 49 members had ADPKD; 36 had been diagnosed previously, of whom 14 had died, and 13 were diagnosed by ultrasound examination during this study. The 18 PKDfamilies included 285 members with ADPKD; 238 had been diagnosed previously, of whom 122 had died, and 47 were identified by ultrasound or DNA-linkage studies during the study. 4 subjects were classified as having ADPKD despite their having fewer cysts than the usually accepted diagnostic criterion 24 2 of these subjects were from non-PKDl1 families; each had 3 cysts in one kidney (at ages 17 and 18 years). 1 subject judged to have PKD1 was a 23-year-old with no renal cysts, and the remaining subject was a 28-year-old with 1 cyst in each kidney. Informative flanking markers in both these individuals showed that their risk of having inherited PKD1was greater than 99 9%.

Results

Among the 52 families, 46 had sufficient family members available for linkage studies to be done. 18 of the 46 families had sufficient informative meioses available to establish that the probability of linkage to PKD1was greater than 095. 5 other families had evidence of non-linkage; ADPKD segregated independently of polymorphic markers flanking the PKD 1 locus (tablesI and II). The other 23 families had

Ht,J. tyl))

Fig 1-Cumulative probability of survival (to onset of end-stage renal disease or death) individuals with ADPKD.

among

PKD1

and

non-PKD1

1332

-

-

J

Fig 2-Number of renal cysts found on ultrasound at diagnosis among PKD1 and non-PKD1 subjectswithADPKDdiagnosed during study. With age at onset of end-stage renal failure or death (whichever occurred first) as the failure time, the cumulative survival probability was greater for non-PKDDthan for PKD1 subjects (median survival 71-5vs 56-0 years; fig 1). A significantly smaller proportion of the non-PKD 1 than of the PKD1 patients with ADPKD had progressed to end-stage renal failure (odds ratio 0-35, 95% CI 0-13-0-92). The age range at onset of end-stage renal failure was 42-80 years in non-PKD1 subjects and 0-3-71 years in PKD1

subjects. The odds ratio for hypertension in non-PKD1 compared with PKD1 affected family members was 0-52 (95% CI 0-25-1-09). However, among these patients, there was a strong association between age and hypertension; the mean age of those with hypertension was 45-6 (13-9) years compared with 29-3 (12-9) years for normotensive subjects (p < 0-001). A random effects logistic regression analysis was done with age and PKD 1 M non-PKD 1 as regression terms with the family of origin treated as a random effect. The aim of this treatment was to adjust simultaneously for age and family of origin. With that adjustment, the odds ratio for hypertension (non-PKD1vs PKD1) was 0-29 (95% CI

0-11-0-80). Among subjects

with newly diagnosed ADPKD, those with non-PKDldisease had fewer renal cysts than those

Age (yr)

Fig 3-Cumulative probability of previously individuals remaining undiagnosed. Vertical bars=age at diagnosis of censored individuals

untested

with PKD1 (fig 2). This difference was larger for kidneys with the smaller number of cysts (median non-PKDl vs PKD 1,3 vs 10-14) than for kidneys with the greater number of cysts (median 6 vs 15). The difference occurred despite the difference in the age at diagnosis (non-PKDl 33.1 [156] vs PKD1 29-0 [13’8] years). Among subjects with newly diagnosed non-PKD 1, the youngest person with more than 15 cysts in both kidneys was 49 years. Among the comparable PKDgroup, the youngest patient with more than 15 cysts in both kidneys was 16 years old. The age at diagnosis, and the factors contributing to the diagnosis, were known in 39 patients with non-PKDland 197 with PKD 1 disease. Survival analysis was done with the age at diagnosis achieved as a result of investigation of a PKD-related symptom as the failure time. Subjects whose diagnosis had been an incidental finding or had been sought only because of the background family history were censored at the time of diagnosis. Among subjects not diagnosed incidentally or because of background family history, the cumulative probability of remaining undiagnosed was higher in the non-PKD1 than in the PKD1group (fig 3). The median age at diagnosis due to investigation of an ADPKD symptom was 69-1years among non-PKDl subjects and 44-8 years among those with PKD1.

Discussion In this study, non-PKD1 patients lived longer and had a lower risk of progressing to renal failure than patients with PKD1-linked disease. In addition, a smaller proportion of non-PKDl subjects were hypertensive, they were diagnosed at an older age, and they had fewer renal cysts at the time of diagnosis. These findings accord with previous reports.11,25 Several very mildly affected PKD1 subjects were identified by DNA-linkage studies and were included in our results. Since we were not able to identify the non-penetrant non-PKDlaffected individuals, our data probably underestimate the difference in severity between the groups. The differences between PKD1 and non-PKD1 affected individuals may simply reflect slower onset of renal disturbance from a lower rate of cyst formation and enlargement in the non-PKDlgroup. However, other differences between these diseases may contribute to the milder phenotype of non-PKDl. Although there was substantial overlap in the numbers of renal cysts in newly diagnosed PKD1and non-PKD1affected subjects, the likelihood of an individual having the non-PKD1 genotype was inversely related to the number of cysts, the trend being most pronounced for the kidney with the fewer cysts. It may become possible to develop predictive tests, based on renal ultrasound results, to distinguish the groups. Clearly, such testing would need to be age-specific and far greater numbers would be needed to assess its usefulness. Other phenotype-genotype correlations may also be helpful in assessing the likely genotype; a detailed family history focusing on age at diagnosis, occurrence and age at onset of end-stage renal failure, and age at death may prove useful. Although most of the families in this study were found by probands’ attending public hospital renal clinics, no nonPKD 1 family was obtained from this source. In Australia, public hospital renal clinics manage most patients with significantly impaired renal function as well as those on dialysis or with renal transplants.2 The two families identified through children managed by a paediatric nephrologist proved to have PKD 1. This finding is

1333

consistent with

a

previous linkage study

heterogeneity of polycystic kidney in Europe. Contrib Nephrol 1992; 97:

of 10 families

ascertained through PKD-affected children,26 and suggests that ADPKD presenting in childhood is likely to be the PKD1type. The 4 non-PKDlprobands individual adult nephrologists were being from referred treated for ADPKD complications other than severe renal impairment. Although these differences in source of ascertainment failed to reach significance, the result is likely to have been limited by the small number of non-PKDl1 families. However, the observed differences are not surprising in the light of the milder phenotypic expression of non-PKD1 and are liable to affect estimation of the prevalence of each genotype. In this study, the high proportion of families ascertained from renal clinics is likely to have led to underestimation of the prevalence of non-PKD 1. For this reason, we did not estimate a. In earlier studies of genetic heterogeneity, the relative contributions of each source of ascertainment were not reported.9,10.27.28 Design of future studies on genotype prevalence must take source of ascertainment into account to ensure fuller identification of milder phenotypes. The existence of non-allelic heterogeneity complicates genetic counselling. If PKD1 linkage can be firmly established within a family, at-risk individuals can be assessed reliably by gene tracking. If PKD1 linkage can be firmly excluded, reliance must be placed upon ultrasound scanning. Unfortunately, little is known about the value of a negative scan in non-PKDl1 families at any particular age. If a family has too few affected individuals to determine involvement of the PKDlocus, possible approaches are ultrasound alone or gene tracking with a Bayesian adjustment for at, the prior probability of PKD 1, though the latter’s reliability depends on the accuracy of the estimate of a.

128-39. 11.

Parfrey PS, Bear JC, Morgan J, et al. The diagnosis and prognosis of autosomal dominant polycystic kidney disease. N Engl J Med 1990;

12.

Breuning MH, Snijdewint FGM, Smits JR, Dauwerse JG, Saris JJ, Van Ommen GJB. A Taq1 polymorphism identified by 26-6 (D16S125) proximal to the locus affecting adult polycystic kidney disease (PKD1)

13.

Snijdewint FGM, Saris JJ, Dauwerse JG, Breuning MH, van Ommen GJN. Probe 218EP6 (D16S246) detects PFLP’s close to the locus affecting adult polycystic kidney disease (PKD1) on chromosome 16.

14.

Hyland VJ, Suthers GK, Friend K, et al. A new probe, VK5B, for linkage studies with the autosomal dominant adult polycystic kidney disease

more severe

We thank the many clinicians who referred the probands whose families included in this study Mr John Donlan for ultrasonography; Dr M. H. Breuning, Dr V. J. Hyland, Dr Y. Nakamura, and Dr D. Higgs for genomic probes; and Dr J. B. Carlin, Ms R. J. Wilson, and Dr R. G. Newcombe for statistical advice. D. R. was the recipient of an Australian National Health and Medical Research Council postgraduate medical scholarship. were

REFERENCES

Geerlings W, Brunner FP, et al. Combined report on regular dialysis and transplantation in Europe, XIX, 1988. Nephrol Dial Transpl 1998; 4 (suppl): 5-29. 2. Disney APS, ed. Thirteenth report of the Australia and New Zealand Combined Dialysis and Transplant Registry (ANZDATA). Queen Elizabeth Hospital, South Australia 1990: 22. 3. Reeders ST, Breuning MH, Davies KE, et al. A highly polymorphic DNA marker linked to adult polycystic kidney disease on chromosome 1. Tufveson G,

16. Nature 1985; 317: 542-44.

4. Breuning MH, Snijdewint FGM, Brunner H, et al. Map in 16 polymorphic loci on the short sarm of chromosome 16 close to the polycystic kidney disease gene (PKD1). J Med Genet 1990; 27: 603-13. 5. Harris PC, Thomas S, Ratcliffe PJ, Breuning MH, Coto E, LopezLarrea C. Rapid genetic analysis of families with polycystic kidney disease 1 by means of a microsatellite marker. Lancet 1991; 338: 1484-87. 6. Kimberling WJ, Fain PR, Kenyon JB, Goldgar D, Sujansky E, Gabow PA. Linkage heterogeneity of autosomal dominant polycystic kidney disease. N Engl J Med 1988; 319: 913-18. 7. Romeo G, Devoto M, Costa G, et al. A second genetic locus for autosomal dominant polycystic kidney disease. Lancet 1988; ii: 8-11. 8. Norby S, Sorensen AWS, Boesen P. Non-allelic genetic heterogeneity of autosomal dominant polycystic kidney disease? In: Bartsocas CS, ed. Genetics of kidney disorders: proceedings of the Fifth International Clinical Genetics Seminar, Rethymno, Crete, October, 1988. Clin Biol Res 1989; 305: 83-89. 9. Kimberling WJ, Pieke SA, Kenyon JB, Gabow PA. An estimate of the proportion of families with autosomal dominant polycystic kidney disease (ADPKD) unliked to chromosome 16. Kidney Int 1990; 37: 249

(abstr). 10. Peters DJM, Sandkuijl LA, Snijdewint FGM,

et

al.

Genetic

323: 1085-90.

on

chromosome 16. Nucleic Acids Res 1990; 18: 3106.

Nucleic Acids Res 1990; 18: 3108.

locus, PKD1. Hum Genet 1990; 84: 286-88. Breuning MH, Reeders ST, Brunner H, et al. Improved early diagnosis of adult polycystic kidney disease with flanking DNA markers. Lancet 1987; ii: 1359-61. 16. Nakamura Y, Martin C, Krapcho K, et al. Isolation and mapping of a

15.

polymorphic DNA sequence (pCMM65) on chromosome 16 (D16S84). Nucleic Acids Res 1988; 16: 3122. 17. Jarman AP, Nicholls RD, Weatgerall DJ, Clegg JB, Higgs DR. Molecular characterisation of a hypervariable region downstream of the alpha-globin gene cluster. Embo J 1986; 5: 1857-63. 18. Thompson AD, Shen Y, Holman K, Sutherland GR, Callen DF, Richards RI. Isolation and characterisation of (AC)n microsatellite genetic markers from human chromosome 16. Genomics (in press). 19. Lathrop GM, Lalouel JM. Easy calculations of lod scores and genetic risks on small computers. Am J Hum Genet 1984; 36: 460-65. 20. Narod S. Counselling under genetic heterogeneity: a practical approach. Clin Genet 1991; 39: 125-31. 21. Weeks DE, Ott J. Risk calculations under heterogeneity. Am J Hum Genet 1989; 45: 819-21. 22. Lathrop GM, Lalouel J-M. Efficient computations in multilocus linkage analysis. Am J Hum Genet 1988; 42: 498-505. 23. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457-81. 24. Bear JC, McManamon P, Morgan J, et al. Age at clinical onset and at ultrasonographic detection of adult polycystic kidney disease: data for genetic counselling. Am J Med Genet 1984; 18: 45-53. 25. Bachner L, Vinet MC, Lacave R, et al. Linkage study of a large family with autosomal dominant polycystic kidney disease with reduced expression. Hum Genet 1990; 85: 221-27. 26. Gal A, Wirth B, Kaariainen H, et al. Childhood manifestations of autosomal dominant polycystic kidney disease: no evidence for genetic heterogeneity. Clin Genet 1989; 35: 13-19. 27. Reeders ST, Breuning MH, Ryynanen MA, et al. A study of genetic linkage heterogeneity in adult polycystic kidney disease. Hum Genet 1987; 76: 348-51. 28. Mandich P, Restagno G, Novelli G, et al. Autosomal dominant polycystic kidney disease: a linkage evaluation of heterogeneity in Italy. Am J Med Genet 1990; 35: 579-81.

From The Lancet Trouble-free travel If there is one feature of modem civilisation more distinctive than another it is the economy of force effected by its inventions and innovations. It is indeed the triumph of the mechanical, in substitution for the personal, factor that the former minimises the forth-putting of the latter. Take one of the most familiar directions in which the individual is called upon to exert himself-that of travel or getting about from place to place. If ancient or mediaeval records or the published experiences of "sixty years since" teach us anything, it is that a journey of any distance used to be a more or less serious undertaking, to be planned and prepared for days or weeks in advance, and sure to absorb the traveller’s mind during every hour of his progress till he found himself safe and sound, weary but thankful, at his destination. Dangers due not merely to the mode of transit, though these were real enough, but to the quality of the resting-places en route, to say nothing of the "gentlemen of the road" whose blackmail exhausted what the "perfidus caupo" had spared, kept the mind preoccupied morning, noon, and night, to the exclusion of that enjoyment shared by every present-day traveller, for whom vicarious exertion or agency minimises all hindrances to free attention, discursive or sustained.

(July 2, 1892)