Nephrol Dial Transplant (996) : -6 Original Article IMephrology Dialysis Transplantation Complications of autosomal dominant polycystic kidney disease in haemodialysed patients. A case-control study J. L. Christophe, C. van Ypersele de Strihou, Y. Pirson, and the U. C. L. Collaborative Group* Service de Nephrologie, Cliniques Universitaires Saint-Luc, Brussels, Belgium Abstract Background. The impact of renal and extrarenal manifestations of autosomal dominant polycystic kidney disease () during chronic haemodialysis (HD) has never been assessed in a paired case-control study. Methods. Comparison of the course of patients with matched control (C) patients who started chronic HD at the same time. Results. Follow-up averaged 48 and 39 months in the and C groups respectively. Actuarial survival was similar in both groups. Prevalence of renal pain (36 vs %, P =.), haematuria (36 vs 6%, /><.3) and renal infection (6 vs %, P<.4) was higher in the than in the C group. Nephrectomy during HD was performed in six (in 4 cases in preparation for transplantation) and in one control patient. Number of patients with coronary and heart valve complications was similar in both groups. Stroke occurred in three patients from both groups. Only two patients experienced a single episode of pain related to liver cyst. Prevalence of severe infection was similar in the group (36%) and the C group (8%). Number and duration of hospitalizations were similar in both groups. Conclusions. The overall outcome of patients on maintenance HD is similar to that of HD patients with other primary renal diseases. Complications related to cystic kidneys are frequent but rarely severe. Extrarenal manifestations of have a limited clinical impact in this short-term study. Key words: autosomal dominant polycystic kidney dis- Correspondence and offprint requests to: Prof. Y. Pirson, Service de Nephrologie, Cliniques Universitaires Saint-Luc, University of Louvain Medical School, Avenue Hippocrate, B- Brussels, Belgium. Members of the U. C. L. Collaborative group who participated in the study: M. Jadoul, E. Goffin (Cliniques Saint-Luc, Bruxelles); P. Bernis (Hopital Saint-Joseph, Gilly); A. Quoidbach, L. Stenuit, H. Demol (Hopital Civil, Jumet); M. Wauthier, J. Jamez (Clinique Saint-Pierre, Ottignies); G. Louie (Centre Hospitalier, Sainte-Ode); A. Cuvelier, J. M. Pochet (Clinique Sainte-Elisabeth, Namur); R. Cuvelier (Clinique du Refuge, Mouscron); J. J. Lafontaine (Clinique Saint-Joseph, Arlon). ease; haemodialysis; extrarenal complications; renal complications. Introduction Autosomal dominant polycystic kidney disease () accounts for 4-% of patients on maintenance dialysis []. In addition to the kidneys, the disease involves several other organs such as the liver, the heart, and the brain [,]. These disorders are expected to progress during renal replacement therapy but their attendant consequences on the fate of patients on haemodialysis (HD) remain to be delineated. We therefore decided to compare the course of patients with that of properly matched patients with end-stage renal failure of another cause who started haemodialysis at the same time. Special attention was paid to the renal and extrarenal complications of. Subjects and methods We retrospectively reviewed the charts of all patients who started chronic HD treatment between January 98 and December 99 in eight Belgian units: Cliniques Saint-Luc, Bruxelles; Hopital Saint-Joseph, Gilly; Hopital Civil, Jumet; Clinique Saint-Pierre, Ottignies; Centre Hospitalier, Sainte-Ode; Clinique Sainte-Elisabeth, Namur; Clinique du Refuge de la Sainte-Famille, Mouscron; Clinique Saint-Joseph, Arlon. Patients dialysed for more than one month in a non-participating centre or previously treated by peritoneal dialysis or kidney transplantation were excluded. Fifty patients were included. The diagnosis of ADPK.D was established before the onset of dialysis by usual criteria: all patients had typical enlarged cystic kidneys on imaging examination (intravenous pyelography, ultrasonography or computed tomography scan); 4 had a family history of or renal failure. Patients dialysed in more than one participating centre were attributed to the reference centre providing the longest treatment time. A control group of non- patients was identified for comparison purpose. For each patient, we selected in his reference centre a patient of the same sex, 996 European Dialysis and Transplant Association-European Renal Association Downloaded from https://academic.oup.com/ndt/article-abstract////8333 on 9 January 8
with a comparable age (± years), who had started HD at the calendar date closest to that of the patient (within ± years). Patients with systemic disease or diabetes were excluded. If no control patient fulfilling these criteria was found in the reference centre, the search was extended to another participating centre, starting with the largest (Cliniques Saint-Luc): this was the case for nine controls. Original nephropathy in the controls included: chronic glomerulonephritis (3), chronic interstitial nephritis (), nephroangiosclerosis (9), obstructive nephropathy () bilateral renal artery thrombosis (), unknown (4). Follow-up extends from initiation of HD until death, transfer to another treatment modality, or December 3 st, 993. Data were collected from the hospital and dialysis charts by one of us (JLC) working with the local nephrologist. The following events were recorded: renal pain: only acute episodes requiring medical advice and/or analgesics; haematuria: only macroscopic; renal infection: acute pyelonephritis or renal cyst infection; cerebrovascular accident (CVA): central nervous system symptoms or signs lasting for more than 4 h, attributed to cerebral ischaemia or haemorrhage; severe infection: proven bacteraemia or infection requiring hospitalization; hospitalization: only if duration exceeded 4 h; liver cyst complication: pain or infection. Data are presented as mean (+SD) or median (range). Differences for continuous variables were tested using Student's t test or Mann-Whitney non-parametric test when necessary. Dichotomous variables were compared using chisquare analysis or Fisher's exact test when needed. Yearly incidence of an event was calculated by dividing the number of events by the number of patient- years at risk. Actuarial patient survival and actuarial risk of complication were calculated by the Kaplan-Meier method and the curves were compared by the log-rank test. Results Patients' characteristics (Table ). As expected from the selection criteria of the control group, mean age and sex are comparable in both groups. Follow-up averaged 4 and 3. years in the and control group respectively. Unilateral nephrectomy had been performed prior to the onset of HD in two patients J. L. Christophe et al. (pyonephrosis, one case; suspicion subsequently invalidated of renal tumour, one case) and in one control patient (pyonephrosis). Patient survival. Actuarial patient survival at and years was 9 and 69% in the group, and 9 and 66% in the control group, respectively. There was no statistical difference between the two survival curves (Figure ). Causes of death are shown in Table. No significant difference emerges from this small series. Renal complications. The prevalence as well as the number of episodes of renal pain, haematuria, and renal infection are significantly higher in patients than in controls (Table 3). The actuarial risk and the yearly incidence of renal complications were calculated in the group. At, 3 and years after starting HD, the actuarial risk was 4, 36 and % for renal pain, 8, 4 and % for haematuria, and 9, and % for renal infection, respectively. For the whole follow- up period ( patient-years), the average yearly incidence of renal pain, haematuria and infection was respectively, and %. The yearly incidence was also calculated separately for the first 3 years of follow-up: during the first, second and third year, this incidence was, respect- SurvivalI 36 8 4 so 3 9 9 controls 8 6 4 " p =. controls Years after HD onset Fig.. Actuarial survival of and control patients. The number of patients at risk are indicated in the upper part of the figure. Table. Causes of death Table. Patients' characteristics Control Control Sex ratio (M F) /3 /3 Age at starting HD (years) 4.6 (±9.) 4.6 (±.) Follow-up (years) mean 4 (±3.) 3. (±.8) median.3 (<.I-.6). (<.I-8) Status at follow-up («patients) Still on HD Transplanted 8 Dead, not significant. Myocardial infarction Cerebrovascular accident Cancer Infection Other" Unknown " group: cachexia, ; abdominal aortic aneurysm rupture, ; acute pulmonary oedema, ; traffic accident, ; voluntary withdrawal from dialjsis,. Control group: chronic respiratory failure, : massive haemoptysis from probably benign bronchial tumour, : haematemesis.. Downloaded from https://academic.oup.com/ndt/article-abstract////8333 on 9 January 8
Complications of polycystic kidney disease in haemodialysis Table 3. Frequency of renal complications Table 4. Number of patients with cardiac complications 3 Control Controls P Overall prevalence (%) Renal pain Haematuria Renal infection Number of episodes Renal pain overall alone + haematuria + renal infection Haematuria overall alone + renal infection + renal pain Renal infection, not significant. 36 36 6 43 8 8 4 8 4 6 <. =.3 =.3 <. =.49 =. =.3 =.4 =. =.3 ively, 6, 9 and 3% for renal pain,, 9 and 3% for haematuria and 9, 8 and % for infection. The yearly incidence of renal complications such as renal pain or haematuria varied widely among patients (Figure ). Whether a history of haematuria or renal infection prior to the onset of dialysis predicted recurrence on HD was examined. Haematuria occurred in of (44%) patients with such a history before HD vs only seven of (8%) without. Similarly, renal infection occurred more than twice more frequently in patients with such a history (4 of 4 patients, (8%) than in those without (4 of 36, (%)). However, none of these differences reached statistical significance. Unilateral nephrectomy was performed during HD in six patients: in preparation for transplantation (n=4) (because of a history of recurrent haematuria in two patients, a history of recurrent renal Yearly incidence - of renal pain Yearly incidence of haematuria os. - 6-4- - - Illlm Him. patients patients Fig.. Average individual yearly incidence of renal pain (panel A) or haematuria (panel B) in patients who had such a complication. Patients are presented by decreasing rate of incidence. Note that the yearly incidence of renal pain exceeded in only one patient and that of haematuria in only two patients. Acute myocardial infarction Coronary artery revascularization Infarction and/or revascularization Heart failure due to valve disease Valve replacement Endocarditis, not significant. 4 8 infection in one patient, and a very large kidney size in the fourth), for suspected -but histologically unconfirmed- renal cancer (n = ) and for severe haematuria («= ). Bilateral nephrectomy was performed in two sessions for urothelial malignancy in one control patient. Cardiac complications. The number of patients with various cardiac events was similar in the two groups (Table 4). Myocardial infarction was fatal in four of five patients, and in two of seven control patients. Heart failure secondary to valve disease (severe mitral stenosis) developed in one control patient but in no patient. One patient underwent valvular replacement for calcified aortic valve stenosis. Another patient had a Staphylococcus aureus mitral endocarditis successfully treated by antibiotics. Hepatic complications. Screening for liver cysts by ultrasonography and/or computed tomodensitometry was performed in 43 patients. Liver cysts were present in 3 patients (6 men) and absent in the other ( men). Only two patients experienced a single episode of pain related to liver cysts. One of them had a similar episode before dialysis. No liver cyst infection was recorded. Cerebrovascular accidents. Three patients from each group had a cerebrovascular accident. None had a cerebral angiography. One patient suffered a cerebellar haemorrhage and survived with neurological sequelae. Another patient had left hemiparesia, right facial nerve paresia, and aphasia. A brain computerized tomography failed to disclose any abnormality. Recovery was almost complete after a few weeks. A third patient presented with right hemiparesia and dysarthria; recovery was progressive; brain imaging was not performed. One control patient had a cerebellar haemorrhage; he recovered completely after a few weeks. Another control patient presented with right hemiplegia and aphasia; brain CT scan was unremarkable; she died several days later of massive gastrointestinal bleeding, after partial neurological improvement. A third control patient presented with a loss of consciousness ascribed to a stroke of the brain stem, three days after starting HD; he was not investigated and died three days later. Infections. There was no significant difference between the two groups both for the number of patients with at least one episode of severe infection and for Downloaded from https://academic.oup.com/ndt/article-abstract////8333 on 9 January 8
4 Table a. Frequency of severe infections: number of patients who had at least one episode, and number of events in parentheses. No significant difference was found for both types of comparison J. L. Christophe et a!. cumulative follow-up time spent as in-patient, did not differ between groups (Table 6). Controls P Discussion Overall Proven bacteraemia Infection requiring hospitalization Table Sb. Origin of severe infections Lower respiratory tract Vascular access Renal Other, identified Unidentified 8(3) 9(4) 4() Table 6. Number, duration, and rate of hospitalization 8 3 4(6) (8) (8) Controls 6 Controls P Total number 83 93 Cumulative duration of hospit- 9 9 alization (days) Mean hospitalization rate" (%).6 (±.4) 8. (+.8) "Number of in hospital days divided by the cumulative duration of follow-up;, not significant. the total number of episodes. This remained true when proven bacteraemia and infection leading to hospitalization are distinguished (Table ). It is of note that in the group renal infection accounted for only eight of 3 episodes (3%) of severe infection occurring in only four patients. No episode of diverticulitis was recorded. Hospitalizations. The number of hospital admissions, the cumulative duration of hospitalization, and the rate of hospitalization, defined as the percentage of the Table. Survival of patients on HD compared to patients with other primary renal diseases Specific complications related to might be expected to compromise survival and outcome of patients treated by maintenance HD. Reassuringly, our case-control study does not support this view, at least in the short-term: is not associated with an increased -year mortality and only marginally affects morbidity in HD patients. The mean age of our patients at the initiation of HD ( years) is within the range reported by others [3-6]. The 69% -year survival rate compares favourably with the corresponding figures of American and European Registries (Table ). Mortality as well as -year survival rates were similar in our and control patients, in agreement with a recent report from Toronto []. Large surveys have even shown a -% higher -year survival rate in patients, than in those with other primary renal diseases (Table ). This was ascribed to a lower rate of cardiac deaths, tentatively related to a higher haematocrit providing some protection against ischaemic heart disease [9]. Taken together, these data establish that by itself does not adversely affect overall survival once HD is started. Despite the large number of patients on maintenance HD, information on renal and extra-renal manifestations of the disease during treatment is scarce. Chester compared patients with randomly selected controls and concluded that patients had a higher haematocrit and a lower blood pressure [3]. Jungers reviewed patients who started HD between 96 and 99 (potential followup similar to ours, - years; mean follow-up not mentioned) but did not compare the recorded complications with those of a control group [4]. Our study is the first to compare the main -associated complications with those observed in a carefully Author Age Group n Survival (%) year years "Mailloux, 988 [] Byrne, 994 [8] b Ritz, 99 [9] Present study All -64 4-9 >6 ± C CGN CGN C SPRD SPRD Controls 48 4 49 6 63 46 43 338 9 9 83 9 9 89 83 9 9 6 49 36 3 8 38 69 66 "Patients undergoing HD for less than <9 days were excluded. b Male patients on renal replacement therapy (most of them on HD) recorded at the EDTA Registry. C CGN, chronic glomerulonephritis: SPRD, standard primary renal diseases (excluding diabetes). Downloaded from https://academic.oup.com/ndt/article-abstract////8333 on 9 January 8
Complications of polycystic kidney disease in haemodialysis matched control group. Only this type of comparison can delineate the actual role of in the occurrence of disorders not necessarily related to the disease, such as cardio-vascular complications. The limitation of our study is the relatively small number of patients, especially when rare complications are evaluated. Renal complications were frequent in our patients. The actuarial risk of pain requiring medical attention or analgesic use reaches % years after HD onset. Gross haematuria was present in 4% of the episodes, an expected finding since acute pain usually results from intracystic haemorrhage []. The prevalence of gross haematuria (36%) is close to that reported (3%) by Jungers [4]. Notably the origin of bleeding was not ascertained in these studies. The 6% prevalence of haematuria found in our control group is of interest: it suggests that some episodes reported in the study group may be unrelated to. Five years after starting HD, half of the patients will have had gross haematuria at least once. Renal infection occurred in 6% of our patients, a lower rate than the 6% reported by Jungers et al. [4]. Another major difference between the two studies is the outcome of this complication. In our study, renal infection required nephrectomy in only (%) patient and was never fatal, whereas the corresponding figures in Jungers' report were 4% for nephrectomy and % for mortality. This improvement certainly reflects the easier control of renal infection by presently available lipophilic antibiotics, such as quinolones, reaching high concentrations in cyst fluid.despite the loss of renal function []. The yearly incidence of renal events in patients on HD provides a practical information to the nephrologist caring for such patients: if he treats patients, he is likely to encounter yearly two episodes of pain, two of gross haematuria, and one of renal infection. May haematuria and infection be predicted by pre-hd history? Our data suggest it but do not achieve statistical significance, whereas in the larger series from Necker, there was indeed, for both events, a significant correlation between pre-hd history and occurrence on HD [4]. Nephrectomy is nowadays rarely needed in patients on HD, outside preparation for transplantation. In the long-term, kidneys may occasionally become so massive that they require nephrectomy []. The absence of renal cancer in our and other series [4, ] confirms the rarity of this complication in [3]. The prevalence of myocardial infarction (%) with eventual death (8%) in our group is close to that found by Jungers et al. (.8 and 6.% respectively) [4]. As compared with controls, the frequency of coronary events tended to be lower in patients. This is in agreement with the abovementioned lower cardiac mortality observed in a large series of patients [9]. The cardiac valve abnormalities did not contribute to death in our patients, in keeping with another report []. It is however to note that one patient required valve replacement and another one medical treatment of endocarditis. Liver cysts, present as expected [4] in 3/4 of our patients, were rarely symptomatic. Only two patients had a single episode of pain ascribed to liver cyst and no infection of liver cyst was recorded. In the absence of haematuria or pyuria, it is sometimes very difficult to distinguish painful/infected renal and liver cysts: we cannot therefore totally exclude that a few episodes diagnosed as renal pain or infection were actually hepatic in origin. In a larger series of 9 patients, most of them on HD, Griinfeld reported liver cyst infection in seven patients, with eventual death in three of them [4]. In our series, an equal number of and agematched control patients experienced a cerebrovascular accident. Roscoe et al, in contrast, reported an increased mortality from cerebrovascular accident among 3 patients and ascribed it to intracranial aneurysm (ICA) rupture []. In this study, however, patients were 4 years older than controls and ICA rupture was not proven, leaving the possibility that age- or disease-related intracerebral haemorrhage or ischaemic stroke contributed to the mortality excess. It should be remembered that ICA is found in only 8% of patients, does not always rupture, and when it does it ruptures at a mean age of 4 years []. Whether patients have an increased risk of non-aneurysmal intracerebral haemorrhage remains to be determined. We did not find an association between and severe infection, especially after exclusion of renal infection. By contrast, an increased rate of fatal infection in patients treated by renal replacement therapy was reported by others [6,, ], suggesting that may be associated with some immunodeficiency []. It should be noted that those reports included both dialysed and transplanted patients, leaving the possibility that this finding was restricted to the transplanted patients. This issue should be carefully examined in case-control study of transplanted patients. Finally, using frequency and duration of hospitalization as an index of serious morbidity, we found no difference between and control patients. The percentage of the total time at risk spent in the hospital by our patients (.6%) is similar to that observed in the Singh et al. series [6]. In conclusion, renal and extrarenal complications of do not impact the short-term mortality and morbidity of patients treated by HD. Complications related to cystic kidneys, especially haematuria and pain, are frequent but rarely severe. Extrarenal lesions, namely liver cysts and intracranial aneurysm, do not significantly affect the short-term prognosis on HD. Whether they impact the longer term remains to be examined in a more extended case-control cohort. Acknowledgements. We are indebted to Dr Genevieve Hanique for statistical advice, Dr Elizabeth Jones for providing EDTA data, and Mrs Rita Jacobs for preparing the manuscript. Downloaded from https://academic.oup.com/ndt/article-abstract////8333 on 9 January 8
6 References. Pirson Y, Grunfeld JP. Autosomal-dominant polycystic kidney disease. In: Cameron JS, Davison AM, Grunfeld JP, Kerr D, Ritz E, eds. Oxford Textbook of Clinical Nephrology. Oxford University Press, Oxford, 99; -88. Gabow PA. Autosomal dominant polycystic kidney disease more than a renal disease. Am J Kidney Dis 993; 6: 43-43 3. Chester AC, Argy WP, Rakowski TA, Schreiner GE. Polycystic kidney disease and chronic hemodialysis. Clin Nephrol 98; : 9-33 4. Jungers P, Albouze G, Aime F et al. Problemes poses par l'hemodialyse periodique et la transplantation renale chez les malades atteints de maladie polykystique renale. In: Kuss R,. Legrain M, ed. Seminaires duro-nephrologie. Masson, Paris, 983; -3. Mendelssohn DC, Harding ME, Cardella CJ, Cook GT, Uldall PR. Management of end-stage autosomal dominant polycystic kidney disease with hemodialysis and transplantation. Clin Nephrol 988; 3: 3-39 6. Barrett BJ, Parfrey PS. Autosomal dominant polycystic kidney disease and end stage renal disease. Semin Dial 99; 4: 6-3. Mailloux LU, Bellucci AG, Mossey RT et al. Predictors of survival in patients undergoing dialysis. Am J Med 988; 84: 8-86 J. L. Christophe et al. 8. Byrne C, Vernon P, Cohen JJ. Effect of age and diagnosis on survival of older patients beginning chronic dialysis. JAMA 994; : 34-36 9. Ritz E, Zeier M, Schneider P, Jones E. Cardiovascular mortality of patients with polycystic kidney disease on dialysis: is there a lesson to learn? Nephron 994; 66: -8. Roscoe JM, Brissenden JE, Williams EA, Chery AL, Silverman M. Autosomal dominant polycystic kidney disease in Toronto. Kidney Int 993; 44: -8. Milutinovic J. Massive growth of kidneys in patients with autosomal dominant polycystic disease treated with chronic hemodialysis. Am J Kidney Dis 989; 4: 36-368. Fick GM, Johnson AM, Hammond WS, Gabow PA. Causes of death in autosomal dominant polycystic kidney disease. J Am Soc Nephrol 99; : 48-6 3. Keith DS, Torres VE, King BF, Zincki H, Farrow GM. Renal cell carcinoma in autosomal dominant polycystic kidney disease. J Am Soc Nephrol 994; 4: 66-669 4. Grunfeld JP, Albouze G, Jungers P et al. Liver changes and complications in adult polycystic kidney disease. Adv Nephrol 98; 4: -. Pirson Y., Chauveau D: Cerebral aneurysms. In: Watson M, Torres V, ed. Polycystic Kidney Disease. Oxford University Press, Oxford, 996; 3-4 6. Singh S, Hariharan S. Renal replacement therapy in autosomal dominant polycystic kidney disease. Nephron 99; : 4-44 Received for publication:..99 Accepted for publication: 4.3 996 Downloaded from https://academic.oup.com/ndt/article-abstract////8333 on 9 January 8