Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
NDT Advance Access published April 5, 2012 Nephrol Dial Transplant (2012) 0: 1–7 doi: 10.1093/ndt/gfs053 Original Article Tonsillectomy has beneficial effects on remission and progression of IgA nephropathy independent of steroid therapy Isseki Maeda1,2, Tomoshige Hayashi1, Kyoko Kogawa Sato1, Mikiko Okumoto Shibata2, Masahiro Hamada2, Masatsugu Kishida2, Chizuko Kitabayashi2, Takashi Morikawa2, Noriyuki Okada2, Michiaki Okumura2, Masayo Konishi3, Yoshio Konishi2, Ginji Endo1 and Masahito Imanishi2 1 Department of Preventive Medicine and Environmental Health, Osaka City University Graduate School of Medicine, Osaka, Japan, Department of Nephrology and Hypertension, Osaka City General Hospital, Osaka, Japan and 3Department of Nephrology, Wakakusa Daiichi Hospital, Osaka, Japan 2 Abstract Background. Indication of tonsillectomy in IgA nephropathy is controversial. The purpose of this study was to examine the efficacy of tonsillectomy on remission and progression of IgA nephropathy. Methods. We conducted a single-center 7-year historical cohort study in 200 patients with biopsy-proven IgA nephropathy. Study outcomes were clinical remission defined as disappearance of urine abnormalities at two consecutive visits, glomerular filtration rate (GFR) decline defined as 30% GFR decrease from baseline and GFR slope during the follow-up. Results. Seventy of the 200 patients received tonsillectomy. Tonsillectomy was associated with increased incidence of clinical remission (P < 0.01, log-rank test) and decreased incidence of GFR decline (P ¼ 0.01, log-rank test). After adjustment for age and gender, hazard ratios in tonsillectomy were 3.90 (95% confidence interval 2.46–6.18) for clinical remission and 0.14 (0.02–1.03) for GFR decline. After further adjustment for laboratory (baseline mean arterial pressure, GFR, 24-h proteinuria and hematuria score), histological (mesangial score, segmental sclerosis or adhesion, endocapillary proliferation and interstitial fibrosis) or treatment variables (steroid and renin–angiotensin system inhibitors), similar results were obtained in each model. Even after exclusion of 69 steroid-treated patients, results did not change. GFR slopes in tonsillectomy and non-tonsillectomy groups were 0.60 6 3.65 and 1.64 6 2.59 mL/min/1.73 m2/year, respectively. In the multiple regression model, tonsillectomy prevented GFR decline during the follow-up period (regression coefficient 2.00, P ¼ 0.01). Conclusion. Tonsillectomy was associated with a favorable renal outcome of IgA nephropathy in terms of clinical remission and delayed renal deterioration even in non-steroid-treated patients. Keywords: clinical remission; IgA nephropathy; renal deterioration; steroid therapy; tonsillectomy Introduction IgA nephropathy is the most common type of primary glomerulonephritis. Clinical course of this disease depends on clinical manifestations and is greatly varied by patient. Ten-year renal survival was estimated ~80–85% [1]. In patients at high risk for progressive disease, such as those with severe proteinuria or arterial hypertension, treatment with renin–angiotensin system (RAS) inhibitors or corticosteroid to resolve urine abnormalities and prevent renal deterioration is established with support from a growing body of evidence [2–7]. However, the therapeutic effect of tonsillectomy in IgA nephropathy is controversial [8]. Although IgA nephropathy has highly variable clinical presentations, gross hematuria at the time of tonsillitis is one of the typical clinical features of this disease. Chronic and recurrent tonsillitis are considered to play an important role in new onset and progression of IgA nephropathy. Although several previous studies have examined whether tonsillectomy had a beneficial effect on IgA nephropathy in regard to clinical remission or renal deterioration, the results were inconsistent [9–12]. Some of the studies showed that tonsillectomy normalized urine abnormalities (clinical remission) in patients with IgA nephropathy, but failed to show that tonsillectomy prevented renal deterioration in the same patients [9, 10]. Because of the slowly progressive nature of IgA nephropathy, renal deterioration could not be detected as study outcomes including end-stage renal disease (ESRD) or need for renal replacement therapy in relatively short-term studies. As the therapeutic target of early-stage IgA nephropathy is delaying loss of renal function, the therapeutic efficacy of tonsillectomy should be measured by continuous values such as glomerular filtration rate (GFR) slope. In addition, because steroid therapy has a strong effect on clinical manifestation of IgA nephropathy, a study in patients who mainly received combination therapy with tonsillectomy and steroid therapy might be inappropriate to evaluate the independent efficacy The Author 2012. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please e-mail: [email protected] Downloaded from http://ndt.oxfordjournals.org/ by guest on April 7, 2012 Correspondence and offprint requests to: Tomoshige Hayashi; E-mail: [email protected] 2 of tonsillectomy. Two retrospective studies examined the effect of tonsillectomy on kidney outcome in the 1970s, when the indication of corticosteroid therapy for IgA nephropathy was limited. However, the results were inconclusive [11, 12]. The main reasons for the discrepancy were the difference of patient characteristics and indication for renal biopsy and treatment. We therefore performed a longitudinal study using historical data from a single center to elucidate the effect of tonsillectomy on the remission and progression of IgA nephropathy. Materials and methods Site and setting Therapeutic intervention Of the 200 patients, 70 patients received tonsillectomy. Indication of tonsillectomy was decided after consultation with otolaryngologists. Standard indications of tonsillectomy were recurrent tonsillitis three times or more per year, repeated episodes of gross hematuria at the time of tonsillitis or chronic tonsillitis with pus in tonsillar crypt [9, 13, 14]. Tonsillectomy was performed within 6 months from the renal biopsy. There were 69 patients who received oral or pulse steroid therapy as their initial treatment. Main indications for steroid therapy were if patients had symptoms of progressive diseases such as high-grade proteinuria, erythrocyte cast, acute renal failure or advanced histological findings including glomerular sclerosis, crescent formation, endocapillary proliferation or interstitial infiltration. In oral steroid therapy, the initial dose of prednisolone was 0.6–0.8 mg/kg/ day. In pulse steroid therapy, 500 mg of methylprednisolone was administered intravenously on three consecutive days followed by oral prednisolone (30 mg/day) on four consecutive days. This course was repeated three times. After that, prednisolone at a dose of 30 mg/day was given orally on alternative days. In both therapies, prednisolone was gradually tapered and finished over 1 year. The protocol of the pulse therapy was established by Hotta et al. [9]. Data collection Clinical, laboratory and histological data were recorded in our database. Patient characteristics included age at biopsy, gender, time of onset, type of onset (gross hematuria, urine abnormalities in medical checkup or other symptoms), history of recurrent tonsillitis, evidence of enlarged tonsils and anthropometric and laboratory data. Blood pressure was measured using a mercury manometer in a supine position after a few minutes of rest. Twenty-four-hour urine collection was performed on three consecutive days. Baseline blood pressure, serum creatinine, 24h proteinuria and hematuria score were calculated as the average of the three measurements. The results of dipstick urinalysis were interpreted as (), (), (1), (21) or (31). Results of () or less in dipstick urinalysis were regarded as normal. To assess the degree of hematuria, the dipstick results were converted into scores: () to 0, () to 0.5, (1) to 1, (21) to 2 and (31) to 3. We calculated estimated GFR using a modified threevariable equation for Japanese, which was validated by the standard inulin clearance techniques, as follows: estimated GFR ¼ 194 3 Age0.287 3 serum creatinine1.094 (mg/dL) 3 0.739 (if female) [15]. Histological evaluation was made according to the Oxford classification of IgA nephropathy [16, 17]. The following pathological variables were evaluated: mesangial hypercellularity score, presence or absence of segmental sclerosis or adhesion, endocapillary hypercellularity and tubular atrophy/interstitial fibrosis. Interstitial lesion was graded as 0: 0–25%, 1: 26–50% and 2: 51% or more. These variables were described as having predictive value for renal outcome in previous articles [16, 17]. During the follow-up period, patients underwent a medical checkup every 1–3 months. Blood pressure, dipstick urinalysis and serum creatinine were measured at every visit. Treatments with oral or intravenous steroids, RAS inhibitors, other anti-hypertensive medications or tonsillectomy were also recorded. Outcomes Clinical remission of IgA nephropathy was defined as normal dipstick examination of hematuria and proteinuria on two consecutive visits at least 3 months apart. GFR decline was defined as >30% loss of estimated GFR from the baseline value. GFR slope of each patient during the follow-up period was obtained by a linear regression model and the principle of least squares. To calculate GFR slope, we used values of estimated GFR at every visit during the entire follow-up period of each patient even if the patient had experienced clinical remission or 30% GFR decline earlier. This value indicated annual GFR loss (mL/min/1.73 m2/year) and was not dependent on the follow-up length of patients. Statistics Baseline characteristics are presented separately according to the presence or absence of tonsillectomy. The values are expressed as mean SD, median (interquartile range) or %. Paired analysis between the groups was performed using unpaired t-test, Mann–Whitney test or chi-square test. To assess the effect of tonsillectomy on clinical remission or GFR decline, we used Kaplan–Meier plot, log-rank test and Cox’s proportional hazards models. Follow-up of each patient for Cox’s proportional hazards models was continued until the date of the first occurrence of each outcome of interest, the final visit or 7 years from diagnostic biopsy, whichever came first. Because we had a limited number (n ¼ 30) of incident cases of GFR decline, adjustment for all baseline variables in one model might cause inaccurate estimation [18]. To avoid this problem, we examined four models in which tonsillectomy was adjusted for the following variables: in Model 1, age and gender; in Model 2 (laboratory variables), mean arterial pressure, GFR, 24-h proteinuria and hematuria score in addition to Model 1; in Model 3 (histological variables), mesangial score, segmental sclerosis or adhesion, endocapillary proliferation and interstitial fibrosis in addition to Model 1 and in Model 4 (treatment variables), RAS inhibitors and steroid therapy in addition to Model 1. Proportional hazard assumption was confirmed by log-log plot. We also examined risk factors affecting GFR slope using multiple linear regression analysis. In this model, we included factors that had a P-value <0.2 in univariate analyses or had clinical relevance. We included age, gender, mean arterial pressure, GFR, 24-h proteinuria, hematuria score, mesangial score, segmental sclerosis or adhesion, interstitial fibrosis, RAS inhibitors and tonsillectomy in the multiple linear regression model. Multicollinearity was tested using variance inflation factor. In both Cox proportional hazards models and multiple linear regression models, non-linear effects of continuous independent variables were evaluated by inserting quadratic and log transformations into the multiple-adjusted models. Because 24-h proteinuria showed a non-linear effect, we converted it into log-transformed form. We calculated 95% confidence Downloaded from http://ndt.oxfordjournals.org/ by guest on April 7, 2012 From April 1997 to December 2007, 1034 patients underwent renal biopsy in the Department of Nephrology and Hypertension, Osaka City General Hospital, Osaka, Japan. In our clinical policy, main indication for performing a renal biopsy was asymptomatic urine abnormalities (persistent microscopic hematuria and mild to moderate proteinuria), recurrent episodes of gross hematuria, nephrotic syndrome or acute renal insufficiency of unknown origin without obvious urological abnormalities. Of these 1034 patients, 340 patients had a diagnosis of IgA nephropathy. Diagnosis of IgA nephropathy was made by detection of mesangial proliferative glomerulonephritis with deposits of immune complex staining predominantly for IgA in light microscopic and immunofluorescence studies. Patients were eligible for the current study if the following criteria were all satisfied: age 15–65 years at study entry; newly diagnosed IgA nephropathy; no clinical or biological evidence of systemic diseases such as Henoch–Schonlein purpura nephritis, systemic lupus erythematosus, other collagen diseases, viral hepatitis, diabetes or malignancy; no previous treatment with corticosteroids or immunosuppressive drugs; baseline serum creatinine level <176.8 mmol/L (2.0 mg/dL) and at least 12 months follow-up at our institute. Of the 340 patients, 99 patients were excluded because of age criteria (n ¼ 18), non-initial biopsy or previous treatment (n ¼ 13), concomitant disease (n ¼ 28), elevated serum creatinine (n ¼ 3) and/or loss to follow up before 12 months (n ¼ 43). Eligible patients consisted of 241 men and women. Then, we further excluded 41 subjects because of incomplete clinical data (n ¼ 18), normal range of hematuria and proteinuria (n ¼ 11) or insufficient biopsy specimen containing less than eight glomeruli (n ¼ 12). Therefore, the cohort for current analyses consisted of 200 men and women. We also performed additional analyses to assess independent effectiveness of tonsillectomy after exclusion of 69 patients who had received steroid therapy as their initial treatment. The protocol of this research was reviewed and approved by the local ethical committee. I. Maeda et al. Tonsillectomy and IgA nephropathy 3 intervals for each hazard ratio and coefficient. P-values were two-tailed. Statistical analyses were performed using PASW Statistics 17.0 (SPSS Inc., Chicago, IL). Results Baseline characteristics Baseline characteristics of study participants are shown in Table 1. Most of the clinical and histological factors were comparable between tonsillectomy and non-tonsillectomy groups. Endocapillary hypercellularity in biopsy specimens was seen more often in patients who received tonsillectomy than in those who did not. The tonsillectomy group received treatment with steroid and RAS inhibitors more frequently than the non-tonsillectomy group. The tonsillectomy group showed more active and less advanced histological changes and was administered aggressive treatments compared to the non-tonsillectomy group. Incidence rates of clinical remission were 9.3 per 100 person-years in the non-tonsillectomy group and 34.2 per 100 person-years in the tonsillectomy group. Tonsillectomy showed a significant positive association with clinical remission (P < 0.01, Figure 1A). In Cox’s proportional hazards models, patients with tonsillectomy revealed a 3.90 times higher hazard ratio for clinical remission than those without tonsillectomy after adjustment for age and gender. This relationship remained after adjustment for laboratory, histological or treatment variables; multiple-adjusted hazard ratios for tonsillectomy were 4.03 (95% confidence interval 2.52–6.44), 3.71 (2.30–5.98) and 3.06 (1.74–5.40), respectively. Lower hematuria score and steroid therapy were also associated with clinical remission (Table 2). GFR decline Incidence rates of GFR decline were 4.8 per 100 person-years in the non-tonsillectomy group and 0.5 per 100 person-years Non-steroid-treated patients To eliminate the therapeutic effect of corticosteroid therapy, we excluded an additional 69 patients who underwent steroid therapy as their initial treatment. Characteristics of the patients without steroid therapy are presented in Table 3. The tonsillectomy group had fewer cases of males, lower 24-h proteinuria and lower mesangial score than the non-tonsillectomy Table 1. Baseline characteristics according to patients with or without tonsillectomy who had IgA nephropathya Age, years Gender, male History of recurrent tonsillitis, yes Gross hematuria, yes Body mass index, kg/m2 Mean arterial pressure, mmHg IgA, mg/dL Estimated GFR, mL/min/1.73 m2 24-h proteinuria, mg/dL Hematuria score Mesangial score Glomeruli with segmental sclerosis or adhesion, % Endocapillary hypercellularity, yes Interstitial fibrosis, 0–25/26–50/51%1 Steroid therapy RAS inhibitor GFR slope, mL/min/1.73 m2/year Tonsillectomy (n ¼ 70) No tonsillectomy (n ¼ 130) 31.0 (25.2–36.7) 19 (27.1%) 51 (72.9%) 26 (37.1%) 21.3 (19.5–24.6) 87.5 (81.8–94.0) 319.0 (267.0–385.3) 89.6 6 25.2 349.7 (129.4–1006.8) 2.00 (1.92–2.38) 0.50 (0.33–0.74) 11.1 (5.4–25.9) 37 (52.9%) 66/3/1 50 (71.4%) 49 (70.0%) 0.03 6 6.03 32.1 (23.0–48.5) 48 (36.9%) 85 (65.4%) 49 (37.7%) 21.9 (19.9–24.2) 87.8 (81.7–98.0) 325.7 (250.4–442.5) 85.4 6 23.2 341.7 (127.9–842.7) 2.50 (1.79–3.00) 0.57 (0.37–0.83) 15.4 (7.0–30.0) 45 (34.5%) 116/9/5 19 (14.6%) 58 (44.6%) 1.31 6 2.73 P-value 0.25 0.21 0.34 1.00 0.40 0.44 0.46 0.25 0.95 0.09 0.18 0.07 0.02 0.46 <0.01 0.01 0.10 a Values are expressed as mean 6 SD, median (interquartile range) or % and compared using unpaired t-test, Mann–Whitney’s U-test or chi-square test, respectively. Downloaded from http://ndt.oxfordjournals.org/ by guest on April 7, 2012 Clinical remission in the tonsillectomy group. Tonsillectomy was associated with a lower incidence of GFR decline (P ¼ 0.01, Figure 1B). In Cox’s proportional hazards models, patient who received tonsillectomy had a lower risk for GFR decline. After adjustment for laboratory, histological or treatment variables, multiple-adjusted hazard ratios were 0.12 (0.02–0.89), 0.12 (0.02–0.89) and 0.10 (0.01–0.85), respectively. Proteinuria was associated with an unfavorable renal outcome. Treatment with RAS inhibitors showed a significantly higher hazard ratio for GFR decline (Table 2). As shown in Table 1, the tonsillectomy group revealed a higher proportion of patients receiving steroid and RAS inhibitor therapy. Since these treatments are known to delay progression of IgA nephropathy, we conducted additional analyses. After further adjustment for the treatments of steroid and RAS inhibitor therapy in Models 1–3 (shown in Table 2), which did not include these treatments, the results did not change (data not shown). Similarly, distributions of segmental glomerulosclerosis and endocapillary hypercellularity differ by groups. When we performed additional analyses to use these histological factors as independent variables in Models 1, 2 and 4 of Table 2 which did not include them, the results did not change (data not shown). Furthermore, even when we defined a rapid disease evolution if a patient had a GFR slope <5.12 mL/min/1.73 m2/year, which was the lowest 10th percentile value of the analytic cohort (n ¼ 200), the proportion of the patients who underwent tonsillectomy was not different between patients with or without rapid disease progression (40.0 versus 34.4%, in patients with or without rapid disease evolution, P ¼ 0.62 in chi-square test). 4 I. Maeda et al. Fig. 1. (A) Kaplan–Meier plot illustrating the probability of clinical remission among all patients who had IgA nephropathy according to tonsillectomy. (B) Kaplan–Meier plot illustrating the probability of renal survival for all patients who had IgA nephropathy according to tonsillectomy. Clinical remissionb HR (95% CI) Model 1 (age- and gender-adjusted model)d Tonsillectomy Model 2 (clinical factor-adjusted model)d Tonsillectomy Mean arterial pressure, 10 mmHg Baseline GFR, 10 mL/min/1.73 m2 Log 24-h proteinuria, mg/dL Hematuria score Model 3 (histological factor-adjusted model)d Tonsillectomy Mesangial score Segmental sclerosis or adhesion Endocapillary proliferation Interstitial fibrosis 26–50% 51%1 Model 4 (treatment factor-adjusted model)d Tonsillectomy Steroid RAS inhibitor GFR declinec P HR (95% CI) P 0.05 3.90 (2.46–6.18) <0.01 0.14 (0.02–1.03) 4.03 (2.52–6.44) 0.95 (0.73–1.23) 0.92 (0.80–1.05) 0.66 (0.42–1.05) 0.64 (0.48–0.85) <0.01 0.68 0.21 0.08 <0.01 0.12 (0.02–0.89) 1.26 (0.81–1.94) 1.26 (0.98–1.60) 5.93 (2.02–17.42) 1.00 (0.61–1.63) 3.71 (2.30–5.98) 0.55 (0.26–1.16) 1.00 (0.53–1.88) 1.30 (0.84–2.00) <0.01 0.12 1.00 0.24 0.12 (0.02–0.89) 1.57 (0.55–4.48) 0.57 (0.16–1.95) 1.78 (0.80–3.97) 0.04 0.40 0.37 0.16 1.14 (0.46–2.81) 0.86 (0.20–3.65) 0.78 0.83 1.86 (0.59–5.86) 2.67 (0.58–12.19) 0.29 0.21 3.06 (1.74–5.40) 2.35 (1.38–4.02) 0.58 (0.36–0.93) <0.01 <0.01 0.02 0.10 (0.01–0.85) 1.01 (0.32–3.16) 2.61 (1.12–6.06) 0.03 0.99 0.03 0.04 0.30 0.07 <0.01 1.00 a CI, confidence interval; HR, hazard ratio. Clinical remission was defined as normal dipstick examination of hematuria and proteinuria on two consecutive visits at least 3 months apart. GFR decline was defined as >30% loss of estimated GFR from baseline. d Age and gender were also adjusted in models 1–4. b c group. The non-tonsillectomy group presented a steeper GFR slope than the tonsillectomy group (1.64 2.59 versus 0.60 3.65, P < 0.01). Incidence rates of clinical remission were 21.8 per 100 person-years in the tonsillectomy group and 8.3 per 100 person-years in the non-tonsillectomy group. The tonsillectomy group showed a higher incidence of clinical remission than the non-tonsillectomy group (P < 0.01, Figure 2A). Although the incidence rate of GFR decline was lower in the tonsillectomy group than in the non-tonsillectomy group, the difference did not reach statistical significance (P ¼ 0.10, Figure 2B). In Cox’s proportional hazards models, the hazard ratio for clinical remission was about three times higher in patients with tonsillectomy than in those with- out tonsillectomy even after adjustment for potential confounders. Lower proteinuria was also associated with a higher incidence of clinical remission (Table 4). Because no patients with tonsillectomy reached the end point of GFR decline during the follow-up, we could not estimate hazard ratios in Cox’s proportional hazards model. Instead, we calculated GFR slope during mean follow-up of 5.2 0.2 years (2.7 0.4 years in tonsillectomy and 5.6 0.2 years in non-tonsillectomy group) and assessed factors relevant to GFR slope. In the multiple linear regression model, tonsillectomy indicated a delaying effect on renal function decline (regression coefficient 2.00, 95% confidence interval 0.58–3.42, P ¼ 0.01). A higher baseline Downloaded from http://ndt.oxfordjournals.org/ by guest on April 7, 2012 Table 2. Multivariable-adjusted hazard ratio for clinical remission or GFR decline in all patients with IgA nephropathy (n ¼ 200)a Tonsillectomy and IgA nephropathy 5 Table 3. Baseline characteristics according to patients with or without tonsillectomy who had IgA nephropathy and who did not receive initial steroid therapya Age, years Gender, male History of recurrent tonsillitis, yes Gross hematuria, yes Body mass index, kg/m2 Mean arterial pressure, mmHg IgA, mg/dL Estimated GFR, mL/min/1.73 m2 24-h proteinuria, mg/dL Hematuria score Mesangial score Glomeruli with segmental sclerosis or adhesion, % Endocapillary hypercellularity, yes Interstitial fibrosis, 0–25/26–50/51%1 RAS inhibitor GFR slope, mL/min/1.73 m2/year Tonsillectomy (n ¼ 20) No tonsillectomy (n ¼ 111) P-value 31.7 (24.5–38.5) 3 (15.0%) 17 (85.0%) 7 (35.0%) 21.5 (19.7–24.2) 87.7 (83.3–95.0) 319.0 (267.0–385.3) 98.5 (71.3–116.3) 138.6 (59.9–461.7) 2.00 (1.67–2.33) 0.37 (0.25–0.50) 9.0 (0.0–26.6) 4 (20.0%) 19/1/0 8 (40.0%) 0.60 6 3.65 34.3 (23.0–49.6) 43 (38.7%) 72 (64.9%) 43 (38.7%) 21.9 (20.0–24.3) 89.3 (82.9–98.0) 325.7 (250.4–442.5) 88.7 (69.5–104.7) 307.7 (119.5–790.2) 2.50 (1.67–3.00) 0.56 (0.38–0.80) 15.4 (6.7–25.0) 31 (27.9%) 100/6/5 49 (44.1%) 1.64 6 2.59 0.36 0.04 0.12 0.81 0.70 0.57 0.17 0.17 0.05 0.07 0.01 0.10 0.59 0.62 0.81 <0.01 a Values are expressed as median (interquartile range) or % and compared using Mann–Whitney’s U-test or chi-square test. GFR was associated with faster renal deterioration (Table 5). Multicollinearity did not affect the result because the variance inflation factor in each variable was <5. Operative findings Although of the 70 patients who received tonsillectomy, 12 patients had no clinical symptoms of recurrent tonsillitis or gross hematuria following tonsillitis before tonsillectomy, a substantial amount of pus attachment to re-sected tonsils was detected among the operative findings in 11 patients. In tonsils of the other one case, inflammatory response was confirmed histologically. Adverse event Serious adverse events were not observed throughout the study. Only one patient experienced hyperglycemia that needed temporary insulin injection. Discussion In this single-center historical cohort study among patients with biopsy-proven IgA nephropathy, we demonstrated that tonsillectomy was significantly associated with favorable renal outcomes such as clinical remission and delayed renal deterioration. Even in patients who did not receive initial steroid therapy, tonsillectomy was significantly associated with clinical remission and a smaller annual GFR loss during the follow-up. This reno-protective effect was independent of the known risk factors including blood pressure, proteinuria and histological findings. Although several reports examined the therapeutic effect of tonsillectomy in IgA nephropathy, the results were not consistent [9–12]. Hotta et al. [9] reported that treatment with tonsillectomy and steroid therapy was associated with clinical remission in 329 patients with IgA nephropathy. Similar results were described in a report by Komatsu et al. Downloaded from http://ndt.oxfordjournals.org/ by guest on April 7, 2012 Fig. 2. (A) Kaplan–Meier plots illustrating the probability of clinical remission among patients who had IgA nephropathy and did not receive initial steroid therapy according to tonsillectomy. (B) Kaplan–Meier plots illustrating the probability of renal survival among patients who had IgA nephropathy and did not receive initial steroid therapy according to tonsillectomy. 6 I. Maeda et al. [10] that combination therapy of tonsillectomy and pulse steroid was superior to pulse steroid alone with regard to remission of proteinuria. However, they failed to demonstrate the delaying effect of the combination therapy on renal deterioration. Since corticosteroid therapy has strong effects on resolution of urine abnormalities and sustention of renal function, evaluating the efficacy of tonsillectomy as combination therapy with steroids might cause indistinguishable results. Two retrospective observations were per- Table 4. Multivariable-adjusted hazard ratio for clinical remission in patients with IgA nephropathy who did not receive initial steroid therapy (n ¼ 131)a Clinical remissionb HR (95% CI) 3.29 (1.53–7.08) <0.01 3.05 (1.37–6.79) 0.90 (0.63–1.29) 0.84 (0.69–1.03) 0.43 (0.22–0.85) 0.70 (0.49–1.00) 0.01 0.56 0.10 0.01 0.05 3.22 (1.41–7.36) 0.47 (0.18–1.26) 0.78 (0.33–1.88) 0.82 (0.41–1.63) 0.01 0.14 0.58 0.58 0.87 (0.19–4.08) 0.49 (0.07–3.73) 0.86 0.49 3.73 (1.68–8.28) <0.01 0.65 (0.35–1.20) 0.17 a CI, confidence interval; HR, hazard ratio. Clinical remission was defined as normal dipstick examination of hematuria and proteinuria on two consecutive visits at least 3 months apart. c Age and gender were also adjusted in models 1–4. b Table 5. Regression coefficients for GFR slope among patients with IgA nephropathy who did not receive initial steroid therapy (n ¼ 131)a GFR slope (mL/min/1.73 m2/year)b b Tonsillectomy, yes Age, per 5 years Gender, male Mean arterial pressure, per 10 mmHg Baseline GFR, per 10 mL/min/1.73 m2 Log 24-h proteinuria, mg/dL Hematuria score, per L Mesangial score, per L Segmental sclerosis or adhesion, yes Interstitial fibrosis 26–50% 51%1 RAS inhibitor, yes a Standardize b P-value 2.00 (0.58–3.42) 0.11 (0.41 to 0.19) 0.57 (1.66 to 0.52) 0.35 (0.96 to 0.26) 0.40 (0.73 to 0.07) 1.21 (2.56 to 0.13) 0.10 (0.57 to 0.77) 0.06 (1.66 to 1.79) 0.20 (1.81 to 1.41) 0.25 0.11 0.09 0.13 0.35 0.22 0.03 0.01 0.02 0.01 0.48 0.30 0.26 0.02 0.08 0.78 0.94 0.81 0.45 (2.93 to 2.03) 0.08 (2.54 to 2.70) 0.02 (1.28 to 1.17) 0.04 0.01 0.01 0.72 0.95 0.97 b, regression coefficient. GFR slope was obtained using linear regression model and values of estimated GFR at every visit during the follow-up (up to 7 years) even if the patient had experienced clinical remission or 30% GFR decline earlier. b Downloaded from http://ndt.oxfordjournals.org/ by guest on April 7, 2012 Model 1 (age- and gender-adjusted model)c Tonsillectomy Model 2 (clinical factor-adjusted model)c Tonsillectomy Mean arterial pressure, 10 mmHg Baseline GFR, 10 mL/min/1.73 m2 Log 24-h proteinuria, mg/dL Hematuria score Model 3 (histological factor-adjusted model)c Tonsillectomy Mesangial score Segmental sclerosis or adhesion Endocapillary proliferation Interstitial fibrosis 26–50% 51%1 Model 4 (treatment factor-adjusted model)c Tonsillectomy RAS inhibitor P formed in the 1970s when the indication of corticosteroid in IgA nephropathy was limited. Rasche et al. [11] described that tonsillectomy had no beneficial effect on preventing ESRD. This might be partly because they included patients with relatively advanced stages: 55% of the participants had hypertension, 35% had elevated serum creatinine (>150 mmol/L) and 62% had high-grade proteinuria (>1.5 g/day). Consequently, they observed that 25% of their subjects reached ESRD in 2.3 years and failed to detect a beneficial effect of tonsillectomy. On the other hand, Xie et al. [12] reported effectiveness of tonsillectomy in a long-term observational study. They included patients with a relatively mild stage of IgA nephropathy: mean serum creatinine was 94.7 mmol/L and 38% of the patients had proteinuria of 0.5 g/day or less at baseline. Therefore, tonsillectomy might indicate a beneficial effect on renal function decline in an early and mild stage of IgA nephropathy. Strength of our study was to demonstrate tonsillectomy was associated with not only amelioration of urinary finding but also delaying renal deterioration in patients treated with both combination therapy and tonsillectomy alone in a large number of subjects with a variety of clinical and histological manifestations. The mechanisms explaining how tonsillectomy improves renal outcome are unclear. As acute deterioration of urinary findings is frequently observed at the time of upper respiratory tract or gastrointestinal infections, IgA nephropathy has been thought to be a disease of mucosal immune system. Because patients with recurrent tonsillitis showed an elevated level of serum immunoglobulin (IgA, IgG and IgM), increased IgA synthesis has been thought to play an important part in the pathogenesis of IgA nephropathy [19]. In addition, anomaly in the glycosylation of IgA molecules may play another important part in the pathogenesis [20]. Altered glycosylation leads to a decrease in the clearance of IgA1 molecules by the liver [21] and increases binding of IgA1 to glomerular mesangium [22]. These two mechanisms are related, and elevated synthesis and reduced clearance of Tonsillectomy and IgA nephropathy Conflict of interest statement. None declared. References 1. D’Amico G. Natural history of idiopathic IgA nephropathy and factors predictive of disease outcome. Semin Nephrol 2004; 24: 179–196 2. Praga M, Gutierrez E, Gonzalez E et al. Treatment of IgA nephropathy with ACE inhibitors: a randomized and controlled trial. J Am Soc Nephrol 2003; 14: 1578–1583 3. Coppo R, Peruzzi L, Amore A et al. IgACE: a placebo-controlled, randomized trial of angiotensin-converting enzyme inhibitors in children and young people with IgA nephropathy and moderate proteinuria. J Am Soc Nephrol 2007; 18: 1880–1888 4. Li PKT, Leung CB, Chow KM et al. Hong Kong study using valsartan in IgA nephropathy (HKVIN): a double-blind, randomized, placebocontrolled study. Am J Kidney Dis 2006; 47: 751–760 5. Pozzi C, Andrulli S, Del Vecchio L et al. Corticosteroid effectiveness in IgA nephropathy: long-term results of a randomized, controlled trial. J Am Soc Nephrol 2004; 15: 157–163 6. Pozzi C, Bolasco PG, Fogazzi GB et al. Corticosteroids in IgA nephropathy: a randomised controlled trial. Lancet 1999; 353: 883–887 7. Samuels JA, Strippoli GFM, Craig JC et al. Immunosuppressive treatments for immunoglobulin A nephropathy: a meta-analysis of randomized controlled trials. Nephrology 2004; 9: 177–185 8. Appel GB, Waldman M. The IgA nephropathy treatment dilemma. Kidney Int 2006; 69: 1939–1944 9. Hotta O, Miyazaki M, Furuta T et al. Tonsillectomy and steroid pulse therapy significantly impact on clinical remission in patients with IgA nephropathy. Am J Kidney Dis 2001; 38: 736–743 10. Komatsu H, Fujimoto S, Hara S et al. Effect of tonsillectomy plus steroid pulse therapy on clinical remission of IgA nephropathy: a controlled study. Clin J Am Soc Nephrol 2008; 3: 1301–1307 11. Rasche FM, Schwarz A, Keller F. Tonsillectomy does not prevent a progressive course in IgA nephropathy. Clin Nephrol 1999; 51: 147–152 12. Xie YS, Nishi S, Ueno M et al. The efficacy of tonsillectomy on longterm renal survival in patients with IgA nephropathy. Kidney Int 2003; 63: 1861–1867 13. AAO-HNS. Clinical indicators: tonsillectomy, adenoidectomy, adenotonillectomy. http://www.entnet.org/practice/products/indicators/ tonsillectomy.html (Accessed July 19th, 2011) 14. Xie YS, Chen XM, Nishi S et al. Relationship between tonsils and IgA nephropathy as well as indications of tonsillectomy. Kidney Int 2004; 65: 1135–1144 15. Matsuo S, Imai E, Horio M et al. Revised equations for estimated GFR from serum creatinine in Japan. Am J Kidney Dis 2009; 53: 982–992 16. Cattran DC, Coppo R, Cook HT et al. The Oxford classification of IgA nephropathy: rationale, clinicopathological correlations, and classification. Kidney Int 2009; 76: 534–545 17. Roberts ISD, Cook HT, Troyanov S et al. The Oxford classification of IgA nephropathy: pathology definitions, correlations, and reproducibility. Kidney Int 2009; 76: 546–556 18. Peduzzi P, Concato J, Feinstein AR et al. Importance of events per independent variable in proportional hazards regression analysis 2: accuracy and precision of regression estimates. J Clin Epidemiol 1995; 48: 1503–1510 19. Lal H, Sachdeva OP, Mehta HR. Serum immunoglobulins in patients with chronic tonsillitis. J Laryngol Otol 1984; 98: 1213–1216 20. Horie A, Hiki Y, Odani H et al. IgA1 molecules produced by tonsillar lymphocytes are under-O-glycosylated in IgA nephropathy. Am J Kidney Dis 2003; 42: 486–496 21. Rifai A, Fadden K, Morrison SL et al. The N-glycans determine the differential blood clearance and hepatic uptake of human immunoglobulin (Ig)A1 and IgA2 isotypes. J Exp Med 2000; 191: 2171–2181 22. Allen AC, Harper SJ, Feehally J. Galactosylation of N-linked and O-linked carbohydrate moieties of IgA1 and IgG in IgA nephropathy. Clin Exp Immunol 1995; 100: 470–474 Received for publication: 22.7.11; Accepted in revised form: 2.2.12 Downloaded from http://ndt.oxfordjournals.org/ by guest on April 7, 2012 abnormal IgA1 may form a vicious circle. Tonsillectomy may provide resolution upstream of the pathogenesis and induce remission of the disease. Our study had several limitations. Firstly, participants of the current study had relatively mild cases of IgA nephropathy compared with those of previous reports. Because all students at schools and subjects at worksites are required to undergo annual health screenings by Japanese law, a mild and early stage of IgA nephropathy might be diagnosed more frequently in Japan than in countries without such a screening system. IgA nephropathy is a slowly and gradually progressive disease with repeated episodes of ‘flare-ups’ at each occurrence of upper respiratory infection. Tonsillectomy can be a useful therapeutic option especially in an early stage of IgA nephropathy with preserved renal function. Secondly, because this study was a historical cohort study, tonsillectomy was not randomly assigned. Tonsillectomy and steroid therapy were frequently performed after 2001, which was the year of publication by Hotta et al. [9]. Their concept was that combination therapy of steroids and tonsillectomy was associated with clinical remission, and a patient who achieved clinical remission sustained renal function over a long period of time. In our institute, tonsillectomy was administered as a part of the combination therapy if a patient wished to undergo this therapeutic option to achieve clinical remission of IgA nephropathy even if their clinical manifestations were not so severe. Relatively recent cases tended to undergo tonsillectomy. This decision-making process could cause the difference of follow-up length between the two groups and substantial selection bias. Thirdly, a significant relationship was not observed between histological findings and renal outcomes. In addition, patients who were prescribed RAS inhibitors showed a higher risk for renal function decline. Mild histological findings and treatment with RAS inhibitors are known as factors associated with favorable renal outcome. They might reflect referral bias. Patients who revealed advanced disease in biopsy specimen or had hypertension or high-grade proteinuria would undergo aggressive treatment. Even if we applied a multivariable analysis, we might not be able to control for these confounders entirely. Finally, difference of the incidence of 30% GFR decline between tonsillectomy and non-tonsillectomy group among non-steroid-treated subjects did not reach a statistical significance in log-rank test. We thought that this was due to the limited number of subjects treated with tonsillectomy alone. In conclusion, tonsillectomy was associated with favorable renal outcomes of clinical remission and delayed renal deterioration. This association was observed not only in patients with combination therapy but also in patients with tonsillectomy alone. This GFR preservation effect was independent of other known factors including treatment with RAS inhibitors. Tonsillectomy should be considered in patients with IgA nephropathy, especially at a mild or early stage, to prevent future renal deterioration. 7