Abstract
Management of obstetrical acute renal failure remains a challenging task. We present data of 100 cases of obstetrical -related acute renal failure of 3-year duration (2007–2009) from Department of Nephrology & Hypertension, Lady Reading Hospital, Peshawar, Pakistan. The study is aimed to look at overall mortality and relationship of oliguria/anuria at presentation to dialysis dependency and renal cortical necrosis (RCN). Evaluation of comorbidity to dialysis dependency and RCN was also considered. While 91 patients required hemodialysis, 9 were managed conservatively; 57 were dialysis dependent whereas 43 remained dialysis independent on discharge; 47 patients had oliguria, 30 had anuria, and 23 had an output of >800 mL per 24 h on admission. RCN was seen in 30 cases, all biopsy confirmed; among these, 26 cases (86.67%) were associated with oliguria/anuria and dialysis dependency right from the beginning (p < 0.0001). However, four (13.33%) with RCN had output >800 mL per 24 h but remained dialysis dependent. Our data showed that out of 30 patients who presented with anuria, only 10 patients (33.33%) were dialysis independent on discharge, whereas out of 47 oliguric patients, 21 patients (44.6%) were dialysis independent upon discharge. Thus dialysis dependency does not correlate with anuria or oliguria at presentation (p = 0.133). Mortality of 7% was recorded; 23% were discharged with normal renal function. Septicemia, operative interventions, retained product of conception, post-partum hemorrhage, and RCN remained important comorbid conditions with regard to survival and dialysis dependency.
INTRODUCTION
Obstetrical acute renal failure (ARF) remains a common nephrological challenge in third-world countries. In this regard, post-partum hemorrhage (PPH), home deliveries, untrained traditional birth attendants (TBAs), grand multiparity, low hemoglobin (Hb), septicemia, delayed referral, and lack of proper initial management at home level have been important contributing factors to the high morbidity and mortality.Citation1,Citation2 This has been clearly shown in the earlier study published from our department regarding obstetrical ARF with maternal mortality of 18%.Citation3
Although ARF in pregnancy can be caused by any factor that affects general population, blood loss due to ante-partum hemorrhage (APH), PPH, placenta previa, and puerperal sepsis usually sets up the ground for injury at the endothelial level. However, this is many times augmented by pre-eclamptic toxemia, eclampsia, and HELLP syndrome (hemolytic anemia, elevated liver enzymes, and low platelet count).Citation4–7 Renal ischemia, once prolonged, would lead to irreversible cortical injury and associated high mortality. Studies from PakistanCitation8,Citation9 have shown mortality of 16–26% over 1-year period in evaluation of 43 and 42 patients, respectively.
Renal cortical necrosis (RCN) remains a dreaded complication of obstetrical ARF, associated with increased dialysis dependency and mortality. Recently, an Indian study has shown decreased incidence of RCN, in relation to obstetrical ARF.Citation10 This is in contrast to our present data which over a 3-year period have shown that 30% of our obstetrical ARF had biopsy-proven RCN. This study is a prospective observational study. It is aimed to assess overall mortality and association of oligo/anuria with dialysis dependency and RCN. The prevalence of RCN and impact of comorbid conditions such as septicemia, low Hb on admission, retained products of conception, amount of blood transfused, duration of dialysis, and RCN on final outcome were also assessed.
MATERIALS AND METHODS
ARF on average represents 18–20% of our total annual admissions. Obstetrical-related ARF constitutes 9–11% of our ARF series. This study was conducted in the Department of Nephrology & Hypertension, Post Graduate Medical Institute (PGMI), Lady Reading Hospital over a period of 3 years from 2007 to 2009. It is an observational hospital-based clinical study aiming at the evaluation of patient survival, mortality, and relationship of urinary output to dialysis dependency. In addition, their association with RCN was also evaluated. Finally, the impact of comorbidities such as septicemia, low hemoglobin on admission, retained products of conception, amount of blood transfused, PPH, duration of dialysis, and RCN on the final outcome were assessed.
Oliguria was defined as urine output <400 mL per 24 h while anuria as no or <50 mL per 24 h. Those with urine output >800 mL per 24 h were grouped separately. Patients who were dependent on dialysis for survival were labeled as dialysis dependent. Those patients who had either regained normal renal function as per hospital labs, that is, serum creatinine <1.5 mg/dL and blood urea <50 mg/dL, or had some degree of renal impairment but did not require dialysis for their survival or to maintain quality of life were labeled as dialysis independent.
Patients who were labeled RCN had a biopsy-confirmed diagnosis. Kidney biopsy was offered to all those patients who remained oligo/anuric >3 weeks or to those who remained dialysis dependent >3 weeks despite having urine output >800 mL per 24 h.
All patients admitted in nephrology department with history of obstetrical-related events from 2007 to 2009 were included in the study. This included patients with normal vaginal delivery, cesarean sections, septic abortions, cesarean hysterectomies, and postnatal septicemia. For a patient to be included in the study, he/she must have an obstetrical event leading to acute kidney injury with serum creatinine >1.5 mg/dL and urea >55 mg/dL within 24 h or 25% rise of urea and creatinine from baseline. In addition, volume of urine per 24 h was also recorded on admission and patients were classified as anuric or oliguric as defined. Routine urine analysis was not possible in anuric patients. However, those with oliguria and urinary output >800 mL per 24 h have shown rather benign sediments including few RBCs, albumin +1/+2, and WBCs 5–10/HPF (high power field).
Patients with history of prior gynecological intervention, renal stones, diabetes mellitus, glomerulonephritis, hypertensive kidney disease, nonsteroidal anti-inflammatory drugs (NSAIDs) abuse, and autoimmune disorders were excluded. Patients with history of proteinuria or hypertension prior to gestation were also excluded so that only cases with acute kidney injury secondary to obstetrical events were included and hence the conclusion remained unskewed.
All the patients included in the study underwent a complete clinical examination. Detailed note was made of clinical and obstetrical history, mode of delivery, degree of blood loss, and urinary output. Basic parameters such as blood pressure, anemia, urea, serum creatinine, serum electrolytes, complete blood count, hepatitis status, liver function tests, abdominal ultrasound, echocardiogram, and disseminated intravascular coagulation screen were recorded. Patients requiring dialysis were dialyzed on Fresenius 4008S machines with On-line Clearance Monitor (OCM). Bicarb dialysis was used with poly-sulfone Gambro kidneys having a size of 1.2–1.5 m.
Hospital outcomes included discharge from hospital with normal renal function tests, discharge with urinary output >800 mL with impaired renal function but dialysis independent, discharge with dialysis dependency, and death.
Statistical Analysis
Chi-square test of independence (ΞCitation2) was used to find the statistical significance of various associations. Results were considered statistically significant with p-value <0.05 at 95% level of significance, α = 0.05.
RESULTS
Of 100 patients evaluated, mean age was found to be 30.35 years with maximum of 50 years and minimum of 17 years. Mean hemoglobin recorded on admission was 8.17 g/dL with a minimum of 3.4 g/dL. Average temperature recorded on admission was 99.97°F. Mean creatinine and urea on admission were 10.14 and 212.3 mg/dL, respectively, whereas creatinine and urea on discharge were 4.57 and 98.32 mg/dL, respectively.
Mode of Delivery
Fifty-three patients (53%) had deliveries assisted by doctors, 28 patients (28%) were assisted by TBAs, and 15 patients (15%) had home deliveries whereas 4 patients (4%) had deliveries assisted by lady health visitors.
Method of Delivery
Sixty-one deliveries (61%) were normal vaginal deliveries and remaining 39 deliveries (39%) had some operative interventions. Among them, 34 patients (87.2%) were operated through C-section, 6 patients (15.4%) underwent hysterectomy, whereas 2 patients (5.1%) had laparotomy. Out of the six patients who underwent hysterectomy, three patients had hysterectomy for severe PPH following C-section.
Urine Output on Admission
Forty-seven patients (47%) had oliguria, 30 patients (30%) had anuria, whereas 23 patients (23%) had urine output greater than 800 mL per 24 h.
Among 47 oliguric patients, 26 patients (55.3%) were dialysis dependent and 21 patients (44.7%) were dialysis independent on discharge. Seventeen patients (36.2%) had RCN on biopsy.
In 30 anuric patients, 20 patients (66.6%) were dialysis dependent and 10 patients (33.3%) were dialysis independent on discharge. Nine patients (30%) had RCN on biopsy.
Twenty-three patients had urine output >800 mL per 24 h and among them 11 patients (47.8%) were dialysis dependent and 12 patients (52.2%) were dialysis independent on discharge. Four patients (17.4%) had RCN on biopsy ( and ).
Figure 1. Comparison of urinary output on admission with dialysis dependency.
Figure 2. Association of renal cortical necrosis with clinical presentation.
Urine Output and Cortical Necrosis
Out of 77 patients who had oligo/anuria on admission, only 26 patients (33.8%) had RCN, whereas among 23 patients with urine output >800 mL per 24 h, only 4 patients (17.4%) had RCN. Thus oliguria/anuria on admission proved to be independent of renal injury and its association with RCN was found to be statistically insignificant (p = 0.133).
Urine Output and Dialysis Dependency
Among 23 patients with urine output >800 mL per 24 h, only 11 patients (47.8%) were dialysis dependent, whereas out of 77 patients with oliguria/anuria, 46 patients (59.7%) were dialysis dependent. However, this association did not prove to be statistically significant (p = 0.311) and thus was not in good agreement with the general consideration that nonoliguric patients have better outcome and less dialysis dependency ().
Table 1. Dialysis dependency versus clinical outcomes
Biopsy-Proven Cortical Necrosis
On biopsy, overall 30 patients (30%) were diagnosed with RCN. Four of them had associated acute tubular necrosis (ATN) and one had additional features of chronic tubular interstitial nephritis. Among them, 26 patients (86.6%) were dialysis dependent and only 4 (13.3%) patients were dialysis independent. On admission, 17 out of 30 patients (56.6%) were oliguric, 9 patients (30%) were anuric, and 4 patients (13.3%) had an output >800 mL per 24 h. RCN turned out to be a bad prognostic feature as 86.6% of these patients were dialysis dependent, and therefore it was significantly associated with dialysis dependency right from the beginning (p < 0.0001).
Operative Interventions
Operative intervention turned out to be a bad prognostic event. Out of 39 operative procedures, 34 patients (80.5%) underwent an emergency C-section. Among them, 24 patients (70.6%) were dialysis dependent on discharge and 9 patients (26.5%) had biopsy-proven RCN. Six patients (15.3%) had total abdominal hysterectomy (three had hysterectomy following C-section), and five of them were dialysis dependent on discharge. Two patients had laparotomy for septic abortion and ectopic pregnancy, and both were dialysis dependent on discharge. Hence, operative intervention turned out to be a comorbid condition as 31 out of 39 patients (79.5%) were dialysis dependent on discharge.
Comorbid Conditions
PPH was seen in 12 patients (12%). Out of them, eight patients (66.7%) were dialysis dependent on discharge. Seven patients (58.3%) had biopsy-proven RCN and associated ATN. Hence, PPH was an additional bad prognostic feature. Sixteen patients were septicemic on admission. Out of them, eight patients (50%) were dialysis dependent on discharge and four patients had biopsy-proven RCN.
Retained product of conception (RPOC) on pelvic ultrasound was seen in 21 patients (21%). Among them, 13 patients (62%) were dialysis dependent on discharge. Eight patients were unconscious on admission. Out of them, seven patients (87.5%) were dialysis dependent on discharge. Out of eight patients, three patients (37.5%) had RCN and three patients (37.5%) died. The duration of dialysis, the number of dialysis sessions, and the number of blood transfusions were not associated with dialysis dependency, RCN, or death.
Overall mortality of 7% during hospital stay was noted. Out of them, three patients were dialysis dependent, whereas the other four were independent of dialysis. Among the dialysis-dependent patients, two patients had RCN. Septicemia, hyperkalemia, and pulmonary edema were the culprits among patients not requiring hemodialysis. Twenty-three patients (23%) were discharged home with normal kidney function.
DISCUSSION
It is claimed that acute renal failure has become a rare complication of pregnancy in western population.Citation2,Citation11–13 This is also in good agreement with a study from the United States, showing only 4% pregnancy-related ARF with a mortality of 1%.Citation14 However, this is in contrast to a country like Pakistan where acute obstetrical renal failure contributes significantly to pregnancy-related morbidity and mortality.
In our setup, obstetrical-related acute renal failure still constitutes 10% of ARF cases. The hospital-based mortality has reduced from 18% in 2004Citation3 to 7% in 2007–2009 in this study. This result is comparable with a study showing reduction in obstetrical-related ARF from 22% to 9% over a 9-year period.Citation15 Studies from AfricaCitation16 and IndiaCitation17 have shown mortality ranging from 9% to 29% in a patient with eclampsia. In a study, evaluation of 325 patients undergoing dialysis over an 11-year period has shown a mortality of 55.3%, quoting late referral, frequent sepsis, and high incidence of bilateral cortical necrosis as the major contributing factors to high mortality.Citation18 Similarly, a study from SIUT, Karachi, Pakistan, has shown a mortality of 23% in their study group over a 1-year period.Citation19 An Italian study has shown that the incidence of pregnancy-related ARF fell from 43% to 0.5% with respect to total number of ARF and from 1/3000 to 1/18000 with respect to total number of pregnancies. Maternal mortality fell from 31% to nil during four successive periods: 1956–1967, 1968–1977 (PubMed), 1978–1987, and 1988–1994.Citation12 A study from South Africa has shown a decreased incidence of pregnancy-related ARF requiring hemodialysis, from 24.6% in 1978 to 16% in 1992.Citation20 Likewise, a number of studies from Southeast Asia and Africa have shown maternal mortality varying from 10% to 55.3%.Citation18,Citation20
It is believed that oliguria/anuria on admission represents a severe degree of renal injury and reflects in increased hospital morbidity and mortality.Citation21,Citation22 However, this did not turn out to be the case in the current data, as out of 30 anuric patients only 20 patients (66.66%) were dialysis dependent whereas in oliguric group, out of 47 patients, 26 patients (55.3%) were dialysis dependent. Therefore, out of a total of 77 oliguric/anuric patients, 46 patients (59.7%) were dialysis dependent on discharge. It is noteworthy that among the 23 patients with urine output >800 mL on admission, 11 patients (47.8%) were dialysis dependent on discharge. This rather interesting observation in our data is important as this clearly proves that oliguria/anuria on admission does not reflect dialysis dependency on discharge (p = 0.311). Biopsy-proven RCN was shown in almost one-third of our cases (30%). This is comparable to RCN of 23.25%, 26.2%, and 23.8% (15 out of 63 patients) shown in various studies, respectively.Citation8,Citation9,Citation23 Massive blood loss requiring 4–7 pints of blood transfusions, APH, operative intervention, PPH, eclampsia, and septicemia were contributing factors for RCN seen in 30% of the cases.
Considering no relationship of dialysis dependency to oliguria/anuria, RCN has shown a very strong correlation to the urinary output on admission in terms of dialysis dependency as 26 out of 30 patients (86.6%) were dialysis dependent on discharge (p < 0.0001). Similar results were shown in a study where total anuria was the commonest presenting feature, noted in 78.8% of patients with RCN.Citation21 A number of comorbid conditions, contributing to dialysis dependency, morbidity, mortality, and RCN, were also evaluated as a part of our study. Operative interventions (cesarean section, hysterectomy, and laparotomy) were associated with bad prognosis in terms of dialysis dependency. Out of 39 patients who had operative intervention, 79.5% (31 patients) had dialysis dependency on discharge.
Our data also showed that out of 12 patients with PPH, 8 patients (66.7%) were dialysis dependent on discharge and 7 patients (58.3%) had biopsy-proven RCN.
In septicemic cases (16 patients), 50% were dialysis dependent and 25% had RCN. All patients in the study had abdominal/pelvis ultrasound on admission. Though not mentioned in literature, it was found that RPOC contributed significantly to dialysis dependency (62%). It is worth mentioning that all of these patients with RPOC had to undergo evacuation and curettage, more than once in few cases. These interventions led to higher chance of septicemia, hypotension, and blood loss, thus increasing patients' risk to dialysis dependency.
Patient's state of consciousness on admission (mostly eclamptic) was also related to higher morbidity and mortality. Seven out of eight patients admitted in state of unconsciousness (87.5%) were dialysis dependent and three patients (37.5%) had RCN.
Number of dialysis sessions, length of dialysis period, and number of blood transfusions needed bore no relationship to mortality, dialysis dependency, or RCN. Twenty-three of our patients were able to go home with normal kidney function, that is, serum creatinine <1.5 mg/dL and blood urea of <50 mg/dL. The overall patient survival was 93% on discharge with seven patients dying during hospital stay. Thus our study highlights decreased hospital mortality of patients with obstetrical-related ARF and this is in consistency with other studies.Citation24
Our study had limitations of shorter duration and prognosis that is based on duration of stay in the hospital. Hence long-term follow-up of these patients would have been a valuable tool to assess long-term morbidity and mortality and dialysis dependency. However, the study clearly shows that hospital-based mortality of obstetrical ARF has improved. This dreadful condition with associated morbidity and mortality of 50%Citation12 can be managed adequately. In this regard, early referral, facilities for blood transfusions, expert nephrology services, and round-the-clock availability of hemodialysis add to the patients' care with reduction in morbidity and mortality. All of these facilities are available in our teaching hospital to account for better outcome of these patients. In addition, aggressive management of eclampsia, adequate obstetrical services, treatment of septicemia, and careful management of fluid and electrolyte balance greatly reduce the mortality.
CONCLUSION
Hospital-based mortality secondary to obstetrical ARF has come down to 7%, and 93% of our patients were discharged home. Oliguria/anuria on admission bears no relationship to dialysis dependency. However, RCN is strongly related to oliguria/anuria at presentation and dialysis dependency (p < 0.0001). We feel that timely referral, full-time gynecological and obstetrical services at the district level, availability of hemodialysis and nephrology care, and timely institution of renal replacement therapy can further reduce morbidity and mortality of obstetrical-related ARF.
ACKNOWLEDGMENTS
The authors acknowledge Dr. Anjum Mehmood for her endless efforts during the process of manuscript preparation. Efforts of Dr. Saadia Izzat during the process of data extraction are also acknowledged.
Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
REFERENCES
- Gammill HS, Jeyabalan A. Acute renal failure in pregnancy. Crit Care Med. 2005;33(10 Suppl.):S372–S384.
- Hachim K, Badahi K, Benghanem M, Obstetrical acute renal failure. Experience of the nephrology department, Central University Hospital ibn Rochd, Casablanca. Nephrologie. 2001;22(1):29–31.
- Ali A, Zafar S, Mehmood A, Nisar A. Obstetrical acute renal failure from Frontier Province: A 3 year prospective study. J Postgrad Med Inst. 2004;18(1):109–116.
- Mjahed K, Alaoui SY, Barrou L. Acute renal failure during eclampsia: Incidence risks factors and outcome in intensive care unit. Ren Fail. 2004;26(3):215–221.
- Drakeley AJ, Le Roux PA, Anthony J, Penny J. Acute renal failure complicating severe preeclampsia requiring admission to an obstetric intensive care unit. Am J Obstet Gynecol. 2002;186(2):253–256.
- Gul A, Aslan H, Cebeci A, Polat I, Ulusoy S, Ceylan Y. Maternal and fetal outcomes in HELLP syndrome complicated with acute renal failure. Ren Fail. 2004;26(5):557–562.
- Silva GB, Jr., Monteiro FA, Mota RM, . Acute kidney injury requiring dialysis in obstetric patients: A series of 55 cases in Brazil. Arch Gynecol Obstet. 2009;279(2):131–137.
- Hassan I, Junejo AM, Dawani ML. Etiology and outcome of acute renal failure in pregnancy. J Coll Physicians Surg Pak. 2009;19(11):714–717.
- Ansari MR, Laghari MS, Solangi KB. Acute renal failure in pregnancy: One year observational study at Liaquat University Hospital, Hyderabad. J Pak Med Assoc. 2008;58(2):61–64.
- Prakash J, Vohra R, Wani IA, Decreasing incidence of renal cortical necrosis in patients with acute renal failure in developing countries: A single-centre experience of 22 years from Eastern India. Nephrol Dial Transplant. 2007;22(4):1213–1217.
- Stratta P, Canavese C, Dogliani M, Pregnancy related acute renal failure. Clin Nephrol. 1989;32(1):14–20.
- Stratta P, Besso L, Canavese C, . Is pregnancy related acute renal failure a disappearing clinical entity? Ren Fail. 1996;18(4):575–584.
- Grunfeld JP, Pertuiset N. Acute renal failure in pregnancy: 1987. Am J Kidney Dis. 1987;9(4):359–362.
- Mattar F, Sibai BM. Eclampsia. VIII. Risk factors for maternal morbidity. Am J Obstet Gynecol. 2000;182(2):307–312.
- Chugh KS, Sakhuja V, Malhotra HS, Pereira BJ. Changing trends in acute renal failure in third-world countries – Chandigarh study. Q J Med. 1989;73(272):1117–1123.
- Oladokun A, Okewole AI, Adewole IF, Babarinsa IA. Evaluation of cases of eclampsia in the University College Hospital, Ibadan over a 10 year period. West Afr J Med. 2000;19(3):192–194.
- Sawhney H, Aggarwal N, Biswas R, Vasishta K, Gopalan S. Maternal mortality associated with eclampsia and severe preeclampsia of pregnancy. J Obstet Gynaecol Res. 2000;26(5):351–356.
- Chugh KS, Singhal PC, Sharma BK, Acute renal failure of obstetric origin. Obstet Gynecol. 1976;48(6):642–646.
- Naqvi R, Akhtar F, Ahmed E, Acute renal failure of obstetrical origin during 1994 at one center. Ren Fail. 1996;18(4):681–683.
- Randeree IG, Czarnocki A, Moodley J, Seedat YK, Naiker IP. Acute renal failure in pregnancy in South Africa. Ren Fail. 1995;17(2):147–153.
- Chugh KS, Jha V, Sakhuja V, Joshi K. Acute renal cortical necrosis – a study of 113 patients. Ren Fail. 1994;16(1):37–47.
- Loo CS, Zainal D. Acute renal failure in a teaching hospital. Singapore Med J. 1995;36(3):278–281.
- Prakash J, Tripathi K, Pandey LK, Gadela SR, Usha. Renal cortical necrosis in pregnancy-related acute renal failure. J Indian Med Assoc. 1996;94(6):227–229.
- Prakash J, Kumar H, Sinha DK, Acute renal failure in pregnancy in a developing country: Twenty years of experience. Ren Fail. 2006;28(4):309–313.