Arch Gynecol Obstet DOI 10.1007/s00404-013-3094-3
GENERAL GYNECOLOGY
Seasonal trend of acute pelvic inflammatory disease Anjeza Xholli • Marianna Cannoletta Angelo Cagnacci
•
Received: 20 July 2013 / Accepted: 11 November 2013 Ó Springer-Verlag Berlin Heidelberg 2013
Abstract Purpose Many infections follow a seasonal trend. Aim of our study was to check whether acute pelvic inflammatory disease (PID) follows a seasonal progress. Methods In a retrospective study on 12,152 hospital records, 158 cases of acute pelvic inflammatory disease were identified. Periodogram analysis was applied to the date of pelvic inflammatory disease admission and to related environmental factors, such as temperature and photoperiod. Results Pelvic inflammatory disease follows a seasonal rhythm with mean to peak variation of 23 % and maximal values in September (±37.2 days). The rhythm, more evident in married women, is related to the rhythm of temperature advanced by 2 months and of photoperiod advanced by 3 months. Cases of pelvic inflammatory disease are more frequent than expected in unmarried (36 vs. 17.3/34,626, p = 0.015), particularly divorced women 30–40 years of age. Conclusions Our study evidences a seasonal trend and confirms unmarried, particularly divorced status, as important risk factor for acute pelvic inflammatory disease. Keywords Biological rhythms Epidemiology Photoperiod Season Pelvic inflammatory disease
Introduction Pelvic inflammatory disease (PID) is a clinical syndrome defined as a spectrum of upper genital tract infections that
A. Xholli M. Cannoletta A. Cagnacci (&) Obstetrics and Gynaecology Unit, Department of Obstetrics Gynaecology and Paediatric, Azienda Policlinico of Modena, Via del Pozzo 71, 41121 Modena, Italy e-mail: [email protected]
includes any combination of endometritis, salpingitis, pyosalpinx, tubo-ovarian abscess and pelvic peritonitis [1]. It is estimated that in US, over 1 million women are affected of an episode of PID, including 11 % of all fertile women [2]. After a single episode of PID, one out of four women will develop a serious complication, such as recurrent PID, chronic abdominal pain, infertility [3], ectopic pregnancy, fixed uterine retroversus-flexion associated with dyspareunia and dysmenorrhea [2]. Reported risk factors for PID include: young age, single or divorced marital status, low socio-economic status, black race, high frequency of partner change, placement of an intrauterine device, sexual transmitted disease, lack of barrier contraception, bacterial vaginosis, recurrent cervicitis [4, 5]. Beside these risk factors, evidences that pathogen transmission and immune activity can be influenced by changes in ultraviolet exposure, in photoperiod or in temperature, suggest that infections may have a seasonal trend [6]. Indeed, most of the sexually transmitted disease shows seasonal fluctuations with peak incidence in the third quarter of the year [7–9] or in sunny months [10]. Others, carried out in the USA have confirmed seasonal trend in the infection by chlamydia and gonorrhea with infection peaks in August and September [11, 12]. In this study, we evaluated whether acute PID requiring hospitalization does follow a seasonal rhythm, similar to the one reported for chlamydia and gonorrhea infection.
Materials and methods Data on 12,152 women hospitalized in the period 2005–2011 were obtained from the database of the Institute
123
Arch Gynecol Obstet
of Obstetrics and Gynecology in Modena, Italy. Data were anonymously extracted by medical personnel licensed to manage the database. According to the letter of discharge, registered in the hospital database, we selected those women with a diagnosis of PID, namely 195 patients. Every clinical record of these patients was checked. Among these we selected only those patients with a diagnosis of PID that were admitted to the hospital, namely 158 women. Acute PID was defined as a patient with fever, high value of PCR, leukocytosis, leucorrhea, and pelvic pain exacerbated by the gynecological exam, requiring prolonged antibiotics therapy with or without surgery [4, 13]. The diagnosis was supported by an ultrasound investigation of the pelvis showing the typical features of acute PID: pyosalpinx with thick walls of the tube, incomplete septa within it, ovoid/retort shaped, intraluminal fluid, ‘cogwheel sign’. Tubal-ovarian involvement: tubo-ovarian complex with sonographic attributes of salpingitis with ovary and tube discernible but that cannot be pushed apart; tubo-ovarian abscess with extreme motion tenderness at the probe pressure, the ovary and the tube are confluent with speckled fluid [14]. For each single woman the following data were collected from the clinical record: date of admission, date of birth, age, menarche, number of spontaneous deliveries, cesarean sections, ectopic pregnancy, spontaneous abortion, voluntary interruption of pregnancy, last menses, nationality, marital status (married, divorced, single), occupation (professional woman, clerk, worker, housewife), previous abdominal surgery, use of medicines, contraceptive therapy (oral contraceptives or intrauterine device), cervical and vaginal swab specimen results, smoking and treatment during hospitalization (medical, surgical). We assessed the seasonal progress of PID during the year confronting it with that of some atmospheric factors such as the daily average temperature per month in °C and photoperiod, calculated by daily average light hours per month. Data on atmospheric factors in our area for the years 2005–2011 were obtained from the ISTAT (National Institute of Statistics) register. To calculate the prevalence of PID in exposed population we extracted from the ISTAT register the number of women living in Modena in the years 2005–2011 and chose those belonging to the group of age present in our cases (18–54 years old). Marital status was also extracted. Circa-annual rhythmic distribution of PID was evaluated by periodogram analysis using the RHYTHM program. The periodogram was adapted to analyze the 12 months rhythm. As originally described by Van Cauter [15], the
123
method initially tests the significance of the observed time fluctuations against the hypothesis of their purely random occurrence via two different tests. The alternative to pure randomness is, for the first test, the existence of local correlations, implying that values at given times depend on values at other times, and for the second test, the existence of periodic fluctuations. When the hypothesis of random occurrence of the data is rejected, the periodogram method is applied to detect and estimate the possible significant components. A sum of sinusoidal components with periods equal to integer divisors of the observation span (i.e. 12 months/1; 12 months/2; 12 months/3; etc.) is fitted on the series. A decision procedure devised by Fisher allows the selection of the significant periodicity underlying the process at a given probability of p \ 0.05. Up to the first three significant periodical components are included in the theoretical description of the profile. A theoretical pattern is computed according to the minimum of the residual sum of squares. The amplitude of the theoretical pattern is defined as half the difference between the maximum and minimum of the theoretical pattern, while the acrophase is the time corresponding to the occurrence of the first global maximum of the theoretical curve. Both the amplitude and the acrophase are furnished with a confidence interval. Periodograms are considered significantly different when described by different periodical components or alternatively when the confidence intervals of either the amplitude or acrophase do not overlap. Statistical analyses were performed by the statistical Package Statview 5.0.1 for Apple Machintosh (SAS Institute Inc, Cary, NC, USA1998). The seasonal progress of PID in the 7 years of the study is expressed as monthly average and related, by regression analysis, to the seasonal course of monthly average temperature and photoperiod evaluated in the same years. Rhythms of temperature and photoperiod were shifted up to 3 months both in advance and in delay to highlight possible phase shift between PID and atmospheric factors. Prevalence of PID in different months of the year and in different group of women was compared by contingency table associated with the Chi-square test. A value of p \ 0.05 was considered as significant.
Results Characteristics of women with acute PID are reported in Table 1. Periodogram analysis showed that the average monthly distribution of acute PID takes a seasonal progress illustrated by a sinusoidal curve that stretches for a period of 12 months with a mean value per month of 1.88, peak in
Arch Gynecol Obstet Table 1 Characteristics of women with pelvic inflammatory disease hospitalized in the period 2005–2011 N = 158
Age
18–54 years
Mean age
35.4 years
Mean age of menarche
12.6 years
Nationality
European 84.8 %
cases/month
Women with PID
2.2
1.8 1.4
Asiatic 2.5 %
TEMPERATURE
African 11.4 %
20
South-American 1.3 % Married 45.6 %
15
ºC
Marital status
PID
2.6
Single 43.7 %
10
Divorced 10.7 % Occupation
Professional woman 16.1 %
5
Clerk 36 % PHOTOPERIOD
Worker 26.7 % Unemployed 6.6 % Parity
Nullipara 41.6 % Multipara 56.6 %
Previous surgery
light hours
Housewife 14.6 %
16
14
12
Appendix 30.4 % Cesarean section 10 %
10
Adnexal surgery 8 %
J
F M A M
Abortion 26.0 % Contraception
Abdominal surgery 2.5 % IUD 6.9 %
J
J
A S O N
D
Months Fig. 1 Seasonal trend of monthly average of pelvic inflammatory disease, temperature and photoperiod
Oral Contraceptive 5 % Smoking
7.6 %
Cervical–vaginal swab
31.6 % Chlamydia trachomatis
60
50
3.7 % Neisseria gonorrhoeae
Divorced Single Married
1.2 % Mycoplasma genitalium 40
Inpatient therapy
Antibiotics 51.9 % Surgery 48.1 %
September (±37.2 days) and amplitude (mean to peak) of 23 % (Fig. 1). Acute PID is significantly lower around nadir (February–April) than peak (August–October) (9.9/1,00,000 vs. 15.9/1,00,000 p = 0.045), with intermediate values in May–July (13.4/1,00,000) and November–January (11.2/ 1,00,000). The August–October peak is evident only in married (34/158 vs. 14/158 of February–April, p = 0.006), but not in unmarried women (Fig. 2). A relation is present among acute PID and seasonal temperature and photoperiod, best correlations being found with the rhythm of temperature advanced by 2 months (R = 0.901 p \ 0.0001), and the rhythm of photoperiod advanced by 3 months (R = 0.898 p \ 0.0001) (Fig. 3).
Cases
0.6 % Ureaplasma urealiticum
30
20
10
F-A
M-J
A-O
N-J
Months Fig. 2 Distribution of pelvic inflammatory disease in groups of 3 months stratified for marital status. Frequency of pelvic inflammatory disease in married women in the period August–October (A–O) was significantly higher than that in the other months (p \ 0.0001). No seasonal difference was observed in unmarried women (single plus divorced), together and separately considered
Cases of acute PID differed by age (p = 0.0001), being 6.9 % in women younger than 20 years, 24.6 % in women 20–30 years, 43.7 % in women 30–40 years, 24.7 % in women older than 40 years of age.
123
Arch Gynecol Obstet
a higher prevalence is observed in divorced women (p = 0.028) (Table 2).
cases/month
2,8
2,4
Discussion 2
R = 0.901 p= 0.0001
1,6 0
5
10
15
20
25
mean temperature/month
cases/month
2,8
2,4
2
R = 0.989 p = 0.0001
1,6 10
11
12
13
14
15
16
17
light hours /month Fig. 3 Linear regression analyses between seasonal trend of pelvic inflammatory disease and atmospheric conditions using seasonal rhythms of temperature advanced by 2 months and photoperiod advanced by 3 months
Comparison of observed vs. expected acute PID in exposed population, stratified by age and marital status is reported in Table 2. A higher than expected prevalence is observed in unmarried (p = 0.015), particularly in divorced women (p = 0.014), 30–40 years of age. A lower prevalence than expected is observed in married women older than 40 years of age (p = 0.016) (Table 2). Overall,
Our study shows a seasonal trend of PID with peak in September and nadir in March. Among the possible mechanisms, seasonal variation in occasional sexual activity may play a role. This possibility seems to be excluded by the evidence that the September peak is evident in married women, and not in single or divorced women, more frequently exposed to occasional relationship. The seasonal rhythm of PID in married women and in the overall population, may thus originate by a different mechanism, possibly endogenous, linked and influenced by environmental factors such as photoperiod and temperature. Small seasonal changes in host or pathogen factors may be sufficient to create large seasonal surges in infectious disease incidence. Most of cervicities caused by chlamydia and Neisseria gonorrhoeae are asymptomatic [16, 17]. These sub-clinical infections may undergo activation in conditions of reduced immune function possibly induced by variations in photoperiod and temperature [18, 19]. Day-length appears to affect immune function in many species [20]. Prolonged summer ultraviolet exposure enhances free radicals and, by favoring release of immunosuppressive cytokines, and disequilibrium of lymphocytes TH1/TH2’s, decreases systemic and local immune response [21, 22]. Prolonged photoperiod also reduces exposure to melatonin that possesses antioxidant properties and favors cytokines release by T-helper cell type 1 (TH1) [23–26]. Temperature, as well, may favor bacterial growth and physiological responses in the host [27–29].
Table 2 Pelvic inflammatory disease prevalence according to marital status and group of age observed and expected in the exposed population
Expected
Age (years)
Married
\20
Observed Expected
20–30
Observed Expected
30–40
Observed Expected
[40
Oserved
Single
Divorced
Unmarried (divorced ? single)
Total
0.2/431
6.6/13,090
0
6.6/13,090
6.8/13,521
0
11
0
11
11
10.9/21,502
26.3/52,227
0.1/229
26.4/25,456
37.3/73,958
12
27
0
27
39
33.5/66,653
15.9/31,703
1.4/2,923
17.3/34,626
50.8/101,279
33
25
11
36
69
p = 0.014
p = 0.015
48.9/9,538
10.1/20,268
4.1/8,059
14.2/28,327
63.1/123,714
27
6
6
12
39
p = 0.016 Expected Obseved
18–54
93.5/1,83,973 72
p = 0.022 58.9/1,17,288 69
5.6/11,211 17 p = 0.028
123
64.5/1,28,499 86
3,12,472 158
Arch Gynecol Obstet
Besides seasonality, our data report that PID is more frequent in women 30–40 years of age than in adolescents or in older women. This is in contrast with what previously reported in Sweden and other countries in the 700 –800 [30, 31]. Improvements in sexual education in the last decades and differences in sexual behavior, between northern countries and our country may explain these different results. Among women 30–40 years of age a higher than expected prevalence of PID is found only in unmarried and particularly divorced women. Single and divorced status represents a major risk factor for PID, as herein and previously [32, 33] reported, and an increased prevalence of unmarried, particularly divorced, women in older ages, may change PID distribution in the general population. This stresses the importance of promoting the idea of a ‘‘healthy’’ sexual behavior in women of all ages. Present study has several limitations. It is a retrospective evaluation on medical records. Analysis is conducted only on acute PID, requiring hospital admission and on a limited number of risk factors because of incomplete data or their unavailability in the ISTAT database. Among available factors we confirm single or divorced marital status, as a risk factor for PID. The analysis performed on 12,152 medical records, yielded a limited number of cases of PID to be spread in the 12 months. The pattern was, however, rather consistent among the years and the results were significant with different types of analysis. In conclusion this study indicates two different risk factors for PID: one is probably linked to sexual behavior and it causes a higher incidence throughout the year of PID in divorced women. The other is linked to environmental factor and it causes seasonal variation in acute PID mainly in low risk women such as are those married. Conflict of interest
None.
References 1. Dayan L (2006) Pelvic inflammatory disease. Aust Fam Physician 35:858–862 2. Centres for Disease Control and Prevention (2010) Sexually transmitted diseases treatment guidelines, 2010. Morb Mortal Wkly Rep 59:63–67 3. Sweet RL (2012) Pelvic inflammatory disease: current concepts of diagnosis and management. Curr Infect Dis Rep 14:194–203 4. Simms I, Stephenson JM, Mallinson H et al (2006) Risk factors associated with pelvic inflammatory disease. Sex Transm Infect 82:452–457 5. Trigg BG, Kerndt PR, Aynalem G (2008) Sexually transmitted infections and pelvic inflammatory disease in women. Med Clin North Am 92:1083–1113 6. Stone L, Olinky R, Huppert A (2007) Seasonal dynamics of recurrent epidemics. Nature 29:533–536
7. Fisman DN (2007) Seasonality of infectious diseases. Annu Rev Public Health 28:127–143 8. Ross JD, Scott GR (1992) Seasonal variation in gonorrhea. Eur J Epidemiol 8:252–255 9. Shah AP, Smolensky MH, Burau KD, Cech IM, Lai D (2007) Recent change in the annual pattern of sexually transmitted diseases in the United States. Chronobiol Int 24:947–960 10. Albus C, Kohler G (1988) Is there a seasonal incidence of acute adnexitis? Zentralblatt fur Gyneakologie 110:824–826 11. Herold AH, Woodard LJ, Roetzheim RG, Pamies RJ, Young DL, Micceri T (1993) Seasonality of Chlamydia trachomatis genital infection in university women. J Am Coll Health 42:117–120 12. Shroeder B, Tetlow P, Sanfilippo JS, Hertweck SP (2001) Is there a seasonal variation in gonorrhea and Chlamydia in adolescents? J Pediatr Adolesc Gynecol 14:25–27 13. Workoski KA, Berman SA (2006) Sexually transmitted disease treatment guidelines, 2006––centers for disease control and prevention recommendations and reports. MWWR 55:56–61 14. Timor-Tritsch IE, Lerner JP, Montegaudo A, Murphy KE, Heller DS (1998) Transvaginal sonographic markers of tubal inflammatory disease. Ultrasound Obstet Gynecol 12:56–66 15. Van Cauter E (1979) Methods for the characterization of 24-h temporal variations of blood components. Am J Physiol 6:255–264 16. Adams OP, Carter AO, Prussia P, Mc Intyre G, Branch SL (2008) Risk behaviour, healthcare access and prevalence of infection with Chlamydia trachomatis and Neisseria gonorrhoeae in a population-based sample of adults in Barbados. Sex Transm Infect 84:192–194 17. Scharbo-DeHaan M, Anderson DG (2003) The CDC 2002 guidelines for the treatment of sexually transmitted disease: implications for women’s health care. J Midwifery Women’s Health 48:97–98 18. Dowell SF (2001) Seasonal variation in host susceptibility and cycles of certain infectious diseases. Emerg Infect Dis 7:369–374 19. Maes M, Stevens W, Scharpe´ S, Scharpe´ S, Bosmans E, De Meyer F, D’Hondt P et al (1994) Seasonal variation in peripheral blood leukocyte subsets and in serum interleukin-6, and soluble interleukin-2 and -6 receptor in normal volunteers. Experentia 50:821–829 20. Nelson RJ, Demas GE (1996) Seasonal changes in immune function. Q Rev Biol 71:511–548 21. Cooper KD, Fox P, Neises G, Katz SI (1985) Effects of ultraviolet radiation on human epidermal cell alloantigen presentation: initial depression of Langerhans cell-dependent function is followed by the appearance of T cells that enhance epidermal alloantigen presentation. J lmmunol 134:129–137 22. Applegate LA, Lay RD, Alcalay J, Kripke ML (1989) Identification of the molecular target for the suppression of contact hypersensitivity by ultraviolet radiation. J Exp Med 170: 1117–1131 23. Yellon SM, Fagoaga OR, Nehlsen-Cannarella SL (1999) Influence of photoperiod on immune cell functions in the male Siberian hamster. Am J Physiol 276:97–102 24. Nelson RJ, Demas GE, Klein SL, Kriegsfeld LJ (1995) The influence of season, photoperiod, and pineal melatonin on the immune function. J Pineal Res 19:149–165 25. Hotchkiss AK, Nelson RJ (2002) Melatonin and immune function: hype or hypothesis? Crit Rev Immunol 22:351–371 26. Walton JC, Weil ZM, Nelson RJ (2011) Influence of melatonin in hormones, behavior, and immune function. Front Neuroendocrinol 32:303–319 27. Linder JE, Owers KA, Promislow DEL (2008) The effects of temperature on host–pathogen interactions in D. Melanogaster: who benefits? J Ins Physiol 54:297–308
123
Arch Gynecol Obstet 28. Checkley W, Epstein LD, Gilman RH, Figueroa D, Cama RI, Patz JA, Black RE (2000) Effect of El Nin˜o and ambient temperature on hospital admissions for diarrhoeal diseases in Peruvian children. Lancet 355:442–450 29. Brenner I, Shek PN, Zamecnik J, Shephard RJ (1998) Stress hormones and the immunological responses to heat and exercise. Int J Sports Med 19:130–143 30. Westro¨m L (1986) Epidemiology of pelvic inflammatory disease. In: Ludwig H, Thomsen K (eds) Gynecology and obstetrics. Springer, New York, pp 659–661
123
31. Kani J, Adler MW (1992) Epidemiology of pelvic inflammatory disease. In: Berger GS, Westro¨m L (eds) Pelvic inflammatory disease. Raven Press, New York, pp 7–22 32. Ness RB, Smith KJ, Chang CH, Schisterman EF, Bass DC, Gynecologic Infection Follow-Through, GIFT, Investigators (2006) Prediction of pelvic inflammatory disease among young single sexually active women. Sex Transm Dis 33:137–142 33. Buchan H, Vessey M (1989) Epidemiology and trends in hospital discharges for pelvic inflammatory disease in England, 1975 to 1985. Br J Obstet Gyneacol 96:1219–1223
GENERAL GYNECOLOGY
Seasonal trend of acute pelvic inflammatory disease Anjeza Xholli • Marianna Cannoletta Angelo Cagnacci
•
Received: 20 July 2013 / Accepted: 11 November 2013 Ó Springer-Verlag Berlin Heidelberg 2013
Abstract Purpose Many infections follow a seasonal trend. Aim of our study was to check whether acute pelvic inflammatory disease (PID) follows a seasonal progress. Methods In a retrospective study on 12,152 hospital records, 158 cases of acute pelvic inflammatory disease were identified. Periodogram analysis was applied to the date of pelvic inflammatory disease admission and to related environmental factors, such as temperature and photoperiod. Results Pelvic inflammatory disease follows a seasonal rhythm with mean to peak variation of 23 % and maximal values in September (±37.2 days). The rhythm, more evident in married women, is related to the rhythm of temperature advanced by 2 months and of photoperiod advanced by 3 months. Cases of pelvic inflammatory disease are more frequent than expected in unmarried (36 vs. 17.3/34,626, p = 0.015), particularly divorced women 30–40 years of age. Conclusions Our study evidences a seasonal trend and confirms unmarried, particularly divorced status, as important risk factor for acute pelvic inflammatory disease. Keywords Biological rhythms Epidemiology Photoperiod Season Pelvic inflammatory disease
Introduction Pelvic inflammatory disease (PID) is a clinical syndrome defined as a spectrum of upper genital tract infections that
A. Xholli M. Cannoletta A. Cagnacci (&) Obstetrics and Gynaecology Unit, Department of Obstetrics Gynaecology and Paediatric, Azienda Policlinico of Modena, Via del Pozzo 71, 41121 Modena, Italy e-mail: [email protected]
includes any combination of endometritis, salpingitis, pyosalpinx, tubo-ovarian abscess and pelvic peritonitis [1]. It is estimated that in US, over 1 million women are affected of an episode of PID, including 11 % of all fertile women [2]. After a single episode of PID, one out of four women will develop a serious complication, such as recurrent PID, chronic abdominal pain, infertility [3], ectopic pregnancy, fixed uterine retroversus-flexion associated with dyspareunia and dysmenorrhea [2]. Reported risk factors for PID include: young age, single or divorced marital status, low socio-economic status, black race, high frequency of partner change, placement of an intrauterine device, sexual transmitted disease, lack of barrier contraception, bacterial vaginosis, recurrent cervicitis [4, 5]. Beside these risk factors, evidences that pathogen transmission and immune activity can be influenced by changes in ultraviolet exposure, in photoperiod or in temperature, suggest that infections may have a seasonal trend [6]. Indeed, most of the sexually transmitted disease shows seasonal fluctuations with peak incidence in the third quarter of the year [7–9] or in sunny months [10]. Others, carried out in the USA have confirmed seasonal trend in the infection by chlamydia and gonorrhea with infection peaks in August and September [11, 12]. In this study, we evaluated whether acute PID requiring hospitalization does follow a seasonal rhythm, similar to the one reported for chlamydia and gonorrhea infection.
Materials and methods Data on 12,152 women hospitalized in the period 2005–2011 were obtained from the database of the Institute
123
Arch Gynecol Obstet
of Obstetrics and Gynecology in Modena, Italy. Data were anonymously extracted by medical personnel licensed to manage the database. According to the letter of discharge, registered in the hospital database, we selected those women with a diagnosis of PID, namely 195 patients. Every clinical record of these patients was checked. Among these we selected only those patients with a diagnosis of PID that were admitted to the hospital, namely 158 women. Acute PID was defined as a patient with fever, high value of PCR, leukocytosis, leucorrhea, and pelvic pain exacerbated by the gynecological exam, requiring prolonged antibiotics therapy with or without surgery [4, 13]. The diagnosis was supported by an ultrasound investigation of the pelvis showing the typical features of acute PID: pyosalpinx with thick walls of the tube, incomplete septa within it, ovoid/retort shaped, intraluminal fluid, ‘cogwheel sign’. Tubal-ovarian involvement: tubo-ovarian complex with sonographic attributes of salpingitis with ovary and tube discernible but that cannot be pushed apart; tubo-ovarian abscess with extreme motion tenderness at the probe pressure, the ovary and the tube are confluent with speckled fluid [14]. For each single woman the following data were collected from the clinical record: date of admission, date of birth, age, menarche, number of spontaneous deliveries, cesarean sections, ectopic pregnancy, spontaneous abortion, voluntary interruption of pregnancy, last menses, nationality, marital status (married, divorced, single), occupation (professional woman, clerk, worker, housewife), previous abdominal surgery, use of medicines, contraceptive therapy (oral contraceptives or intrauterine device), cervical and vaginal swab specimen results, smoking and treatment during hospitalization (medical, surgical). We assessed the seasonal progress of PID during the year confronting it with that of some atmospheric factors such as the daily average temperature per month in °C and photoperiod, calculated by daily average light hours per month. Data on atmospheric factors in our area for the years 2005–2011 were obtained from the ISTAT (National Institute of Statistics) register. To calculate the prevalence of PID in exposed population we extracted from the ISTAT register the number of women living in Modena in the years 2005–2011 and chose those belonging to the group of age present in our cases (18–54 years old). Marital status was also extracted. Circa-annual rhythmic distribution of PID was evaluated by periodogram analysis using the RHYTHM program. The periodogram was adapted to analyze the 12 months rhythm. As originally described by Van Cauter [15], the
123
method initially tests the significance of the observed time fluctuations against the hypothesis of their purely random occurrence via two different tests. The alternative to pure randomness is, for the first test, the existence of local correlations, implying that values at given times depend on values at other times, and for the second test, the existence of periodic fluctuations. When the hypothesis of random occurrence of the data is rejected, the periodogram method is applied to detect and estimate the possible significant components. A sum of sinusoidal components with periods equal to integer divisors of the observation span (i.e. 12 months/1; 12 months/2; 12 months/3; etc.) is fitted on the series. A decision procedure devised by Fisher allows the selection of the significant periodicity underlying the process at a given probability of p \ 0.05. Up to the first three significant periodical components are included in the theoretical description of the profile. A theoretical pattern is computed according to the minimum of the residual sum of squares. The amplitude of the theoretical pattern is defined as half the difference between the maximum and minimum of the theoretical pattern, while the acrophase is the time corresponding to the occurrence of the first global maximum of the theoretical curve. Both the amplitude and the acrophase are furnished with a confidence interval. Periodograms are considered significantly different when described by different periodical components or alternatively when the confidence intervals of either the amplitude or acrophase do not overlap. Statistical analyses were performed by the statistical Package Statview 5.0.1 for Apple Machintosh (SAS Institute Inc, Cary, NC, USA1998). The seasonal progress of PID in the 7 years of the study is expressed as monthly average and related, by regression analysis, to the seasonal course of monthly average temperature and photoperiod evaluated in the same years. Rhythms of temperature and photoperiod were shifted up to 3 months both in advance and in delay to highlight possible phase shift between PID and atmospheric factors. Prevalence of PID in different months of the year and in different group of women was compared by contingency table associated with the Chi-square test. A value of p \ 0.05 was considered as significant.
Results Characteristics of women with acute PID are reported in Table 1. Periodogram analysis showed that the average monthly distribution of acute PID takes a seasonal progress illustrated by a sinusoidal curve that stretches for a period of 12 months with a mean value per month of 1.88, peak in
Arch Gynecol Obstet Table 1 Characteristics of women with pelvic inflammatory disease hospitalized in the period 2005–2011 N = 158
Age
18–54 years
Mean age
35.4 years
Mean age of menarche
12.6 years
Nationality
European 84.8 %
cases/month
Women with PID
2.2
1.8 1.4
Asiatic 2.5 %
TEMPERATURE
African 11.4 %
20
South-American 1.3 % Married 45.6 %
15
ºC
Marital status
PID
2.6
Single 43.7 %
10
Divorced 10.7 % Occupation
Professional woman 16.1 %
5
Clerk 36 % PHOTOPERIOD
Worker 26.7 % Unemployed 6.6 % Parity
Nullipara 41.6 % Multipara 56.6 %
Previous surgery
light hours
Housewife 14.6 %
16
14
12
Appendix 30.4 % Cesarean section 10 %
10
Adnexal surgery 8 %
J
F M A M
Abortion 26.0 % Contraception
Abdominal surgery 2.5 % IUD 6.9 %
J
J
A S O N
D
Months Fig. 1 Seasonal trend of monthly average of pelvic inflammatory disease, temperature and photoperiod
Oral Contraceptive 5 % Smoking
7.6 %
Cervical–vaginal swab
31.6 % Chlamydia trachomatis
60
50
3.7 % Neisseria gonorrhoeae
Divorced Single Married
1.2 % Mycoplasma genitalium 40
Inpatient therapy
Antibiotics 51.9 % Surgery 48.1 %
September (±37.2 days) and amplitude (mean to peak) of 23 % (Fig. 1). Acute PID is significantly lower around nadir (February–April) than peak (August–October) (9.9/1,00,000 vs. 15.9/1,00,000 p = 0.045), with intermediate values in May–July (13.4/1,00,000) and November–January (11.2/ 1,00,000). The August–October peak is evident only in married (34/158 vs. 14/158 of February–April, p = 0.006), but not in unmarried women (Fig. 2). A relation is present among acute PID and seasonal temperature and photoperiod, best correlations being found with the rhythm of temperature advanced by 2 months (R = 0.901 p \ 0.0001), and the rhythm of photoperiod advanced by 3 months (R = 0.898 p \ 0.0001) (Fig. 3).
Cases
0.6 % Ureaplasma urealiticum
30
20
10
F-A
M-J
A-O
N-J
Months Fig. 2 Distribution of pelvic inflammatory disease in groups of 3 months stratified for marital status. Frequency of pelvic inflammatory disease in married women in the period August–October (A–O) was significantly higher than that in the other months (p \ 0.0001). No seasonal difference was observed in unmarried women (single plus divorced), together and separately considered
Cases of acute PID differed by age (p = 0.0001), being 6.9 % in women younger than 20 years, 24.6 % in women 20–30 years, 43.7 % in women 30–40 years, 24.7 % in women older than 40 years of age.
123
Arch Gynecol Obstet
a higher prevalence is observed in divorced women (p = 0.028) (Table 2).
cases/month
2,8
2,4
Discussion 2
R = 0.901 p= 0.0001
1,6 0
5
10
15
20
25
mean temperature/month
cases/month
2,8
2,4
2
R = 0.989 p = 0.0001
1,6 10
11
12
13
14
15
16
17
light hours /month Fig. 3 Linear regression analyses between seasonal trend of pelvic inflammatory disease and atmospheric conditions using seasonal rhythms of temperature advanced by 2 months and photoperiod advanced by 3 months
Comparison of observed vs. expected acute PID in exposed population, stratified by age and marital status is reported in Table 2. A higher than expected prevalence is observed in unmarried (p = 0.015), particularly in divorced women (p = 0.014), 30–40 years of age. A lower prevalence than expected is observed in married women older than 40 years of age (p = 0.016) (Table 2). Overall,
Our study shows a seasonal trend of PID with peak in September and nadir in March. Among the possible mechanisms, seasonal variation in occasional sexual activity may play a role. This possibility seems to be excluded by the evidence that the September peak is evident in married women, and not in single or divorced women, more frequently exposed to occasional relationship. The seasonal rhythm of PID in married women and in the overall population, may thus originate by a different mechanism, possibly endogenous, linked and influenced by environmental factors such as photoperiod and temperature. Small seasonal changes in host or pathogen factors may be sufficient to create large seasonal surges in infectious disease incidence. Most of cervicities caused by chlamydia and Neisseria gonorrhoeae are asymptomatic [16, 17]. These sub-clinical infections may undergo activation in conditions of reduced immune function possibly induced by variations in photoperiod and temperature [18, 19]. Day-length appears to affect immune function in many species [20]. Prolonged summer ultraviolet exposure enhances free radicals and, by favoring release of immunosuppressive cytokines, and disequilibrium of lymphocytes TH1/TH2’s, decreases systemic and local immune response [21, 22]. Prolonged photoperiod also reduces exposure to melatonin that possesses antioxidant properties and favors cytokines release by T-helper cell type 1 (TH1) [23–26]. Temperature, as well, may favor bacterial growth and physiological responses in the host [27–29].
Table 2 Pelvic inflammatory disease prevalence according to marital status and group of age observed and expected in the exposed population
Expected
Age (years)
Married
\20
Observed Expected
20–30
Observed Expected
30–40
Observed Expected
[40
Oserved
Single
Divorced
Unmarried (divorced ? single)
Total
0.2/431
6.6/13,090
0
6.6/13,090
6.8/13,521
0
11
0
11
11
10.9/21,502
26.3/52,227
0.1/229
26.4/25,456
37.3/73,958
12
27
0
27
39
33.5/66,653
15.9/31,703
1.4/2,923
17.3/34,626
50.8/101,279
33
25
11
36
69
p = 0.014
p = 0.015
48.9/9,538
10.1/20,268
4.1/8,059
14.2/28,327
63.1/123,714
27
6
6
12
39
p = 0.016 Expected Obseved
18–54
93.5/1,83,973 72
p = 0.022 58.9/1,17,288 69
5.6/11,211 17 p = 0.028
123
64.5/1,28,499 86
3,12,472 158
Arch Gynecol Obstet
Besides seasonality, our data report that PID is more frequent in women 30–40 years of age than in adolescents or in older women. This is in contrast with what previously reported in Sweden and other countries in the 700 –800 [30, 31]. Improvements in sexual education in the last decades and differences in sexual behavior, between northern countries and our country may explain these different results. Among women 30–40 years of age a higher than expected prevalence of PID is found only in unmarried and particularly divorced women. Single and divorced status represents a major risk factor for PID, as herein and previously [32, 33] reported, and an increased prevalence of unmarried, particularly divorced, women in older ages, may change PID distribution in the general population. This stresses the importance of promoting the idea of a ‘‘healthy’’ sexual behavior in women of all ages. Present study has several limitations. It is a retrospective evaluation on medical records. Analysis is conducted only on acute PID, requiring hospital admission and on a limited number of risk factors because of incomplete data or their unavailability in the ISTAT database. Among available factors we confirm single or divorced marital status, as a risk factor for PID. The analysis performed on 12,152 medical records, yielded a limited number of cases of PID to be spread in the 12 months. The pattern was, however, rather consistent among the years and the results were significant with different types of analysis. In conclusion this study indicates two different risk factors for PID: one is probably linked to sexual behavior and it causes a higher incidence throughout the year of PID in divorced women. The other is linked to environmental factor and it causes seasonal variation in acute PID mainly in low risk women such as are those married. Conflict of interest
None.
References 1. Dayan L (2006) Pelvic inflammatory disease. Aust Fam Physician 35:858–862 2. Centres for Disease Control and Prevention (2010) Sexually transmitted diseases treatment guidelines, 2010. Morb Mortal Wkly Rep 59:63–67 3. Sweet RL (2012) Pelvic inflammatory disease: current concepts of diagnosis and management. Curr Infect Dis Rep 14:194–203 4. Simms I, Stephenson JM, Mallinson H et al (2006) Risk factors associated with pelvic inflammatory disease. Sex Transm Infect 82:452–457 5. Trigg BG, Kerndt PR, Aynalem G (2008) Sexually transmitted infections and pelvic inflammatory disease in women. Med Clin North Am 92:1083–1113 6. Stone L, Olinky R, Huppert A (2007) Seasonal dynamics of recurrent epidemics. Nature 29:533–536
7. Fisman DN (2007) Seasonality of infectious diseases. Annu Rev Public Health 28:127–143 8. Ross JD, Scott GR (1992) Seasonal variation in gonorrhea. Eur J Epidemiol 8:252–255 9. Shah AP, Smolensky MH, Burau KD, Cech IM, Lai D (2007) Recent change in the annual pattern of sexually transmitted diseases in the United States. Chronobiol Int 24:947–960 10. Albus C, Kohler G (1988) Is there a seasonal incidence of acute adnexitis? Zentralblatt fur Gyneakologie 110:824–826 11. Herold AH, Woodard LJ, Roetzheim RG, Pamies RJ, Young DL, Micceri T (1993) Seasonality of Chlamydia trachomatis genital infection in university women. J Am Coll Health 42:117–120 12. Shroeder B, Tetlow P, Sanfilippo JS, Hertweck SP (2001) Is there a seasonal variation in gonorrhea and Chlamydia in adolescents? J Pediatr Adolesc Gynecol 14:25–27 13. Workoski KA, Berman SA (2006) Sexually transmitted disease treatment guidelines, 2006––centers for disease control and prevention recommendations and reports. MWWR 55:56–61 14. Timor-Tritsch IE, Lerner JP, Montegaudo A, Murphy KE, Heller DS (1998) Transvaginal sonographic markers of tubal inflammatory disease. Ultrasound Obstet Gynecol 12:56–66 15. Van Cauter E (1979) Methods for the characterization of 24-h temporal variations of blood components. Am J Physiol 6:255–264 16. Adams OP, Carter AO, Prussia P, Mc Intyre G, Branch SL (2008) Risk behaviour, healthcare access and prevalence of infection with Chlamydia trachomatis and Neisseria gonorrhoeae in a population-based sample of adults in Barbados. Sex Transm Infect 84:192–194 17. Scharbo-DeHaan M, Anderson DG (2003) The CDC 2002 guidelines for the treatment of sexually transmitted disease: implications for women’s health care. J Midwifery Women’s Health 48:97–98 18. Dowell SF (2001) Seasonal variation in host susceptibility and cycles of certain infectious diseases. Emerg Infect Dis 7:369–374 19. Maes M, Stevens W, Scharpe´ S, Scharpe´ S, Bosmans E, De Meyer F, D’Hondt P et al (1994) Seasonal variation in peripheral blood leukocyte subsets and in serum interleukin-6, and soluble interleukin-2 and -6 receptor in normal volunteers. Experentia 50:821–829 20. Nelson RJ, Demas GE (1996) Seasonal changes in immune function. Q Rev Biol 71:511–548 21. Cooper KD, Fox P, Neises G, Katz SI (1985) Effects of ultraviolet radiation on human epidermal cell alloantigen presentation: initial depression of Langerhans cell-dependent function is followed by the appearance of T cells that enhance epidermal alloantigen presentation. J lmmunol 134:129–137 22. Applegate LA, Lay RD, Alcalay J, Kripke ML (1989) Identification of the molecular target for the suppression of contact hypersensitivity by ultraviolet radiation. J Exp Med 170: 1117–1131 23. Yellon SM, Fagoaga OR, Nehlsen-Cannarella SL (1999) Influence of photoperiod on immune cell functions in the male Siberian hamster. Am J Physiol 276:97–102 24. Nelson RJ, Demas GE, Klein SL, Kriegsfeld LJ (1995) The influence of season, photoperiod, and pineal melatonin on the immune function. J Pineal Res 19:149–165 25. Hotchkiss AK, Nelson RJ (2002) Melatonin and immune function: hype or hypothesis? Crit Rev Immunol 22:351–371 26. Walton JC, Weil ZM, Nelson RJ (2011) Influence of melatonin in hormones, behavior, and immune function. Front Neuroendocrinol 32:303–319 27. Linder JE, Owers KA, Promislow DEL (2008) The effects of temperature on host–pathogen interactions in D. Melanogaster: who benefits? J Ins Physiol 54:297–308
123
Arch Gynecol Obstet 28. Checkley W, Epstein LD, Gilman RH, Figueroa D, Cama RI, Patz JA, Black RE (2000) Effect of El Nin˜o and ambient temperature on hospital admissions for diarrhoeal diseases in Peruvian children. Lancet 355:442–450 29. Brenner I, Shek PN, Zamecnik J, Shephard RJ (1998) Stress hormones and the immunological responses to heat and exercise. Int J Sports Med 19:130–143 30. Westro¨m L (1986) Epidemiology of pelvic inflammatory disease. In: Ludwig H, Thomsen K (eds) Gynecology and obstetrics. Springer, New York, pp 659–661
123
31. Kani J, Adler MW (1992) Epidemiology of pelvic inflammatory disease. In: Berger GS, Westro¨m L (eds) Pelvic inflammatory disease. Raven Press, New York, pp 7–22 32. Ness RB, Smith KJ, Chang CH, Schisterman EF, Bass DC, Gynecologic Infection Follow-Through, GIFT, Investigators (2006) Prediction of pelvic inflammatory disease among young single sexually active women. Sex Transm Dis 33:137–142 33. Buchan H, Vessey M (1989) Epidemiology and trends in hospital discharges for pelvic inflammatory disease in England, 1975 to 1985. Br J Obstet Gyneacol 96:1219–1223
